* [PATCH 01/45] drm/msm/dp: dont call dp_catalog_ctrl_mainlink_ctrl in dp_ctrl_configure_source_params()
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 17:50 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 02/45] drm/msm/dp: disable the opp table request even for dp_ctrl_off_link() Abhinav Kumar
` (45 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Once the link has already been setup there is no need to call
dp_catalog_ctrl_mainlink_ctrl() as this does a reset on the mainlink
thereby tearing down the link briefly.
Fixes: c943b4948b58 ("drm/msm/dp: add displayPort driver support")
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index bc2ca8133b790fc049e18ab3b37a629558664dd4..a8069f7c4773f087229727dc38c9af0d6d84f863 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -178,7 +178,6 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
u32 cc, tb;
msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
- msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, true);
msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
msm_dp_ctrl_config_ctrl(ctrl);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 01/45] drm/msm/dp: dont call dp_catalog_ctrl_mainlink_ctrl in dp_ctrl_configure_source_params()
2024-12-06 4:31 ` [PATCH 01/45] drm/msm/dp: dont call dp_catalog_ctrl_mainlink_ctrl in dp_ctrl_configure_source_params() Abhinav Kumar
@ 2024-12-08 17:50 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 17:50 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:32PM -0800, Abhinav Kumar wrote:
> Once the link has already been setup there is no need to call
> dp_catalog_ctrl_mainlink_ctrl() as this does a reset on the mainlink
> thereby tearing down the link briefly.
>
> Fixes: c943b4948b58 ("drm/msm/dp: add displayPort driver support")
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 1 -
> 1 file changed, 1 deletion(-)
>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 02/45] drm/msm/dp: disable the opp table request even for dp_ctrl_off_link()
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
2024-12-06 4:31 ` [PATCH 01/45] drm/msm/dp: dont call dp_catalog_ctrl_mainlink_ctrl in dp_ctrl_configure_source_params() Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 8:38 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces Abhinav Kumar
` (44 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
dp_ctrl_off_link() was created to handle a case where we received
a cable connect and then get a cable disconnect without the corresponding
dp_display_enable(). For such cases the pixel clock will be off but the
link clock will still be on. dp_ctrl_off_link() handles this case by
turning off the link clock only.
However, the vote removal to the opp table for this case was missed.
Remove the opp table vote in dp_ctrl_off_link().
Fixes: 375a126090b9 ("drm/msm/dp: tear down main link at unplug handle immediately")
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index a8069f7c4773f087229727dc38c9af0d6d84f863..9c463ae2f8fae916661fef1c7e225f55c1026478 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -2070,6 +2070,7 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
+ dev_pm_opp_set_rate(ctrl->dev, 0);
msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
DRM_DEBUG_DP("Before, phy=%p init_count=%d power_on=%d\n",
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 02/45] drm/msm/dp: disable the opp table request even for dp_ctrl_off_link()
2024-12-06 4:31 ` [PATCH 02/45] drm/msm/dp: disable the opp table request even for dp_ctrl_off_link() Abhinav Kumar
@ 2024-12-06 8:38 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 8:38 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:33PM -0800, Abhinav Kumar wrote:
> dp_ctrl_off_link() was created to handle a case where we received
> a cable connect and then get a cable disconnect without the corresponding
> dp_display_enable(). For such cases the pixel clock will be off but the
> link clock will still be on. dp_ctrl_off_link() handles this case by
> turning off the link clock only.
>
> However, the vote removal to the opp table for this case was missed.
> Remove the opp table vote in dp_ctrl_off_link().
>
> Fixes: 375a126090b9 ("drm/msm/dp: tear down main link at unplug handle immediately")
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 1 +
> 1 file changed, 1 insertion(+)
>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
2024-12-06 4:31 ` [PATCH 01/45] drm/msm/dp: dont call dp_catalog_ctrl_mainlink_ctrl in dp_ctrl_configure_source_params() Abhinav Kumar
2024-12-06 4:31 ` [PATCH 02/45] drm/msm/dp: disable the opp table request even for dp_ctrl_off_link() Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 8:41 ` Dmitry Baryshkov
2025-01-07 0:55 ` Bjorn Andersson
2024-12-06 4:31 ` [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts Abhinav Kumar
` (43 subsequent siblings)
46 siblings, 2 replies; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Interface type of MST interfaces is currently INTF_NONE.
Fix this to INTF_DP.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
index 907b4d7ceb470b0391d2bbbab3ce520efa2b3263..2509e28e3d6b582cd837c6aea167b3f4ad877383 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
+++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
@@ -375,7 +375,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
.name = "intf_3", .id = INTF_3,
.base = 0x37000, .len = 0x280,
.features = INTF_SC7280_MASK,
- .type = INTF_NONE,
+ .type = INTF_DP,
.controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
.prog_fetch_lines_worst_case = 24,
.intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 30),
@@ -393,7 +393,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
.name = "intf_6", .id = INTF_6,
.base = 0x3A000, .len = 0x280,
.features = INTF_SC7280_MASK,
- .type = INTF_NONE,
+ .type = INTF_DP,
.controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
.prog_fetch_lines_worst_case = 24,
.intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 17),
@@ -402,7 +402,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
.name = "intf_7", .id = INTF_7,
.base = 0x3b000, .len = 0x280,
.features = INTF_SC7280_MASK,
- .type = INTF_NONE,
+ .type = INTF_DP,
.controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
.prog_fetch_lines_worst_case = 24,
.intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 18),
@@ -411,7 +411,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
.name = "intf_8", .id = INTF_8,
.base = 0x3c000, .len = 0x280,
.features = INTF_SC7280_MASK,
- .type = INTF_NONE,
+ .type = INTF_DP,
.controller_id = MSM_DP_CONTROLLER_1, /* pair with intf_4 for DP MST */
.prog_fetch_lines_worst_case = 24,
.intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12),
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces
2024-12-06 4:31 ` [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces Abhinav Kumar
@ 2024-12-06 8:41 ` Dmitry Baryshkov
2025-01-07 0:55 ` Bjorn Andersson
1 sibling, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 8:41 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:34PM -0800, Abhinav Kumar wrote:
> Interface type of MST interfaces is currently INTF_NONE.
> Fix this to INTF_DP.
Neither subject nor commit message tell that this is limited to sa8775p
only. Please consider toggling all catalog files which are supposed to
handle MST.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
> index 907b4d7ceb470b0391d2bbbab3ce520efa2b3263..2509e28e3d6b582cd837c6aea167b3f4ad877383 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
> +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
> @@ -375,7 +375,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_3", .id = INTF_3,
> .base = 0x37000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 30),
> @@ -393,7 +393,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_6", .id = INTF_6,
> .base = 0x3A000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 17),
> @@ -402,7 +402,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_7", .id = INTF_7,
> .base = 0x3b000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 18),
> @@ -411,7 +411,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_8", .id = INTF_8,
> .base = 0x3c000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_1, /* pair with intf_4 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12),
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces
2024-12-06 4:31 ` [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces Abhinav Kumar
2024-12-06 8:41 ` Dmitry Baryshkov
@ 2025-01-07 0:55 ` Bjorn Andersson
2025-05-29 10:06 ` Yongxing Mou
1 sibling, 1 reply; 111+ messages in thread
From: Bjorn Andersson @ 2025-01-07 0:55 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:34PM -0800, Abhinav Kumar wrote:
> Interface type of MST interfaces is currently INTF_NONE.
> Fix this to INTF_DP.
>
Wouldn't it make sense to introduce this later in the series, once the
implementation would actually handle this case? Or could/should we have
left these INTF_DP from the start?
Regards,
Bjorn
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
> index 907b4d7ceb470b0391d2bbbab3ce520efa2b3263..2509e28e3d6b582cd837c6aea167b3f4ad877383 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
> +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
> @@ -375,7 +375,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_3", .id = INTF_3,
> .base = 0x37000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 30),
> @@ -393,7 +393,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_6", .id = INTF_6,
> .base = 0x3A000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 17),
> @@ -402,7 +402,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_7", .id = INTF_7,
> .base = 0x3b000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 18),
> @@ -411,7 +411,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
> .name = "intf_8", .id = INTF_8,
> .base = 0x3c000, .len = 0x280,
> .features = INTF_SC7280_MASK,
> - .type = INTF_NONE,
> + .type = INTF_DP,
> .controller_id = MSM_DP_CONTROLLER_1, /* pair with intf_4 for DP MST */
> .prog_fetch_lines_worst_case = 24,
> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12),
>
> --
> 2.34.1
>
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces
2025-01-07 0:55 ` Bjorn Andersson
@ 2025-05-29 10:06 ` Yongxing Mou
0 siblings, 0 replies; 111+ messages in thread
From: Yongxing Mou @ 2025-05-29 10:06 UTC (permalink / raw)
To: Bjorn Andersson, Abhinav Kumar
Cc: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On 2025/1/7 8:55, Bjorn Andersson wrote:
> On Thu, Dec 05, 2024 at 08:31:34PM -0800, Abhinav Kumar wrote:
>> Interface type of MST interfaces is currently INTF_NONE.
>> Fix this to INTF_DP.
>>
>
> Wouldn't it make sense to introduce this later in the series, once the
> implementation would actually handle this case? Or could/should we have
> left these INTF_DP from the start?
>
> Regards,
> Bjorn
>
Yes, it's better to place this patch at the end of the series, once the
platform is already capable of handling MST.
>> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
>> ---
>> drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h | 8 ++++----
>> 1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
>> index 907b4d7ceb470b0391d2bbbab3ce520efa2b3263..2509e28e3d6b582cd837c6aea167b3f4ad877383 100644
>> --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
>> +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h
>> @@ -375,7 +375,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
>> .name = "intf_3", .id = INTF_3,
>> .base = 0x37000, .len = 0x280,
>> .features = INTF_SC7280_MASK,
>> - .type = INTF_NONE,
>> + .type = INTF_DP,
>> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
>> .prog_fetch_lines_worst_case = 24,
>> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 30),
>> @@ -393,7 +393,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
>> .name = "intf_6", .id = INTF_6,
>> .base = 0x3A000, .len = 0x280,
>> .features = INTF_SC7280_MASK,
>> - .type = INTF_NONE,
>> + .type = INTF_DP,
>> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
>> .prog_fetch_lines_worst_case = 24,
>> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 17),
>> @@ -402,7 +402,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
>> .name = "intf_7", .id = INTF_7,
>> .base = 0x3b000, .len = 0x280,
>> .features = INTF_SC7280_MASK,
>> - .type = INTF_NONE,
>> + .type = INTF_DP,
>> .controller_id = MSM_DP_CONTROLLER_0, /* pair with intf_0 for DP MST */
>> .prog_fetch_lines_worst_case = 24,
>> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 18),
>> @@ -411,7 +411,7 @@ static const struct dpu_intf_cfg sa8775p_intf[] = {
>> .name = "intf_8", .id = INTF_8,
>> .base = 0x3c000, .len = 0x280,
>> .features = INTF_SC7280_MASK,
>> - .type = INTF_NONE,
>> + .type = INTF_DP,
>> .controller_id = MSM_DP_CONTROLLER_1, /* pair with intf_4 for DP MST */
>> .prog_fetch_lines_worst_case = 24,
>> .intr_underrun = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 12),
>>
>> --
>> 2.34.1
>>
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (2 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 03/45] drm/msm/dp: fix the intf_type of MST interfaces Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 8:51 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 05/45] drm/msm/dp: add a helper to read mst caps for dp_panel Abhinav Kumar
` (42 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
In preparation of DP MST where link caps are read for the
immediate downstream device and the edid is read through
sideband messaging, split the msm_dp_panel_read_sink_caps() into
two parts which read the link parameters and the edid parts
respectively.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 6 +++++-
drivers/gpu/drm/msm/dp/dp_panel.c | 25 +++++++++++++++++--------
drivers/gpu/drm/msm/dp/dp_panel.h | 5 ++++-
3 files changed, 26 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index be26064af9febf4f4761e21ea7db85ab1ac66081..052db80c6a365f53c2c0a37d3b69ea2b627aea1f 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -372,7 +372,11 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
const struct drm_display_info *info = &connector->display_info;
int rc = 0;
- rc = msm_dp_panel_read_sink_caps(dp->panel, connector);
+ rc = msm_dp_panel_read_link_caps(dp->panel, connector);
+ if (rc)
+ goto end;
+
+ rc = msm_dp_panel_read_edid(dp->panel, connector);
if (rc)
goto end;
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
index 5d7eaa31bf3176566f40f01ff636bee64e81c64f..d277e9b2cbc03688976b6aa481ee724b186bab51 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.c
+++ b/drivers/gpu/drm/msm/dp/dp_panel.c
@@ -108,8 +108,8 @@ static u32 msm_dp_panel_get_supported_bpp(struct msm_dp_panel *msm_dp_panel,
return min_supported_bpp;
}
-int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
- struct drm_connector *connector)
+int msm_dp_panel_read_link_caps(struct msm_dp_panel *msm_dp_panel,
+ struct drm_connector *connector)
{
int rc, bw_code;
int count;
@@ -150,8 +150,19 @@ int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
rc = drm_dp_read_downstream_info(panel->aux, msm_dp_panel->dpcd,
msm_dp_panel->downstream_ports);
- if (rc)
- return rc;
+ return rc;
+}
+
+int msm_dp_panel_read_edid(struct msm_dp_panel *msm_dp_panel, struct drm_connector *connector)
+{
+ struct msm_dp_panel_private *panel;
+
+ if (!msm_dp_panel || !connector) {
+ DRM_ERROR("invalid input\n");
+ return -EINVAL;
+ }
+
+ panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
drm_edid_free(msm_dp_panel->drm_edid);
@@ -163,13 +174,11 @@ int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
DRM_ERROR("panel edid read failed\n");
/* check edid read fail is due to unplug */
if (!msm_dp_catalog_link_is_connected(panel->catalog)) {
- rc = -ETIMEDOUT;
- goto end;
+ return -ETIMEDOUT;
}
}
-end:
- return rc;
+ return 0;
}
u32 msm_dp_panel_get_mode_bpp(struct msm_dp_panel *msm_dp_panel,
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
index 0e944db3adf2f187f313664fe80cf540ec7a19f2..7a38655c443af597c84fb78c6702b2a3ef9822ed 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.h
+++ b/drivers/gpu/drm/msm/dp/dp_panel.h
@@ -59,7 +59,10 @@ void msm_dp_panel_dump_regs(struct msm_dp_panel *msm_dp_panel);
int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
struct drm_connector *connector);
u32 msm_dp_panel_get_mode_bpp(struct msm_dp_panel *msm_dp_panel, u32 mode_max_bpp,
- u32 mode_pclk_khz);
+ u32 mode_pclk_khz);
+int msm_dp_panel_read_link_caps(struct msm_dp_panel *dp_panel,
+ struct drm_connector *connector);
+int msm_dp_panel_read_edid(struct msm_dp_panel *dp_panel, struct drm_connector *connector);
int msm_dp_panel_get_modes(struct msm_dp_panel *msm_dp_panel,
struct drm_connector *connector);
void msm_dp_panel_handle_sink_request(struct msm_dp_panel *msm_dp_panel);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts
2024-12-06 4:31 ` [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts Abhinav Kumar
@ 2024-12-06 8:51 ` Dmitry Baryshkov
2025-05-22 9:51 ` Yongxing Mou
0 siblings, 1 reply; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 8:51 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:35PM -0800, Abhinav Kumar wrote:
> In preparation of DP MST where link caps are read for the
> immediate downstream device and the edid is read through
> sideband messaging, split the msm_dp_panel_read_sink_caps() into
> two parts which read the link parameters and the edid parts
> respectively.
As you are touching this part, could you please refactor the code
instead by dropping the msm_dp_panel->drm_edid? There should be no need
to store EDID in the panel structure.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 6 +++++-
> drivers/gpu/drm/msm/dp/dp_panel.c | 25 +++++++++++++++++--------
> drivers/gpu/drm/msm/dp/dp_panel.h | 5 ++++-
> 3 files changed, 26 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index be26064af9febf4f4761e21ea7db85ab1ac66081..052db80c6a365f53c2c0a37d3b69ea2b627aea1f 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -372,7 +372,11 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> const struct drm_display_info *info = &connector->display_info;
> int rc = 0;
>
> - rc = msm_dp_panel_read_sink_caps(dp->panel, connector);
> + rc = msm_dp_panel_read_link_caps(dp->panel, connector);
> + if (rc)
> + goto end;
> +
> + rc = msm_dp_panel_read_edid(dp->panel, connector);
> if (rc)
> goto end;
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
> index 5d7eaa31bf3176566f40f01ff636bee64e81c64f..d277e9b2cbc03688976b6aa481ee724b186bab51 100644
> --- a/drivers/gpu/drm/msm/dp/dp_panel.c
> +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
> @@ -108,8 +108,8 @@ static u32 msm_dp_panel_get_supported_bpp(struct msm_dp_panel *msm_dp_panel,
> return min_supported_bpp;
> }
>
> -int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
> - struct drm_connector *connector)
> +int msm_dp_panel_read_link_caps(struct msm_dp_panel *msm_dp_panel,
> + struct drm_connector *connector)
This function doesn't require connector anymore.
> {
> int rc, bw_code;
> int count;
> @@ -150,8 +150,19 @@ int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
>
> rc = drm_dp_read_downstream_info(panel->aux, msm_dp_panel->dpcd,
> msm_dp_panel->downstream_ports);
> - if (rc)
> - return rc;
> + return rc;
> +}
> +
> +int msm_dp_panel_read_edid(struct msm_dp_panel *msm_dp_panel, struct drm_connector *connector)
> +{
> + struct msm_dp_panel_private *panel;
> +
> + if (!msm_dp_panel || !connector) {
> + DRM_ERROR("invalid input\n");
> + return -EINVAL;
> + }
Neither panel nor connector can be NULL here, please drop.
> +
> + panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
>
> drm_edid_free(msm_dp_panel->drm_edid);
>
> @@ -163,13 +174,11 @@ int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
> DRM_ERROR("panel edid read failed\n");
> /* check edid read fail is due to unplug */
> if (!msm_dp_catalog_link_is_connected(panel->catalog)) {
> - rc = -ETIMEDOUT;
> - goto end;
> + return -ETIMEDOUT;
> }
> }
>
> -end:
> - return rc;
> + return 0;
> }
>
> u32 msm_dp_panel_get_mode_bpp(struct msm_dp_panel *msm_dp_panel,
> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
> index 0e944db3adf2f187f313664fe80cf540ec7a19f2..7a38655c443af597c84fb78c6702b2a3ef9822ed 100644
> --- a/drivers/gpu/drm/msm/dp/dp_panel.h
> +++ b/drivers/gpu/drm/msm/dp/dp_panel.h
> @@ -59,7 +59,10 @@ void msm_dp_panel_dump_regs(struct msm_dp_panel *msm_dp_panel);
> int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
> struct drm_connector *connector);
> u32 msm_dp_panel_get_mode_bpp(struct msm_dp_panel *msm_dp_panel, u32 mode_max_bpp,
> - u32 mode_pclk_khz);
> + u32 mode_pclk_khz);
> +int msm_dp_panel_read_link_caps(struct msm_dp_panel *dp_panel,
> + struct drm_connector *connector);
> +int msm_dp_panel_read_edid(struct msm_dp_panel *dp_panel, struct drm_connector *connector);
> int msm_dp_panel_get_modes(struct msm_dp_panel *msm_dp_panel,
> struct drm_connector *connector);
> void msm_dp_panel_handle_sink_request(struct msm_dp_panel *msm_dp_panel);
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts
2024-12-06 8:51 ` Dmitry Baryshkov
@ 2025-05-22 9:51 ` Yongxing Mou
2025-05-22 12:38 ` Dmitry Baryshkov
0 siblings, 1 reply; 111+ messages in thread
From: Yongxing Mou @ 2025-05-22 9:51 UTC (permalink / raw)
To: Dmitry Baryshkov, Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On 2024/12/6 16:51, Dmitry Baryshkov wrote:
> On Thu, Dec 05, 2024 at 08:31:35PM -0800, Abhinav Kumar wrote:
>> In preparation of DP MST where link caps are read for the
>> immediate downstream device and the edid is read through
>> sideband messaging, split the msm_dp_panel_read_sink_caps() into
>> two parts which read the link parameters and the edid parts
>> respectively.
>
> As you are touching this part, could you please refactor the code
> instead by dropping the msm_dp_panel->drm_edid? There should be no need
> to store EDID in the panel structure.
>
Hi, Dmitry, Abhinav will be leaving the company and will no longer be
responsible for updating and address the comments. I will take over
handling MST patch series. Regarding this comments, I don't got that
where the drm_edid should be stored. In MST cases, where multiple panels
exist, i think that there should be a separate drm_edid saved for each
panel.
>>
>> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
>> ---
>> drivers/gpu/drm/msm/dp/dp_display.c | 6 +++++-
>> drivers/gpu/drm/msm/dp/dp_panel.c | 25 +++++++++++++++++--------
>> drivers/gpu/drm/msm/dp/dp_panel.h | 5 ++++-
>> 3 files changed, 26 insertions(+), 10 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
>> index be26064af9febf4f4761e21ea7db85ab1ac66081..052db80c6a365f53c2c0a37d3b69ea2b627aea1f 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_display.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
>> @@ -372,7 +372,11 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
>> const struct drm_display_info *info = &connector->display_info;
>> int rc = 0;
>>
>> - rc = msm_dp_panel_read_sink_caps(dp->panel, connector);
>> + rc = msm_dp_panel_read_link_caps(dp->panel, connector);
>> + if (rc)
>> + goto end;
>> +
>> + rc = msm_dp_panel_read_edid(dp->panel, connector);
>> if (rc)
>> goto end;
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
>> index 5d7eaa31bf3176566f40f01ff636bee64e81c64f..d277e9b2cbc03688976b6aa481ee724b186bab51 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_panel.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
>> @@ -108,8 +108,8 @@ static u32 msm_dp_panel_get_supported_bpp(struct msm_dp_panel *msm_dp_panel,
>> return min_supported_bpp;
>> }
>>
>> -int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
>> - struct drm_connector *connector)
>> +int msm_dp_panel_read_link_caps(struct msm_dp_panel *msm_dp_panel,
>> + struct drm_connector *connector)
>
> This function doesn't require connector anymore.
Right, will remove connector in function.
>
>> {
>> int rc, bw_code;
>> int count;
>> @@ -150,8 +150,19 @@ int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
>>
>> rc = drm_dp_read_downstream_info(panel->aux, msm_dp_panel->dpcd,
>> msm_dp_panel->downstream_ports);
>> - if (rc)
>> - return rc;
>> + return rc;
>> +}
>> +
>> +int msm_dp_panel_read_edid(struct msm_dp_panel *msm_dp_panel, struct drm_connector *connector)
>> +{
>> + struct msm_dp_panel_private *panel;
>> +
>> + if (!msm_dp_panel || !connector) {
>> + DRM_ERROR("invalid input\n");
>> + return -EINVAL;
>> + }
>
> Neither panel nor connector can be NULL here, please drop.
>
>> +
>> + panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
>>
>> drm_edid_free(msm_dp_panel->drm_edid);
>>
>> @@ -163,13 +174,11 @@ int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
>> DRM_ERROR("panel edid read failed\n");
>> /* check edid read fail is due to unplug */
>> if (!msm_dp_catalog_link_is_connected(panel->catalog)) {
>> - rc = -ETIMEDOUT;
>> - goto end;
>> + return -ETIMEDOUT;
>> }
>> }
>>
>> -end:
>> - return rc;
>> + return 0;
>> }
>>
>> u32 msm_dp_panel_get_mode_bpp(struct msm_dp_panel *msm_dp_panel,
>> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
>> index 0e944db3adf2f187f313664fe80cf540ec7a19f2..7a38655c443af597c84fb78c6702b2a3ef9822ed 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_panel.h
>> +++ b/drivers/gpu/drm/msm/dp/dp_panel.h
>> @@ -59,7 +59,10 @@ void msm_dp_panel_dump_regs(struct msm_dp_panel *msm_dp_panel);
>> int msm_dp_panel_read_sink_caps(struct msm_dp_panel *msm_dp_panel,
>> struct drm_connector *connector);
>> u32 msm_dp_panel_get_mode_bpp(struct msm_dp_panel *msm_dp_panel, u32 mode_max_bpp,
>> - u32 mode_pclk_khz);
>> + u32 mode_pclk_khz);
>> +int msm_dp_panel_read_link_caps(struct msm_dp_panel *dp_panel,
>> + struct drm_connector *connector);
>> +int msm_dp_panel_read_edid(struct msm_dp_panel *dp_panel, struct drm_connector *connector);
>> int msm_dp_panel_get_modes(struct msm_dp_panel *msm_dp_panel,
>> struct drm_connector *connector);
>> void msm_dp_panel_handle_sink_request(struct msm_dp_panel *msm_dp_panel);
>>
>> --
>> 2.34.1
>>
>
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts
2025-05-22 9:51 ` Yongxing Mou
@ 2025-05-22 12:38 ` Dmitry Baryshkov
2025-05-26 12:21 ` Yongxing Mou
0 siblings, 1 reply; 111+ messages in thread
From: Dmitry Baryshkov @ 2025-05-22 12:38 UTC (permalink / raw)
To: Yongxing Mou
Cc: Abhinav Kumar, Rob Clark, Sean Paul, Marijn Suijten, David Airlie,
Simona Vetter, Stephen Boyd, Chandan Uddaraju, Guenter Roeck,
Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On Thu, May 22, 2025 at 05:51:52PM +0800, Yongxing Mou wrote:
>
>
> On 2024/12/6 16:51, Dmitry Baryshkov wrote:
> > On Thu, Dec 05, 2024 at 08:31:35PM -0800, Abhinav Kumar wrote:
> > > In preparation of DP MST where link caps are read for the
> > > immediate downstream device and the edid is read through
> > > sideband messaging, split the msm_dp_panel_read_sink_caps() into
> > > two parts which read the link parameters and the edid parts
> > > respectively.
> >
> > As you are touching this part, could you please refactor the code
> > instead by dropping the msm_dp_panel->drm_edid? There should be no need
> > to store EDID in the panel structure.
> >
> Hi, Dmitry, Abhinav will be leaving the company and will no longer be
> responsible for updating and address the comments. I will take over handling
> MST patch series. Regarding this comments, I don't got that where the
> drm_edid should be stored. In MST cases, where multiple panels exist, i
> think that there should be a separate drm_edid saved for each panel.
Why do we need to store EDID at all?
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* Re: [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts
2025-05-22 12:38 ` Dmitry Baryshkov
@ 2025-05-26 12:21 ` Yongxing Mou
0 siblings, 0 replies; 111+ messages in thread
From: Yongxing Mou @ 2025-05-26 12:21 UTC (permalink / raw)
To: Dmitry Baryshkov
Cc: Abhinav Kumar, Rob Clark, Sean Paul, Marijn Suijten, David Airlie,
Simona Vetter, Stephen Boyd, Chandan Uddaraju, Guenter Roeck,
Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On 2025/5/22 20:38, Dmitry Baryshkov wrote:
> On Thu, May 22, 2025 at 05:51:52PM +0800, Yongxing Mou wrote:
>>
>>
>> On 2024/12/6 16:51, Dmitry Baryshkov wrote:
>>> On Thu, Dec 05, 2024 at 08:31:35PM -0800, Abhinav Kumar wrote:
>>>> In preparation of DP MST where link caps are read for the
>>>> immediate downstream device and the edid is read through
>>>> sideband messaging, split the msm_dp_panel_read_sink_caps() into
>>>> two parts which read the link parameters and the edid parts
>>>> respectively.
>>>
>>> As you are touching this part, could you please refactor the code
>>> instead by dropping the msm_dp_panel->drm_edid? There should be no need
>>> to store EDID in the panel structure.
>>>
>> Hi, Dmitry, Abhinav will be leaving the company and will no longer be
>> responsible for updating and address the comments. I will take over handling
>> MST patch series. Regarding this comments, I don't got that where the
>> drm_edid should be stored. In MST cases, where multiple panels exist, i
>> think that there should be a separate drm_edid saved for each panel.
>
> Why do we need to store EDID at all?
>
Got it. Will try to drop it.
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 05/45] drm/msm/dp: add a helper to read mst caps for dp_panel
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (3 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 04/45] drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 8:52 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 06/45] drm/msm/dp: remove dp_display's dp_mode and use dp_panel's instead Abhinav Kumar
` (41 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Add a helper to check whether a dp_panel is mst capable.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_aux.h | 1 +
drivers/gpu/drm/msm/dp/dp_panel.c | 14 ++++++++++++++
drivers/gpu/drm/msm/dp/dp_panel.h | 1 +
3 files changed, 16 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_aux.h b/drivers/gpu/drm/msm/dp/dp_aux.h
index 39c5b4c8596ab28d822493a6b4d479f5f786cdee..cb97a73cdd6ea74b612053bec578247a42214f23 100644
--- a/drivers/gpu/drm/msm/dp/dp_aux.h
+++ b/drivers/gpu/drm/msm/dp/dp_aux.h
@@ -8,6 +8,7 @@
#include "dp_catalog.h"
#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_dp_mst_helper.h>
int msm_dp_aux_register(struct drm_dp_aux *msm_dp_aux);
void msm_dp_aux_unregister(struct drm_dp_aux *msm_dp_aux);
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
index d277e9b2cbc03688976b6aa481ee724b186bab51..172de804dec445cb08ad8e3f058407f483cd6684 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.c
+++ b/drivers/gpu/drm/msm/dp/dp_panel.c
@@ -108,6 +108,20 @@ static u32 msm_dp_panel_get_supported_bpp(struct msm_dp_panel *msm_dp_panel,
return min_supported_bpp;
}
+bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *msm_dp_panel)
+{
+ struct msm_dp_panel_private *panel;
+
+ if (!msm_dp_panel) {
+ DRM_ERROR("invalid input\n");
+ return 0;
+ }
+
+ panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
+
+ return drm_dp_read_mst_cap(panel->aux, msm_dp_panel->dpcd);
+}
+
int msm_dp_panel_read_link_caps(struct msm_dp_panel *msm_dp_panel,
struct drm_connector *connector)
{
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
index 7a38655c443af597c84fb78c6702b2a3ef9822ed..363b416e4cbe290f9c0e6171d6c0c5170f9fea62 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.h
+++ b/drivers/gpu/drm/msm/dp/dp_panel.h
@@ -67,6 +67,7 @@ int msm_dp_panel_get_modes(struct msm_dp_panel *msm_dp_panel,
struct drm_connector *connector);
void msm_dp_panel_handle_sink_request(struct msm_dp_panel *msm_dp_panel);
void msm_dp_panel_tpg_config(struct msm_dp_panel *msm_dp_panel, bool enable);
+bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *dp_panel);
/**
* is_link_rate_valid() - validates the link rate
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 05/45] drm/msm/dp: add a helper to read mst caps for dp_panel
2024-12-06 4:31 ` [PATCH 05/45] drm/msm/dp: add a helper to read mst caps for dp_panel Abhinav Kumar
@ 2024-12-06 8:52 ` Dmitry Baryshkov
2025-05-26 12:26 ` Yongxing Mou
0 siblings, 1 reply; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 8:52 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:36PM -0800, Abhinav Kumar wrote:
> Add a helper to check whether a dp_panel is mst capable.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_aux.h | 1 +
> drivers/gpu/drm/msm/dp/dp_panel.c | 14 ++++++++++++++
> drivers/gpu/drm/msm/dp/dp_panel.h | 1 +
> 3 files changed, 16 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_aux.h b/drivers/gpu/drm/msm/dp/dp_aux.h
> index 39c5b4c8596ab28d822493a6b4d479f5f786cdee..cb97a73cdd6ea74b612053bec578247a42214f23 100644
> --- a/drivers/gpu/drm/msm/dp/dp_aux.h
> +++ b/drivers/gpu/drm/msm/dp/dp_aux.h
> @@ -8,6 +8,7 @@
>
> #include "dp_catalog.h"
> #include <drm/display/drm_dp_helper.h>
> +#include <drm/display/drm_dp_mst_helper.h>
>
> int msm_dp_aux_register(struct drm_dp_aux *msm_dp_aux);
> void msm_dp_aux_unregister(struct drm_dp_aux *msm_dp_aux);
> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
> index d277e9b2cbc03688976b6aa481ee724b186bab51..172de804dec445cb08ad8e3f058407f483cd6684 100644
> --- a/drivers/gpu/drm/msm/dp/dp_panel.c
> +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
> @@ -108,6 +108,20 @@ static u32 msm_dp_panel_get_supported_bpp(struct msm_dp_panel *msm_dp_panel,
> return min_supported_bpp;
> }
>
> +bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *msm_dp_panel)
> +{
> + struct msm_dp_panel_private *panel;
> +
> + if (!msm_dp_panel) {
> + DRM_ERROR("invalid input\n");
> + return 0;
> + }
> +
> + panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
> +
> + return drm_dp_read_mst_cap(panel->aux, msm_dp_panel->dpcd);
So, it's a one-line wrapper. Do we actually need it?
> +}
> +
> int msm_dp_panel_read_link_caps(struct msm_dp_panel *msm_dp_panel,
> struct drm_connector *connector)
> {
> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
> index 7a38655c443af597c84fb78c6702b2a3ef9822ed..363b416e4cbe290f9c0e6171d6c0c5170f9fea62 100644
> --- a/drivers/gpu/drm/msm/dp/dp_panel.h
> +++ b/drivers/gpu/drm/msm/dp/dp_panel.h
> @@ -67,6 +67,7 @@ int msm_dp_panel_get_modes(struct msm_dp_panel *msm_dp_panel,
> struct drm_connector *connector);
> void msm_dp_panel_handle_sink_request(struct msm_dp_panel *msm_dp_panel);
> void msm_dp_panel_tpg_config(struct msm_dp_panel *msm_dp_panel, bool enable);
> +bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *dp_panel);
>
> /**
> * is_link_rate_valid() - validates the link rate
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 05/45] drm/msm/dp: add a helper to read mst caps for dp_panel
2024-12-06 8:52 ` Dmitry Baryshkov
@ 2025-05-26 12:26 ` Yongxing Mou
2025-05-28 16:13 ` Dmitry Baryshkov
0 siblings, 1 reply; 111+ messages in thread
From: Yongxing Mou @ 2025-05-26 12:26 UTC (permalink / raw)
To: Dmitry Baryshkov, Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On 2024/12/6 16:52, Dmitry Baryshkov wrote:
> On Thu, Dec 05, 2024 at 08:31:36PM -0800, Abhinav Kumar wrote:
>> Add a helper to check whether a dp_panel is mst capable.
>>
>> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
>> ---
>> drivers/gpu/drm/msm/dp/dp_aux.h | 1 +
>> drivers/gpu/drm/msm/dp/dp_panel.c | 14 ++++++++++++++
>> drivers/gpu/drm/msm/dp/dp_panel.h | 1 +
>> 3 files changed, 16 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_aux.h b/drivers/gpu/drm/msm/dp/dp_aux.h
>> index 39c5b4c8596ab28d822493a6b4d479f5f786cdee..cb97a73cdd6ea74b612053bec578247a42214f23 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_aux.h
>> +++ b/drivers/gpu/drm/msm/dp/dp_aux.h
>> @@ -8,6 +8,7 @@
>>
>> #include "dp_catalog.h"
>> #include <drm/display/drm_dp_helper.h>
>> +#include <drm/display/drm_dp_mst_helper.h>
>>
>> int msm_dp_aux_register(struct drm_dp_aux *msm_dp_aux);
>> void msm_dp_aux_unregister(struct drm_dp_aux *msm_dp_aux);
>> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
>> index d277e9b2cbc03688976b6aa481ee724b186bab51..172de804dec445cb08ad8e3f058407f483cd6684 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_panel.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
>> @@ -108,6 +108,20 @@ static u32 msm_dp_panel_get_supported_bpp(struct msm_dp_panel *msm_dp_panel,
>> return min_supported_bpp;
>> }
>>
>> +bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *msm_dp_panel)
>> +{
>> + struct msm_dp_panel_private *panel;
>> +
>> + if (!msm_dp_panel) {
>> + DRM_ERROR("invalid input\n");
>> + return 0;
>> + }
>> +
>> + panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
>> +
>> + return drm_dp_read_mst_cap(panel->aux, msm_dp_panel->dpcd);
>
> So, it's a one-line wrapper. Do we actually need it?
It beacuse the point of aux is in msm_dp_panel_private, so if we want to
call drm_dp_read_mst_cap in other file, we need this wrapper.
>
>> +}
>> +
>> int msm_dp_panel_read_link_caps(struct msm_dp_panel *msm_dp_panel,
>> struct drm_connector *connector)
>> {
>> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
>> index 7a38655c443af597c84fb78c6702b2a3ef9822ed..363b416e4cbe290f9c0e6171d6c0c5170f9fea62 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_panel.h
>> +++ b/drivers/gpu/drm/msm/dp/dp_panel.h
>> @@ -67,6 +67,7 @@ int msm_dp_panel_get_modes(struct msm_dp_panel *msm_dp_panel,
>> struct drm_connector *connector);
>> void msm_dp_panel_handle_sink_request(struct msm_dp_panel *msm_dp_panel);
>> void msm_dp_panel_tpg_config(struct msm_dp_panel *msm_dp_panel, bool enable);
>> +bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *dp_panel);
>>
>> /**
>> * is_link_rate_valid() - validates the link rate
>>
>> --
>> 2.34.1
>>
>
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 05/45] drm/msm/dp: add a helper to read mst caps for dp_panel
2025-05-26 12:26 ` Yongxing Mou
@ 2025-05-28 16:13 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2025-05-28 16:13 UTC (permalink / raw)
To: Yongxing Mou
Cc: Abhinav Kumar, Rob Clark, Sean Paul, Marijn Suijten, David Airlie,
Simona Vetter, Stephen Boyd, Chandan Uddaraju, Guenter Roeck,
Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On Mon, May 26, 2025 at 08:26:47PM +0800, Yongxing Mou wrote:
>
>
> On 2024/12/6 16:52, Dmitry Baryshkov wrote:
> > On Thu, Dec 05, 2024 at 08:31:36PM -0800, Abhinav Kumar wrote:
> > > Add a helper to check whether a dp_panel is mst capable.
> > >
> > > Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> > > ---
> > > drivers/gpu/drm/msm/dp/dp_aux.h | 1 +
> > > drivers/gpu/drm/msm/dp/dp_panel.c | 14 ++++++++++++++
> > > drivers/gpu/drm/msm/dp/dp_panel.h | 1 +
> > > 3 files changed, 16 insertions(+)
> > >
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_aux.h b/drivers/gpu/drm/msm/dp/dp_aux.h
> > > index 39c5b4c8596ab28d822493a6b4d479f5f786cdee..cb97a73cdd6ea74b612053bec578247a42214f23 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_aux.h
> > > +++ b/drivers/gpu/drm/msm/dp/dp_aux.h
> > > @@ -8,6 +8,7 @@
> > > #include "dp_catalog.h"
> > > #include <drm/display/drm_dp_helper.h>
> > > +#include <drm/display/drm_dp_mst_helper.h>
> > > int msm_dp_aux_register(struct drm_dp_aux *msm_dp_aux);
> > > void msm_dp_aux_unregister(struct drm_dp_aux *msm_dp_aux);
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
> > > index d277e9b2cbc03688976b6aa481ee724b186bab51..172de804dec445cb08ad8e3f058407f483cd6684 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_panel.c
> > > +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
> > > @@ -108,6 +108,20 @@ static u32 msm_dp_panel_get_supported_bpp(struct msm_dp_panel *msm_dp_panel,
> > > return min_supported_bpp;
> > > }
> > > +bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *msm_dp_panel)
> > > +{
> > > + struct msm_dp_panel_private *panel;
> > > +
> > > + if (!msm_dp_panel) {
> > > + DRM_ERROR("invalid input\n");
> > > + return 0;
> > > + }
> > > +
> > > + panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
> > > +
> > > + return drm_dp_read_mst_cap(panel->aux, msm_dp_panel->dpcd);
> >
> > So, it's a one-line wrapper. Do we actually need it?
> It beacuse the point of aux is in msm_dp_panel_private, so if we want to
> call drm_dp_read_mst_cap in other file, we need this wrapper.
It is being used from dp_display, which also ->aux pointer.
> >
> > > +}
> > > +
> > > int msm_dp_panel_read_link_caps(struct msm_dp_panel *msm_dp_panel,
> > > struct drm_connector *connector)
> > > {
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
> > > index 7a38655c443af597c84fb78c6702b2a3ef9822ed..363b416e4cbe290f9c0e6171d6c0c5170f9fea62 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_panel.h
> > > +++ b/drivers/gpu/drm/msm/dp/dp_panel.h
> > > @@ -67,6 +67,7 @@ int msm_dp_panel_get_modes(struct msm_dp_panel *msm_dp_panel,
> > > struct drm_connector *connector);
> > > void msm_dp_panel_handle_sink_request(struct msm_dp_panel *msm_dp_panel);
> > > void msm_dp_panel_tpg_config(struct msm_dp_panel *msm_dp_panel, bool enable);
> > > +bool msm_dp_panel_read_mst_cap(struct msm_dp_panel *dp_panel);
> > > /**
> > > * is_link_rate_valid() - validates the link rate
> > >
> > > --
> > > 2.34.1
> > >
> >
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 06/45] drm/msm/dp: remove dp_display's dp_mode and use dp_panel's instead
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (4 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 05/45] drm/msm/dp: add a helper to read mst caps for dp_panel Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 9:01 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 07/45] drm/msm/dp: break up dp_display_enable into two parts Abhinav Kumar
` (40 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
dp_display caches the current display mode and then passes it onto
the panel to be used for programming the panel params. Remove this
two level passing and directly populated the panel's dp_display_mode
instead.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 46 ++++++++++++++++---------------------
1 file changed, 20 insertions(+), 26 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 052db80c6a365f53c2c0a37d3b69ea2b627aea1f..4bd85ae754429333aa423c985368344cd03c7752 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -90,7 +90,6 @@ struct msm_dp_display_private {
struct msm_dp_panel *panel;
struct msm_dp_ctrl *ctrl;
- struct msm_dp_display_mode msm_dp_mode;
struct msm_dp msm_dp_display;
/* wait for audio signaling */
@@ -1436,10 +1435,13 @@ bool msm_dp_needs_periph_flush(const struct msm_dp *msm_dp_display,
bool msm_dp_wide_bus_available(const struct msm_dp *msm_dp_display)
{
struct msm_dp_display_private *dp;
+ struct msm_dp_panel *dp_panel;
dp = container_of(msm_dp_display, struct msm_dp_display_private, msm_dp_display);
- if (dp->msm_dp_mode.out_fmt_is_yuv_420)
+ dp_panel = dp->panel;
+
+ if (dp_panel->msm_dp_mode.out_fmt_is_yuv_420)
return false;
return dp->wide_bus_supported;
@@ -1501,10 +1503,6 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
bool force_link_train = false;
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
- if (!msm_dp_display->msm_dp_mode.drm_mode.clock) {
- DRM_ERROR("invalid params\n");
- return;
- }
if (dp->is_edp)
msm_dp_hpd_plug_handle(msm_dp_display, 0);
@@ -1516,13 +1514,6 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
return;
}
- rc = msm_dp_display_set_mode(dp, &msm_dp_display->msm_dp_mode);
- if (rc) {
- DRM_ERROR("Failed to perform a mode set, rc=%d\n", rc);
- mutex_unlock(&msm_dp_display->event_mutex);
- return;
- }
-
state = msm_dp_display->hpd_state;
if (state == ST_CONNECTED && !dp->power_on) {
@@ -1599,37 +1590,40 @@ void msm_dp_bridge_mode_set(struct drm_bridge *drm_bridge,
struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
struct msm_dp_display_private *msm_dp_display;
struct msm_dp_panel *msm_dp_panel;
+ struct msm_dp_display_mode msm_dp_mode;
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
msm_dp_panel = msm_dp_display->panel;
- memset(&msm_dp_display->msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
+ memset(&msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
if (msm_dp_display_check_video_test(dp))
- msm_dp_display->msm_dp_mode.bpp = msm_dp_display_get_test_bpp(dp);
+ msm_dp_mode.bpp = msm_dp_display_get_test_bpp(dp);
else /* Default num_components per pixel = 3 */
- msm_dp_display->msm_dp_mode.bpp = dp->connector->display_info.bpc * 3;
+ msm_dp_mode.bpp = dp->connector->display_info.bpc * 3;
- if (!msm_dp_display->msm_dp_mode.bpp)
- msm_dp_display->msm_dp_mode.bpp = 24; /* Default bpp */
+ if (!msm_dp_mode.bpp)
+ msm_dp_mode.bpp = 24; /* Default bpp */
- drm_mode_copy(&msm_dp_display->msm_dp_mode.drm_mode, adjusted_mode);
+ drm_mode_copy(&msm_dp_mode.drm_mode, adjusted_mode);
- msm_dp_display->msm_dp_mode.v_active_low =
- !!(msm_dp_display->msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
+ msm_dp_mode.v_active_low =
+ !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
- msm_dp_display->msm_dp_mode.h_active_low =
- !!(msm_dp_display->msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
+ msm_dp_mode.h_active_low =
+ !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
- msm_dp_display->msm_dp_mode.out_fmt_is_yuv_420 =
+ msm_dp_mode.out_fmt_is_yuv_420 =
drm_mode_is_420_only(&dp->connector->display_info, adjusted_mode) &&
msm_dp_panel->vsc_sdp_supported;
+ msm_dp_display_set_mode(dp, &msm_dp_mode);
+
/* populate wide_bus_support to different layers */
msm_dp_display->ctrl->wide_bus_en =
- msm_dp_display->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
+ msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
msm_dp_display->catalog->wide_bus_en =
- msm_dp_display->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
+ msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
}
void msm_dp_bridge_hpd_enable(struct drm_bridge *bridge)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 06/45] drm/msm/dp: remove dp_display's dp_mode and use dp_panel's instead
2024-12-06 4:31 ` [PATCH 06/45] drm/msm/dp: remove dp_display's dp_mode and use dp_panel's instead Abhinav Kumar
@ 2024-12-06 9:01 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:01 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:37PM -0800, Abhinav Kumar wrote:
> dp_display caches the current display mode and then passes it onto
> the panel to be used for programming the panel params. Remove this
> two level passing and directly populated the panel's dp_display_mode
> instead.
Remove both and use the one from crtc_state?
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 46 ++++++++++++++++---------------------
> 1 file changed, 20 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 052db80c6a365f53c2c0a37d3b69ea2b627aea1f..4bd85ae754429333aa423c985368344cd03c7752 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -90,7 +90,6 @@ struct msm_dp_display_private {
> struct msm_dp_panel *panel;
> struct msm_dp_ctrl *ctrl;
>
> - struct msm_dp_display_mode msm_dp_mode;
> struct msm_dp msm_dp_display;
>
> /* wait for audio signaling */
> @@ -1436,10 +1435,13 @@ bool msm_dp_needs_periph_flush(const struct msm_dp *msm_dp_display,
> bool msm_dp_wide_bus_available(const struct msm_dp *msm_dp_display)
> {
> struct msm_dp_display_private *dp;
> + struct msm_dp_panel *dp_panel;
>
> dp = container_of(msm_dp_display, struct msm_dp_display_private, msm_dp_display);
>
> - if (dp->msm_dp_mode.out_fmt_is_yuv_420)
> + dp_panel = dp->panel;
> +
> + if (dp_panel->msm_dp_mode.out_fmt_is_yuv_420)
> return false;
>
> return dp->wide_bus_supported;
> @@ -1501,10 +1503,6 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
> bool force_link_train = false;
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
> - if (!msm_dp_display->msm_dp_mode.drm_mode.clock) {
> - DRM_ERROR("invalid params\n");
> - return;
> - }
>
> if (dp->is_edp)
> msm_dp_hpd_plug_handle(msm_dp_display, 0);
> @@ -1516,13 +1514,6 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
> return;
> }
>
> - rc = msm_dp_display_set_mode(dp, &msm_dp_display->msm_dp_mode);
> - if (rc) {
> - DRM_ERROR("Failed to perform a mode set, rc=%d\n", rc);
> - mutex_unlock(&msm_dp_display->event_mutex);
> - return;
> - }
> -
> state = msm_dp_display->hpd_state;
>
> if (state == ST_CONNECTED && !dp->power_on) {
> @@ -1599,37 +1590,40 @@ void msm_dp_bridge_mode_set(struct drm_bridge *drm_bridge,
> struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
> struct msm_dp_display_private *msm_dp_display;
> struct msm_dp_panel *msm_dp_panel;
> + struct msm_dp_display_mode msm_dp_mode;
No need to allocate it on stack just to copy it later on. Please write
the data directly to a proper location from the beginning.
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
> msm_dp_panel = msm_dp_display->panel;
>
> - memset(&msm_dp_display->msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
> + memset(&msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
>
> if (msm_dp_display_check_video_test(dp))
> - msm_dp_display->msm_dp_mode.bpp = msm_dp_display_get_test_bpp(dp);
> + msm_dp_mode.bpp = msm_dp_display_get_test_bpp(dp);
> else /* Default num_components per pixel = 3 */
> - msm_dp_display->msm_dp_mode.bpp = dp->connector->display_info.bpc * 3;
> + msm_dp_mode.bpp = dp->connector->display_info.bpc * 3;
>
> - if (!msm_dp_display->msm_dp_mode.bpp)
> - msm_dp_display->msm_dp_mode.bpp = 24; /* Default bpp */
> + if (!msm_dp_mode.bpp)
> + msm_dp_mode.bpp = 24; /* Default bpp */
The msm_dp_mode.bpp gets rewritten by msm_dp_panel_init_panel_info()
after being set here. Is this code part redundant?
>
> - drm_mode_copy(&msm_dp_display->msm_dp_mode.drm_mode, adjusted_mode);
> + drm_mode_copy(&msm_dp_mode.drm_mode, adjusted_mode);
>
> - msm_dp_display->msm_dp_mode.v_active_low =
> - !!(msm_dp_display->msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
> + msm_dp_mode.v_active_low =
> + !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
>
> - msm_dp_display->msm_dp_mode.h_active_low =
> - !!(msm_dp_display->msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
> + msm_dp_mode.h_active_low =
> + !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
>
> - msm_dp_display->msm_dp_mode.out_fmt_is_yuv_420 =
> + msm_dp_mode.out_fmt_is_yuv_420 =
> drm_mode_is_420_only(&dp->connector->display_info, adjusted_mode) &&
> msm_dp_panel->vsc_sdp_supported;
>
> + msm_dp_display_set_mode(dp, &msm_dp_mode);
> +
> /* populate wide_bus_support to different layers */
> msm_dp_display->ctrl->wide_bus_en =
> - msm_dp_display->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
> + msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
> msm_dp_display->catalog->wide_bus_en =
> - msm_dp_display->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
> + msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
> }
>
> void msm_dp_bridge_hpd_enable(struct drm_bridge *bridge)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 07/45] drm/msm/dp: break up dp_display_enable into two parts
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (5 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 06/45] drm/msm/dp: remove dp_display's dp_mode and use dp_panel's instead Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 9:04 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 08/45] drm/msm/dp: re-arrange dp_display_disable() into functional parts Abhinav Kumar
` (39 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
dp_display_enable() currently re-trains the link if needed
and then enables the pixel clock, programs the controller to
start sending the pixel stream. Break up these two parts into
separate APIs to distinguish these two parts better.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 57 ++++++++++++++--------
drivers/gpu/drm/msm/dp/dp_ctrl.h | 3 +-
drivers/gpu/drm/msm/dp/dp_display.c | 97 ++++++++++++++++++++++++++++---------
drivers/gpu/drm/msm/dp/dp_display.h | 1 +
4 files changed, 112 insertions(+), 46 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 9c463ae2f8fae916661fef1c7e225f55c1026478..763bd58c24d29d49caafb76eab32b493e1618e7b 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -1941,40 +1941,61 @@ static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl)
return msm_dp_ctrl_setup_main_link(ctrl, &training_step);
}
-int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train)
+int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train)
{
int ret = 0;
- bool mainlink_ready = false;
struct msm_dp_ctrl_private *ctrl;
- unsigned long pixel_rate;
- unsigned long pixel_rate_orig;
if (!msm_dp_ctrl)
return -EINVAL;
ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
- pixel_rate = pixel_rate_orig = ctrl->panel->msm_dp_mode.drm_mode.clock;
-
- if (msm_dp_ctrl->wide_bus_en || ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
- pixel_rate >>= 1;
-
- drm_dbg_dp(ctrl->drm_dev, "rate=%d, num_lanes=%d, pixel_rate=%lu\n",
- ctrl->link->link_params.rate,
- ctrl->link->link_params.num_lanes, pixel_rate);
+ drm_dbg_dp(ctrl->drm_dev, "rate=%d, num_lanes=%d\n",
+ ctrl->link->link_params.rate,
+ ctrl->link->link_params.num_lanes);
drm_dbg_dp(ctrl->drm_dev,
- "core_clk_on=%d link_clk_on=%d stream_clk_on=%d\n",
- ctrl->core_clks_on, ctrl->link_clks_on, ctrl->stream_clks_on);
+ "core_clk_on=%d link_clk_on=%d stream_clk_on=%d\n",
+ ctrl->core_clks_on, ctrl->link_clks_on, ctrl->stream_clks_on);
if (!ctrl->link_clks_on) { /* link clk is off */
ret = msm_dp_ctrl_enable_mainlink_clocks(ctrl);
if (ret) {
DRM_ERROR("Failed to start link clocks. ret=%d\n", ret);
- goto end;
+ return ret;
}
}
+ if (force_link_train || !msm_dp_ctrl_channel_eq_ok(ctrl))
+ msm_dp_ctrl_link_retrain(ctrl);
+
+ /* stop txing train pattern to end link training */
+ msm_dp_ctrl_clear_training_pattern(ctrl);
+
+ return ret;
+}
+
+int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
+{
+ int ret = 0;
+ bool mainlink_ready = false;
+ struct msm_dp_ctrl_private *ctrl;
+ unsigned long pixel_rate;
+ unsigned long pixel_rate_orig;
+
+ if (!msm_dp_ctrl)
+ return -EINVAL;
+
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+
+ pixel_rate = pixel_rate_orig = ctrl->panel->msm_dp_mode.drm_mode.clock;
+
+ if (msm_dp_ctrl->wide_bus_en || ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
+ pixel_rate >>= 1;
+
+ drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
+
ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
if (ret) {
DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
@@ -1992,12 +2013,6 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train
ctrl->stream_clks_on = true;
}
- if (force_link_train || !msm_dp_ctrl_channel_eq_ok(ctrl))
- msm_dp_ctrl_link_retrain(ctrl);
-
- /* stop txing train pattern to end link training */
- msm_dp_ctrl_clear_training_pattern(ctrl);
-
/*
* Set up transfer unit values and set controller state to send
* video.
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index b7abfedbf5749c25877a0b8ba3af3d8ed4b23d67..42745c912adbad7221c78f5cecefa730bfda1e75 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -18,7 +18,8 @@ struct msm_dp_ctrl {
struct phy;
int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl);
-int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train);
+int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl);
+int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 4bd85ae754429333aa423c985368344cd03c7752..434380b442ec84c12c240dab6538ccdf31963cea 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -811,7 +811,37 @@ static int msm_dp_display_set_mode(struct msm_dp *msm_dp_display,
return 0;
}
-static int msm_dp_display_enable(struct msm_dp_display_private *dp, bool force_link_train)
+static int msm_dp_display_prepare(struct msm_dp_display_private *dp)
+{
+ int rc = 0;
+ struct msm_dp *msm_dp_display = &dp->msm_dp_display;
+ bool force_link_train = false;
+
+ drm_dbg_dp(dp->drm_dev, "sink_count=%d\n", dp->link->sink_count);
+ if (msm_dp_display->prepared) {
+ drm_dbg_dp(dp->drm_dev, "Link already setup, return\n");
+ return 0;
+ }
+
+ rc = pm_runtime_resume_and_get(&msm_dp_display->pdev->dev);
+ if (rc) {
+ DRM_ERROR("failed to pm_runtime_resume\n");
+ return rc;
+ }
+
+ if (dp->hpd_state == ST_CONNECTED && !msm_dp_display->power_on) {
+ msm_dp_display_host_phy_init(dp);
+ force_link_train = true;
+ }
+
+ rc = msm_dp_ctrl_prepare_stream_on(dp->ctrl, force_link_train);
+ if (!rc)
+ msm_dp_display->prepared = true;
+
+ return rc;
+}
+
+static int msm_dp_display_enable(struct msm_dp_display_private *dp)
{
int rc = 0;
struct msm_dp *msm_dp_display = &dp->msm_dp_display;
@@ -822,7 +852,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp, bool force_l
return 0;
}
- rc = msm_dp_ctrl_on_stream(dp->ctrl, force_link_train);
+ rc = msm_dp_ctrl_on_stream(dp->ctrl);
if (!rc)
msm_dp_display->power_on = true;
@@ -852,13 +882,10 @@ static int msm_dp_display_post_enable(struct msm_dp *msm_dp_display)
return 0;
}
-static int msm_dp_display_disable(struct msm_dp_display_private *dp)
+static void msm_dp_display_audio_notify_disable(struct msm_dp_display_private *dp)
{
struct msm_dp *msm_dp_display = &dp->msm_dp_display;
- if (!msm_dp_display->power_on)
- return 0;
-
/* wait only if audio was enabled */
if (msm_dp_display->audio_enabled) {
/* signal the disconnect event */
@@ -869,6 +896,14 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
}
msm_dp_display->audio_enabled = false;
+}
+
+static int msm_dp_display_disable(struct msm_dp_display_private *dp)
+{
+ struct msm_dp *msm_dp_display = &dp->msm_dp_display;
+
+ if (!msm_dp_display->power_on)
+ return 0;
if (dp->link->sink_count == 0) {
/*
@@ -1498,9 +1533,8 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(drm_bridge);
struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
int rc = 0;
+
struct msm_dp_display_private *msm_dp_display;
- u32 state;
- bool force_link_train = false;
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
@@ -1508,25 +1542,23 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
msm_dp_hpd_plug_handle(msm_dp_display, 0);
mutex_lock(&msm_dp_display->event_mutex);
- if (pm_runtime_resume_and_get(&dp->pdev->dev)) {
- DRM_ERROR("failed to pm_runtime_resume\n");
+
+ rc = msm_dp_display_prepare(msm_dp_display);
+ if (rc) {
+ DRM_ERROR("DP display prepare failed, rc=%d\n", rc);
mutex_unlock(&msm_dp_display->event_mutex);
return;
}
- state = msm_dp_display->hpd_state;
-
- if (state == ST_CONNECTED && !dp->power_on) {
- msm_dp_display_host_phy_init(msm_dp_display);
- force_link_train = true;
- }
-
- msm_dp_display_enable(msm_dp_display, force_link_train);
-
- rc = msm_dp_display_post_enable(dp);
- if (rc) {
- DRM_ERROR("DP display post enable failed, rc=%d\n", rc);
- msm_dp_display_disable(msm_dp_display);
+ if (dp->prepared) {
+ rc = msm_dp_display_enable(msm_dp_display);
+ if (rc)
+ DRM_ERROR("DP display enable failed, rc=%d\n", rc);
+ rc = msm_dp_display_post_enable(dp);
+ if (rc) {
+ DRM_ERROR("DP display post enable failed, rc=%d\n", rc);
+ msm_dp_display_disable(msm_dp_display);
+ }
}
/* completed connection */
@@ -1548,6 +1580,20 @@ void msm_dp_bridge_atomic_disable(struct drm_bridge *drm_bridge,
msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
}
+static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
+{
+ struct msm_dp *msm_dp_display = &dp->msm_dp_display;
+
+ if (!msm_dp_display->prepared) {
+ drm_dbg_dp(dp->drm_dev, "Link already setup, return\n");
+ return;
+ }
+
+ pm_runtime_put_sync(&msm_dp_display->pdev->dev);
+
+ msm_dp_display->prepared = false;
+}
+
void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
struct drm_bridge_state *old_bridge_state)
{
@@ -1568,6 +1614,8 @@ void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
drm_dbg_dp(dp->drm_dev, "type=%d wrong hpd_state=%d\n",
dp->connector_type, state);
+ msm_dp_display_audio_notify_disable(msm_dp_display);
+
msm_dp_display_disable(msm_dp_display);
state = msm_dp_display->hpd_state;
@@ -1576,9 +1624,10 @@ void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
msm_dp_display->hpd_state = ST_DISCONNECTED;
}
+ msm_dp_display_unprepare(msm_dp_display);
+
drm_dbg_dp(dp->drm_dev, "type=%d Done\n", dp->connector_type);
- pm_runtime_put_sync(&dp->pdev->dev);
mutex_unlock(&msm_dp_display->event_mutex);
}
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index ecbc2d92f546a346ee53adcf1b060933e4f54317..0165a80863e65b6eea6d2d8d1c5c08cd51fed101 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -21,6 +21,7 @@ struct msm_dp {
bool link_ready;
bool audio_enabled;
bool power_on;
+ bool prepared;
unsigned int connector_type;
bool is_edp;
bool internal_hpd;
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 07/45] drm/msm/dp: break up dp_display_enable into two parts
2024-12-06 4:31 ` [PATCH 07/45] drm/msm/dp: break up dp_display_enable into two parts Abhinav Kumar
@ 2024-12-06 9:04 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:04 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:38PM -0800, Abhinav Kumar wrote:
> dp_display_enable() currently re-trains the link if needed
> and then enables the pixel clock, programs the controller to
> start sending the pixel stream. Break up these two parts into
> separate APIs to distinguish these two parts better.
This does more than just breaking dp_display_enable(). It moves parts of
the code sequence to the .prepare() callback, etc. Please review this
again.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 57 ++++++++++++++--------
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 3 +-
> drivers/gpu/drm/msm/dp/dp_display.c | 97 ++++++++++++++++++++++++++++---------
> drivers/gpu/drm/msm/dp/dp_display.h | 1 +
> 4 files changed, 112 insertions(+), 46 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 9c463ae2f8fae916661fef1c7e225f55c1026478..763bd58c24d29d49caafb76eab32b493e1618e7b 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -1941,40 +1941,61 @@ static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl)
> return msm_dp_ctrl_setup_main_link(ctrl, &training_step);
> }
>
> -int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train)
> +int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train)
> {
> int ret = 0;
> - bool mainlink_ready = false;
> struct msm_dp_ctrl_private *ctrl;
> - unsigned long pixel_rate;
> - unsigned long pixel_rate_orig;
>
> if (!msm_dp_ctrl)
> return -EINVAL;
>
> ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
>
> - pixel_rate = pixel_rate_orig = ctrl->panel->msm_dp_mode.drm_mode.clock;
> -
> - if (msm_dp_ctrl->wide_bus_en || ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
> - pixel_rate >>= 1;
> -
> - drm_dbg_dp(ctrl->drm_dev, "rate=%d, num_lanes=%d, pixel_rate=%lu\n",
> - ctrl->link->link_params.rate,
> - ctrl->link->link_params.num_lanes, pixel_rate);
> + drm_dbg_dp(ctrl->drm_dev, "rate=%d, num_lanes=%d\n",
> + ctrl->link->link_params.rate,
> + ctrl->link->link_params.num_lanes);
>
> drm_dbg_dp(ctrl->drm_dev,
> - "core_clk_on=%d link_clk_on=%d stream_clk_on=%d\n",
> - ctrl->core_clks_on, ctrl->link_clks_on, ctrl->stream_clks_on);
> + "core_clk_on=%d link_clk_on=%d stream_clk_on=%d\n",
> + ctrl->core_clks_on, ctrl->link_clks_on, ctrl->stream_clks_on);
>
> if (!ctrl->link_clks_on) { /* link clk is off */
> ret = msm_dp_ctrl_enable_mainlink_clocks(ctrl);
> if (ret) {
> DRM_ERROR("Failed to start link clocks. ret=%d\n", ret);
> - goto end;
> + return ret;
> }
> }
>
> + if (force_link_train || !msm_dp_ctrl_channel_eq_ok(ctrl))
> + msm_dp_ctrl_link_retrain(ctrl);
> +
> + /* stop txing train pattern to end link training */
> + msm_dp_ctrl_clear_training_pattern(ctrl);
> +
> + return ret;
> +}
> +
> +int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
> +{
> + int ret = 0;
> + bool mainlink_ready = false;
> + struct msm_dp_ctrl_private *ctrl;
> + unsigned long pixel_rate;
> + unsigned long pixel_rate_orig;
> +
> + if (!msm_dp_ctrl)
> + return -EINVAL;
> +
> + ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> +
> + pixel_rate = pixel_rate_orig = ctrl->panel->msm_dp_mode.drm_mode.clock;
> +
> + if (msm_dp_ctrl->wide_bus_en || ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
> + pixel_rate >>= 1;
> +
> + drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
> +
> ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
> if (ret) {
> DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
> @@ -1992,12 +2013,6 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train
> ctrl->stream_clks_on = true;
> }
>
> - if (force_link_train || !msm_dp_ctrl_channel_eq_ok(ctrl))
> - msm_dp_ctrl_link_retrain(ctrl);
> -
> - /* stop txing train pattern to end link training */
> - msm_dp_ctrl_clear_training_pattern(ctrl);
> -
> /*
> * Set up transfer unit values and set controller state to send
> * video.
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index b7abfedbf5749c25877a0b8ba3af3d8ed4b23d67..42745c912adbad7221c78f5cecefa730bfda1e75 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -18,7 +18,8 @@ struct msm_dp_ctrl {
> struct phy;
>
> int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl);
> -int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train);
> +int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> +int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
> void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl);
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 4bd85ae754429333aa423c985368344cd03c7752..434380b442ec84c12c240dab6538ccdf31963cea 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -811,7 +811,37 @@ static int msm_dp_display_set_mode(struct msm_dp *msm_dp_display,
> return 0;
> }
>
> -static int msm_dp_display_enable(struct msm_dp_display_private *dp, bool force_link_train)
> +static int msm_dp_display_prepare(struct msm_dp_display_private *dp)
> +{
> + int rc = 0;
> + struct msm_dp *msm_dp_display = &dp->msm_dp_display;
> + bool force_link_train = false;
> +
> + drm_dbg_dp(dp->drm_dev, "sink_count=%d\n", dp->link->sink_count);
> + if (msm_dp_display->prepared) {
> + drm_dbg_dp(dp->drm_dev, "Link already setup, return\n");
> + return 0;
> + }
> +
> + rc = pm_runtime_resume_and_get(&msm_dp_display->pdev->dev);
> + if (rc) {
> + DRM_ERROR("failed to pm_runtime_resume\n");
> + return rc;
> + }
> +
> + if (dp->hpd_state == ST_CONNECTED && !msm_dp_display->power_on) {
> + msm_dp_display_host_phy_init(dp);
> + force_link_train = true;
> + }
> +
> + rc = msm_dp_ctrl_prepare_stream_on(dp->ctrl, force_link_train);
> + if (!rc)
> + msm_dp_display->prepared = true;
> +
> + return rc;
> +}
> +
> +static int msm_dp_display_enable(struct msm_dp_display_private *dp)
> {
> int rc = 0;
> struct msm_dp *msm_dp_display = &dp->msm_dp_display;
> @@ -822,7 +852,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp, bool force_l
> return 0;
> }
>
> - rc = msm_dp_ctrl_on_stream(dp->ctrl, force_link_train);
> + rc = msm_dp_ctrl_on_stream(dp->ctrl);
> if (!rc)
> msm_dp_display->power_on = true;
>
> @@ -852,13 +882,10 @@ static int msm_dp_display_post_enable(struct msm_dp *msm_dp_display)
> return 0;
> }
>
> -static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> +static void msm_dp_display_audio_notify_disable(struct msm_dp_display_private *dp)
> {
> struct msm_dp *msm_dp_display = &dp->msm_dp_display;
>
> - if (!msm_dp_display->power_on)
> - return 0;
> -
> /* wait only if audio was enabled */
> if (msm_dp_display->audio_enabled) {
> /* signal the disconnect event */
> @@ -869,6 +896,14 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> }
>
> msm_dp_display->audio_enabled = false;
> +}
> +
> +static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> +{
> + struct msm_dp *msm_dp_display = &dp->msm_dp_display;
> +
> + if (!msm_dp_display->power_on)
> + return 0;
>
> if (dp->link->sink_count == 0) {
> /*
> @@ -1498,9 +1533,8 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
> struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(drm_bridge);
> struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
> int rc = 0;
> +
> struct msm_dp_display_private *msm_dp_display;
> - u32 state;
> - bool force_link_train = false;
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
>
> @@ -1508,25 +1542,23 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
> msm_dp_hpd_plug_handle(msm_dp_display, 0);
>
> mutex_lock(&msm_dp_display->event_mutex);
> - if (pm_runtime_resume_and_get(&dp->pdev->dev)) {
> - DRM_ERROR("failed to pm_runtime_resume\n");
> +
> + rc = msm_dp_display_prepare(msm_dp_display);
> + if (rc) {
> + DRM_ERROR("DP display prepare failed, rc=%d\n", rc);
> mutex_unlock(&msm_dp_display->event_mutex);
> return;
> }
>
> - state = msm_dp_display->hpd_state;
> -
> - if (state == ST_CONNECTED && !dp->power_on) {
> - msm_dp_display_host_phy_init(msm_dp_display);
> - force_link_train = true;
> - }
> -
> - msm_dp_display_enable(msm_dp_display, force_link_train);
> -
> - rc = msm_dp_display_post_enable(dp);
> - if (rc) {
> - DRM_ERROR("DP display post enable failed, rc=%d\n", rc);
> - msm_dp_display_disable(msm_dp_display);
> + if (dp->prepared) {
> + rc = msm_dp_display_enable(msm_dp_display);
> + if (rc)
> + DRM_ERROR("DP display enable failed, rc=%d\n", rc);
> + rc = msm_dp_display_post_enable(dp);
> + if (rc) {
> + DRM_ERROR("DP display post enable failed, rc=%d\n", rc);
> + msm_dp_display_disable(msm_dp_display);
> + }
> }
>
> /* completed connection */
> @@ -1548,6 +1580,20 @@ void msm_dp_bridge_atomic_disable(struct drm_bridge *drm_bridge,
> msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
> }
>
> +static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
> +{
> + struct msm_dp *msm_dp_display = &dp->msm_dp_display;
> +
> + if (!msm_dp_display->prepared) {
> + drm_dbg_dp(dp->drm_dev, "Link already setup, return\n");
> + return;
> + }
> +
> + pm_runtime_put_sync(&msm_dp_display->pdev->dev);
> +
> + msm_dp_display->prepared = false;
> +}
> +
> void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
> struct drm_bridge_state *old_bridge_state)
> {
> @@ -1568,6 +1614,8 @@ void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
> drm_dbg_dp(dp->drm_dev, "type=%d wrong hpd_state=%d\n",
> dp->connector_type, state);
>
> + msm_dp_display_audio_notify_disable(msm_dp_display);
> +
> msm_dp_display_disable(msm_dp_display);
>
> state = msm_dp_display->hpd_state;
> @@ -1576,9 +1624,10 @@ void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
> msm_dp_display->hpd_state = ST_DISCONNECTED;
> }
>
> + msm_dp_display_unprepare(msm_dp_display);
> +
> drm_dbg_dp(dp->drm_dev, "type=%d Done\n", dp->connector_type);
>
> - pm_runtime_put_sync(&dp->pdev->dev);
> mutex_unlock(&msm_dp_display->event_mutex);
> }
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> index ecbc2d92f546a346ee53adcf1b060933e4f54317..0165a80863e65b6eea6d2d8d1c5c08cd51fed101 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.h
> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> @@ -21,6 +21,7 @@ struct msm_dp {
> bool link_ready;
> bool audio_enabled;
> bool power_on;
> + bool prepared;
> unsigned int connector_type;
> bool is_edp;
> bool internal_hpd;
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 08/45] drm/msm/dp: re-arrange dp_display_disable() into functional parts
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (6 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 07/45] drm/msm/dp: break up dp_display_enable into two parts Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 12:03 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 09/45] drm/msm/dp: allow dp_ctrl stream APIs to use any panel passed to it Abhinav Kumar
` (38 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
dp_display_disable() handles special case of when monitor is
disconnected from the dongle while the dongle stays connected
thereby needing a separate function dp_ctrl_off_link_stream()
for this. However with a slight rework this can still be handled
by keeping common paths same for regular and special case.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 29 +++++++++++++++--------------
drivers/gpu/drm/msm/dp/dp_ctrl.h | 4 ++++
drivers/gpu/drm/msm/dp/dp_display.c | 25 ++++++++++++-------------
3 files changed, 31 insertions(+), 27 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 763bd58c24d29d49caafb76eab32b493e1618e7b..9e08996be0cb969cb96d9a3019c445ab4dfc92ef 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -2042,30 +2042,31 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
return ret;
}
-void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl)
+void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl)
{
struct msm_dp_ctrl_private *ctrl;
- struct phy *phy;
ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
- phy = ctrl->phy;
-
msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
+}
- /* set dongle to D3 (power off) mode */
- msm_dp_link_psm_config(ctrl->link, &ctrl->panel->link_info, true);
+void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl)
+{
+ struct msm_dp_ctrl_private *ctrl;
- msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
- if (ctrl->stream_clks_on) {
- clk_disable_unprepare(ctrl->pixel_clk);
- ctrl->stream_clks_on = false;
- }
+ /* set dongle to D3 (power off) mode */
+ msm_dp_link_psm_config(ctrl->link, &ctrl->panel->link_info, true);
+}
- dev_pm_opp_set_rate(ctrl->dev, 0);
- msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
+void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl)
+{
+ struct msm_dp_ctrl_private *ctrl;
+ struct phy *phy;
- phy_power_off(phy);
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+ phy = ctrl->phy;
/* aux channel down, reinit phy */
phy_exit(phy);
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index 42745c912adbad7221c78f5cecefa730bfda1e75..0f58b63c5c7c5aab43c0db2a697ba491959b79d2 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -42,4 +42,8 @@ void msm_dp_ctrl_config_psr(struct msm_dp_ctrl *msm_dp_ctrl);
int msm_dp_ctrl_core_clk_enable(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_core_clk_disable(struct msm_dp_ctrl *msm_dp_ctrl);
+void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl);
+void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl);
+void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl);
+
#endif /* _DP_CTRL_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 434380b442ec84c12c240dab6538ccdf31963cea..bbce8ca09ff70059458231982f002e1f22d2c3ab 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -905,20 +905,19 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
if (!msm_dp_display->power_on)
return 0;
- if (dp->link->sink_count == 0) {
- /*
- * irq_hpd with sink_count = 0
- * hdmi unplugged out of dongle
- */
- msm_dp_ctrl_off_link_stream(dp->ctrl);
- } else {
- /*
- * unplugged interrupt
- * dongle unplugged out of DUT
- */
- msm_dp_ctrl_off(dp->ctrl);
+ msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl);
+
+ /* dongle is still connected but sinks are disconnected */
+ if (dp->link->sink_count == 0)
+ msm_dp_ctrl_psm_config(dp->ctrl);
+
+ msm_dp_ctrl_off(dp->ctrl);
+
+ /* re-init the PHY so that we can listen to Dongle disconnect */
+ if (dp->link->sink_count == 0)
+ msm_dp_ctrl_reinit_phy(dp->ctrl);
+ else
msm_dp_display_host_phy_exit(dp);
- }
msm_dp_display->power_on = false;
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 08/45] drm/msm/dp: re-arrange dp_display_disable() into functional parts
2024-12-06 4:31 ` [PATCH 08/45] drm/msm/dp: re-arrange dp_display_disable() into functional parts Abhinav Kumar
@ 2024-12-08 12:03 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 12:03 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:39PM -0800, Abhinav Kumar wrote:
> dp_display_disable() handles special case of when monitor is
> disconnected from the dongle while the dongle stays connected
> thereby needing a separate function dp_ctrl_off_link_stream()
> for this. However with a slight rework this can still be handled
> by keeping common paths same for regular and special case.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 29 +++++++++++++++--------------
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 4 ++++
> drivers/gpu/drm/msm/dp/dp_display.c | 25 ++++++++++++-------------
> 3 files changed, 31 insertions(+), 27 deletions(-)
>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Nevertheless,
> @@ -905,20 +905,19 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> if (!msm_dp_display->power_on)
> return 0;
>
> - if (dp->link->sink_count == 0) {
> - /*
> - * irq_hpd with sink_count = 0
> - * hdmi unplugged out of dongle
> - */
> - msm_dp_ctrl_off_link_stream(dp->ctrl);
> - } else {
> - /*
> - * unplugged interrupt
> - * dongle unplugged out of DUT
> - */
> - msm_dp_ctrl_off(dp->ctrl);
> + msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl);
> +
> + /* dongle is still connected but sinks are disconnected */
> + if (dp->link->sink_count == 0)
> + msm_dp_ctrl_psm_config(dp->ctrl);
> +
> + msm_dp_ctrl_off(dp->ctrl);
> +
> + /* re-init the PHY so that we can listen to Dongle disconnect */
> + if (dp->link->sink_count == 0)
It might be better to have just two codepaths:
if (sink_count == 0) {
msm_dp_ctrl_clear_vsc_sdp_pkt()
msm_dp_ctrl_psm_config()
msm_dp_ctrl_off()
msm_dp_ctrl_reinit_phy()
} else {
msm_dp_ctrl_clear_vsc_sdp_pkt()
msm_dp_ctrl_off()
msm_dp_display_host_phy_exit()
}
> + msm_dp_ctrl_reinit_phy(dp->ctrl);
> + else
> msm_dp_display_host_phy_exit(dp);
> - }
>
> msm_dp_display->power_on = false;
>
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 09/45] drm/msm/dp: allow dp_ctrl stream APIs to use any panel passed to it
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (7 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 08/45] drm/msm/dp: re-arrange dp_display_disable() into functional parts Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 9:09 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 10/45] drm/msm/dp: move the pixel clock control to its own API Abhinav Kumar
` (37 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Currently, the dp_ctrl stream APIs operate on their own dp_panel
which is cached inside the dp_ctrl's private struct. However with MST,
the cached panel represents the fixed link and not the sinks which
are hotplugged. Allow the stream related APIs to work on the panel
which is passed to them rather than the cached one. For SST cases,
this shall continue to use the cached dp_panel.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 37 ++++++++++++++++++++-----------------
drivers/gpu/drm/msm/dp/dp_ctrl.h | 5 +++--
drivers/gpu/drm/msm/dp/dp_display.c | 4 ++--
3 files changed, 25 insertions(+), 21 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 9e08996be0cb969cb96d9a3019c445ab4dfc92ef..0bed85b5c8e8133ffa8c74d5de22668905396d09 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -134,7 +134,8 @@ void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl)
drm_dbg_dp(ctrl->drm_dev, "mainlink off\n");
}
-static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
+static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl,
+ struct msm_dp_panel *msm_dp_panel)
{
u32 config = 0, tbd;
const u8 *dpcd = ctrl->panel->dpcd;
@@ -142,7 +143,7 @@ static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
/* Default-> LSCLK DIV: 1/4 LCLK */
config |= (2 << DP_CONFIGURATION_CTRL_LSCLK_DIV_SHIFT);
- if (ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
+ if (msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420)
config |= DP_CONFIGURATION_CTRL_RGB_YUV; /* YUV420 */
/* Scrambler reset enable */
@@ -150,7 +151,7 @@ static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
config |= DP_CONFIGURATION_CTRL_ASSR;
tbd = msm_dp_link_get_test_bits_depth(ctrl->link,
- ctrl->panel->msm_dp_mode.bpp);
+ msm_dp_panel->msm_dp_mode.bpp);
config |= tbd << DP_CONFIGURATION_CTRL_BPC_SHIFT;
@@ -173,20 +174,21 @@ static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
msm_dp_catalog_ctrl_config_ctrl(ctrl->catalog, config);
}
-static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl)
+static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl,
+ struct msm_dp_panel *msm_dp_panel)
{
u32 cc, tb;
msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
- msm_dp_ctrl_config_ctrl(ctrl);
+ msm_dp_ctrl_config_ctrl(ctrl, msm_dp_panel);
tb = msm_dp_link_get_test_bits_depth(ctrl->link,
- ctrl->panel->msm_dp_mode.bpp);
+ msm_dp_panel->msm_dp_mode.bpp);
cc = msm_dp_link_get_colorimetry_config(ctrl->link);
msm_dp_catalog_ctrl_config_misc(ctrl->catalog, cc, tb);
- msm_dp_panel_timing_cfg(ctrl->panel);
+ msm_dp_panel_timing_cfg(msm_dp_panel);
}
/*
@@ -1279,7 +1281,7 @@ static int msm_dp_ctrl_link_train(struct msm_dp_ctrl_private *ctrl,
u8 assr;
struct msm_dp_link_info link_info = {0};
- msm_dp_ctrl_config_ctrl(ctrl);
+ msm_dp_ctrl_config_ctrl(ctrl, ctrl->panel);
link_info.num_lanes = ctrl->link->link_params.num_lanes;
link_info.rate = ctrl->link->link_params.rate;
@@ -1696,7 +1698,8 @@ static bool msm_dp_ctrl_send_phy_test_pattern(struct msm_dp_ctrl_private *ctrl)
return success;
}
-static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl)
+static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl,
+ struct msm_dp_panel *msm_dp_panel)
{
int ret;
unsigned long pixel_rate;
@@ -1720,7 +1723,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
return ret;
}
- pixel_rate = ctrl->panel->msm_dp_mode.drm_mode.clock;
+ pixel_rate = msm_dp_panel->msm_dp_mode.drm_mode.clock;
ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
if (ret) {
DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
@@ -1758,7 +1761,7 @@ void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl)
if (sink_request & DP_TEST_LINK_PHY_TEST_PATTERN) {
drm_dbg_dp(ctrl->drm_dev, "PHY_TEST_PATTERN request\n");
- if (msm_dp_ctrl_process_phy_test_request(ctrl)) {
+ if (msm_dp_ctrl_process_phy_test_request(ctrl, ctrl->panel)) {
DRM_ERROR("process phy_test_req failed\n");
return;
}
@@ -1976,7 +1979,7 @@ int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_li
return ret;
}
-int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
+int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
{
int ret = 0;
bool mainlink_ready = false;
@@ -1989,9 +1992,9 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
- pixel_rate = pixel_rate_orig = ctrl->panel->msm_dp_mode.drm_mode.clock;
+ pixel_rate = pixel_rate_orig = msm_dp_panel->msm_dp_mode.drm_mode.clock;
- if (msm_dp_ctrl->wide_bus_en || ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
+ if (msm_dp_ctrl->wide_bus_en || msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420)
pixel_rate >>= 1;
drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
@@ -2019,12 +2022,12 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
*/
reinit_completion(&ctrl->video_comp);
- msm_dp_ctrl_configure_source_params(ctrl);
+ msm_dp_ctrl_configure_source_params(ctrl, msm_dp_panel);
msm_dp_catalog_ctrl_config_msa(ctrl->catalog,
ctrl->link->link_params.rate,
pixel_rate_orig,
- ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
+ msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420);
msm_dp_ctrl_setup_tr_unit(ctrl);
@@ -2042,7 +2045,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
return ret;
}
-void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl)
+void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *dp_panel)
{
struct msm_dp_ctrl_private *ctrl;
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index 0f58b63c5c7c5aab43c0db2a697ba491959b79d2..547155ffa50fbe2f3a1f2c2e1ee17420daf0f3da 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -18,7 +18,7 @@ struct msm_dp_ctrl {
struct phy;
int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl);
-int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl);
+int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
@@ -42,7 +42,8 @@ void msm_dp_ctrl_config_psr(struct msm_dp_ctrl *msm_dp_ctrl);
int msm_dp_ctrl_core_clk_enable(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_core_clk_disable(struct msm_dp_ctrl *msm_dp_ctrl);
-void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl);
+void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl,
+ struct msm_dp_panel *msm_dp_panel);
void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index bbce8ca09ff70059458231982f002e1f22d2c3ab..c059f749c1f204deac9dfb0c56f537f5545d9acb 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -852,7 +852,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp)
return 0;
}
- rc = msm_dp_ctrl_on_stream(dp->ctrl);
+ rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel);
if (!rc)
msm_dp_display->power_on = true;
@@ -905,7 +905,7 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
if (!msm_dp_display->power_on)
return 0;
- msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl);
+ msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, dp->panel);
/* dongle is still connected but sinks are disconnected */
if (dp->link->sink_count == 0)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 09/45] drm/msm/dp: allow dp_ctrl stream APIs to use any panel passed to it
2024-12-06 4:31 ` [PATCH 09/45] drm/msm/dp: allow dp_ctrl stream APIs to use any panel passed to it Abhinav Kumar
@ 2024-12-06 9:09 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:09 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:40PM -0800, Abhinav Kumar wrote:
> Currently, the dp_ctrl stream APIs operate on their own dp_panel
> which is cached inside the dp_ctrl's private struct. However with MST,
> the cached panel represents the fixed link and not the sinks which
> are hotplugged. Allow the stream related APIs to work on the panel
> which is passed to them rather than the cached one. For SST cases,
> this shall continue to use the cached dp_panel.
"cached inside the dp_ctrl's private struct" usually means that dp_ctrl
stores a copy of the msm_dp_panel, not just a pointer to it, so the
commit message needs rephrasing.
Also if the stored pointer is (or will be) no longer applicable, can we
drop it completely and always pass the dp_panel by reference? Otherwise
it's hard to know when to pass a new one and when to use ctrl->panel
instead.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 37 ++++++++++++++++++++-----------------
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 5 +++--
> drivers/gpu/drm/msm/dp/dp_display.c | 4 ++--
> 3 files changed, 25 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 9e08996be0cb969cb96d9a3019c445ab4dfc92ef..0bed85b5c8e8133ffa8c74d5de22668905396d09 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -134,7 +134,8 @@ void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl)
> drm_dbg_dp(ctrl->drm_dev, "mainlink off\n");
> }
>
> -static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
> +static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl,
> + struct msm_dp_panel *msm_dp_panel)
> {
> u32 config = 0, tbd;
> const u8 *dpcd = ctrl->panel->dpcd;
> @@ -142,7 +143,7 @@ static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
> /* Default-> LSCLK DIV: 1/4 LCLK */
> config |= (2 << DP_CONFIGURATION_CTRL_LSCLK_DIV_SHIFT);
>
> - if (ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
> + if (msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420)
> config |= DP_CONFIGURATION_CTRL_RGB_YUV; /* YUV420 */
>
> /* Scrambler reset enable */
> @@ -150,7 +151,7 @@ static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
> config |= DP_CONFIGURATION_CTRL_ASSR;
>
> tbd = msm_dp_link_get_test_bits_depth(ctrl->link,
> - ctrl->panel->msm_dp_mode.bpp);
> + msm_dp_panel->msm_dp_mode.bpp);
>
> config |= tbd << DP_CONFIGURATION_CTRL_BPC_SHIFT;
>
> @@ -173,20 +174,21 @@ static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl)
> msm_dp_catalog_ctrl_config_ctrl(ctrl->catalog, config);
> }
>
> -static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl)
> +static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl,
> + struct msm_dp_panel *msm_dp_panel)
> {
> u32 cc, tb;
>
> msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
> msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
>
> - msm_dp_ctrl_config_ctrl(ctrl);
> + msm_dp_ctrl_config_ctrl(ctrl, msm_dp_panel);
>
> tb = msm_dp_link_get_test_bits_depth(ctrl->link,
> - ctrl->panel->msm_dp_mode.bpp);
> + msm_dp_panel->msm_dp_mode.bpp);
> cc = msm_dp_link_get_colorimetry_config(ctrl->link);
> msm_dp_catalog_ctrl_config_misc(ctrl->catalog, cc, tb);
> - msm_dp_panel_timing_cfg(ctrl->panel);
> + msm_dp_panel_timing_cfg(msm_dp_panel);
> }
>
> /*
> @@ -1279,7 +1281,7 @@ static int msm_dp_ctrl_link_train(struct msm_dp_ctrl_private *ctrl,
> u8 assr;
> struct msm_dp_link_info link_info = {0};
>
> - msm_dp_ctrl_config_ctrl(ctrl);
> + msm_dp_ctrl_config_ctrl(ctrl, ctrl->panel);
>
> link_info.num_lanes = ctrl->link->link_params.num_lanes;
> link_info.rate = ctrl->link->link_params.rate;
> @@ -1696,7 +1698,8 @@ static bool msm_dp_ctrl_send_phy_test_pattern(struct msm_dp_ctrl_private *ctrl)
> return success;
> }
>
> -static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl)
> +static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl,
> + struct msm_dp_panel *msm_dp_panel)
> {
> int ret;
> unsigned long pixel_rate;
> @@ -1720,7 +1723,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
> return ret;
> }
>
> - pixel_rate = ctrl->panel->msm_dp_mode.drm_mode.clock;
> + pixel_rate = msm_dp_panel->msm_dp_mode.drm_mode.clock;
> ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
> if (ret) {
> DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
> @@ -1758,7 +1761,7 @@ void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl)
>
> if (sink_request & DP_TEST_LINK_PHY_TEST_PATTERN) {
> drm_dbg_dp(ctrl->drm_dev, "PHY_TEST_PATTERN request\n");
> - if (msm_dp_ctrl_process_phy_test_request(ctrl)) {
> + if (msm_dp_ctrl_process_phy_test_request(ctrl, ctrl->panel)) {
> DRM_ERROR("process phy_test_req failed\n");
> return;
> }
> @@ -1976,7 +1979,7 @@ int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_li
> return ret;
> }
>
> -int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
> +int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
> {
> int ret = 0;
> bool mainlink_ready = false;
> @@ -1989,9 +1992,9 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
>
> ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
>
> - pixel_rate = pixel_rate_orig = ctrl->panel->msm_dp_mode.drm_mode.clock;
> + pixel_rate = pixel_rate_orig = msm_dp_panel->msm_dp_mode.drm_mode.clock;
>
> - if (msm_dp_ctrl->wide_bus_en || ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420)
> + if (msm_dp_ctrl->wide_bus_en || msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420)
> pixel_rate >>= 1;
>
> drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
> @@ -2019,12 +2022,12 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
> */
> reinit_completion(&ctrl->video_comp);
>
> - msm_dp_ctrl_configure_source_params(ctrl);
> + msm_dp_ctrl_configure_source_params(ctrl, msm_dp_panel);
>
> msm_dp_catalog_ctrl_config_msa(ctrl->catalog,
> ctrl->link->link_params.rate,
> pixel_rate_orig,
> - ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
> + msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420);
>
> msm_dp_ctrl_setup_tr_unit(ctrl);
>
> @@ -2042,7 +2045,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl)
> return ret;
> }
>
> -void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl)
> +void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *dp_panel)
> {
> struct msm_dp_ctrl_private *ctrl;
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index 0f58b63c5c7c5aab43c0db2a697ba491959b79d2..547155ffa50fbe2f3a1f2c2e1ee17420daf0f3da 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -18,7 +18,7 @@ struct msm_dp_ctrl {
> struct phy;
>
> int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl);
> -int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> +int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
> int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
> void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
> @@ -42,7 +42,8 @@ void msm_dp_ctrl_config_psr(struct msm_dp_ctrl *msm_dp_ctrl);
> int msm_dp_ctrl_core_clk_enable(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_core_clk_disable(struct msm_dp_ctrl *msm_dp_ctrl);
>
> -void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl);
> +void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl,
> + struct msm_dp_panel *msm_dp_panel);
> void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl);
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index bbce8ca09ff70059458231982f002e1f22d2c3ab..c059f749c1f204deac9dfb0c56f537f5545d9acb 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -852,7 +852,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp)
> return 0;
> }
>
> - rc = msm_dp_ctrl_on_stream(dp->ctrl);
> + rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel);
> if (!rc)
> msm_dp_display->power_on = true;
>
> @@ -905,7 +905,7 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> if (!msm_dp_display->power_on)
> return 0;
>
> - msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl);
> + msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, dp->panel);
>
> /* dongle is still connected but sinks are disconnected */
> if (dp->link->sink_count == 0)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 10/45] drm/msm/dp: move the pixel clock control to its own API
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (8 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 09/45] drm/msm/dp: allow dp_ctrl stream APIs to use any panel passed to it Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 9:11 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 11/45] drm/msm/dp: split dp_ctrl_off() into stream and link parts Abhinav Kumar
` (36 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Enable/Disable of DP pixel clock happens in multiple code paths
leading to code duplication. Move it into individual helpers so that
the helpers can be called wherever necessary.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 76 ++++++++++++++++++++++------------------
1 file changed, 41 insertions(+), 35 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 0bed85b5c8e8133ffa8c74d5de22668905396d09..118f5ed83e464f9f27f813eb39624f9c3189f5ac 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -1698,6 +1698,30 @@ static bool msm_dp_ctrl_send_phy_test_pattern(struct msm_dp_ctrl_private *ctrl)
return success;
}
+static int msm_dp_ctrl_stream_clk_on(struct msm_dp_ctrl_private *ctrl, unsigned long pixel_rate)
+{
+ int ret;
+
+ ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
+ if (ret) {
+ DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
+ return ret;
+ }
+
+ if (ctrl->stream_clks_on) {
+ drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
+ } else {
+ ret = clk_prepare_enable(ctrl->pixel_clk);
+ if (ret) {
+ DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
+ return ret;
+ }
+ ctrl->stream_clks_on = true;
+ }
+
+ return ret;
+}
+
static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl,
struct msm_dp_panel *msm_dp_panel)
{
@@ -1724,22 +1748,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
}
pixel_rate = msm_dp_panel->msm_dp_mode.drm_mode.clock;
- ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
- if (ret) {
- DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
- return ret;
- }
-
- if (ctrl->stream_clks_on) {
- drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
- } else {
- ret = clk_prepare_enable(ctrl->pixel_clk);
- if (ret) {
- DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
- return ret;
- }
- ctrl->stream_clks_on = true;
- }
+ ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
msm_dp_ctrl_send_phy_test_pattern(ctrl);
@@ -1999,21 +2008,10 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
- ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
+ ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
if (ret) {
- DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
- goto end;
- }
-
- if (ctrl->stream_clks_on) {
- drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
- } else {
- ret = clk_prepare_enable(ctrl->pixel_clk);
- if (ret) {
- DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
- goto end;
- }
- ctrl->stream_clks_on = true;
+ DRM_ERROR("failed to enable stream pixel clk\n");
+ return ret;
}
/*
@@ -2041,10 +2039,21 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
drm_dbg_dp(ctrl->drm_dev,
"mainlink %s\n", mainlink_ready ? "READY" : "NOT READY");
-end:
return ret;
}
+static void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
+{
+ struct msm_dp_ctrl_private *ctrl;
+
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+
+ if (ctrl->stream_clks_on) {
+ clk_disable_unprepare(ctrl->pixel_clk);
+ ctrl->stream_clks_on = false;
+ }
+}
+
void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *dp_panel)
{
struct msm_dp_ctrl_private *ctrl;
@@ -2115,10 +2124,7 @@ void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl)
msm_dp_catalog_ctrl_reset(ctrl->catalog);
- if (ctrl->stream_clks_on) {
- clk_disable_unprepare(ctrl->pixel_clk);
- ctrl->stream_clks_on = false;
- }
+ msm_dp_ctrl_stream_clk_off(msm_dp_ctrl);
dev_pm_opp_set_rate(ctrl->dev, 0);
msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 10/45] drm/msm/dp: move the pixel clock control to its own API
2024-12-06 4:31 ` [PATCH 10/45] drm/msm/dp: move the pixel clock control to its own API Abhinav Kumar
@ 2024-12-06 9:11 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:11 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:41PM -0800, Abhinav Kumar wrote:
> Enable/Disable of DP pixel clock happens in multiple code paths
> leading to code duplication. Move it into individual helpers so that
> the helpers can be called wherever necessary.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 76 ++++++++++++++++++++++------------------
> 1 file changed, 41 insertions(+), 35 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 0bed85b5c8e8133ffa8c74d5de22668905396d09..118f5ed83e464f9f27f813eb39624f9c3189f5ac 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -1698,6 +1698,30 @@ static bool msm_dp_ctrl_send_phy_test_pattern(struct msm_dp_ctrl_private *ctrl)
> return success;
> }
>
> +static int msm_dp_ctrl_stream_clk_on(struct msm_dp_ctrl_private *ctrl, unsigned long pixel_rate)
> +{
> + int ret;
> +
> + ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
> + if (ret) {
> + DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
> + return ret;
> + }
> +
> + if (ctrl->stream_clks_on) {
> + drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
> + } else {
> + ret = clk_prepare_enable(ctrl->pixel_clk);
> + if (ret) {
> + DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
> + return ret;
> + }
> + ctrl->stream_clks_on = true;
> + }
> +
> + return ret;
> +}
> +
> static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl,
> struct msm_dp_panel *msm_dp_panel)
> {
> @@ -1724,22 +1748,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
> }
>
> pixel_rate = msm_dp_panel->msm_dp_mode.drm_mode.clock;
> - ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
> - if (ret) {
> - DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
> - return ret;
> - }
> -
> - if (ctrl->stream_clks_on) {
> - drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
> - } else {
> - ret = clk_prepare_enable(ctrl->pixel_clk);
> - if (ret) {
> - DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
> - return ret;
> - }
> - ctrl->stream_clks_on = true;
> - }
> + ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
>
> msm_dp_ctrl_send_phy_test_pattern(ctrl);
>
> @@ -1999,21 +2008,10 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
>
> drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
>
> - ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
> + ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
> if (ret) {
> - DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
> - goto end;
> - }
> -
> - if (ctrl->stream_clks_on) {
> - drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
> - } else {
> - ret = clk_prepare_enable(ctrl->pixel_clk);
> - if (ret) {
> - DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
> - goto end;
> - }
> - ctrl->stream_clks_on = true;
> + DRM_ERROR("failed to enable stream pixel clk\n");
> + return ret;
> }
>
> /*
> @@ -2041,10 +2039,21 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> drm_dbg_dp(ctrl->drm_dev,
> "mainlink %s\n", mainlink_ready ? "READY" : "NOT READY");
>
> -end:
> return ret;
> }
>
> +static void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
At least this should come right after msm_dp_ctrl_stream_clk_on().
> +{
> + struct msm_dp_ctrl_private *ctrl;
> +
> + ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> +
> + if (ctrl->stream_clks_on) {
> + clk_disable_unprepare(ctrl->pixel_clk);
> + ctrl->stream_clks_on = false;
> + }
> +}
> +
> void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *dp_panel)
> {
> struct msm_dp_ctrl_private *ctrl;
> @@ -2115,10 +2124,7 @@ void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl)
>
> msm_dp_catalog_ctrl_reset(ctrl->catalog);
>
> - if (ctrl->stream_clks_on) {
> - clk_disable_unprepare(ctrl->pixel_clk);
> - ctrl->stream_clks_on = false;
> - }
> + msm_dp_ctrl_stream_clk_off(msm_dp_ctrl);
>
> dev_pm_opp_set_rate(ctrl->dev, 0);
> msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 11/45] drm/msm/dp: split dp_ctrl_off() into stream and link parts
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (9 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 10/45] drm/msm/dp: move the pixel clock control to its own API Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 9:14 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 12/45] drm/msm/dp: make bridge helpers use dp_display to allow re-use Abhinav Kumar
` (35 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Split dp_ctrl_off() into stream and link parts so that for MST
cases we can control the link and pixel parts separately.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 29 +++--------------------------
drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 +-
drivers/gpu/drm/msm/dp/dp_display.c | 4 +++-
3 files changed, 7 insertions(+), 28 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 118f5ed83e464f9f27f813eb39624f9c3189f5ac..485339eb998cc6c8c1e8ab0a88b5c5d6ef300a1f 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -1739,7 +1739,8 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
* running. Add the global reset just before disabling the
* link clocks and core clocks.
*/
- msm_dp_ctrl_off(&ctrl->msm_dp_ctrl);
+ msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl);
+ msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
if (ret) {
@@ -2042,7 +2043,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
return ret;
}
-static void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
+void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
{
struct msm_dp_ctrl_private *ctrl;
@@ -2110,30 +2111,6 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
phy, phy->init_count, phy->power_count);
}
-void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl)
-{
- struct msm_dp_ctrl_private *ctrl;
- struct phy *phy;
-
- ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
- phy = ctrl->phy;
-
- msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
-
- msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
-
- msm_dp_catalog_ctrl_reset(ctrl->catalog);
-
- msm_dp_ctrl_stream_clk_off(msm_dp_ctrl);
-
- dev_pm_opp_set_rate(ctrl->dev, 0);
- msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
-
- phy_power_off(phy);
- drm_dbg_dp(ctrl->drm_dev, "phy=%p init=%d power_on=%d\n",
- phy, phy->init_count, phy->power_count);
-}
-
irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
{
struct msm_dp_ctrl_private *ctrl;
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index 547155ffa50fbe2f3a1f2c2e1ee17420daf0f3da..887cf5a866f07cb9038887a0634d3e1a0375879c 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -22,7 +22,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
-void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl);
+void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl);
irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index c059f749c1f204deac9dfb0c56f537f5545d9acb..b0458bbc89e934ca33ed5af3f2a8ebca30b50824 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -911,7 +911,9 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
if (dp->link->sink_count == 0)
msm_dp_ctrl_psm_config(dp->ctrl);
- msm_dp_ctrl_off(dp->ctrl);
+ msm_dp_ctrl_stream_clk_off(dp->ctrl);
+
+ msm_dp_ctrl_off_link(dp->ctrl);
/* re-init the PHY so that we can listen to Dongle disconnect */
if (dp->link->sink_count == 0)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 11/45] drm/msm/dp: split dp_ctrl_off() into stream and link parts
2024-12-06 4:31 ` [PATCH 11/45] drm/msm/dp: split dp_ctrl_off() into stream and link parts Abhinav Kumar
@ 2024-12-06 9:14 ` Dmitry Baryshkov
2025-05-26 12:47 ` Yongxing Mou
0 siblings, 1 reply; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:14 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:42PM -0800, Abhinav Kumar wrote:
> Split dp_ctrl_off() into stream and link parts so that for MST
> cases we can control the link and pixel parts separately.
Please start by describing the problem.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 29 +++--------------------------
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 +-
> drivers/gpu/drm/msm/dp/dp_display.c | 4 +++-
> 3 files changed, 7 insertions(+), 28 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 118f5ed83e464f9f27f813eb39624f9c3189f5ac..485339eb998cc6c8c1e8ab0a88b5c5d6ef300a1f 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -1739,7 +1739,8 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
> * running. Add the global reset just before disabling the
> * link clocks and core clocks.
> */
> - msm_dp_ctrl_off(&ctrl->msm_dp_ctrl);
> + msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl);
> + msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
Huh? What happened with the rest of the msm_dp_ctrl_off() code sequence?
It got dropped, but the commit message tells nothing about it.
>
> ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
> if (ret) {
> @@ -2042,7 +2043,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> return ret;
> }
>
> -static void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
> +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
> {
> struct msm_dp_ctrl_private *ctrl;
>
> @@ -2110,30 +2111,6 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
> phy, phy->init_count, phy->power_count);
> }
>
> -void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl)
> -{
> - struct msm_dp_ctrl_private *ctrl;
> - struct phy *phy;
> -
> - ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> - phy = ctrl->phy;
> -
> - msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
> -
> - msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
> -
> - msm_dp_catalog_ctrl_reset(ctrl->catalog);
> -
> - msm_dp_ctrl_stream_clk_off(msm_dp_ctrl);
> -
> - dev_pm_opp_set_rate(ctrl->dev, 0);
> - msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
> -
> - phy_power_off(phy);
> - drm_dbg_dp(ctrl->drm_dev, "phy=%p init=%d power_on=%d\n",
> - phy, phy->init_count, phy->power_count);
> -}
> -
> irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
> {
> struct msm_dp_ctrl_private *ctrl;
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index 547155ffa50fbe2f3a1f2c2e1ee17420daf0f3da..887cf5a866f07cb9038887a0634d3e1a0375879c 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -22,7 +22,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
> void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
> -void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl);
> +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl);
> irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl);
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index c059f749c1f204deac9dfb0c56f537f5545d9acb..b0458bbc89e934ca33ed5af3f2a8ebca30b50824 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -911,7 +911,9 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> if (dp->link->sink_count == 0)
> msm_dp_ctrl_psm_config(dp->ctrl);
>
> - msm_dp_ctrl_off(dp->ctrl);
> + msm_dp_ctrl_stream_clk_off(dp->ctrl);
> +
> + msm_dp_ctrl_off_link(dp->ctrl);
>
> /* re-init the PHY so that we can listen to Dongle disconnect */
> if (dp->link->sink_count == 0)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 11/45] drm/msm/dp: split dp_ctrl_off() into stream and link parts
2024-12-06 9:14 ` Dmitry Baryshkov
@ 2025-05-26 12:47 ` Yongxing Mou
2025-05-28 16:15 ` Dmitry Baryshkov
0 siblings, 1 reply; 111+ messages in thread
From: Yongxing Mou @ 2025-05-26 12:47 UTC (permalink / raw)
To: Dmitry Baryshkov, Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On 2024/12/6 17:14, Dmitry Baryshkov wrote:
> On Thu, Dec 05, 2024 at 08:31:42PM -0800, Abhinav Kumar wrote:
>> Split dp_ctrl_off() into stream and link parts so that for MST
>> cases we can control the link and pixel parts separately.
>
> Please start by describing the problem.
Got it.
>
>>
>> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
>> ---
>> drivers/gpu/drm/msm/dp/dp_ctrl.c | 29 +++--------------------------
>> drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 +-
>> drivers/gpu/drm/msm/dp/dp_display.c | 4 +++-
>> 3 files changed, 7 insertions(+), 28 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
>> index 118f5ed83e464f9f27f813eb39624f9c3189f5ac..485339eb998cc6c8c1e8ab0a88b5c5d6ef300a1f 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
>> @@ -1739,7 +1739,8 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
>> * running. Add the global reset just before disabling the
>> * link clocks and core clocks.
>> */
>> - msm_dp_ctrl_off(&ctrl->msm_dp_ctrl);
>> + msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl);
>> + msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
>
> Huh? What happened with the rest of the msm_dp_ctrl_off() code sequence?
> It got dropped, but the commit message tells nothing about it.
>
The function msm_dp_ctrl_off has been split into two parts,
stream_clk_off and off_link, so it got dropped. This part is a bit
confusing, will make it clearer.
>>
>> ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
>> if (ret) {
>> @@ -2042,7 +2043,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
>> return ret;
>> }
>>
>> -static void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
>> +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
>> {
>> struct msm_dp_ctrl_private *ctrl;
>>
>> @@ -2110,30 +2111,6 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
>> phy, phy->init_count, phy->power_count);
>> }
>>
>> -void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl)
>> -{
>> - struct msm_dp_ctrl_private *ctrl;
>> - struct phy *phy;
>> -
>> - ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
>> - phy = ctrl->phy;
>> -
>> - msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
>> -
>> - msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
>> -
>> - msm_dp_catalog_ctrl_reset(ctrl->catalog);
>> -
>> - msm_dp_ctrl_stream_clk_off(msm_dp_ctrl);
>> -
>> - dev_pm_opp_set_rate(ctrl->dev, 0);
>> - msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
>> -
>> - phy_power_off(phy);
>> - drm_dbg_dp(ctrl->drm_dev, "phy=%p init=%d power_on=%d\n",
>> - phy, phy->init_count, phy->power_count);
>> -}
>> -
>> irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
>> {
>> struct msm_dp_ctrl_private *ctrl;
>> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
>> index 547155ffa50fbe2f3a1f2c2e1ee17420daf0f3da..887cf5a866f07cb9038887a0634d3e1a0375879c 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
>> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
>> @@ -22,7 +22,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
>> int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
>> void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
>> void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
>> -void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl);
>> +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl);
>> void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl);
>> irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl);
>> void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl);
>> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
>> index c059f749c1f204deac9dfb0c56f537f5545d9acb..b0458bbc89e934ca33ed5af3f2a8ebca30b50824 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_display.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
>> @@ -911,7 +911,9 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
>> if (dp->link->sink_count == 0)
>> msm_dp_ctrl_psm_config(dp->ctrl);
>>
>> - msm_dp_ctrl_off(dp->ctrl);
>> + msm_dp_ctrl_stream_clk_off(dp->ctrl);
>> +
>> + msm_dp_ctrl_off_link(dp->ctrl);
>>
>> /* re-init the PHY so that we can listen to Dongle disconnect */
>> if (dp->link->sink_count == 0)
>>
>> --
>> 2.34.1
>>
>
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 11/45] drm/msm/dp: split dp_ctrl_off() into stream and link parts
2025-05-26 12:47 ` Yongxing Mou
@ 2025-05-28 16:15 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2025-05-28 16:15 UTC (permalink / raw)
To: Yongxing Mou
Cc: Abhinav Kumar, Rob Clark, Sean Paul, Marijn Suijten, David Airlie,
Simona Vetter, Stephen Boyd, Chandan Uddaraju, Guenter Roeck,
Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On Mon, May 26, 2025 at 08:47:22PM +0800, Yongxing Mou wrote:
>
>
> On 2024/12/6 17:14, Dmitry Baryshkov wrote:
> > On Thu, Dec 05, 2024 at 08:31:42PM -0800, Abhinav Kumar wrote:
> > > Split dp_ctrl_off() into stream and link parts so that for MST
> > > cases we can control the link and pixel parts separately.
> >
> > Please start by describing the problem.
> Got it.
> >
> > >
> > > Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> > > ---
> > > drivers/gpu/drm/msm/dp/dp_ctrl.c | 29 +++--------------------------
> > > drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 +-
> > > drivers/gpu/drm/msm/dp/dp_display.c | 4 +++-
> > > 3 files changed, 7 insertions(+), 28 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> > > index 118f5ed83e464f9f27f813eb39624f9c3189f5ac..485339eb998cc6c8c1e8ab0a88b5c5d6ef300a1f 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> > > +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> > > @@ -1739,7 +1739,8 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
> > > * running. Add the global reset just before disabling the
> > > * link clocks and core clocks.
> > > */
> > > - msm_dp_ctrl_off(&ctrl->msm_dp_ctrl);
> > > + msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl);
> > > + msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
> >
> > Huh? What happened with the rest of the msm_dp_ctrl_off() code sequence?
> > It got dropped, but the commit message tells nothing about it.
> >
> The function msm_dp_ctrl_off has been split into two parts, stream_clk_off
> and off_link, so it got dropped. This part is a bit confusing, will make it
> clearer.
Then msm_dp_ctrl_off_link() should be a part of this patch.
> > > ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
> > > if (ret) {
> > > @@ -2042,7 +2043,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> > > return ret;
> > > }
> > > -static void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
> > > +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
> > > {
> > > struct msm_dp_ctrl_private *ctrl;
> > > @@ -2110,30 +2111,6 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
> > > phy, phy->init_count, phy->power_count);
> > > }
> > > -void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl)
> > > -{
> > > - struct msm_dp_ctrl_private *ctrl;
> > > - struct phy *phy;
> > > -
> > > - ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> > > - phy = ctrl->phy;
> > > -
> > > - msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
> > > -
> > > - msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
> > > -
> > > - msm_dp_catalog_ctrl_reset(ctrl->catalog);
> > > -
> > > - msm_dp_ctrl_stream_clk_off(msm_dp_ctrl);
> > > -
> > > - dev_pm_opp_set_rate(ctrl->dev, 0);
> > > - msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
> > > -
> > > - phy_power_off(phy);
> > > - drm_dbg_dp(ctrl->drm_dev, "phy=%p init=%d power_on=%d\n",
> > > - phy, phy->init_count, phy->power_count);
> > > -}
> > > -
> > > irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
> > > {
> > > struct msm_dp_ctrl_private *ctrl;
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> > > index 547155ffa50fbe2f3a1f2c2e1ee17420daf0f3da..887cf5a866f07cb9038887a0634d3e1a0375879c 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> > > +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> > > @@ -22,7 +22,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> > > int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
> > > void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> > > void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
> > > -void msm_dp_ctrl_off(struct msm_dp_ctrl *msm_dp_ctrl);
> > > +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl);
> > > void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl);
> > > irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl);
> > > void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl);
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> > > index c059f749c1f204deac9dfb0c56f537f5545d9acb..b0458bbc89e934ca33ed5af3f2a8ebca30b50824 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_display.c
> > > +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> > > @@ -911,7 +911,9 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> > > if (dp->link->sink_count == 0)
> > > msm_dp_ctrl_psm_config(dp->ctrl);
> > > - msm_dp_ctrl_off(dp->ctrl);
> > > + msm_dp_ctrl_stream_clk_off(dp->ctrl);
> > > +
> > > + msm_dp_ctrl_off_link(dp->ctrl);
> > > /* re-init the PHY so that we can listen to Dongle disconnect */
> > > if (dp->link->sink_count == 0)
> > >
> > > --
> > > 2.34.1
> > >
> >
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 12/45] drm/msm/dp: make bridge helpers use dp_display to allow re-use
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (10 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 11/45] drm/msm/dp: split dp_ctrl_off() into stream and link parts Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:07 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 13/45] drm/msm/dp: separate dp_display_prepare() into its own API Abhinav Kumar
` (34 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
dp_bridge helpers take drm_bridge as an input and extract the
dp_display object to be used in the dp_display module. Rather than
doing it in a roundabout way, directly pass the dp_display object
to these helpers so that the MST bridge can also re-use the same
helpers.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 34 ++++++++-----------------
drivers/gpu/drm/msm/dp/dp_display.h | 9 +++++++
drivers/gpu/drm/msm/dp/dp_drm.c | 49 ++++++++++++++++++++++++++++++++++++-
drivers/gpu/drm/msm/dp/dp_drm.h | 12 ---------
4 files changed, 67 insertions(+), 37 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index b0458bbc89e934ca33ed5af3f2a8ebca30b50824..2f8650d60202deaa90de1a5e0dd6d8bc50f09782 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -942,24 +942,21 @@ int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display,
/**
* msm_dp_bridge_mode_valid - callback to determine if specified mode is valid
- * @bridge: Pointer to drm bridge structure
+ * @dp: Pointer to dp display structure
* @info: display info
* @mode: Pointer to drm mode structure
* Returns: Validity status for specified mode
*/
-enum drm_mode_status msm_dp_bridge_mode_valid(struct drm_bridge *bridge,
- const struct drm_display_info *info,
- const struct drm_display_mode *mode)
+enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
+ const struct drm_display_info *info,
+ const struct drm_display_mode *mode)
{
const u32 num_components = 3, default_bpp = 24;
struct msm_dp_display_private *msm_dp_display;
struct msm_dp_link_info *link_info;
u32 mode_rate_khz = 0, supported_rate_khz = 0, mode_bpp = 0;
- struct msm_dp *dp;
int mode_pclk_khz = mode->clock;
- dp = to_dp_bridge(bridge)->msm_dp_display;
-
if (!dp || !mode_pclk_khz || !dp->connector) {
DRM_ERROR("invalid params\n");
return -EINVAL;
@@ -1528,11 +1525,8 @@ int msm_dp_modeset_init(struct msm_dp *msm_dp_display, struct drm_device *dev,
return 0;
}
-void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
- struct drm_bridge_state *old_bridge_state)
+void msm_dp_display_atomic_enable(struct msm_dp *dp)
{
- struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(drm_bridge);
- struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
int rc = 0;
struct msm_dp_display_private *msm_dp_display;
@@ -1569,11 +1563,8 @@ void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
mutex_unlock(&msm_dp_display->event_mutex);
}
-void msm_dp_bridge_atomic_disable(struct drm_bridge *drm_bridge,
- struct drm_bridge_state *old_bridge_state)
+void msm_dp_display_atomic_disable(struct msm_dp *dp)
{
- struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(drm_bridge);
- struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
struct msm_dp_display_private *msm_dp_display;
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
@@ -1595,11 +1586,8 @@ static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
msm_dp_display->prepared = false;
}
-void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
- struct drm_bridge_state *old_bridge_state)
+void msm_dp_display_atomic_post_disable(struct msm_dp *dp)
{
- struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(drm_bridge);
- struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
u32 state;
struct msm_dp_display_private *msm_dp_display;
@@ -1632,12 +1620,10 @@ void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
mutex_unlock(&msm_dp_display->event_mutex);
}
-void msm_dp_bridge_mode_set(struct drm_bridge *drm_bridge,
- const struct drm_display_mode *mode,
- const struct drm_display_mode *adjusted_mode)
+void msm_dp_display_mode_set(struct msm_dp *dp,
+ const struct drm_display_mode *mode,
+ const struct drm_display_mode *adjusted_mode)
{
- struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(drm_bridge);
- struct msm_dp *dp = msm_dp_bridge->msm_dp_display;
struct msm_dp_display_private *msm_dp_display;
struct msm_dp_panel *msm_dp_panel;
struct msm_dp_display_mode msm_dp_mode;
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 0165a80863e65b6eea6d2d8d1c5c08cd51fed101..2a800164cd9c74c29db80dbad15a2dff9fcb93d6 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -41,5 +41,14 @@ void msm_dp_display_signal_audio_start(struct msm_dp *msm_dp_display);
void msm_dp_display_signal_audio_complete(struct msm_dp *msm_dp_display);
void msm_dp_display_set_psr(struct msm_dp *dp, bool enter);
void msm_dp_display_debugfs_init(struct msm_dp *msm_dp_display, struct dentry *dentry, bool is_edp);
+void msm_dp_display_atomic_post_disable(struct msm_dp *dp_display);
+void msm_dp_display_atomic_disable(struct msm_dp *dp_display);
+void msm_dp_display_atomic_enable(struct msm_dp *dp_display);
+void msm_dp_display_mode_set(struct msm_dp *dp,
+ const struct drm_display_mode *mode,
+ const struct drm_display_mode *adjusted_mode);
+enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
+ const struct drm_display_info *info,
+ const struct drm_display_mode *mode);
#endif /* _DP_DISPLAY_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_drm.c b/drivers/gpu/drm/msm/dp/dp_drm.c
index d3e241ea6941615b8e274dd17426c2f8557f09b5..4ef3f16a68890bc220d147ac3d04f53ef911f004 100644
--- a/drivers/gpu/drm/msm/dp/dp_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_drm.c
@@ -97,6 +97,53 @@ static void msm_dp_bridge_debugfs_init(struct drm_bridge *bridge, struct dentry
msm_dp_display_debugfs_init(dp, root, false);
}
+static void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
+ struct drm_bridge_state *old_bridge_state)
+{
+ struct msm_dp_bridge *dp_bridge = to_dp_bridge(drm_bridge);
+ struct msm_dp *dp = dp_bridge->msm_dp_display;
+
+ msm_dp_display_atomic_enable(dp);
+}
+
+static void msm_dp_bridge_atomic_disable(struct drm_bridge *drm_bridge,
+ struct drm_bridge_state *old_bridge_state)
+{
+ struct msm_dp_bridge *dp_bridge = to_dp_bridge(drm_bridge);
+ struct msm_dp *dp = dp_bridge->msm_dp_display;
+
+ msm_dp_display_atomic_disable(dp);
+}
+
+static void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
+ struct drm_bridge_state *old_bridge_state)
+{
+ struct msm_dp_bridge *dp_bridge = to_dp_bridge(drm_bridge);
+ struct msm_dp *dp = dp_bridge->msm_dp_display;
+
+ msm_dp_display_atomic_post_disable(dp);
+}
+
+static void msm_dp_bridge_mode_set(struct drm_bridge *drm_bridge,
+ const struct drm_display_mode *mode,
+ const struct drm_display_mode *adjusted_mode)
+{
+ struct msm_dp_bridge *dp_bridge = to_dp_bridge(drm_bridge);
+ struct msm_dp *dp = dp_bridge->msm_dp_display;
+
+ msm_dp_display_mode_set(dp, mode, adjusted_mode);
+}
+
+static enum drm_mode_status msm_dp_bridge_mode_valid(struct drm_bridge *drm_bridge,
+ const struct drm_display_info *info,
+ const struct drm_display_mode *mode)
+{
+ struct msm_dp_bridge *dp_bridge = to_dp_bridge(drm_bridge);
+ struct msm_dp *dp = dp_bridge->msm_dp_display;
+
+ return msm_dp_display_mode_valid(dp, info, mode);
+}
+
static const struct drm_bridge_funcs msm_dp_bridge_ops = {
.atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
.atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
@@ -163,7 +210,7 @@ static void msm_edp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
return;
}
- msm_dp_bridge_atomic_enable(drm_bridge, old_bridge_state);
+ msm_dp_display_atomic_enable(dp);
}
static void msm_edp_bridge_atomic_disable(struct drm_bridge *drm_bridge,
diff --git a/drivers/gpu/drm/msm/dp/dp_drm.h b/drivers/gpu/drm/msm/dp/dp_drm.h
index 8eae2f74839f71cd3ef8511df4bb1aaa235abf1b..81e628c32279210f6e09f2fc68b2aeafd76fd1fe 100644
--- a/drivers/gpu/drm/msm/dp/dp_drm.h
+++ b/drivers/gpu/drm/msm/dp/dp_drm.h
@@ -25,18 +25,6 @@ int msm_dp_bridge_init(struct msm_dp *msm_dp_display, struct drm_device *dev,
struct drm_encoder *encoder,
bool yuv_supported);
-void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
- struct drm_bridge_state *old_bridge_state);
-void msm_dp_bridge_atomic_disable(struct drm_bridge *drm_bridge,
- struct drm_bridge_state *old_bridge_state);
-void msm_dp_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
- struct drm_bridge_state *old_bridge_state);
-enum drm_mode_status msm_dp_bridge_mode_valid(struct drm_bridge *bridge,
- const struct drm_display_info *info,
- const struct drm_display_mode *mode);
-void msm_dp_bridge_mode_set(struct drm_bridge *drm_bridge,
- const struct drm_display_mode *mode,
- const struct drm_display_mode *adjusted_mode);
void msm_dp_bridge_hpd_enable(struct drm_bridge *bridge);
void msm_dp_bridge_hpd_disable(struct drm_bridge *bridge);
void msm_dp_bridge_hpd_notify(struct drm_bridge *bridge,
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 12/45] drm/msm/dp: make bridge helpers use dp_display to allow re-use
2024-12-06 4:31 ` [PATCH 12/45] drm/msm/dp: make bridge helpers use dp_display to allow re-use Abhinav Kumar
@ 2024-12-08 11:07 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:07 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:43PM -0800, Abhinav Kumar wrote:
> dp_bridge helpers take drm_bridge as an input and extract the
> dp_display object to be used in the dp_display module. Rather than
> doing it in a roundabout way, directly pass the dp_display object
> to these helpers so that the MST bridge can also re-use the same
> helpers.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 34 ++++++++-----------------
> drivers/gpu/drm/msm/dp/dp_display.h | 9 +++++++
> drivers/gpu/drm/msm/dp/dp_drm.c | 49 ++++++++++++++++++++++++++++++++++++-
> drivers/gpu/drm/msm/dp/dp_drm.h | 12 ---------
> 4 files changed, 67 insertions(+), 37 deletions(-)
>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 13/45] drm/msm/dp: separate dp_display_prepare() into its own API
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (11 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 12/45] drm/msm/dp: make bridge helpers use dp_display to allow re-use Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 12:12 ` Stephan Gerhold
2024-12-06 4:31 ` [PATCH 14/45] drm/msm/dp: introduce stream_id for each DP panel Abhinav Kumar
` (33 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
dp_display_prepare() only prepares the link in case its not
already ready before dp_display_enable(). Hence separate it into
its own API.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 24 +++++++++++++++++-------
drivers/gpu/drm/msm/dp/dp_display.h | 1 +
drivers/gpu/drm/msm/dp/dp_drm.c | 2 ++
3 files changed, 20 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 2f8650d60202deaa90de1a5e0dd6d8bc50f09782..02282f58f1b31594601692b406215cee4ca41032 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -1525,26 +1525,36 @@ int msm_dp_modeset_init(struct msm_dp *msm_dp_display, struct drm_device *dev,
return 0;
}
-void msm_dp_display_atomic_enable(struct msm_dp *dp)
+void msm_dp_display_atomic_prepare(struct msm_dp *dp)
{
int rc = 0;
-
struct msm_dp_display_private *msm_dp_display;
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
- if (dp->is_edp)
- msm_dp_hpd_plug_handle(msm_dp_display, 0);
-
mutex_lock(&msm_dp_display->event_mutex);
rc = msm_dp_display_prepare(msm_dp_display);
if (rc) {
DRM_ERROR("DP display prepare failed, rc=%d\n", rc);
- mutex_unlock(&msm_dp_display->event_mutex);
- return;
}
+ mutex_unlock(&msm_dp_display->event_mutex);
+}
+
+void msm_dp_display_atomic_enable(struct msm_dp *dp)
+{
+ int rc = 0;
+
+ struct msm_dp_display_private *msm_dp_display;
+
+ msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
+
+ if (dp->is_edp)
+ msm_dp_hpd_plug_handle(msm_dp_display, 0);
+
+ mutex_lock(&msm_dp_display->event_mutex);
+
if (dp->prepared) {
rc = msm_dp_display_enable(msm_dp_display);
if (rc)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 2a800164cd9c74c29db80dbad15a2dff9fcb93d6..46912a9855b512d9dc6a4edff91ffd21df46e22a 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -44,6 +44,7 @@ void msm_dp_display_debugfs_init(struct msm_dp *msm_dp_display, struct dentry *d
void msm_dp_display_atomic_post_disable(struct msm_dp *dp_display);
void msm_dp_display_atomic_disable(struct msm_dp *dp_display);
void msm_dp_display_atomic_enable(struct msm_dp *dp_display);
+void msm_dp_display_atomic_prepare(struct msm_dp *dp);
void msm_dp_display_mode_set(struct msm_dp *dp,
const struct drm_display_mode *mode,
const struct drm_display_mode *adjusted_mode);
diff --git a/drivers/gpu/drm/msm/dp/dp_drm.c b/drivers/gpu/drm/msm/dp/dp_drm.c
index 4ef3f16a68890bc220d147ac3d04f53ef911f004..920392b3c688821bccdc66d50fb7052ac3a9a85a 100644
--- a/drivers/gpu/drm/msm/dp/dp_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_drm.c
@@ -103,6 +103,7 @@ static void msm_dp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
struct msm_dp_bridge *dp_bridge = to_dp_bridge(drm_bridge);
struct msm_dp *dp = dp_bridge->msm_dp_display;
+ msm_dp_display_atomic_prepare(dp);
msm_dp_display_atomic_enable(dp);
}
@@ -210,6 +211,7 @@ static void msm_edp_bridge_atomic_enable(struct drm_bridge *drm_bridge,
return;
}
+ msm_dp_display_atomic_prepare(dp);
msm_dp_display_atomic_enable(dp);
}
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 13/45] drm/msm/dp: separate dp_display_prepare() into its own API
2024-12-06 4:31 ` [PATCH 13/45] drm/msm/dp: separate dp_display_prepare() into its own API Abhinav Kumar
@ 2024-12-06 12:12 ` Stephan Gerhold
0 siblings, 0 replies; 111+ messages in thread
From: Stephan Gerhold @ 2024-12-06 12:12 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Vara Reddy,
Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel, freedreno,
linux-kernel, devicetree, Jessica Zhang, Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:44PM -0800, Abhinav Kumar wrote:
> dp_display_prepare() only prepares the link in case its not
> already ready before dp_display_enable(). Hence separate it into
> its own API.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 24 +++++++++++++++++-------
> drivers/gpu/drm/msm/dp/dp_display.h | 1 +
> drivers/gpu/drm/msm/dp/dp_drm.c | 2 ++
> 3 files changed, 20 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 2f8650d60202deaa90de1a5e0dd6d8bc50f09782..02282f58f1b31594601692b406215cee4ca41032 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -1525,26 +1525,36 @@ int msm_dp_modeset_init(struct msm_dp *msm_dp_display, struct drm_device *dev,
> return 0;
> }
>
> -void msm_dp_display_atomic_enable(struct msm_dp *dp)
> +void msm_dp_display_atomic_prepare(struct msm_dp *dp)
> {
> int rc = 0;
> -
> struct msm_dp_display_private *msm_dp_display;
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
>
> - if (dp->is_edp)
> - msm_dp_hpd_plug_handle(msm_dp_display, 0);
> -
> mutex_lock(&msm_dp_display->event_mutex);
>
> rc = msm_dp_display_prepare(msm_dp_display);
> if (rc) {
> DRM_ERROR("DP display prepare failed, rc=%d\n", rc);
> - mutex_unlock(&msm_dp_display->event_mutex);
> - return;
> }
FWIW: This patch breaks the eDP panel on the X1E80100 CRD for me. If you
don't do the msm_dp_hpd_plug_handle() before msm_dp_display_prepare(),
then the link_params (rate/num_lanes etc) are not initialized. Moving it
back here seems to fix it.
Maybe I'm missing some dependent patches or so, I was just experimenting
a bit. :-)
Thanks,
Stephan
[ 17.724076] phy phy-aec5a00.phy.15: phy poweron failed --> -22
[ 17.724698] ------------[ cut here ]------------
[ 17.724699] disp_cc_mdss_dptx3_link_clk status stuck at 'off'
[ 17.724709] WARNING: CPU: 9 PID: 705 at drivers/clk/qcom/clk-branch.c:88 clk_branch_toggle+0x124/0x16c
[ 17.724877] CPU: 9 UID: 0 PID: 705 Comm: (udev-worker) Not tainted 6.13.0-rc1 #1
[ 17.724883] pstate: 614000c5 (nZCv daIF +PAN -UAO -TCO +DIT -SSBS BTYPE=--)
[ 17.724887] pc : clk_branch_toggle+0x124/0x16c
[ 17.724889] lr : clk_branch_toggle+0x120/0x16c
[ 17.724927] Call trace:
[ 17.724929] clk_branch_toggle+0x124/0x16c (P)
[ 17.724933] clk_branch_toggle+0x120/0x16c (L)
[ 17.724935] clk_branch2_enable+0x1c/0x28
[ 17.724938] clk_core_enable+0x78/0xb4
[ 17.724944] clk_core_enable_lock+0x88/0x118
[ 17.724947] clk_enable+0x1c/0x28
[ 17.724950] clk_bulk_enable+0x38/0xb0
[ 17.724953] msm_dp_ctrl_enable_mainlink_clocks+0x140/0x234 [msm]
[ 17.724974] msm_dp_ctrl_prepare_stream_on+0x10c/0x19c [msm]
[ 17.724985] msm_dp_display_atomic_prepare+0x9c/0x1b4 [msm]
[ 17.724994] msm_edp_bridge_atomic_enable+0x60/0x78 [msm]
[...]
[ 17.725216] ---[ end trace 0000000000000000 ]---
[ 17.725218] Failed to enable clk 'ctrl_link': -16
[ 17.725220] [drm:msm_dp_ctrl_enable_mainlink_clocks [msm]] *ERROR* Unable to start link clocks. ret=-16
[ 17.725231] [drm:msm_dp_ctrl_prepare_stream_on [msm]] *ERROR* Failed to start link clocks. ret=-16
[ 17.725240] [drm:msm_dp_display_atomic_prepare [msm]] *ERROR* DP display prepare failed, rc=-16
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 14/45] drm/msm/dp: introduce stream_id for each DP panel
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (12 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 13/45] drm/msm/dp: separate dp_display_prepare() into its own API Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:22 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 15/45] drm/msm/dp: convert dp_display_set_mode() to use dp_panel argument Abhinav Kumar
` (32 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
With MST, each DP controller can handle multiple streams.
There shall be one dp_panel for each stream but the dp_display
object shall be shared among them. To represent this abstraction,
create a stream_id for each DP panel which shall be set by the
MST stream. For SST, default this to stream 0.
Use the stream ID to control the pixel clock of that respective
stream by extending the clock handles and state tracking of the
DP pixel clock to an array of max supported streams.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.h | 9 +++++++
drivers/gpu/drm/msm/dp/dp_ctrl.c | 50 +++++++++++++++++++++++--------------
drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 +-
drivers/gpu/drm/msm/dp/dp_display.c | 28 ++++++++++++++++++++-
drivers/gpu/drm/msm/dp/dp_display.h | 3 +++
drivers/gpu/drm/msm/dp/dp_panel.h | 2 ++
6 files changed, 73 insertions(+), 21 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
index e932b17eecbf514070cd8cd0b98ca0fefbe81ab7..edeebf1f313f50e9c54feee1e5aa6aa2dbba3058 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.h
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
@@ -51,6 +51,15 @@ struct msm_dp_catalog {
bool wide_bus_en;
};
+/* stream id */
+enum msm_dp_stream_id {
+ DP_STREAM_0,
+ DP_STREAM_1,
+ DP_STREAM_2,
+ DP_STREAM_3,
+ DP_STREAM_MAX,
+};
+
/* Debug module */
void msm_dp_catalog_snapshot(struct msm_dp_catalog *msm_dp_catalog, struct msm_disp_state *disp_state);
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 485339eb998cc6c8c1e8ab0a88b5c5d6ef300a1f..0648831df956dfc7afa1cbfb0dea2c32b02ff74e 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -86,7 +86,8 @@ struct msm_dp_ctrl_private {
unsigned int num_link_clks;
struct clk_bulk_data *link_clks;
- struct clk *pixel_clk;
+ struct clk *pixel_clk[DP_STREAM_MAX];
+ unsigned int num_pixel_clks;
union phy_configure_opts phy_opts;
@@ -96,7 +97,7 @@ struct msm_dp_ctrl_private {
bool core_clks_on;
bool link_clks_on;
- bool stream_clks_on;
+ bool stream_clks_on[DP_STREAM_MAX];
};
static int msm_dp_aux_link_configure(struct drm_dp_aux *aux,
@@ -1698,25 +1699,26 @@ static bool msm_dp_ctrl_send_phy_test_pattern(struct msm_dp_ctrl_private *ctrl)
return success;
}
-static int msm_dp_ctrl_stream_clk_on(struct msm_dp_ctrl_private *ctrl, unsigned long pixel_rate)
+static int msm_dp_ctrl_stream_clk_on(struct msm_dp_ctrl_private *ctrl, unsigned long pixel_rate,
+ enum msm_dp_stream_id stream_id)
{
int ret;
- ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
+ ret = clk_set_rate(ctrl->pixel_clk[stream_id], pixel_rate * 1000);
if (ret) {
DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
return ret;
}
- if (ctrl->stream_clks_on) {
+ if (ctrl->stream_clks_on[stream_id]) {
drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
} else {
- ret = clk_prepare_enable(ctrl->pixel_clk);
+ ret = clk_prepare_enable(ctrl->pixel_clk[stream_id]);
if (ret) {
DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
return ret;
}
- ctrl->stream_clks_on = true;
+ ctrl->stream_clks_on[stream_id] = true;
}
return ret;
@@ -1739,7 +1741,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
* running. Add the global reset just before disabling the
* link clocks and core clocks.
*/
- msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl);
+ msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl, msm_dp_panel);
msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
@@ -1749,7 +1751,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
}
pixel_rate = msm_dp_panel->msm_dp_mode.drm_mode.clock;
- ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
+ ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate, msm_dp_panel->stream_id);
msm_dp_ctrl_send_phy_test_pattern(ctrl);
@@ -1969,8 +1971,8 @@ int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_li
ctrl->link->link_params.num_lanes);
drm_dbg_dp(ctrl->drm_dev,
- "core_clk_on=%d link_clk_on=%d stream_clk_on=%d\n",
- ctrl->core_clks_on, ctrl->link_clks_on, ctrl->stream_clks_on);
+ "core_clk_on=%d link_clk_on=%d\n",
+ ctrl->core_clks_on, ctrl->link_clks_on);
if (!ctrl->link_clks_on) { /* link clk is off */
ret = msm_dp_ctrl_enable_mainlink_clocks(ctrl);
@@ -2009,7 +2011,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
- ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
+ ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate, msm_dp_panel->stream_id);
if (ret) {
DRM_ERROR("failed to enable stream pixel clk\n");
return ret;
@@ -2043,15 +2045,15 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
return ret;
}
-void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
+void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
{
struct msm_dp_ctrl_private *ctrl;
ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
- if (ctrl->stream_clks_on) {
- clk_disable_unprepare(ctrl->pixel_clk);
- ctrl->stream_clks_on = false;
+ if (ctrl->stream_clks_on[msm_dp_panel->stream_id]) {
+ clk_disable_unprepare(ctrl->pixel_clk[msm_dp_panel->stream_id]);
+ ctrl->stream_clks_on[msm_dp_panel->stream_id] = false;
}
}
@@ -2199,9 +2201,19 @@ static int msm_dp_ctrl_clk_init(struct msm_dp_ctrl *msm_dp_ctrl)
if (rc)
return rc;
- ctrl->pixel_clk = devm_clk_get(dev, "stream_pixel");
- if (IS_ERR(ctrl->pixel_clk))
- return PTR_ERR(ctrl->pixel_clk);
+ ctrl->num_pixel_clks = 0;
+
+ ctrl->pixel_clk[DP_STREAM_0] = devm_clk_get(dev, "stream_pixel");
+ if (IS_ERR(ctrl->pixel_clk[DP_STREAM_0]))
+ return PTR_ERR(ctrl->pixel_clk[DP_STREAM_0]);
+
+ ctrl->num_pixel_clks++;
+
+ ctrl->pixel_clk[DP_STREAM_1] = devm_clk_get(dev, "stream_1_pixel");
+ if (IS_ERR(ctrl->pixel_clk[DP_STREAM_1]))
+ DRM_ERROR("failed to get stream_1_pixel clock");
+ else
+ ctrl->num_pixel_clks++;
return 0;
}
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index 887cf5a866f07cb9038887a0634d3e1a0375879c..d422fd683b65d6f5e459710d0327e472a12c30b0 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -22,7 +22,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
-void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl);
+void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl);
irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 02282f58f1b31594601692b406215cee4ca41032..b506159191184a2a2c83d0735260ac040a33be98 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -911,7 +911,7 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
if (dp->link->sink_count == 0)
msm_dp_ctrl_psm_config(dp->ctrl);
- msm_dp_ctrl_stream_clk_off(dp->ctrl);
+ msm_dp_ctrl_stream_clk_off(dp->ctrl, dp->panel);
msm_dp_ctrl_off_link(dp->ctrl);
@@ -940,6 +940,30 @@ int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display,
return 0;
}
+int msm_dp_display_set_stream_id(struct msm_dp *dp,
+ struct msm_dp_panel *panel, u32 strm_id)
+{
+ int rc = 0;
+ struct msm_dp_display_private *msm_dp_display;
+
+ msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
+
+ if (!msm_dp_display) {
+ DRM_ERROR("invalid input\n");
+ return -EINVAL;
+ }
+
+ if (strm_id >= DP_STREAM_MAX) {
+ DRM_ERROR("invalid stream id:%d\n", strm_id);
+ return -EINVAL;
+ }
+
+ if (panel)
+ panel->stream_id = strm_id;
+
+ return rc;
+}
+
/**
* msm_dp_bridge_mode_valid - callback to determine if specified mode is valid
* @dp: Pointer to dp display structure
@@ -1555,6 +1579,8 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
mutex_lock(&msm_dp_display->event_mutex);
+ msm_dp_display_set_stream_id(dp, msm_dp_display->panel, 0);
+
if (dp->prepared) {
rc = msm_dp_display_enable(msm_dp_display);
if (rc)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 46912a9855b512d9dc6a4edff91ffd21df46e22a..8ce8ba254b1bfe5b4c000df83eceef5823772780 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -34,6 +34,7 @@ struct msm_dp {
int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display,
hdmi_codec_plugged_cb fn, struct device *codec_dev);
+
int msm_dp_display_get_modes(struct msm_dp *msm_dp_display);
bool msm_dp_display_check_video_test(struct msm_dp *msm_dp_display);
int msm_dp_display_get_test_bpp(struct msm_dp *msm_dp_display);
@@ -51,5 +52,7 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
const struct drm_display_info *info,
const struct drm_display_mode *mode);
+int msm_dp_display_set_stream_id(struct msm_dp *dp,
+ struct msm_dp_panel *panel, u32 strm_id);
#endif /* _DP_DISPLAY_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
index 363b416e4cbe290f9c0e6171d6c0c5170f9fea62..9af2272da697e7aa49377c02abdb97e72f07c0bd 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.h
+++ b/drivers/gpu/drm/msm/dp/dp_panel.h
@@ -46,6 +46,8 @@ struct msm_dp_panel {
bool video_test;
bool vsc_sdp_supported;
+ enum msm_dp_stream_id stream_id;
+
u32 max_dp_lanes;
u32 max_dp_link_rate;
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 14/45] drm/msm/dp: introduce stream_id for each DP panel
2024-12-06 4:31 ` [PATCH 14/45] drm/msm/dp: introduce stream_id for each DP panel Abhinav Kumar
@ 2024-12-08 11:22 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:22 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:45PM -0800, Abhinav Kumar wrote:
> With MST, each DP controller can handle multiple streams.
> There shall be one dp_panel for each stream but the dp_display
> object shall be shared among them. To represent this abstraction,
> create a stream_id for each DP panel which shall be set by the
> MST stream. For SST, default this to stream 0.
>
> Use the stream ID to control the pixel clock of that respective
> stream by extending the clock handles and state tracking of the
> DP pixel clock to an array of max supported streams.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.h | 9 +++++++
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 50 +++++++++++++++++++++++--------------
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 +-
> drivers/gpu/drm/msm/dp/dp_display.c | 28 ++++++++++++++++++++-
> drivers/gpu/drm/msm/dp/dp_display.h | 3 +++
> drivers/gpu/drm/msm/dp/dp_panel.h | 2 ++
> 6 files changed, 73 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
> index e932b17eecbf514070cd8cd0b98ca0fefbe81ab7..edeebf1f313f50e9c54feee1e5aa6aa2dbba3058 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
> @@ -51,6 +51,15 @@ struct msm_dp_catalog {
> bool wide_bus_en;
> };
>
> +/* stream id */
> +enum msm_dp_stream_id {
> + DP_STREAM_0,
> + DP_STREAM_1,
> + DP_STREAM_2,
> + DP_STREAM_3,
> + DP_STREAM_MAX,
> +};
> +
> /* Debug module */
> void msm_dp_catalog_snapshot(struct msm_dp_catalog *msm_dp_catalog, struct msm_disp_state *disp_state);
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 485339eb998cc6c8c1e8ab0a88b5c5d6ef300a1f..0648831df956dfc7afa1cbfb0dea2c32b02ff74e 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -86,7 +86,8 @@ struct msm_dp_ctrl_private {
> unsigned int num_link_clks;
> struct clk_bulk_data *link_clks;
>
> - struct clk *pixel_clk;
> + struct clk *pixel_clk[DP_STREAM_MAX];
> + unsigned int num_pixel_clks;
>
> union phy_configure_opts phy_opts;
>
> @@ -96,7 +97,7 @@ struct msm_dp_ctrl_private {
>
> bool core_clks_on;
> bool link_clks_on;
> - bool stream_clks_on;
> + bool stream_clks_on[DP_STREAM_MAX];
> };
>
> static int msm_dp_aux_link_configure(struct drm_dp_aux *aux,
> @@ -1698,25 +1699,26 @@ static bool msm_dp_ctrl_send_phy_test_pattern(struct msm_dp_ctrl_private *ctrl)
> return success;
> }
>
> -static int msm_dp_ctrl_stream_clk_on(struct msm_dp_ctrl_private *ctrl, unsigned long pixel_rate)
> +static int msm_dp_ctrl_stream_clk_on(struct msm_dp_ctrl_private *ctrl, unsigned long pixel_rate,
> + enum msm_dp_stream_id stream_id)
> {
> int ret;
>
> - ret = clk_set_rate(ctrl->pixel_clk, pixel_rate * 1000);
> + ret = clk_set_rate(ctrl->pixel_clk[stream_id], pixel_rate * 1000);
> if (ret) {
> DRM_ERROR("Failed to set pixel clock rate. ret=%d\n", ret);
> return ret;
> }
>
> - if (ctrl->stream_clks_on) {
> + if (ctrl->stream_clks_on[stream_id]) {
> drm_dbg_dp(ctrl->drm_dev, "pixel clks already enabled\n");
> } else {
> - ret = clk_prepare_enable(ctrl->pixel_clk);
> + ret = clk_prepare_enable(ctrl->pixel_clk[stream_id]);
> if (ret) {
> DRM_ERROR("Failed to start pixel clocks. ret=%d\n", ret);
> return ret;
> }
> - ctrl->stream_clks_on = true;
> + ctrl->stream_clks_on[stream_id] = true;
> }
>
> return ret;
> @@ -1739,7 +1741,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
> * running. Add the global reset just before disabling the
> * link clocks and core clocks.
> */
> - msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl);
> + msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl, msm_dp_panel);
> msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
>
> ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
> @@ -1749,7 +1751,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
> }
>
> pixel_rate = msm_dp_panel->msm_dp_mode.drm_mode.clock;
> - ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
> + ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate, msm_dp_panel->stream_id);
Why does msm_dp_ctrl_stream_clk_on() get stream_id, while
msm_dp_ctrl_stream_clk_off() gets the panel? I'd expect that they are
symmetrical.
>
> msm_dp_ctrl_send_phy_test_pattern(ctrl);
>
> @@ -1969,8 +1971,8 @@ int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_li
> ctrl->link->link_params.num_lanes);
>
> drm_dbg_dp(ctrl->drm_dev,
> - "core_clk_on=%d link_clk_on=%d stream_clk_on=%d\n",
> - ctrl->core_clks_on, ctrl->link_clks_on, ctrl->stream_clks_on);
> + "core_clk_on=%d link_clk_on=%d\n",
> + ctrl->core_clks_on, ctrl->link_clks_on);
>
> if (!ctrl->link_clks_on) { /* link clk is off */
> ret = msm_dp_ctrl_enable_mainlink_clocks(ctrl);
> @@ -2009,7 +2011,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
>
> drm_dbg_dp(ctrl->drm_dev, "pixel_rate=%lu\n", pixel_rate);
>
> - ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate);
> + ret = msm_dp_ctrl_stream_clk_on(ctrl, pixel_rate, msm_dp_panel->stream_id);
> if (ret) {
> DRM_ERROR("failed to enable stream pixel clk\n");
> return ret;
> @@ -2043,15 +2045,15 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> return ret;
> }
>
> -void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl)
> +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
> {
> struct msm_dp_ctrl_private *ctrl;
>
> ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
>
> - if (ctrl->stream_clks_on) {
> - clk_disable_unprepare(ctrl->pixel_clk);
> - ctrl->stream_clks_on = false;
> + if (ctrl->stream_clks_on[msm_dp_panel->stream_id]) {
> + clk_disable_unprepare(ctrl->pixel_clk[msm_dp_panel->stream_id]);
> + ctrl->stream_clks_on[msm_dp_panel->stream_id] = false;
> }
> }
>
> @@ -2199,9 +2201,19 @@ static int msm_dp_ctrl_clk_init(struct msm_dp_ctrl *msm_dp_ctrl)
> if (rc)
> return rc;
>
> - ctrl->pixel_clk = devm_clk_get(dev, "stream_pixel");
> - if (IS_ERR(ctrl->pixel_clk))
> - return PTR_ERR(ctrl->pixel_clk);
> + ctrl->num_pixel_clks = 0;
> +
> + ctrl->pixel_clk[DP_STREAM_0] = devm_clk_get(dev, "stream_pixel");
> + if (IS_ERR(ctrl->pixel_clk[DP_STREAM_0]))
> + return PTR_ERR(ctrl->pixel_clk[DP_STREAM_0]);
> +
> + ctrl->num_pixel_clks++;
> +
> + ctrl->pixel_clk[DP_STREAM_1] = devm_clk_get(dev, "stream_1_pixel");
> + if (IS_ERR(ctrl->pixel_clk[DP_STREAM_1]))
> + DRM_ERROR("failed to get stream_1_pixel clock");
> + else
> + ctrl->num_pixel_clks++;
>
> return 0;
> }
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index 887cf5a866f07cb9038887a0634d3e1a0375879c..d422fd683b65d6f5e459710d0327e472a12c30b0 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -22,7 +22,7 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
> void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
> -void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl);
> +void msm_dp_ctrl_stream_clk_off(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
> void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl);
> irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_handle_sink_request(struct msm_dp_ctrl *msm_dp_ctrl);
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 02282f58f1b31594601692b406215cee4ca41032..b506159191184a2a2c83d0735260ac040a33be98 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -911,7 +911,7 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> if (dp->link->sink_count == 0)
> msm_dp_ctrl_psm_config(dp->ctrl);
>
> - msm_dp_ctrl_stream_clk_off(dp->ctrl);
> + msm_dp_ctrl_stream_clk_off(dp->ctrl, dp->panel);
>
> msm_dp_ctrl_off_link(dp->ctrl);
>
> @@ -940,6 +940,30 @@ int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display,
> return 0;
> }
>
> +int msm_dp_display_set_stream_id(struct msm_dp *dp,
> + struct msm_dp_panel *panel, u32 strm_id)
stream_id, there is no need be cryptic.
> +{
> + int rc = 0;
> + struct msm_dp_display_private *msm_dp_display;
> +
> + msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
> +
> + if (!msm_dp_display) {
> + DRM_ERROR("invalid input\n");
> + return -EINVAL;
> + }
> +
> + if (strm_id >= DP_STREAM_MAX) {
> + DRM_ERROR("invalid stream id:%d\n", strm_id);
> + return -EINVAL;
Who can actually check for this error? Please try moving stream IDs to
the state and assign them during atomic_check().
> + }
> +
> + if (panel)
> + panel->stream_id = strm_id;
> +
> + return rc;
> +}
> +
> /**
> * msm_dp_bridge_mode_valid - callback to determine if specified mode is valid
> * @dp: Pointer to dp display structure
> @@ -1555,6 +1579,8 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
>
> mutex_lock(&msm_dp_display->event_mutex);
>
> + msm_dp_display_set_stream_id(dp, msm_dp_display->panel, 0);
> +
> if (dp->prepared) {
> rc = msm_dp_display_enable(msm_dp_display);
> if (rc)
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> index 46912a9855b512d9dc6a4edff91ffd21df46e22a..8ce8ba254b1bfe5b4c000df83eceef5823772780 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.h
> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> @@ -34,6 +34,7 @@ struct msm_dp {
>
> int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display,
> hdmi_codec_plugged_cb fn, struct device *codec_dev);
> +
> int msm_dp_display_get_modes(struct msm_dp *msm_dp_display);
> bool msm_dp_display_check_video_test(struct msm_dp *msm_dp_display);
> int msm_dp_display_get_test_bpp(struct msm_dp *msm_dp_display);
> @@ -51,5 +52,7 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
> enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
> const struct drm_display_info *info,
> const struct drm_display_mode *mode);
> +int msm_dp_display_set_stream_id(struct msm_dp *dp,
> + struct msm_dp_panel *panel, u32 strm_id);
>
> #endif /* _DP_DISPLAY_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
> index 363b416e4cbe290f9c0e6171d6c0c5170f9fea62..9af2272da697e7aa49377c02abdb97e72f07c0bd 100644
> --- a/drivers/gpu/drm/msm/dp/dp_panel.h
> +++ b/drivers/gpu/drm/msm/dp/dp_panel.h
> @@ -46,6 +46,8 @@ struct msm_dp_panel {
> bool video_test;
> bool vsc_sdp_supported;
>
> + enum msm_dp_stream_id stream_id;
> +
> u32 max_dp_lanes;
> u32 max_dp_link_rate;
>
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 15/45] drm/msm/dp: convert dp_display_set_mode() to use dp_panel argument
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (13 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 14/45] drm/msm/dp: introduce stream_id for each DP panel Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:39 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 16/45] drm/msm/dp: add support for programming p1 register block Abhinav Kumar
` (31 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Convert dp_display_set_mode() to use the dp_panel passed to it
as an argument rather than the cached one in dp_display_private.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 60 ++++++++++++++++++-------------------
1 file changed, 30 insertions(+), 30 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index b506159191184a2a2c83d0735260ac040a33be98..5fa6c003cf6c51eae77573549a555a00dc33f476 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -798,16 +798,38 @@ static int msm_dp_init_sub_modules(struct msm_dp_display_private *dp)
}
static int msm_dp_display_set_mode(struct msm_dp *msm_dp_display,
- struct msm_dp_display_mode *mode)
+ const struct drm_display_mode *adjusted_mode,
+ struct msm_dp_panel *msm_dp_panel)
{
- struct msm_dp_display_private *dp;
+ struct msm_dp_display_mode msm_dp_mode;
- dp = container_of(msm_dp_display, struct msm_dp_display_private, msm_dp_display);
+ memset(&msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
+
+ if (msm_dp_display_check_video_test(msm_dp_display))
+ msm_dp_mode.bpp = msm_dp_display_get_test_bpp(msm_dp_display);
+ else /* Default num_components per pixel = 3 */
+ msm_dp_mode.bpp = msm_dp_panel->connector->display_info.bpc * 3;
+
+ if (!msm_dp_mode.bpp)
+ msm_dp_mode.bpp = 24; /* Default bpp */
+
+ drm_mode_copy(&msm_dp_mode.drm_mode, adjusted_mode);
+
+ msm_dp_mode.v_active_low =
+ !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
+
+ msm_dp_mode.h_active_low =
+ !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
+
+ msm_dp_mode.out_fmt_is_yuv_420 =
+ drm_mode_is_420_only(&msm_dp_display->connector->display_info, adjusted_mode) &&
+ msm_dp_panel->vsc_sdp_supported;
+
+ drm_mode_copy(&msm_dp_panel->msm_dp_mode.drm_mode, &msm_dp_mode.drm_mode);
+ msm_dp_panel->msm_dp_mode.bpp = msm_dp_mode.bpp;
+ msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 = msm_dp_mode.out_fmt_is_yuv_420;
+ msm_dp_panel_init_panel_info(msm_dp_panel);
- drm_mode_copy(&dp->panel->msm_dp_mode.drm_mode, &mode->drm_mode);
- dp->panel->msm_dp_mode.bpp = mode->bpp;
- dp->panel->msm_dp_mode.out_fmt_is_yuv_420 = mode->out_fmt_is_yuv_420;
- msm_dp_panel_init_panel_info(dp->panel);
return 0;
}
@@ -1662,34 +1684,12 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
{
struct msm_dp_display_private *msm_dp_display;
struct msm_dp_panel *msm_dp_panel;
- struct msm_dp_display_mode msm_dp_mode;
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
msm_dp_panel = msm_dp_display->panel;
- memset(&msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
-
- if (msm_dp_display_check_video_test(dp))
- msm_dp_mode.bpp = msm_dp_display_get_test_bpp(dp);
- else /* Default num_components per pixel = 3 */
- msm_dp_mode.bpp = dp->connector->display_info.bpc * 3;
-
- if (!msm_dp_mode.bpp)
- msm_dp_mode.bpp = 24; /* Default bpp */
-
- drm_mode_copy(&msm_dp_mode.drm_mode, adjusted_mode);
-
- msm_dp_mode.v_active_low =
- !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
-
- msm_dp_mode.h_active_low =
- !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
-
- msm_dp_mode.out_fmt_is_yuv_420 =
- drm_mode_is_420_only(&dp->connector->display_info, adjusted_mode) &&
- msm_dp_panel->vsc_sdp_supported;
- msm_dp_display_set_mode(dp, &msm_dp_mode);
+ msm_dp_display_set_mode(dp, adjusted_mode, msm_dp_panel);
/* populate wide_bus_support to different layers */
msm_dp_display->ctrl->wide_bus_en =
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 15/45] drm/msm/dp: convert dp_display_set_mode() to use dp_panel argument
2024-12-06 4:31 ` [PATCH 15/45] drm/msm/dp: convert dp_display_set_mode() to use dp_panel argument Abhinav Kumar
@ 2024-12-08 11:39 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:39 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:46PM -0800, Abhinav Kumar wrote:
> Convert dp_display_set_mode() to use the dp_panel passed to it
> as an argument rather than the cached one in dp_display_private.
Why?
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 60 ++++++++++++++++++-------------------
> 1 file changed, 30 insertions(+), 30 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index b506159191184a2a2c83d0735260ac040a33be98..5fa6c003cf6c51eae77573549a555a00dc33f476 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -798,16 +798,38 @@ static int msm_dp_init_sub_modules(struct msm_dp_display_private *dp)
> }
>
> static int msm_dp_display_set_mode(struct msm_dp *msm_dp_display,
> - struct msm_dp_display_mode *mode)
> + const struct drm_display_mode *adjusted_mode,
> + struct msm_dp_panel *msm_dp_panel)
> {
> - struct msm_dp_display_private *dp;
> + struct msm_dp_display_mode msm_dp_mode;
>
> - dp = container_of(msm_dp_display, struct msm_dp_display_private, msm_dp_display);
> + memset(&msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
> +
> + if (msm_dp_display_check_video_test(msm_dp_display))
> + msm_dp_mode.bpp = msm_dp_display_get_test_bpp(msm_dp_display);
> + else /* Default num_components per pixel = 3 */
> + msm_dp_mode.bpp = msm_dp_panel->connector->display_info.bpc * 3;
> +
> + if (!msm_dp_mode.bpp)
> + msm_dp_mode.bpp = 24; /* Default bpp */
> +
> + drm_mode_copy(&msm_dp_mode.drm_mode, adjusted_mode);
> +
> + msm_dp_mode.v_active_low =
> + !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
> +
> + msm_dp_mode.h_active_low =
> + !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
> +
> + msm_dp_mode.out_fmt_is_yuv_420 =
> + drm_mode_is_420_only(&msm_dp_display->connector->display_info, adjusted_mode) &&
> + msm_dp_panel->vsc_sdp_supported;
> +
> + drm_mode_copy(&msm_dp_panel->msm_dp_mode.drm_mode, &msm_dp_mode.drm_mode);
> + msm_dp_panel->msm_dp_mode.bpp = msm_dp_mode.bpp;
> + msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 = msm_dp_mode.out_fmt_is_yuv_420;
> + msm_dp_panel_init_panel_info(msm_dp_panel);
>
> - drm_mode_copy(&dp->panel->msm_dp_mode.drm_mode, &mode->drm_mode);
> - dp->panel->msm_dp_mode.bpp = mode->bpp;
> - dp->panel->msm_dp_mode.out_fmt_is_yuv_420 = mode->out_fmt_is_yuv_420;
> - msm_dp_panel_init_panel_info(dp->panel);
> return 0;
> }
>
> @@ -1662,34 +1684,12 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
> {
> struct msm_dp_display_private *msm_dp_display;
> struct msm_dp_panel *msm_dp_panel;
> - struct msm_dp_display_mode msm_dp_mode;
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
> msm_dp_panel = msm_dp_display->panel;
>
> - memset(&msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
> -
> - if (msm_dp_display_check_video_test(dp))
> - msm_dp_mode.bpp = msm_dp_display_get_test_bpp(dp);
> - else /* Default num_components per pixel = 3 */
> - msm_dp_mode.bpp = dp->connector->display_info.bpc * 3;
> -
> - if (!msm_dp_mode.bpp)
> - msm_dp_mode.bpp = 24; /* Default bpp */
> -
> - drm_mode_copy(&msm_dp_mode.drm_mode, adjusted_mode);
> -
> - msm_dp_mode.v_active_low =
> - !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NVSYNC);
> -
> - msm_dp_mode.h_active_low =
> - !!(msm_dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
> -
> - msm_dp_mode.out_fmt_is_yuv_420 =
> - drm_mode_is_420_only(&dp->connector->display_info, adjusted_mode) &&
> - msm_dp_panel->vsc_sdp_supported;
>
> - msm_dp_display_set_mode(dp, &msm_dp_mode);
> + msm_dp_display_set_mode(dp, adjusted_mode, msm_dp_panel);
>
> /* populate wide_bus_support to different layers */
> msm_dp_display->ctrl->wide_bus_en =
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 16/45] drm/msm/dp: add support for programming p1 register block
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (14 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 15/45] drm/msm/dp: convert dp_display_set_mode() to use dp_panel argument Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-06 9:39 ` Dmitry Baryshkov
` (2 more replies)
2024-12-06 4:31 ` [PATCH 17/45] drm/msm/dp: use stream_id to change offsets in dp_catalog Abhinav Kumar
` (30 subsequent siblings)
46 siblings, 3 replies; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
p1 register block is needed for the second mst stream.
Add support in the catalog to be able to program this block.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.c | 29 +++++++++++++++++++++++++++++
1 file changed, 29 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index b4c8856fb25d01dd1b30c5ec33ce821aafa9551d..ee7f2d0b23aa034428a01ef2c9752f51013c5e01 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -73,6 +73,7 @@ struct dss_io_data {
struct dss_io_region aux;
struct dss_io_region link;
struct dss_io_region p0;
+ struct dss_io_region p1;
};
struct msm_dp_catalog_private {
@@ -93,6 +94,8 @@ void msm_dp_catalog_snapshot(struct msm_dp_catalog *msm_dp_catalog, struct msm_d
msm_disp_snapshot_add_block(disp_state, dss->aux.len, dss->aux.base, "dp_aux");
msm_disp_snapshot_add_block(disp_state, dss->link.len, dss->link.base, "dp_link");
msm_disp_snapshot_add_block(disp_state, dss->p0.len, dss->p0.base, "dp_p0");
+
+ msm_disp_snapshot_add_block(disp_state, dss->p1.len, dss->p0.base, "dp_p1");
}
static inline u32 msm_dp_read_aux(struct msm_dp_catalog_private *catalog, u32 offset)
@@ -145,6 +148,26 @@ static inline u32 msm_dp_read_p0(struct msm_dp_catalog_private *catalog,
return readl_relaxed(catalog->io.p0.base + offset);
}
+static inline void msm_dp_write_p1(struct msm_dp_catalog_private *catalog,
+ u32 offset, u32 data)
+{
+ /*
+ * To make sure interface reg writes happens before any other operation,
+ * this function uses writel() instread of writel_relaxed()
+ */
+ writel(data, catalog->io.p1.base + offset);
+}
+
+static inline u32 msm_dp_read_p1(struct msm_dp_catalog_private *catalog,
+ u32 offset)
+{
+ /*
+ * To make sure interface reg writes happens before any other operation,
+ * this function uses writel() instread of writel_relaxed()
+ */
+ return readl_relaxed(catalog->io.p1.base + offset);
+}
+
static inline u32 msm_dp_read_link(struct msm_dp_catalog_private *catalog, u32 offset)
{
return readl_relaxed(catalog->io.link.base + offset);
@@ -1137,6 +1160,12 @@ static int msm_dp_catalog_get_io(struct msm_dp_catalog_private *catalog)
DRM_ERROR("unable to remap p0 region: %pe\n", dss->p0.base);
return PTR_ERR(dss->p0.base);
}
+
+ dss->p1.base = msm_dp_ioremap(pdev, 4, &dss->p1.len);
+ if (IS_ERR(dss->p1.base)) {
+ DRM_ERROR("unable to remap p1 region: %pe\n", dss->p1.base);
+ return PTR_ERR(dss->p1.base);
+ }
}
return 0;
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 16/45] drm/msm/dp: add support for programming p1 register block
2024-12-06 4:31 ` [PATCH 16/45] drm/msm/dp: add support for programming p1 register block Abhinav Kumar
@ 2024-12-06 9:39 ` Dmitry Baryshkov
2024-12-06 9:42 ` Dmitry Baryshkov
2024-12-08 6:22 ` Dmitry Baryshkov
2 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:39 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:47PM -0800, Abhinav Kumar wrote:
> p1 register block is needed for the second mst stream.
> Add support in the catalog to be able to program this block.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 29 +++++++++++++++++++++++++++++
> 1 file changed, 29 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
> index b4c8856fb25d01dd1b30c5ec33ce821aafa9551d..ee7f2d0b23aa034428a01ef2c9752f51013c5e01 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> @@ -73,6 +73,7 @@ struct dss_io_data {
> struct dss_io_region aux;
> struct dss_io_region link;
> struct dss_io_region p0;
> + struct dss_io_region p1;
> };
>
> struct msm_dp_catalog_private {
> @@ -93,6 +94,8 @@ void msm_dp_catalog_snapshot(struct msm_dp_catalog *msm_dp_catalog, struct msm_d
> msm_disp_snapshot_add_block(disp_state, dss->aux.len, dss->aux.base, "dp_aux");
> msm_disp_snapshot_add_block(disp_state, dss->link.len, dss->link.base, "dp_link");
> msm_disp_snapshot_add_block(disp_state, dss->p0.len, dss->p0.base, "dp_p0");
> +
Drop extra empty line, please
> + msm_disp_snapshot_add_block(disp_state, dss->p1.len, dss->p0.base, "dp_p1");
> }
>
> static inline u32 msm_dp_read_aux(struct msm_dp_catalog_private *catalog, u32 offset)
> @@ -145,6 +148,26 @@ static inline u32 msm_dp_read_p0(struct msm_dp_catalog_private *catalog,
> return readl_relaxed(catalog->io.p0.base + offset);
> }
>
> +static inline void msm_dp_write_p1(struct msm_dp_catalog_private *catalog,
> + u32 offset, u32 data)
> +{
> + /*
> + * To make sure interface reg writes happens before any other operation,
> + * this function uses writel() instread of writel_relaxed()
> + */
> + writel(data, catalog->io.p1.base + offset);
> +}
> +
> +static inline u32 msm_dp_read_p1(struct msm_dp_catalog_private *catalog,
> + u32 offset)
> +{
> + /*
> + * To make sure interface reg writes happens before any other operation,
> + * this function uses writel() instread of writel_relaxed()
> + */
Not applicable to the actual function.
> + return readl_relaxed(catalog->io.p1.base + offset);
> +}
> +
> static inline u32 msm_dp_read_link(struct msm_dp_catalog_private *catalog, u32 offset)
> {
> return readl_relaxed(catalog->io.link.base + offset);
> @@ -1137,6 +1160,12 @@ static int msm_dp_catalog_get_io(struct msm_dp_catalog_private *catalog)
> DRM_ERROR("unable to remap p0 region: %pe\n", dss->p0.base);
> return PTR_ERR(dss->p0.base);
> }
> +
> + dss->p1.base = msm_dp_ioremap(pdev, 4, &dss->p1.len);
p1 is not populated for eDP case, it wasn't always present in DT, etc.
So please make it optional.
> + if (IS_ERR(dss->p1.base)) {
> + DRM_ERROR("unable to remap p1 region: %pe\n", dss->p1.base);
> + return PTR_ERR(dss->p1.base);
> + }
> }
>
> return 0;
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 16/45] drm/msm/dp: add support for programming p1 register block
2024-12-06 4:31 ` [PATCH 16/45] drm/msm/dp: add support for programming p1 register block Abhinav Kumar
2024-12-06 9:39 ` Dmitry Baryshkov
@ 2024-12-06 9:42 ` Dmitry Baryshkov
2024-12-08 6:22 ` Dmitry Baryshkov
2 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:42 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:47PM -0800, Abhinav Kumar wrote:
> p1 register block is needed for the second mst stream.
> Add support in the catalog to be able to program this block.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 29 +++++++++++++++++++++++++++++
> 1 file changed, 29 insertions(+)
>
> @@ -1137,6 +1160,12 @@ static int msm_dp_catalog_get_io(struct msm_dp_catalog_private *catalog)
> DRM_ERROR("unable to remap p0 region: %pe\n", dss->p0.base);
> return PTR_ERR(dss->p0.base);
> }
> +
> + dss->p1.base = msm_dp_ioremap(pdev, 4, &dss->p1.len);
> + if (IS_ERR(dss->p1.base)) {
> + DRM_ERROR("unable to remap p1 region: %pe\n", dss->p1.base);
> + return PTR_ERR(dss->p1.base);
> + }
Forgot to mention, please also map p1 in the legacy bingdings branch in
this function.
> }
>
> return 0;
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 16/45] drm/msm/dp: add support for programming p1 register block
2024-12-06 4:31 ` [PATCH 16/45] drm/msm/dp: add support for programming p1 register block Abhinav Kumar
2024-12-06 9:39 ` Dmitry Baryshkov
2024-12-06 9:42 ` Dmitry Baryshkov
@ 2024-12-08 6:22 ` Dmitry Baryshkov
2 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 6:22 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:47PM -0800, Abhinav Kumar wrote:
> p1 register block is needed for the second mst stream.
> Add support in the catalog to be able to program this block.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 29 +++++++++++++++++++++++++++++
> 1 file changed, 29 insertions(+)
> @@ -145,6 +148,26 @@ static inline u32 msm_dp_read_p0(struct msm_dp_catalog_private *catalog,
> return readl_relaxed(catalog->io.p0.base + offset);
> }
>
> +static inline void msm_dp_write_p1(struct msm_dp_catalog_private *catalog,
> + u32 offset, u32 data)
> +{
> + /*
> + * To make sure interface reg writes happens before any other operation,
> + * this function uses writel() instread of writel_relaxed()
> + */
> + writel(data, catalog->io.p1.base + offset);
> +}
> +
> +static inline u32 msm_dp_read_p1(struct msm_dp_catalog_private *catalog,
> + u32 offset)
> +{
> + /*
> + * To make sure interface reg writes happens before any other operation,
> + * this function uses writel() instread of writel_relaxed()
> + */
> + return readl_relaxed(catalog->io.p1.base + offset);
> +}
After looking at the actual code, please implement
msm_dp_read_pn(stream_id) / msm_dp_write_pn(stream_id)
> +
> static inline u32 msm_dp_read_link(struct msm_dp_catalog_private *catalog, u32 offset)
> {
> return readl_relaxed(catalog->io.link.base + offset);
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 17/45] drm/msm/dp: use stream_id to change offsets in dp_catalog
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (15 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 16/45] drm/msm/dp: add support for programming p1 register block Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 5:42 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 18/45] drm/msm/dp: add support to send ACT packets for MST Abhinav Kumar
` (29 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Use the dp_panel's stream_id to adjust the offsets for stream 1
which will be used for MST in the dp_catalog. Also add additional
register defines for stream 1.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.c | 99 ++++++++++++++++++++++++++++---------
drivers/gpu/drm/msm/dp/dp_catalog.h | 9 ++--
drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 ++
drivers/gpu/drm/msm/dp/dp_panel.c | 2 +
drivers/gpu/drm/msm/dp/dp_reg.h | 13 ++++-
5 files changed, 99 insertions(+), 27 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index ee7f2d0b23aa034428a01ef2c9752f51013c5e01..e6f6edf617898241c74580eb0ae6bc58f06a154f 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -457,10 +457,20 @@ void msm_dp_catalog_ctrl_config_misc(struct msm_dp_catalog *msm_dp_catalog,
u32 test_bits_depth)
{
u32 misc_val;
+ u32 reg_offset = 0;
+
struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
struct msm_dp_catalog_private, msm_dp_catalog);
- misc_val = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0);
+ if (msm_dp_catalog->stream_id >= DP_STREAM_MAX) {
+ DRM_ERROR("invalid stream_id:%d\n", msm_dp_catalog->stream_id);
+ return;
+ }
+
+ if (msm_dp_catalog->stream_id == DP_STREAM_1)
+ reg_offset = REG_DP1_MISC1_MISC0 - REG_DP_MISC1_MISC0;
+
+ misc_val = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0 + reg_offset);
/* clear bpp bits */
misc_val &= ~(0x07 << DP_MISC0_TEST_BITS_DEPTH_SHIFT);
@@ -470,7 +480,7 @@ void msm_dp_catalog_ctrl_config_misc(struct msm_dp_catalog *msm_dp_catalog,
misc_val |= DP_MISC0_SYNCHRONOUS_CLK;
drm_dbg_dp(catalog->drm_dev, "misc settings = 0x%x\n", misc_val);
- msm_dp_write_link(catalog, REG_DP_MISC1_MISC0, misc_val);
+ msm_dp_write_link(catalog, REG_DP_MISC1_MISC0 + reg_offset, misc_val);
}
void msm_dp_catalog_setup_peripheral_flush(struct msm_dp_catalog *msm_dp_catalog)
@@ -500,10 +510,21 @@ void msm_dp_catalog_ctrl_config_msa(struct msm_dp_catalog *msm_dp_catalog,
u32 const link_rate_hbr2 = 540000;
u32 const link_rate_hbr3 = 810000;
unsigned long den, num;
+ u32 mvid_reg_off = 0, nvid_reg_off = 0;
struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
struct msm_dp_catalog_private, msm_dp_catalog);
+ if (msm_dp_catalog->stream_id >= DP_STREAM_MAX) {
+ DRM_ERROR("invalid stream_id:%d\n", msm_dp_catalog->stream_id);
+ return;
+ }
+
+ if (msm_dp_catalog->stream_id == DP_STREAM_1) {
+ mvid_reg_off = REG_DP1_SOFTWARE_MVID - REG_DP_SOFTWARE_MVID;
+ nvid_reg_off = REG_DP1_SOFTWARE_NVID - REG_DP_SOFTWARE_NVID;
+ }
+
if (rate == link_rate_hbr3)
pixel_div = 6;
else if (rate == 162000 || rate == 270000)
@@ -545,9 +566,14 @@ void msm_dp_catalog_ctrl_config_msa(struct msm_dp_catalog *msm_dp_catalog,
nvid *= 3;
drm_dbg_dp(catalog->drm_dev, "mvid=0x%x, nvid=0x%x\n", mvid, nvid);
- msm_dp_write_link(catalog, REG_DP_SOFTWARE_MVID, mvid);
- msm_dp_write_link(catalog, REG_DP_SOFTWARE_NVID, nvid);
- msm_dp_write_p0(catalog, MMSS_DP_DSC_DTO, 0x0);
+
+ msm_dp_write_link(catalog, REG_DP_SOFTWARE_MVID + mvid_reg_off, mvid);
+ msm_dp_write_link(catalog, REG_DP_SOFTWARE_NVID + nvid_reg_off, nvid);
+
+ if (msm_dp_catalog->stream_id == DP_STREAM_0)
+ msm_dp_write_p0(catalog, MMSS_DP_DSC_DTO, 0x0);
+ else
+ msm_dp_write_p1(catalog, MMSS_DP_DSC_DTO, 0x0);
}
int msm_dp_catalog_ctrl_set_pattern_state_bit(struct msm_dp_catalog *msm_dp_catalog,
@@ -910,13 +936,20 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
struct msm_dp_catalog_private, msm_dp_catalog);
u32 reg;
+ u32 offset = 0;
+
+ if (msm_dp_catalog->stream_id == DP_STREAM_1)
+ offset = REG_DP1_TOTAL_HOR_VER - REG_DP_TOTAL_HOR_VER;
- msm_dp_write_link(catalog, REG_DP_TOTAL_HOR_VER, total);
- msm_dp_write_link(catalog, REG_DP_START_HOR_VER_FROM_SYNC, sync_start);
- msm_dp_write_link(catalog, REG_DP_HSYNC_VSYNC_WIDTH_POLARITY, width_blanking);
- msm_dp_write_link(catalog, REG_DP_ACTIVE_HOR_VER, msm_dp_active);
+ msm_dp_write_link(catalog, REG_DP_TOTAL_HOR_VER + offset, total);
+ msm_dp_write_link(catalog, REG_DP_START_HOR_VER_FROM_SYNC + offset, sync_start);
+ msm_dp_write_link(catalog, REG_DP_HSYNC_VSYNC_WIDTH_POLARITY + offset, width_blanking);
+ msm_dp_write_link(catalog, REG_DP_ACTIVE_HOR_VER + offset, msm_dp_active);
- reg = msm_dp_read_p0(catalog, MMSS_DP_INTF_CONFIG);
+ if (msm_dp_catalog->stream_id == DP_STREAM_0)
+ reg = msm_dp_read_p0(catalog, MMSS_DP_INTF_CONFIG);
+ else
+ reg = msm_dp_read_p1(catalog, MMSS_DP_INTF_CONFIG);
if (msm_dp_catalog->wide_bus_en)
reg |= DP_INTF_CONFIG_DATABUS_WIDEN;
@@ -926,7 +959,11 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
DRM_DEBUG_DP("wide_bus_en=%d reg=%#x\n", msm_dp_catalog->wide_bus_en, reg);
- msm_dp_write_p0(catalog, MMSS_DP_INTF_CONFIG, reg);
+ if (msm_dp_catalog->stream_id == DP_STREAM_0)
+ msm_dp_write_p0(catalog, MMSS_DP_INTF_CONFIG, reg);
+ else
+ msm_dp_write_p1(catalog, MMSS_DP_INTF_CONFIG, reg);
+
return 0;
}
@@ -936,18 +973,22 @@ static void msm_dp_catalog_panel_send_vsc_sdp(struct msm_dp_catalog *msm_dp_cata
u32 header[2];
u32 val;
int i;
+ u32 msm_dp_generic_offset = 0;
catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
+ if (msm_dp_catalog->stream_id == DP_STREAM_1)
+ msm_dp_generic_offset = MMSS_DP1_GENERIC0_0 - MMSS_DP_GENERIC0_0;
+
msm_dp_utils_pack_sdp_header(&vsc_sdp->sdp_header, header);
- msm_dp_write_link(catalog, MMSS_DP_GENERIC0_0, header[0]);
- msm_dp_write_link(catalog, MMSS_DP_GENERIC0_1, header[1]);
+ msm_dp_write_link(catalog, MMSS_DP_GENERIC0_0 + msm_dp_generic_offset, header[0]);
+ msm_dp_write_link(catalog, MMSS_DP_GENERIC0_1 + msm_dp_generic_offset, header[1]);
for (i = 0; i < sizeof(vsc_sdp->db); i += 4) {
val = ((vsc_sdp->db[i]) | (vsc_sdp->db[i + 1] << 8) | (vsc_sdp->db[i + 2] << 16) |
(vsc_sdp->db[i + 3] << 24));
- msm_dp_write_link(catalog, MMSS_DP_GENERIC0_2 + i, val);
+ msm_dp_write_link(catalog, MMSS_DP_GENERIC0_2 + i + msm_dp_generic_offset, val);
}
}
@@ -955,13 +996,17 @@ static void msm_dp_catalog_panel_update_sdp(struct msm_dp_catalog *msm_dp_catalo
{
struct msm_dp_catalog_private *catalog;
u32 hw_revision;
+ u32 sdp_cfg3_offset = 0;
catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
+ if (msm_dp_catalog->stream_id == DP_STREAM_1)
+ sdp_cfg3_offset = MMSS_DP1_SDP_CFG3 - MMSS_DP_SDP_CFG3;
+
hw_revision = msm_dp_catalog_hw_revision(msm_dp_catalog);
if (hw_revision < DP_HW_VERSION_1_2 && hw_revision >= DP_HW_VERSION_1_0) {
- msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3, 0x01);
- msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3, 0x00);
+ msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3 + sdp_cfg3_offset, 0x01);
+ msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3 + sdp_cfg3_offset, 0x00);
}
}
@@ -969,18 +1014,27 @@ void msm_dp_catalog_panel_enable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog,
{
struct msm_dp_catalog_private *catalog;
u32 cfg, cfg2, misc;
+ u32 misc_reg_offset = 0;
+ u32 sdp_cfg_offset = 0;
+ u32 sdp_cfg2_offset = 0;
catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
- cfg = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG);
- cfg2 = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG2);
- misc = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0);
+ if (msm_dp_catalog->stream_id == DP_STREAM_1) {
+ misc_reg_offset = REG_DP1_MISC1_MISC0 - REG_DP_MISC1_MISC0;
+ sdp_cfg_offset = MMSS_DP1_SDP_CFG - MMSS_DP_SDP_CFG;
+ sdp_cfg2_offset = MMSS_DP1_SDP_CFG2 - MMSS_DP_SDP_CFG2;
+ }
+
+ cfg = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG + sdp_cfg_offset);
+ cfg2 = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG2 + sdp_cfg2_offset);
+ misc = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0 + misc_reg_offset);
cfg |= GEN0_SDP_EN;
- msm_dp_write_link(catalog, MMSS_DP_SDP_CFG, cfg);
+ msm_dp_write_link(catalog, MMSS_DP_SDP_CFG + sdp_cfg_offset, cfg);
cfg2 |= GENERIC0_SDPSIZE_VALID;
- msm_dp_write_link(catalog, MMSS_DP_SDP_CFG2, cfg2);
+ msm_dp_write_link(catalog, MMSS_DP_SDP_CFG2 + sdp_cfg2_offset, cfg2);
msm_dp_catalog_panel_send_vsc_sdp(msm_dp_catalog, vsc_sdp);
@@ -990,7 +1044,8 @@ void msm_dp_catalog_panel_enable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog,
drm_dbg_dp(catalog->drm_dev, "vsc sdp enable=1\n");
pr_debug("misc settings = 0x%x\n", misc);
- msm_dp_write_link(catalog, REG_DP_MISC1_MISC0, misc);
+
+ msm_dp_write_link(catalog, REG_DP_MISC1_MISC0 + misc_reg_offset, misc);
msm_dp_catalog_panel_update_sdp(msm_dp_catalog);
}
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
index edeebf1f313f50e9c54feee1e5aa6aa2dbba3058..c020b7cfa008241e937f6a53764b136431f1dbd9 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.h
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
@@ -47,10 +47,6 @@ enum msm_dp_catalog_audio_header_type {
DP_AUDIO_SDP_HEADER_MAX,
};
-struct msm_dp_catalog {
- bool wide_bus_en;
-};
-
/* stream id */
enum msm_dp_stream_id {
DP_STREAM_0,
@@ -60,6 +56,11 @@ enum msm_dp_stream_id {
DP_STREAM_MAX,
};
+struct msm_dp_catalog {
+ bool wide_bus_en;
+ enum msm_dp_stream_id stream_id;
+};
+
/* Debug module */
void msm_dp_catalog_snapshot(struct msm_dp_catalog *msm_dp_catalog, struct msm_disp_state *disp_state);
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 0648831df956dfc7afa1cbfb0dea2c32b02ff74e..ba39b009032dd6f5cb708988963cd6acb6838e4a 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -179,6 +179,7 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
struct msm_dp_panel *msm_dp_panel)
{
u32 cc, tb;
+ ctrl->catalog->stream_id = msm_dp_panel->stream_id;
msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
@@ -2062,7 +2063,9 @@ void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_d
struct msm_dp_ctrl_private *ctrl;
ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+ ctrl->catalog->stream_id = dp_panel->stream_id;
msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
+
}
void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl)
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
index 172de804dec445cb08ad8e3f058407f483cd6684..662bf02b8b1a5165f927835bef3c11ac091ddce6 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.c
+++ b/drivers/gpu/drm/msm/dp/dp_panel.c
@@ -309,7 +309,9 @@ static int msm_dp_panel_setup_vsc_sdp_yuv_420(struct msm_dp_panel *msm_dp_panel)
panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
catalog = panel->catalog;
+
msm_dp_mode = &msm_dp_panel->msm_dp_mode;
+ catalog->stream_id = msm_dp_panel->stream_id;
memset(&vsc_sdp_data, 0, sizeof(vsc_sdp_data));
diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
index 3835c7f5cb984406f8fc52ea765ef2315e0d175b..6c534fde6034fced2cb428e9a29de31ed5c5fcc4 100644
--- a/drivers/gpu/drm/msm/dp/dp_reg.h
+++ b/drivers/gpu/drm/msm/dp/dp_reg.h
@@ -138,13 +138,17 @@
#define DP_CONFIGURATION_CTRL_LSCLK_DIV_SHIFT (0x0D)
#define REG_DP_SOFTWARE_MVID (0x00000010)
+#define REG_DP1_SOFTWARE_MVID (0x00000414)
#define REG_DP_SOFTWARE_NVID (0x00000018)
+#define REG_DP1_SOFTWARE_NVID (0x00000418)
#define REG_DP_TOTAL_HOR_VER (0x0000001C)
+#define REG_DP1_TOTAL_HOR_VER (0x0000041C)
#define REG_DP_START_HOR_VER_FROM_SYNC (0x00000020)
#define REG_DP_HSYNC_VSYNC_WIDTH_POLARITY (0x00000024)
#define REG_DP_ACTIVE_HOR_VER (0x00000028)
-
#define REG_DP_MISC1_MISC0 (0x0000002C)
+#define REG_DP1_MISC1_MISC0 (0x0000042C)
+
#define DP_MISC0_SYNCHRONOUS_CLK (0x00000001)
#define DP_MISC0_COLORIMETRY_CFG_SHIFT (0x00000001)
#define DP_MISC0_TEST_BITS_DEPTH_SHIFT (0x00000005)
@@ -211,8 +215,11 @@
#define MMSS_DP_AUDIO_CTRL_RESET (0x00000214)
#define MMSS_DP_SDP_CFG (0x00000228)
+#define MMSS_DP1_SDP_CFG (0x000004E0)
#define GEN0_SDP_EN (0x00020000)
#define MMSS_DP_SDP_CFG2 (0x0000022C)
+#define MMSS_DP1_SDP_CFG2 (0x000004E4)
+
#define MMSS_DP_AUDIO_TIMESTAMP_0 (0x00000230)
#define MMSS_DP_AUDIO_TIMESTAMP_1 (0x00000234)
#define GENERIC0_SDPSIZE_VALID (0x00010000)
@@ -221,6 +228,8 @@
#define MMSS_DP_AUDIO_STREAM_1 (0x00000244)
#define MMSS_DP_SDP_CFG3 (0x0000024c)
+#define MMSS_DP1_SDP_CFG3 (0x000004E8)
+
#define UPDATE_SDP (0x00000001)
#define MMSS_DP_EXTENSION_0 (0x00000250)
@@ -270,6 +279,8 @@
#define MMSS_DP_GENERIC1_8 (0x00000348)
#define MMSS_DP_GENERIC1_9 (0x0000034C)
+#define MMSS_DP1_GENERIC0_0 (0x00000490)
+
#define MMSS_DP_VSCEXT_0 (0x000002D0)
#define MMSS_DP_VSCEXT_1 (0x000002D4)
#define MMSS_DP_VSCEXT_2 (0x000002D8)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 17/45] drm/msm/dp: use stream_id to change offsets in dp_catalog
2024-12-06 4:31 ` [PATCH 17/45] drm/msm/dp: use stream_id to change offsets in dp_catalog Abhinav Kumar
@ 2024-12-08 5:42 ` Dmitry Baryshkov
2025-05-26 12:57 ` Yongxing Mou
0 siblings, 1 reply; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 5:42 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:48PM -0800, Abhinav Kumar wrote:
> Use the dp_panel's stream_id to adjust the offsets for stream 1
> which will be used for MST in the dp_catalog. Also add additional
> register defines for stream 1.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 99 ++++++++++++++++++++++++++++---------
> drivers/gpu/drm/msm/dp/dp_catalog.h | 9 ++--
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 ++
> drivers/gpu/drm/msm/dp/dp_panel.c | 2 +
> drivers/gpu/drm/msm/dp/dp_reg.h | 13 ++++-
> 5 files changed, 99 insertions(+), 27 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
> index ee7f2d0b23aa034428a01ef2c9752f51013c5e01..e6f6edf617898241c74580eb0ae6bc58f06a154f 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> @@ -457,10 +457,20 @@ void msm_dp_catalog_ctrl_config_misc(struct msm_dp_catalog *msm_dp_catalog,
> u32 test_bits_depth)
> {
> u32 misc_val;
> + u32 reg_offset = 0;
> +
> struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
> struct msm_dp_catalog_private, msm_dp_catalog);
>
> - misc_val = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0);
> + if (msm_dp_catalog->stream_id >= DP_STREAM_MAX) {
> + DRM_ERROR("invalid stream_id:%d\n", msm_dp_catalog->stream_id);
> + return;
> + }
Please drop extra-protective handling. How can stream_id become invalid?
> +
> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
> + reg_offset = REG_DP1_MISC1_MISC0 - REG_DP_MISC1_MISC0;
> +
> + misc_val = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0 + reg_offset);
>
> /* clear bpp bits */
> misc_val &= ~(0x07 << DP_MISC0_TEST_BITS_DEPTH_SHIFT);
> @@ -470,7 +480,7 @@ void msm_dp_catalog_ctrl_config_misc(struct msm_dp_catalog *msm_dp_catalog,
> misc_val |= DP_MISC0_SYNCHRONOUS_CLK;
>
> drm_dbg_dp(catalog->drm_dev, "misc settings = 0x%x\n", misc_val);
> - msm_dp_write_link(catalog, REG_DP_MISC1_MISC0, misc_val);
> + msm_dp_write_link(catalog, REG_DP_MISC1_MISC0 + reg_offset, misc_val);
> }
>
> void msm_dp_catalog_setup_peripheral_flush(struct msm_dp_catalog *msm_dp_catalog)
> @@ -500,10 +510,21 @@ void msm_dp_catalog_ctrl_config_msa(struct msm_dp_catalog *msm_dp_catalog,
> u32 const link_rate_hbr2 = 540000;
> u32 const link_rate_hbr3 = 810000;
> unsigned long den, num;
> + u32 mvid_reg_off = 0, nvid_reg_off = 0;
>
> struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
> struct msm_dp_catalog_private, msm_dp_catalog);
>
> + if (msm_dp_catalog->stream_id >= DP_STREAM_MAX) {
> + DRM_ERROR("invalid stream_id:%d\n", msm_dp_catalog->stream_id);
> + return;
> + }
> +
> + if (msm_dp_catalog->stream_id == DP_STREAM_1) {
> + mvid_reg_off = REG_DP1_SOFTWARE_MVID - REG_DP_SOFTWARE_MVID;
> + nvid_reg_off = REG_DP1_SOFTWARE_NVID - REG_DP_SOFTWARE_NVID;
> + }
> +
> if (rate == link_rate_hbr3)
> pixel_div = 6;
> else if (rate == 162000 || rate == 270000)
> @@ -545,9 +566,14 @@ void msm_dp_catalog_ctrl_config_msa(struct msm_dp_catalog *msm_dp_catalog,
> nvid *= 3;
>
> drm_dbg_dp(catalog->drm_dev, "mvid=0x%x, nvid=0x%x\n", mvid, nvid);
> - msm_dp_write_link(catalog, REG_DP_SOFTWARE_MVID, mvid);
> - msm_dp_write_link(catalog, REG_DP_SOFTWARE_NVID, nvid);
> - msm_dp_write_p0(catalog, MMSS_DP_DSC_DTO, 0x0);
> +
> + msm_dp_write_link(catalog, REG_DP_SOFTWARE_MVID + mvid_reg_off, mvid);
> + msm_dp_write_link(catalog, REG_DP_SOFTWARE_NVID + nvid_reg_off, nvid);
> +
> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
> + msm_dp_write_p0(catalog, MMSS_DP_DSC_DTO, 0x0);
> + else
> + msm_dp_write_p1(catalog, MMSS_DP_DSC_DTO, 0x0);
> }
>
> int msm_dp_catalog_ctrl_set_pattern_state_bit(struct msm_dp_catalog *msm_dp_catalog,
> @@ -910,13 +936,20 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
> struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
> struct msm_dp_catalog_private, msm_dp_catalog);
> u32 reg;
> + u32 offset = 0;
> +
> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
> + offset = REG_DP1_TOTAL_HOR_VER - REG_DP_TOTAL_HOR_VER;
>
> - msm_dp_write_link(catalog, REG_DP_TOTAL_HOR_VER, total);
> - msm_dp_write_link(catalog, REG_DP_START_HOR_VER_FROM_SYNC, sync_start);
> - msm_dp_write_link(catalog, REG_DP_HSYNC_VSYNC_WIDTH_POLARITY, width_blanking);
> - msm_dp_write_link(catalog, REG_DP_ACTIVE_HOR_VER, msm_dp_active);
> + msm_dp_write_link(catalog, REG_DP_TOTAL_HOR_VER + offset, total);
> + msm_dp_write_link(catalog, REG_DP_START_HOR_VER_FROM_SYNC + offset, sync_start);
> + msm_dp_write_link(catalog, REG_DP_HSYNC_VSYNC_WIDTH_POLARITY + offset, width_blanking);
> + msm_dp_write_link(catalog, REG_DP_ACTIVE_HOR_VER + offset, msm_dp_active);
>
> - reg = msm_dp_read_p0(catalog, MMSS_DP_INTF_CONFIG);
> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
> + reg = msm_dp_read_p0(catalog, MMSS_DP_INTF_CONFIG);
> + else
> + reg = msm_dp_read_p1(catalog, MMSS_DP_INTF_CONFIG);
>
> if (msm_dp_catalog->wide_bus_en)
> reg |= DP_INTF_CONFIG_DATABUS_WIDEN;
> @@ -926,7 +959,11 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
>
> DRM_DEBUG_DP("wide_bus_en=%d reg=%#x\n", msm_dp_catalog->wide_bus_en, reg);
>
> - msm_dp_write_p0(catalog, MMSS_DP_INTF_CONFIG, reg);
> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
> + msm_dp_write_p0(catalog, MMSS_DP_INTF_CONFIG, reg);
> + else
> + msm_dp_write_p1(catalog, MMSS_DP_INTF_CONFIG, reg);
> +
> return 0;
> }
>
> @@ -936,18 +973,22 @@ static void msm_dp_catalog_panel_send_vsc_sdp(struct msm_dp_catalog *msm_dp_cata
> u32 header[2];
> u32 val;
> int i;
> + u32 msm_dp_generic_offset = 0;
>
> catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
>
> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
> + msm_dp_generic_offset = MMSS_DP1_GENERIC0_0 - MMSS_DP_GENERIC0_0;
> +
> msm_dp_utils_pack_sdp_header(&vsc_sdp->sdp_header, header);
>
> - msm_dp_write_link(catalog, MMSS_DP_GENERIC0_0, header[0]);
> - msm_dp_write_link(catalog, MMSS_DP_GENERIC0_1, header[1]);
> + msm_dp_write_link(catalog, MMSS_DP_GENERIC0_0 + msm_dp_generic_offset, header[0]);
> + msm_dp_write_link(catalog, MMSS_DP_GENERIC0_1 + msm_dp_generic_offset, header[1]);
>
> for (i = 0; i < sizeof(vsc_sdp->db); i += 4) {
> val = ((vsc_sdp->db[i]) | (vsc_sdp->db[i + 1] << 8) | (vsc_sdp->db[i + 2] << 16) |
> (vsc_sdp->db[i + 3] << 24));
> - msm_dp_write_link(catalog, MMSS_DP_GENERIC0_2 + i, val);
> + msm_dp_write_link(catalog, MMSS_DP_GENERIC0_2 + i + msm_dp_generic_offset, val);
> }
> }
>
> @@ -955,13 +996,17 @@ static void msm_dp_catalog_panel_update_sdp(struct msm_dp_catalog *msm_dp_catalo
> {
> struct msm_dp_catalog_private *catalog;
> u32 hw_revision;
> + u32 sdp_cfg3_offset = 0;
>
> catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
>
> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
> + sdp_cfg3_offset = MMSS_DP1_SDP_CFG3 - MMSS_DP_SDP_CFG3;
> +
> hw_revision = msm_dp_catalog_hw_revision(msm_dp_catalog);
> if (hw_revision < DP_HW_VERSION_1_2 && hw_revision >= DP_HW_VERSION_1_0) {
> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3, 0x01);
> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3, 0x00);
> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3 + sdp_cfg3_offset, 0x01);
> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3 + sdp_cfg3_offset, 0x00);
> }
> }
>
> @@ -969,18 +1014,27 @@ void msm_dp_catalog_panel_enable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog,
> {
> struct msm_dp_catalog_private *catalog;
> u32 cfg, cfg2, misc;
> + u32 misc_reg_offset = 0;
> + u32 sdp_cfg_offset = 0;
> + u32 sdp_cfg2_offset = 0;
>
> catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
>
> - cfg = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG);
> - cfg2 = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG2);
> - misc = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0);
> + if (msm_dp_catalog->stream_id == DP_STREAM_1) {
> + misc_reg_offset = REG_DP1_MISC1_MISC0 - REG_DP_MISC1_MISC0;
> + sdp_cfg_offset = MMSS_DP1_SDP_CFG - MMSS_DP_SDP_CFG;
> + sdp_cfg2_offset = MMSS_DP1_SDP_CFG2 - MMSS_DP_SDP_CFG2;
> + }
> +
> + cfg = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG + sdp_cfg_offset);
> + cfg2 = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG2 + sdp_cfg2_offset);
> + misc = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0 + misc_reg_offset);
>
> cfg |= GEN0_SDP_EN;
> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG, cfg);
> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG + sdp_cfg_offset, cfg);
>
> cfg2 |= GENERIC0_SDPSIZE_VALID;
> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG2, cfg2);
> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG2 + sdp_cfg2_offset, cfg2);
>
> msm_dp_catalog_panel_send_vsc_sdp(msm_dp_catalog, vsc_sdp);
>
> @@ -990,7 +1044,8 @@ void msm_dp_catalog_panel_enable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog,
> drm_dbg_dp(catalog->drm_dev, "vsc sdp enable=1\n");
>
> pr_debug("misc settings = 0x%x\n", misc);
> - msm_dp_write_link(catalog, REG_DP_MISC1_MISC0, misc);
> +
> + msm_dp_write_link(catalog, REG_DP_MISC1_MISC0 + misc_reg_offset, misc);
>
> msm_dp_catalog_panel_update_sdp(msm_dp_catalog);
> }
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
> index edeebf1f313f50e9c54feee1e5aa6aa2dbba3058..c020b7cfa008241e937f6a53764b136431f1dbd9 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
> @@ -47,10 +47,6 @@ enum msm_dp_catalog_audio_header_type {
> DP_AUDIO_SDP_HEADER_MAX,
> };
>
> -struct msm_dp_catalog {
> - bool wide_bus_en;
> -};
> -
> /* stream id */
> enum msm_dp_stream_id {
> DP_STREAM_0,
> @@ -60,6 +56,11 @@ enum msm_dp_stream_id {
> DP_STREAM_MAX,
> };
>
> +struct msm_dp_catalog {
> + bool wide_bus_en;
> + enum msm_dp_stream_id stream_id;
> +};
> +
The same can be achieved by moving enum msm_dp_stream_id up in one of
the earlier patches.
> /* Debug module */
> void msm_dp_catalog_snapshot(struct msm_dp_catalog *msm_dp_catalog, struct msm_disp_state *disp_state);
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 0648831df956dfc7afa1cbfb0dea2c32b02ff74e..ba39b009032dd6f5cb708988963cd6acb6838e4a 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -179,6 +179,7 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
> struct msm_dp_panel *msm_dp_panel)
> {
> u32 cc, tb;
> + ctrl->catalog->stream_id = msm_dp_panel->stream_id;
>
> msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
> msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
> @@ -2062,7 +2063,9 @@ void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_d
> struct msm_dp_ctrl_private *ctrl;
>
> ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> + ctrl->catalog->stream_id = dp_panel->stream_id;
> msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
> +
> }
>
> void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl)
> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
> index 172de804dec445cb08ad8e3f058407f483cd6684..662bf02b8b1a5165f927835bef3c11ac091ddce6 100644
> --- a/drivers/gpu/drm/msm/dp/dp_panel.c
> +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
> @@ -309,7 +309,9 @@ static int msm_dp_panel_setup_vsc_sdp_yuv_420(struct msm_dp_panel *msm_dp_panel)
>
> panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
> catalog = panel->catalog;
> +
> msm_dp_mode = &msm_dp_panel->msm_dp_mode;
> + catalog->stream_id = msm_dp_panel->stream_id;
Why is it a proper place to set catalog->stream_id? It doesn't looks
like it to me.
>
> memset(&vsc_sdp_data, 0, sizeof(vsc_sdp_data));
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
> index 3835c7f5cb984406f8fc52ea765ef2315e0d175b..6c534fde6034fced2cb428e9a29de31ed5c5fcc4 100644
> --- a/drivers/gpu/drm/msm/dp/dp_reg.h
> +++ b/drivers/gpu/drm/msm/dp/dp_reg.h
> @@ -138,13 +138,17 @@
> #define DP_CONFIGURATION_CTRL_LSCLK_DIV_SHIFT (0x0D)
>
> #define REG_DP_SOFTWARE_MVID (0x00000010)
> +#define REG_DP1_SOFTWARE_MVID (0x00000414)
> #define REG_DP_SOFTWARE_NVID (0x00000018)
> +#define REG_DP1_SOFTWARE_NVID (0x00000418)
> #define REG_DP_TOTAL_HOR_VER (0x0000001C)
> +#define REG_DP1_TOTAL_HOR_VER (0x0000041C)
> #define REG_DP_START_HOR_VER_FROM_SYNC (0x00000020)
> #define REG_DP_HSYNC_VSYNC_WIDTH_POLARITY (0x00000024)
> #define REG_DP_ACTIVE_HOR_VER (0x00000028)
> -
> #define REG_DP_MISC1_MISC0 (0x0000002C)
> +#define REG_DP1_MISC1_MISC0 (0x0000042C)
> +
> #define DP_MISC0_SYNCHRONOUS_CLK (0x00000001)
> #define DP_MISC0_COLORIMETRY_CFG_SHIFT (0x00000001)
> #define DP_MISC0_TEST_BITS_DEPTH_SHIFT (0x00000005)
> @@ -211,8 +215,11 @@
> #define MMSS_DP_AUDIO_CTRL_RESET (0x00000214)
>
> #define MMSS_DP_SDP_CFG (0x00000228)
> +#define MMSS_DP1_SDP_CFG (0x000004E0)
> #define GEN0_SDP_EN (0x00020000)
> #define MMSS_DP_SDP_CFG2 (0x0000022C)
> +#define MMSS_DP1_SDP_CFG2 (0x000004E4)
> +
> #define MMSS_DP_AUDIO_TIMESTAMP_0 (0x00000230)
> #define MMSS_DP_AUDIO_TIMESTAMP_1 (0x00000234)
> #define GENERIC0_SDPSIZE_VALID (0x00010000)
> @@ -221,6 +228,8 @@
> #define MMSS_DP_AUDIO_STREAM_1 (0x00000244)
>
> #define MMSS_DP_SDP_CFG3 (0x0000024c)
> +#define MMSS_DP1_SDP_CFG3 (0x000004E8)
> +
> #define UPDATE_SDP (0x00000001)
>
> #define MMSS_DP_EXTENSION_0 (0x00000250)
> @@ -270,6 +279,8 @@
> #define MMSS_DP_GENERIC1_8 (0x00000348)
> #define MMSS_DP_GENERIC1_9 (0x0000034C)
>
> +#define MMSS_DP1_GENERIC0_0 (0x00000490)
> +
> #define MMSS_DP_VSCEXT_0 (0x000002D0)
> #define MMSS_DP_VSCEXT_1 (0x000002D4)
> #define MMSS_DP_VSCEXT_2 (0x000002D8)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 17/45] drm/msm/dp: use stream_id to change offsets in dp_catalog
2024-12-08 5:42 ` Dmitry Baryshkov
@ 2025-05-26 12:57 ` Yongxing Mou
2025-05-28 16:17 ` Dmitry Baryshkov
0 siblings, 1 reply; 111+ messages in thread
From: Yongxing Mou @ 2025-05-26 12:57 UTC (permalink / raw)
To: Dmitry Baryshkov, Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On 2024/12/8 13:42, Dmitry Baryshkov wrote:
> On Thu, Dec 05, 2024 at 08:31:48PM -0800, Abhinav Kumar wrote:
>> Use the dp_panel's stream_id to adjust the offsets for stream 1
>> which will be used for MST in the dp_catalog. Also add additional
>> register defines for stream 1.
>>
>> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
>> ---
>> drivers/gpu/drm/msm/dp/dp_catalog.c | 99 ++++++++++++++++++++++++++++---------
>> drivers/gpu/drm/msm/dp/dp_catalog.h | 9 ++--
>> drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 ++
>> drivers/gpu/drm/msm/dp/dp_panel.c | 2 +
>> drivers/gpu/drm/msm/dp/dp_reg.h | 13 ++++-
>> 5 files changed, 99 insertions(+), 27 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
>> index ee7f2d0b23aa034428a01ef2c9752f51013c5e01..e6f6edf617898241c74580eb0ae6bc58f06a154f 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
>> @@ -457,10 +457,20 @@ void msm_dp_catalog_ctrl_config_misc(struct msm_dp_catalog *msm_dp_catalog,
>> u32 test_bits_depth)
>> {
>> u32 misc_val;
>> + u32 reg_offset = 0;
>> +
>> struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
>> struct msm_dp_catalog_private, msm_dp_catalog);
>>
>> - misc_val = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0);
>> + if (msm_dp_catalog->stream_id >= DP_STREAM_MAX) {
>> + DRM_ERROR("invalid stream_id:%d\n", msm_dp_catalog->stream_id);
>> + return;
>> + }
>
> Please drop extra-protective handling. How can stream_id become invalid?
>
>> +
>> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
>> + reg_offset = REG_DP1_MISC1_MISC0 - REG_DP_MISC1_MISC0;
>> +
>> + misc_val = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0 + reg_offset);
>>
>> /* clear bpp bits */
>> misc_val &= ~(0x07 << DP_MISC0_TEST_BITS_DEPTH_SHIFT);
>> @@ -470,7 +480,7 @@ void msm_dp_catalog_ctrl_config_misc(struct msm_dp_catalog *msm_dp_catalog,
>> misc_val |= DP_MISC0_SYNCHRONOUS_CLK;
>>
>> drm_dbg_dp(catalog->drm_dev, "misc settings = 0x%x\n", misc_val);
>> - msm_dp_write_link(catalog, REG_DP_MISC1_MISC0, misc_val);
>> + msm_dp_write_link(catalog, REG_DP_MISC1_MISC0 + reg_offset, misc_val);
>> }
>>
>> void msm_dp_catalog_setup_peripheral_flush(struct msm_dp_catalog *msm_dp_catalog)
>> @@ -500,10 +510,21 @@ void msm_dp_catalog_ctrl_config_msa(struct msm_dp_catalog *msm_dp_catalog,
>> u32 const link_rate_hbr2 = 540000;
>> u32 const link_rate_hbr3 = 810000;
>> unsigned long den, num;
>> + u32 mvid_reg_off = 0, nvid_reg_off = 0;
>>
>> struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
>> struct msm_dp_catalog_private, msm_dp_catalog);
>>
>> + if (msm_dp_catalog->stream_id >= DP_STREAM_MAX) {
>> + DRM_ERROR("invalid stream_id:%d\n", msm_dp_catalog->stream_id);
>> + return;
>> + }
>> +
>> + if (msm_dp_catalog->stream_id == DP_STREAM_1) {
>> + mvid_reg_off = REG_DP1_SOFTWARE_MVID - REG_DP_SOFTWARE_MVID;
>> + nvid_reg_off = REG_DP1_SOFTWARE_NVID - REG_DP_SOFTWARE_NVID;
>> + }
>> +
>> if (rate == link_rate_hbr3)
>> pixel_div = 6;
>> else if (rate == 162000 || rate == 270000)
>> @@ -545,9 +566,14 @@ void msm_dp_catalog_ctrl_config_msa(struct msm_dp_catalog *msm_dp_catalog,
>> nvid *= 3;
>>
>> drm_dbg_dp(catalog->drm_dev, "mvid=0x%x, nvid=0x%x\n", mvid, nvid);
>> - msm_dp_write_link(catalog, REG_DP_SOFTWARE_MVID, mvid);
>> - msm_dp_write_link(catalog, REG_DP_SOFTWARE_NVID, nvid);
>> - msm_dp_write_p0(catalog, MMSS_DP_DSC_DTO, 0x0);
>> +
>> + msm_dp_write_link(catalog, REG_DP_SOFTWARE_MVID + mvid_reg_off, mvid);
>> + msm_dp_write_link(catalog, REG_DP_SOFTWARE_NVID + nvid_reg_off, nvid);
>> +
>> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
>> + msm_dp_write_p0(catalog, MMSS_DP_DSC_DTO, 0x0);
>> + else
>> + msm_dp_write_p1(catalog, MMSS_DP_DSC_DTO, 0x0);
>> }
>>
>> int msm_dp_catalog_ctrl_set_pattern_state_bit(struct msm_dp_catalog *msm_dp_catalog,
>> @@ -910,13 +936,20 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
>> struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
>> struct msm_dp_catalog_private, msm_dp_catalog);
>> u32 reg;
>> + u32 offset = 0;
>> +
>> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
>> + offset = REG_DP1_TOTAL_HOR_VER - REG_DP_TOTAL_HOR_VER;
>>
>> - msm_dp_write_link(catalog, REG_DP_TOTAL_HOR_VER, total);
>> - msm_dp_write_link(catalog, REG_DP_START_HOR_VER_FROM_SYNC, sync_start);
>> - msm_dp_write_link(catalog, REG_DP_HSYNC_VSYNC_WIDTH_POLARITY, width_blanking);
>> - msm_dp_write_link(catalog, REG_DP_ACTIVE_HOR_VER, msm_dp_active);
>> + msm_dp_write_link(catalog, REG_DP_TOTAL_HOR_VER + offset, total);
>> + msm_dp_write_link(catalog, REG_DP_START_HOR_VER_FROM_SYNC + offset, sync_start);
>> + msm_dp_write_link(catalog, REG_DP_HSYNC_VSYNC_WIDTH_POLARITY + offset, width_blanking);
>> + msm_dp_write_link(catalog, REG_DP_ACTIVE_HOR_VER + offset, msm_dp_active);
>>
>> - reg = msm_dp_read_p0(catalog, MMSS_DP_INTF_CONFIG);
>> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
>> + reg = msm_dp_read_p0(catalog, MMSS_DP_INTF_CONFIG);
>> + else
>> + reg = msm_dp_read_p1(catalog, MMSS_DP_INTF_CONFIG);
>>
>> if (msm_dp_catalog->wide_bus_en)
>> reg |= DP_INTF_CONFIG_DATABUS_WIDEN;
>> @@ -926,7 +959,11 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
>>
>> DRM_DEBUG_DP("wide_bus_en=%d reg=%#x\n", msm_dp_catalog->wide_bus_en, reg);
>>
>> - msm_dp_write_p0(catalog, MMSS_DP_INTF_CONFIG, reg);
>> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
>> + msm_dp_write_p0(catalog, MMSS_DP_INTF_CONFIG, reg);
>> + else
>> + msm_dp_write_p1(catalog, MMSS_DP_INTF_CONFIG, reg);
>> +
>> return 0;
>> }
>>
>> @@ -936,18 +973,22 @@ static void msm_dp_catalog_panel_send_vsc_sdp(struct msm_dp_catalog *msm_dp_cata
>> u32 header[2];
>> u32 val;
>> int i;
>> + u32 msm_dp_generic_offset = 0;
>>
>> catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
>>
>> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
>> + msm_dp_generic_offset = MMSS_DP1_GENERIC0_0 - MMSS_DP_GENERIC0_0;
>> +
>> msm_dp_utils_pack_sdp_header(&vsc_sdp->sdp_header, header);
>>
>> - msm_dp_write_link(catalog, MMSS_DP_GENERIC0_0, header[0]);
>> - msm_dp_write_link(catalog, MMSS_DP_GENERIC0_1, header[1]);
>> + msm_dp_write_link(catalog, MMSS_DP_GENERIC0_0 + msm_dp_generic_offset, header[0]);
>> + msm_dp_write_link(catalog, MMSS_DP_GENERIC0_1 + msm_dp_generic_offset, header[1]);
>>
>> for (i = 0; i < sizeof(vsc_sdp->db); i += 4) {
>> val = ((vsc_sdp->db[i]) | (vsc_sdp->db[i + 1] << 8) | (vsc_sdp->db[i + 2] << 16) |
>> (vsc_sdp->db[i + 3] << 24));
>> - msm_dp_write_link(catalog, MMSS_DP_GENERIC0_2 + i, val);
>> + msm_dp_write_link(catalog, MMSS_DP_GENERIC0_2 + i + msm_dp_generic_offset, val);
>> }
>> }
>>
>> @@ -955,13 +996,17 @@ static void msm_dp_catalog_panel_update_sdp(struct msm_dp_catalog *msm_dp_catalo
>> {
>> struct msm_dp_catalog_private *catalog;
>> u32 hw_revision;
>> + u32 sdp_cfg3_offset = 0;
>>
>> catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
>>
>> + if (msm_dp_catalog->stream_id == DP_STREAM_1)
>> + sdp_cfg3_offset = MMSS_DP1_SDP_CFG3 - MMSS_DP_SDP_CFG3;
>> +
>> hw_revision = msm_dp_catalog_hw_revision(msm_dp_catalog);
>> if (hw_revision < DP_HW_VERSION_1_2 && hw_revision >= DP_HW_VERSION_1_0) {
>> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3, 0x01);
>> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3, 0x00);
>> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3 + sdp_cfg3_offset, 0x01);
>> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG3 + sdp_cfg3_offset, 0x00);
>> }
>> }
>>
>> @@ -969,18 +1014,27 @@ void msm_dp_catalog_panel_enable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog,
>> {
>> struct msm_dp_catalog_private *catalog;
>> u32 cfg, cfg2, misc;
>> + u32 misc_reg_offset = 0;
>> + u32 sdp_cfg_offset = 0;
>> + u32 sdp_cfg2_offset = 0;
>>
>> catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
>>
>> - cfg = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG);
>> - cfg2 = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG2);
>> - misc = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0);
>> + if (msm_dp_catalog->stream_id == DP_STREAM_1) {
>> + misc_reg_offset = REG_DP1_MISC1_MISC0 - REG_DP_MISC1_MISC0;
>> + sdp_cfg_offset = MMSS_DP1_SDP_CFG - MMSS_DP_SDP_CFG;
>> + sdp_cfg2_offset = MMSS_DP1_SDP_CFG2 - MMSS_DP_SDP_CFG2;
>> + }
>> +
>> + cfg = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG + sdp_cfg_offset);
>> + cfg2 = msm_dp_read_link(catalog, MMSS_DP_SDP_CFG2 + sdp_cfg2_offset);
>> + misc = msm_dp_read_link(catalog, REG_DP_MISC1_MISC0 + misc_reg_offset);
>>
>> cfg |= GEN0_SDP_EN;
>> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG, cfg);
>> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG + sdp_cfg_offset, cfg);
>>
>> cfg2 |= GENERIC0_SDPSIZE_VALID;
>> - msm_dp_write_link(catalog, MMSS_DP_SDP_CFG2, cfg2);
>> + msm_dp_write_link(catalog, MMSS_DP_SDP_CFG2 + sdp_cfg2_offset, cfg2);
>>
>> msm_dp_catalog_panel_send_vsc_sdp(msm_dp_catalog, vsc_sdp);
>>
>> @@ -990,7 +1044,8 @@ void msm_dp_catalog_panel_enable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog,
>> drm_dbg_dp(catalog->drm_dev, "vsc sdp enable=1\n");
>>
>> pr_debug("misc settings = 0x%x\n", misc);
>> - msm_dp_write_link(catalog, REG_DP_MISC1_MISC0, misc);
>> +
>> + msm_dp_write_link(catalog, REG_DP_MISC1_MISC0 + misc_reg_offset, misc);
>>
>> msm_dp_catalog_panel_update_sdp(msm_dp_catalog);
>> }
>> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
>> index edeebf1f313f50e9c54feee1e5aa6aa2dbba3058..c020b7cfa008241e937f6a53764b136431f1dbd9 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
>> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
>> @@ -47,10 +47,6 @@ enum msm_dp_catalog_audio_header_type {
>> DP_AUDIO_SDP_HEADER_MAX,
>> };
>>
>> -struct msm_dp_catalog {
>> - bool wide_bus_en;
>> -};
>> -
>> /* stream id */
>> enum msm_dp_stream_id {
>> DP_STREAM_0,
>> @@ -60,6 +56,11 @@ enum msm_dp_stream_id {
>> DP_STREAM_MAX,
>> };
>>
>> +struct msm_dp_catalog {
>> + bool wide_bus_en;
>> + enum msm_dp_stream_id stream_id;
>> +};
>> +
>
> The same can be achieved by moving enum msm_dp_stream_id up in one of
> the earlier patches.
>
>> /* Debug module */
>> void msm_dp_catalog_snapshot(struct msm_dp_catalog *msm_dp_catalog, struct msm_disp_state *disp_state);
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
>> index 0648831df956dfc7afa1cbfb0dea2c32b02ff74e..ba39b009032dd6f5cb708988963cd6acb6838e4a 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
>> @@ -179,6 +179,7 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
>> struct msm_dp_panel *msm_dp_panel)
>> {
>> u32 cc, tb;
>> + ctrl->catalog->stream_id = msm_dp_panel->stream_id;
>>
>> msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
>> msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
>> @@ -2062,7 +2063,9 @@ void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_d
>> struct msm_dp_ctrl_private *ctrl;
>>
>> ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
>> + ctrl->catalog->stream_id = dp_panel->stream_id;
>> msm_dp_catalog_panel_disable_vsc_sdp(ctrl->catalog);
>> +
>> }
>>
>> void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl)
>> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
>> index 172de804dec445cb08ad8e3f058407f483cd6684..662bf02b8b1a5165f927835bef3c11ac091ddce6 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_panel.c
>> +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
>> @@ -309,7 +309,9 @@ static int msm_dp_panel_setup_vsc_sdp_yuv_420(struct msm_dp_panel *msm_dp_panel)
>>
>> panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
>> catalog = panel->catalog;
>> +
>> msm_dp_mode = &msm_dp_panel->msm_dp_mode;
>> + catalog->stream_id = msm_dp_panel->stream_id;
>
> Why is it a proper place to set catalog->stream_id? It doesn't looks
> like it to me.
Ok, maybe msm_dp_display_set_stream_id is more proper place. Or can we
drop stream_id in catalog totally, and f the stream_id is needed in the
catalog function, pass it as a parameter to the catalog function. just
like that:
int msm_dp_ctrl_***(struct msm_dp_ctrl *ctrl, enum msm_dp_stream_id
stream_id,***);
>
>>
>> memset(&vsc_sdp_data, 0, sizeof(vsc_sdp_data));
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
>> index 3835c7f5cb984406f8fc52ea765ef2315e0d175b..6c534fde6034fced2cb428e9a29de31ed5c5fcc4 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_reg.h
>> +++ b/drivers/gpu/drm/msm/dp/dp_reg.h
>> @@ -138,13 +138,17 @@
>> #define DP_CONFIGURATION_CTRL_LSCLK_DIV_SHIFT (0x0D)
>>
>> #define REG_DP_SOFTWARE_MVID (0x00000010)
>> +#define REG_DP1_SOFTWARE_MVID (0x00000414)
>> #define REG_DP_SOFTWARE_NVID (0x00000018)
>> +#define REG_DP1_SOFTWARE_NVID (0x00000418)
>> #define REG_DP_TOTAL_HOR_VER (0x0000001C)
>> +#define REG_DP1_TOTAL_HOR_VER (0x0000041C)
>> #define REG_DP_START_HOR_VER_FROM_SYNC (0x00000020)
>> #define REG_DP_HSYNC_VSYNC_WIDTH_POLARITY (0x00000024)
>> #define REG_DP_ACTIVE_HOR_VER (0x00000028)
>> -
>> #define REG_DP_MISC1_MISC0 (0x0000002C)
>> +#define REG_DP1_MISC1_MISC0 (0x0000042C)
>> +
>> #define DP_MISC0_SYNCHRONOUS_CLK (0x00000001)
>> #define DP_MISC0_COLORIMETRY_CFG_SHIFT (0x00000001)
>> #define DP_MISC0_TEST_BITS_DEPTH_SHIFT (0x00000005)
>> @@ -211,8 +215,11 @@
>> #define MMSS_DP_AUDIO_CTRL_RESET (0x00000214)
>>
>> #define MMSS_DP_SDP_CFG (0x00000228)
>> +#define MMSS_DP1_SDP_CFG (0x000004E0)
>> #define GEN0_SDP_EN (0x00020000)
>> #define MMSS_DP_SDP_CFG2 (0x0000022C)
>> +#define MMSS_DP1_SDP_CFG2 (0x000004E4)
>> +
>> #define MMSS_DP_AUDIO_TIMESTAMP_0 (0x00000230)
>> #define MMSS_DP_AUDIO_TIMESTAMP_1 (0x00000234)
>> #define GENERIC0_SDPSIZE_VALID (0x00010000)
>> @@ -221,6 +228,8 @@
>> #define MMSS_DP_AUDIO_STREAM_1 (0x00000244)
>>
>> #define MMSS_DP_SDP_CFG3 (0x0000024c)
>> +#define MMSS_DP1_SDP_CFG3 (0x000004E8)
>> +
>> #define UPDATE_SDP (0x00000001)
>>
>> #define MMSS_DP_EXTENSION_0 (0x00000250)
>> @@ -270,6 +279,8 @@
>> #define MMSS_DP_GENERIC1_8 (0x00000348)
>> #define MMSS_DP_GENERIC1_9 (0x0000034C)
>>
>> +#define MMSS_DP1_GENERIC0_0 (0x00000490)
>> +
>> #define MMSS_DP_VSCEXT_0 (0x000002D0)
>> #define MMSS_DP_VSCEXT_1 (0x000002D4)
>> #define MMSS_DP_VSCEXT_2 (0x000002D8)
>>
>> --
>> 2.34.1
>>
>
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 17/45] drm/msm/dp: use stream_id to change offsets in dp_catalog
2025-05-26 12:57 ` Yongxing Mou
@ 2025-05-28 16:17 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2025-05-28 16:17 UTC (permalink / raw)
To: Yongxing Mou
Cc: Abhinav Kumar, Rob Clark, Sean Paul, Marijn Suijten, David Airlie,
Simona Vetter, Stephen Boyd, Chandan Uddaraju, Guenter Roeck,
Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On Mon, May 26, 2025 at 08:57:35PM +0800, Yongxing Mou wrote:
>
>
> On 2024/12/8 13:42, Dmitry Baryshkov wrote:
> > On Thu, Dec 05, 2024 at 08:31:48PM -0800, Abhinav Kumar wrote:
> > > Use the dp_panel's stream_id to adjust the offsets for stream 1
> > > which will be used for MST in the dp_catalog. Also add additional
> > > register defines for stream 1.
> > >
> > > Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> > > ---
> > > drivers/gpu/drm/msm/dp/dp_catalog.c | 99 ++++++++++++++++++++++++++++---------
> > > drivers/gpu/drm/msm/dp/dp_catalog.h | 9 ++--
> > > drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 ++
> > > drivers/gpu/drm/msm/dp/dp_panel.c | 2 +
> > > drivers/gpu/drm/msm/dp/dp_reg.h | 13 ++++-
> > > 5 files changed, 99 insertions(+), 27 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_panel.c b/drivers/gpu/drm/msm/dp/dp_panel.c
> > > index 172de804dec445cb08ad8e3f058407f483cd6684..662bf02b8b1a5165f927835bef3c11ac091ddce6 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_panel.c
> > > +++ b/drivers/gpu/drm/msm/dp/dp_panel.c
> > > @@ -309,7 +309,9 @@ static int msm_dp_panel_setup_vsc_sdp_yuv_420(struct msm_dp_panel *msm_dp_panel)
> > > panel = container_of(msm_dp_panel, struct msm_dp_panel_private, msm_dp_panel);
> > > catalog = panel->catalog;
> > > +
> > > msm_dp_mode = &msm_dp_panel->msm_dp_mode;
> > > + catalog->stream_id = msm_dp_panel->stream_id;
> >
> > Why is it a proper place to set catalog->stream_id? It doesn't looks
> > like it to me.
> Ok, maybe msm_dp_display_set_stream_id is more proper place. Or can we drop
> stream_id in catalog totally, and f the stream_id is needed in the catalog
> function, pass it as a parameter to the catalog function. just like that:
> int msm_dp_ctrl_***(struct msm_dp_ctrl *ctrl, enum msm_dp_stream_id
> stream_id,***);
LGTM.
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 18/45] drm/msm/dp: add support to send ACT packets for MST
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (16 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 17/45] drm/msm/dp: use stream_id to change offsets in dp_catalog Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 5:45 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 19/45] drm/msm/dp: add support to program mst support in mainlink Abhinav Kumar
` (28 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Whenever virtual channel slot allocation changes, the DP
source must send the action control trigger sequence to notify
the sink about the same. This would be applicable during the
start and stop of the pixel stream. Add the infrastructure
to be able to send ACT packets for the DP controller when
operating in MST mode.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.c | 21 +++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_catalog.h | 4 ++++
drivers/gpu/drm/msm/dp/dp_ctrl.c | 36 ++++++++++++++++++++++++++++++++++--
drivers/gpu/drm/msm/dp/dp_ctrl.h | 3 ++-
drivers/gpu/drm/msm/dp/dp_display.c | 6 +++++-
drivers/gpu/drm/msm/dp/dp_display.h | 1 +
drivers/gpu/drm/msm/dp/dp_reg.h | 3 +++
7 files changed, 70 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index e6f6edf617898241c74580eb0ae6bc58f06a154f..88d6262a972ef2d30c467ef5ff5c58ef3299ae7d 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -1078,6 +1078,27 @@ void msm_dp_catalog_panel_disable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog)
msm_dp_catalog_panel_update_sdp(msm_dp_catalog);
}
+void msm_dp_catalog_trigger_act(struct msm_dp_catalog *msm_dp_catalog)
+{
+ struct msm_dp_catalog_private *catalog;
+
+ catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
+
+ msm_dp_write_link(catalog, REG_DP_MST_ACT, 0x1);
+
+ /* make sure ACT signal is performed */
+ wmb();
+}
+
+bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *msm_dp_catalog)
+{
+ struct msm_dp_catalog_private *catalog;
+
+ catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
+
+ return msm_dp_read_link(catalog, REG_DP_MST_ACT);
+}
+
void msm_dp_catalog_panel_tpg_enable(struct msm_dp_catalog *msm_dp_catalog,
struct drm_display_mode *drm_mode)
{
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
index c020b7cfa008241e937f6a53764b136431f1dbd9..c91c52d40209b8bcb63db9c0256f6ef721dace8a 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.h
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
@@ -135,4 +135,8 @@ void msm_dp_catalog_audio_config_sdp(struct msm_dp_catalog *catalog);
void msm_dp_catalog_audio_init(struct msm_dp_catalog *catalog);
void msm_dp_catalog_audio_sfe_level(struct msm_dp_catalog *catalog, u32 safe_to_exit_level);
+/* DP MST APIs */
+void msm_dp_catalog_trigger_act(struct msm_dp_catalog *dp_catalog);
+bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *dp_catalog);
+
#endif /* _DP_CATALOG_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index ba39b009032dd6f5cb708988963cd6acb6838e4a..2bfe2aac3c02b02b12713dbd98e79ed4a75b85d0 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -97,6 +97,8 @@ struct msm_dp_ctrl_private {
bool core_clks_on;
bool link_clks_on;
+
+ bool mst_active;
bool stream_clks_on[DP_STREAM_MAX];
};
@@ -1625,6 +1627,26 @@ static int msm_dp_ctrl_deinitialize_mainlink(struct msm_dp_ctrl_private *ctrl)
return 0;
}
+int msm_dp_ctrl_mst_send_act(struct msm_dp_ctrl *msm_dp_ctrl)
+{
+ struct msm_dp_ctrl_private *ctrl;
+ bool act_complete;
+
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+
+ msm_dp_catalog_trigger_act(ctrl->catalog);
+ msleep(20); /* needs 1 frame time */
+
+ act_complete = msm_dp_catalog_read_act_complete_sts(ctrl->catalog);
+
+ if (!act_complete)
+ DRM_ERROR("mst ACT trigger complete SUCCESS\n");
+ else
+ drm_dbg_dp(ctrl->drm_dev, "mst ACT trigger complete failed\n");
+
+ return 0;
+}
+
static int msm_dp_ctrl_link_maintenance(struct msm_dp_ctrl_private *ctrl)
{
int ret = 0;
@@ -1643,6 +1665,9 @@ static int msm_dp_ctrl_link_maintenance(struct msm_dp_ctrl_private *ctrl)
msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
+ if (ctrl->mst_active)
+ msm_dp_ctrl_mst_send_act(&ctrl->msm_dp_ctrl);
+
ret = msm_dp_ctrl_wait4video_ready(ctrl);
end:
return ret;
@@ -1745,7 +1770,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl, msm_dp_panel);
msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
- ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
+ ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl, ctrl->mst_active);
if (ret) {
DRM_ERROR("failed to enable DP link controller\n");
return ret;
@@ -1825,7 +1850,7 @@ static bool msm_dp_ctrl_channel_eq_ok(struct msm_dp_ctrl_private *ctrl)
return drm_dp_channel_eq_ok(link_status, num_lanes);
}
-int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl)
+int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl, bool mst_active)
{
int rc = 0;
struct msm_dp_ctrl_private *ctrl;
@@ -1844,6 +1869,7 @@ int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl)
rate = ctrl->panel->link_info.rate;
pixel_rate = ctrl->panel->msm_dp_mode.drm_mode.clock;
+ ctrl->mst_active = mst_active;
msm_dp_ctrl_core_clk_enable(&ctrl->msm_dp_ctrl);
if (ctrl->link->sink_request & DP_TEST_LINK_PHY_TEST_PATTERN) {
@@ -2035,6 +2061,9 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
+ if (ctrl->mst_active)
+ msm_dp_ctrl_mst_send_act(msm_dp_ctrl);
+
ret = msm_dp_ctrl_wait4video_ready(ctrl);
if (ret)
return ret;
@@ -2104,6 +2133,8 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
+ ctrl->mst_active = false;
+
dev_pm_opp_set_rate(ctrl->dev, 0);
msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
@@ -2264,6 +2295,7 @@ struct msm_dp_ctrl *msm_dp_ctrl_get(struct device *dev, struct msm_dp_link *link
ctrl->catalog = catalog;
ctrl->dev = dev;
ctrl->phy = phy;
+ ctrl->mst_active = false;
ret = msm_dp_ctrl_clk_init(&ctrl->msm_dp_ctrl);
if (ret) {
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index d422fd683b65d6f5e459710d0327e472a12c30b0..81c05b1b2baac63e1f1888f3f517e62a98e230a7 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -17,7 +17,7 @@ struct msm_dp_ctrl {
struct phy;
-int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl);
+int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl, bool mst_active);
int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
@@ -46,5 +46,6 @@ void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl,
struct msm_dp_panel *msm_dp_panel);
void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl);
+int msm_dp_ctrl_mst_send_act(struct msm_dp_ctrl *ctrl);
#endif /* _DP_CTRL_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 5fa6c003cf6c51eae77573549a555a00dc33f476..e19860ef3493fb100afbf04b09d14a136fd6b887 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -370,6 +370,7 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
struct drm_connector *connector = dp->msm_dp_display.connector;
const struct drm_display_info *info = &connector->display_info;
int rc = 0;
+ struct msm_dp *dp_display = &dp->msm_dp_display;
rc = msm_dp_panel_read_link_caps(dp->panel, connector);
if (rc)
@@ -399,7 +400,7 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
msm_dp_link_psm_config(dp->link, &dp->panel->link_info, false);
msm_dp_link_reset_phy_params_vx_px(dp->link);
- rc = msm_dp_ctrl_on_link(dp->ctrl);
+ rc = msm_dp_ctrl_on_link(dp->ctrl, dp_display->mst_active);
if (rc) {
DRM_ERROR("failed to complete DP link training\n");
goto end;
@@ -1628,6 +1629,9 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
+
+ if (dp->mst_active)
+ msm_dp_ctrl_mst_send_act(msm_dp_display->ctrl);
}
static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 8ce8ba254b1bfe5b4c000df83eceef5823772780..38ca25491b1ccfd95e027a9c8f659abb3cd576d3 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -22,6 +22,7 @@ struct msm_dp {
bool audio_enabled;
bool power_on;
bool prepared;
+ bool mst_active;
unsigned int connector_type;
bool is_edp;
bool internal_hpd;
diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
index 6c534fde6034fced2cb428e9a29de31ed5c5fcc4..46400973eb499066e9e805b16df759b1db34cf22 100644
--- a/drivers/gpu/drm/msm/dp/dp_reg.h
+++ b/drivers/gpu/drm/msm/dp/dp_reg.h
@@ -330,6 +330,9 @@
#define REG_DP_PHY_AUX_BIST_CFG (0x00000050)
#define REG_DP_PHY_AUX_INTERRUPT_STATUS (0x000000BC)
+/* DP MST related registers */
+#define REG_DP_MST_ACT (0x00000500)
+
/* DP HDCP 1.3 registers */
#define DP_HDCP_CTRL (0x0A0)
#define DP_HDCP_STATUS (0x0A4)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 18/45] drm/msm/dp: add support to send ACT packets for MST
2024-12-06 4:31 ` [PATCH 18/45] drm/msm/dp: add support to send ACT packets for MST Abhinav Kumar
@ 2024-12-08 5:45 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 5:45 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:49PM -0800, Abhinav Kumar wrote:
> Whenever virtual channel slot allocation changes, the DP
> source must send the action control trigger sequence to notify
> the sink about the same. This would be applicable during the
> start and stop of the pixel stream. Add the infrastructure
> to be able to send ACT packets for the DP controller when
> operating in MST mode.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 21 +++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_catalog.h | 4 ++++
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 36 ++++++++++++++++++++++++++++++++++--
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 3 ++-
> drivers/gpu/drm/msm/dp/dp_display.c | 6 +++++-
> drivers/gpu/drm/msm/dp/dp_display.h | 1 +
> drivers/gpu/drm/msm/dp/dp_reg.h | 3 +++
> 7 files changed, 70 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
> index e6f6edf617898241c74580eb0ae6bc58f06a154f..88d6262a972ef2d30c467ef5ff5c58ef3299ae7d 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> @@ -1078,6 +1078,27 @@ void msm_dp_catalog_panel_disable_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog)
> msm_dp_catalog_panel_update_sdp(msm_dp_catalog);
> }
>
> +void msm_dp_catalog_trigger_act(struct msm_dp_catalog *msm_dp_catalog)
> +{
> + struct msm_dp_catalog_private *catalog;
> +
> + catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
> +
> + msm_dp_write_link(catalog, REG_DP_MST_ACT, 0x1);
> +
> + /* make sure ACT signal is performed */
> + wmb();
> +}
> +
> +bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *msm_dp_catalog)
> +{
> + struct msm_dp_catalog_private *catalog;
> +
> + catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
> +
> + return msm_dp_read_link(catalog, REG_DP_MST_ACT);
> +}
> +
Hopefully after [1] this becomes inline in the corresponding dp_ctrl
functions.
[1] https://patchwork.freedesktop.org/patch/626816/?series=141074&rev=2
> void msm_dp_catalog_panel_tpg_enable(struct msm_dp_catalog *msm_dp_catalog,
> struct drm_display_mode *drm_mode)
> {
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
> index c020b7cfa008241e937f6a53764b136431f1dbd9..c91c52d40209b8bcb63db9c0256f6ef721dace8a 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
> @@ -135,4 +135,8 @@ void msm_dp_catalog_audio_config_sdp(struct msm_dp_catalog *catalog);
> void msm_dp_catalog_audio_init(struct msm_dp_catalog *catalog);
> void msm_dp_catalog_audio_sfe_level(struct msm_dp_catalog *catalog, u32 safe_to_exit_level);
>
> +/* DP MST APIs */
> +void msm_dp_catalog_trigger_act(struct msm_dp_catalog *dp_catalog);
> +bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *dp_catalog);
> +
> #endif /* _DP_CATALOG_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index ba39b009032dd6f5cb708988963cd6acb6838e4a..2bfe2aac3c02b02b12713dbd98e79ed4a75b85d0 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -97,6 +97,8 @@ struct msm_dp_ctrl_private {
>
> bool core_clks_on;
> bool link_clks_on;
> +
> + bool mst_active;
> bool stream_clks_on[DP_STREAM_MAX];
> };
>
> @@ -1625,6 +1627,26 @@ static int msm_dp_ctrl_deinitialize_mainlink(struct msm_dp_ctrl_private *ctrl)
> return 0;
> }
>
> +int msm_dp_ctrl_mst_send_act(struct msm_dp_ctrl *msm_dp_ctrl)
> +{
> + struct msm_dp_ctrl_private *ctrl;
> + bool act_complete;
> +
> + ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> +
> + msm_dp_catalog_trigger_act(ctrl->catalog);
> + msleep(20); /* needs 1 frame time */
> +
> + act_complete = msm_dp_catalog_read_act_complete_sts(ctrl->catalog);
> +
> + if (!act_complete)
> + DRM_ERROR("mst ACT trigger complete SUCCESS\n");
> + else
> + drm_dbg_dp(ctrl->drm_dev, "mst ACT trigger complete failed\n");
> +
> + return 0;
> +}
> +
> static int msm_dp_ctrl_link_maintenance(struct msm_dp_ctrl_private *ctrl)
> {
> int ret = 0;
> @@ -1643,6 +1665,9 @@ static int msm_dp_ctrl_link_maintenance(struct msm_dp_ctrl_private *ctrl)
>
> msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
>
> + if (ctrl->mst_active)
> + msm_dp_ctrl_mst_send_act(&ctrl->msm_dp_ctrl);
> +
> ret = msm_dp_ctrl_wait4video_ready(ctrl);
> end:
> return ret;
> @@ -1745,7 +1770,7 @@ static int msm_dp_ctrl_process_phy_test_request(struct msm_dp_ctrl_private *ctrl
> msm_dp_ctrl_stream_clk_off(&ctrl->msm_dp_ctrl, msm_dp_panel);
> msm_dp_ctrl_off_link(&ctrl->msm_dp_ctrl);
>
> - ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl);
> + ret = msm_dp_ctrl_on_link(&ctrl->msm_dp_ctrl, ctrl->mst_active);
> if (ret) {
> DRM_ERROR("failed to enable DP link controller\n");
> return ret;
> @@ -1825,7 +1850,7 @@ static bool msm_dp_ctrl_channel_eq_ok(struct msm_dp_ctrl_private *ctrl)
> return drm_dp_channel_eq_ok(link_status, num_lanes);
> }
>
> -int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl)
> +int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl, bool mst_active)
> {
> int rc = 0;
> struct msm_dp_ctrl_private *ctrl;
> @@ -1844,6 +1869,7 @@ int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl)
> rate = ctrl->panel->link_info.rate;
> pixel_rate = ctrl->panel->msm_dp_mode.drm_mode.clock;
>
> + ctrl->mst_active = mst_active;
> msm_dp_ctrl_core_clk_enable(&ctrl->msm_dp_ctrl);
>
> if (ctrl->link->sink_request & DP_TEST_LINK_PHY_TEST_PATTERN) {
> @@ -2035,6 +2061,9 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
>
> msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
>
> + if (ctrl->mst_active)
> + msm_dp_ctrl_mst_send_act(msm_dp_ctrl);
> +
> ret = msm_dp_ctrl_wait4video_ready(ctrl);
> if (ret)
> return ret;
> @@ -2104,6 +2133,8 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
>
> msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
>
> + ctrl->mst_active = false;
> +
> dev_pm_opp_set_rate(ctrl->dev, 0);
> msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
>
> @@ -2264,6 +2295,7 @@ struct msm_dp_ctrl *msm_dp_ctrl_get(struct device *dev, struct msm_dp_link *link
> ctrl->catalog = catalog;
> ctrl->dev = dev;
> ctrl->phy = phy;
> + ctrl->mst_active = false;
>
> ret = msm_dp_ctrl_clk_init(&ctrl->msm_dp_ctrl);
> if (ret) {
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index d422fd683b65d6f5e459710d0327e472a12c30b0..81c05b1b2baac63e1f1888f3f517e62a98e230a7 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -17,7 +17,7 @@ struct msm_dp_ctrl {
>
> struct phy;
>
> -int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl);
> +int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl, bool mst_active);
> int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
> int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
> void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> @@ -46,5 +46,6 @@ void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl,
> struct msm_dp_panel *msm_dp_panel);
> void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl);
> +int msm_dp_ctrl_mst_send_act(struct msm_dp_ctrl *ctrl);
>
> #endif /* _DP_CTRL_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 5fa6c003cf6c51eae77573549a555a00dc33f476..e19860ef3493fb100afbf04b09d14a136fd6b887 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -370,6 +370,7 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> struct drm_connector *connector = dp->msm_dp_display.connector;
> const struct drm_display_info *info = &connector->display_info;
> int rc = 0;
> + struct msm_dp *dp_display = &dp->msm_dp_display;
>
> rc = msm_dp_panel_read_link_caps(dp->panel, connector);
> if (rc)
> @@ -399,7 +400,7 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> msm_dp_link_psm_config(dp->link, &dp->panel->link_info, false);
>
> msm_dp_link_reset_phy_params_vx_px(dp->link);
> - rc = msm_dp_ctrl_on_link(dp->ctrl);
> + rc = msm_dp_ctrl_on_link(dp->ctrl, dp_display->mst_active);
> if (rc) {
> DRM_ERROR("failed to complete DP link training\n");
> goto end;
> @@ -1628,6 +1629,9 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
>
> msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
> +
> + if (dp->mst_active)
> + msm_dp_ctrl_mst_send_act(msm_dp_display->ctrl);
> }
>
> static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> index 8ce8ba254b1bfe5b4c000df83eceef5823772780..38ca25491b1ccfd95e027a9c8f659abb3cd576d3 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.h
> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> @@ -22,6 +22,7 @@ struct msm_dp {
> bool audio_enabled;
> bool power_on;
> bool prepared;
> + bool mst_active;
> unsigned int connector_type;
> bool is_edp;
> bool internal_hpd;
> diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
> index 6c534fde6034fced2cb428e9a29de31ed5c5fcc4..46400973eb499066e9e805b16df759b1db34cf22 100644
> --- a/drivers/gpu/drm/msm/dp/dp_reg.h
> +++ b/drivers/gpu/drm/msm/dp/dp_reg.h
> @@ -330,6 +330,9 @@
> #define REG_DP_PHY_AUX_BIST_CFG (0x00000050)
> #define REG_DP_PHY_AUX_INTERRUPT_STATUS (0x000000BC)
>
> +/* DP MST related registers */
> +#define REG_DP_MST_ACT (0x00000500)
This looks completely out of place.
> +
> /* DP HDCP 1.3 registers */
> #define DP_HDCP_CTRL (0x0A0)
> #define DP_HDCP_STATUS (0x0A4)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 19/45] drm/msm/dp: add support to program mst support in mainlink
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (17 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 18/45] drm/msm/dp: add support to send ACT packets for MST Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:42 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 20/45] drm/msm/dp: no need to update tu calculation for mst Abhinav Kumar
` (27 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Add support to program the MST enabled bit in the mainlink
control when a mst session is active and disabled.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.c | 17 +++++++++++++++++
drivers/gpu/drm/msm/dp/dp_catalog.h | 1 +
drivers/gpu/drm/msm/dp/dp_ctrl.c | 4 ++++
3 files changed, 22 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index 88d6262a972ef2d30c467ef5ff5c58ef3299ae7d..bdc66e5cab640c351708ba1a1bc3bca21784df6e 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -417,6 +417,23 @@ void msm_dp_catalog_ctrl_psr_mainlink_enable(struct msm_dp_catalog *msm_dp_catal
msm_dp_write_link(catalog, REG_DP_MAINLINK_CTRL, val);
}
+void msm_dp_catalog_mst_config(struct msm_dp_catalog *msm_dp_catalog, bool enable)
+{
+ struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
+ struct msm_dp_catalog_private,
+ msm_dp_catalog);
+
+ u32 mainlink_ctrl;
+
+ mainlink_ctrl = msm_dp_read_link(catalog, REG_DP_MAINLINK_CTRL);
+ if (enable)
+ mainlink_ctrl |= (0x04000100);
+ else
+ mainlink_ctrl &= ~(0x04000100);
+
+ msm_dp_write_link(catalog, REG_DP_MAINLINK_CTRL, mainlink_ctrl);
+}
+
void msm_dp_catalog_ctrl_mainlink_ctrl(struct msm_dp_catalog *msm_dp_catalog,
bool enable)
{
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
index c91c52d40209b8bcb63db9c0256f6ef721dace8a..07284f484e2861aeae12b115cd05a94afed1c9cb 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.h
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
@@ -138,5 +138,6 @@ void msm_dp_catalog_audio_sfe_level(struct msm_dp_catalog *catalog, u32 safe_to_
/* DP MST APIs */
void msm_dp_catalog_trigger_act(struct msm_dp_catalog *dp_catalog);
bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *dp_catalog);
+void msm_dp_catalog_mst_config(struct msm_dp_catalog *dp_catalog, bool enable);
#endif /* _DP_CATALOG_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 2bfe2aac3c02b02b12713dbd98e79ed4a75b85d0..3839f1e8e1aeb2a14a7f59c546693141a0df6323 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -186,6 +186,9 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
+ if (ctrl->mst_active)
+ msm_dp_catalog_mst_config(ctrl->catalog, true);
+
msm_dp_ctrl_config_ctrl(ctrl, msm_dp_panel);
tb = msm_dp_link_get_test_bits_depth(ctrl->link,
@@ -2132,6 +2135,7 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
phy = ctrl->phy;
msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
+ msm_dp_catalog_mst_config(ctrl->catalog, false);
ctrl->mst_active = false;
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 19/45] drm/msm/dp: add support to program mst support in mainlink
2024-12-06 4:31 ` [PATCH 19/45] drm/msm/dp: add support to program mst support in mainlink Abhinav Kumar
@ 2024-12-08 11:42 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:42 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:50PM -0800, Abhinav Kumar wrote:
> Add support to program the MST enabled bit in the mainlink
> control when a mst session is active and disabled.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 17 +++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_catalog.h | 1 +
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 4 ++++
> 3 files changed, 22 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
> index 88d6262a972ef2d30c467ef5ff5c58ef3299ae7d..bdc66e5cab640c351708ba1a1bc3bca21784df6e 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> @@ -417,6 +417,23 @@ void msm_dp_catalog_ctrl_psr_mainlink_enable(struct msm_dp_catalog *msm_dp_catal
> msm_dp_write_link(catalog, REG_DP_MAINLINK_CTRL, val);
> }
>
> +void msm_dp_catalog_mst_config(struct msm_dp_catalog *msm_dp_catalog, bool enable)
Can this be merged into msm_dp_catalog_ctrl_mainlink_ctrl() ? Or is that
function called too early, when we do not know yet if we need MST or
not?
> +{
> + struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
> + struct msm_dp_catalog_private,
> + msm_dp_catalog);
> +
> + u32 mainlink_ctrl;
> +
> + mainlink_ctrl = msm_dp_read_link(catalog, REG_DP_MAINLINK_CTRL);
> + if (enable)
> + mainlink_ctrl |= (0x04000100);
> + else
> + mainlink_ctrl &= ~(0x04000100);
#define
> +
> + msm_dp_write_link(catalog, REG_DP_MAINLINK_CTRL, mainlink_ctrl);
> +}
> +
> void msm_dp_catalog_ctrl_mainlink_ctrl(struct msm_dp_catalog *msm_dp_catalog,
> bool enable)
> {
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
> index c91c52d40209b8bcb63db9c0256f6ef721dace8a..07284f484e2861aeae12b115cd05a94afed1c9cb 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
> @@ -138,5 +138,6 @@ void msm_dp_catalog_audio_sfe_level(struct msm_dp_catalog *catalog, u32 safe_to_
> /* DP MST APIs */
> void msm_dp_catalog_trigger_act(struct msm_dp_catalog *dp_catalog);
> bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *dp_catalog);
> +void msm_dp_catalog_mst_config(struct msm_dp_catalog *dp_catalog, bool enable);
>
> #endif /* _DP_CATALOG_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 2bfe2aac3c02b02b12713dbd98e79ed4a75b85d0..3839f1e8e1aeb2a14a7f59c546693141a0df6323 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -186,6 +186,9 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
> msm_dp_catalog_ctrl_lane_mapping(ctrl->catalog);
> msm_dp_catalog_setup_peripheral_flush(ctrl->catalog);
>
> + if (ctrl->mst_active)
> + msm_dp_catalog_mst_config(ctrl->catalog, true);
> +
> msm_dp_ctrl_config_ctrl(ctrl, msm_dp_panel);
>
> tb = msm_dp_link_get_test_bits_depth(ctrl->link,
> @@ -2132,6 +2135,7 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
> phy = ctrl->phy;
>
> msm_dp_catalog_ctrl_mainlink_ctrl(ctrl->catalog, false);
> + msm_dp_catalog_mst_config(ctrl->catalog, false);
>
> ctrl->mst_active = false;
>
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 20/45] drm/msm/dp: no need to update tu calculation for mst
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (18 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 19/45] drm/msm/dp: add support to program mst support in mainlink Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 5:45 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 21/45] drm/msm/dp: add support for mst channel slot allocation Abhinav Kumar
` (26 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
DP stream is transmitted in transfer units only for SST
case there is no need to calculate and program TU parameters
for MST case. Skip the TU programming for MST cases.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 3839f1e8e1aeb2a14a7f59c546693141a0df6323..1c313cf33f398dffc2ad349d7d1bc995fb4b45b3 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -2060,7 +2060,8 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
pixel_rate_orig,
msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420);
- msm_dp_ctrl_setup_tr_unit(ctrl);
+ if (!ctrl->mst_active)
+ msm_dp_ctrl_setup_tr_unit(ctrl);
msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 20/45] drm/msm/dp: no need to update tu calculation for mst
2024-12-06 4:31 ` [PATCH 20/45] drm/msm/dp: no need to update tu calculation for mst Abhinav Kumar
@ 2024-12-08 5:45 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 5:45 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:51PM -0800, Abhinav Kumar wrote:
> DP stream is transmitted in transfer units only for SST
> case there is no need to calculate and program TU parameters
> for MST case. Skip the TU programming for MST cases.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 21/45] drm/msm/dp: add support for mst channel slot allocation
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (19 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 20/45] drm/msm/dp: no need to update tu calculation for mst Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 6:13 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 22/45] drm/msm/dp: add support to send vcpf packets in dp controller Abhinav Kumar
` (25 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
For DP MST streams, the 64 MTP slots are time-shared between
the streams. Add the support to calculate the rate governor,
slots and reservation of the slots to the DP controller. Each
DP MST stream shall reserve its streams by calling the
dp_display_set_stream_info() from its bridge calls.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.c | 77 +++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_catalog.h | 7 ++
drivers/gpu/drm/msm/dp/dp_ctrl.c | 148 ++++++++++++++++++++++++++++++++++--
drivers/gpu/drm/msm/dp/dp_ctrl.h | 7 +-
drivers/gpu/drm/msm/dp/dp_display.c | 33 ++++++--
drivers/gpu/drm/msm/dp/dp_display.h | 5 +-
drivers/gpu/drm/msm/dp/dp_panel.h | 7 ++
drivers/gpu/drm/msm/dp/dp_reg.h | 6 ++
8 files changed, 276 insertions(+), 14 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index bdc66e5cab640c351708ba1a1bc3bca21784df6e..f9d21444d7891bcd043d282b31ae75711add4817 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -1116,6 +1116,83 @@ bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *msm_dp_catalog)
return msm_dp_read_link(catalog, REG_DP_MST_ACT);
}
+void msm_dp_catalog_mst_channel_alloc(struct msm_dp_catalog *msm_dp_catalog,
+ u32 ch, u32 ch_start_slot, u32 tot_slot_cnt)
+{
+ struct msm_dp_catalog_private *catalog;
+ u32 i, slot_reg_1, slot_reg_2, slot;
+ u32 reg_off = 0;
+ int const num_slots_per_reg = 32;
+
+ if (!msm_dp_catalog || ch >= DP_STREAM_MAX) {
+ DRM_ERROR("invalid input. ch %d\n", ch);
+ return;
+ }
+
+ if (ch_start_slot > DP_MAX_TIME_SLOTS ||
+ (ch_start_slot + tot_slot_cnt > DP_MAX_TIME_SLOTS)) {
+ DRM_ERROR("invalid slots start %d, tot %d\n",
+ ch_start_slot, tot_slot_cnt);
+ return;
+ }
+
+ catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
+
+ drm_dbg_dp(catalog->drm_dev, "ch %d, start_slot %d, tot_slot %d\n",
+ ch, ch_start_slot, tot_slot_cnt);
+
+ if (ch == DP_STREAM_1)
+ reg_off = REG_DP_DP1_TIMESLOT_1_32 - REG_DP_DP0_TIMESLOT_1_32;
+
+ slot_reg_1 = 0;
+ slot_reg_2 = 0;
+
+ if (ch_start_slot && tot_slot_cnt) {
+ ch_start_slot--;
+ for (i = 0; i < tot_slot_cnt; i++) {
+ if (ch_start_slot < num_slots_per_reg) {
+ slot_reg_1 |= BIT(ch_start_slot);
+ } else {
+ slot = ch_start_slot - num_slots_per_reg;
+ slot_reg_2 |= BIT(slot);
+ }
+ ch_start_slot++;
+ }
+ }
+
+ drm_dbg_dp(catalog->drm_dev, "ch:%d slot_reg_1:%d, slot_reg_2:%d\n", ch,
+ slot_reg_1, slot_reg_2);
+
+ msm_dp_write_link(catalog, REG_DP_DP0_TIMESLOT_1_32 + reg_off, slot_reg_1);
+ msm_dp_write_link(catalog, REG_DP_DP0_TIMESLOT_33_63 + reg_off, slot_reg_2);
+}
+
+void msm_dp_catalog_ctrl_update_rg(struct msm_dp_catalog *msm_dp_catalog, u32 stream,
+ u32 x_int, u32 y_frac_enum)
+{
+ struct msm_dp_catalog_private *catalog;
+
+ u32 rg, reg_off = 0;
+
+ if (!msm_dp_catalog || stream >= DP_STREAM_MAX) {
+ DRM_ERROR("invalid input. stream %d\n", stream);
+ return;
+ }
+
+ catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
+
+ rg = y_frac_enum;
+ rg |= (x_int << 16);
+
+ drm_dbg_dp(catalog->drm_dev, "stream: %d x_int:%d y_frac_enum:%d rg:%d\n",
+ stream, x_int, y_frac_enum, rg);
+
+ if (stream == DP_STREAM_1)
+ reg_off = REG_DP_DP1_RG - REG_DP_DP0_RG;
+
+ msm_dp_write_link(catalog, REG_DP_DP0_RG + reg_off, rg);
+}
+
void msm_dp_catalog_panel_tpg_enable(struct msm_dp_catalog *msm_dp_catalog,
struct drm_display_mode *drm_mode)
{
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
index 07284f484e2861aeae12b115cd05a94afed1c9cb..560016e2f929d4b92d6ea764d81a099c09c0e668 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.h
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
@@ -31,6 +31,8 @@
#define DP_HW_VERSION_1_0 0x10000000
#define DP_HW_VERSION_1_2 0x10020000
+#define DP_MAX_TIME_SLOTS 64
+
enum msm_dp_catalog_audio_sdp_type {
DP_AUDIO_SDP_STREAM,
DP_AUDIO_SDP_TIMESTAMP,
@@ -140,4 +142,9 @@ void msm_dp_catalog_trigger_act(struct msm_dp_catalog *dp_catalog);
bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *dp_catalog);
void msm_dp_catalog_mst_config(struct msm_dp_catalog *dp_catalog, bool enable);
+void msm_dp_catalog_mst_channel_alloc(struct msm_dp_catalog *ctrl,
+ u32 ch, u32 ch_start_slot, u32 tot_slot_cnt);
+void msm_dp_catalog_ctrl_update_rg(struct msm_dp_catalog *ctrl, u32 stream,
+ u32 x_int, u32 y_frac_enum);
+
#endif /* _DP_CATALOG_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 1c313cf33f398dffc2ad349d7d1bc995fb4b45b3..14562def1e70b769434243d1ce72661a7b4d4c6b 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -69,6 +69,11 @@ struct msm_dp_vc_tu_mapping_table {
u8 tu_size_minus1;
};
+struct msm_dp_mst_ch_slot_info {
+ u32 start_slot;
+ u32 tot_slots;
+};
+
struct msm_dp_ctrl_private {
struct msm_dp_ctrl msm_dp_ctrl;
struct drm_device *drm_dev;
@@ -100,6 +105,8 @@ struct msm_dp_ctrl_private {
bool mst_active;
bool stream_clks_on[DP_STREAM_MAX];
+
+ struct msm_dp_mst_ch_slot_info mst_ch_info[DP_STREAM_MAX];
};
static int msm_dp_aux_link_configure(struct drm_dp_aux *aux,
@@ -2021,7 +2028,103 @@ int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_li
return ret;
}
-int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
+static void msm_dp_ctrl_mst_calculate_rg(struct msm_dp_ctrl_private *ctrl,
+ struct msm_dp_panel *panel,
+ u32 *p_x_int, u32 *p_y_frac_enum)
+{
+ u64 min_slot_cnt, max_slot_cnt;
+ u64 raw_target_sc, target_sc_fixp;
+ u64 ts_denom, ts_enum, ts_int;
+ u64 pclk = panel->msm_dp_mode.drm_mode.clock;
+ u64 lclk = 0;
+ u64 lanes = ctrl->link->link_params.num_lanes;
+ u64 bpp = panel->msm_dp_mode.bpp;
+ u64 pbn = panel->mst_caps.pbn;
+ u64 numerator, denominator, temp, temp1, temp2;
+ u32 x_int = 0, y_frac_enum = 0;
+ u64 target_strm_sym, ts_int_fixp, ts_frac_fixp, y_frac_enum_fixp;
+
+ lclk = ctrl->link->link_params.rate;
+
+ /* min_slot_cnt */
+ numerator = pclk * bpp * 64 * 1000;
+ denominator = lclk * lanes * 8 * 1000;
+ min_slot_cnt = drm_fixp_from_fraction(numerator, denominator);
+
+ /* max_slot_cnt */
+ numerator = pbn * 54 * 1000;
+ denominator = lclk * lanes;
+ max_slot_cnt = drm_fixp_from_fraction(numerator, denominator);
+
+ /* raw_target_sc */
+ numerator = max_slot_cnt + min_slot_cnt;
+ denominator = drm_fixp_from_fraction(2, 1);
+ raw_target_sc = drm_fixp_div(numerator, denominator);
+
+ /* target_sc */
+ temp = drm_fixp_from_fraction(256 * lanes, 1);
+ numerator = drm_fixp_mul(raw_target_sc, temp);
+ denominator = drm_fixp_from_fraction(256 * lanes, 1);
+ target_sc_fixp = drm_fixp_div(numerator, denominator);
+
+ ts_enum = 256 * lanes;
+ ts_denom = drm_fixp_from_fraction(256 * lanes, 1);
+ ts_int = drm_fixp2int(target_sc_fixp);
+
+ temp = drm_fixp2int_ceil(raw_target_sc);
+ if (temp != ts_int) {
+ temp = drm_fixp_from_fraction(ts_int, 1);
+ temp1 = raw_target_sc - temp;
+ temp2 = drm_fixp_mul(temp1, ts_denom);
+ ts_enum = drm_fixp2int(temp2);
+ }
+
+ /* target_strm_sym */
+ ts_int_fixp = drm_fixp_from_fraction(ts_int, 1);
+ ts_frac_fixp = drm_fixp_from_fraction(ts_enum, drm_fixp2int(ts_denom));
+ temp = ts_int_fixp + ts_frac_fixp;
+ temp1 = drm_fixp_from_fraction(lanes, 1);
+ target_strm_sym = drm_fixp_mul(temp, temp1);
+
+ /* x_int */
+ x_int = drm_fixp2int(target_strm_sym);
+
+ /* y_enum_frac */
+ temp = drm_fixp_from_fraction(x_int, 1);
+ temp1 = target_strm_sym - temp;
+ temp2 = drm_fixp_from_fraction(256, 1);
+ y_frac_enum_fixp = drm_fixp_mul(temp1, temp2);
+
+ temp1 = drm_fixp2int(y_frac_enum_fixp);
+ temp2 = drm_fixp2int_ceil(y_frac_enum_fixp);
+
+ y_frac_enum = (u32)((temp1 == temp2) ? temp1 : temp1 + 1);
+
+ *p_x_int = x_int;
+ *p_y_frac_enum = y_frac_enum;
+
+ drm_dbg_dp(ctrl->drm_dev, "mst lane_cnt:%llu, rate:%llu x_int:%d, y_frac:%d\n",
+ lanes, lclk, x_int, y_frac_enum);
+}
+
+static void msm_dp_ctrl_mst_stream_setup(struct msm_dp_ctrl_private *ctrl,
+ struct msm_dp_panel *panel,
+ u32 max_streams)
+{
+ u32 x_int, y_frac_enum;
+
+ drm_dbg_dp(ctrl->drm_dev, "mst stream channel allocation\n");
+
+ msm_dp_ctrl_mst_stream_channel_slot_setup(&ctrl->msm_dp_ctrl, max_streams);
+
+ msm_dp_ctrl_mst_calculate_rg(ctrl, panel, &x_int, &y_frac_enum);
+
+ msm_dp_catalog_ctrl_update_rg(ctrl->catalog, panel->stream_id,
+ x_int, y_frac_enum);
+}
+
+int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl,
+ struct msm_dp_panel *msm_dp_panel, u32 max_streams)
{
int ret = 0;
bool mainlink_ready = false;
@@ -2063,6 +2166,9 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
if (!ctrl->mst_active)
msm_dp_ctrl_setup_tr_unit(ctrl);
+ if (ctrl->mst_active)
+ msm_dp_ctrl_mst_stream_setup(ctrl, msm_dp_panel, max_streams);
+
msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
if (ctrl->mst_active)
@@ -2143,13 +2249,45 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
dev_pm_opp_set_rate(ctrl->dev, 0);
msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
- DRM_DEBUG_DP("Before, phy=%p init_count=%d power_on=%d\n",
- phy, phy->init_count, phy->power_count);
+ drm_dbg_dp(ctrl->drm_dev, "Before, phy=%p init_count=%d power_on=%d\n",
+ phy, phy->init_count, phy->power_count);
phy_power_off(phy);
- DRM_DEBUG_DP("After, phy=%p init_count=%d power_on=%d\n",
- phy, phy->init_count, phy->power_count);
+ drm_dbg_dp(ctrl->drm_dev, "After, phy=%p init_count=%d power_on=%d\n",
+ phy, phy->init_count, phy->power_count);
+}
+
+void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
+ enum msm_dp_stream_id strm,
+ u32 start_slot, u32 tot_slots)
+{
+ struct msm_dp_ctrl_private *ctrl;
+
+ if (!msm_dp_ctrl || strm >= DP_STREAM_MAX) {
+ DRM_ERROR("invalid input\n");
+ return;
+ }
+
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+
+ ctrl->mst_ch_info[strm].start_slot = start_slot;
+ ctrl->mst_ch_info[strm].tot_slots = tot_slots;
+}
+
+// TO-DO : Check if we can do a dealloc instead of this one during teardown
+void msm_dp_ctrl_mst_stream_channel_slot_setup(struct msm_dp_ctrl *msm_dp_ctrl, u32 max_streams)
+{
+ struct msm_dp_ctrl_private *ctrl;
+ int i;
+
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+
+ for (i = DP_STREAM_0; i < max_streams; i++) {
+ msm_dp_catalog_mst_channel_alloc(ctrl->catalog,
+ i, ctrl->mst_ch_info[i].start_slot,
+ ctrl->mst_ch_info[i].tot_slots);
+ }
}
irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index 81c05b1b2baac63e1f1888f3f517e62a98e230a7..b126651da24b3abdaf540268758b37dca9fe1291 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -18,7 +18,8 @@ struct msm_dp_ctrl {
struct phy;
int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl, bool mst_active);
-int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
+int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl,
+ struct msm_dp_panel *msm_dp_panel, u32 max_streams);
int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
@@ -47,5 +48,9 @@ void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl,
void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl);
void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl);
int msm_dp_ctrl_mst_send_act(struct msm_dp_ctrl *ctrl);
+void msm_dp_ctrl_mst_stream_channel_slot_setup(struct msm_dp_ctrl *msm_dp_ctrl, u32 max_streams);
+void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
+ enum msm_dp_stream_id strm,
+ u32 start_slot, u32 tot_slots);
#endif /* _DP_CTRL_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index e19860ef3493fb100afbf04b09d14a136fd6b887..2a4a79317153817cb24537ea95fad07c9bc20715 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -33,6 +33,8 @@ MODULE_PARM_DESC(psr_enabled, "enable PSR for eDP and DP displays");
#define HPD_STRING_SIZE 30
+#define DEFAULT_STREAM_COUNT 1
+
enum {
ISR_DISCONNECTED,
ISR_CONNECT_PENDING,
@@ -95,6 +97,8 @@ struct msm_dp_display_private {
/* wait for audio signaling */
struct completion audio_comp;
+ unsigned int max_stream;
+
/* event related only access by event thread */
struct mutex event_mutex;
wait_queue_head_t event_q;
@@ -875,7 +879,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp)
return 0;
}
- rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel);
+ rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel, dp->max_stream);
if (!rc)
msm_dp_display->power_on = true;
@@ -963,11 +967,14 @@ int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display,
return 0;
}
-int msm_dp_display_set_stream_id(struct msm_dp *dp,
- struct msm_dp_panel *panel, u32 strm_id)
+int msm_dp_display_set_stream_info(struct msm_dp *dp,
+ struct msm_dp_panel *panel, u32 strm_id, u32 start_slot,
+ u32 num_slots, u32 pbn, int vcpi)
{
int rc = 0;
struct msm_dp_display_private *msm_dp_display;
+ const int max_slots = 64;
+
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
@@ -981,8 +988,18 @@ int msm_dp_display_set_stream_id(struct msm_dp *dp,
return -EINVAL;
}
- if (panel)
+ if (start_slot + num_slots > max_slots) {
+ DRM_ERROR("invalid channel info received. start:%d, slots:%d\n",
+ start_slot, num_slots);
+ return -EINVAL;
+ }
+
+ msm_dp_ctrl_set_mst_channel_info(msm_dp_display->ctrl, strm_id, start_slot, num_slots);
+
+ if (panel) {
panel->stream_id = strm_id;
+ panel->mst_caps.pbn = pbn;
+ }
return rc;
}
@@ -1370,6 +1387,7 @@ static int msm_dp_display_probe(struct platform_device *pdev)
dp->msm_dp_display.is_edp =
(dp->msm_dp_display.connector_type == DRM_MODE_CONNECTOR_eDP);
+ dp->max_stream = DEFAULT_STREAM_COUNT;
rc = msm_dp_init_sub_modules(dp);
if (rc) {
DRM_ERROR("init sub module failed\n");
@@ -1602,7 +1620,7 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
mutex_lock(&msm_dp_display->event_mutex);
- msm_dp_display_set_stream_id(dp, msm_dp_display->panel, 0);
+ msm_dp_display_set_stream_info(dp, msm_dp_display->panel, 0, 0, 0, 0, 0);
if (dp->prepared) {
rc = msm_dp_display_enable(msm_dp_display);
@@ -1630,8 +1648,11 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
- if (dp->mst_active)
+ if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT) {
+ msm_dp_ctrl_mst_stream_channel_slot_setup(msm_dp_display->ctrl,
+ msm_dp_display->max_stream);
msm_dp_ctrl_mst_send_act(msm_dp_display->ctrl);
+ }
}
static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 38ca25491b1ccfd95e027a9c8f659abb3cd576d3..258c240de580b634c05cf5895a8e52160449eba1 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -53,7 +53,8 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
const struct drm_display_info *info,
const struct drm_display_mode *mode);
-int msm_dp_display_set_stream_id(struct msm_dp *dp,
- struct msm_dp_panel *panel, u32 strm_id);
+int msm_dp_display_set_stream_info(struct msm_dp *dp,
+ struct msm_dp_panel *panel, u32 strm_id,
+ u32 start_slot, u32 num_slots, u32 pbn, int vcpi);
#endif /* _DP_DISPLAY_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
index 9af2272da697e7aa49377c02abdb97e72f07c0bd..b4f6efaff7ed227d6e3fc846986aba375cdbbadb 100644
--- a/drivers/gpu/drm/msm/dp/dp_panel.h
+++ b/drivers/gpu/drm/msm/dp/dp_panel.h
@@ -33,6 +33,11 @@ struct msm_dp_panel_psr {
u8 capabilities;
};
+struct mst_caps {
+ u32 pbn_no_overhead;
+ u32 pbn;
+};
+
struct msm_dp_panel {
/* dpcd raw data */
u8 dpcd[DP_RECEIVER_CAP_SIZE];
@@ -52,6 +57,8 @@ struct msm_dp_panel {
u32 max_dp_link_rate;
u32 max_bw_code;
+
+ struct mst_caps mst_caps;
};
int msm_dp_panel_init_panel_info(struct msm_dp_panel *msm_dp_panel);
diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
index 46400973eb499066e9e805b16df759b1db34cf22..8bc2a431462fc1fb45b1fe8e43a0a0ec7f75e5b1 100644
--- a/drivers/gpu/drm/msm/dp/dp_reg.h
+++ b/drivers/gpu/drm/msm/dp/dp_reg.h
@@ -332,6 +332,12 @@
/* DP MST related registers */
#define REG_DP_MST_ACT (0x00000500)
+#define REG_DP_DP0_RG (0x000004F8)
+#define REG_DP_DP1_RG (0x000004FC)
+#define REG_DP_DP0_TIMESLOT_1_32 (0x00000404)
+#define REG_DP_DP0_TIMESLOT_33_63 (0x00000408)
+#define REG_DP_DP1_TIMESLOT_1_32 (0x0000040C)
+#define REG_DP_DP1_TIMESLOT_33_63 (0x00000410)
/* DP HDCP 1.3 registers */
#define DP_HDCP_CTRL (0x0A0)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 21/45] drm/msm/dp: add support for mst channel slot allocation
2024-12-06 4:31 ` [PATCH 21/45] drm/msm/dp: add support for mst channel slot allocation Abhinav Kumar
@ 2024-12-08 6:13 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 6:13 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:52PM -0800, Abhinav Kumar wrote:
> For DP MST streams, the 64 MTP slots are time-shared between
> the streams. Add the support to calculate the rate governor,
> slots and reservation of the slots to the DP controller. Each
> DP MST stream shall reserve its streams by calling the
> dp_display_set_stream_info() from its bridge calls.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 77 +++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_catalog.h | 7 ++
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 148 ++++++++++++++++++++++++++++++++++--
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 7 +-
> drivers/gpu/drm/msm/dp/dp_display.c | 33 ++++++--
> drivers/gpu/drm/msm/dp/dp_display.h | 5 +-
> drivers/gpu/drm/msm/dp/dp_panel.h | 7 ++
> drivers/gpu/drm/msm/dp/dp_reg.h | 6 ++
> 8 files changed, 276 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
> index bdc66e5cab640c351708ba1a1bc3bca21784df6e..f9d21444d7891bcd043d282b31ae75711add4817 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> @@ -1116,6 +1116,83 @@ bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *msm_dp_catalog)
> return msm_dp_read_link(catalog, REG_DP_MST_ACT);
> }
>
> +void msm_dp_catalog_mst_channel_alloc(struct msm_dp_catalog *msm_dp_catalog,
> + u32 ch, u32 ch_start_slot, u32 tot_slot_cnt)
> +{
> + struct msm_dp_catalog_private *catalog;
> + u32 i, slot_reg_1, slot_reg_2, slot;
> + u32 reg_off = 0;
> + int const num_slots_per_reg = 32;
> +
> + if (!msm_dp_catalog || ch >= DP_STREAM_MAX) {
> + DRM_ERROR("invalid input. ch %d\n", ch);
> + return;
> + }
> +
> + if (ch_start_slot > DP_MAX_TIME_SLOTS ||
> + (ch_start_slot + tot_slot_cnt > DP_MAX_TIME_SLOTS)) {
> + DRM_ERROR("invalid slots start %d, tot %d\n",
> + ch_start_slot, tot_slot_cnt);
> + return;
> + }
This should be handled by the generic MST code.
> +
> + catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
> +
> + drm_dbg_dp(catalog->drm_dev, "ch %d, start_slot %d, tot_slot %d\n",
> + ch, ch_start_slot, tot_slot_cnt);
> +
> + if (ch == DP_STREAM_1)
> + reg_off = REG_DP_DP1_TIMESLOT_1_32 - REG_DP_DP0_TIMESLOT_1_32;
> +
> + slot_reg_1 = 0;
> + slot_reg_2 = 0;
> +
> + if (ch_start_slot && tot_slot_cnt) {
> + ch_start_slot--;
> + for (i = 0; i < tot_slot_cnt; i++) {
> + if (ch_start_slot < num_slots_per_reg) {
> + slot_reg_1 |= BIT(ch_start_slot);
> + } else {
> + slot = ch_start_slot - num_slots_per_reg;
> + slot_reg_2 |= BIT(slot);
> + }
> + ch_start_slot++;
> + }
> + }
> +
> + drm_dbg_dp(catalog->drm_dev, "ch:%d slot_reg_1:%d, slot_reg_2:%d\n", ch,
> + slot_reg_1, slot_reg_2);
> +
> + msm_dp_write_link(catalog, REG_DP_DP0_TIMESLOT_1_32 + reg_off, slot_reg_1);
> + msm_dp_write_link(catalog, REG_DP_DP0_TIMESLOT_33_63 + reg_off, slot_reg_2);
The more I look at the msm_dp_write_link(catalog, REG_DP_foo +
reg_offset), more I think that the I'd prefer to have explicit
(stream_id == STREAM_1 ? REG_DP_DP1_TIMESLOT_1_32 :
REG_DP_DP0_TIMESLOT_1_32). There is a difference between selecting the
offset between two register spaces and overusing offset-like model to
select between two registers.
> +}
> +
> +void msm_dp_catalog_ctrl_update_rg(struct msm_dp_catalog *msm_dp_catalog, u32 stream,
> + u32 x_int, u32 y_frac_enum)
> +{
> + struct msm_dp_catalog_private *catalog;
> +
> + u32 rg, reg_off = 0;
> +
> + if (!msm_dp_catalog || stream >= DP_STREAM_MAX) {
> + DRM_ERROR("invalid input. stream %d\n", stream);
> + return;
> + }
> +
> + catalog = container_of(msm_dp_catalog, struct msm_dp_catalog_private, msm_dp_catalog);
> +
> + rg = y_frac_enum;
> + rg |= (x_int << 16);
> +
> + drm_dbg_dp(catalog->drm_dev, "stream: %d x_int:%d y_frac_enum:%d rg:%d\n",
> + stream, x_int, y_frac_enum, rg);
> +
> + if (stream == DP_STREAM_1)
> + reg_off = REG_DP_DP1_RG - REG_DP_DP0_RG;
> +
> + msm_dp_write_link(catalog, REG_DP_DP0_RG + reg_off, rg);
> +}
> +
> void msm_dp_catalog_panel_tpg_enable(struct msm_dp_catalog *msm_dp_catalog,
> struct drm_display_mode *drm_mode)
> {
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
> index 07284f484e2861aeae12b115cd05a94afed1c9cb..560016e2f929d4b92d6ea764d81a099c09c0e668 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
> @@ -31,6 +31,8 @@
> #define DP_HW_VERSION_1_0 0x10000000
> #define DP_HW_VERSION_1_2 0x10020000
>
> +#define DP_MAX_TIME_SLOTS 64
Doesn't it depend on the encoding format?
> +
> enum msm_dp_catalog_audio_sdp_type {
> DP_AUDIO_SDP_STREAM,
> DP_AUDIO_SDP_TIMESTAMP,
> @@ -140,4 +142,9 @@ void msm_dp_catalog_trigger_act(struct msm_dp_catalog *dp_catalog);
> bool msm_dp_catalog_read_act_complete_sts(struct msm_dp_catalog *dp_catalog);
> void msm_dp_catalog_mst_config(struct msm_dp_catalog *dp_catalog, bool enable);
>
> +void msm_dp_catalog_mst_channel_alloc(struct msm_dp_catalog *ctrl,
> + u32 ch, u32 ch_start_slot, u32 tot_slot_cnt);
> +void msm_dp_catalog_ctrl_update_rg(struct msm_dp_catalog *ctrl, u32 stream,
> + u32 x_int, u32 y_frac_enum);
> +
> #endif /* _DP_CATALOG_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 1c313cf33f398dffc2ad349d7d1bc995fb4b45b3..14562def1e70b769434243d1ce72661a7b4d4c6b 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -69,6 +69,11 @@ struct msm_dp_vc_tu_mapping_table {
> u8 tu_size_minus1;
> };
>
> +struct msm_dp_mst_ch_slot_info {
> + u32 start_slot;
> + u32 tot_slots;
> +};
> +
> struct msm_dp_ctrl_private {
> struct msm_dp_ctrl msm_dp_ctrl;
> struct drm_device *drm_dev;
> @@ -100,6 +105,8 @@ struct msm_dp_ctrl_private {
>
> bool mst_active;
> bool stream_clks_on[DP_STREAM_MAX];
> +
> + struct msm_dp_mst_ch_slot_info mst_ch_info[DP_STREAM_MAX];
> };
>
> static int msm_dp_aux_link_configure(struct drm_dp_aux *aux,
> @@ -2021,7 +2028,103 @@ int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *msm_dp_ctrl, bool force_li
> return ret;
> }
>
> -int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
> +static void msm_dp_ctrl_mst_calculate_rg(struct msm_dp_ctrl_private *ctrl,
> + struct msm_dp_panel *panel,
> + u32 *p_x_int, u32 *p_y_frac_enum)
> +{
> + u64 min_slot_cnt, max_slot_cnt;
> + u64 raw_target_sc, target_sc_fixp;
> + u64 ts_denom, ts_enum, ts_int;
> + u64 pclk = panel->msm_dp_mode.drm_mode.clock;
> + u64 lclk = 0;
> + u64 lanes = ctrl->link->link_params.num_lanes;
> + u64 bpp = panel->msm_dp_mode.bpp;
> + u64 pbn = panel->mst_caps.pbn;
> + u64 numerator, denominator, temp, temp1, temp2;
> + u32 x_int = 0, y_frac_enum = 0;
> + u64 target_strm_sym, ts_int_fixp, ts_frac_fixp, y_frac_enum_fixp;
> +
> + lclk = ctrl->link->link_params.rate;
> +
> + /* min_slot_cnt */
> + numerator = pclk * bpp * 64 * 1000;
> + denominator = lclk * lanes * 8 * 1000;
> + min_slot_cnt = drm_fixp_from_fraction(numerator, denominator);
> +
> + /* max_slot_cnt */
> + numerator = pbn * 54 * 1000;
> + denominator = lclk * lanes;
> + max_slot_cnt = drm_fixp_from_fraction(numerator, denominator);
> +
> + /* raw_target_sc */
> + numerator = max_slot_cnt + min_slot_cnt;
> + denominator = drm_fixp_from_fraction(2, 1);
> + raw_target_sc = drm_fixp_div(numerator, denominator);
> +
> + /* target_sc */
> + temp = drm_fixp_from_fraction(256 * lanes, 1);
> + numerator = drm_fixp_mul(raw_target_sc, temp);
> + denominator = drm_fixp_from_fraction(256 * lanes, 1);
> + target_sc_fixp = drm_fixp_div(numerator, denominator);
> +
> + ts_enum = 256 * lanes;
> + ts_denom = drm_fixp_from_fraction(256 * lanes, 1);
> + ts_int = drm_fixp2int(target_sc_fixp);
> +
> + temp = drm_fixp2int_ceil(raw_target_sc);
> + if (temp != ts_int) {
> + temp = drm_fixp_from_fraction(ts_int, 1);
> + temp1 = raw_target_sc - temp;
> + temp2 = drm_fixp_mul(temp1, ts_denom);
> + ts_enum = drm_fixp2int(temp2);
> + }
> +
> + /* target_strm_sym */
> + ts_int_fixp = drm_fixp_from_fraction(ts_int, 1);
> + ts_frac_fixp = drm_fixp_from_fraction(ts_enum, drm_fixp2int(ts_denom));
> + temp = ts_int_fixp + ts_frac_fixp;
> + temp1 = drm_fixp_from_fraction(lanes, 1);
> + target_strm_sym = drm_fixp_mul(temp, temp1);
> +
> + /* x_int */
> + x_int = drm_fixp2int(target_strm_sym);
> +
> + /* y_enum_frac */
> + temp = drm_fixp_from_fraction(x_int, 1);
> + temp1 = target_strm_sym - temp;
> + temp2 = drm_fixp_from_fraction(256, 1);
> + y_frac_enum_fixp = drm_fixp_mul(temp1, temp2);
> +
> + temp1 = drm_fixp2int(y_frac_enum_fixp);
> + temp2 = drm_fixp2int_ceil(y_frac_enum_fixp);
> +
> + y_frac_enum = (u32)((temp1 == temp2) ? temp1 : temp1 + 1);
> +
> + *p_x_int = x_int;
> + *p_y_frac_enum = y_frac_enum;
> +
> + drm_dbg_dp(ctrl->drm_dev, "mst lane_cnt:%llu, rate:%llu x_int:%d, y_frac:%d\n",
> + lanes, lclk, x_int, y_frac_enum);
> +}
> +
> +static void msm_dp_ctrl_mst_stream_setup(struct msm_dp_ctrl_private *ctrl,
> + struct msm_dp_panel *panel,
> + u32 max_streams)
> +{
> + u32 x_int, y_frac_enum;
> +
> + drm_dbg_dp(ctrl->drm_dev, "mst stream channel allocation\n");
> +
> + msm_dp_ctrl_mst_stream_channel_slot_setup(&ctrl->msm_dp_ctrl, max_streams);
> +
> + msm_dp_ctrl_mst_calculate_rg(ctrl, panel, &x_int, &y_frac_enum);
Use S16.16 (or U16.16) instead and pass required value as is
> +
> + msm_dp_catalog_ctrl_update_rg(ctrl->catalog, panel->stream_id,
> + x_int, y_frac_enum);
> +}
> +
> +int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl,
> + struct msm_dp_panel *msm_dp_panel, u32 max_streams)
> {
> int ret = 0;
> bool mainlink_ready = false;
> @@ -2063,6 +2166,9 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *
> if (!ctrl->mst_active)
> msm_dp_ctrl_setup_tr_unit(ctrl);
>
> + if (ctrl->mst_active)
> + msm_dp_ctrl_mst_stream_setup(ctrl, msm_dp_panel, max_streams);
> +
> msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, DP_STATE_CTRL_SEND_VIDEO);
>
> if (ctrl->mst_active)
> @@ -2143,13 +2249,45 @@ void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl)
> dev_pm_opp_set_rate(ctrl->dev, 0);
> msm_dp_ctrl_link_clk_disable(&ctrl->msm_dp_ctrl);
>
> - DRM_DEBUG_DP("Before, phy=%p init_count=%d power_on=%d\n",
> - phy, phy->init_count, phy->power_count);
> + drm_dbg_dp(ctrl->drm_dev, "Before, phy=%p init_count=%d power_on=%d\n",
> + phy, phy->init_count, phy->power_count);
>
> phy_power_off(phy);
>
> - DRM_DEBUG_DP("After, phy=%p init_count=%d power_on=%d\n",
> - phy, phy->init_count, phy->power_count);
> + drm_dbg_dp(ctrl->drm_dev, "After, phy=%p init_count=%d power_on=%d\n",
> + phy, phy->init_count, phy->power_count);
> +}
> +
> +void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
> + enum msm_dp_stream_id strm,
> + u32 start_slot, u32 tot_slots)
> +{
> + struct msm_dp_ctrl_private *ctrl;
> +
> + if (!msm_dp_ctrl || strm >= DP_STREAM_MAX) {
> + DRM_ERROR("invalid input\n");
> + return;
> + }
> +
> + ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> +
> + ctrl->mst_ch_info[strm].start_slot = start_slot;
> + ctrl->mst_ch_info[strm].tot_slots = tot_slots;
> +}
> +
> +// TO-DO : Check if we can do a dealloc instead of this one during teardown
> +void msm_dp_ctrl_mst_stream_channel_slot_setup(struct msm_dp_ctrl *msm_dp_ctrl, u32 max_streams)
> +{
> + struct msm_dp_ctrl_private *ctrl;
> + int i;
> +
> + ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> +
> + for (i = DP_STREAM_0; i < max_streams; i++) {
> + msm_dp_catalog_mst_channel_alloc(ctrl->catalog,
> + i, ctrl->mst_ch_info[i].start_slot,
> + ctrl->mst_ch_info[i].tot_slots);
> + }
> }
>
> irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index 81c05b1b2baac63e1f1888f3f517e62a98e230a7..b126651da24b3abdaf540268758b37dca9fe1291 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -18,7 +18,8 @@ struct msm_dp_ctrl {
> struct phy;
>
> int msm_dp_ctrl_on_link(struct msm_dp_ctrl *msm_dp_ctrl, bool mst_active);
> -int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel);
> +int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl,
> + struct msm_dp_panel *msm_dp_panel, u32 max_streams);
> int msm_dp_ctrl_prepare_stream_on(struct msm_dp_ctrl *dp_ctrl, bool force_link_train);
> void msm_dp_ctrl_off_link_stream(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_off_link(struct msm_dp_ctrl *msm_dp_ctrl);
> @@ -47,5 +48,9 @@ void msm_dp_ctrl_clear_vsc_sdp_pkt(struct msm_dp_ctrl *msm_dp_ctrl,
> void msm_dp_ctrl_psm_config(struct msm_dp_ctrl *msm_dp_ctrl);
> void msm_dp_ctrl_reinit_phy(struct msm_dp_ctrl *msm_dp_ctrl);
> int msm_dp_ctrl_mst_send_act(struct msm_dp_ctrl *ctrl);
> +void msm_dp_ctrl_mst_stream_channel_slot_setup(struct msm_dp_ctrl *msm_dp_ctrl, u32 max_streams);
> +void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
> + enum msm_dp_stream_id strm,
> + u32 start_slot, u32 tot_slots);
>
> #endif /* _DP_CTRL_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index e19860ef3493fb100afbf04b09d14a136fd6b887..2a4a79317153817cb24537ea95fad07c9bc20715 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -33,6 +33,8 @@ MODULE_PARM_DESC(psr_enabled, "enable PSR for eDP and DP displays");
>
> #define HPD_STRING_SIZE 30
>
> +#define DEFAULT_STREAM_COUNT 1
> +
> enum {
> ISR_DISCONNECTED,
> ISR_CONNECT_PENDING,
> @@ -95,6 +97,8 @@ struct msm_dp_display_private {
> /* wait for audio signaling */
> struct completion audio_comp;
>
> + unsigned int max_stream;
> +
> /* event related only access by event thread */
> struct mutex event_mutex;
> wait_queue_head_t event_q;
> @@ -875,7 +879,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp)
> return 0;
> }
>
> - rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel);
> + rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel, dp->max_stream);
> if (!rc)
> msm_dp_display->power_on = true;
>
> @@ -963,11 +967,14 @@ int msm_dp_display_set_plugged_cb(struct msm_dp *msm_dp_display,
> return 0;
> }
>
> -int msm_dp_display_set_stream_id(struct msm_dp *dp,
> - struct msm_dp_panel *panel, u32 strm_id)
> +int msm_dp_display_set_stream_info(struct msm_dp *dp,
> + struct msm_dp_panel *panel, u32 strm_id, u32 start_slot,
> + u32 num_slots, u32 pbn, int vcpi)
Where is this used with anything else than 0 start / num / pbn/ vcpi?
> {
> int rc = 0;
> struct msm_dp_display_private *msm_dp_display;
> + const int max_slots = 64;
> +
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
>
> @@ -981,8 +988,18 @@ int msm_dp_display_set_stream_id(struct msm_dp *dp,
> return -EINVAL;
> }
>
> - if (panel)
> + if (start_slot + num_slots > max_slots) {
> + DRM_ERROR("invalid channel info received. start:%d, slots:%d\n",
> + start_slot, num_slots);
> + return -EINVAL;
> + }
> +
> + msm_dp_ctrl_set_mst_channel_info(msm_dp_display->ctrl, strm_id, start_slot, num_slots);
> +
> + if (panel) {
> panel->stream_id = strm_id;
> + panel->mst_caps.pbn = pbn;
> + }
>
> return rc;
> }
> @@ -1370,6 +1387,7 @@ static int msm_dp_display_probe(struct platform_device *pdev)
> dp->msm_dp_display.is_edp =
> (dp->msm_dp_display.connector_type == DRM_MODE_CONNECTOR_eDP);
>
> + dp->max_stream = DEFAULT_STREAM_COUNT;
> rc = msm_dp_init_sub_modules(dp);
> if (rc) {
> DRM_ERROR("init sub module failed\n");
> @@ -1602,7 +1620,7 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
>
> mutex_lock(&msm_dp_display->event_mutex);
>
> - msm_dp_display_set_stream_id(dp, msm_dp_display->panel, 0);
> + msm_dp_display_set_stream_info(dp, msm_dp_display->panel, 0, 0, 0, 0, 0);
>
> if (dp->prepared) {
> rc = msm_dp_display_enable(msm_dp_display);
> @@ -1630,8 +1648,11 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
>
> msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
>
> - if (dp->mst_active)
> + if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT) {
Why? I liked the previous condition more. There is no connection between
_default_ and supporting only SST.
> + msm_dp_ctrl_mst_stream_channel_slot_setup(msm_dp_display->ctrl,
> + msm_dp_display->max_stream);
> msm_dp_ctrl_mst_send_act(msm_dp_display->ctrl);
> + }
> }
>
> static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> index 38ca25491b1ccfd95e027a9c8f659abb3cd576d3..258c240de580b634c05cf5895a8e52160449eba1 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.h
> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> @@ -53,7 +53,8 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
> enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
> const struct drm_display_info *info,
> const struct drm_display_mode *mode);
> -int msm_dp_display_set_stream_id(struct msm_dp *dp,
> - struct msm_dp_panel *panel, u32 strm_id);
> +int msm_dp_display_set_stream_info(struct msm_dp *dp,
> + struct msm_dp_panel *panel, u32 strm_id,
> + u32 start_slot, u32 num_slots, u32 pbn, int vcpi);
>
> #endif /* _DP_DISPLAY_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_panel.h b/drivers/gpu/drm/msm/dp/dp_panel.h
> index 9af2272da697e7aa49377c02abdb97e72f07c0bd..b4f6efaff7ed227d6e3fc846986aba375cdbbadb 100644
> --- a/drivers/gpu/drm/msm/dp/dp_panel.h
> +++ b/drivers/gpu/drm/msm/dp/dp_panel.h
> @@ -33,6 +33,11 @@ struct msm_dp_panel_psr {
> u8 capabilities;
> };
>
> +struct mst_caps {
> + u32 pbn_no_overhead;
Doesn't seem to be used
> + u32 pbn;
Why is it a part of capabilities structure?
> +};
> +
> struct msm_dp_panel {
> /* dpcd raw data */
> u8 dpcd[DP_RECEIVER_CAP_SIZE];
> @@ -52,6 +57,8 @@ struct msm_dp_panel {
> u32 max_dp_link_rate;
>
> u32 max_bw_code;
> +
> + struct mst_caps mst_caps;
> };
>
> int msm_dp_panel_init_panel_info(struct msm_dp_panel *msm_dp_panel);
> diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
> index 46400973eb499066e9e805b16df759b1db34cf22..8bc2a431462fc1fb45b1fe8e43a0a0ec7f75e5b1 100644
> --- a/drivers/gpu/drm/msm/dp/dp_reg.h
> +++ b/drivers/gpu/drm/msm/dp/dp_reg.h
> @@ -332,6 +332,12 @@
>
> /* DP MST related registers */
> #define REG_DP_MST_ACT (0x00000500)
> +#define REG_DP_DP0_RG (0x000004F8)
> +#define REG_DP_DP1_RG (0x000004FC)
> +#define REG_DP_DP0_TIMESLOT_1_32 (0x00000404)
> +#define REG_DP_DP0_TIMESLOT_33_63 (0x00000408)
> +#define REG_DP_DP1_TIMESLOT_1_32 (0x0000040C)
> +#define REG_DP_DP1_TIMESLOT_33_63 (0x00000410)
- sort by register
- lowercase hex numbers
>
> /* DP HDCP 1.3 registers */
> #define DP_HDCP_CTRL (0x0A0)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 22/45] drm/msm/dp: add support to send vcpf packets in dp controller
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (20 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 21/45] drm/msm/dp: add support for mst channel slot allocation Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 6:20 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 23/45] drm/msm/dp: always program MST_FIFO_CONSTANT_FILL for MST Abhinav Kumar
` (24 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
VC payload fill sequence is inserted by the DP controller in the
absence of stream symbols that is before stream is disabled. Add
support to send the VCPF sequence for msm dp controller.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.c | 25 +++++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_catalog.h | 4 ++++
drivers/gpu/drm/msm/dp/dp_ctrl.c | 40 +++++++++++++++++++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_ctrl.h | 1 +
drivers/gpu/drm/msm/dp/dp_display.c | 5 ++++-
drivers/gpu/drm/msm/dp/dp_reg.h | 3 ++-
6 files changed, 76 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index f9d21444d7891bcd043d282b31ae75711add4817..4826a698979ce7c37112812299879411c5743fa9 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -50,6 +50,11 @@
(PSR_UPDATE_INT | PSR_CAPTURE_INT | PSR_EXIT_INT | \
PSR_UPDATE_ERROR_INT | PSR_WAKE_ERROR_INT)
+#define DP_INTERRUPT_STATUS5 \
+ (DP_INTR_DP0_VCPF_SENT | DP_INTR_DP1_VCPF_SENT)
+#define DP_INTERRUPT_STATUS5_MASK \
+ (DP_INTERRUPT_STATUS5 << DP_INTERRUPT_STATUS_MASK_SHIFT)
+
#define DP_INTERRUPT_MASK4 \
(PSR_UPDATE_MASK | PSR_CAPTURE_MASK | PSR_EXIT_MASK | \
PSR_UPDATE_ERROR_MASK | PSR_WAKE_ERROR_MASK)
@@ -694,9 +699,12 @@ void msm_dp_catalog_ctrl_enable_irq(struct msm_dp_catalog *msm_dp_catalog,
DP_INTERRUPT_STATUS1_MASK);
msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS2,
DP_INTERRUPT_STATUS2_MASK);
+ msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS5,
+ DP_INTERRUPT_STATUS5_MASK);
} else {
msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS, 0x00);
msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS2, 0x00);
+ msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS5, 0x00);
}
}
@@ -850,6 +858,23 @@ int msm_dp_catalog_ctrl_get_interrupt(struct msm_dp_catalog *msm_dp_catalog)
return intr;
}
+int msm_dp_catalog_ctrl_get_interrupt_5(struct msm_dp_catalog *msm_dp_catalog)
+{
+ struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
+ struct msm_dp_catalog_private,
+ msm_dp_catalog);
+ u32 intr, intr_ack;
+
+ intr = msm_dp_read_ahb(catalog, REG_DP_INTR_STATUS5);
+ intr &= ~DP_INTERRUPT_STATUS5_MASK;
+ intr_ack = (intr & DP_INTERRUPT_STATUS5)
+ << DP_INTERRUPT_STATUS_ACK_SHIFT;
+ msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS5,
+ intr_ack | DP_INTERRUPT_STATUS5_MASK);
+
+ return intr;
+}
+
void msm_dp_catalog_ctrl_phy_reset(struct msm_dp_catalog *msm_dp_catalog)
{
struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
index 560016e2f929d4b92d6ea764d81a099c09c0e668..323858c587f85996d296156c7b8b201cdb7b7eb4 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.h
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
@@ -28,6 +28,9 @@
#define DP_INTR_FRAME_END BIT(6)
#define DP_INTR_CRC_UPDATED BIT(9)
+#define DP_INTR_DP0_VCPF_SENT BIT(0)
+#define DP_INTR_DP1_VCPF_SENT BIT(3)
+
#define DP_HW_VERSION_1_0 0x10000000
#define DP_HW_VERSION_1_2 0x10020000
@@ -103,6 +106,7 @@ u32 msm_dp_catalog_link_is_connected(struct msm_dp_catalog *msm_dp_catalog);
u32 msm_dp_catalog_hpd_get_intr_status(struct msm_dp_catalog *msm_dp_catalog);
void msm_dp_catalog_ctrl_phy_reset(struct msm_dp_catalog *msm_dp_catalog);
int msm_dp_catalog_ctrl_get_interrupt(struct msm_dp_catalog *msm_dp_catalog);
+int msm_dp_catalog_ctrl_get_interrupt_5(struct msm_dp_catalog *msm_dp_catalog);
u32 msm_dp_catalog_ctrl_read_psr_interrupt_status(struct msm_dp_catalog *msm_dp_catalog);
void msm_dp_catalog_ctrl_update_transfer_unit(struct msm_dp_catalog *msm_dp_catalog,
u32 msm_dp_tu, u32 valid_boundary,
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 14562def1e70b769434243d1ce72661a7b4d4c6b..2288c379283c721a01c81302f8d307d0b3c76527 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -27,6 +27,11 @@
#define DP_CTRL_INTR_READY_FOR_VIDEO BIT(0)
#define DP_CTRL_INTR_IDLE_PATTERN_SENT BIT(3)
+#define DP_CTRL_INTR_DP0_VCPF_SENT BIT(0)
+#define DP_CTRL_INTR_DP1_VCPF_SENT BIT(3)
+
+#define MST_DP0_PUSH_VCPF BIT(12)
+#define MST_DP1_PUSH_VCPF BIT(14)
#define MR_LINK_TRAINING1 0x8
#define MR_LINK_SYMBOL_ERM 0x80
@@ -144,6 +149,34 @@ void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl)
drm_dbg_dp(ctrl->drm_dev, "mainlink off\n");
}
+void msm_dp_ctrl_push_vcpf(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
+{
+ u32 state = 0x0;
+ struct msm_dp_ctrl_private *ctrl;
+
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+
+ if (msm_dp_panel->stream_id >= DP_STREAM_MAX) {
+ DRM_ERROR("invalid input\n");
+ return;
+ }
+
+ if (msm_dp_panel->stream_id == DP_STREAM_0)
+ state |= MST_DP0_PUSH_VCPF;
+ else
+ state |= MST_DP1_PUSH_VCPF;
+
+ reinit_completion(&ctrl->idle_comp);
+
+ msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, state);
+
+ if (!wait_for_completion_timeout(&ctrl->idle_comp,
+ IDLE_PATTERN_COMPLETION_TIMEOUT_JIFFIES))
+ pr_warn("PUSH_VCPF pattern timedout\n");
+
+ drm_dbg_dp(ctrl->drm_dev, "mainlink off\n");
+}
+
static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl,
struct msm_dp_panel *msm_dp_panel)
{
@@ -2332,6 +2365,13 @@ irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
ret = IRQ_HANDLED;
}
+ isr = msm_dp_catalog_ctrl_get_interrupt_5(ctrl->catalog);
+ if (isr & (DP_INTR_DP0_VCPF_SENT | DP_INTR_DP1_VCPF_SENT)) {
+ drm_dbg_dp(ctrl->drm_dev, "vcpf sent\n");
+ complete(&ctrl->idle_comp);
+ ret = IRQ_HANDLED;
+ }
+
return ret;
}
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index b126651da24b3abdaf540268758b37dca9fe1291..9ad7022d6217572395d69294c3cc4d4dbaddf0ac 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -52,5 +52,6 @@ void msm_dp_ctrl_mst_stream_channel_slot_setup(struct msm_dp_ctrl *msm_dp_ctrl,
void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
enum msm_dp_stream_id strm,
u32 start_slot, u32 tot_slots);
+void msm_dp_ctrl_push_vcpf(struct msm_dp_ctrl *dp_ctrl, struct msm_dp_panel *msm_dp_panel);
#endif /* _DP_CTRL_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 2a4a79317153817cb24537ea95fad07c9bc20715..1dfc82211c50bb4ed239f9730b91c33c4897c78f 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -1646,7 +1646,10 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
- msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
+ if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT)
+ msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_display->panel);
+ else
+ msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT) {
msm_dp_ctrl_mst_stream_channel_slot_setup(msm_dp_display->ctrl,
diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
index 8bc2a431462fc1fb45b1fe8e43a0a0ec7f75e5b1..c7532217b369c6235b2fe5fe9c86642d5c2712cb 100644
--- a/drivers/gpu/drm/msm/dp/dp_reg.h
+++ b/drivers/gpu/drm/msm/dp/dp_reg.h
@@ -24,8 +24,9 @@
#define REG_DP_INTR_STATUS (0x00000020)
#define REG_DP_INTR_STATUS2 (0x00000024)
#define REG_DP_INTR_STATUS3 (0x00000028)
-
#define REG_DP_INTR_STATUS4 (0x0000002C)
+#define REG_DP_INTR_STATUS5 (0x00000034)
+
#define PSR_UPDATE_INT (0x00000001)
#define PSR_CAPTURE_INT (0x00000004)
#define PSR_EXIT_INT (0x00000010)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 22/45] drm/msm/dp: add support to send vcpf packets in dp controller
2024-12-06 4:31 ` [PATCH 22/45] drm/msm/dp: add support to send vcpf packets in dp controller Abhinav Kumar
@ 2024-12-08 6:20 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 6:20 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:53PM -0800, Abhinav Kumar wrote:
> VC payload fill sequence is inserted by the DP controller in the
> absence of stream symbols that is before stream is disabled. Add
> support to send the VCPF sequence for msm dp controller.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 25 +++++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_catalog.h | 4 ++++
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 40 +++++++++++++++++++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 1 +
> drivers/gpu/drm/msm/dp/dp_display.c | 5 ++++-
> drivers/gpu/drm/msm/dp/dp_reg.h | 3 ++-
> 6 files changed, 76 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
> index f9d21444d7891bcd043d282b31ae75711add4817..4826a698979ce7c37112812299879411c5743fa9 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> @@ -50,6 +50,11 @@
> (PSR_UPDATE_INT | PSR_CAPTURE_INT | PSR_EXIT_INT | \
> PSR_UPDATE_ERROR_INT | PSR_WAKE_ERROR_INT)
>
> +#define DP_INTERRUPT_STATUS5 \
> + (DP_INTR_DP0_VCPF_SENT | DP_INTR_DP1_VCPF_SENT)
> +#define DP_INTERRUPT_STATUS5_MASK \
> + (DP_INTERRUPT_STATUS5 << DP_INTERRUPT_STATUS_MASK_SHIFT)
> +
> #define DP_INTERRUPT_MASK4 \
> (PSR_UPDATE_MASK | PSR_CAPTURE_MASK | PSR_EXIT_MASK | \
> PSR_UPDATE_ERROR_MASK | PSR_WAKE_ERROR_MASK)
> @@ -694,9 +699,12 @@ void msm_dp_catalog_ctrl_enable_irq(struct msm_dp_catalog *msm_dp_catalog,
> DP_INTERRUPT_STATUS1_MASK);
> msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS2,
> DP_INTERRUPT_STATUS2_MASK);
> + msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS5,
> + DP_INTERRUPT_STATUS5_MASK);
> } else {
> msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS, 0x00);
> msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS2, 0x00);
> + msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS5, 0x00);
> }
> }
>
> @@ -850,6 +858,23 @@ int msm_dp_catalog_ctrl_get_interrupt(struct msm_dp_catalog *msm_dp_catalog)
> return intr;
> }
>
> +int msm_dp_catalog_ctrl_get_interrupt_5(struct msm_dp_catalog *msm_dp_catalog)
> +{
> + struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
> + struct msm_dp_catalog_private,
> + msm_dp_catalog);
> + u32 intr, intr_ack;
> +
> + intr = msm_dp_read_ahb(catalog, REG_DP_INTR_STATUS5);
> + intr &= ~DP_INTERRUPT_STATUS5_MASK;
> + intr_ack = (intr & DP_INTERRUPT_STATUS5)
> + << DP_INTERRUPT_STATUS_ACK_SHIFT;
> + msm_dp_write_ahb(catalog, REG_DP_INTR_STATUS5,
> + intr_ack | DP_INTERRUPT_STATUS5_MASK);
> +
> + return intr;
> +}
> +
> void msm_dp_catalog_ctrl_phy_reset(struct msm_dp_catalog *msm_dp_catalog)
> {
> struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
> index 560016e2f929d4b92d6ea764d81a099c09c0e668..323858c587f85996d296156c7b8b201cdb7b7eb4 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
> @@ -28,6 +28,9 @@
> #define DP_INTR_FRAME_END BIT(6)
> #define DP_INTR_CRC_UPDATED BIT(9)
>
> +#define DP_INTR_DP0_VCPF_SENT BIT(0)
> +#define DP_INTR_DP1_VCPF_SENT BIT(3)
> +
> #define DP_HW_VERSION_1_0 0x10000000
> #define DP_HW_VERSION_1_2 0x10020000
>
> @@ -103,6 +106,7 @@ u32 msm_dp_catalog_link_is_connected(struct msm_dp_catalog *msm_dp_catalog);
> u32 msm_dp_catalog_hpd_get_intr_status(struct msm_dp_catalog *msm_dp_catalog);
> void msm_dp_catalog_ctrl_phy_reset(struct msm_dp_catalog *msm_dp_catalog);
> int msm_dp_catalog_ctrl_get_interrupt(struct msm_dp_catalog *msm_dp_catalog);
> +int msm_dp_catalog_ctrl_get_interrupt_5(struct msm_dp_catalog *msm_dp_catalog);
> u32 msm_dp_catalog_ctrl_read_psr_interrupt_status(struct msm_dp_catalog *msm_dp_catalog);
I'd prefer some uniformity here. read_psr_interrupt_status() returns
REG_DP_INTR_STATUS4, get_interrupt() returns REG_DP_INTR_STATUS2 value,
Now you are adding get_interrupt_5(). Could you please make that similar
and logical?
> void msm_dp_catalog_ctrl_update_transfer_unit(struct msm_dp_catalog *msm_dp_catalog,
> u32 msm_dp_tu, u32 valid_boundary,
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 14562def1e70b769434243d1ce72661a7b4d4c6b..2288c379283c721a01c81302f8d307d0b3c76527 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -27,6 +27,11 @@
>
> #define DP_CTRL_INTR_READY_FOR_VIDEO BIT(0)
> #define DP_CTRL_INTR_IDLE_PATTERN_SENT BIT(3)
> +#define DP_CTRL_INTR_DP0_VCPF_SENT BIT(0)
> +#define DP_CTRL_INTR_DP1_VCPF_SENT BIT(3)
> +
> +#define MST_DP0_PUSH_VCPF BIT(12)
> +#define MST_DP1_PUSH_VCPF BIT(14)
>
> #define MR_LINK_TRAINING1 0x8
> #define MR_LINK_SYMBOL_ERM 0x80
> @@ -144,6 +149,34 @@ void msm_dp_ctrl_push_idle(struct msm_dp_ctrl *msm_dp_ctrl)
> drm_dbg_dp(ctrl->drm_dev, "mainlink off\n");
> }
>
> +void msm_dp_ctrl_push_vcpf(struct msm_dp_ctrl *msm_dp_ctrl, struct msm_dp_panel *msm_dp_panel)
> +{
> + u32 state = 0x0;
> + struct msm_dp_ctrl_private *ctrl;
> +
> + ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> +
> + if (msm_dp_panel->stream_id >= DP_STREAM_MAX) {
> + DRM_ERROR("invalid input\n");
> + return;
> + }
> +
> + if (msm_dp_panel->stream_id == DP_STREAM_0)
> + state |= MST_DP0_PUSH_VCPF;
> + else
> + state |= MST_DP1_PUSH_VCPF;
> +
> + reinit_completion(&ctrl->idle_comp);
> +
> + msm_dp_catalog_ctrl_state_ctrl(ctrl->catalog, state);
> +
> + if (!wait_for_completion_timeout(&ctrl->idle_comp,
> + IDLE_PATTERN_COMPLETION_TIMEOUT_JIFFIES))
> + pr_warn("PUSH_VCPF pattern timedout\n");
> +
> + drm_dbg_dp(ctrl->drm_dev, "mainlink off\n");
??
The comment seems misplaced or wrong to me. The msm_dp_ctrl_push_vcpf()
doesn't turn link off.
> +}
> +
> static void msm_dp_ctrl_config_ctrl(struct msm_dp_ctrl_private *ctrl,
> struct msm_dp_panel *msm_dp_panel)
> {
> @@ -2332,6 +2365,13 @@ irqreturn_t msm_dp_ctrl_isr(struct msm_dp_ctrl *msm_dp_ctrl)
> ret = IRQ_HANDLED;
> }
>
> + isr = msm_dp_catalog_ctrl_get_interrupt_5(ctrl->catalog);
> + if (isr & (DP_INTR_DP0_VCPF_SENT | DP_INTR_DP1_VCPF_SENT)) {
> + drm_dbg_dp(ctrl->drm_dev, "vcpf sent\n");
> + complete(&ctrl->idle_comp);
> + ret = IRQ_HANDLED;
> + }
> +
> return ret;
> }
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index b126651da24b3abdaf540268758b37dca9fe1291..9ad7022d6217572395d69294c3cc4d4dbaddf0ac 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -52,5 +52,6 @@ void msm_dp_ctrl_mst_stream_channel_slot_setup(struct msm_dp_ctrl *msm_dp_ctrl,
> void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
> enum msm_dp_stream_id strm,
> u32 start_slot, u32 tot_slots);
> +void msm_dp_ctrl_push_vcpf(struct msm_dp_ctrl *dp_ctrl, struct msm_dp_panel *msm_dp_panel);
>
> #endif /* _DP_CTRL_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 2a4a79317153817cb24537ea95fad07c9bc20715..1dfc82211c50bb4ed239f9730b91c33c4897c78f 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -1646,7 +1646,10 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
>
> - msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
> + if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT)
Same comment regarding default vs SST
> + msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_display->panel);
> + else
> + msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
>
> if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT) {
> msm_dp_ctrl_mst_stream_channel_slot_setup(msm_dp_display->ctrl,
> diff --git a/drivers/gpu/drm/msm/dp/dp_reg.h b/drivers/gpu/drm/msm/dp/dp_reg.h
> index 8bc2a431462fc1fb45b1fe8e43a0a0ec7f75e5b1..c7532217b369c6235b2fe5fe9c86642d5c2712cb 100644
> --- a/drivers/gpu/drm/msm/dp/dp_reg.h
> +++ b/drivers/gpu/drm/msm/dp/dp_reg.h
> @@ -24,8 +24,9 @@
> #define REG_DP_INTR_STATUS (0x00000020)
> #define REG_DP_INTR_STATUS2 (0x00000024)
> #define REG_DP_INTR_STATUS3 (0x00000028)
> -
> #define REG_DP_INTR_STATUS4 (0x0000002C)
> +#define REG_DP_INTR_STATUS5 (0x00000034)
> +
> #define PSR_UPDATE_INT (0x00000001)
> #define PSR_CAPTURE_INT (0x00000004)
> #define PSR_EXIT_INT (0x00000010)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 23/45] drm/msm/dp: always program MST_FIFO_CONSTANT_FILL for MST
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (21 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 22/45] drm/msm/dp: add support to send vcpf packets in dp controller Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:44 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 24/45] drm/msm/dp: abstract out the dp_display stream helpers to accept a panel Abhinav Kumar
` (23 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
As required by the hardware programming guide, always program
the MST_FIFO_CONSTANT_FILL for MST use-cases.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_catalog.c | 24 ++++++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_catalog.h | 1 +
drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 +++
3 files changed, 28 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
index 4826a698979ce7c37112812299879411c5743fa9..3cfa9fd0c6f5e664a65e6d9b74e5c9f22ad711f2 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.c
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
@@ -1009,6 +1009,30 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
return 0;
}
+int msm_dp_catalog_mst_async_fifo(struct msm_dp_catalog *msm_dp_catalog)
+{
+ struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
+ struct msm_dp_catalog_private,
+ msm_dp_catalog);
+
+ u32 reg;
+
+ if (msm_dp_catalog->stream_id == DP_STREAM_0)
+ reg = msm_dp_read_p0(catalog, MMSS_DP_ASYNC_FIFO_CONFIG);
+ else
+ reg = msm_dp_read_p1(catalog, MMSS_DP_ASYNC_FIFO_CONFIG);
+
+ /* enable MST_FIFO_CONSTANT_FILL */
+ reg |= BIT(0);
+
+ if (msm_dp_catalog->stream_id == DP_STREAM_0)
+ msm_dp_write_p0(catalog, MMSS_DP_ASYNC_FIFO_CONFIG, reg);
+ else
+ msm_dp_write_p1(catalog, MMSS_DP_ASYNC_FIFO_CONFIG, reg);
+
+ return 0;
+}
+
static void msm_dp_catalog_panel_send_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog, struct dp_sdp *vsc_sdp)
{
struct msm_dp_catalog_private *catalog;
diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
index 323858c587f85996d296156c7b8b201cdb7b7eb4..0025ecc0adb2f351c44f10af82332a6622749416 100644
--- a/drivers/gpu/drm/msm/dp/dp_catalog.h
+++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
@@ -150,5 +150,6 @@ void msm_dp_catalog_mst_channel_alloc(struct msm_dp_catalog *ctrl,
u32 ch, u32 ch_start_slot, u32 tot_slot_cnt);
void msm_dp_catalog_ctrl_update_rg(struct msm_dp_catalog *ctrl, u32 stream,
u32 x_int, u32 y_frac_enum);
+int msm_dp_catalog_mst_async_fifo(struct msm_dp_catalog *dp_catalog);
#endif /* _DP_CATALOG_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index 2288c379283c721a01c81302f8d307d0b3c76527..d4915a962f97b3d2a347456e197265a5dc043eb0 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -236,6 +236,9 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
cc = msm_dp_link_get_colorimetry_config(ctrl->link);
msm_dp_catalog_ctrl_config_misc(ctrl->catalog, cc, tb);
msm_dp_panel_timing_cfg(msm_dp_panel);
+
+ if (ctrl->mst_active)
+ msm_dp_catalog_mst_async_fifo(ctrl->catalog);
}
/*
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 23/45] drm/msm/dp: always program MST_FIFO_CONSTANT_FILL for MST
2024-12-06 4:31 ` [PATCH 23/45] drm/msm/dp: always program MST_FIFO_CONSTANT_FILL for MST Abhinav Kumar
@ 2024-12-08 11:44 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:44 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:54PM -0800, Abhinav Kumar wrote:
> As required by the hardware programming guide, always program
> the MST_FIFO_CONSTANT_FILL for MST use-cases.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_catalog.c | 24 ++++++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_catalog.h | 1 +
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 3 +++
> 3 files changed, 28 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c b/drivers/gpu/drm/msm/dp/dp_catalog.c
> index 4826a698979ce7c37112812299879411c5743fa9..3cfa9fd0c6f5e664a65e6d9b74e5c9f22ad711f2 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> @@ -1009,6 +1009,30 @@ int msm_dp_catalog_panel_timing_cfg(struct msm_dp_catalog *msm_dp_catalog, u32 t
> return 0;
> }
>
> +int msm_dp_catalog_mst_async_fifo(struct msm_dp_catalog *msm_dp_catalog)
> +{
> + struct msm_dp_catalog_private *catalog = container_of(msm_dp_catalog,
> + struct msm_dp_catalog_private,
> + msm_dp_catalog);
> +
> + u32 reg;
> +
> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
> + reg = msm_dp_read_p0(catalog, MMSS_DP_ASYNC_FIFO_CONFIG);
> + else
> + reg = msm_dp_read_p1(catalog, MMSS_DP_ASYNC_FIFO_CONFIG);
> +
> + /* enable MST_FIFO_CONSTANT_FILL */
> + reg |= BIT(0);
#define
> +
> + if (msm_dp_catalog->stream_id == DP_STREAM_0)
> + msm_dp_write_p0(catalog, MMSS_DP_ASYNC_FIFO_CONFIG, reg);
> + else
> + msm_dp_write_p1(catalog, MMSS_DP_ASYNC_FIFO_CONFIG, reg);
> +
> + return 0;
> +}
> +
> static void msm_dp_catalog_panel_send_vsc_sdp(struct msm_dp_catalog *msm_dp_catalog, struct dp_sdp *vsc_sdp)
> {
> struct msm_dp_catalog_private *catalog;
> diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.h b/drivers/gpu/drm/msm/dp/dp_catalog.h
> index 323858c587f85996d296156c7b8b201cdb7b7eb4..0025ecc0adb2f351c44f10af82332a6622749416 100644
> --- a/drivers/gpu/drm/msm/dp/dp_catalog.h
> +++ b/drivers/gpu/drm/msm/dp/dp_catalog.h
> @@ -150,5 +150,6 @@ void msm_dp_catalog_mst_channel_alloc(struct msm_dp_catalog *ctrl,
> u32 ch, u32 ch_start_slot, u32 tot_slot_cnt);
> void msm_dp_catalog_ctrl_update_rg(struct msm_dp_catalog *ctrl, u32 stream,
> u32 x_int, u32 y_frac_enum);
> +int msm_dp_catalog_mst_async_fifo(struct msm_dp_catalog *dp_catalog);
>
> #endif /* _DP_CATALOG_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index 2288c379283c721a01c81302f8d307d0b3c76527..d4915a962f97b3d2a347456e197265a5dc043eb0 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -236,6 +236,9 @@ static void msm_dp_ctrl_configure_source_params(struct msm_dp_ctrl_private *ctrl
> cc = msm_dp_link_get_colorimetry_config(ctrl->link);
> msm_dp_catalog_ctrl_config_misc(ctrl->catalog, cc, tb);
> msm_dp_panel_timing_cfg(msm_dp_panel);
> +
> + if (ctrl->mst_active)
> + msm_dp_catalog_mst_async_fifo(ctrl->catalog);
Should it be disabled further on in a non-MST case?
> }
>
> /*
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 24/45] drm/msm/dp: abstract out the dp_display stream helpers to accept a panel
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (22 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 23/45] drm/msm/dp: always program MST_FIFO_CONSTANT_FILL for MST Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:46 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 25/45] drm/msm/dp: move link related operations to dp_display_unprepare() Abhinav Kumar
` (22 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Currently the dp_display bridge helpers, in particular the
dp_display_enable()/dp_display_disable() use the cached panel.
To be able to re-use these helpers for MST use-case abstract the
helpers to use the panel which is passed in to them.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 96 ++++++++++++++++++++++++++-----------
drivers/gpu/drm/msm/dp/dp_display.h | 8 ++++
2 files changed, 75 insertions(+), 29 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 1dfc82211c50bb4ed239f9730b91c33c4897c78f..e169cd22db960c0c30707ddbe6a79999dc2a273d 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -868,7 +868,8 @@ static int msm_dp_display_prepare(struct msm_dp_display_private *dp)
return rc;
}
-static int msm_dp_display_enable(struct msm_dp_display_private *dp)
+static int msm_dp_display_enable(struct msm_dp_display_private *dp,
+ struct msm_dp_panel *msm_dp_panel)
{
int rc = 0;
struct msm_dp *msm_dp_display = &dp->msm_dp_display;
@@ -879,7 +880,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp)
return 0;
}
- rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel, dp->max_stream);
+ rc = msm_dp_ctrl_on_stream(dp->ctrl, msm_dp_panel, dp->max_stream);
if (!rc)
msm_dp_display->power_on = true;
@@ -925,20 +926,21 @@ static void msm_dp_display_audio_notify_disable(struct msm_dp_display_private *d
msm_dp_display->audio_enabled = false;
}
-static int msm_dp_display_disable(struct msm_dp_display_private *dp)
+static int msm_dp_display_disable(struct msm_dp_display_private *dp,
+ struct msm_dp_panel *msm_dp_panel)
{
struct msm_dp *msm_dp_display = &dp->msm_dp_display;
if (!msm_dp_display->power_on)
return 0;
- msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, dp->panel);
+ msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, msm_dp_panel);
/* dongle is still connected but sinks are disconnected */
if (dp->link->sink_count == 0)
msm_dp_ctrl_psm_config(dp->ctrl);
- msm_dp_ctrl_stream_clk_off(dp->ctrl, dp->panel);
+ msm_dp_ctrl_stream_clk_off(dp->ctrl, msm_dp_panel);
msm_dp_ctrl_off_link(dp->ctrl);
@@ -1607,7 +1609,7 @@ void msm_dp_display_atomic_prepare(struct msm_dp *dp)
mutex_unlock(&msm_dp_display->event_mutex);
}
-void msm_dp_display_atomic_enable(struct msm_dp *dp)
+void msm_dp_display_enable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_dp_panel)
{
int rc = 0;
@@ -1620,16 +1622,14 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
mutex_lock(&msm_dp_display->event_mutex);
- msm_dp_display_set_stream_info(dp, msm_dp_display->panel, 0, 0, 0, 0, 0);
-
if (dp->prepared) {
- rc = msm_dp_display_enable(msm_dp_display);
+ rc = msm_dp_display_enable(msm_dp_display, msm_dp_panel);
if (rc)
DRM_ERROR("DP display enable failed, rc=%d\n", rc);
rc = msm_dp_display_post_enable(dp);
if (rc) {
DRM_ERROR("DP display post enable failed, rc=%d\n", rc);
- msm_dp_display_disable(msm_dp_display);
+ msm_dp_display_disable(msm_dp_display, msm_dp_panel);
}
}
@@ -1640,14 +1640,25 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
mutex_unlock(&msm_dp_display->event_mutex);
}
-void msm_dp_display_atomic_disable(struct msm_dp *dp)
+void msm_dp_display_atomic_enable(struct msm_dp *msm_dp)
+{
+ struct msm_dp_display_private *msm_dp_display;
+
+ msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
+
+ msm_dp_display_set_stream_info(msm_dp, msm_dp_display->panel, 0, 0, 0, 0, 0);
+
+ msm_dp_display_enable_helper(msm_dp, msm_dp_display->panel);
+}
+
+void msm_dp_display_disable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_dp_panel)
{
struct msm_dp_display_private *msm_dp_display;
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT)
- msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_display->panel);
+ msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_panel);
else
msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
@@ -1658,21 +1669,30 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
}
}
-static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
+void msm_dp_display_atomic_disable(struct msm_dp *msm_dp)
{
- struct msm_dp *msm_dp_display = &dp->msm_dp_display;
+ struct msm_dp_display_private *msm_dp_display;
- if (!msm_dp_display->prepared) {
- drm_dbg_dp(dp->drm_dev, "Link already setup, return\n");
+ msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
+
+ msm_dp_display_disable_helper(msm_dp, msm_dp_display->panel);
+}
+
+static void msm_dp_display_unprepare(struct msm_dp_display_private *msm_dp_display_priv)
+{
+ struct msm_dp *msm_dp = &msm_dp_display_priv->msm_dp_display;
+
+ if (!msm_dp->prepared) {
+ drm_dbg_dp(msm_dp->drm_dev, "Link already setup, return\n");
return;
}
- pm_runtime_put_sync(&msm_dp_display->pdev->dev);
+ pm_runtime_put_sync(&msm_dp->pdev->dev);
- msm_dp_display->prepared = false;
+ msm_dp->prepared = false;
}
-void msm_dp_display_atomic_post_disable(struct msm_dp *dp)
+void msm_dp_display_atomic_post_disable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_dp_panel)
{
u32 state;
struct msm_dp_display_private *msm_dp_display;
@@ -1691,7 +1711,7 @@ void msm_dp_display_atomic_post_disable(struct msm_dp *dp)
msm_dp_display_audio_notify_disable(msm_dp_display);
- msm_dp_display_disable(msm_dp_display);
+ msm_dp_display_disable(msm_dp_display, msm_dp_panel);
state = msm_dp_display->hpd_state;
if (state == ST_DISCONNECT_PENDING) {
@@ -1699,25 +1719,32 @@ void msm_dp_display_atomic_post_disable(struct msm_dp *dp)
msm_dp_display->hpd_state = ST_DISCONNECTED;
}
- msm_dp_display_unprepare(msm_dp_display);
-
drm_dbg_dp(dp->drm_dev, "type=%d Done\n", dp->connector_type);
mutex_unlock(&msm_dp_display->event_mutex);
}
-void msm_dp_display_mode_set(struct msm_dp *dp,
- const struct drm_display_mode *mode,
- const struct drm_display_mode *adjusted_mode)
+void msm_dp_display_atomic_post_disable(struct msm_dp *msm_dp)
{
struct msm_dp_display_private *msm_dp_display;
- struct msm_dp_panel *msm_dp_panel;
- msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
- msm_dp_panel = msm_dp_display->panel;
+ msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
+
+ msm_dp_display_atomic_post_disable_helper(msm_dp, msm_dp_display->panel);
+
+ msm_dp_display_unprepare(msm_dp_display);
+}
+
+void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
+ const struct drm_display_mode *mode,
+ const struct drm_display_mode *adjusted_mode,
+ struct msm_dp_panel *msm_dp_panel)
+{
+ struct msm_dp_display_private *msm_dp_display;
+ msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
- msm_dp_display_set_mode(dp, adjusted_mode, msm_dp_panel);
+ msm_dp_display_set_mode(msm_dp, adjusted_mode, msm_dp_panel);
/* populate wide_bus_support to different layers */
msm_dp_display->ctrl->wide_bus_en =
@@ -1726,6 +1753,17 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
}
+void msm_dp_display_mode_set(struct msm_dp *msm_dp,
+ const struct drm_display_mode *mode,
+ const struct drm_display_mode *adjusted_mode)
+{
+ struct msm_dp_display_private *msm_dp_display;
+
+ msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
+
+ msm_dp_display_mode_set_helper(msm_dp, mode, adjusted_mode, msm_dp_display->panel);
+}
+
void msm_dp_bridge_hpd_enable(struct drm_bridge *bridge)
{
struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(bridge);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 258c240de580b634c05cf5895a8e52160449eba1..2b23f2bf7535d3fd513d40a8411a1903fcd560b0 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -56,5 +56,13 @@ enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
int msm_dp_display_set_stream_info(struct msm_dp *dp,
struct msm_dp_panel *panel, u32 strm_id,
u32 start_slot, u32 num_slots, u32 pbn, int vcpi);
+void msm_dp_display_enable_helper(struct msm_dp *msm_dp, struct msm_dp_panel *msm_dp_panel);
+void msm_dp_display_disable_helper(struct msm_dp *msm_dp, struct msm_dp_panel *msm_dp_panel);
+void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
+ const struct drm_display_mode *mode,
+ const struct drm_display_mode *adjusted_mode,
+ struct msm_dp_panel *msm_dp_panel);
+void msm_dp_display_atomic_post_disable_helper(struct msm_dp *msm_dp,
+ struct msm_dp_panel *msm_dp_panel);
#endif /* _DP_DISPLAY_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 24/45] drm/msm/dp: abstract out the dp_display stream helpers to accept a panel
2024-12-06 4:31 ` [PATCH 24/45] drm/msm/dp: abstract out the dp_display stream helpers to accept a panel Abhinav Kumar
@ 2024-12-08 11:46 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:46 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:55PM -0800, Abhinav Kumar wrote:
> Currently the dp_display bridge helpers, in particular the
> dp_display_enable()/dp_display_disable() use the cached panel.
> To be able to re-use these helpers for MST use-case abstract the
> helpers to use the panel which is passed in to them.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 96 ++++++++++++++++++++++++++-----------
> drivers/gpu/drm/msm/dp/dp_display.h | 8 ++++
> 2 files changed, 75 insertions(+), 29 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 1dfc82211c50bb4ed239f9730b91c33c4897c78f..e169cd22db960c0c30707ddbe6a79999dc2a273d 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -868,7 +868,8 @@ static int msm_dp_display_prepare(struct msm_dp_display_private *dp)
> return rc;
> }
>
> -static int msm_dp_display_enable(struct msm_dp_display_private *dp)
> +static int msm_dp_display_enable(struct msm_dp_display_private *dp,
> + struct msm_dp_panel *msm_dp_panel)
> {
> int rc = 0;
> struct msm_dp *msm_dp_display = &dp->msm_dp_display;
> @@ -879,7 +880,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp)
> return 0;
> }
>
> - rc = msm_dp_ctrl_on_stream(dp->ctrl, dp->panel, dp->max_stream);
> + rc = msm_dp_ctrl_on_stream(dp->ctrl, msm_dp_panel, dp->max_stream);
> if (!rc)
> msm_dp_display->power_on = true;
>
> @@ -925,20 +926,21 @@ static void msm_dp_display_audio_notify_disable(struct msm_dp_display_private *d
> msm_dp_display->audio_enabled = false;
> }
>
> -static int msm_dp_display_disable(struct msm_dp_display_private *dp)
> +static int msm_dp_display_disable(struct msm_dp_display_private *dp,
> + struct msm_dp_panel *msm_dp_panel)
> {
> struct msm_dp *msm_dp_display = &dp->msm_dp_display;
>
> if (!msm_dp_display->power_on)
> return 0;
>
> - msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, dp->panel);
> + msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, msm_dp_panel);
>
> /* dongle is still connected but sinks are disconnected */
> if (dp->link->sink_count == 0)
> msm_dp_ctrl_psm_config(dp->ctrl);
>
> - msm_dp_ctrl_stream_clk_off(dp->ctrl, dp->panel);
> + msm_dp_ctrl_stream_clk_off(dp->ctrl, msm_dp_panel);
>
> msm_dp_ctrl_off_link(dp->ctrl);
>
> @@ -1607,7 +1609,7 @@ void msm_dp_display_atomic_prepare(struct msm_dp *dp)
> mutex_unlock(&msm_dp_display->event_mutex);
> }
>
> -void msm_dp_display_atomic_enable(struct msm_dp *dp)
> +void msm_dp_display_enable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_dp_panel)
> {
> int rc = 0;
>
> @@ -1620,16 +1622,14 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
>
> mutex_lock(&msm_dp_display->event_mutex);
>
> - msm_dp_display_set_stream_info(dp, msm_dp_display->panel, 0, 0, 0, 0, 0);
> -
Why? You have added it few patches ago.
> if (dp->prepared) {
> - rc = msm_dp_display_enable(msm_dp_display);
> + rc = msm_dp_display_enable(msm_dp_display, msm_dp_panel);
> if (rc)
> DRM_ERROR("DP display enable failed, rc=%d\n", rc);
> rc = msm_dp_display_post_enable(dp);
> if (rc) {
> DRM_ERROR("DP display post enable failed, rc=%d\n", rc);
> - msm_dp_display_disable(msm_dp_display);
> + msm_dp_display_disable(msm_dp_display, msm_dp_panel);
> }
> }
>
> @@ -1640,14 +1640,25 @@ void msm_dp_display_atomic_enable(struct msm_dp *dp)
> mutex_unlock(&msm_dp_display->event_mutex);
> }
>
> -void msm_dp_display_atomic_disable(struct msm_dp *dp)
> +void msm_dp_display_atomic_enable(struct msm_dp *msm_dp)
> +{
> + struct msm_dp_display_private *msm_dp_display;
> +
> + msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
> +
> + msm_dp_display_set_stream_info(msm_dp, msm_dp_display->panel, 0, 0, 0, 0, 0);
> +
> + msm_dp_display_enable_helper(msm_dp, msm_dp_display->panel);
> +}
> +
> +void msm_dp_display_disable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_dp_panel)
> {
> struct msm_dp_display_private *msm_dp_display;
>
> msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
>
> if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT)
> - msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_display->panel);
> + msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_panel);
> else
> msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
>
> @@ -1658,21 +1669,30 @@ void msm_dp_display_atomic_disable(struct msm_dp *dp)
> }
> }
>
> -static void msm_dp_display_unprepare(struct msm_dp_display_private *dp)
> +void msm_dp_display_atomic_disable(struct msm_dp *msm_dp)
> {
> - struct msm_dp *msm_dp_display = &dp->msm_dp_display;
> + struct msm_dp_display_private *msm_dp_display;
>
> - if (!msm_dp_display->prepared) {
> - drm_dbg_dp(dp->drm_dev, "Link already setup, return\n");
> + msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
> +
> + msm_dp_display_disable_helper(msm_dp, msm_dp_display->panel);
> +}
> +
> +static void msm_dp_display_unprepare(struct msm_dp_display_private *msm_dp_display_priv)
> +{
> + struct msm_dp *msm_dp = &msm_dp_display_priv->msm_dp_display;
> +
> + if (!msm_dp->prepared) {
> + drm_dbg_dp(msm_dp->drm_dev, "Link already setup, return\n");
> return;
> }
>
> - pm_runtime_put_sync(&msm_dp_display->pdev->dev);
> + pm_runtime_put_sync(&msm_dp->pdev->dev);
>
> - msm_dp_display->prepared = false;
> + msm_dp->prepared = false;
> }
>
> -void msm_dp_display_atomic_post_disable(struct msm_dp *dp)
> +void msm_dp_display_atomic_post_disable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_dp_panel)
> {
> u32 state;
> struct msm_dp_display_private *msm_dp_display;
> @@ -1691,7 +1711,7 @@ void msm_dp_display_atomic_post_disable(struct msm_dp *dp)
>
> msm_dp_display_audio_notify_disable(msm_dp_display);
>
> - msm_dp_display_disable(msm_dp_display);
> + msm_dp_display_disable(msm_dp_display, msm_dp_panel);
>
> state = msm_dp_display->hpd_state;
> if (state == ST_DISCONNECT_PENDING) {
> @@ -1699,25 +1719,32 @@ void msm_dp_display_atomic_post_disable(struct msm_dp *dp)
> msm_dp_display->hpd_state = ST_DISCONNECTED;
> }
>
> - msm_dp_display_unprepare(msm_dp_display);
> -
Please split refactoring to a separate patch.
> drm_dbg_dp(dp->drm_dev, "type=%d Done\n", dp->connector_type);
>
> mutex_unlock(&msm_dp_display->event_mutex);
> }
>
> -void msm_dp_display_mode_set(struct msm_dp *dp,
> - const struct drm_display_mode *mode,
> - const struct drm_display_mode *adjusted_mode)
> +void msm_dp_display_atomic_post_disable(struct msm_dp *msm_dp)
> {
> struct msm_dp_display_private *msm_dp_display;
> - struct msm_dp_panel *msm_dp_panel;
>
> - msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
> - msm_dp_panel = msm_dp_display->panel;
> + msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
> +
> + msm_dp_display_atomic_post_disable_helper(msm_dp, msm_dp_display->panel);
> +
> + msm_dp_display_unprepare(msm_dp_display);
> +}
> +
> +void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
> + const struct drm_display_mode *mode,
> + const struct drm_display_mode *adjusted_mode,
> + struct msm_dp_panel *msm_dp_panel)
> +{
> + struct msm_dp_display_private *msm_dp_display;
>
> + msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
>
> - msm_dp_display_set_mode(dp, adjusted_mode, msm_dp_panel);
> + msm_dp_display_set_mode(msm_dp, adjusted_mode, msm_dp_panel);
>
> /* populate wide_bus_support to different layers */
> msm_dp_display->ctrl->wide_bus_en =
> @@ -1726,6 +1753,17 @@ void msm_dp_display_mode_set(struct msm_dp *dp,
> msm_dp_panel->msm_dp_mode.out_fmt_is_yuv_420 ? false : msm_dp_display->wide_bus_supported;
> }
>
> +void msm_dp_display_mode_set(struct msm_dp *msm_dp,
> + const struct drm_display_mode *mode,
> + const struct drm_display_mode *adjusted_mode)
> +{
> + struct msm_dp_display_private *msm_dp_display;
> +
> + msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
> +
> + msm_dp_display_mode_set_helper(msm_dp, mode, adjusted_mode, msm_dp_display->panel);
> +}
> +
> void msm_dp_bridge_hpd_enable(struct drm_bridge *bridge)
> {
> struct msm_dp_bridge *msm_dp_bridge = to_dp_bridge(bridge);
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> index 258c240de580b634c05cf5895a8e52160449eba1..2b23f2bf7535d3fd513d40a8411a1903fcd560b0 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.h
> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> @@ -56,5 +56,13 @@ enum drm_mode_status msm_dp_display_mode_valid(struct msm_dp *dp,
> int msm_dp_display_set_stream_info(struct msm_dp *dp,
> struct msm_dp_panel *panel, u32 strm_id,
> u32 start_slot, u32 num_slots, u32 pbn, int vcpi);
> +void msm_dp_display_enable_helper(struct msm_dp *msm_dp, struct msm_dp_panel *msm_dp_panel);
> +void msm_dp_display_disable_helper(struct msm_dp *msm_dp, struct msm_dp_panel *msm_dp_panel);
> +void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
> + const struct drm_display_mode *mode,
> + const struct drm_display_mode *adjusted_mode,
> + struct msm_dp_panel *msm_dp_panel);
> +void msm_dp_display_atomic_post_disable_helper(struct msm_dp *msm_dp,
> + struct msm_dp_panel *msm_dp_panel);
>
> #endif /* _DP_DISPLAY_H_ */
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 25/45] drm/msm/dp: move link related operations to dp_display_unprepare()
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (23 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 24/45] drm/msm/dp: abstract out the dp_display stream helpers to accept a panel Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:48 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 26/45] drm/msm/dp: replace power_on with active_stream_cnt for dp_display Abhinav Kumar
` (21 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Move the link related operations to dp_display_unprepare() and keep
only stream related operations in dp_display_disable().
Make dp_display_unprepare() available to other clients such as DP MST.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 31 ++++++++++++++++---------------
drivers/gpu/drm/msm/dp/dp_display.h | 2 ++
2 files changed, 18 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index e169cd22db960c0c30707ddbe6a79999dc2a273d..d5b8fd1d4d736ffa7929b9798601dcef0dea5211 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -936,20 +936,8 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp,
msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, msm_dp_panel);
- /* dongle is still connected but sinks are disconnected */
- if (dp->link->sink_count == 0)
- msm_dp_ctrl_psm_config(dp->ctrl);
-
msm_dp_ctrl_stream_clk_off(dp->ctrl, msm_dp_panel);
- msm_dp_ctrl_off_link(dp->ctrl);
-
- /* re-init the PHY so that we can listen to Dongle disconnect */
- if (dp->link->sink_count == 0)
- msm_dp_ctrl_reinit_phy(dp->ctrl);
- else
- msm_dp_display_host_phy_exit(dp);
-
msm_dp_display->power_on = false;
drm_dbg_dp(dp->drm_dev, "sink count: %d\n", dp->link->sink_count);
@@ -1678,15 +1666,28 @@ void msm_dp_display_atomic_disable(struct msm_dp *msm_dp)
msm_dp_display_disable_helper(msm_dp, msm_dp_display->panel);
}
-static void msm_dp_display_unprepare(struct msm_dp_display_private *msm_dp_display_priv)
+void msm_dp_display_unprepare(struct msm_dp *msm_dp)
{
- struct msm_dp *msm_dp = &msm_dp_display_priv->msm_dp_display;
+ struct msm_dp_display_private *msm_dp_display;
+ msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
if (!msm_dp->prepared) {
drm_dbg_dp(msm_dp->drm_dev, "Link already setup, return\n");
return;
}
+ /* dongle is still connected but sinks are disconnected */
+ if (msm_dp_display->link->sink_count == 0)
+ msm_dp_ctrl_psm_config(msm_dp_display->ctrl);
+
+ msm_dp_ctrl_off_link(msm_dp_display->ctrl);
+
+ /* re-init the PHY so that we can listen to Dongle disconnect */
+ if (msm_dp_display->link->sink_count == 0)
+ msm_dp_ctrl_reinit_phy(msm_dp_display->ctrl);
+ else
+ msm_dp_display_host_phy_exit(msm_dp_display);
+
pm_runtime_put_sync(&msm_dp->pdev->dev);
msm_dp->prepared = false;
@@ -1732,7 +1733,7 @@ void msm_dp_display_atomic_post_disable(struct msm_dp *msm_dp)
msm_dp_display_atomic_post_disable_helper(msm_dp, msm_dp_display->panel);
- msm_dp_display_unprepare(msm_dp_display);
+ msm_dp_display_unprepare(msm_dp);
}
void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 2b23f2bf7535d3fd513d40a8411a1903fcd560b0..82eb1c6ed1467b21742bda8eaae9c51d3207e997 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -65,4 +65,6 @@ void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
void msm_dp_display_atomic_post_disable_helper(struct msm_dp *msm_dp,
struct msm_dp_panel *msm_dp_panel);
+void msm_dp_display_unprepare(struct msm_dp *dp);
+
#endif /* _DP_DISPLAY_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 25/45] drm/msm/dp: move link related operations to dp_display_unprepare()
2024-12-06 4:31 ` [PATCH 25/45] drm/msm/dp: move link related operations to dp_display_unprepare() Abhinav Kumar
@ 2024-12-08 11:48 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:48 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:56PM -0800, Abhinav Kumar wrote:
> Move the link related operations to dp_display_unprepare() and keep
> only stream related operations in dp_display_disable().
Why, no what
>
> Make dp_display_unprepare() available to other clients such as DP MST.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 31 ++++++++++++++++---------------
> drivers/gpu/drm/msm/dp/dp_display.h | 2 ++
> 2 files changed, 18 insertions(+), 15 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index e169cd22db960c0c30707ddbe6a79999dc2a273d..d5b8fd1d4d736ffa7929b9798601dcef0dea5211 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -936,20 +936,8 @@ static int msm_dp_display_disable(struct msm_dp_display_private *dp,
>
> msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, msm_dp_panel);
>
> - /* dongle is still connected but sinks are disconnected */
> - if (dp->link->sink_count == 0)
> - msm_dp_ctrl_psm_config(dp->ctrl);
> -
> msm_dp_ctrl_stream_clk_off(dp->ctrl, msm_dp_panel);
>
> - msm_dp_ctrl_off_link(dp->ctrl);
> -
> - /* re-init the PHY so that we can listen to Dongle disconnect */
> - if (dp->link->sink_count == 0)
> - msm_dp_ctrl_reinit_phy(dp->ctrl);
> - else
> - msm_dp_display_host_phy_exit(dp);
> -
This changes the meaning of msm_dp_display->power_on. I'll have to
review corresponding code carefully. Please note it in the commit
message.
> msm_dp_display->power_on = false;
>
> drm_dbg_dp(dp->drm_dev, "sink count: %d\n", dp->link->sink_count);
> @@ -1678,15 +1666,28 @@ void msm_dp_display_atomic_disable(struct msm_dp *msm_dp)
> msm_dp_display_disable_helper(msm_dp, msm_dp_display->panel);
> }
>
> -static void msm_dp_display_unprepare(struct msm_dp_display_private *msm_dp_display_priv)
> +void msm_dp_display_unprepare(struct msm_dp *msm_dp)
> {
> - struct msm_dp *msm_dp = &msm_dp_display_priv->msm_dp_display;
> + struct msm_dp_display_private *msm_dp_display;
>
> + msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
Set it it at the assignment time, please. Or at least add an empty line
afterwards.
> if (!msm_dp->prepared) {
> drm_dbg_dp(msm_dp->drm_dev, "Link already setup, return\n");
> return;
> }
>
> + /* dongle is still connected but sinks are disconnected */
> + if (msm_dp_display->link->sink_count == 0)
> + msm_dp_ctrl_psm_config(msm_dp_display->ctrl);
> +
> + msm_dp_ctrl_off_link(msm_dp_display->ctrl);
> +
> + /* re-init the PHY so that we can listen to Dongle disconnect */
> + if (msm_dp_display->link->sink_count == 0)
> + msm_dp_ctrl_reinit_phy(msm_dp_display->ctrl);
> + else
> + msm_dp_display_host_phy_exit(msm_dp_display);
> +
> pm_runtime_put_sync(&msm_dp->pdev->dev);
>
> msm_dp->prepared = false;
> @@ -1732,7 +1733,7 @@ void msm_dp_display_atomic_post_disable(struct msm_dp *msm_dp)
>
> msm_dp_display_atomic_post_disable_helper(msm_dp, msm_dp_display->panel);
>
> - msm_dp_display_unprepare(msm_dp_display);
> + msm_dp_display_unprepare(msm_dp);
> }
>
> void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> index 2b23f2bf7535d3fd513d40a8411a1903fcd560b0..82eb1c6ed1467b21742bda8eaae9c51d3207e997 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.h
> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> @@ -65,4 +65,6 @@ void msm_dp_display_mode_set_helper(struct msm_dp *msm_dp,
> void msm_dp_display_atomic_post_disable_helper(struct msm_dp *msm_dp,
> struct msm_dp_panel *msm_dp_panel);
>
> +void msm_dp_display_unprepare(struct msm_dp *dp);
> +
> #endif /* _DP_DISPLAY_H_ */
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 26/45] drm/msm/dp: replace power_on with active_stream_cnt for dp_display
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (24 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 25/45] drm/msm/dp: move link related operations to dp_display_unprepare() Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:50 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 27/45] drm/msm/dp: make the SST bridge disconnected when mst is active Abhinav Kumar
` (20 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
For DP MST, the link clock and power domain resources stay on until
both streams have been disabled OR we receive hotplug. Introduce an
active_stream_cnt to track the number of active streams and necessary
state handling. Replace the power_on variable with active_stream_cnt
as power_on boolean works only for a single stream.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_audio.c | 2 +-
drivers/gpu/drm/msm/dp/dp_display.c | 42 ++++++++++++++++++++++++-------------
drivers/gpu/drm/msm/dp/dp_display.h | 3 ++-
3 files changed, 31 insertions(+), 16 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_audio.c b/drivers/gpu/drm/msm/dp/dp_audio.c
index 74e01a5dd4195d5e0e04250663886f1116f25711..3aad4c6f90382bd5c7cad7bcab38335ef81fc869 100644
--- a/drivers/gpu/drm/msm/dp/dp_audio.c
+++ b/drivers/gpu/drm/msm/dp/dp_audio.c
@@ -441,7 +441,7 @@ int msm_dp_audio_hw_params(struct device *dev,
* such cases check for connection status and bail out if not
* connected.
*/
- if (!msm_dp_display->power_on) {
+ if (!msm_dp_display_get_active_stream_cnt(msm_dp_display)) {
rc = -EINVAL;
goto end;
}
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index d5b8fd1d4d736ffa7929b9798601dcef0dea5211..e69cdfbbe10484e47475ef189849f8ff6628bd7a 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -111,6 +111,8 @@ struct msm_dp_display_private {
bool wide_bus_supported;
+ u32 active_stream_cnt;
+
struct msm_dp_audio *audio;
};
@@ -187,6 +189,15 @@ static const struct of_device_id msm_dp_dt_match[] = {
{}
};
+int msm_dp_display_get_active_stream_cnt(struct msm_dp *msm_dp)
+{
+ struct msm_dp_display_private *msm_dp_display;
+
+ msm_dp_display = container_of(msm_dp, struct msm_dp_display_private, msm_dp_display);
+
+ return msm_dp_display->active_stream_cnt;
+}
+
static struct msm_dp_display_private *dev_get_dp_display_private(struct device *dev)
{
struct msm_dp *dp = dev_get_drvdata(dev);
@@ -662,7 +673,7 @@ static int msm_dp_hpd_unplug_handle(struct msm_dp_display_private *dp, u32 data)
*/
msm_dp_display_notify_disconnect(&dp->msm_dp_display.pdev->dev);
- if (!dp->msm_dp_display.power_on) {
+ if (!dp->active_stream_cnt) {
dp->hpd_state = ST_DISCONNECTED;
} else {
dp->hpd_state = ST_DISCONNECT_PENDING;
@@ -856,7 +867,7 @@ static int msm_dp_display_prepare(struct msm_dp_display_private *dp)
return rc;
}
- if (dp->hpd_state == ST_CONNECTED && !msm_dp_display->power_on) {
+ if (dp->hpd_state == ST_CONNECTED && !dp->active_stream_cnt) {
msm_dp_display_host_phy_init(dp);
force_link_train = true;
}
@@ -872,17 +883,10 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp,
struct msm_dp_panel *msm_dp_panel)
{
int rc = 0;
- struct msm_dp *msm_dp_display = &dp->msm_dp_display;
drm_dbg_dp(dp->drm_dev, "sink_count=%d\n", dp->link->sink_count);
- if (msm_dp_display->power_on) {
- drm_dbg_dp(dp->drm_dev, "Link already setup, return\n");
- return 0;
- }
rc = msm_dp_ctrl_on_stream(dp->ctrl, msm_dp_panel, dp->max_stream);
- if (!rc)
- msm_dp_display->power_on = true;
return rc;
}
@@ -929,16 +933,14 @@ static void msm_dp_display_audio_notify_disable(struct msm_dp_display_private *d
static int msm_dp_display_disable(struct msm_dp_display_private *dp,
struct msm_dp_panel *msm_dp_panel)
{
- struct msm_dp *msm_dp_display = &dp->msm_dp_display;
-
- if (!msm_dp_display->power_on)
+ if (!dp->active_stream_cnt)
return 0;
msm_dp_ctrl_clear_vsc_sdp_pkt(dp->ctrl, msm_dp_panel);
msm_dp_ctrl_stream_clk_off(dp->ctrl, msm_dp_panel);
- msm_dp_display->power_on = false;
+ dp->active_stream_cnt--;
drm_dbg_dp(dp->drm_dev, "sink count: %d\n", dp->link->sink_count);
return 0;
@@ -1096,7 +1098,7 @@ void msm_dp_snapshot(struct msm_disp_state *disp_state, struct msm_dp *dp)
*/
mutex_lock(&msm_dp_display->event_mutex);
- if (!dp->power_on) {
+ if (!msm_dp_display->active_stream_cnt) {
mutex_unlock(&msm_dp_display->event_mutex);
return;
}
@@ -1621,6 +1623,8 @@ void msm_dp_display_enable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_dp
}
}
+ msm_dp_display->active_stream_cnt++;
+
/* completed connection */
msm_dp_display->hpd_state = ST_CONNECTED;
@@ -1645,6 +1649,11 @@ void msm_dp_display_disable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_d
msm_dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
+ if (!msm_dp_display->active_stream_cnt) {
+ drm_dbg_dp(dp->drm_dev, "no active streams\n");
+ return;
+ }
+
if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT)
msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_panel);
else
@@ -1676,6 +1685,11 @@ void msm_dp_display_unprepare(struct msm_dp *msm_dp)
return;
}
+ if (msm_dp_display->active_stream_cnt) {
+ drm_dbg_dp(msm_dp->drm_dev, "stream still active, return\n");
+ return;
+ }
+
/* dongle is still connected but sinks are disconnected */
if (msm_dp_display->link->sink_count == 0)
msm_dp_ctrl_psm_config(msm_dp_display->ctrl);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 82eb1c6ed1467b21742bda8eaae9c51d3207e997..f1142b7b96c372b49e8e45dd0378eb507b081915 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -20,7 +20,6 @@ struct msm_dp {
struct drm_bridge *next_bridge;
bool link_ready;
bool audio_enabled;
- bool power_on;
bool prepared;
bool mst_active;
unsigned int connector_type;
@@ -67,4 +66,6 @@ void msm_dp_display_atomic_post_disable_helper(struct msm_dp *msm_dp,
void msm_dp_display_unprepare(struct msm_dp *dp);
+int msm_dp_display_get_active_stream_cnt(struct msm_dp *msm_dp);
+
#endif /* _DP_DISPLAY_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 26/45] drm/msm/dp: replace power_on with active_stream_cnt for dp_display
2024-12-06 4:31 ` [PATCH 26/45] drm/msm/dp: replace power_on with active_stream_cnt for dp_display Abhinav Kumar
@ 2024-12-08 11:50 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:50 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:57PM -0800, Abhinav Kumar wrote:
> For DP MST, the link clock and power domain resources stay on until
> both streams have been disabled OR we receive hotplug. Introduce an
> active_stream_cnt to track the number of active streams and necessary
> state handling. Replace the power_on variable with active_stream_cnt
> as power_on boolean works only for a single stream.
Okay, this answers one of my previous questions. Swapping these two
patches might be beneficial.
For this patch:
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_audio.c | 2 +-
> drivers/gpu/drm/msm/dp/dp_display.c | 42 ++++++++++++++++++++++++-------------
> drivers/gpu/drm/msm/dp/dp_display.h | 3 ++-
> 3 files changed, 31 insertions(+), 16 deletions(-)
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 27/45] drm/msm/dp: make the SST bridge disconnected when mst is active
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (25 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 26/45] drm/msm/dp: replace power_on with active_stream_cnt for dp_display Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:51 ` Dmitry Baryshkov
2024-12-06 4:31 ` [PATCH 28/45] drm/msm/dp: add an API to initialize MST on sink side Abhinav Kumar
` (19 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_drm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_drm.c b/drivers/gpu/drm/msm/dp/dp_drm.c
index 920392b3c688821bccdc66d50fb7052ac3a9a85a..225858c6240512cf2c5ca3b5eb52cf9b7f4db3e3 100644
--- a/drivers/gpu/drm/msm/dp/dp_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_drm.c
@@ -27,7 +27,7 @@ static enum drm_connector_status msm_dp_bridge_detect(struct drm_bridge *bridge)
drm_dbg_dp(dp->drm_dev, "link_ready = %s\n",
(dp->link_ready) ? "true" : "false");
- return (dp->link_ready) ? connector_status_connected :
+ return (dp->link_ready && !dp->mst_active) ? connector_status_connected :
connector_status_disconnected;
}
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 27/45] drm/msm/dp: make the SST bridge disconnected when mst is active
2024-12-06 4:31 ` [PATCH 27/45] drm/msm/dp: make the SST bridge disconnected when mst is active Abhinav Kumar
@ 2024-12-08 11:51 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:51 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:58PM -0800, Abhinav Kumar wrote:
missing commit message
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_drm.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_drm.c b/drivers/gpu/drm/msm/dp/dp_drm.c
> index 920392b3c688821bccdc66d50fb7052ac3a9a85a..225858c6240512cf2c5ca3b5eb52cf9b7f4db3e3 100644
> --- a/drivers/gpu/drm/msm/dp/dp_drm.c
> +++ b/drivers/gpu/drm/msm/dp/dp_drm.c
> @@ -27,7 +27,7 @@ static enum drm_connector_status msm_dp_bridge_detect(struct drm_bridge *bridge)
> drm_dbg_dp(dp->drm_dev, "link_ready = %s\n",
> (dp->link_ready) ? "true" : "false");
>
> - return (dp->link_ready) ? connector_status_connected :
> + return (dp->link_ready && !dp->mst_active) ? connector_status_connected :
> connector_status_disconnected;
> }
>
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 28/45] drm/msm/dp: add an API to initialize MST on sink side
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (26 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 27/45] drm/msm/dp: make the SST bridge disconnected when mst is active Abhinav Kumar
@ 2024-12-06 4:31 ` Abhinav Kumar
2024-12-08 11:56 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 29/45] drm/msm/dp: skip reading the EDID for MST cases Abhinav Kumar
` (18 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:31 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
If the DP controller is capable of supporting multiple streams
then initialize the DP sink in MST mode by programming the DP_MSTM_CTRL
DPCD register to enable MST mode.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 32 ++++++++++++++++++++++++++++++++
1 file changed, 32 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index e69cdfbbe10484e47475ef189849f8ff6628bd7a..033d238e956263c1212fce45aab01316ef341edb 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -380,6 +380,35 @@ static int msm_dp_display_send_hpd_notification(struct msm_dp_display_private *d
return 0;
}
+static void msm_dp_display_mst_init(struct msm_dp_display_private *dp)
+{
+ const unsigned long clear_mstm_ctrl_timeout_us = 100000;
+ u8 old_mstm_ctrl;
+ struct msm_dp *msm_dp = &dp->msm_dp_display;
+ int ret;
+
+ /* clear sink mst state */
+ drm_dp_dpcd_readb(dp->aux, DP_MSTM_CTRL, &old_mstm_ctrl);
+ drm_dp_dpcd_writeb(dp->aux, DP_MSTM_CTRL, 0);
+
+ /* add extra delay if MST state is not cleared */
+ if (old_mstm_ctrl) {
+ drm_dbg_dp(dp->drm_dev, "MSTM_CTRL is not cleared, wait %luus\n",
+ clear_mstm_ctrl_timeout_us);
+ usleep_range(clear_mstm_ctrl_timeout_us,
+ clear_mstm_ctrl_timeout_us + 1000);
+ }
+
+ ret = drm_dp_dpcd_writeb(dp->aux, DP_MSTM_CTRL,
+ DP_MST_EN | DP_UP_REQ_EN | DP_UPSTREAM_IS_SRC);
+ if (ret < 0) {
+ DRM_ERROR("sink mst enablement failed\n");
+ return;
+ }
+
+ msm_dp->mst_active = true;
+}
+
static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
{
struct drm_connector *connector = dp->msm_dp_display.connector;
@@ -414,6 +443,9 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
*/
msm_dp_link_psm_config(dp->link, &dp->panel->link_info, false);
+ if (dp->max_stream > DEFAULT_STREAM_COUNT && msm_dp_panel_read_mst_cap(dp->panel))
+ msm_dp_display_mst_init(dp);
+
msm_dp_link_reset_phy_params_vx_px(dp->link);
rc = msm_dp_ctrl_on_link(dp->ctrl, dp_display->mst_active);
if (rc) {
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 28/45] drm/msm/dp: add an API to initialize MST on sink side
2024-12-06 4:31 ` [PATCH 28/45] drm/msm/dp: add an API to initialize MST on sink side Abhinav Kumar
@ 2024-12-08 11:56 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:56 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:31:59PM -0800, Abhinav Kumar wrote:
> If the DP controller is capable of supporting multiple streams
> then initialize the DP sink in MST mode by programming the DP_MSTM_CTRL
> DPCD register to enable MST mode.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 32 ++++++++++++++++++++++++++++++++
> 1 file changed, 32 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index e69cdfbbe10484e47475ef189849f8ff6628bd7a..033d238e956263c1212fce45aab01316ef341edb 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -380,6 +380,35 @@ static int msm_dp_display_send_hpd_notification(struct msm_dp_display_private *d
> return 0;
> }
>
> +static void msm_dp_display_mst_init(struct msm_dp_display_private *dp)
> +{
> + const unsigned long clear_mstm_ctrl_timeout_us = 100000;
> + u8 old_mstm_ctrl;
> + struct msm_dp *msm_dp = &dp->msm_dp_display;
> + int ret;
> +
> + /* clear sink mst state */
> + drm_dp_dpcd_readb(dp->aux, DP_MSTM_CTRL, &old_mstm_ctrl);
> + drm_dp_dpcd_writeb(dp->aux, DP_MSTM_CTRL, 0);
> +
> + /* add extra delay if MST state is not cleared */
> + if (old_mstm_ctrl) {
> + drm_dbg_dp(dp->drm_dev, "MSTM_CTRL is not cleared, wait %luus\n",
> + clear_mstm_ctrl_timeout_us);
"is not cleared" usually means that we wrote 0 to it, but the value
still stays on. Please rephrase the message, drop wait time.
> + usleep_range(clear_mstm_ctrl_timeout_us,
> + clear_mstm_ctrl_timeout_us + 1000);
> + }
> +
> + ret = drm_dp_dpcd_writeb(dp->aux, DP_MSTM_CTRL,
> + DP_MST_EN | DP_UP_REQ_EN | DP_UPSTREAM_IS_SRC);
> + if (ret < 0) {
if (ret != 1) ?
> + DRM_ERROR("sink mst enablement failed\n");
> + return;
> + }
> +
> + msm_dp->mst_active = true;
> +}
> +
> static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> {
> struct drm_connector *connector = dp->msm_dp_display.connector;
> @@ -414,6 +443,9 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> */
> msm_dp_link_psm_config(dp->link, &dp->panel->link_info, false);
>
> + if (dp->max_stream > DEFAULT_STREAM_COUNT && msm_dp_panel_read_mst_cap(dp->panel))
> + msm_dp_display_mst_init(dp);
> +
> msm_dp_link_reset_phy_params_vx_px(dp->link);
> rc = msm_dp_ctrl_on_link(dp->ctrl, dp_display->mst_active);
> if (rc) {
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 29/45] drm/msm/dp: skip reading the EDID for MST cases
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (27 preceding siblings ...)
2024-12-06 4:31 ` [PATCH 28/45] drm/msm/dp: add an API to initialize MST on sink side Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-06 9:32 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 30/45] drm/msm/dp: add dp_display_get_panel() to initialize DP panel Abhinav Kumar
` (17 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
For MST cases, EDID is handled through AUX sideband messaging.
Skip the EDID read during hotplug handle for MST cases.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 033d238e956263c1212fce45aab01316ef341edb..a67bc7c1b83a5a9996435804ff7337f72dae93a0 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -420,9 +420,11 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
if (rc)
goto end;
- rc = msm_dp_panel_read_edid(dp->panel, connector);
- if (rc)
- goto end;
+ if (dp->max_stream <= DEFAULT_STREAM_COUNT || !msm_dp_panel_read_mst_cap(dp->panel)) {
+ rc = msm_dp_panel_read_edid(dp->panel, connector);
+ if (rc)
+ goto end;
+ }
msm_dp_link_process_request(dp->link);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 29/45] drm/msm/dp: skip reading the EDID for MST cases
2024-12-06 4:32 ` [PATCH 29/45] drm/msm/dp: skip reading the EDID for MST cases Abhinav Kumar
@ 2024-12-06 9:32 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:32 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:00PM -0800, Abhinav Kumar wrote:
> For MST cases, EDID is handled through AUX sideband messaging.
> Skip the EDID read during hotplug handle for MST cases.
But why? Isn't EDID being read at the hotplug time to update
drm_connector's data?
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 033d238e956263c1212fce45aab01316ef341edb..a67bc7c1b83a5a9996435804ff7337f72dae93a0 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -420,9 +420,11 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> if (rc)
> goto end;
>
> - rc = msm_dp_panel_read_edid(dp->panel, connector);
> - if (rc)
> - goto end;
> + if (dp->max_stream <= DEFAULT_STREAM_COUNT || !msm_dp_panel_read_mst_cap(dp->panel)) {
> + rc = msm_dp_panel_read_edid(dp->panel, connector);
> + if (rc)
> + goto end;
> + }
>
> msm_dp_link_process_request(dp->link);
>
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 30/45] drm/msm/dp: add dp_display_get_panel() to initialize DP panel
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (28 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 29/45] drm/msm/dp: skip reading the EDID for MST cases Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 8:51 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations Abhinav Kumar
` (16 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Add an API dp_display_get_panel() to initialize and return a DP
panel to be used by DP MST module. Since some of the fields of
DP panel are private, dp_display module needs to initialize these
parts and return the panel back.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 28 ++++++++++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_display.h | 2 ++
2 files changed, 30 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index a67bc7c1b83a5a9996435804ff7337f72dae93a0..60316633cc6bcfdfad8ab356d803642d353add61 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -753,6 +753,34 @@ static int msm_dp_irq_hpd_handle(struct msm_dp_display_private *dp, u32 data)
return 0;
}
+struct msm_dp_panel *msm_dp_display_get_panel(struct msm_dp *dp_display)
+{
+ struct msm_dp_display_private *dp;
+ struct msm_dp_panel *dp_panel;
+
+ struct msm_dp_panel_in panel_in;
+
+ dp = container_of(dp_display, struct msm_dp_display_private, msm_dp_display);
+
+ panel_in.dev = &dp_display->pdev->dev;
+ panel_in.aux = dp->aux;
+ panel_in.catalog = dp->catalog;
+ panel_in.link = dp->link;
+
+ dp_panel = msm_dp_panel_get(&panel_in);
+
+ if (IS_ERR(dp->panel)) {
+ DRM_ERROR("failed to initialize panel\n");
+ return NULL;
+ }
+
+ memcpy(dp_panel->dpcd, dp->panel->dpcd, DP_RECEIVER_CAP_SIZE + 1);
+ memcpy(&dp_panel->link_info, &dp->panel->link_info,
+ sizeof(dp->panel->link_info));
+
+ return dp_panel;
+}
+
static void msm_dp_display_deinit_sub_modules(struct msm_dp_display_private *dp)
{
msm_dp_audio_put(dp->audio);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index f1142b7b96c372b49e8e45dd0378eb507b081915..6ab14e969bce0fd07b3a550bae17e99652479232 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -68,4 +68,6 @@ void msm_dp_display_unprepare(struct msm_dp *dp);
int msm_dp_display_get_active_stream_cnt(struct msm_dp *msm_dp);
+struct msm_dp_panel *msm_dp_display_get_panel(struct msm_dp *dp_display);
+
#endif /* _DP_DISPLAY_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 30/45] drm/msm/dp: add dp_display_get_panel() to initialize DP panel
2024-12-06 4:32 ` [PATCH 30/45] drm/msm/dp: add dp_display_get_panel() to initialize DP panel Abhinav Kumar
@ 2024-12-08 8:51 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 8:51 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:01PM -0800, Abhinav Kumar wrote:
> Add an API dp_display_get_panel() to initialize and return a DP
> panel to be used by DP MST module. Since some of the fields of
> DP panel are private, dp_display module needs to initialize these
> parts and return the panel back.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 28 ++++++++++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_display.h | 2 ++
> 2 files changed, 30 insertions(+)
>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (29 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 30/45] drm/msm/dp: add dp_display_get_panel() to initialize DP panel Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-06 10:12 ` Dmitry Baryshkov
2024-12-08 7:19 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 32/45] drm/msm/dp: add connector abstraction for DP MST Abhinav Kumar
` (15 subsequent siblings)
46 siblings, 2 replies; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Add a new file dp_mst_drm to manage the DP MST bridge operations
similar to the dp_drm file which manages the SST bridge operations.
Each MST encoder creates one bridge and each bridge is bound to its
own dp_panel abstraction to manage the operations of its pipeline.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/Makefile | 3 +-
drivers/gpu/drm/msm/dp/dp_display.h | 2 +
drivers/gpu/drm/msm/dp/dp_mst_drm.c | 490 ++++++++++++++++++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_mst_drm.h | 102 ++++++++
4 files changed, 596 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile
index f274d9430cc311405f890074c1466ffe2ec45ac9..b1e01b3123d9afc4818f059c5d4e7ca70dca3754 100644
--- a/drivers/gpu/drm/msm/Makefile
+++ b/drivers/gpu/drm/msm/Makefile
@@ -142,7 +142,8 @@ msm-display-$(CONFIG_DRM_MSM_DP)+= dp/dp_aux.o \
dp/dp_link.o \
dp/dp_panel.o \
dp/dp_audio.o \
- dp/dp_utils.o
+ dp/dp_utils.o \
+ dp/dp_mst_drm.o
msm-display-$(CONFIG_DRM_MSM_HDMI_HDCP) += hdmi/hdmi_hdcp.o
diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
index 6ab14e969bce0fd07b3a550bae17e99652479232..a5d4893f689c6afbbe622c9b7dfa98d23d754831 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.h
+++ b/drivers/gpu/drm/msm/dp/dp_display.h
@@ -7,6 +7,7 @@
#define _DP_DISPLAY_H_
#include "dp_panel.h"
+#include "dp_mst_drm.h"
#include <sound/hdmi-codec.h>
#include "disp/msm_disp_snapshot.h"
@@ -26,6 +27,7 @@ struct msm_dp {
bool is_edp;
bool internal_hpd;
+ struct msm_dp_mst *msm_dp_mst;
hdmi_codec_plugged_cb plugged_cb;
struct msm_dp_audio *msm_dp_audio;
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
new file mode 100644
index 0000000000000000000000000000000000000000..e66bd1e565aeb4da3d636eb5f4aa75504d60fd40
--- /dev/null
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
@@ -0,0 +1,490 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2022-2024 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2018-2021, The Linux Foundation. All rights reserved.
+ */
+
+/*
+ * Copyright © 2014 Red Hat.
+ *
+ * Permission to use, copy, modify, distribute, and sell this software and its
+ * documentation for any purpose is hereby granted without fee, provided that
+ * the above copyright notice appear in all copies and that both that copyright
+ * notice and this permission notice appear in supporting documentation, and
+ * that the name of the copyright holders not be used in advertising or
+ * publicity pertaining to distribution of the software without specific,
+ * written prior permission. The copyright holders make no representations
+ * about the suitability of this software for any purpose. It is provided "as
+ * is" without express or implied warranty.
+ *
+ * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
+ * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
+ * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
+ * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
+ * OF THIS SOFTWARE.
+ */
+
+#include "dp_mst_drm.h"
+
+static struct drm_private_state *msm_dp_mst_duplicate_bridge_state(struct drm_private_obj *obj)
+{
+ struct msm_dp_mst_bridge_state *state;
+
+ state = kmemdup(obj->state, sizeof(*state), GFP_KERNEL);
+ if (!state)
+ return NULL;
+
+ __drm_atomic_helper_private_obj_duplicate_state(obj, &state->base);
+
+ return &state->base;
+}
+
+static void msm_dp_mst_destroy_bridge_state(struct drm_private_obj *obj,
+ struct drm_private_state *state)
+{
+ struct msm_dp_mst_bridge_state *priv_state =
+ to_msm_dp_mst_bridge_priv_state(state);
+
+ kfree(priv_state);
+}
+
+static const struct drm_private_state_funcs msm_dp_mst_bridge_state_funcs = {
+ .atomic_duplicate_state = msm_dp_mst_duplicate_bridge_state,
+ .atomic_destroy_state = msm_dp_mst_destroy_bridge_state,
+};
+
+/**
+ * dp_mst_find_vcpi_slots() - Find VCPI slots for this PBN value
+ * @mgr: manager to use
+ * @pbn: payload bandwidth to convert into slots.
+ *
+ * Calculate the number of VCPI slots that will be required for the given PBN
+ * value.
+ *
+ * RETURNS:
+ * The total slots required for this port, or error.
+ */
+static int msm_dp_mst_find_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr, int pbn)
+{
+ int num_slots;
+ struct drm_dp_mst_topology_state *state;
+
+ state = to_drm_dp_mst_topology_state(mgr->base.state);
+ num_slots = DIV_ROUND_UP(pbn, dfixed_trunc(state->pbn_div));
+
+ /* max. time slots - one slot for MTP header */
+ if (num_slots > 63)
+ return -ENOSPC;
+ return num_slots;
+}
+
+static void _msm_dp_mst_update_timeslots(struct msm_dp_mst *mst,
+ struct msm_dp_mst_bridge *mst_bridge,
+ struct drm_dp_mst_port *port)
+{
+ int i;
+ struct msm_dp_mst_bridge *msm_dp_bridge;
+ struct drm_dp_mst_topology_state *mst_state;
+ struct drm_dp_mst_atomic_payload *payload;
+ int prev_start = 0;
+ int prev_slots = 0;
+
+ mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
+ payload = drm_atomic_get_mst_payload_state(mst_state, port);
+
+ if (!payload) {
+ DRM_ERROR("mst bridge [%d] update_timeslots failed, null payload\n",
+ mst_bridge->id);
+ return;
+ }
+
+ for (i = 0; i < mst->max_streams; i++) {
+ msm_dp_bridge = &mst->mst_bridge[i];
+ if (mst_bridge == msm_dp_bridge) {
+ /*
+ * When a payload was removed make sure to move any payloads after it
+ * to the left so all payloads are aligned to the left.
+ */
+ if (payload->vc_start_slot < 0) {
+ // cache the payload
+ prev_start = msm_dp_bridge->start_slot;
+ prev_slots = msm_dp_bridge->num_slots;
+ msm_dp_bridge->pbn = 0;
+ msm_dp_bridge->start_slot = 1;
+ msm_dp_bridge->num_slots = 0;
+ msm_dp_bridge->vcpi = 0;
+ } else { //add payload
+ msm_dp_bridge->pbn = payload->pbn;
+ msm_dp_bridge->start_slot = payload->vc_start_slot;
+ msm_dp_bridge->num_slots = payload->time_slots;
+ msm_dp_bridge->vcpi = payload->vcpi;
+ }
+ }
+ }
+
+ // Now commit all the updated payloads
+ for (i = 0; i < mst->max_streams; i++) {
+ msm_dp_bridge = &mst->mst_bridge[i];
+
+ //Shift payloads to the left if there was a removed payload.
+ if (payload->vc_start_slot < 0 && msm_dp_bridge->start_slot > prev_start)
+ msm_dp_bridge->start_slot -= prev_slots;
+
+ msm_dp_display_set_stream_info(mst->msm_dp, msm_dp_bridge->msm_dp_panel,
+ msm_dp_bridge->id, msm_dp_bridge->start_slot,
+ msm_dp_bridge->num_slots,
+ msm_dp_bridge->pbn, msm_dp_bridge->vcpi);
+ drm_dbg_dp(mst->msm_dp->drm_dev,
+ "conn:%d vcpi:%d start_slot:%d num_slots:%d, pbn:%d\n",
+ DP_MST_CONN_ID(msm_dp_bridge), msm_dp_bridge->vcpi,
+ msm_dp_bridge->start_slot,
+ msm_dp_bridge->num_slots, msm_dp_bridge->pbn);
+ }
+}
+
+static int _msm_dp_mst_bridge_pre_enable_part1(struct msm_dp_mst_bridge *dp_bridge,
+ struct drm_bridge_state *bridge_state)
+{
+ struct msm_dp *msm_dp = dp_bridge->display;
+ struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
+ struct drm_dp_mst_port *port = mst_conn->mst_port;
+ struct drm_dp_mst_topology_state *mst_state;
+ struct drm_dp_mst_atomic_payload *payload;
+ struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
+ int pbn, slots;
+ int rc = 0;
+
+ mst_state = drm_atomic_get_new_mst_topology_state(bridge_state->base.state,
+ &mst->mst_mgr);
+
+ pbn = drm_dp_calc_pbn_mode(dp_panel->msm_dp_mode.drm_mode.clock,
+ (mst_conn->connector.display_info.bpc * 3) << 4);
+
+ slots = msm_dp_mst_find_vcpi_slots(&mst->mst_mgr, pbn);
+
+ drm_dbg_dp(msm_dp->drm_dev, "conn:%d pbn:%d, slots:%d\n", DP_MST_CONN_ID(dp_bridge),
+ pbn, slots);
+
+ payload = drm_atomic_get_mst_payload_state(mst_state, port);
+ if (!payload || payload->time_slots <= 0) {
+ DRM_ERROR("time slots not allocated for conn:%d\n", DP_MST_CONN_ID(dp_bridge));
+ rc = -EINVAL;
+ return rc;
+ }
+
+ drm_dp_mst_update_slots(mst_state, DP_CAP_ANSI_8B10B);
+
+ rc = drm_dp_add_payload_part1(&mst->mst_mgr, mst_state, payload);
+ if (rc) {
+ DRM_ERROR("payload allocation failure for conn:%d\n", DP_MST_CONN_ID(dp_bridge));
+ return rc;
+ }
+
+ _msm_dp_mst_update_timeslots(mst, dp_bridge, port);
+
+ return rc;
+}
+
+static void _msm_dp_mst_bridge_pre_enable_part2(struct msm_dp_mst_bridge *dp_bridge,
+ struct drm_bridge_state *bridge_state)
+{
+ struct msm_dp *msm_dp = dp_bridge->display;
+ struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
+ struct drm_dp_mst_port *port = mst_conn->mst_port;
+ struct drm_dp_mst_topology_state *mst_state;
+ struct drm_dp_mst_atomic_payload *payload;
+
+ drm_dp_check_act_status(&mst->mst_mgr);
+
+ mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
+ payload = drm_atomic_get_mst_payload_state(mst_state, port);
+
+ if (!payload) {
+ DRM_ERROR("mst bridge [%d] null payload\n", dp_bridge->id);
+ return;
+ }
+
+ if (!payload->port) {
+ DRM_ERROR("mst bridge [%d] null port\n", dp_bridge->id);
+ return;
+ }
+
+ if (!payload->port->connector) {
+ DRM_ERROR("mst bridge [%d] part-2 failed, null connector\n",
+ dp_bridge->id);
+ return;
+ }
+
+ if (payload->vc_start_slot == -1) {
+ DRM_ERROR("mst bridge [%d] part-2 failed, payload alloc part 1 failed\n",
+ dp_bridge->id);
+ return;
+ }
+
+ drm_dp_add_payload_part2(&mst->mst_mgr, payload);
+
+ drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre enable part-2 complete\n",
+ dp_bridge->id);
+}
+
+static void _msm_dp_mst_bridge_pre_disable_part1(struct msm_dp_mst_bridge *dp_bridge,
+ struct drm_bridge_state *bridge_state)
+{
+ struct msm_dp *msm_dp = dp_bridge->display;
+ struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
+ struct drm_dp_mst_port *port = mst_conn->mst_port;
+ struct drm_dp_mst_topology_state *old_mst_state;
+ struct drm_dp_mst_topology_state *new_mst_state;
+ const struct drm_dp_mst_atomic_payload *old_payload;
+ struct drm_dp_mst_atomic_payload *new_payload;
+
+ old_mst_state = drm_atomic_get_old_mst_topology_state(bridge_state->base.state,
+ &mst->mst_mgr);
+
+ new_mst_state = drm_atomic_get_new_mst_topology_state(bridge_state->base.state,
+ &mst->mst_mgr);
+
+ old_payload = drm_atomic_get_mst_payload_state(old_mst_state, port);
+ new_payload = drm_atomic_get_mst_payload_state(new_mst_state, port);
+
+ if (!old_payload || !new_payload) {
+ DRM_ERROR("mst bridge [%d] _pre disable part-1 failed, null payload\n",
+ dp_bridge->id);
+ return;
+ }
+
+ drm_dp_remove_payload_part1(&mst->mst_mgr, new_mst_state, new_payload);
+ drm_dp_remove_payload_part2(&mst->mst_mgr, new_mst_state, old_payload, new_payload);
+
+ _msm_dp_mst_update_timeslots(mst, dp_bridge, port);
+
+ drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre disable part-1 complete\n",
+ dp_bridge->id);
+}
+
+static void _msm_dp_mst_bridge_pre_disable_part2(struct msm_dp_mst_bridge *dp_bridge)
+{
+ struct msm_dp *msm_dp = dp_bridge->display;
+ struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
+
+ drm_dp_check_act_status(&mst->mst_mgr);
+
+ drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre disable part-2 complete\n",
+ dp_bridge->id);
+}
+
+static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
+ struct drm_bridge_state *old_bridge_state)
+{
+ int rc = 0;
+ struct msm_dp_mst_bridge *bridge;
+ struct msm_dp *dp;
+ struct msm_dp_mst_bridge_state *msm_dp_bridge_state;
+
+ if (!drm_bridge) {
+ DRM_ERROR("Invalid params\n");
+ return;
+ }
+
+ bridge = to_msm_dp_mst_bridge(drm_bridge);
+ msm_dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
+ dp = bridge->display;
+
+ /* to cover cases of bridge_disable/bridge_enable without modeset */
+ bridge->connector = msm_dp_bridge_state->connector;
+ bridge->msm_dp_panel = msm_dp_bridge_state->msm_dp_panel;
+
+ if (!bridge->connector) {
+ DRM_ERROR("Invalid connector\n");
+ return;
+ }
+
+ msm_dp_display_atomic_prepare(dp);
+
+ rc = _msm_dp_mst_bridge_pre_enable_part1(bridge, old_bridge_state);
+ if (rc) {
+ DRM_ERROR("[%d] DP display pre-enable failed, rc=%d\n", bridge->id, rc);
+ msm_dp_display_unprepare(dp);
+ return;
+ }
+
+ msm_dp_display_enable_helper(dp, bridge->msm_dp_panel);
+
+ _msm_dp_mst_bridge_pre_enable_part2(bridge, old_bridge_state);
+
+ drm_dbg_dp(dp->drm_dev, "conn:%d mode:%s fps:%d vcpi:%d slots:%d to %d\n",
+ DP_MST_CONN_ID(bridge), bridge->drm_mode.name,
+ drm_mode_vrefresh(&bridge->drm_mode),
+ bridge->vcpi, bridge->start_slot,
+ bridge->start_slot + bridge->num_slots);
+}
+
+static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
+ struct drm_bridge_state *old_bridge_state)
+{
+ struct msm_dp_mst_bridge *bridge;
+ struct msm_dp *dp;
+
+ if (!drm_bridge) {
+ DRM_ERROR("Invalid params\n");
+ return;
+ }
+
+ bridge = to_msm_dp_mst_bridge(drm_bridge);
+ if (!bridge->connector) {
+ DRM_ERROR("Invalid connector\n");
+ return;
+ }
+
+ dp = bridge->display;
+
+ _msm_dp_mst_bridge_pre_disable_part1(bridge, old_bridge_state);
+
+ msm_dp_display_disable_helper(dp, bridge->msm_dp_panel);
+
+ _msm_dp_mst_bridge_pre_disable_part2(bridge);
+
+ drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d disable complete\n", bridge->id,
+ DP_MST_CONN_ID(bridge));
+}
+
+static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
+ struct drm_bridge_state *old_bridge_state)
+{
+ int conn = 0;
+ struct msm_dp_mst_bridge *bridge;
+ struct msm_dp *dp;
+
+ if (!drm_bridge) {
+ DRM_ERROR("Invalid params\n");
+ return;
+ }
+
+ bridge = to_msm_dp_mst_bridge(drm_bridge);
+ if (!bridge->connector) {
+ DRM_ERROR("Invalid connector\n");
+ return;
+ }
+
+ conn = DP_MST_CONN_ID(bridge);
+
+ dp = bridge->display;
+
+ msm_dp_display_atomic_post_disable_helper(dp, bridge->msm_dp_panel);
+
+ if (!dp->mst_active)
+ msm_dp_display_unprepare(dp);
+
+ bridge->connector = NULL;
+ bridge->msm_dp_panel = NULL;
+
+ drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d post disable complete\n",
+ bridge->id, conn);
+}
+
+static void msm_dp_mst_bridge_mode_set(struct drm_bridge *drm_bridge,
+ const struct drm_display_mode *mode,
+ const struct drm_display_mode *adjusted_mode)
+{
+ struct msm_dp_mst_bridge *bridge;
+ struct msm_dp_mst_bridge_state *dp_bridge_state;
+ struct msm_dp *dp;
+ struct msm_dp_panel *msm_dp_panel;
+
+ if (!drm_bridge || !mode || !adjusted_mode) {
+ DRM_ERROR("Invalid params\n");
+ return;
+ }
+
+ bridge = to_msm_dp_mst_bridge(drm_bridge);
+
+ dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
+ bridge->connector = dp_bridge_state->connector;
+ bridge->msm_dp_panel = dp_bridge_state->msm_dp_panel;
+
+ msm_dp_panel = bridge->msm_dp_panel;
+ dp = bridge->display;
+
+ memset(&bridge->msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
+ memcpy(&bridge->drm_mode, adjusted_mode, sizeof(bridge->drm_mode));
+ msm_dp_display_mode_set_helper(dp, mode, adjusted_mode, bridge->msm_dp_panel);
+ msm_dp_panel->mst_caps.pbn = drm_dp_calc_pbn_mode(msm_dp_panel->msm_dp_mode.drm_mode.clock,
+ (msm_dp_panel->msm_dp_mode.bpp << 4));
+ memcpy(&bridge->msm_dp_mode, &bridge->msm_dp_panel->msm_dp_mode,
+ sizeof(struct msm_dp_display_mode));
+ drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d mode set complete %s\n", bridge->id,
+ DP_MST_CONN_ID(bridge), mode->name);
+}
+
+/* DP MST Bridge APIs */
+static const struct drm_bridge_funcs msm_dp_mst_bridge_ops = {
+ .atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
+ .atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
+ .atomic_reset = drm_atomic_helper_bridge_reset,
+ .atomic_pre_enable = msm_dp_mst_bridge_atomic_pre_enable,
+ .atomic_disable = msm_dp_mst_bridge_atomic_disable,
+ .atomic_post_disable = msm_dp_mst_bridge_atomic_post_disable,
+ .mode_set = msm_dp_mst_bridge_mode_set,
+};
+
+int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder)
+{
+ int rc = 0;
+ struct msm_dp_mst_bridge *bridge = NULL;
+ struct msm_dp_mst_bridge_state *state;
+ struct drm_device *dev;
+ struct msm_dp_mst *mst = dp->msm_dp_mst;
+ int i;
+
+ for (i = 0; i < mst->max_streams; i++) {
+ if (!mst->mst_bridge[i].in_use) {
+ bridge = &mst->mst_bridge[i];
+ bridge->encoder = encoder;
+ bridge->in_use = true;
+ bridge->id = i;
+ break;
+ }
+ }
+
+ if (i == mst->max_streams) {
+ DRM_ERROR("mst supports only %d bridges\n", i);
+ rc = -EACCES;
+ goto end;
+ }
+
+ dev = dp->drm_dev;
+ bridge->display = dp;
+ bridge->base.funcs = &msm_dp_mst_bridge_ops;
+ bridge->base.encoder = encoder;
+ bridge->base.type = dp->connector_type;
+ bridge->base.ops = DRM_BRIDGE_OP_MODES;
+ drm_bridge_add(&bridge->base);
+
+ rc = drm_bridge_attach(encoder, &bridge->base, NULL, 0);
+ if (rc) {
+ DRM_ERROR("failed to attach bridge, rc=%d\n", rc);
+ goto end;
+ }
+
+ state = kzalloc(sizeof(*state), GFP_KERNEL);
+ if (!state) {
+ rc = -ENOMEM;
+ goto end;
+ }
+
+ drm_atomic_private_obj_init(dev, &bridge->obj,
+ &state->base,
+ &msm_dp_mst_bridge_state_funcs);
+
+ drm_dbg_dp(dp->drm_dev, "mst drm bridge init. bridge id:%d\n", i);
+
+ return 0;
+
+end:
+ return rc;
+}
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
new file mode 100644
index 0000000000000000000000000000000000000000..c2a82cd3c6d6e1951a8e5905d3aa39dfc691023b
--- /dev/null
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
@@ -0,0 +1,102 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2024 Qualcomm Innovation Center, Inc. All rights reserved.
+ *
+ * Permission to use, copy, modify, distribute, and sell this software and its
+ * documentation for any purpose is hereby granted without fee, provided that
+ * the above copyright notice appear in all copies and that both that copyright
+ * notice and this permission notice appear in supporting documentation, and
+ * that the name of the copyright holders not be used in advertising or
+ * publicity pertaining to distribution of the software without specific,
+ * written prior permission. The copyright holders make no representations
+ * about the suitability of this software for any purpose. It is provided "as
+ * is" without express or implied warranty.
+ *
+ * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
+ * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
+ * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
+ * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
+ * OF THIS SOFTWARE.
+ */
+
+#ifndef _DP_MST_DRM_H_
+#define _DP_MST_DRM_H_
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/version.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_bridge.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_fixed.h>
+#include <drm/drm_connector.h>
+#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_dp_mst_helper.h>
+
+#include "dp_panel.h"
+#include "dp_display.h"
+
+#define DP_MST_CONN_ID(bridge) ((bridge)->connector ? \
+ (bridge)->connector->base.id : 0)
+
+struct msm_dp_mst_bridge {
+ struct drm_bridge base;
+ struct drm_private_obj obj;
+ u32 id;
+
+ bool in_use;
+
+ struct msm_dp *display;
+ struct drm_encoder *encoder;
+
+ struct drm_display_mode drm_mode;
+ struct msm_dp_display_mode msm_dp_mode;
+ struct drm_connector *connector;
+ struct msm_dp_panel *msm_dp_panel;
+
+ int vcpi;
+ int pbn;
+ int num_slots;
+ int start_slot;
+};
+
+struct msm_dp_mst_bridge_state {
+ struct drm_private_state base;
+ struct drm_connector *connector;
+ struct msm_dp_panel *msm_dp_panel;
+ int num_slots;
+};
+
+struct msm_dp_mst {
+ bool mst_initialized;
+ struct drm_dp_mst_topology_mgr mst_mgr;
+ struct msm_dp_mst_bridge *mst_bridge;
+ struct msm_dp *msm_dp;
+ bool mst_session_hpd_state;
+ u32 max_streams;
+};
+
+struct msm_dp_mst_connector {
+ struct drm_connector connector;
+ struct drm_dp_mst_port *mst_port;
+ struct msm_dp *msm_dp;
+ struct msm_dp_panel *dp_panel;
+};
+
+#define to_msm_dp_mst_bridge(x) container_of((x), struct msm_dp_mst_bridge, base)
+#define to_msm_dp_mst_bridge_priv(x) \
+ container_of((x), struct msm_dp_mst_bridge, obj)
+#define to_msm_dp_mst_bridge_priv_state(x) \
+ container_of((x), struct msm_dp_mst_bridge_state, base)
+#define to_msm_dp_mst_bridge_state(x) \
+ to_msm_dp_mst_bridge_priv_state((x)->obj.state)
+#define to_msm_dp_mst_connector(x) \
+ container_of((x), struct msm_dp_mst_connector, connector)
+int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
+
+#endif /* _DP_MST_DRM_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations
2024-12-06 4:32 ` [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations Abhinav Kumar
@ 2024-12-06 10:12 ` Dmitry Baryshkov
2025-05-27 10:29 ` Yongxing Mou
2024-12-08 7:19 ` Dmitry Baryshkov
1 sibling, 1 reply; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 10:12 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:02PM -0800, Abhinav Kumar wrote:
> Add a new file dp_mst_drm to manage the DP MST bridge operations
> similar to the dp_drm file which manages the SST bridge operations.
> Each MST encoder creates one bridge and each bridge is bound to its
> own dp_panel abstraction to manage the operations of its pipeline.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/Makefile | 3 +-
> drivers/gpu/drm/msm/dp/dp_display.h | 2 +
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 490 ++++++++++++++++++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 102 ++++++++
> 4 files changed, 596 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile
> index f274d9430cc311405f890074c1466ffe2ec45ac9..b1e01b3123d9afc4818f059c5d4e7ca70dca3754 100644
> --- a/drivers/gpu/drm/msm/Makefile
> +++ b/drivers/gpu/drm/msm/Makefile
> @@ -142,7 +142,8 @@ msm-display-$(CONFIG_DRM_MSM_DP)+= dp/dp_aux.o \
> dp/dp_link.o \
> dp/dp_panel.o \
> dp/dp_audio.o \
> - dp/dp_utils.o
> + dp/dp_utils.o \
> + dp/dp_mst_drm.o
>
> msm-display-$(CONFIG_DRM_MSM_HDMI_HDCP) += hdmi/hdmi_hdcp.o
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> index 6ab14e969bce0fd07b3a550bae17e99652479232..a5d4893f689c6afbbe622c9b7dfa98d23d754831 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.h
> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> @@ -7,6 +7,7 @@
> #define _DP_DISPLAY_H_
>
> #include "dp_panel.h"
> +#include "dp_mst_drm.h"
> #include <sound/hdmi-codec.h>
> #include "disp/msm_disp_snapshot.h"
>
> @@ -26,6 +27,7 @@ struct msm_dp {
> bool is_edp;
> bool internal_hpd;
>
> + struct msm_dp_mst *msm_dp_mst;
> hdmi_codec_plugged_cb plugged_cb;
>
> struct msm_dp_audio *msm_dp_audio;
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> new file mode 100644
> index 0000000000000000000000000000000000000000..e66bd1e565aeb4da3d636eb5f4aa75504d60fd40
> --- /dev/null
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> @@ -0,0 +1,490 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (c) 2022-2024 Qualcomm Innovation Center, Inc. All rights reserved.
> + * Copyright (c) 2018-2021, The Linux Foundation. All rights reserved.
> + */
> +
> +/*
> + * Copyright © 2014 Red Hat.
> + *
> + * Permission to use, copy, modify, distribute, and sell this software and its
> + * documentation for any purpose is hereby granted without fee, provided that
> + * the above copyright notice appear in all copies and that both that copyright
> + * notice and this permission notice appear in supporting documentation, and
> + * that the name of the copyright holders not be used in advertising or
> + * publicity pertaining to distribution of the software without specific,
> + * written prior permission. The copyright holders make no representations
> + * about the suitability of this software for any purpose. It is provided "as
> + * is" without express or implied warranty.
> + *
> + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
> + * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
> + * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
> + * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
> + * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
> + * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
> + * OF THIS SOFTWARE.
> + */
> +
> +#include "dp_mst_drm.h"
> +
> +static struct drm_private_state *msm_dp_mst_duplicate_bridge_state(struct drm_private_obj *obj)
> +{
> + struct msm_dp_mst_bridge_state *state;
> +
> + state = kmemdup(obj->state, sizeof(*state), GFP_KERNEL);
> + if (!state)
> + return NULL;
> +
> + __drm_atomic_helper_private_obj_duplicate_state(obj, &state->base);
> +
> + return &state->base;
> +}
> +
> +static void msm_dp_mst_destroy_bridge_state(struct drm_private_obj *obj,
> + struct drm_private_state *state)
> +{
> + struct msm_dp_mst_bridge_state *priv_state =
> + to_msm_dp_mst_bridge_priv_state(state);
> +
> + kfree(priv_state);
> +}
> +
> +static const struct drm_private_state_funcs msm_dp_mst_bridge_state_funcs = {
> + .atomic_duplicate_state = msm_dp_mst_duplicate_bridge_state,
> + .atomic_destroy_state = msm_dp_mst_destroy_bridge_state,
> +};
> +
> +/**
> + * dp_mst_find_vcpi_slots() - Find VCPI slots for this PBN value
> + * @mgr: manager to use
> + * @pbn: payload bandwidth to convert into slots.
> + *
> + * Calculate the number of VCPI slots that will be required for the given PBN
> + * value.
> + *
> + * RETURNS:
> + * The total slots required for this port, or error.
> + */
> +static int msm_dp_mst_find_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr, int pbn)
> +{
> + int num_slots;
> + struct drm_dp_mst_topology_state *state;
> +
> + state = to_drm_dp_mst_topology_state(mgr->base.state);
> + num_slots = DIV_ROUND_UP(pbn, dfixed_trunc(state->pbn_div));
drm_dp_atomic_find_time_slots() uses slightly different maths here, with
the different precision. Can we use the data that is set by that function
instead (payload->time_slots)?
> +
> + /* max. time slots - one slot for MTP header */
> + if (num_slots > 63)
> + return -ENOSPC;
> + return num_slots;
> +}
> +
> +static void _msm_dp_mst_update_timeslots(struct msm_dp_mst *mst,
> + struct msm_dp_mst_bridge *mst_bridge,
> + struct drm_dp_mst_port *port)
> +{
> + int i;
> + struct msm_dp_mst_bridge *msm_dp_bridge;
> + struct drm_dp_mst_topology_state *mst_state;
> + struct drm_dp_mst_atomic_payload *payload;
> + int prev_start = 0;
> + int prev_slots = 0;
> +
> + mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
> + payload = drm_atomic_get_mst_payload_state(mst_state, port);
> +
> + if (!payload) {
> + DRM_ERROR("mst bridge [%d] update_timeslots failed, null payload\n",
> + mst_bridge->id);
> + return;
> + }
> +
> + for (i = 0; i < mst->max_streams; i++) {
> + msm_dp_bridge = &mst->mst_bridge[i];
> + if (mst_bridge == msm_dp_bridge) {
> + /*
> + * When a payload was removed make sure to move any payloads after it
> + * to the left so all payloads are aligned to the left.
> + */
Please don't. drm_dp_remove_payload_part2() should take care of that for
us. What is the reason for caching the data if we have to manually
handle the cache?
> + if (payload->vc_start_slot < 0) {
> + // cache the payload
> + prev_start = msm_dp_bridge->start_slot;
> + prev_slots = msm_dp_bridge->num_slots;
> + msm_dp_bridge->pbn = 0;
> + msm_dp_bridge->start_slot = 1;
> + msm_dp_bridge->num_slots = 0;
> + msm_dp_bridge->vcpi = 0;
> + } else { //add payload
> + msm_dp_bridge->pbn = payload->pbn;
> + msm_dp_bridge->start_slot = payload->vc_start_slot;
> + msm_dp_bridge->num_slots = payload->time_slots;
> + msm_dp_bridge->vcpi = payload->vcpi;
> + }
> + }
> + }
> +
> + // Now commit all the updated payloads
> + for (i = 0; i < mst->max_streams; i++) {
> + msm_dp_bridge = &mst->mst_bridge[i];
> +
> + //Shift payloads to the left if there was a removed payload.
> + if (payload->vc_start_slot < 0 && msm_dp_bridge->start_slot > prev_start)
> + msm_dp_bridge->start_slot -= prev_slots;
> +
> + msm_dp_display_set_stream_info(mst->msm_dp, msm_dp_bridge->msm_dp_panel,
> + msm_dp_bridge->id, msm_dp_bridge->start_slot,
> + msm_dp_bridge->num_slots,
> + msm_dp_bridge->pbn, msm_dp_bridge->vcpi);
> + drm_dbg_dp(mst->msm_dp->drm_dev,
> + "conn:%d vcpi:%d start_slot:%d num_slots:%d, pbn:%d\n",
> + DP_MST_CONN_ID(msm_dp_bridge), msm_dp_bridge->vcpi,
> + msm_dp_bridge->start_slot,
> + msm_dp_bridge->num_slots, msm_dp_bridge->pbn);
> + }
> +}
> +
> +static int _msm_dp_mst_bridge_pre_enable_part1(struct msm_dp_mst_bridge *dp_bridge,
> + struct drm_bridge_state *bridge_state)
> +{
> + struct msm_dp *msm_dp = dp_bridge->display;
> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
> + struct drm_dp_mst_port *port = mst_conn->mst_port;
> + struct drm_dp_mst_topology_state *mst_state;
> + struct drm_dp_mst_atomic_payload *payload;
> + struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
> + int pbn, slots;
> + int rc = 0;
> +
> + mst_state = drm_atomic_get_new_mst_topology_state(bridge_state->base.state,
> + &mst->mst_mgr);
> +
> + pbn = drm_dp_calc_pbn_mode(dp_panel->msm_dp_mode.drm_mode.clock,
> + (mst_conn->connector.display_info.bpc * 3) << 4);
> +
> + slots = msm_dp_mst_find_vcpi_slots(&mst->mst_mgr, pbn);
> +
> + drm_dbg_dp(msm_dp->drm_dev, "conn:%d pbn:%d, slots:%d\n", DP_MST_CONN_ID(dp_bridge),
> + pbn, slots);
> +
> + payload = drm_atomic_get_mst_payload_state(mst_state, port);
> + if (!payload || payload->time_slots <= 0) {
> + DRM_ERROR("time slots not allocated for conn:%d\n", DP_MST_CONN_ID(dp_bridge));
> + rc = -EINVAL;
> + return rc;
> + }
> +
> + drm_dp_mst_update_slots(mst_state, DP_CAP_ANSI_8B10B);
> +
> + rc = drm_dp_add_payload_part1(&mst->mst_mgr, mst_state, payload);
> + if (rc) {
> + DRM_ERROR("payload allocation failure for conn:%d\n", DP_MST_CONN_ID(dp_bridge));
> + return rc;
> + }
> +
> + _msm_dp_mst_update_timeslots(mst, dp_bridge, port);
> +
> + return rc;
> +}
> +
> +static void _msm_dp_mst_bridge_pre_enable_part2(struct msm_dp_mst_bridge *dp_bridge,
> + struct drm_bridge_state *bridge_state)
> +{
> + struct msm_dp *msm_dp = dp_bridge->display;
> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
> + struct drm_dp_mst_port *port = mst_conn->mst_port;
> + struct drm_dp_mst_topology_state *mst_state;
> + struct drm_dp_mst_atomic_payload *payload;
> +
> + drm_dp_check_act_status(&mst->mst_mgr);
> +
> + mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
> + payload = drm_atomic_get_mst_payload_state(mst_state, port);
> +
> + if (!payload) {
> + DRM_ERROR("mst bridge [%d] null payload\n", dp_bridge->id);
> + return;
> + }
> +
> + if (!payload->port) {
> + DRM_ERROR("mst bridge [%d] null port\n", dp_bridge->id);
> + return;
> + }
> +
> + if (!payload->port->connector) {
> + DRM_ERROR("mst bridge [%d] part-2 failed, null connector\n",
> + dp_bridge->id);
> + return;
> + }
> +
> + if (payload->vc_start_slot == -1) {
> + DRM_ERROR("mst bridge [%d] part-2 failed, payload alloc part 1 failed\n",
> + dp_bridge->id);
> + return;
> + }
> +
> + drm_dp_add_payload_part2(&mst->mst_mgr, payload);
> +
> + drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre enable part-2 complete\n",
> + dp_bridge->id);
> +}
> +
> +static void _msm_dp_mst_bridge_pre_disable_part1(struct msm_dp_mst_bridge *dp_bridge,
> + struct drm_bridge_state *bridge_state)
> +{
> + struct msm_dp *msm_dp = dp_bridge->display;
> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
> + struct drm_dp_mst_port *port = mst_conn->mst_port;
> + struct drm_dp_mst_topology_state *old_mst_state;
> + struct drm_dp_mst_topology_state *new_mst_state;
> + const struct drm_dp_mst_atomic_payload *old_payload;
> + struct drm_dp_mst_atomic_payload *new_payload;
> +
> + old_mst_state = drm_atomic_get_old_mst_topology_state(bridge_state->base.state,
> + &mst->mst_mgr);
> +
> + new_mst_state = drm_atomic_get_new_mst_topology_state(bridge_state->base.state,
> + &mst->mst_mgr);
> +
> + old_payload = drm_atomic_get_mst_payload_state(old_mst_state, port);
> + new_payload = drm_atomic_get_mst_payload_state(new_mst_state, port);
> +
> + if (!old_payload || !new_payload) {
> + DRM_ERROR("mst bridge [%d] _pre disable part-1 failed, null payload\n",
> + dp_bridge->id);
> + return;
> + }
> +
> + drm_dp_remove_payload_part1(&mst->mst_mgr, new_mst_state, new_payload);
> + drm_dp_remove_payload_part2(&mst->mst_mgr, new_mst_state, old_payload, new_payload);
> +
> + _msm_dp_mst_update_timeslots(mst, dp_bridge, port);
> +
> + drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre disable part-1 complete\n",
> + dp_bridge->id);
> +}
> +
> +static void _msm_dp_mst_bridge_pre_disable_part2(struct msm_dp_mst_bridge *dp_bridge)
> +{
> + struct msm_dp *msm_dp = dp_bridge->display;
> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
> +
> + drm_dp_check_act_status(&mst->mst_mgr);
> +
> + drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre disable part-2 complete\n",
> + dp_bridge->id);
> +}
> +
> +static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
> + struct drm_bridge_state *old_bridge_state)
> +{
> + int rc = 0;
> + struct msm_dp_mst_bridge *bridge;
> + struct msm_dp *dp;
> + struct msm_dp_mst_bridge_state *msm_dp_bridge_state;
> +
> + if (!drm_bridge) {
> + DRM_ERROR("Invalid params\n");
> + return;
> + }
> +
> + bridge = to_msm_dp_mst_bridge(drm_bridge);
> + msm_dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
> + dp = bridge->display;
> +
> + /* to cover cases of bridge_disable/bridge_enable without modeset */
> + bridge->connector = msm_dp_bridge_state->connector;
> + bridge->msm_dp_panel = msm_dp_bridge_state->msm_dp_panel;
> +
> + if (!bridge->connector) {
> + DRM_ERROR("Invalid connector\n");
> + return;
> + }
> +
> + msm_dp_display_atomic_prepare(dp);
> +
> + rc = _msm_dp_mst_bridge_pre_enable_part1(bridge, old_bridge_state);
> + if (rc) {
> + DRM_ERROR("[%d] DP display pre-enable failed, rc=%d\n", bridge->id, rc);
> + msm_dp_display_unprepare(dp);
> + return;
> + }
> +
> + msm_dp_display_enable_helper(dp, bridge->msm_dp_panel);
> +
> + _msm_dp_mst_bridge_pre_enable_part2(bridge, old_bridge_state);
> +
> + drm_dbg_dp(dp->drm_dev, "conn:%d mode:%s fps:%d vcpi:%d slots:%d to %d\n",
> + DP_MST_CONN_ID(bridge), bridge->drm_mode.name,
> + drm_mode_vrefresh(&bridge->drm_mode),
> + bridge->vcpi, bridge->start_slot,
> + bridge->start_slot + bridge->num_slots);
> +}
> +
> +static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
> + struct drm_bridge_state *old_bridge_state)
> +{
> + struct msm_dp_mst_bridge *bridge;
> + struct msm_dp *dp;
> +
> + if (!drm_bridge) {
> + DRM_ERROR("Invalid params\n");
> + return;
> + }
> +
> + bridge = to_msm_dp_mst_bridge(drm_bridge);
> + if (!bridge->connector) {
> + DRM_ERROR("Invalid connector\n");
> + return;
> + }
> +
> + dp = bridge->display;
> +
> + _msm_dp_mst_bridge_pre_disable_part1(bridge, old_bridge_state);
> +
> + msm_dp_display_disable_helper(dp, bridge->msm_dp_panel);
> +
> + _msm_dp_mst_bridge_pre_disable_part2(bridge);
> +
> + drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d disable complete\n", bridge->id,
> + DP_MST_CONN_ID(bridge));
> +}
> +
> +static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
> + struct drm_bridge_state *old_bridge_state)
> +{
> + int conn = 0;
> + struct msm_dp_mst_bridge *bridge;
> + struct msm_dp *dp;
> +
> + if (!drm_bridge) {
> + DRM_ERROR("Invalid params\n");
> + return;
> + }
> +
> + bridge = to_msm_dp_mst_bridge(drm_bridge);
> + if (!bridge->connector) {
> + DRM_ERROR("Invalid connector\n");
> + return;
> + }
> +
> + conn = DP_MST_CONN_ID(bridge);
> +
> + dp = bridge->display;
> +
> + msm_dp_display_atomic_post_disable_helper(dp, bridge->msm_dp_panel);
> +
> + if (!dp->mst_active)
> + msm_dp_display_unprepare(dp);
> +
> + bridge->connector = NULL;
> + bridge->msm_dp_panel = NULL;
> +
> + drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d post disable complete\n",
> + bridge->id, conn);
> +}
> +
> +static void msm_dp_mst_bridge_mode_set(struct drm_bridge *drm_bridge,
> + const struct drm_display_mode *mode,
> + const struct drm_display_mode *adjusted_mode)
> +{
> + struct msm_dp_mst_bridge *bridge;
> + struct msm_dp_mst_bridge_state *dp_bridge_state;
> + struct msm_dp *dp;
> + struct msm_dp_panel *msm_dp_panel;
> +
> + if (!drm_bridge || !mode || !adjusted_mode) {
> + DRM_ERROR("Invalid params\n");
> + return;
> + }
> +
> + bridge = to_msm_dp_mst_bridge(drm_bridge);
> +
> + dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
> + bridge->connector = dp_bridge_state->connector;
> + bridge->msm_dp_panel = dp_bridge_state->msm_dp_panel;
> +
> + msm_dp_panel = bridge->msm_dp_panel;
> + dp = bridge->display;
> +
> + memset(&bridge->msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
> + memcpy(&bridge->drm_mode, adjusted_mode, sizeof(bridge->drm_mode));
> + msm_dp_display_mode_set_helper(dp, mode, adjusted_mode, bridge->msm_dp_panel);
> + msm_dp_panel->mst_caps.pbn = drm_dp_calc_pbn_mode(msm_dp_panel->msm_dp_mode.drm_mode.clock,
> + (msm_dp_panel->msm_dp_mode.bpp << 4));
> + memcpy(&bridge->msm_dp_mode, &bridge->msm_dp_panel->msm_dp_mode,
> + sizeof(struct msm_dp_display_mode));
No, you can't just memcpy the drm_mode struct.
> + drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d mode set complete %s\n", bridge->id,
> + DP_MST_CONN_ID(bridge), mode->name);
> +}
> +
> +/* DP MST Bridge APIs */
> +static const struct drm_bridge_funcs msm_dp_mst_bridge_ops = {
> + .atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
> + .atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
> + .atomic_reset = drm_atomic_helper_bridge_reset,
> + .atomic_pre_enable = msm_dp_mst_bridge_atomic_pre_enable,
> + .atomic_disable = msm_dp_mst_bridge_atomic_disable,
> + .atomic_post_disable = msm_dp_mst_bridge_atomic_post_disable,
> + .mode_set = msm_dp_mst_bridge_mode_set,
> +};
> +
> +int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder)
> +{
> + int rc = 0;
> + struct msm_dp_mst_bridge *bridge = NULL;
> + struct msm_dp_mst_bridge_state *state;
> + struct drm_device *dev;
> + struct msm_dp_mst *mst = dp->msm_dp_mst;
> + int i;
> +
> + for (i = 0; i < mst->max_streams; i++) {
> + if (!mst->mst_bridge[i].in_use) {
> + bridge = &mst->mst_bridge[i];
> + bridge->encoder = encoder;
> + bridge->in_use = true;
> + bridge->id = i;
> + break;
> + }
> + }
> +
> + if (i == mst->max_streams) {
> + DRM_ERROR("mst supports only %d bridges\n", i);
> + rc = -EACCES;
> + goto end;
> + }
> +
> + dev = dp->drm_dev;
> + bridge->display = dp;
> + bridge->base.funcs = &msm_dp_mst_bridge_ops;
> + bridge->base.encoder = encoder;
> + bridge->base.type = dp->connector_type;
> + bridge->base.ops = DRM_BRIDGE_OP_MODES;
> + drm_bridge_add(&bridge->base);
> +
> + rc = drm_bridge_attach(encoder, &bridge->base, NULL, 0);
> + if (rc) {
> + DRM_ERROR("failed to attach bridge, rc=%d\n", rc);
> + goto end;
> + }
> +
> + state = kzalloc(sizeof(*state), GFP_KERNEL);
> + if (!state) {
> + rc = -ENOMEM;
> + goto end;
> + }
> +
> + drm_atomic_private_obj_init(dev, &bridge->obj,
> + &state->base,
> + &msm_dp_mst_bridge_state_funcs);
> +
> + drm_dbg_dp(dp->drm_dev, "mst drm bridge init. bridge id:%d\n", i);
> +
> + return 0;
> +
> +end:
> + return rc;
> +}
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> new file mode 100644
> index 0000000000000000000000000000000000000000..c2a82cd3c6d6e1951a8e5905d3aa39dfc691023b
> --- /dev/null
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> @@ -0,0 +1,102 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (c) 2024 Qualcomm Innovation Center, Inc. All rights reserved.
> + *
> + * Permission to use, copy, modify, distribute, and sell this software and its
> + * documentation for any purpose is hereby granted without fee, provided that
> + * the above copyright notice appear in all copies and that both that copyright
> + * notice and this permission notice appear in supporting documentation, and
> + * that the name of the copyright holders not be used in advertising or
> + * publicity pertaining to distribution of the software without specific,
> + * written prior permission. The copyright holders make no representations
> + * about the suitability of this software for any purpose. It is provided "as
> + * is" without express or implied warranty.
> + *
> + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
> + * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
> + * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
> + * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
> + * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
> + * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
> + * OF THIS SOFTWARE.
> + */
> +
> +#ifndef _DP_MST_DRM_H_
> +#define _DP_MST_DRM_H_
> +
> +#include <linux/kernel.h>
> +#include <linux/init.h>
> +#include <linux/errno.h>
> +#include <linux/version.h>
> +
> +#include <drm/drm_atomic_helper.h>
> +#include <drm/drm_atomic.h>
> +#include <drm/drm_bridge.h>
> +#include <drm/drm_crtc.h>
> +#include <drm/drm_fixed.h>
> +#include <drm/drm_connector.h>
> +#include <drm/display/drm_dp_helper.h>
> +#include <drm/display/drm_dp_mst_helper.h>
> +
> +#include "dp_panel.h"
> +#include "dp_display.h"
> +
> +#define DP_MST_CONN_ID(bridge) ((bridge)->connector ? \
> + (bridge)->connector->base.id : 0)
locking? Also connectors can easily be freed, so you can't just read
into it.
> +
> +struct msm_dp_mst_bridge {
> + struct drm_bridge base;
> + struct drm_private_obj obj;
> + u32 id;
> +
> + bool in_use;
> +
> + struct msm_dp *display;
> + struct drm_encoder *encoder;
> +
> + struct drm_display_mode drm_mode;
Why? Where is this being set? Why can't you use the state objects instead?
> + struct msm_dp_display_mode msm_dp_mode;
I don't see this being used. Please drop.
> + struct drm_connector *connector;
So, you have connector here and a connector in the state. Please drop
one (I'd guess this one).
> + struct msm_dp_panel *msm_dp_panel;
> +
> + int vcpi;
> + int pbn;
> + int num_slots;
> + int start_slot;
Which of the fields (including in_use) are long-lived and which are a
part of the current state? Can we move all state ones to bridge's state?
> +};
> +
> +struct msm_dp_mst_bridge_state {
> + struct drm_private_state base;
> + struct drm_connector *connector;
> + struct msm_dp_panel *msm_dp_panel;
> + int num_slots;
> +};
> +
> +struct msm_dp_mst {
> + bool mst_initialized;
> + struct drm_dp_mst_topology_mgr mst_mgr;
> + struct msm_dp_mst_bridge *mst_bridge;
> + struct msm_dp *msm_dp;
> + bool mst_session_hpd_state;
> + u32 max_streams;
> +};
> +
> +struct msm_dp_mst_connector {
> + struct drm_connector connector;
> + struct drm_dp_mst_port *mst_port;
> + struct msm_dp *msm_dp;
> + struct msm_dp_panel *dp_panel;
> +};
> +
> +#define to_msm_dp_mst_bridge(x) container_of((x), struct msm_dp_mst_bridge, base)
> +#define to_msm_dp_mst_bridge_priv(x) \
> + container_of((x), struct msm_dp_mst_bridge, obj)
> +#define to_msm_dp_mst_bridge_priv_state(x) \
> + container_of((x), struct msm_dp_mst_bridge_state, base)
> +#define to_msm_dp_mst_bridge_state(x) \
> + to_msm_dp_mst_bridge_priv_state((x)->obj.state)
> +#define to_msm_dp_mst_connector(x) \
> + container_of((x), struct msm_dp_mst_connector, connector)
> +int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
> +
> +#endif /* _DP_MST_DRM_H_ */
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations
2024-12-06 10:12 ` Dmitry Baryshkov
@ 2025-05-27 10:29 ` Yongxing Mou
2025-05-28 18:02 ` Dmitry Baryshkov
0 siblings, 1 reply; 111+ messages in thread
From: Yongxing Mou @ 2025-05-27 10:29 UTC (permalink / raw)
To: Dmitry Baryshkov, Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On 2024/12/6 18:12, Dmitry Baryshkov wrote:
> On Thu, Dec 05, 2024 at 08:32:02PM -0800, Abhinav Kumar wrote:
>> Add a new file dp_mst_drm to manage the DP MST bridge operations
>> similar to the dp_drm file which manages the SST bridge operations.
>> Each MST encoder creates one bridge and each bridge is bound to its
>> own dp_panel abstraction to manage the operations of its pipeline.
>>
>> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
>> ---
>> drivers/gpu/drm/msm/Makefile | 3 +-
>> drivers/gpu/drm/msm/dp/dp_display.h | 2 +
>> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 490 ++++++++++++++++++++++++++++++++++++
>> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 102 ++++++++
>> 4 files changed, 596 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile
>> index f274d9430cc311405f890074c1466ffe2ec45ac9..b1e01b3123d9afc4818f059c5d4e7ca70dca3754 100644
>> --- a/drivers/gpu/drm/msm/Makefile
>> +++ b/drivers/gpu/drm/msm/Makefile
>> @@ -142,7 +142,8 @@ msm-display-$(CONFIG_DRM_MSM_DP)+= dp/dp_aux.o \
>> dp/dp_link.o \
>> dp/dp_panel.o \
>> dp/dp_audio.o \
>> - dp/dp_utils.o
>> + dp/dp_utils.o \
>> + dp/dp_mst_drm.o
>>
>> msm-display-$(CONFIG_DRM_MSM_HDMI_HDCP) += hdmi/hdmi_hdcp.o
>>
>> diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
>> index 6ab14e969bce0fd07b3a550bae17e99652479232..a5d4893f689c6afbbe622c9b7dfa98d23d754831 100644
>> --- a/drivers/gpu/drm/msm/dp/dp_display.h
>> +++ b/drivers/gpu/drm/msm/dp/dp_display.h
>> @@ -7,6 +7,7 @@
>> #define _DP_DISPLAY_H_
>>
>> #include "dp_panel.h"
>> +#include "dp_mst_drm.h"
>> #include <sound/hdmi-codec.h>
>> #include "disp/msm_disp_snapshot.h"
>>
>> @@ -26,6 +27,7 @@ struct msm_dp {
>> bool is_edp;
>> bool internal_hpd;
>>
>> + struct msm_dp_mst *msm_dp_mst;
>> hdmi_codec_plugged_cb plugged_cb;
>>
>> struct msm_dp_audio *msm_dp_audio;
>> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
>> new file mode 100644
>> index 0000000000000000000000000000000000000000..e66bd1e565aeb4da3d636eb5f4aa75504d60fd40
>> --- /dev/null
>> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
>> @@ -0,0 +1,490 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (c) 2022-2024 Qualcomm Innovation Center, Inc. All rights reserved.
>> + * Copyright (c) 2018-2021, The Linux Foundation. All rights reserved.
>> + */
>> +
>> +/*
>> + * Copyright © 2014 Red Hat.
>> + *
>> + * Permission to use, copy, modify, distribute, and sell this software and its
>> + * documentation for any purpose is hereby granted without fee, provided that
>> + * the above copyright notice appear in all copies and that both that copyright
>> + * notice and this permission notice appear in supporting documentation, and
>> + * that the name of the copyright holders not be used in advertising or
>> + * publicity pertaining to distribution of the software without specific,
>> + * written prior permission. The copyright holders make no representations
>> + * about the suitability of this software for any purpose. It is provided "as
>> + * is" without express or implied warranty.
>> + *
>> + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
>> + * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
>> + * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
>> + * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
>> + * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
>> + * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
>> + * OF THIS SOFTWARE.
>> + */
>> +
>> +#include "dp_mst_drm.h"
>> +
>> +static struct drm_private_state *msm_dp_mst_duplicate_bridge_state(struct drm_private_obj *obj)
>> +{
>> + struct msm_dp_mst_bridge_state *state;
>> +
>> + state = kmemdup(obj->state, sizeof(*state), GFP_KERNEL);
>> + if (!state)
>> + return NULL;
>> +
>> + __drm_atomic_helper_private_obj_duplicate_state(obj, &state->base);
>> +
>> + return &state->base;
>> +}
>> +
>> +static void msm_dp_mst_destroy_bridge_state(struct drm_private_obj *obj,
>> + struct drm_private_state *state)
>> +{
>> + struct msm_dp_mst_bridge_state *priv_state =
>> + to_msm_dp_mst_bridge_priv_state(state);
>> +
>> + kfree(priv_state);
>> +}
>> +
>> +static const struct drm_private_state_funcs msm_dp_mst_bridge_state_funcs = {
>> + .atomic_duplicate_state = msm_dp_mst_duplicate_bridge_state,
>> + .atomic_destroy_state = msm_dp_mst_destroy_bridge_state,
>> +};
>> +
>> +/**
>> + * dp_mst_find_vcpi_slots() - Find VCPI slots for this PBN value
>> + * @mgr: manager to use
>> + * @pbn: payload bandwidth to convert into slots.
>> + *
>> + * Calculate the number of VCPI slots that will be required for the given PBN
>> + * value.
>> + *
>> + * RETURNS:
>> + * The total slots required for this port, or error.
>> + */
>> +static int msm_dp_mst_find_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr, int pbn)
>> +{
>> + int num_slots;
>> + struct drm_dp_mst_topology_state *state;
>> +
>> + state = to_drm_dp_mst_topology_state(mgr->base.state);
>> + num_slots = DIV_ROUND_UP(pbn, dfixed_trunc(state->pbn_div));
>
> drm_dp_atomic_find_time_slots() uses slightly different maths here, with
> the different precision. Can we use the data that is set by that function
> instead (payload->time_slots)?
>
Note that drm_dp_atomic_find_time_slots all call in atomic_check func,
not in other place.So can we call this func in atomic_pre_enable? Also,
amg driver also have similar usage pattern.
>> +
>> + /* max. time slots - one slot for MTP header */
>> + if (num_slots > 63)
>> + return -ENOSPC;
>> + return num_slots;
>> +}
>> +
>> +static void _msm_dp_mst_update_timeslots(struct msm_dp_mst *mst,
>> + struct msm_dp_mst_bridge *mst_bridge,
>> + struct drm_dp_mst_port *port)
>> +{
>> + int i;
>> + struct msm_dp_mst_bridge *msm_dp_bridge;
>> + struct drm_dp_mst_topology_state *mst_state;
>> + struct drm_dp_mst_atomic_payload *payload;
>> + int prev_start = 0;
>> + int prev_slots = 0;
>> +
>> + mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
>> + payload = drm_atomic_get_mst_payload_state(mst_state, port);
>> +
>> + if (!payload) {
>> + DRM_ERROR("mst bridge [%d] update_timeslots failed, null payload\n",
>> + mst_bridge->id);
>> + return;
>> + }
>> +
>> + for (i = 0; i < mst->max_streams; i++) {
>> + msm_dp_bridge = &mst->mst_bridge[i];
>> + if (mst_bridge == msm_dp_bridge) {
>> + /*
>> + * When a payload was removed make sure to move any payloads after it
>> + * to the left so all payloads are aligned to the left.
>> + */
>
> Please don't. drm_dp_remove_payload_part2() should take care of that for
> us. What is the reason for caching the data if we have to manually
> handle the cache?
>
MST framework is managing the port's bandwidth, but we have a bridhge
for each stream, so we need to keep track of the payload allcation
status for each bridge, how much we used and how much we left. So maybe
they are manage two different part.
>> + if (payload->vc_start_slot < 0) {
>> + // cache the payload
>> + prev_start = msm_dp_bridge->start_slot;
>> + prev_slots = msm_dp_bridge->num_slots;
>> + msm_dp_bridge->pbn = 0;
>> + msm_dp_bridge->start_slot = 1;
>> + msm_dp_bridge->num_slots = 0;
>> + msm_dp_bridge->vcpi = 0;
>> + } else { //add payload
>> + msm_dp_bridge->pbn = payload->pbn;
>> + msm_dp_bridge->start_slot = payload->vc_start_slot;
>> + msm_dp_bridge->num_slots = payload->time_slots;
>> + msm_dp_bridge->vcpi = payload->vcpi;
>> + }
>> + }
>> + }
>> +
>> + // Now commit all the updated payloads
>> + for (i = 0; i < mst->max_streams; i++) {
>> + msm_dp_bridge = &mst->mst_bridge[i];
>> +
>> + //Shift payloads to the left if there was a removed payload.
>> + if (payload->vc_start_slot < 0 && msm_dp_bridge->start_slot > prev_start)
>> + msm_dp_bridge->start_slot -= prev_slots;
>> +
>> + msm_dp_display_set_stream_info(mst->msm_dp, msm_dp_bridge->msm_dp_panel,
>> + msm_dp_bridge->id, msm_dp_bridge->start_slot,
>> + msm_dp_bridge->num_slots,
>> + msm_dp_bridge->pbn, msm_dp_bridge->vcpi);
>> + drm_dbg_dp(mst->msm_dp->drm_dev,
>> + "conn:%d vcpi:%d start_slot:%d num_slots:%d, pbn:%d\n",
>> + DP_MST_CONN_ID(msm_dp_bridge), msm_dp_bridge->vcpi,
>> + msm_dp_bridge->start_slot,
>> + msm_dp_bridge->num_slots, msm_dp_bridge->pbn);
>> + }
>> +}
>> +
>> +static int _msm_dp_mst_bridge_pre_enable_part1(struct msm_dp_mst_bridge *dp_bridge,
>> + struct drm_bridge_state *bridge_state)
>> +{
>> + struct msm_dp *msm_dp = dp_bridge->display;
>> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
>> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
>> + struct drm_dp_mst_port *port = mst_conn->mst_port;
>> + struct drm_dp_mst_topology_state *mst_state;
>> + struct drm_dp_mst_atomic_payload *payload;
>> + struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
>> + int pbn, slots;
>> + int rc = 0;
>> +
>> + mst_state = drm_atomic_get_new_mst_topology_state(bridge_state->base.state,
>> + &mst->mst_mgr);
>> +
>> + pbn = drm_dp_calc_pbn_mode(dp_panel->msm_dp_mode.drm_mode.clock,
>> + (mst_conn->connector.display_info.bpc * 3) << 4);
>> +
>> + slots = msm_dp_mst_find_vcpi_slots(&mst->mst_mgr, pbn);
>> +
>> + drm_dbg_dp(msm_dp->drm_dev, "conn:%d pbn:%d, slots:%d\n", DP_MST_CONN_ID(dp_bridge),
>> + pbn, slots);
>> +
>> + payload = drm_atomic_get_mst_payload_state(mst_state, port);
>> + if (!payload || payload->time_slots <= 0) {
>> + DRM_ERROR("time slots not allocated for conn:%d\n", DP_MST_CONN_ID(dp_bridge));
>> + rc = -EINVAL;
>> + return rc;
>> + }
>> +
>> + drm_dp_mst_update_slots(mst_state, DP_CAP_ANSI_8B10B);
>> +
>> + rc = drm_dp_add_payload_part1(&mst->mst_mgr, mst_state, payload);
>> + if (rc) {
>> + DRM_ERROR("payload allocation failure for conn:%d\n", DP_MST_CONN_ID(dp_bridge));
>> + return rc;
>> + }
>> +
>> + _msm_dp_mst_update_timeslots(mst, dp_bridge, port);
>> +
>> + return rc;
>> +}
>> +
>> +static void _msm_dp_mst_bridge_pre_enable_part2(struct msm_dp_mst_bridge *dp_bridge,
>> + struct drm_bridge_state *bridge_state)
>> +{
>> + struct msm_dp *msm_dp = dp_bridge->display;
>> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
>> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
>> + struct drm_dp_mst_port *port = mst_conn->mst_port;
>> + struct drm_dp_mst_topology_state *mst_state;
>> + struct drm_dp_mst_atomic_payload *payload;
>> +
>> + drm_dp_check_act_status(&mst->mst_mgr);
>> +
>> + mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
>> + payload = drm_atomic_get_mst_payload_state(mst_state, port);
>> +
>> + if (!payload) {
>> + DRM_ERROR("mst bridge [%d] null payload\n", dp_bridge->id);
>> + return;
>> + }
>> +
>> + if (!payload->port) {
>> + DRM_ERROR("mst bridge [%d] null port\n", dp_bridge->id);
>> + return;
>> + }
>> +
>> + if (!payload->port->connector) {
>> + DRM_ERROR("mst bridge [%d] part-2 failed, null connector\n",
>> + dp_bridge->id);
>> + return;
>> + }
>> +
>> + if (payload->vc_start_slot == -1) {
>> + DRM_ERROR("mst bridge [%d] part-2 failed, payload alloc part 1 failed\n",
>> + dp_bridge->id);
>> + return;
>> + }
>> +
>> + drm_dp_add_payload_part2(&mst->mst_mgr, payload);
>> +
>> + drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre enable part-2 complete\n",
>> + dp_bridge->id);
>> +}
>> +
>> +static void _msm_dp_mst_bridge_pre_disable_part1(struct msm_dp_mst_bridge *dp_bridge,
>> + struct drm_bridge_state *bridge_state)
>> +{
>> + struct msm_dp *msm_dp = dp_bridge->display;
>> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
>> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
>> + struct drm_dp_mst_port *port = mst_conn->mst_port;
>> + struct drm_dp_mst_topology_state *old_mst_state;
>> + struct drm_dp_mst_topology_state *new_mst_state;
>> + const struct drm_dp_mst_atomic_payload *old_payload;
>> + struct drm_dp_mst_atomic_payload *new_payload;
>> +
>> + old_mst_state = drm_atomic_get_old_mst_topology_state(bridge_state->base.state,
>> + &mst->mst_mgr);
>> +
>> + new_mst_state = drm_atomic_get_new_mst_topology_state(bridge_state->base.state,
>> + &mst->mst_mgr);
>> +
>> + old_payload = drm_atomic_get_mst_payload_state(old_mst_state, port);
>> + new_payload = drm_atomic_get_mst_payload_state(new_mst_state, port);
>> +
>> + if (!old_payload || !new_payload) {
>> + DRM_ERROR("mst bridge [%d] _pre disable part-1 failed, null payload\n",
>> + dp_bridge->id);
>> + return;
>> + }
>> +
>> + drm_dp_remove_payload_part1(&mst->mst_mgr, new_mst_state, new_payload);
>> + drm_dp_remove_payload_part2(&mst->mst_mgr, new_mst_state, old_payload, new_payload);
>> +
>> + _msm_dp_mst_update_timeslots(mst, dp_bridge, port);
>> +
>> + drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre disable part-1 complete\n",
>> + dp_bridge->id);
>> +}
>> +
>> +static void _msm_dp_mst_bridge_pre_disable_part2(struct msm_dp_mst_bridge *dp_bridge)
>> +{
>> + struct msm_dp *msm_dp = dp_bridge->display;
>> + struct msm_dp_mst *mst = msm_dp->msm_dp_mst;
>> +
>> + drm_dp_check_act_status(&mst->mst_mgr);
>> +
>> + drm_dbg_dp(msm_dp->drm_dev, "mst bridge [%d] _pre disable part-2 complete\n",
>> + dp_bridge->id);
>> +}
>> +
>> +static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
>> + struct drm_bridge_state *old_bridge_state)
>> +{
>> + int rc = 0;
>> + struct msm_dp_mst_bridge *bridge;
>> + struct msm_dp *dp;
>> + struct msm_dp_mst_bridge_state *msm_dp_bridge_state;
>> +
>> + if (!drm_bridge) {
>> + DRM_ERROR("Invalid params\n");
>> + return;
>> + }
>> +
>> + bridge = to_msm_dp_mst_bridge(drm_bridge);
>> + msm_dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
>> + dp = bridge->display;
>> +
>> + /* to cover cases of bridge_disable/bridge_enable without modeset */
>> + bridge->connector = msm_dp_bridge_state->connector;
>> + bridge->msm_dp_panel = msm_dp_bridge_state->msm_dp_panel;
>> +
>> + if (!bridge->connector) {
>> + DRM_ERROR("Invalid connector\n");
>> + return;
>> + }
>> +
>> + msm_dp_display_atomic_prepare(dp);
>> +
>> + rc = _msm_dp_mst_bridge_pre_enable_part1(bridge, old_bridge_state);
>> + if (rc) {
>> + DRM_ERROR("[%d] DP display pre-enable failed, rc=%d\n", bridge->id, rc);
>> + msm_dp_display_unprepare(dp);
>> + return;
>> + }
>> +
>> + msm_dp_display_enable_helper(dp, bridge->msm_dp_panel);
>> +
>> + _msm_dp_mst_bridge_pre_enable_part2(bridge, old_bridge_state);
>> +
>> + drm_dbg_dp(dp->drm_dev, "conn:%d mode:%s fps:%d vcpi:%d slots:%d to %d\n",
>> + DP_MST_CONN_ID(bridge), bridge->drm_mode.name,
>> + drm_mode_vrefresh(&bridge->drm_mode),
>> + bridge->vcpi, bridge->start_slot,
>> + bridge->start_slot + bridge->num_slots);
>> +}
>> +
>> +static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
>> + struct drm_bridge_state *old_bridge_state)
>> +{
>> + struct msm_dp_mst_bridge *bridge;
>> + struct msm_dp *dp;
>> +
>> + if (!drm_bridge) {
>> + DRM_ERROR("Invalid params\n");
>> + return;
>> + }
>> +
>> + bridge = to_msm_dp_mst_bridge(drm_bridge);
>> + if (!bridge->connector) {
>> + DRM_ERROR("Invalid connector\n");
>> + return;
>> + }
>> +
>> + dp = bridge->display;
>> +
>> + _msm_dp_mst_bridge_pre_disable_part1(bridge, old_bridge_state);
>> +
>> + msm_dp_display_disable_helper(dp, bridge->msm_dp_panel);
>> +
>> + _msm_dp_mst_bridge_pre_disable_part2(bridge);
>> +
>> + drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d disable complete\n", bridge->id,
>> + DP_MST_CONN_ID(bridge));
>> +}
>> +
>> +static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
>> + struct drm_bridge_state *old_bridge_state)
>> +{
>> + int conn = 0;
>> + struct msm_dp_mst_bridge *bridge;
>> + struct msm_dp *dp;
>> +
>> + if (!drm_bridge) {
>> + DRM_ERROR("Invalid params\n");
>> + return;
>> + }
>> +
>> + bridge = to_msm_dp_mst_bridge(drm_bridge);
>> + if (!bridge->connector) {
>> + DRM_ERROR("Invalid connector\n");
>> + return;
>> + }
>> +
>> + conn = DP_MST_CONN_ID(bridge);
>> +
>> + dp = bridge->display;
>> +
>> + msm_dp_display_atomic_post_disable_helper(dp, bridge->msm_dp_panel);
>> +
>> + if (!dp->mst_active)
>> + msm_dp_display_unprepare(dp);
>> +
>> + bridge->connector = NULL;
>> + bridge->msm_dp_panel = NULL;
>> +
>> + drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d post disable complete\n",
>> + bridge->id, conn);
>> +}
>> +
>> +static void msm_dp_mst_bridge_mode_set(struct drm_bridge *drm_bridge,
>> + const struct drm_display_mode *mode,
>> + const struct drm_display_mode *adjusted_mode)
>> +{
>> + struct msm_dp_mst_bridge *bridge;
>> + struct msm_dp_mst_bridge_state *dp_bridge_state;
>> + struct msm_dp *dp;
>> + struct msm_dp_panel *msm_dp_panel;
>> +
>> + if (!drm_bridge || !mode || !adjusted_mode) {
>> + DRM_ERROR("Invalid params\n");
>> + return;
>> + }
>> +
>> + bridge = to_msm_dp_mst_bridge(drm_bridge);
>> +
>> + dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
>> + bridge->connector = dp_bridge_state->connector;
>> + bridge->msm_dp_panel = dp_bridge_state->msm_dp_panel;
>> +
>> + msm_dp_panel = bridge->msm_dp_panel;
>> + dp = bridge->display;
>> +
>> + memset(&bridge->msm_dp_mode, 0x0, sizeof(struct msm_dp_display_mode));
>> + memcpy(&bridge->drm_mode, adjusted_mode, sizeof(bridge->drm_mode));
>> + msm_dp_display_mode_set_helper(dp, mode, adjusted_mode, bridge->msm_dp_panel);
>> + msm_dp_panel->mst_caps.pbn = drm_dp_calc_pbn_mode(msm_dp_panel->msm_dp_mode.drm_mode.clock,
>> + (msm_dp_panel->msm_dp_mode.bpp << 4));
>> + memcpy(&bridge->msm_dp_mode, &bridge->msm_dp_panel->msm_dp_mode,
>> + sizeof(struct msm_dp_display_mode));
>
> No, you can't just memcpy the drm_mode struct.
>
Got it.
>> + drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d mode set complete %s\n", bridge->id,
>> + DP_MST_CONN_ID(bridge), mode->name);
>> +}
>> +
>> +/* DP MST Bridge APIs */
>> +static const struct drm_bridge_funcs msm_dp_mst_bridge_ops = {
>> + .atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
>> + .atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
>> + .atomic_reset = drm_atomic_helper_bridge_reset,
>> + .atomic_pre_enable = msm_dp_mst_bridge_atomic_pre_enable,
>> + .atomic_disable = msm_dp_mst_bridge_atomic_disable,
>> + .atomic_post_disable = msm_dp_mst_bridge_atomic_post_disable,
>> + .mode_set = msm_dp_mst_bridge_mode_set,
>> +};
>> +
>> +int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder)
>> +{
>> + int rc = 0;
>> + struct msm_dp_mst_bridge *bridge = NULL;
>> + struct msm_dp_mst_bridge_state *state;
>> + struct drm_device *dev;
>> + struct msm_dp_mst *mst = dp->msm_dp_mst;
>> + int i;
>> +
>> + for (i = 0; i < mst->max_streams; i++) {
>> + if (!mst->mst_bridge[i].in_use) {
>> + bridge = &mst->mst_bridge[i];
>> + bridge->encoder = encoder;
>> + bridge->in_use = true;
>> + bridge->id = i;
>> + break;
>> + }
>> + }
>> +
>> + if (i == mst->max_streams) {
>> + DRM_ERROR("mst supports only %d bridges\n", i);
>> + rc = -EACCES;
>> + goto end;
>> + }
>> +
>> + dev = dp->drm_dev;
>> + bridge->display = dp;
>> + bridge->base.funcs = &msm_dp_mst_bridge_ops;
>> + bridge->base.encoder = encoder;
>> + bridge->base.type = dp->connector_type;
>> + bridge->base.ops = DRM_BRIDGE_OP_MODES;
>> + drm_bridge_add(&bridge->base);
>> +
>> + rc = drm_bridge_attach(encoder, &bridge->base, NULL, 0);
>> + if (rc) {
>> + DRM_ERROR("failed to attach bridge, rc=%d\n", rc);
>> + goto end;
>> + }
>> +
>> + state = kzalloc(sizeof(*state), GFP_KERNEL);
>> + if (!state) {
>> + rc = -ENOMEM;
>> + goto end;
>> + }
>> +
>> + drm_atomic_private_obj_init(dev, &bridge->obj,
>> + &state->base,
>> + &msm_dp_mst_bridge_state_funcs);
>> +
>> + drm_dbg_dp(dp->drm_dev, "mst drm bridge init. bridge id:%d\n", i);
>> +
>> + return 0;
>> +
>> +end:
>> + return rc;
>> +}
>> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
>> new file mode 100644
>> index 0000000000000000000000000000000000000000..c2a82cd3c6d6e1951a8e5905d3aa39dfc691023b
>> --- /dev/null
>> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
>> @@ -0,0 +1,102 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (c) 2024 Qualcomm Innovation Center, Inc. All rights reserved.
>> + *
>> + * Permission to use, copy, modify, distribute, and sell this software and its
>> + * documentation for any purpose is hereby granted without fee, provided that
>> + * the above copyright notice appear in all copies and that both that copyright
>> + * notice and this permission notice appear in supporting documentation, and
>> + * that the name of the copyright holders not be used in advertising or
>> + * publicity pertaining to distribution of the software without specific,
>> + * written prior permission. The copyright holders make no representations
>> + * about the suitability of this software for any purpose. It is provided "as
>> + * is" without express or implied warranty.
>> + *
>> + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
>> + * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
>> + * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
>> + * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
>> + * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
>> + * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
>> + * OF THIS SOFTWARE.
>> + */
>> +
>> +#ifndef _DP_MST_DRM_H_
>> +#define _DP_MST_DRM_H_
>> +
>> +#include <linux/kernel.h>
>> +#include <linux/init.h>
>> +#include <linux/errno.h>
>> +#include <linux/version.h>
>> +
>> +#include <drm/drm_atomic_helper.h>
>> +#include <drm/drm_atomic.h>
>> +#include <drm/drm_bridge.h>
>> +#include <drm/drm_crtc.h>
>> +#include <drm/drm_fixed.h>
>> +#include <drm/drm_connector.h>
>> +#include <drm/display/drm_dp_helper.h>
>> +#include <drm/display/drm_dp_mst_helper.h>
>> +
>> +#include "dp_panel.h"
>> +#include "dp_display.h"
>> +
>> +#define DP_MST_CONN_ID(bridge) ((bridge)->connector ? \
>> + (bridge)->connector->base.id : 0)
>
> locking? Also connectors can easily be freed, so you can't just read
> into it.
Got it.
>> +
>> +struct msm_dp_mst_bridge {
>> + struct drm_bridge base;
>> + struct drm_private_obj obj;
>> + u32 id;
>> +
>> + bool in_use;
>> +
>> + struct msm_dp *display;
>> + struct drm_encoder *encoder;
>> +
>> + struct drm_display_mode drm_mode;
>
> Why? Where is this being set? Why can't you use the state objects instead?
>
>> + struct msm_dp_display_mode msm_dp_mode;
>
> I don't see this being used. Please drop.
Got it. It stores a backup of msm_dp_mode from dp_panel, but there's
really no need for two drm_mode instances here
>
>> + struct drm_connector *connector;
>
> So, you have connector here and a connector in the state. Please drop
> one (I'd guess this one).
>
Got it.
>> + struct msm_dp_panel *msm_dp_panel;
>> +
>> + int vcpi;
>> + int pbn;
>> + int num_slots;
>> + int start_slot;
>
> Which of the fields (including in_use) are long-lived and which are a
> part of the current state? Can we move all state ones to bridge's state?
>
in_use only used in bridge_init, so it is long-lived. Looking at it
together with the next patch, only num_slots changes during
atomic_check, so it is in bridge_state. pbn/vcpi/start_slots only change
during bridge enable/disable, so they are placed in the bridge.
>> +};
>> +
>> +struct msm_dp_mst_bridge_state {
>> + struct drm_private_state base;
>> + struct drm_connector *connector;
>> + struct msm_dp_panel *msm_dp_panel;
>> + int num_slots;
>> +};
>> +
>> +struct msm_dp_mst {
>> + bool mst_initialized;
>> + struct drm_dp_mst_topology_mgr mst_mgr;
>> + struct msm_dp_mst_bridge *mst_bridge;
>> + struct msm_dp *msm_dp;
>> + bool mst_session_hpd_state;
>> + u32 max_streams;
>> +};
>> +
>> +struct msm_dp_mst_connector {
>> + struct drm_connector connector;
>> + struct drm_dp_mst_port *mst_port;
>> + struct msm_dp *msm_dp;
>> + struct msm_dp_panel *dp_panel;
>> +};
>> +
>> +#define to_msm_dp_mst_bridge(x) container_of((x), struct msm_dp_mst_bridge, base)
>> +#define to_msm_dp_mst_bridge_priv(x) \
>> + container_of((x), struct msm_dp_mst_bridge, obj)
>> +#define to_msm_dp_mst_bridge_priv_state(x) \
>> + container_of((x), struct msm_dp_mst_bridge_state, base)
>> +#define to_msm_dp_mst_bridge_state(x) \
>> + to_msm_dp_mst_bridge_priv_state((x)->obj.state)
>> +#define to_msm_dp_mst_connector(x) \
>> + container_of((x), struct msm_dp_mst_connector, connector)
>> +int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
>> +
>> +#endif /* _DP_MST_DRM_H_ */
>>
>> --
>> 2.34.1
>>
>
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations
2025-05-27 10:29 ` Yongxing Mou
@ 2025-05-28 18:02 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2025-05-28 18:02 UTC (permalink / raw)
To: Yongxing Mou
Cc: Abhinav Kumar, Rob Clark, Sean Paul, Marijn Suijten, David Airlie,
Simona Vetter, Stephen Boyd, Chandan Uddaraju, Guenter Roeck,
Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart
On Tue, May 27, 2025 at 06:29:49PM +0800, Yongxing Mou wrote:
>
>
> On 2024/12/6 18:12, Dmitry Baryshkov wrote:
> > On Thu, Dec 05, 2024 at 08:32:02PM -0800, Abhinav Kumar wrote:
> > > Add a new file dp_mst_drm to manage the DP MST bridge operations
> > > similar to the dp_drm file which manages the SST bridge operations.
> > > Each MST encoder creates one bridge and each bridge is bound to its
> > > own dp_panel abstraction to manage the operations of its pipeline.
> > >
> > > Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> > > ---
> > > drivers/gpu/drm/msm/Makefile | 3 +-
> > > drivers/gpu/drm/msm/dp/dp_display.h | 2 +
> > > drivers/gpu/drm/msm/dp/dp_mst_drm.c | 490 ++++++++++++++++++++++++++++++++++++
> > > drivers/gpu/drm/msm/dp/dp_mst_drm.h | 102 ++++++++
> > > 4 files changed, 596 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile
> > > index f274d9430cc311405f890074c1466ffe2ec45ac9..b1e01b3123d9afc4818f059c5d4e7ca70dca3754 100644
> > > --- a/drivers/gpu/drm/msm/Makefile
> > > +++ b/drivers/gpu/drm/msm/Makefile
> > > @@ -142,7 +142,8 @@ msm-display-$(CONFIG_DRM_MSM_DP)+= dp/dp_aux.o \
> > > dp/dp_link.o \
> > > dp/dp_panel.o \
> > > dp/dp_audio.o \
> > > - dp/dp_utils.o
> > > + dp/dp_utils.o \
> > > + dp/dp_mst_drm.o
> > > msm-display-$(CONFIG_DRM_MSM_HDMI_HDCP) += hdmi/hdmi_hdcp.o
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> > > index 6ab14e969bce0fd07b3a550bae17e99652479232..a5d4893f689c6afbbe622c9b7dfa98d23d754831 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_display.h
> > > +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> > > @@ -7,6 +7,7 @@
> > > #define _DP_DISPLAY_H_
> > > #include "dp_panel.h"
> > > +#include "dp_mst_drm.h"
> > > #include <sound/hdmi-codec.h>
> > > #include "disp/msm_disp_snapshot.h"
> > > @@ -26,6 +27,7 @@ struct msm_dp {
> > > bool is_edp;
> > > bool internal_hpd;
> > > + struct msm_dp_mst *msm_dp_mst;
> > > hdmi_codec_plugged_cb plugged_cb;
> > > struct msm_dp_audio *msm_dp_audio;
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> > > new file mode 100644
> > > index 0000000000000000000000000000000000000000..e66bd1e565aeb4da3d636eb5f4aa75504d60fd40
> > > --- /dev/null
> > > +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> > > @@ -0,0 +1,490 @@
> > > +// SPDX-License-Identifier: GPL-2.0-only
> > > +/*
> > > + * Copyright (c) 2022-2024 Qualcomm Innovation Center, Inc. All rights reserved.
> > > + * Copyright (c) 2018-2021, The Linux Foundation. All rights reserved.
> > > + */
> > > +
> > > +/*
> > > + * Copyright © 2014 Red Hat.
> > > + *
> > > + * Permission to use, copy, modify, distribute, and sell this software and its
> > > + * documentation for any purpose is hereby granted without fee, provided that
> > > + * the above copyright notice appear in all copies and that both that copyright
> > > + * notice and this permission notice appear in supporting documentation, and
> > > + * that the name of the copyright holders not be used in advertising or
> > > + * publicity pertaining to distribution of the software without specific,
> > > + * written prior permission. The copyright holders make no representations
> > > + * about the suitability of this software for any purpose. It is provided "as
> > > + * is" without express or implied warranty.
> > > + *
> > > + * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
> > > + * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
> > > + * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
> > > + * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
> > > + * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
> > > + * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
> > > + * OF THIS SOFTWARE.
> > > + */
> > > +
> > > +#include "dp_mst_drm.h"
> > > +
> > > +static struct drm_private_state *msm_dp_mst_duplicate_bridge_state(struct drm_private_obj *obj)
> > > +{
> > > + struct msm_dp_mst_bridge_state *state;
> > > +
> > > + state = kmemdup(obj->state, sizeof(*state), GFP_KERNEL);
> > > + if (!state)
> > > + return NULL;
> > > +
> > > + __drm_atomic_helper_private_obj_duplicate_state(obj, &state->base);
> > > +
> > > + return &state->base;
> > > +}
> > > +
> > > +static void msm_dp_mst_destroy_bridge_state(struct drm_private_obj *obj,
> > > + struct drm_private_state *state)
> > > +{
> > > + struct msm_dp_mst_bridge_state *priv_state =
> > > + to_msm_dp_mst_bridge_priv_state(state);
> > > +
> > > + kfree(priv_state);
> > > +}
> > > +
> > > +static const struct drm_private_state_funcs msm_dp_mst_bridge_state_funcs = {
> > > + .atomic_duplicate_state = msm_dp_mst_duplicate_bridge_state,
> > > + .atomic_destroy_state = msm_dp_mst_destroy_bridge_state,
> > > +};
> > > +
> > > +/**
> > > + * dp_mst_find_vcpi_slots() - Find VCPI slots for this PBN value
> > > + * @mgr: manager to use
> > > + * @pbn: payload bandwidth to convert into slots.
> > > + *
> > > + * Calculate the number of VCPI slots that will be required for the given PBN
> > > + * value.
> > > + *
> > > + * RETURNS:
> > > + * The total slots required for this port, or error.
> > > + */
> > > +static int msm_dp_mst_find_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr, int pbn)
> > > +{
> > > + int num_slots;
> > > + struct drm_dp_mst_topology_state *state;
> > > +
> > > + state = to_drm_dp_mst_topology_state(mgr->base.state);
> > > + num_slots = DIV_ROUND_UP(pbn, dfixed_trunc(state->pbn_div));
> >
> > drm_dp_atomic_find_time_slots() uses slightly different maths here, with
> > the different precision. Can we use the data that is set by that function
> > instead (payload->time_slots)?
> >
> Note that drm_dp_atomic_find_time_slots all call in atomic_check func, not
> in other place.So can we call this func in atomic_pre_enable? Also, amg
> driver also have similar usage pattern.
Well, granted that this function can return ENOSPC, it is an error to
call it in atomic_pre_enable(). Nothing in atomic_pre_enable() /
atomic_enable() is allowed to fail.
I think this answers my question: drm_dp_atomic_find_time_slots()
should be called from atomic_check(), this function must be dropped.
> > > +
> > > + /* max. time slots - one slot for MTP header */
> > > + if (num_slots > 63)
> > > + return -ENOSPC;
> > > + return num_slots;
> > > +}
> > > +
> > > +static void _msm_dp_mst_update_timeslots(struct msm_dp_mst *mst,
> > > + struct msm_dp_mst_bridge *mst_bridge,
> > > + struct drm_dp_mst_port *port)
> > > +{
> > > + int i;
> > > + struct msm_dp_mst_bridge *msm_dp_bridge;
> > > + struct drm_dp_mst_topology_state *mst_state;
> > > + struct drm_dp_mst_atomic_payload *payload;
> > > + int prev_start = 0;
> > > + int prev_slots = 0;
> > > +
> > > + mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
> > > + payload = drm_atomic_get_mst_payload_state(mst_state, port);
> > > +
> > > + if (!payload) {
> > > + DRM_ERROR("mst bridge [%d] update_timeslots failed, null payload\n",
> > > + mst_bridge->id);
> > > + return;
> > > + }
> > > +
> > > + for (i = 0; i < mst->max_streams; i++) {
> > > + msm_dp_bridge = &mst->mst_bridge[i];
> > > + if (mst_bridge == msm_dp_bridge) {
> > > + /*
> > > + * When a payload was removed make sure to move any payloads after it
> > > + * to the left so all payloads are aligned to the left.
> > > + */
> >
> > Please don't. drm_dp_remove_payload_part2() should take care of that for
> > us. What is the reason for caching the data if we have to manually
> > handle the cache?
> >
> MST framework is managing the port's bandwidth, but we have a bridhge for
> each stream, so we need to keep track of the payload allcation status for
> each bridge, how much we used and how much we left. So maybe they are manage
> two different part.
Well, still no. MST topology master should handle all payload
allocation. If you need any data, enhance its API instead of duplicating
its functionality.
> > > + if (payload->vc_start_slot < 0) {
> > > + // cache the payload
> > > + prev_start = msm_dp_bridge->start_slot;
> > > + prev_slots = msm_dp_bridge->num_slots;
> > > + msm_dp_bridge->pbn = 0;
> > > + msm_dp_bridge->start_slot = 1;
> > > + msm_dp_bridge->num_slots = 0;
> > > + msm_dp_bridge->vcpi = 0;
> > > + } else { //add payload
> > > + msm_dp_bridge->pbn = payload->pbn;
> > > + msm_dp_bridge->start_slot = payload->vc_start_slot;
> > > + msm_dp_bridge->num_slots = payload->time_slots;
> > > + msm_dp_bridge->vcpi = payload->vcpi;
> > > + }
> > > + }
> > > + }
> > > +
> > > + // Now commit all the updated payloads
> > > + for (i = 0; i < mst->max_streams; i++) {
> > > + msm_dp_bridge = &mst->mst_bridge[i];
> > > +
> > > + //Shift payloads to the left if there was a removed payload.
> > > + if (payload->vc_start_slot < 0 && msm_dp_bridge->start_slot > prev_start)
> > > + msm_dp_bridge->start_slot -= prev_slots;
> > > +
> > > + msm_dp_display_set_stream_info(mst->msm_dp, msm_dp_bridge->msm_dp_panel,
> > > + msm_dp_bridge->id, msm_dp_bridge->start_slot,
> > > + msm_dp_bridge->num_slots,
> > > + msm_dp_bridge->pbn, msm_dp_bridge->vcpi);
> > > + drm_dbg_dp(mst->msm_dp->drm_dev,
> > > + "conn:%d vcpi:%d start_slot:%d num_slots:%d, pbn:%d\n",
> > > + DP_MST_CONN_ID(msm_dp_bridge), msm_dp_bridge->vcpi,
> > > + msm_dp_bridge->start_slot,
> > > + msm_dp_bridge->num_slots, msm_dp_bridge->pbn);
> > > + }
> > > +}
> > > +
[...]
> > > + struct msm_dp_panel *msm_dp_panel;
> > > +
> > > + int vcpi;
> > > + int pbn;
> > > + int num_slots;
> > > + int start_slot;
> >
> > Which of the fields (including in_use) are long-lived and which are a
> > part of the current state? Can we move all state ones to bridge's state?
> >
> in_use only used in bridge_init, so it is long-lived. Looking at it together
> with the next patch, only num_slots changes during atomic_check, so it is in
> bridge_state. pbn/vcpi/start_slots only change during bridge enable/disable,
> so they are placed in the bridge.
Let's look how it will look after refactoring. I'd still push all
changing fields to state. It make a lot of things much easier.
> > > +};
> > > +
> > > +struct msm_dp_mst_bridge_state {
> > > + struct drm_private_state base;
> > > + struct drm_connector *connector;
> > > + struct msm_dp_panel *msm_dp_panel;
> > > + int num_slots;
> > > +};
> > > +
> > > +struct msm_dp_mst {
> > > + bool mst_initialized;
> > > + struct drm_dp_mst_topology_mgr mst_mgr;
> > > + struct msm_dp_mst_bridge *mst_bridge;
> > > + struct msm_dp *msm_dp;
> > > + bool mst_session_hpd_state;
> > > + u32 max_streams;
> > > +};
> > > +
> > > +struct msm_dp_mst_connector {
> > > + struct drm_connector connector;
> > > + struct drm_dp_mst_port *mst_port;
> > > + struct msm_dp *msm_dp;
> > > + struct msm_dp_panel *dp_panel;
> > > +};
> > > +
> > > +#define to_msm_dp_mst_bridge(x) container_of((x), struct msm_dp_mst_bridge, base)
> > > +#define to_msm_dp_mst_bridge_priv(x) \
> > > + container_of((x), struct msm_dp_mst_bridge, obj)
> > > +#define to_msm_dp_mst_bridge_priv_state(x) \
> > > + container_of((x), struct msm_dp_mst_bridge_state, base)
> > > +#define to_msm_dp_mst_bridge_state(x) \
> > > + to_msm_dp_mst_bridge_priv_state((x)->obj.state)
> > > +#define to_msm_dp_mst_connector(x) \
> > > + container_of((x), struct msm_dp_mst_connector, connector)
> > > +int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
> > > +
> > > +#endif /* _DP_MST_DRM_H_ */
> > >
> > > --
> > > 2.34.1
> > >
> >
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* Re: [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations
2024-12-06 4:32 ` [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations Abhinav Kumar
2024-12-06 10:12 ` Dmitry Baryshkov
@ 2024-12-08 7:19 ` Dmitry Baryshkov
1 sibling, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 7:19 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:02PM -0800, Abhinav Kumar wrote:
> Add a new file dp_mst_drm to manage the DP MST bridge operations
> similar to the dp_drm file which manages the SST bridge operations.
> Each MST encoder creates one bridge and each bridge is bound to its
> own dp_panel abstraction to manage the operations of its pipeline.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/Makefile | 3 +-
> drivers/gpu/drm/msm/dp/dp_display.h | 2 +
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 490 ++++++++++++++++++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 102 ++++++++
> 4 files changed, 596 insertions(+), 1 deletion(-)
>
> +
> +struct msm_dp_mst_bridge {
> + struct drm_bridge base;
> + struct drm_private_obj obj;
it occured to me, while reviewing the next patch. Why do you need
another drm_private_obj???
> + u32 id;
> +
> + bool in_use;
> +
> + struct msm_dp *display;
> + struct drm_encoder *encoder;
> +
> + struct drm_display_mode drm_mode;
> + struct msm_dp_display_mode msm_dp_mode;
> + struct drm_connector *connector;
> + struct msm_dp_panel *msm_dp_panel;
> +
> + int vcpi;
> + int pbn;
> + int num_slots;
> + int start_slot;
> +};
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 32/45] drm/msm/dp: add connector abstraction for DP MST
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (30 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 31/45] drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 7:24 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 33/45] drm/msm/dp: add irq hpd callback for dp mst Abhinav Kumar
` (14 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Add connector abstraction for the DP MST. Each MST encoder
is connected through a DRM bridge to a MST connector and each
MST connector has a DP panel abstraction attached to it.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_mst_drm.c | 517 ++++++++++++++++++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_mst_drm.h | 4 +
2 files changed, 521 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
index e66bd1e565aeb4da3d636eb5f4aa75504d60fd40..2934fe6a93ff8b927b035c01b6007f3f4ec91a3f 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
@@ -26,6 +26,7 @@
* OF THIS SOFTWARE.
*/
+#include <drm/drm_edid.h>
#include "dp_mst_drm.h"
static struct drm_private_state *msm_dp_mst_duplicate_bridge_state(struct drm_private_obj *obj)
@@ -80,6 +81,64 @@ static int msm_dp_mst_find_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr, int p
return num_slots;
}
+static int msm_dp_mst_get_mst_pbn_div(struct msm_dp_panel *msm_dp_panel)
+{
+ struct msm_dp_link_info *link_info;
+
+ link_info = &msm_dp_panel->link_info;
+
+ return link_info->rate * link_info->num_lanes / 54000;
+}
+
+static int _msm_dp_mst_compute_config(struct drm_atomic_state *state,
+ struct msm_dp_mst *mst, struct drm_connector *connector,
+ struct drm_display_mode *mode)
+{
+ int slots = 0, pbn;
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+ int rc = 0;
+ struct drm_dp_mst_topology_state *mst_state;
+ int pbn_div;
+ struct msm_dp *dp_display = mst->msm_dp;
+ u32 bpp;
+
+ bpp = connector->display_info.bpc * 3;
+ //default to 24
+ if (!bpp)
+ bpp = 24;
+
+ pbn = drm_dp_calc_pbn_mode(mode->clock, bpp << 4);
+
+ mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
+
+ if (!dfixed_trunc(mst_state->pbn_div)) {
+ pbn_div = msm_dp_mst_get_mst_pbn_div(mst_conn->dp_panel);
+ mst_state->pbn_div.full = dfixed_const(pbn_div);
+ }
+
+ rc = drm_dp_atomic_find_time_slots(state, &mst->mst_mgr, mst_conn->mst_port, pbn);
+ if (rc < 0) {
+ DRM_ERROR("conn:%d failed to find vcpi slots. pbn:%d, rc:%d\n",
+ connector->base.id, pbn, rc);
+ goto end;
+ }
+
+ slots = rc;
+
+ rc = drm_dp_mst_atomic_check(state);
+ if (rc) {
+ DRM_ERROR("conn:%d mst atomic check failed: rc=%d\n", connector->base.id, rc);
+ slots = 0;
+ goto end;
+ }
+
+ drm_dbg_dp(dp_display->drm_dev, "conn:%d pbn:%d slots:%d rc:%d\n",
+ connector->base.id, pbn, slots, rc);
+
+end:
+ return (rc < 0 ? rc : slots);
+}
+
static void _msm_dp_mst_update_timeslots(struct msm_dp_mst *mst,
struct msm_dp_mst_bridge *mst_bridge,
struct drm_dp_mst_port *port)
@@ -488,3 +547,461 @@ int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder)
end:
return rc;
}
+
+static struct msm_dp_mst_bridge_state *msm_dp_mst_br_priv_state(struct drm_atomic_state *st,
+ struct msm_dp_mst_bridge *bridge)
+{
+ struct drm_device *dev = bridge->base.dev;
+
+ WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+
+ return to_msm_dp_mst_bridge_priv_state(drm_atomic_get_private_obj_state(st,
+ &bridge->obj));
+}
+
+/* DP MST Connector OPs */
+static int
+msm_dp_mst_connector_detect(struct drm_connector *connector,
+ struct drm_modeset_acquire_ctx *ctx,
+ bool force)
+{
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+ struct msm_dp *dp_display = mst_conn->msm_dp;
+ struct msm_dp_mst *mst = dp_display->msm_dp_mst;
+ enum drm_connector_status status = connector_status_disconnected;
+
+ if (dp_display->link_ready)
+ status = drm_dp_mst_detect_port(connector,
+ ctx, &mst->mst_mgr, mst_conn->mst_port);
+
+ drm_dbg_dp(dp_display->drm_dev, "conn:%d status:%d\n", connector->base.id, status);
+
+ return (int)status;
+}
+
+static int msm_dp_mst_connector_get_modes(struct drm_connector *connector)
+{
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+ struct msm_dp *dp_display = mst_conn->msm_dp;
+ struct msm_dp_mst *mst = dp_display->msm_dp_mst;
+ struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
+
+ drm_edid_free(dp_panel->drm_edid);
+
+ dp_panel->drm_edid = drm_dp_mst_edid_read(connector, &mst->mst_mgr, mst_conn->mst_port);
+ if (!dp_panel->drm_edid) {
+ DRM_ERROR("get edid failed. id: %d\n", connector->base.id);
+ return -EINVAL;
+ }
+
+ drm_edid_connector_update(connector, dp_panel->drm_edid);
+
+ return drm_edid_connector_add_modes(connector);
+}
+
+static enum drm_mode_status msm_dp_mst_connector_mode_valid(struct drm_connector *connector,
+ struct drm_display_mode *mode)
+{
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+ struct msm_dp *dp_display = mst_conn->msm_dp;
+ struct drm_dp_mst_port *mst_port;
+ struct msm_dp_panel *dp_panel;
+ struct msm_dp_mst *mst;
+ u16 full_pbn, required_pbn;
+ int available_slots, required_slots;
+ struct msm_dp_mst_bridge_state *dp_bridge_state;
+ int i, slots_in_use = 0, active_enc_cnt = 0;
+ const u32 tot_slots = 63;
+
+ if (!connector || !mode || !dp_display) {
+ DRM_ERROR("invalid input\n");
+ return 0;
+ }
+
+ mst = dp_display->msm_dp_mst;
+ mst_conn = to_msm_dp_mst_connector(connector);
+ mst_port = mst_conn->mst_port;
+ dp_panel = mst_conn->dp_panel;
+
+ if (!dp_panel || !mst_port)
+ return MODE_ERROR;
+
+ for (i = 0; i < mst->max_streams; i++) {
+ dp_bridge_state = to_msm_dp_mst_bridge_state(&mst->mst_bridge[i]);
+ if (dp_bridge_state->connector &&
+ dp_bridge_state->connector != connector) {
+ active_enc_cnt++;
+ slots_in_use += dp_bridge_state->num_slots;
+ }
+ }
+
+ if (active_enc_cnt < DP_STREAM_MAX) {
+ full_pbn = mst_port->full_pbn;
+ available_slots = tot_slots - slots_in_use;
+ } else {
+ DRM_ERROR("all mst streams are active\n");
+ return MODE_BAD;
+ }
+
+ required_pbn = drm_dp_calc_pbn_mode(mode->clock, (connector->display_info.bpc * 3) << 4);
+
+ required_slots = msm_dp_mst_find_vcpi_slots(&mst->mst_mgr, required_pbn);
+
+ if (required_pbn > full_pbn || required_slots > available_slots) {
+ drm_dbg_dp(dp_display->drm_dev,
+ "mode:%s not supported. pbn %d vs %d slots %d vs %d\n",
+ mode->name, required_pbn, full_pbn,
+ required_slots, available_slots);
+ return MODE_BAD;
+ }
+
+ return msm_dp_display_mode_valid(dp_display, &dp_display->connector->display_info, mode);
+}
+
+static struct drm_encoder *
+msm_dp_mst_atomic_best_encoder(struct drm_connector *connector, struct drm_atomic_state *state)
+{
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+ struct msm_dp *dp_display = mst_conn->msm_dp;
+ struct msm_dp_mst *mst = dp_display->msm_dp_mst;
+ struct drm_encoder *enc = NULL;
+ struct msm_dp_mst_bridge_state *bridge_state;
+ u32 i;
+ struct drm_connector_state *conn_state = drm_atomic_get_new_connector_state(state,
+ connector);
+
+ if (conn_state && conn_state->best_encoder)
+ return conn_state->best_encoder;
+
+ for (i = 0; i < mst->max_streams; i++) {
+ bridge_state = msm_dp_mst_br_priv_state(state, &mst->mst_bridge[i]);
+ if (IS_ERR(bridge_state))
+ goto end;
+
+ if (bridge_state->connector == connector) {
+ enc = mst->mst_bridge[i].encoder;
+ goto end;
+ }
+ }
+
+ for (i = 0; i < mst->max_streams; i++) {
+ bridge_state = msm_dp_mst_br_priv_state(state, &mst->mst_bridge[i]);
+
+ if (!bridge_state->connector) {
+ bridge_state->connector = connector;
+ bridge_state->msm_dp_panel = mst_conn->dp_panel;
+ enc = mst->mst_bridge[i].encoder;
+ break;
+ }
+ }
+
+end:
+ if (enc)
+ drm_dbg_dp(dp_display->drm_dev, "mst connector:%d atomic best encoder:%d\n",
+ connector->base.id, i);
+ else
+ drm_dbg_dp(dp_display->drm_dev, "mst connector:%d atomic best encoder failed\n",
+ connector->base.id);
+
+ return enc;
+}
+
+static int msm_dp_mst_connector_atomic_check(struct drm_connector *connector,
+ struct drm_atomic_state *state)
+{
+ int rc = 0, slots, i;
+ bool vcpi_released = false;
+ struct drm_connector_state *old_conn_state;
+ struct drm_connector_state *new_conn_state;
+ struct drm_crtc *old_crtc;
+ struct drm_crtc_state *crtc_state;
+ struct msm_dp_mst_bridge *bridge;
+ struct msm_dp_mst_bridge_state *bridge_state;
+ struct drm_bridge *drm_bridge;
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+ struct msm_dp *dp_display = mst_conn->msm_dp;
+ struct msm_dp_mst *mst = dp_display->msm_dp_mst;
+
+ if (!state)
+ return rc;
+
+ new_conn_state = drm_atomic_get_new_connector_state(state, connector);
+ if (!new_conn_state)
+ return rc;
+
+ old_conn_state = drm_atomic_get_old_connector_state(state, connector);
+ if (!old_conn_state)
+ goto mode_set;
+
+ old_crtc = old_conn_state->crtc;
+ if (!old_crtc)
+ goto mode_set;
+
+ crtc_state = drm_atomic_get_new_crtc_state(state, old_crtc);
+
+ for (i = 0; i < mst->max_streams; i++) {
+ bridge = &mst->mst_bridge[i];
+ drm_dbg_dp(dp_display->drm_dev, "bridge id:%d, vcpi:%d, pbn:%d, slots:%d\n",
+ bridge->id, bridge->vcpi, bridge->pbn,
+ bridge->num_slots);
+ }
+
+ /*attempt to release vcpi slots on a modeset change for crtc state*/
+ if (drm_atomic_crtc_needs_modeset(crtc_state)) {
+ if (WARN_ON(!old_conn_state->best_encoder)) {
+ rc = -EINVAL;
+ goto end;
+ }
+
+ drm_bridge = drm_bridge_chain_get_first_bridge(old_conn_state->best_encoder);
+ if (WARN_ON(!drm_bridge)) {
+ rc = -EINVAL;
+ goto end;
+ }
+ bridge = to_msm_dp_mst_bridge(drm_bridge);
+
+ bridge_state = msm_dp_mst_br_priv_state(state, bridge);
+ if (IS_ERR(bridge_state)) {
+ rc = PTR_ERR(bridge_state);
+ goto end;
+ }
+
+ if (WARN_ON(bridge_state->connector != connector)) {
+ rc = -EINVAL;
+ goto end;
+ }
+
+ slots = bridge_state->num_slots;
+ if (slots > 0) {
+ rc = drm_dp_atomic_release_time_slots(state,
+ &mst->mst_mgr,
+ mst_conn->mst_port);
+ if (rc) {
+ DRM_ERROR("failed releasing %d vcpi slots %d\n", slots, rc);
+ goto end;
+ }
+ vcpi_released = true;
+ }
+
+ if (!new_conn_state->crtc) {
+ /* for cases where crtc is not disabled the slots are not
+ * freed by drm_dp_atomic_release_time_slots. this results
+ * in subsequent atomic_check failing since internal slots
+ * were freed but not the dp mst mgr's
+ */
+ bridge_state->num_slots = 0;
+ bridge_state->connector = NULL;
+ bridge_state->msm_dp_panel = NULL;
+
+ drm_dbg_dp(dp_display->drm_dev, "clear best encoder: %d\n", bridge->id);
+ }
+ }
+
+mode_set:
+ if (!new_conn_state->crtc)
+ goto end;
+
+ crtc_state = drm_atomic_get_new_crtc_state(state, new_conn_state->crtc);
+
+ if (drm_atomic_crtc_needs_modeset(crtc_state) && crtc_state->active) {
+ if (WARN_ON(!new_conn_state->best_encoder)) {
+ rc = -EINVAL;
+ goto end;
+ }
+
+ drm_bridge = drm_bridge_chain_get_first_bridge(new_conn_state->best_encoder);
+ if (WARN_ON(!drm_bridge)) {
+ rc = -EINVAL;
+ goto end;
+ }
+ bridge = to_msm_dp_mst_bridge(drm_bridge);
+
+ bridge_state = msm_dp_mst_br_priv_state(state, bridge);
+ if (IS_ERR(bridge_state)) {
+ rc = PTR_ERR(bridge_state);
+ goto end;
+ }
+
+ if (WARN_ON(bridge_state->connector != connector)) {
+ rc = -EINVAL;
+ goto end;
+ }
+
+ /*
+ * check if vcpi slots are trying to get allocated in same phase
+ * as deallocation. If so, go to end to avoid allocation.
+ */
+ if (vcpi_released) {
+ drm_dbg_dp(dp_display->drm_dev,
+ "skipping allocation since vcpi was released in the same state\n");
+ goto end;
+ }
+
+ if (WARN_ON(bridge_state->num_slots)) {
+ rc = -EINVAL;
+ goto end;
+ }
+
+ slots = _msm_dp_mst_compute_config(state, mst, connector, &crtc_state->mode);
+ if (slots < 0) {
+ rc = slots;
+ goto end;
+ }
+
+ bridge_state->num_slots = slots;
+ }
+
+end:
+ drm_dbg_dp(dp_display->drm_dev, "mst connector:%d atomic check ret %d\n",
+ connector->base.id, rc);
+ return rc;
+}
+
+static void dp_mst_connector_destroy(struct drm_connector *connector)
+{
+ struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+
+ drm_connector_cleanup(connector);
+ drm_dp_mst_put_port_malloc(mst_conn->mst_port);
+ msm_dp_panel_put(mst_conn->dp_panel);
+}
+
+/* DRM MST callbacks */
+static const struct drm_connector_helper_funcs msm_dp_drm_mst_connector_helper_funcs = {
+ .get_modes = msm_dp_mst_connector_get_modes,
+ .detect_ctx = msm_dp_mst_connector_detect,
+ .mode_valid = msm_dp_mst_connector_mode_valid,
+ .atomic_best_encoder = msm_dp_mst_atomic_best_encoder,
+ .atomic_check = msm_dp_mst_connector_atomic_check,
+};
+
+static const struct drm_connector_funcs msm_dp_drm_mst_connector_funcs = {
+ .reset = drm_atomic_helper_connector_reset,
+ .destroy = dp_mst_connector_destroy,
+ .fill_modes = drm_helper_probe_single_connector_modes,
+ .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+ .atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+};
+
+static struct drm_connector *
+msm_dp_mst_add_connector(struct drm_dp_mst_topology_mgr *mgr,
+ struct drm_dp_mst_port *port, const char *pathprop)
+{
+ struct msm_dp_mst *dp_mst;
+ struct drm_device *dev;
+ struct msm_dp *dp_display;
+ struct msm_dp_mst_connector *mst_connector;
+ struct drm_connector *connector;
+ int rc, i;
+
+ dp_mst = container_of(mgr, struct msm_dp_mst, mst_mgr);
+
+ dp_display = dp_mst->msm_dp;
+ dev = dp_display->drm_dev;
+
+ mst_connector = devm_kzalloc(dev->dev, sizeof(*mst_connector), GFP_KERNEL);
+
+ drm_modeset_lock_all(dev);
+
+ rc = drm_connector_init(dev, &mst_connector->connector, &msm_dp_drm_mst_connector_funcs,
+ DRM_MODE_CONNECTOR_DisplayPort);
+ if (rc) {
+ drm_modeset_unlock_all(dev);
+ return NULL;
+ }
+
+ mst_connector->dp_panel = msm_dp_display_get_panel(dp_display);
+ if (!mst_connector->dp_panel) {
+ DRM_ERROR("failed to get dp_panel for connector\n");
+ drm_modeset_unlock_all(dev);
+ return NULL;
+ }
+
+ mst_connector->dp_panel->connector = &mst_connector->connector;
+ mst_connector->msm_dp = dp_display;
+ connector = &mst_connector->connector;
+ drm_connector_helper_add(&mst_connector->connector, &msm_dp_drm_mst_connector_helper_funcs);
+
+ if (connector->funcs->reset)
+ connector->funcs->reset(connector);
+
+ /* add all encoders as possible encoders */
+ for (i = 0; i < dp_mst->max_streams; i++) {
+ rc = drm_connector_attach_encoder(&mst_connector->connector,
+ dp_mst->mst_bridge[i].encoder);
+ if (rc) {
+ DRM_ERROR("failed to attach encoder to connector, %d\n", rc);
+ drm_modeset_unlock_all(dev);
+ return NULL;
+ }
+ }
+
+ mst_connector->mst_port = port;
+ drm_dp_mst_get_port_malloc(mst_connector->mst_port);
+
+ drm_object_attach_property(&mst_connector->connector.base,
+ dev->mode_config.path_property, 0);
+ drm_object_attach_property(&mst_connector->connector.base,
+ dev->mode_config.tile_property, 0);
+
+ drm_modeset_unlock_all(dev);
+
+ drm_dbg_dp(dp_display->drm_dev, "add mst connector id:%d\n",
+ mst_connector->connector.base.id);
+
+ return &mst_connector->connector;
+}
+
+static const struct drm_dp_mst_topology_cbs msm_dp_mst_drm_cbs = {
+ .add_connector = msm_dp_mst_add_connector,
+};
+
+int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams, u32 max_dpcd_transaction_bytes,
+ struct drm_dp_aux *drm_aux)
+{
+ struct drm_device *dev;
+ int conn_base_id = 0;
+ int ret;
+ struct msm_dp_mst *msm_dp_mst;
+
+ if (!dp_display) {
+ DRM_ERROR("invalid params\n");
+ return 0;
+ }
+
+ dev = dp_display->drm_dev;
+
+ msm_dp_mst = devm_kzalloc(dev->dev, sizeof(*msm_dp_mst), GFP_KERNEL);
+ if (!msm_dp_mst)
+ return -ENOMEM;
+
+ memset(&msm_dp_mst->mst_mgr, 0, sizeof(msm_dp_mst->mst_mgr));
+ msm_dp_mst->mst_mgr.cbs = &msm_dp_mst_drm_cbs;
+ conn_base_id = dp_display->connector->base.id;
+ msm_dp_mst->msm_dp = dp_display;
+ msm_dp_mst->max_streams = max_streams;
+
+ msm_dp_mst->mst_bridge = devm_kzalloc(dev->dev,
+ max_streams * sizeof(struct msm_dp_mst_bridge),
+ GFP_KERNEL);
+
+ msm_dp_mst->dp_aux = drm_aux;
+
+ ret = drm_dp_mst_topology_mgr_init(&msm_dp_mst->mst_mgr, dev,
+ drm_aux,
+ max_dpcd_transaction_bytes,
+ max_streams,
+ conn_base_id);
+ if (ret) {
+ DRM_ERROR("dp drm mst topology manager init failed\n");
+ return ret;
+ }
+
+ dp_display->msm_dp_mst = msm_dp_mst;
+
+ msm_dp_mst->mst_initialized = true;
+
+ drm_dbg_dp(dp_display->drm_dev, "dp drm mst topology manager init completed\n");
+
+ return ret;
+}
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
index c2a82cd3c6d6e1951a8e5905d3aa39dfc691023b..9a8535b97193d911592b44fdfcf43ca2f252cea0 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
@@ -77,6 +77,7 @@ struct msm_dp_mst {
struct drm_dp_mst_topology_mgr mst_mgr;
struct msm_dp_mst_bridge *mst_bridge;
struct msm_dp *msm_dp;
+ struct drm_dp_aux *dp_aux;
bool mst_session_hpd_state;
u32 max_streams;
};
@@ -99,4 +100,7 @@ struct msm_dp_mst_connector {
container_of((x), struct msm_dp_mst_connector, connector)
int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
+int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams,
+ u32 max_dpcd_transaction_bytes, struct drm_dp_aux *drm_aux);
+
#endif /* _DP_MST_DRM_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 32/45] drm/msm/dp: add connector abstraction for DP MST
2024-12-06 4:32 ` [PATCH 32/45] drm/msm/dp: add connector abstraction for DP MST Abhinav Kumar
@ 2024-12-08 7:24 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 7:24 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:03PM -0800, Abhinav Kumar wrote:
> Add connector abstraction for the DP MST. Each MST encoder
> is connected through a DRM bridge to a MST connector and each
> MST connector has a DP panel abstraction attached to it.
Why do we need an extra DRM bridge? Can we merge its functionality into
the DRM MST connector?
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 517 ++++++++++++++++++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 4 +
> 2 files changed, 521 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> index e66bd1e565aeb4da3d636eb5f4aa75504d60fd40..2934fe6a93ff8b927b035c01b6007f3f4ec91a3f 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> @@ -26,6 +26,7 @@
> * OF THIS SOFTWARE.
> */
>
> +#include <drm/drm_edid.h>
> #include "dp_mst_drm.h"
>
> static struct drm_private_state *msm_dp_mst_duplicate_bridge_state(struct drm_private_obj *obj)
> @@ -80,6 +81,64 @@ static int msm_dp_mst_find_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr, int p
> return num_slots;
> }
>
> +static int msm_dp_mst_get_mst_pbn_div(struct msm_dp_panel *msm_dp_panel)
> +{
> + struct msm_dp_link_info *link_info;
> +
> + link_info = &msm_dp_panel->link_info;
> +
> + return link_info->rate * link_info->num_lanes / 54000;
> +}
> +
> +static int _msm_dp_mst_compute_config(struct drm_atomic_state *state,
> + struct msm_dp_mst *mst, struct drm_connector *connector,
> + struct drm_display_mode *mode)
No single underscore functions, please.
> +{
> + int slots = 0, pbn;
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> + int rc = 0;
> + struct drm_dp_mst_topology_state *mst_state;
> + int pbn_div;
> + struct msm_dp *dp_display = mst->msm_dp;
> + u32 bpp;
> +
> + bpp = connector->display_info.bpc * 3;
> + //default to 24
> + if (!bpp)
> + bpp = 24;
> +
> + pbn = drm_dp_calc_pbn_mode(mode->clock, bpp << 4);
> +
> + mst_state = to_drm_dp_mst_topology_state(mst->mst_mgr.base.state);
Here from atomic_check() you are accessing existing state. That doesn't
sound correct. Please take a look at i915 driver.
> +
> + if (!dfixed_trunc(mst_state->pbn_div)) {
> + pbn_div = msm_dp_mst_get_mst_pbn_div(mst_conn->dp_panel);
> + mst_state->pbn_div.full = dfixed_const(pbn_div);
> + }
> +
> + rc = drm_dp_atomic_find_time_slots(state, &mst->mst_mgr, mst_conn->mst_port, pbn);
> + if (rc < 0) {
> + DRM_ERROR("conn:%d failed to find vcpi slots. pbn:%d, rc:%d\n",
> + connector->base.id, pbn, rc);
> + goto end;
> + }
> +
> + slots = rc;
> +
> + rc = drm_dp_mst_atomic_check(state);
> + if (rc) {
> + DRM_ERROR("conn:%d mst atomic check failed: rc=%d\n", connector->base.id, rc);
> + slots = 0;
> + goto end;
> + }
> +
> + drm_dbg_dp(dp_display->drm_dev, "conn:%d pbn:%d slots:%d rc:%d\n",
> + connector->base.id, pbn, slots, rc);
> +
> +end:
> + return (rc < 0 ? rc : slots);
> +}
> +
> static void _msm_dp_mst_update_timeslots(struct msm_dp_mst *mst,
> struct msm_dp_mst_bridge *mst_bridge,
> struct drm_dp_mst_port *port)
> @@ -488,3 +547,461 @@ int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder)
> end:
> return rc;
> }
> +
> +static struct msm_dp_mst_bridge_state *msm_dp_mst_br_priv_state(struct drm_atomic_state *st,
> + struct msm_dp_mst_bridge *bridge)
> +{
> + struct drm_device *dev = bridge->base.dev;
> +
> + WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
> +
> + return to_msm_dp_mst_bridge_priv_state(drm_atomic_get_private_obj_state(st,
> + &bridge->obj));
drm_atomic_get_bridge_state() NIH?
> +}
> +
> +/* DP MST Connector OPs */
> +static int
> +msm_dp_mst_connector_detect(struct drm_connector *connector,
> + struct drm_modeset_acquire_ctx *ctx,
> + bool force)
> +{
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> + struct msm_dp *dp_display = mst_conn->msm_dp;
> + struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> + enum drm_connector_status status = connector_status_disconnected;
> +
Missing check for unregistered connector.
> + if (dp_display->link_ready)
I really want to get rid of link_ready checks in the detect paths.
Please refactor the HPD / detect handling to report actual plug status
rather than link being ready.
> + status = drm_dp_mst_detect_port(connector,
> + ctx, &mst->mst_mgr, mst_conn->mst_port);
> +
> + drm_dbg_dp(dp_display->drm_dev, "conn:%d status:%d\n", connector->base.id, status);
> +
> + return (int)status;
No need for typecasting. enum is int underneath. Also just return
drm_dp_mst_detect_port() as everybody else does.
> +}
> +
> +static int msm_dp_mst_connector_get_modes(struct drm_connector *connector)
> +{
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> + struct msm_dp *dp_display = mst_conn->msm_dp;
> + struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> + struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
Missing unregistered check
> +
> + drm_edid_free(dp_panel->drm_edid);
> +
> + dp_panel->drm_edid = drm_dp_mst_edid_read(connector, &mst->mst_mgr, mst_conn->mst_port);
> + if (!dp_panel->drm_edid) {
Stop caching the EDID. Let it go.
> + DRM_ERROR("get edid failed. id: %d\n", connector->base.id);
> + return -EINVAL;
> + }
> +
> + drm_edid_connector_update(connector, dp_panel->drm_edid);
> +
> + return drm_edid_connector_add_modes(connector);
> +}
> +
> +static enum drm_mode_status msm_dp_mst_connector_mode_valid(struct drm_connector *connector,
> + struct drm_display_mode *mode)
> +{
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> + struct msm_dp *dp_display = mst_conn->msm_dp;
> + struct drm_dp_mst_port *mst_port;
> + struct msm_dp_panel *dp_panel;
> + struct msm_dp_mst *mst;
> + u16 full_pbn, required_pbn;
> + int available_slots, required_slots;
> + struct msm_dp_mst_bridge_state *dp_bridge_state;
> + int i, slots_in_use = 0, active_enc_cnt = 0;
> + const u32 tot_slots = 63;
Missing unregistered check
> +
> + if (!connector || !mode || !dp_display) {
> + DRM_ERROR("invalid input\n");
> + return 0;
> + }
Useless, please drop
> +
> + mst = dp_display->msm_dp_mst;
> + mst_conn = to_msm_dp_mst_connector(connector);
> + mst_port = mst_conn->mst_port;
> + dp_panel = mst_conn->dp_panel;
> +
> + if (!dp_panel || !mst_port)
> + return MODE_ERROR;
> +
> + for (i = 0; i < mst->max_streams; i++) {
> + dp_bridge_state = to_msm_dp_mst_bridge_state(&mst->mst_bridge[i]);
Can we just order them by stream_id?
> + if (dp_bridge_state->connector &&
> + dp_bridge_state->connector != connector) {
> + active_enc_cnt++;
> + slots_in_use += dp_bridge_state->num_slots;
> + }
> + }
> +
> + if (active_enc_cnt < DP_STREAM_MAX) {
> + full_pbn = mst_port->full_pbn;
> + available_slots = tot_slots - slots_in_use;
> + } else {
> + DRM_ERROR("all mst streams are active\n");
No! mode_valid() checks if the mode is valid at all. It should perform
the same checks as an SST controller is doing. None of the PBN / slots
checks belong here. Intel driver does slightly different check: it
verifies that the mode can fit at all, not that it can fit into the
available slots. And there is a very simple logic for that: the modes
for a connector are not revalidated if another connector changes state
or mode.
Also, if we think about it, how are you accessing the current slots
state without taking the modesetting lock? What if the underlying object
changes while you are running mode_valid() checks? Note, i915 driver
uses different callback and it also uses the context.
> + return MODE_BAD;
> + }
> +
> + required_pbn = drm_dp_calc_pbn_mode(mode->clock, (connector->display_info.bpc * 3) << 4);
> +
> + required_slots = msm_dp_mst_find_vcpi_slots(&mst->mst_mgr, required_pbn);
> +
> + if (required_pbn > full_pbn || required_slots > available_slots) {
> + drm_dbg_dp(dp_display->drm_dev,
> + "mode:%s not supported. pbn %d vs %d slots %d vs %d\n",
> + mode->name, required_pbn, full_pbn,
> + required_slots, available_slots);
> + return MODE_BAD;
> + }
> +
> + return msm_dp_display_mode_valid(dp_display, &dp_display->connector->display_info, mode);
> +}
> +
> +static struct drm_encoder *
> +msm_dp_mst_atomic_best_encoder(struct drm_connector *connector, struct drm_atomic_state *state)
> +{
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> + struct msm_dp *dp_display = mst_conn->msm_dp;
> + struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> + struct drm_encoder *enc = NULL;
> + struct msm_dp_mst_bridge_state *bridge_state;
> + u32 i;
> + struct drm_connector_state *conn_state = drm_atomic_get_new_connector_state(state,
> + connector);
> +
> + if (conn_state && conn_state->best_encoder)
> + return conn_state->best_encoder;
> +
> + for (i = 0; i < mst->max_streams; i++) {
> + bridge_state = msm_dp_mst_br_priv_state(state, &mst->mst_bridge[i]);
> + if (IS_ERR(bridge_state))
> + goto end;
> +
> + if (bridge_state->connector == connector) {
> + enc = mst->mst_bridge[i].encoder;
> + goto end;
> + }
> + }
> +
> + for (i = 0; i < mst->max_streams; i++) {
> + bridge_state = msm_dp_mst_br_priv_state(state, &mst->mst_bridge[i]);
> +
mst_bridges should be ordered by stream_id, save you from the loop, etc.
And, I think, allows us to drop the possibly-lingering pointer to the
drm_connector.
> + if (!bridge_state->connector) {
> + bridge_state->connector = connector;
> + bridge_state->msm_dp_panel = mst_conn->dp_panel;
> + enc = mst->mst_bridge[i].encoder;
> + break;
> + }
> + }
> +
> +end:
> + if (enc)
> + drm_dbg_dp(dp_display->drm_dev, "mst connector:%d atomic best encoder:%d\n",
> + connector->base.id, i);
> + else
> + drm_dbg_dp(dp_display->drm_dev, "mst connector:%d atomic best encoder failed\n",
> + connector->base.id);
> +
> + return enc;
> +}
> +
> +static int msm_dp_mst_connector_atomic_check(struct drm_connector *connector,
> + struct drm_atomic_state *state)
> +{
> + int rc = 0, slots, i;
> + bool vcpi_released = false;
> + struct drm_connector_state *old_conn_state;
> + struct drm_connector_state *new_conn_state;
> + struct drm_crtc *old_crtc;
> + struct drm_crtc_state *crtc_state;
> + struct msm_dp_mst_bridge *bridge;
> + struct msm_dp_mst_bridge_state *bridge_state;
> + struct drm_bridge *drm_bridge;
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> + struct msm_dp *dp_display = mst_conn->msm_dp;
> + struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> +
> + if (!state)
> + return rc;
> +
> + new_conn_state = drm_atomic_get_new_connector_state(state, connector);
> + if (!new_conn_state)
> + return rc;
> +
> + old_conn_state = drm_atomic_get_old_connector_state(state, connector);
> + if (!old_conn_state)
> + goto mode_set;
> +
> + old_crtc = old_conn_state->crtc;
> + if (!old_crtc)
> + goto mode_set;
> +
> + crtc_state = drm_atomic_get_new_crtc_state(state, old_crtc);
> +
> + for (i = 0; i < mst->max_streams; i++) {
> + bridge = &mst->mst_bridge[i];
> + drm_dbg_dp(dp_display->drm_dev, "bridge id:%d, vcpi:%d, pbn:%d, slots:%d\n",
> + bridge->id, bridge->vcpi, bridge->pbn,
> + bridge->num_slots);
> + }
> +
> + /*attempt to release vcpi slots on a modeset change for crtc state*/
> + if (drm_atomic_crtc_needs_modeset(crtc_state)) {
> + if (WARN_ON(!old_conn_state->best_encoder)) {
> + rc = -EINVAL;
> + goto end;
> + }
> +
> + drm_bridge = drm_bridge_chain_get_first_bridge(old_conn_state->best_encoder);
> + if (WARN_ON(!drm_bridge)) {
> + rc = -EINVAL;
> + goto end;
> + }
> + bridge = to_msm_dp_mst_bridge(drm_bridge);
> +
> + bridge_state = msm_dp_mst_br_priv_state(state, bridge);
> + if (IS_ERR(bridge_state)) {
> + rc = PTR_ERR(bridge_state);
> + goto end;
> + }
> +
> + if (WARN_ON(bridge_state->connector != connector)) {
> + rc = -EINVAL;
> + goto end;
> + }
> +
> + slots = bridge_state->num_slots;
> + if (slots > 0) {
> + rc = drm_dp_atomic_release_time_slots(state,
> + &mst->mst_mgr,
> + mst_conn->mst_port);
> + if (rc) {
> + DRM_ERROR("failed releasing %d vcpi slots %d\n", slots, rc);
> + goto end;
> + }
> + vcpi_released = true;
> + }
> +
> + if (!new_conn_state->crtc) {
> + /* for cases where crtc is not disabled the slots are not
> + * freed by drm_dp_atomic_release_time_slots. this results
> + * in subsequent atomic_check failing since internal slots
> + * were freed but not the dp mst mgr's
> + */
Huh?
> + bridge_state->num_slots = 0;
> + bridge_state->connector = NULL;
> + bridge_state->msm_dp_panel = NULL;
> +
> + drm_dbg_dp(dp_display->drm_dev, "clear best encoder: %d\n", bridge->id);
> + }
> + }
> +
> +mode_set:
> + if (!new_conn_state->crtc)
> + goto end;
> +
> + crtc_state = drm_atomic_get_new_crtc_state(state, new_conn_state->crtc);
> +
> + if (drm_atomic_crtc_needs_modeset(crtc_state) && crtc_state->active) {
> + if (WARN_ON(!new_conn_state->best_encoder)) {
> + rc = -EINVAL;
> + goto end;
> + }
> +
> + drm_bridge = drm_bridge_chain_get_first_bridge(new_conn_state->best_encoder);
> + if (WARN_ON(!drm_bridge)) {
> + rc = -EINVAL;
> + goto end;
> + }
> + bridge = to_msm_dp_mst_bridge(drm_bridge);
> +
> + bridge_state = msm_dp_mst_br_priv_state(state, bridge);
> + if (IS_ERR(bridge_state)) {
> + rc = PTR_ERR(bridge_state);
> + goto end;
> + }
> +
> + if (WARN_ON(bridge_state->connector != connector)) {
> + rc = -EINVAL;
> + goto end;
> + }
> +
> + /*
> + * check if vcpi slots are trying to get allocated in same phase
> + * as deallocation. If so, go to end to avoid allocation.
> + */
I don't see this in any other driver. Do you have a pointer?
> + if (vcpi_released) {
> + drm_dbg_dp(dp_display->drm_dev,
> + "skipping allocation since vcpi was released in the same state\n");
> + goto end;
> + }
> +
> + if (WARN_ON(bridge_state->num_slots)) {
> + rc = -EINVAL;
> + goto end;
> + }
> +
> + slots = _msm_dp_mst_compute_config(state, mst, connector, &crtc_state->mode);
> + if (slots < 0) {
> + rc = slots;
> + goto end;
> + }
> +
> + bridge_state->num_slots = slots;
> + }
> +
> +end:
> + drm_dbg_dp(dp_display->drm_dev, "mst connector:%d atomic check ret %d\n",
> + connector->base.id, rc);
> + return rc;
> +}
> +
> +static void dp_mst_connector_destroy(struct drm_connector *connector)
> +{
> + struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> +
> + drm_connector_cleanup(connector);
> + drm_dp_mst_put_port_malloc(mst_conn->mst_port);
> + msm_dp_panel_put(mst_conn->dp_panel);
> +}
> +
> +/* DRM MST callbacks */
> +static const struct drm_connector_helper_funcs msm_dp_drm_mst_connector_helper_funcs = {
> + .get_modes = msm_dp_mst_connector_get_modes,
> + .detect_ctx = msm_dp_mst_connector_detect,
> + .mode_valid = msm_dp_mst_connector_mode_valid,
> + .atomic_best_encoder = msm_dp_mst_atomic_best_encoder,
> + .atomic_check = msm_dp_mst_connector_atomic_check,
> +};
> +
> +static const struct drm_connector_funcs msm_dp_drm_mst_connector_funcs = {
> + .reset = drm_atomic_helper_connector_reset,
> + .destroy = dp_mst_connector_destroy,
> + .fill_modes = drm_helper_probe_single_connector_modes,
> + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
> + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
> +};
> +
> +static struct drm_connector *
> +msm_dp_mst_add_connector(struct drm_dp_mst_topology_mgr *mgr,
> + struct drm_dp_mst_port *port, const char *pathprop)
> +{
> + struct msm_dp_mst *dp_mst;
> + struct drm_device *dev;
> + struct msm_dp *dp_display;
> + struct msm_dp_mst_connector *mst_connector;
> + struct drm_connector *connector;
> + int rc, i;
> +
> + dp_mst = container_of(mgr, struct msm_dp_mst, mst_mgr);
> +
> + dp_display = dp_mst->msm_dp;
> + dev = dp_display->drm_dev;
> +
> + mst_connector = devm_kzalloc(dev->dev, sizeof(*mst_connector), GFP_KERNEL);
> +
> + drm_modeset_lock_all(dev);
> +
> + rc = drm_connector_init(dev, &mst_connector->connector, &msm_dp_drm_mst_connector_funcs,
> + DRM_MODE_CONNECTOR_DisplayPort);
> + if (rc) {
> + drm_modeset_unlock_all(dev);
> + return NULL;
> + }
> +
> + mst_connector->dp_panel = msm_dp_display_get_panel(dp_display);
> + if (!mst_connector->dp_panel) {
> + DRM_ERROR("failed to get dp_panel for connector\n");
> + drm_modeset_unlock_all(dev);
> + return NULL;
> + }
> +
> + mst_connector->dp_panel->connector = &mst_connector->connector;
> + mst_connector->msm_dp = dp_display;
> + connector = &mst_connector->connector;
> + drm_connector_helper_add(&mst_connector->connector, &msm_dp_drm_mst_connector_helper_funcs);
> +
> + if (connector->funcs->reset)
> + connector->funcs->reset(connector);
> +
> + /* add all encoders as possible encoders */
> + for (i = 0; i < dp_mst->max_streams; i++) {
> + rc = drm_connector_attach_encoder(&mst_connector->connector,
> + dp_mst->mst_bridge[i].encoder);
> + if (rc) {
> + DRM_ERROR("failed to attach encoder to connector, %d\n", rc);
> + drm_modeset_unlock_all(dev);
> + return NULL;
> + }
> + }
> +
> + mst_connector->mst_port = port;
> + drm_dp_mst_get_port_malloc(mst_connector->mst_port);
> +
> + drm_object_attach_property(&mst_connector->connector.base,
> + dev->mode_config.path_property, 0);
> + drm_object_attach_property(&mst_connector->connector.base,
> + dev->mode_config.tile_property, 0);
> +
> + drm_modeset_unlock_all(dev);
> +
> + drm_dbg_dp(dp_display->drm_dev, "add mst connector id:%d\n",
> + mst_connector->connector.base.id);
> +
> + return &mst_connector->connector;
> +}
> +
> +static const struct drm_dp_mst_topology_cbs msm_dp_mst_drm_cbs = {
> + .add_connector = msm_dp_mst_add_connector,
> +};
> +
> +int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams, u32 max_dpcd_transaction_bytes,
> + struct drm_dp_aux *drm_aux)
> +{
> + struct drm_device *dev;
> + int conn_base_id = 0;
> + int ret;
> + struct msm_dp_mst *msm_dp_mst;
> +
> + if (!dp_display) {
> + DRM_ERROR("invalid params\n");
> + return 0;
> + }
> +
> + dev = dp_display->drm_dev;
> +
> + msm_dp_mst = devm_kzalloc(dev->dev, sizeof(*msm_dp_mst), GFP_KERNEL);
> + if (!msm_dp_mst)
> + return -ENOMEM;
> +
> + memset(&msm_dp_mst->mst_mgr, 0, sizeof(msm_dp_mst->mst_mgr));
> + msm_dp_mst->mst_mgr.cbs = &msm_dp_mst_drm_cbs;
> + conn_base_id = dp_display->connector->base.id;
> + msm_dp_mst->msm_dp = dp_display;
> + msm_dp_mst->max_streams = max_streams;
> +
> + msm_dp_mst->mst_bridge = devm_kzalloc(dev->dev,
devm_ seems incorrect here. drmm_ sounds more appropriate.
> + max_streams * sizeof(struct msm_dp_mst_bridge),
> + GFP_KERNEL);
> +
> + msm_dp_mst->dp_aux = drm_aux;
> +
> + ret = drm_dp_mst_topology_mgr_init(&msm_dp_mst->mst_mgr, dev,
> + drm_aux,
> + max_dpcd_transaction_bytes,
> + max_streams,
> + conn_base_id);
> + if (ret) {
> + DRM_ERROR("dp drm mst topology manager init failed\n");
> + return ret;
> + }
> +
> + dp_display->msm_dp_mst = msm_dp_mst;
> +
> + msm_dp_mst->mst_initialized = true;
Can it exist in a non-initialized state?
> +
> + drm_dbg_dp(dp_display->drm_dev, "dp drm mst topology manager init completed\n");
> +
> + return ret;
> +}
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> index c2a82cd3c6d6e1951a8e5905d3aa39dfc691023b..9a8535b97193d911592b44fdfcf43ca2f252cea0 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> @@ -77,6 +77,7 @@ struct msm_dp_mst {
> struct drm_dp_mst_topology_mgr mst_mgr;
> struct msm_dp_mst_bridge *mst_bridge;
> struct msm_dp *msm_dp;
> + struct drm_dp_aux *dp_aux;
> bool mst_session_hpd_state;
> u32 max_streams;
> };
> @@ -99,4 +100,7 @@ struct msm_dp_mst_connector {
> container_of((x), struct msm_dp_mst_connector, connector)
> int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
>
> +int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams,
> + u32 max_dpcd_transaction_bytes, struct drm_dp_aux *drm_aux);
> +
> #endif /* _DP_MST_DRM_H_ */
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 33/45] drm/msm/dp: add irq hpd callback for dp mst
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (31 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 32/45] drm/msm/dp: add connector abstraction for DP MST Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 6:44 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 34/45] drm/msm/dp: add support to re-use and clear the panel edid Abhinav Kumar
` (13 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Add irq hpd callback for the dp mst module which shall be
invoked from the dp_display's irq hpd handler to perform
mst specific operations in case of irq hpd. In MST case, route
the IRQ HPD messages to MST module.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 17 +++++++++++++----
drivers/gpu/drm/msm/dp/dp_mst_drm.c | 36 +++++++++++++++++++++++++++++++++++-
drivers/gpu/drm/msm/dp/dp_mst_drm.h | 2 ++
3 files changed, 50 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 60316633cc6bcfdfad8ab356d803642d353add61..97f8228042773f51f23a9d39fc009de0798059d7 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -583,7 +583,9 @@ static int msm_dp_display_usbpd_attention_cb(struct device *dev)
{
int rc = 0;
u32 sink_request;
+
struct msm_dp_display_private *dp = dev_get_dp_display_private(dev);
+ struct msm_dp *msm_dp_display = &dp->msm_dp_display;
/* check for any test request issued by sink */
rc = msm_dp_link_process_request(dp->link);
@@ -591,12 +593,18 @@ static int msm_dp_display_usbpd_attention_cb(struct device *dev)
sink_request = dp->link->sink_request;
drm_dbg_dp(dp->drm_dev, "hpd_state=%d sink_request=%d\n",
dp->hpd_state, sink_request);
- if (sink_request & DS_PORT_STATUS_CHANGED)
- rc = msm_dp_display_handle_port_status_changed(dp);
- else
+ if (sink_request & DS_PORT_STATUS_CHANGED) {
+ if (!msm_dp_display->mst_active)
+ rc = msm_dp_display_handle_port_status_changed(dp);
+ } else {
rc = msm_dp_display_handle_irq_hpd(dp);
+ }
}
+ /* let MST specific IRQ events be handled by its callback */
+ if (msm_dp_display->mst_active)
+ msm_dp_mst_display_hpd_irq(&dp->msm_dp_display);
+
return rc;
}
@@ -736,7 +744,8 @@ static int msm_dp_irq_hpd_handle(struct msm_dp_display_private *dp, u32 data)
drm_dbg_dp(dp->drm_dev, "Before, type=%d hpd_state=%d\n",
dp->msm_dp_display.connector_type, state);
- if (state == ST_MAINLINK_READY || state == ST_DISCONNECT_PENDING) {
+ if ((state == ST_MAINLINK_READY || state == ST_DISCONNECT_PENDING) &&
+ !dp->msm_dp_display.mst_active) {
/* wait until ST_CONNECTED */
msm_dp_add_event(dp, EV_IRQ_HPD_INT, 0, 1); /* delay = 1 */
mutex_unlock(&dp->event_mutex);
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
index 2934fe6a93ff8b927b035c01b6007f3f4ec91a3f..2d92084586b466d4953429e8a6fbf766d081cb9f 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
@@ -559,6 +559,40 @@ static struct msm_dp_mst_bridge_state *msm_dp_mst_br_priv_state(struct drm_atomi
&bridge->obj));
}
+/* DP MST HPD IRQ callback */
+void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display)
+{
+ int rc;
+ struct msm_dp_mst *mst = dp_display->msm_dp_mst;
+ u8 ack[8] = {};
+ u8 esi[14];
+ unsigned int esi_res = DP_SINK_COUNT_ESI + 1;
+ bool handled;
+
+ rc = drm_dp_dpcd_read(mst->dp_aux, DP_SINK_COUNT_ESI,
+ esi, 14);
+ if (rc != 14) {
+ DRM_ERROR("dpcd sink status read failed, rlen=%d\n", rc);
+ return;
+ }
+
+ drm_dbg_dp(dp_display->drm_dev, "mst irq: esi1[0x%x] esi2[0x%x] esi3[%x]\n",
+ esi[1], esi[2], esi[3]);
+
+ rc = drm_dp_mst_hpd_irq_handle_event(&mst->mst_mgr, esi, ack, &handled);
+
+ /* ack the request */
+ if (handled) {
+ rc = drm_dp_dpcd_writeb(mst->dp_aux, esi_res, ack[1]);
+
+ if (rc != 1)
+ DRM_ERROR("dpcd esi_res failed. rc=%d\n", rc);
+
+ drm_dp_mst_hpd_irq_send_new_request(&mst->mst_mgr);
+ }
+ drm_dbg_dp(dp_display->drm_dev, "mst display hpd_irq handled:%d rc:%d\n", handled, rc);
+}
+
/* DP MST Connector OPs */
static int
msm_dp_mst_connector_detect(struct drm_connector *connector,
@@ -570,7 +604,7 @@ msm_dp_mst_connector_detect(struct drm_connector *connector,
struct msm_dp_mst *mst = dp_display->msm_dp_mst;
enum drm_connector_status status = connector_status_disconnected;
- if (dp_display->link_ready)
+ if (dp_display->link_ready && dp_display->mst_active)
status = drm_dp_mst_detect_port(connector,
ctx, &mst->mst_mgr, mst_conn->mst_port);
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
index 9a8535b97193d911592b44fdfcf43ca2f252cea0..5fe5dc7596086467e9a3b3d7d04a665853fbb3d7 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
@@ -103,4 +103,6 @@ int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams,
u32 max_dpcd_transaction_bytes, struct drm_dp_aux *drm_aux);
+void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display);
+
#endif /* _DP_MST_DRM_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 33/45] drm/msm/dp: add irq hpd callback for dp mst
2024-12-06 4:32 ` [PATCH 33/45] drm/msm/dp: add irq hpd callback for dp mst Abhinav Kumar
@ 2024-12-08 6:44 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 6:44 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:04PM -0800, Abhinav Kumar wrote:
> Add irq hpd callback for the dp mst module which shall be
> invoked from the dp_display's irq hpd handler to perform
> mst specific operations in case of irq hpd. In MST case, route
Nit: MST, IRQ, HPD
> the IRQ HPD messages to MST module.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 17 +++++++++++++----
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 36 +++++++++++++++++++++++++++++++++++-
> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 2 ++
> 3 files changed, 50 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 60316633cc6bcfdfad8ab356d803642d353add61..97f8228042773f51f23a9d39fc009de0798059d7 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -583,7 +583,9 @@ static int msm_dp_display_usbpd_attention_cb(struct device *dev)
> {
> int rc = 0;
> u32 sink_request;
> +
Drop.
> struct msm_dp_display_private *dp = dev_get_dp_display_private(dev);
> + struct msm_dp *msm_dp_display = &dp->msm_dp_display;
>
> /* check for any test request issued by sink */
> rc = msm_dp_link_process_request(dp->link);
> @@ -591,12 +593,18 @@ static int msm_dp_display_usbpd_attention_cb(struct device *dev)
> sink_request = dp->link->sink_request;
> drm_dbg_dp(dp->drm_dev, "hpd_state=%d sink_request=%d\n",
> dp->hpd_state, sink_request);
> - if (sink_request & DS_PORT_STATUS_CHANGED)
> - rc = msm_dp_display_handle_port_status_changed(dp);
> - else
> + if (sink_request & DS_PORT_STATUS_CHANGED) {
> + if (!msm_dp_display->mst_active)
> + rc = msm_dp_display_handle_port_status_changed(dp);
Why?
> + } else {
> rc = msm_dp_display_handle_irq_hpd(dp);
> + }
> }
>
> + /* let MST specific IRQ events be handled by its callback */
> + if (msm_dp_display->mst_active)
> + msm_dp_mst_display_hpd_irq(&dp->msm_dp_display);
> +
> return rc;
> }
>
> @@ -736,7 +744,8 @@ static int msm_dp_irq_hpd_handle(struct msm_dp_display_private *dp, u32 data)
> drm_dbg_dp(dp->drm_dev, "Before, type=%d hpd_state=%d\n",
> dp->msm_dp_display.connector_type, state);
>
> - if (state == ST_MAINLINK_READY || state == ST_DISCONNECT_PENDING) {
> + if ((state == ST_MAINLINK_READY || state == ST_DISCONNECT_PENDING) &&
> + !dp->msm_dp_display.mst_active) {
Why?
> /* wait until ST_CONNECTED */
> msm_dp_add_event(dp, EV_IRQ_HPD_INT, 0, 1); /* delay = 1 */
> mutex_unlock(&dp->event_mutex);
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> index 2934fe6a93ff8b927b035c01b6007f3f4ec91a3f..2d92084586b466d4953429e8a6fbf766d081cb9f 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> @@ -559,6 +559,40 @@ static struct msm_dp_mst_bridge_state *msm_dp_mst_br_priv_state(struct drm_atomi
> &bridge->obj));
> }
>
> +/* DP MST HPD IRQ callback */
> +void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display)
> +{
> + int rc;
> + struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> + u8 ack[8] = {};
4 should be enough
> + u8 esi[14];
> + unsigned int esi_res = DP_SINK_COUNT_ESI + 1;
inline
> + bool handled;
> +
> + rc = drm_dp_dpcd_read(mst->dp_aux, DP_SINK_COUNT_ESI,
> + esi, 14);
Why do we need 14 bytes? Other drivers read smaller chunks.
drm_dp_mst_hpd_irq_handle_event() explicitly documents just 4 bytes
there. Please consider extracting generic helper funciton that just
handles ESI read, drm_dp_mst_hpd_irq_handle_event() and writing of ACK.
> + if (rc != 14) {
> + DRM_ERROR("dpcd sink status read failed, rlen=%d\n", rc);
> + return;
> + }
> +
> + drm_dbg_dp(dp_display->drm_dev, "mst irq: esi1[0x%x] esi2[0x%x] esi3[%x]\n",
> + esi[1], esi[2], esi[3]);
> +
> + rc = drm_dp_mst_hpd_irq_handle_event(&mst->mst_mgr, esi, ack, &handled);
> +
> + /* ack the request */
> + if (handled) {
> + rc = drm_dp_dpcd_writeb(mst->dp_aux, esi_res, ack[1]);
> +
> + if (rc != 1)
> + DRM_ERROR("dpcd esi_res failed. rc=%d\n", rc);
> +
> + drm_dp_mst_hpd_irq_send_new_request(&mst->mst_mgr);
> + }
> + drm_dbg_dp(dp_display->drm_dev, "mst display hpd_irq handled:%d rc:%d\n", handled, rc);
> +}
> +
> /* DP MST Connector OPs */
> static int
> msm_dp_mst_connector_detect(struct drm_connector *connector,
> @@ -570,7 +604,7 @@ msm_dp_mst_connector_detect(struct drm_connector *connector,
> struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> enum drm_connector_status status = connector_status_disconnected;
>
> - if (dp_display->link_ready)
> + if (dp_display->link_ready && dp_display->mst_active)
Why is it a part of this patch?
> status = drm_dp_mst_detect_port(connector,
> ctx, &mst->mst_mgr, mst_conn->mst_port);
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> index 9a8535b97193d911592b44fdfcf43ca2f252cea0..5fe5dc7596086467e9a3b3d7d04a665853fbb3d7 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> @@ -103,4 +103,6 @@ int msm_dp_mst_drm_bridge_init(struct msm_dp *dp, struct drm_encoder *encoder);
> int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams,
> u32 max_dpcd_transaction_bytes, struct drm_dp_aux *drm_aux);
>
> +void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display);
> +
> #endif /* _DP_MST_DRM_H_ */
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 34/45] drm/msm/dp: add support to re-use and clear the panel edid
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (32 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 33/45] drm/msm/dp: add irq hpd callback for dp mst Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 8:53 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 35/45] drm/msm/dp: add a mst session mutex to protect bridge ops Abhinav Kumar
` (12 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
During certain cases, the dp mst connector's panel edid can be
re-used such as getting multiple get_modes() without irq_hpd and
should be cleared in cases when the connector is destroyed or when
irq hpd is received. Add support to handle these cases for the
mst_connector's panel edid.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_mst_drm.c | 39 +++++++++++++++++++++++++++++++++++++
1 file changed, 39 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
index 2d92084586b466d4953429e8a6fbf766d081cb9f..15c61fd37c418889074222c0f576778adadf51c9 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
@@ -559,6 +559,34 @@ static struct msm_dp_mst_bridge_state *msm_dp_mst_br_priv_state(struct drm_atomi
&bridge->obj));
}
+static void msm_dp_mst_clear_panel_edid(struct msm_dp *dp_display)
+{
+ struct msm_dp_mst *mst = dp_display->msm_dp_mst;
+ struct msm_dp_mst_connector *mst_conn;
+ struct msm_dp_panel *dp_panel;
+ struct msm_dp_mst_bridge *dp_bridge;
+ int i;
+
+ if (!dp_display) {
+ DRM_ERROR("invalid input\n");
+ return;
+ }
+
+ for (i = 0; i < mst->max_streams; i++) {
+ dp_bridge = &mst->mst_bridge[i];
+ mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
+ dp_panel = dp_bridge->msm_dp_panel;
+
+ if (!dp_panel || !mst_conn || !mst_conn->mst_port)
+ continue;
+
+ if (dp_panel->drm_edid) {
+ drm_edid_free(dp_panel->drm_edid);
+ dp_panel->drm_edid = NULL;
+ }
+ }
+}
+
/* DP MST HPD IRQ callback */
void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display)
{
@@ -585,6 +613,9 @@ void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display)
if (handled) {
rc = drm_dp_dpcd_writeb(mst->dp_aux, esi_res, ack[1]);
+ if (ack[1] & DP_UP_REQ_MSG_RDY)
+ msm_dp_mst_clear_panel_edid(dp_display);
+
if (rc != 1)
DRM_ERROR("dpcd esi_res failed. rc=%d\n", rc);
@@ -620,6 +651,9 @@ static int msm_dp_mst_connector_get_modes(struct drm_connector *connector)
struct msm_dp_mst *mst = dp_display->msm_dp_mst;
struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
+ if (dp_panel->drm_edid)
+ goto duplicate_edid;
+
drm_edid_free(dp_panel->drm_edid);
dp_panel->drm_edid = drm_dp_mst_edid_read(connector, &mst->mst_mgr, mst_conn->mst_port);
@@ -628,6 +662,7 @@ static int msm_dp_mst_connector_get_modes(struct drm_connector *connector)
return -EINVAL;
}
+duplicate_edid:
drm_edid_connector_update(connector, dp_panel->drm_edid);
return drm_edid_connector_add_modes(connector);
@@ -894,6 +929,10 @@ static int msm_dp_mst_connector_atomic_check(struct drm_connector *connector,
static void dp_mst_connector_destroy(struct drm_connector *connector)
{
struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
+ struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
+
+ drm_edid_free(dp_panel->drm_edid);
+ dp_panel->drm_edid = NULL;
drm_connector_cleanup(connector);
drm_dp_mst_put_port_malloc(mst_conn->mst_port);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 34/45] drm/msm/dp: add support to re-use and clear the panel edid
2024-12-06 4:32 ` [PATCH 34/45] drm/msm/dp: add support to re-use and clear the panel edid Abhinav Kumar
@ 2024-12-08 8:53 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 8:53 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:05PM -0800, Abhinav Kumar wrote:
> During certain cases, the dp mst connector's panel edid can be
> re-used such as getting multiple get_modes() without irq_hpd and
> should be cleared in cases when the connector is destroyed or when
> irq hpd is received. Add support to handle these cases for the
> mst_connector's panel edid.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 39 +++++++++++++++++++++++++++++++++++++
> 1 file changed, 39 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> index 2d92084586b466d4953429e8a6fbf766d081cb9f..15c61fd37c418889074222c0f576778adadf51c9 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> @@ -559,6 +559,34 @@ static struct msm_dp_mst_bridge_state *msm_dp_mst_br_priv_state(struct drm_atomi
> &bridge->obj));
> }
>
> +static void msm_dp_mst_clear_panel_edid(struct msm_dp *dp_display)
> +{
> + struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> + struct msm_dp_mst_connector *mst_conn;
> + struct msm_dp_panel *dp_panel;
> + struct msm_dp_mst_bridge *dp_bridge;
> + int i;
> +
> + if (!dp_display) {
> + DRM_ERROR("invalid input\n");
> + return;
> + }
Drop
> +
> + for (i = 0; i < mst->max_streams; i++) {
> + dp_bridge = &mst->mst_bridge[i];
> + mst_conn = to_msm_dp_mst_connector(dp_bridge->connector);
> + dp_panel = dp_bridge->msm_dp_panel;
> +
> + if (!dp_panel || !mst_conn || !mst_conn->mst_port)
> + continue;
> +
> + if (dp_panel->drm_edid) {
> + drm_edid_free(dp_panel->drm_edid);
> + dp_panel->drm_edid = NULL;
> + }
> + }
> +}
> +
> /* DP MST HPD IRQ callback */
> void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display)
> {
> @@ -585,6 +613,9 @@ void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display)
> if (handled) {
> rc = drm_dp_dpcd_writeb(mst->dp_aux, esi_res, ack[1]);
>
> + if (ack[1] & DP_UP_REQ_MSG_RDY)
> + msm_dp_mst_clear_panel_edid(dp_display);
> +
> if (rc != 1)
> DRM_ERROR("dpcd esi_res failed. rc=%d\n", rc);
>
> @@ -620,6 +651,9 @@ static int msm_dp_mst_connector_get_modes(struct drm_connector *connector)
> struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
>
> + if (dp_panel->drm_edid)
> + goto duplicate_edid;
> +
> drm_edid_free(dp_panel->drm_edid);
>
> dp_panel->drm_edid = drm_dp_mst_edid_read(connector, &mst->mst_mgr, mst_conn->mst_port);
> @@ -628,6 +662,7 @@ static int msm_dp_mst_connector_get_modes(struct drm_connector *connector)
> return -EINVAL;
> }
>
> +duplicate_edid:
> drm_edid_connector_update(connector, dp_panel->drm_edid);
No need to update connector. Goto to the next line instead.
>
> return drm_edid_connector_add_modes(connector);
> @@ -894,6 +929,10 @@ static int msm_dp_mst_connector_atomic_check(struct drm_connector *connector,
> static void dp_mst_connector_destroy(struct drm_connector *connector)
> {
> struct msm_dp_mst_connector *mst_conn = to_msm_dp_mst_connector(connector);
> + struct msm_dp_panel *dp_panel = mst_conn->dp_panel;
> +
> + drm_edid_free(dp_panel->drm_edid);
> + dp_panel->drm_edid = NULL;
>
> drm_connector_cleanup(connector);
> drm_dp_mst_put_port_malloc(mst_conn->mst_port);
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 35/45] drm/msm/dp: add a mst session mutex to protect bridge ops
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (33 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 34/45] drm/msm/dp: add support to re-use and clear the panel edid Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 8:54 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 36/45] drm/msm/dp: propagate hpd state changes to dp mst module Abhinav Kumar
` (11 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
To protect against concurrent access of the dp mst bridges
introduce a session mutex.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_mst_drm.c | 20 ++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_mst_drm.h | 1 +
2 files changed, 21 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
index 15c61fd37c418889074222c0f576778adadf51c9..313eb63b9a35cbbb36db2d7d8f0a85e4441f2998 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
@@ -344,6 +344,7 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
struct msm_dp_mst_bridge *bridge;
struct msm_dp *dp;
struct msm_dp_mst_bridge_state *msm_dp_bridge_state;
+ struct msm_dp_mst *dp_mst;
if (!drm_bridge) {
DRM_ERROR("Invalid params\n");
@@ -353,6 +354,7 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
bridge = to_msm_dp_mst_bridge(drm_bridge);
msm_dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
dp = bridge->display;
+ dp_mst = dp->msm_dp_mst;
/* to cover cases of bridge_disable/bridge_enable without modeset */
bridge->connector = msm_dp_bridge_state->connector;
@@ -363,12 +365,14 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
return;
}
+ mutex_lock(&dp_mst->mst_lock);
msm_dp_display_atomic_prepare(dp);
rc = _msm_dp_mst_bridge_pre_enable_part1(bridge, old_bridge_state);
if (rc) {
DRM_ERROR("[%d] DP display pre-enable failed, rc=%d\n", bridge->id, rc);
msm_dp_display_unprepare(dp);
+ mutex_unlock(&dp_mst->mst_lock);
return;
}
@@ -381,6 +385,8 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
drm_mode_vrefresh(&bridge->drm_mode),
bridge->vcpi, bridge->start_slot,
bridge->start_slot + bridge->num_slots);
+
+ mutex_unlock(&dp_mst->mst_lock);
}
static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
@@ -388,6 +394,7 @@ static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
{
struct msm_dp_mst_bridge *bridge;
struct msm_dp *dp;
+ struct msm_dp_mst *mst;
if (!drm_bridge) {
DRM_ERROR("Invalid params\n");
@@ -401,6 +408,9 @@ static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
}
dp = bridge->display;
+ mst = dp->msm_dp_mst;
+
+ mutex_lock(&mst->mst_lock);
_msm_dp_mst_bridge_pre_disable_part1(bridge, old_bridge_state);
@@ -410,6 +420,8 @@ static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d disable complete\n", bridge->id,
DP_MST_CONN_ID(bridge));
+
+ mutex_unlock(&mst->mst_lock);
}
static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
@@ -418,6 +430,7 @@ static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
int conn = 0;
struct msm_dp_mst_bridge *bridge;
struct msm_dp *dp;
+ struct msm_dp_mst *mst;
if (!drm_bridge) {
DRM_ERROR("Invalid params\n");
@@ -433,6 +446,9 @@ static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
conn = DP_MST_CONN_ID(bridge);
dp = bridge->display;
+ mst = dp->msm_dp_mst;
+
+ mutex_lock(&mst->mst_lock);
msm_dp_display_atomic_post_disable_helper(dp, bridge->msm_dp_panel);
@@ -444,6 +460,8 @@ static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d post disable complete\n",
bridge->id, conn);
+
+ mutex_unlock(&mst->mst_lock);
}
static void msm_dp_mst_bridge_mode_set(struct drm_bridge *drm_bridge,
@@ -1072,6 +1090,8 @@ int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams, u32 max_dpcd_tra
dp_display->msm_dp_mst = msm_dp_mst;
+ mutex_init(&msm_dp_mst->mst_lock);
+
msm_dp_mst->mst_initialized = true;
drm_dbg_dp(dp_display->drm_dev, "dp drm mst topology manager init completed\n");
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
index 5fe5dc7596086467e9a3b3d7d04a665853fbb3d7..b1adb8a61115d4809107553809206bb2ed3c6c3d 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
@@ -80,6 +80,7 @@ struct msm_dp_mst {
struct drm_dp_aux *dp_aux;
bool mst_session_hpd_state;
u32 max_streams;
+ struct mutex mst_lock;
};
struct msm_dp_mst_connector {
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 35/45] drm/msm/dp: add a mst session mutex to protect bridge ops
2024-12-06 4:32 ` [PATCH 35/45] drm/msm/dp: add a mst session mutex to protect bridge ops Abhinav Kumar
@ 2024-12-08 8:54 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 8:54 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:06PM -0800, Abhinav Kumar wrote:
> To protect against concurrent access of the dp mst bridges
> introduce a session mutex.
Why is it necessary? What kind of concurrent access is expected?
Why wasn't it introduced together with struct msm_dp_mst?
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 20 ++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 1 +
> 2 files changed, 21 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> index 15c61fd37c418889074222c0f576778adadf51c9..313eb63b9a35cbbb36db2d7d8f0a85e4441f2998 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> @@ -344,6 +344,7 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
> struct msm_dp_mst_bridge *bridge;
> struct msm_dp *dp;
> struct msm_dp_mst_bridge_state *msm_dp_bridge_state;
> + struct msm_dp_mst *dp_mst;
>
> if (!drm_bridge) {
> DRM_ERROR("Invalid params\n");
> @@ -353,6 +354,7 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
> bridge = to_msm_dp_mst_bridge(drm_bridge);
> msm_dp_bridge_state = to_msm_dp_mst_bridge_state(bridge);
> dp = bridge->display;
> + dp_mst = dp->msm_dp_mst;
>
> /* to cover cases of bridge_disable/bridge_enable without modeset */
> bridge->connector = msm_dp_bridge_state->connector;
> @@ -363,12 +365,14 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
> return;
> }
>
> + mutex_lock(&dp_mst->mst_lock);
> msm_dp_display_atomic_prepare(dp);
>
> rc = _msm_dp_mst_bridge_pre_enable_part1(bridge, old_bridge_state);
> if (rc) {
> DRM_ERROR("[%d] DP display pre-enable failed, rc=%d\n", bridge->id, rc);
> msm_dp_display_unprepare(dp);
> + mutex_unlock(&dp_mst->mst_lock);
> return;
> }
>
> @@ -381,6 +385,8 @@ static void msm_dp_mst_bridge_atomic_pre_enable(struct drm_bridge *drm_bridge,
> drm_mode_vrefresh(&bridge->drm_mode),
> bridge->vcpi, bridge->start_slot,
> bridge->start_slot + bridge->num_slots);
> +
> + mutex_unlock(&dp_mst->mst_lock);
> }
>
> static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
> @@ -388,6 +394,7 @@ static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
> {
> struct msm_dp_mst_bridge *bridge;
> struct msm_dp *dp;
> + struct msm_dp_mst *mst;
>
> if (!drm_bridge) {
> DRM_ERROR("Invalid params\n");
> @@ -401,6 +408,9 @@ static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
> }
>
> dp = bridge->display;
> + mst = dp->msm_dp_mst;
> +
> + mutex_lock(&mst->mst_lock);
>
> _msm_dp_mst_bridge_pre_disable_part1(bridge, old_bridge_state);
>
> @@ -410,6 +420,8 @@ static void msm_dp_mst_bridge_atomic_disable(struct drm_bridge *drm_bridge,
>
> drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d disable complete\n", bridge->id,
> DP_MST_CONN_ID(bridge));
> +
> + mutex_unlock(&mst->mst_lock);
> }
>
> static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
> @@ -418,6 +430,7 @@ static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
> int conn = 0;
> struct msm_dp_mst_bridge *bridge;
> struct msm_dp *dp;
> + struct msm_dp_mst *mst;
>
> if (!drm_bridge) {
> DRM_ERROR("Invalid params\n");
> @@ -433,6 +446,9 @@ static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
> conn = DP_MST_CONN_ID(bridge);
>
> dp = bridge->display;
> + mst = dp->msm_dp_mst;
> +
> + mutex_lock(&mst->mst_lock);
>
> msm_dp_display_atomic_post_disable_helper(dp, bridge->msm_dp_panel);
>
> @@ -444,6 +460,8 @@ static void msm_dp_mst_bridge_atomic_post_disable(struct drm_bridge *drm_bridge,
>
> drm_dbg_dp(dp->drm_dev, "mst bridge:%d conn:%d post disable complete\n",
> bridge->id, conn);
> +
> + mutex_unlock(&mst->mst_lock);
> }
>
> static void msm_dp_mst_bridge_mode_set(struct drm_bridge *drm_bridge,
> @@ -1072,6 +1090,8 @@ int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams, u32 max_dpcd_tra
>
> dp_display->msm_dp_mst = msm_dp_mst;
>
> + mutex_init(&msm_dp_mst->mst_lock);
> +
> msm_dp_mst->mst_initialized = true;
>
> drm_dbg_dp(dp_display->drm_dev, "dp drm mst topology manager init completed\n");
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> index 5fe5dc7596086467e9a3b3d7d04a665853fbb3d7..b1adb8a61115d4809107553809206bb2ed3c6c3d 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> @@ -80,6 +80,7 @@ struct msm_dp_mst {
> struct drm_dp_aux *dp_aux;
> bool mst_session_hpd_state;
> u32 max_streams;
> + struct mutex mst_lock;
> };
>
> struct msm_dp_mst_connector {
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 36/45] drm/msm/dp: propagate hpd state changes to dp mst module
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (34 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 35/45] drm/msm/dp: add a mst session mutex to protect bridge ops Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 8:58 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 37/45] drm/msm: add support for non-blocking commits Abhinav Kumar
` (10 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar, Yongxing Mou
From: Yongxing Mou <quic_yongmou@quicinc.com>
Propagate the hpd state changes to dp mst module so that it
can be synchronized with the cable connect/disconnect.
Signed-off-by: Yongxing Mou <quic_yongmou@quicinc.com>
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 20 ++++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_mst_drm.c | 18 ++++++++++++++++++
drivers/gpu/drm/msm/dp/dp_mst_drm.h | 1 +
3 files changed, 39 insertions(+)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 97f8228042773f51f23a9d39fc009de0798059d7..80df79a7c2077d49184cdeb7b801bf0699ff4ece 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -26,6 +26,7 @@
#include "dp_drm.h"
#include "dp_audio.h"
#include "dp_debug.h"
+#include "dp_mst_drm.h"
static bool psr_enabled = false;
module_param(psr_enabled, bool, 0);
@@ -409,6 +410,17 @@ static void msm_dp_display_mst_init(struct msm_dp_display_private *dp)
msm_dp->mst_active = true;
}
+static void msm_dp_display_set_mst_mgr_state(struct msm_dp_display_private *dp,
+ bool state)
+{
+ if (!dp->msm_dp_display.mst_active)
+ return;
+
+ msm_dp_mst_display_set_mgr_state(&dp->msm_dp_display, state);
+
+ drm_dbg_dp(dp->drm_dev, "mst_mgr_state: %d\n", state);
+}
+
static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
{
struct drm_connector *connector = dp->msm_dp_display.connector;
@@ -455,6 +467,8 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
goto end;
}
+ msm_dp_display_set_mst_mgr_state(dp, true);
+
msm_dp_add_event(dp, EV_USER_NOTIFICATION, true, 0);
end:
@@ -521,6 +535,12 @@ static int msm_dp_display_usbpd_configure_cb(struct device *dev)
static int msm_dp_display_notify_disconnect(struct device *dev)
{
struct msm_dp_display_private *dp = dev_get_dp_display_private(dev);
+ struct msm_dp *dp_display = &dp->msm_dp_display;
+
+ if (dp_display->mst_active) {
+ msm_dp_mst_display_set_mgr_state(&dp->msm_dp_display, false);
+ dp_display->mst_active = false;
+ }
msm_dp_add_event(dp, EV_USER_NOTIFICATION, false, 0);
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
index 313eb63b9a35cbbb36db2d7d8f0a85e4441f2998..1149af71d01f99ba5326870fa69e30ae081d6101 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
@@ -1043,6 +1043,24 @@ msm_dp_mst_add_connector(struct drm_dp_mst_topology_mgr *mgr,
return &mst_connector->connector;
}
+int msm_dp_mst_display_set_mgr_state(struct msm_dp *dp_display, bool state)
+{
+ int rc;
+ struct msm_dp_mst *mst = dp_display->msm_dp_mst;
+
+ if (state)
+ mst->mst_session_hpd_state = state;
+
+ rc = drm_dp_mst_topology_mgr_set_mst(&mst->mst_mgr, state);
+ if (rc < 0) {
+ DRM_ERROR("failed to set topology mgr state to %d. rc %d\n",
+ state, rc);
+ }
+
+ drm_dbg_dp(dp_display->drm_dev, "dp_mst_display_set_mgr_state state:%d\n", state);
+ return rc;
+}
+
static const struct drm_dp_mst_topology_cbs msm_dp_mst_drm_cbs = {
.add_connector = msm_dp_mst_add_connector,
};
diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
index b1adb8a61115d4809107553809206bb2ed3c6c3d..b89913ef7b343d449e0003f56b96df049fa36e89 100644
--- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
+++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
@@ -105,5 +105,6 @@ int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams,
u32 max_dpcd_transaction_bytes, struct drm_dp_aux *drm_aux);
void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display);
+int msm_dp_mst_display_set_mgr_state(struct msm_dp *dp_display, bool state);
#endif /* _DP_MST_DRM_H_ */
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 36/45] drm/msm/dp: propagate hpd state changes to dp mst module
2024-12-06 4:32 ` [PATCH 36/45] drm/msm/dp: propagate hpd state changes to dp mst module Abhinav Kumar
@ 2024-12-08 8:58 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 8:58 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Yongxing Mou
On Thu, Dec 05, 2024 at 08:32:07PM -0800, Abhinav Kumar wrote:
> From: Yongxing Mou <quic_yongmou@quicinc.com>
>
> Propagate the hpd state changes to dp mst module so that it
please fix upcase.
> can be synchronized with the cable connect/disconnect.
It doesn't propagate HPD state. Instead it updates MST state when the
device is plugged / unplugged.
>
> Signed-off-by: Yongxing Mou <quic_yongmou@quicinc.com>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 20 ++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 18 ++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 1 +
> 3 files changed, 39 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 97f8228042773f51f23a9d39fc009de0798059d7..80df79a7c2077d49184cdeb7b801bf0699ff4ece 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -26,6 +26,7 @@
> #include "dp_drm.h"
> #include "dp_audio.h"
> #include "dp_debug.h"
> +#include "dp_mst_drm.h"
>
> static bool psr_enabled = false;
> module_param(psr_enabled, bool, 0);
> @@ -409,6 +410,17 @@ static void msm_dp_display_mst_init(struct msm_dp_display_private *dp)
> msm_dp->mst_active = true;
> }
>
> +static void msm_dp_display_set_mst_mgr_state(struct msm_dp_display_private *dp,
> + bool state)
> +{
> + if (!dp->msm_dp_display.mst_active)
> + return;
> +
> + msm_dp_mst_display_set_mgr_state(&dp->msm_dp_display, state);
> +
> + drm_dbg_dp(dp->drm_dev, "mst_mgr_state: %d\n", state);
> +}
> +
> static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> {
> struct drm_connector *connector = dp->msm_dp_display.connector;
> @@ -455,6 +467,8 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> goto end;
> }
>
> + msm_dp_display_set_mst_mgr_state(dp, true);
> +
> msm_dp_add_event(dp, EV_USER_NOTIFICATION, true, 0);
>
> end:
> @@ -521,6 +535,12 @@ static int msm_dp_display_usbpd_configure_cb(struct device *dev)
> static int msm_dp_display_notify_disconnect(struct device *dev)
> {
> struct msm_dp_display_private *dp = dev_get_dp_display_private(dev);
> + struct msm_dp *dp_display = &dp->msm_dp_display;
> +
> + if (dp_display->mst_active) {
> + msm_dp_mst_display_set_mgr_state(&dp->msm_dp_display, false);
> + dp_display->mst_active = false;
> + }
>
> msm_dp_add_event(dp, EV_USER_NOTIFICATION, false, 0);
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.c b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> index 313eb63b9a35cbbb36db2d7d8f0a85e4441f2998..1149af71d01f99ba5326870fa69e30ae081d6101 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.c
> @@ -1043,6 +1043,24 @@ msm_dp_mst_add_connector(struct drm_dp_mst_topology_mgr *mgr,
> return &mst_connector->connector;
> }
>
> +int msm_dp_mst_display_set_mgr_state(struct msm_dp *dp_display, bool state)
> +{
> + int rc;
> + struct msm_dp_mst *mst = dp_display->msm_dp_mst;
> +
> + if (state)
> + mst->mst_session_hpd_state = state;
Where is mst_session_hpd_state cleared?
> +
> + rc = drm_dp_mst_topology_mgr_set_mst(&mst->mst_mgr, state);
> + if (rc < 0) {
> + DRM_ERROR("failed to set topology mgr state to %d. rc %d\n",
> + state, rc);
> + }
> +
> + drm_dbg_dp(dp_display->drm_dev, "dp_mst_display_set_mgr_state state:%d\n", state);
> + return rc;
> +}
> +
> static const struct drm_dp_mst_topology_cbs msm_dp_mst_drm_cbs = {
> .add_connector = msm_dp_mst_add_connector,
> };
> diff --git a/drivers/gpu/drm/msm/dp/dp_mst_drm.h b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> index b1adb8a61115d4809107553809206bb2ed3c6c3d..b89913ef7b343d449e0003f56b96df049fa36e89 100644
> --- a/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> +++ b/drivers/gpu/drm/msm/dp/dp_mst_drm.h
> @@ -105,5 +105,6 @@ int msm_dp_mst_init(struct msm_dp *dp_display, u32 max_streams,
> u32 max_dpcd_transaction_bytes, struct drm_dp_aux *drm_aux);
>
> void msm_dp_mst_display_hpd_irq(struct msm_dp *dp_display);
> +int msm_dp_mst_display_set_mgr_state(struct msm_dp *dp_display, bool state);
>
> #endif /* _DP_MST_DRM_H_ */
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 37/45] drm/msm: add support for non-blocking commits
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (35 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 36/45] drm/msm/dp: propagate hpd state changes to dp mst module Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-06 9:27 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 38/45] drm/msm: initialize DRM MST encoders for DP controllers Abhinav Kumar
` (9 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Hook up the mst framework APIs with atomic_commit_setup() and
atomic_commit_tail() APIs to handle non-blocking commits.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/msm_atomic.c | 2 ++
drivers/gpu/drm/msm/msm_drv.h | 1 +
drivers/gpu/drm/msm/msm_kms.c | 1 +
3 files changed, 4 insertions(+)
diff --git a/drivers/gpu/drm/msm/msm_atomic.c b/drivers/gpu/drm/msm/msm_atomic.c
index 9c45d641b5212c11078ab38c13a519663d85e10a..801399419c3d26f68d9b0a65d41fc4e1706c70be 100644
--- a/drivers/gpu/drm/msm/msm_atomic.c
+++ b/drivers/gpu/drm/msm/msm_atomic.c
@@ -210,6 +210,8 @@ void msm_atomic_commit_tail(struct drm_atomic_state *state)
trace_msm_atomic_commit_tail_start(async, crtc_mask);
+ drm_dp_mst_atomic_wait_for_dependencies(state);
+
kms->funcs->enable_commit(kms);
/*
diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
index d8c9a1b192632d3e29ff125bd7bb2d0bb491275d..1616a4682795f6b9b30cc0bef2baf448ccc62bc0 100644
--- a/drivers/gpu/drm/msm/msm_drv.h
+++ b/drivers/gpu/drm/msm/msm_drv.h
@@ -30,6 +30,7 @@
#include <drm/drm_atomic_helper.h>
#include <drm/drm_probe_helper.h>
#include <drm/display/drm_dsc.h>
+#include <drm/display/drm_dp_mst_helper.h>
#include <drm/msm_drm.h>
#include <drm/drm_gem.h>
diff --git a/drivers/gpu/drm/msm/msm_kms.c b/drivers/gpu/drm/msm/msm_kms.c
index f3326d09bdbce19d40d0b48549c330c2b836476f..343ad9e9988f6c8d99c5867cf8e81ae625aaa90d 100644
--- a/drivers/gpu/drm/msm/msm_kms.c
+++ b/drivers/gpu/drm/msm/msm_kms.c
@@ -28,6 +28,7 @@ static const struct drm_mode_config_funcs mode_config_funcs = {
static const struct drm_mode_config_helper_funcs mode_config_helper_funcs = {
.atomic_commit_tail = msm_atomic_commit_tail,
+ .atomic_commit_setup = drm_dp_mst_atomic_setup_commit,
};
static irqreturn_t msm_irq(int irq, void *arg)
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 37/45] drm/msm: add support for non-blocking commits
2024-12-06 4:32 ` [PATCH 37/45] drm/msm: add support for non-blocking commits Abhinav Kumar
@ 2024-12-06 9:27 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:27 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:08PM -0800, Abhinav Kumar wrote:
> Hook up the mst framework APIs with atomic_commit_setup() and
> atomic_commit_tail() APIs to handle non-blocking commits.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/msm_atomic.c | 2 ++
> drivers/gpu/drm/msm/msm_drv.h | 1 +
> drivers/gpu/drm/msm/msm_kms.c | 1 +
> 3 files changed, 4 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/msm_atomic.c b/drivers/gpu/drm/msm/msm_atomic.c
> index 9c45d641b5212c11078ab38c13a519663d85e10a..801399419c3d26f68d9b0a65d41fc4e1706c70be 100644
> --- a/drivers/gpu/drm/msm/msm_atomic.c
> +++ b/drivers/gpu/drm/msm/msm_atomic.c
> @@ -210,6 +210,8 @@ void msm_atomic_commit_tail(struct drm_atomic_state *state)
>
> trace_msm_atomic_commit_tail_start(async, crtc_mask);
>
> + drm_dp_mst_atomic_wait_for_dependencies(state);
> +
> kms->funcs->enable_commit(kms);
>
> /*
> diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
> index d8c9a1b192632d3e29ff125bd7bb2d0bb491275d..1616a4682795f6b9b30cc0bef2baf448ccc62bc0 100644
> --- a/drivers/gpu/drm/msm/msm_drv.h
> +++ b/drivers/gpu/drm/msm/msm_drv.h
> @@ -30,6 +30,7 @@
> #include <drm/drm_atomic_helper.h>
> #include <drm/drm_probe_helper.h>
> #include <drm/display/drm_dsc.h>
> +#include <drm/display/drm_dp_mst_helper.h>
Please don't bring extra dependencies to the global list. Individual
files can perfectly include the header on their own.
> #include <drm/msm_drm.h>
> #include <drm/drm_gem.h>
>
> diff --git a/drivers/gpu/drm/msm/msm_kms.c b/drivers/gpu/drm/msm/msm_kms.c
> index f3326d09bdbce19d40d0b48549c330c2b836476f..343ad9e9988f6c8d99c5867cf8e81ae625aaa90d 100644
> --- a/drivers/gpu/drm/msm/msm_kms.c
> +++ b/drivers/gpu/drm/msm/msm_kms.c
> @@ -28,6 +28,7 @@ static const struct drm_mode_config_funcs mode_config_funcs = {
>
> static const struct drm_mode_config_helper_funcs mode_config_helper_funcs = {
> .atomic_commit_tail = msm_atomic_commit_tail,
> + .atomic_commit_setup = drm_dp_mst_atomic_setup_commit,
> };
>
> static irqreturn_t msm_irq(int irq, void *arg)
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 38/45] drm/msm: initialize DRM MST encoders for DP controllers
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (36 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 37/45] drm/msm: add support for non-blocking commits Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 6:25 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 39/45] drm/msm/dp: initialize dp_mst module for each DP MST controller Abhinav Kumar
` (8 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Initiliaze a DPMST encoder for each MST capable DP controller
and the number of encoders it supports depends on the number
of streams it supports. Replace the opencoded instances of max_stream
with the newly introduced API to centralize the usage.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h | 2 ++
drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 23 ++++++++++++++++++++++-
drivers/gpu/drm/msm/dp/dp_display.c | 26 +++++++++++++++++++++-----
drivers/gpu/drm/msm/msm_drv.h | 14 ++++++++++++++
4 files changed, 59 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
index 92b5ee390788d16e85e195a664417896a2bf1cae..618a5b6f8222882ed8c972a78a26f8c25ca389a8 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
@@ -28,6 +28,7 @@
* @h_tile_instance: Controller instance used per tile. Number of elements is
* based on num_of_h_tiles
* @is_cmd_mode Boolean to indicate if the CMD mode is requested
+ * @stream_id stream id for which the interface needs to be acquired
* @vsync_source: Source of the TE signal for DSI CMD devices
*/
struct msm_display_info {
@@ -35,6 +36,7 @@ struct msm_display_info {
uint32_t num_of_h_tiles;
uint32_t h_tile_instance[MAX_H_TILES_PER_DISPLAY];
bool is_cmd_mode;
+ int stream_id;
enum dpu_vsync_source vsync_source;
};
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
index 8b251f87a0520da0807b9b7aed17493990e41627..359de04abf4bbead3daa5e8b357a3c34216e3e65 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
@@ -636,7 +636,8 @@ static int _dpu_kms_initialize_displayport(struct drm_device *dev,
struct msm_display_info info;
bool yuv_supported;
int rc;
- int i;
+ int i, stream_id;
+ int stream_cnt;
for (i = 0; i < ARRAY_SIZE(priv->dp); i++) {
if (!priv->dp[i])
@@ -659,6 +660,26 @@ static int _dpu_kms_initialize_displayport(struct drm_device *dev,
DPU_ERROR("modeset_init failed for DP, rc = %d\n", rc);
return rc;
}
+
+ stream_cnt = msm_dp_get_mst_max_stream(priv->dp[i]);
+
+ if (stream_cnt > 1) {
+ for (stream_id = 0; stream_id < stream_cnt; stream_id++) {
+ info.stream_id = stream_id;
+ encoder = dpu_encoder_init(dev, DRM_MODE_ENCODER_DPMST, &info);
+ if (IS_ERR(encoder)) {
+ DPU_ERROR("encoder init failed for dp mst display\n");
+ return PTR_ERR(encoder);
+ }
+
+ rc = msm_dp_mst_bridge_init(priv->dp[i], encoder);
+ if (rc) {
+ DPU_ERROR("dp mst bridge %d init failed, %d\n",
+ stream_id, rc);
+ continue;
+ }
+ }
+ }
}
return 0;
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 80df79a7c2077d49184cdeb7b801bf0699ff4ece..eafec9ab4f83cb44e861687e7550748b4d9b7ece 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -432,7 +432,8 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
if (rc)
goto end;
- if (dp->max_stream <= DEFAULT_STREAM_COUNT || !msm_dp_panel_read_mst_cap(dp->panel)) {
+ if (msm_dp_get_mst_max_stream(dp_display) <= DEFAULT_STREAM_COUNT ||
+ !msm_dp_panel_read_mst_cap(dp->panel)) {
rc = msm_dp_panel_read_edid(dp->panel, connector);
if (rc)
goto end;
@@ -457,7 +458,8 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
*/
msm_dp_link_psm_config(dp->link, &dp->panel->link_info, false);
- if (dp->max_stream > DEFAULT_STREAM_COUNT && msm_dp_panel_read_mst_cap(dp->panel))
+ if (msm_dp_get_mst_max_stream(dp_display) > DEFAULT_STREAM_COUNT &&
+ msm_dp_panel_read_mst_cap(dp->panel))
msm_dp_display_mst_init(dp);
msm_dp_link_reset_phy_params_vx_px(dp->link);
@@ -977,7 +979,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp,
drm_dbg_dp(dp->drm_dev, "sink_count=%d\n", dp->link->sink_count);
- rc = msm_dp_ctrl_on_stream(dp->ctrl, msm_dp_panel, dp->max_stream);
+ rc = msm_dp_ctrl_on_stream(dp->ctrl, msm_dp_panel, msm_dp_get_mst_max_stream(&dp->msm_dp_display));
return rc;
}
@@ -1444,6 +1446,20 @@ static int msm_dp_display_get_connector_type(struct platform_device *pdev,
return connector_type;
}
+int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display)
+{
+ struct msm_dp_display_private *dp_priv;
+
+ dp_priv = container_of(dp_display, struct msm_dp_display_private, msm_dp_display);
+
+ return dp_priv->max_stream;
+}
+
+int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
+{
+ return msm_dp_mst_drm_bridge_init(dp_display, encoder);
+}
+
static int msm_dp_display_probe(struct platform_device *pdev)
{
int rc = 0;
@@ -1745,12 +1761,12 @@ void msm_dp_display_disable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_d
return;
}
- if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT)
+ if (msm_dp_get_mst_max_stream(dp) > DEFAULT_STREAM_COUNT)
msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_panel);
else
msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
- if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT) {
+ if (msm_dp_get_mst_max_stream(dp) > DEFAULT_STREAM_COUNT) {
msm_dp_ctrl_mst_stream_channel_slot_setup(msm_dp_display->ctrl,
msm_dp_display->max_stream);
msm_dp_ctrl_mst_send_act(msm_dp_display->ctrl);
diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
index 1616a4682795f6b9b30cc0bef2baf448ccc62bc0..12b50a797772f574122481cd8a1c7c88aacb8250 100644
--- a/drivers/gpu/drm/msm/msm_drv.h
+++ b/drivers/gpu/drm/msm/msm_drv.h
@@ -372,6 +372,10 @@ bool msm_dp_needs_periph_flush(const struct msm_dp *dp_display,
const struct drm_display_mode *mode);
bool msm_dp_wide_bus_available(const struct msm_dp *dp_display);
+int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display);
+
+int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder);
+
#else
static inline int __init msm_dp_register(void)
{
@@ -388,6 +392,16 @@ static inline int msm_dp_modeset_init(struct msm_dp *dp_display,
return -EINVAL;
}
+static inline int msm_dp_get_mst_max_stream(struct msm_dp *dp_display)
+{
+ return -EINVAL;
+}
+
+int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
+{
+ return -EINVAL;
+}
+
static inline void msm_dp_snapshot(struct msm_disp_state *disp_state, struct msm_dp *dp_display)
{
}
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 38/45] drm/msm: initialize DRM MST encoders for DP controllers
2024-12-06 4:32 ` [PATCH 38/45] drm/msm: initialize DRM MST encoders for DP controllers Abhinav Kumar
@ 2024-12-08 6:25 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 6:25 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:09PM -0800, Abhinav Kumar wrote:
> Initiliaze a DPMST encoder for each MST capable DP controller
> and the number of encoders it supports depends on the number
> of streams it supports. Replace the opencoded instances of max_stream
> with the newly introduced API to centralize the usage.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h | 2 ++
> drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 23 ++++++++++++++++++++++-
> drivers/gpu/drm/msm/dp/dp_display.c | 26 +++++++++++++++++++++-----
> drivers/gpu/drm/msm/msm_drv.h | 14 ++++++++++++++
Split into two commits
> 4 files changed, 59 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
> index 92b5ee390788d16e85e195a664417896a2bf1cae..618a5b6f8222882ed8c972a78a26f8c25ca389a8 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
> @@ -28,6 +28,7 @@
> * @h_tile_instance: Controller instance used per tile. Number of elements is
> * based on num_of_h_tiles
> * @is_cmd_mode Boolean to indicate if the CMD mode is requested
> + * @stream_id stream id for which the interface needs to be acquired
> * @vsync_source: Source of the TE signal for DSI CMD devices
> */
> struct msm_display_info {
> @@ -35,6 +36,7 @@ struct msm_display_info {
> uint32_t num_of_h_tiles;
> uint32_t h_tile_instance[MAX_H_TILES_PER_DISPLAY];
> bool is_cmd_mode;
> + int stream_id;
> enum dpu_vsync_source vsync_source;
> };
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> index 8b251f87a0520da0807b9b7aed17493990e41627..359de04abf4bbead3daa5e8b357a3c34216e3e65 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> @@ -636,7 +636,8 @@ static int _dpu_kms_initialize_displayport(struct drm_device *dev,
> struct msm_display_info info;
> bool yuv_supported;
> int rc;
> - int i;
> + int i, stream_id;
> + int stream_cnt;
>
> for (i = 0; i < ARRAY_SIZE(priv->dp); i++) {
> if (!priv->dp[i])
> @@ -659,6 +660,26 @@ static int _dpu_kms_initialize_displayport(struct drm_device *dev,
> DPU_ERROR("modeset_init failed for DP, rc = %d\n", rc);
> return rc;
> }
> +
> + stream_cnt = msm_dp_get_mst_max_stream(priv->dp[i]);
Is there any reason for this not being a part of the DPU catalog? DPU
can support required number of DPMST streams even if DP doesn't
> +
> + if (stream_cnt > 1) {
> + for (stream_id = 0; stream_id < stream_cnt; stream_id++) {
> + info.stream_id = stream_id;
> + encoder = dpu_encoder_init(dev, DRM_MODE_ENCODER_DPMST, &info);
> + if (IS_ERR(encoder)) {
> + DPU_ERROR("encoder init failed for dp mst display\n");
> + return PTR_ERR(encoder);
> + }
> +
> + rc = msm_dp_mst_bridge_init(priv->dp[i], encoder);
> + if (rc) {
> + DPU_ERROR("dp mst bridge %d init failed, %d\n",
> + stream_id, rc);
> + continue;
> + }
> + }
> + }
> }
>
> return 0;
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 80df79a7c2077d49184cdeb7b801bf0699ff4ece..eafec9ab4f83cb44e861687e7550748b4d9b7ece 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -432,7 +432,8 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> if (rc)
> goto end;
>
> - if (dp->max_stream <= DEFAULT_STREAM_COUNT || !msm_dp_panel_read_mst_cap(dp->panel)) {
> + if (msm_dp_get_mst_max_stream(dp_display) <= DEFAULT_STREAM_COUNT ||
> + !msm_dp_panel_read_mst_cap(dp->panel)) {
> rc = msm_dp_panel_read_edid(dp->panel, connector);
> if (rc)
> goto end;
> @@ -457,7 +458,8 @@ static int msm_dp_display_process_hpd_high(struct msm_dp_display_private *dp)
> */
> msm_dp_link_psm_config(dp->link, &dp->panel->link_info, false);
>
> - if (dp->max_stream > DEFAULT_STREAM_COUNT && msm_dp_panel_read_mst_cap(dp->panel))
> + if (msm_dp_get_mst_max_stream(dp_display) > DEFAULT_STREAM_COUNT &&
> + msm_dp_panel_read_mst_cap(dp->panel))
> msm_dp_display_mst_init(dp);
>
> msm_dp_link_reset_phy_params_vx_px(dp->link);
> @@ -977,7 +979,7 @@ static int msm_dp_display_enable(struct msm_dp_display_private *dp,
>
> drm_dbg_dp(dp->drm_dev, "sink_count=%d\n", dp->link->sink_count);
>
> - rc = msm_dp_ctrl_on_stream(dp->ctrl, msm_dp_panel, dp->max_stream);
> + rc = msm_dp_ctrl_on_stream(dp->ctrl, msm_dp_panel, msm_dp_get_mst_max_stream(&dp->msm_dp_display));
>
> return rc;
> }
> @@ -1444,6 +1446,20 @@ static int msm_dp_display_get_connector_type(struct platform_device *pdev,
> return connector_type;
> }
>
> +int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display)
> +{
> + struct msm_dp_display_private *dp_priv;
> +
> + dp_priv = container_of(dp_display, struct msm_dp_display_private, msm_dp_display);
> +
> + return dp_priv->max_stream;
> +}
> +
> +int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
> +{
> + return msm_dp_mst_drm_bridge_init(dp_display, encoder);
> +}
> +
> static int msm_dp_display_probe(struct platform_device *pdev)
> {
> int rc = 0;
> @@ -1745,12 +1761,12 @@ void msm_dp_display_disable_helper(struct msm_dp *dp, struct msm_dp_panel *msm_d
> return;
> }
>
> - if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT)
> + if (msm_dp_get_mst_max_stream(dp) > DEFAULT_STREAM_COUNT)
> msm_dp_ctrl_push_vcpf(msm_dp_display->ctrl, msm_dp_panel);
> else
> msm_dp_ctrl_push_idle(msm_dp_display->ctrl);
>
> - if (msm_dp_display->max_stream > DEFAULT_STREAM_COUNT) {
> + if (msm_dp_get_mst_max_stream(dp) > DEFAULT_STREAM_COUNT) {
> msm_dp_ctrl_mst_stream_channel_slot_setup(msm_dp_display->ctrl,
> msm_dp_display->max_stream);
> msm_dp_ctrl_mst_send_act(msm_dp_display->ctrl);
> diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
> index 1616a4682795f6b9b30cc0bef2baf448ccc62bc0..12b50a797772f574122481cd8a1c7c88aacb8250 100644
> --- a/drivers/gpu/drm/msm/msm_drv.h
> +++ b/drivers/gpu/drm/msm/msm_drv.h
> @@ -372,6 +372,10 @@ bool msm_dp_needs_periph_flush(const struct msm_dp *dp_display,
> const struct drm_display_mode *mode);
> bool msm_dp_wide_bus_available(const struct msm_dp *dp_display);
>
> +int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display);
> +
> +int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder);
> +
> #else
> static inline int __init msm_dp_register(void)
> {
> @@ -388,6 +392,16 @@ static inline int msm_dp_modeset_init(struct msm_dp *dp_display,
> return -EINVAL;
> }
>
> +static inline int msm_dp_get_mst_max_stream(struct msm_dp *dp_display)
> +{
> + return -EINVAL;
> +}
> +
> +int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
> +{
> + return -EINVAL;
> +}
> +
> static inline void msm_dp_snapshot(struct msm_disp_state *disp_state, struct msm_dp *dp_display)
> {
> }
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 39/45] drm/msm/dp: initialize dp_mst module for each DP MST controller
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (37 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 38/45] drm/msm: initialize DRM MST encoders for DP controllers Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 6:31 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 40/45] drm/msm: add a stream to intf map for DP controller Abhinav Kumar
` (7 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
For each MST capable DP controller, initialize a dp_mst module to
manage its DP MST operations. The DP MST module for each controller
is the central entity to manage its topology related operations as
well as interfacing with the rest of the DP driver.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 6 ++++++
drivers/gpu/drm/msm/dp/dp_display.c | 12 ++++++++++++
drivers/gpu/drm/msm/msm_drv.h | 7 +++++++
3 files changed, 25 insertions(+)
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
index 359de04abf4bbead3daa5e8b357a3c34216e3e65..734d8972bbd65153778d5d70a55ac09dfc693ac9 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
@@ -664,6 +664,12 @@ static int _dpu_kms_initialize_displayport(struct drm_device *dev,
stream_cnt = msm_dp_get_mst_max_stream(priv->dp[i]);
if (stream_cnt > 1) {
+ rc = msm_dp_mst_register(priv->dp[i]);
+ if (rc) {
+ DPU_ERROR("dp_mst_init failed for DP, rc = %d\n", rc);
+ return rc;
+ }
+
for (stream_id = 0; stream_id < stream_cnt; stream_id++) {
info.stream_id = stream_id;
encoder = dpu_encoder_init(dev, DRM_MODE_ENCODER_DPMST, &info);
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index eafec9ab4f83cb44e861687e7550748b4d9b7ece..7f2eace17c126e3758c68bb0dee67662463a6e05 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -69,6 +69,8 @@ enum {
#define WAIT_FOR_RESUME_TIMEOUT_JIFFIES (HZ / 2)
+#define MAX_DPCD_TRANSACTION_BYTES 16
+
struct msm_dp_event {
u32 event_id;
u32 data;
@@ -1689,6 +1691,16 @@ int msm_dp_modeset_init(struct msm_dp *msm_dp_display, struct drm_device *dev,
return 0;
}
+int msm_dp_mst_register(struct msm_dp *dp)
+{
+ struct msm_dp_display_private *dp_display;
+
+ dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
+
+ return msm_dp_mst_init(dp, dp_display->max_stream,
+ MAX_DPCD_TRANSACTION_BYTES, dp_display->aux);
+}
+
void msm_dp_display_atomic_prepare(struct msm_dp *dp)
{
int rc = 0;
diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
index 12b50a797772f574122481cd8a1c7c88aacb8250..7ed0e25d6c2bc9e4e3d78895742226d22d103e4c 100644
--- a/drivers/gpu/drm/msm/msm_drv.h
+++ b/drivers/gpu/drm/msm/msm_drv.h
@@ -376,6 +376,8 @@ int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display);
int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder);
+int msm_dp_mst_register(struct msm_dp *dp_display);
+
#else
static inline int __init msm_dp_register(void)
{
@@ -397,6 +399,11 @@ static inline int msm_dp_get_mst_max_stream(struct msm_dp *dp_display)
return -EINVAL;
}
+static inline int msm_dp_mst_register(struct msm_dp *dp_display)
+{
+ return -EINVAL;
+}
+
int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
{
return -EINVAL;
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 39/45] drm/msm/dp: initialize dp_mst module for each DP MST controller
2024-12-06 4:32 ` [PATCH 39/45] drm/msm/dp: initialize dp_mst module for each DP MST controller Abhinav Kumar
@ 2024-12-08 6:31 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 6:31 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:10PM -0800, Abhinav Kumar wrote:
> For each MST capable DP controller, initialize a dp_mst module to
> manage its DP MST operations. The DP MST module for each controller
> is the central entity to manage its topology related operations as
> well as interfacing with the rest of the DP driver.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 6 ++++++
> drivers/gpu/drm/msm/dp/dp_display.c | 12 ++++++++++++
> drivers/gpu/drm/msm/msm_drv.h | 7 +++++++
> 3 files changed, 25 insertions(+)
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> index 359de04abf4bbead3daa5e8b357a3c34216e3e65..734d8972bbd65153778d5d70a55ac09dfc693ac9 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> @@ -664,6 +664,12 @@ static int _dpu_kms_initialize_displayport(struct drm_device *dev,
> stream_cnt = msm_dp_get_mst_max_stream(priv->dp[i]);
>
> if (stream_cnt > 1) {
> + rc = msm_dp_mst_register(priv->dp[i]);
> + if (rc) {
> + DPU_ERROR("dp_mst_init failed for DP, rc = %d\n", rc);
> + return rc;
> + }
> +
This seems to be put backwards. I'd prefer if DPU driver allocated all
MST encoders and then called into the DP driver, letting it handle all
MST-related internals.
> for (stream_id = 0; stream_id < stream_cnt; stream_id++) {
> info.stream_id = stream_id;
> encoder = dpu_encoder_init(dev, DRM_MODE_ENCODER_DPMST, &info);
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index eafec9ab4f83cb44e861687e7550748b4d9b7ece..7f2eace17c126e3758c68bb0dee67662463a6e05 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -69,6 +69,8 @@ enum {
>
> #define WAIT_FOR_RESUME_TIMEOUT_JIFFIES (HZ / 2)
>
> +#define MAX_DPCD_TRANSACTION_BYTES 16
> +
> struct msm_dp_event {
> u32 event_id;
> u32 data;
> @@ -1689,6 +1691,16 @@ int msm_dp_modeset_init(struct msm_dp *msm_dp_display, struct drm_device *dev,
> return 0;
> }
>
> +int msm_dp_mst_register(struct msm_dp *dp)
> +{
> + struct msm_dp_display_private *dp_display;
> +
> + dp_display = container_of(dp, struct msm_dp_display_private, msm_dp_display);
> +
> + return msm_dp_mst_init(dp, dp_display->max_stream,
> + MAX_DPCD_TRANSACTION_BYTES, dp_display->aux);
This doesn't seem to change between plaforms. Please push it to
msm_dp_mst_init() instead.
> +}
> +
> void msm_dp_display_atomic_prepare(struct msm_dp *dp)
> {
> int rc = 0;
> diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
> index 12b50a797772f574122481cd8a1c7c88aacb8250..7ed0e25d6c2bc9e4e3d78895742226d22d103e4c 100644
> --- a/drivers/gpu/drm/msm/msm_drv.h
> +++ b/drivers/gpu/drm/msm/msm_drv.h
> @@ -376,6 +376,8 @@ int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display);
>
> int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder);
>
> +int msm_dp_mst_register(struct msm_dp *dp_display);
> +
> #else
> static inline int __init msm_dp_register(void)
> {
> @@ -397,6 +399,11 @@ static inline int msm_dp_get_mst_max_stream(struct msm_dp *dp_display)
> return -EINVAL;
> }
>
> +static inline int msm_dp_mst_register(struct msm_dp *dp_display)
> +{
> + return -EINVAL;
> +}
> +
> int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
> {
> return -EINVAL;
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 40/45] drm/msm: add a stream to intf map for DP controller
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (38 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 39/45] drm/msm/dp: initialize dp_mst module for each DP MST controller Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-06 9:20 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 41/45] drm/msm/dpu: use msm_dp_get_mst_intf_id() to get the intf id Abhinav Kumar
` (6 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Each DP controller capable of MST can support multiple streams
and each of the streams maps to an interface block ID which can
vary based on chipset. Add a stream to interface map for MST capable
DP controllers.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 46 +++++++++++++++++++++++++++++++++++--
drivers/gpu/drm/msm/msm_drv.h | 7 ++++++
2 files changed, 51 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 7f2eace17c126e3758c68bb0dee67662463a6e05..caac0cd3ec94e7be1389d8129fbd506998cf77da 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -116,6 +116,8 @@ struct msm_dp_display_private {
u32 active_stream_cnt;
+ const unsigned int *intf_map;
+
struct msm_dp_audio *audio;
};
@@ -123,11 +125,36 @@ struct msm_dp_desc {
phys_addr_t io_start;
unsigned int id;
bool wide_bus_supported;
+ const unsigned int *intf_map;
+};
+
+/* to be kept in sync with enum dpu_intf of dpu_hw_mdss.h */
+enum dp_mst_intf {
+ INTF_0 = 1,
+ INTF_1,
+ INTF_2,
+ INTF_3,
+ INTF_4,
+ INTF_5,
+ INTF_6,
+ INTF_7,
+ INTF_8,
+ INTF_MAX
+};
+
+static const unsigned int stream_intf_map_sa_8775p[][DP_STREAM_MAX] = {
+ {INTF_0, INTF_3},
+ {INTF_4, INTF_8},
+ {}
};
static const struct msm_dp_desc msm_dp_desc_sa8775p[] = {
- { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true },
- { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true },
+ { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true,
+ .intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_0],
+ },
+ { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true,
+ .intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_1],
+ },
{ .io_start = 0x22154000, .id = MSM_DP_CONTROLLER_2, .wide_bus_supported = true },
{ .io_start = 0x2215c000, .id = MSM_DP_CONTROLLER_3, .wide_bus_supported = true },
{}
@@ -1489,6 +1516,9 @@ static int msm_dp_display_probe(struct platform_device *pdev)
(dp->msm_dp_display.connector_type == DRM_MODE_CONNECTOR_eDP);
dp->max_stream = DEFAULT_STREAM_COUNT;
+
+ dp->intf_map = desc->intf_map;
+
rc = msm_dp_init_sub_modules(dp);
if (rc) {
DRM_ERROR("init sub module failed\n");
@@ -1646,6 +1676,18 @@ bool msm_dp_wide_bus_available(const struct msm_dp *msm_dp_display)
return dp->wide_bus_supported;
}
+int msm_dp_get_mst_intf_id(struct msm_dp *dp_display, int stream_id)
+{
+ struct msm_dp_display_private *dp;
+
+ dp = container_of(dp_display, struct msm_dp_display_private, msm_dp_display);
+
+ if (dp->intf_map)
+ return dp->intf_map[stream_id];
+
+ return 0;
+}
+
void msm_dp_display_debugfs_init(struct msm_dp *msm_dp_display, struct dentry *root, bool is_edp)
{
struct msm_dp_display_private *dp;
diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
index 7ed0e25d6c2bc9e4e3d78895742226d22d103e4c..50719e188732acd3652e4a7063d1ba1e2963b48a 100644
--- a/drivers/gpu/drm/msm/msm_drv.h
+++ b/drivers/gpu/drm/msm/msm_drv.h
@@ -378,6 +378,8 @@ int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encode
int msm_dp_mst_register(struct msm_dp *dp_display);
+int msm_dp_get_mst_intf_id(struct msm_dp *dp_display, int stream_id);
+
#else
static inline int __init msm_dp_register(void)
{
@@ -430,6 +432,11 @@ static inline bool msm_dp_wide_bus_available(const struct msm_dp *dp_display)
return false;
}
+int msm_dp_get_mst_intf_id(struct msm_dp *dp_display, int stream_id)
+{
+ return -EINVAL;
+}
+
#endif
#ifdef CONFIG_DRM_MSM_MDP4
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 40/45] drm/msm: add a stream to intf map for DP controller
2024-12-06 4:32 ` [PATCH 40/45] drm/msm: add a stream to intf map for DP controller Abhinav Kumar
@ 2024-12-06 9:20 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:20 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:11PM -0800, Abhinav Kumar wrote:
> Each DP controller capable of MST can support multiple streams
> and each of the streams maps to an interface block ID which can
> vary based on chipset. Add a stream to interface map for MST capable
> DP controllers.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 46 +++++++++++++++++++++++++++++++++++--
> drivers/gpu/drm/msm/msm_drv.h | 7 ++++++
> 2 files changed, 51 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 7f2eace17c126e3758c68bb0dee67662463a6e05..caac0cd3ec94e7be1389d8129fbd506998cf77da 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -116,6 +116,8 @@ struct msm_dp_display_private {
>
> u32 active_stream_cnt;
>
> + const unsigned int *intf_map;
> +
> struct msm_dp_audio *audio;
> };
>
> @@ -123,11 +125,36 @@ struct msm_dp_desc {
> phys_addr_t io_start;
> unsigned int id;
> bool wide_bus_supported;
> + const unsigned int *intf_map;
> +};
> +
> +/* to be kept in sync with enum dpu_intf of dpu_hw_mdss.h */
This points out that it's not the best place to handle the mapping.
Please move the mapping to the DPU's hw_catalog instead.
> +enum dp_mst_intf {
> + INTF_0 = 1,
> + INTF_1,
> + INTF_2,
> + INTF_3,
> + INTF_4,
> + INTF_5,
> + INTF_6,
> + INTF_7,
> + INTF_8,
> + INTF_MAX
> +};
> +
> +static const unsigned int stream_intf_map_sa_8775p[][DP_STREAM_MAX] = {
> + {INTF_0, INTF_3},
> + {INTF_4, INTF_8},
> + {}
> };
>
> static const struct msm_dp_desc msm_dp_desc_sa8775p[] = {
> - { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true },
> - { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true },
> + { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true,
> + .intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_0],
> + },
> + { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true,
> + .intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_1],
> + },
> { .io_start = 0x22154000, .id = MSM_DP_CONTROLLER_2, .wide_bus_supported = true },
> { .io_start = 0x2215c000, .id = MSM_DP_CONTROLLER_3, .wide_bus_supported = true },
> {}
> @@ -1489,6 +1516,9 @@ static int msm_dp_display_probe(struct platform_device *pdev)
> (dp->msm_dp_display.connector_type == DRM_MODE_CONNECTOR_eDP);
>
> dp->max_stream = DEFAULT_STREAM_COUNT;
> +
> + dp->intf_map = desc->intf_map;
> +
> rc = msm_dp_init_sub_modules(dp);
> if (rc) {
> DRM_ERROR("init sub module failed\n");
> @@ -1646,6 +1676,18 @@ bool msm_dp_wide_bus_available(const struct msm_dp *msm_dp_display)
> return dp->wide_bus_supported;
> }
>
> +int msm_dp_get_mst_intf_id(struct msm_dp *dp_display, int stream_id)
> +{
> + struct msm_dp_display_private *dp;
> +
> + dp = container_of(dp_display, struct msm_dp_display_private, msm_dp_display);
> +
> + if (dp->intf_map)
> + return dp->intf_map[stream_id];
> +
> + return 0;
> +}
> +
> void msm_dp_display_debugfs_init(struct msm_dp *msm_dp_display, struct dentry *root, bool is_edp)
> {
> struct msm_dp_display_private *dp;
> diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
> index 7ed0e25d6c2bc9e4e3d78895742226d22d103e4c..50719e188732acd3652e4a7063d1ba1e2963b48a 100644
> --- a/drivers/gpu/drm/msm/msm_drv.h
> +++ b/drivers/gpu/drm/msm/msm_drv.h
> @@ -378,6 +378,8 @@ int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encode
>
> int msm_dp_mst_register(struct msm_dp *dp_display);
>
> +int msm_dp_get_mst_intf_id(struct msm_dp *dp_display, int stream_id);
> +
> #else
> static inline int __init msm_dp_register(void)
> {
> @@ -430,6 +432,11 @@ static inline bool msm_dp_wide_bus_available(const struct msm_dp *dp_display)
> return false;
> }
>
> +int msm_dp_get_mst_intf_id(struct msm_dp *dp_display, int stream_id)
> +{
> + return -EINVAL;
> +}
> +
> #endif
>
> #ifdef CONFIG_DRM_MSM_MDP4
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 41/45] drm/msm/dpu: use msm_dp_get_mst_intf_id() to get the intf id
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (39 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 40/45] drm/msm: add a stream to intf map for DP controller Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-06 9:25 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 42/45] drm/msm/dp: mark ST_DISCONNECTED only if all streams are disabled Abhinav Kumar
` (5 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Use msm_dp_get_mst_intf_id() to get the intf id for the DP MST
controller as the intf_id is unique for each MST stream of each
DP controller.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 25 ++++++++++++++++++++-----
1 file changed, 20 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
index 83de7564e2c1fe14fcf8c4f82335cafc937e1b99..ce2f0fa8ebb1efd1a229a99543593965dbccd752 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
@@ -2512,17 +2512,23 @@ static int dpu_encoder_virt_add_phys_encs(
static int dpu_encoder_setup_display(struct dpu_encoder_virt *dpu_enc,
struct dpu_kms *dpu_kms,
- struct msm_display_info *disp_info)
+ struct msm_display_info *disp_info,
+ int drm_enc_mode)
{
int ret = 0;
int i = 0;
struct dpu_enc_phys_init_params phys_params;
+ unsigned int intf_id;
+ struct msm_drm_private *priv;
+ struct drm_encoder *drm_enc;
if (!dpu_enc) {
DPU_ERROR("invalid arg(s), enc %d\n", dpu_enc != NULL);
return -EINVAL;
}
+ drm_enc = &dpu_enc->base;
+ priv = drm_enc->dev->dev_private;
dpu_enc->cur_master = NULL;
memset(&phys_params, 0, sizeof(phys_params));
@@ -2559,9 +2565,18 @@ static int dpu_encoder_setup_display(struct dpu_encoder_virt *dpu_enc,
DPU_DEBUG("h_tile_instance %d = %d, split_role %d\n",
i, controller_id, phys_params.split_role);
- phys_params.hw_intf = dpu_encoder_get_intf(dpu_kms->catalog, &dpu_kms->rm,
- disp_info->intf_type,
- controller_id);
+ if (drm_enc_mode == DRM_MODE_ENCODER_DPMST) {
+ intf_id = msm_dp_get_mst_intf_id(priv->dp[controller_id],
+ disp_info->stream_id);
+ DPU_DEBUG("intf_id %d for disp_info->stream_id = %d\n", intf_id,
+ disp_info->stream_id);
+ phys_params.hw_intf = dpu_rm_get_intf(&dpu_kms->rm, intf_id);
+
+ } else {
+ phys_params.hw_intf = dpu_encoder_get_intf(dpu_kms->catalog, &dpu_kms->rm,
+ disp_info->intf_type,
+ controller_id);
+ }
if (disp_info->intf_type == INTF_WB && controller_id < WB_MAX)
phys_params.hw_wb = dpu_rm_get_wb(&dpu_kms->rm, controller_id);
@@ -2662,7 +2677,7 @@ struct drm_encoder *dpu_encoder_init(struct drm_device *dev,
mutex_init(&dpu_enc->enc_lock);
mutex_init(&dpu_enc->rc_lock);
- ret = dpu_encoder_setup_display(dpu_enc, dpu_kms, disp_info);
+ ret = dpu_encoder_setup_display(dpu_enc, dpu_kms, disp_info, drm_enc_mode);
if (ret) {
DPU_ERROR("failed to setup encoder\n");
return ERR_PTR(-ENOMEM);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 41/45] drm/msm/dpu: use msm_dp_get_mst_intf_id() to get the intf id
2024-12-06 4:32 ` [PATCH 41/45] drm/msm/dpu: use msm_dp_get_mst_intf_id() to get the intf id Abhinav Kumar
@ 2024-12-06 9:25 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-06 9:25 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:12PM -0800, Abhinav Kumar wrote:
> Use msm_dp_get_mst_intf_id() to get the intf id for the DP MST
> controller as the intf_id is unique for each MST stream of each
> DP controller.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 25 ++++++++++++++++++++-----
> 1 file changed, 20 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> index 83de7564e2c1fe14fcf8c4f82335cafc937e1b99..ce2f0fa8ebb1efd1a229a99543593965dbccd752 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> @@ -2512,17 +2512,23 @@ static int dpu_encoder_virt_add_phys_encs(
>
> static int dpu_encoder_setup_display(struct dpu_encoder_virt *dpu_enc,
> struct dpu_kms *dpu_kms,
> - struct msm_display_info *disp_info)
> + struct msm_display_info *disp_info,
> + int drm_enc_mode)
No need to, use dpu_enc->base.encoder_type
> {
> int ret = 0;
> int i = 0;
> struct dpu_enc_phys_init_params phys_params;
> + unsigned int intf_id;
> + struct msm_drm_private *priv;
> + struct drm_encoder *drm_enc;
>
> if (!dpu_enc) {
> DPU_ERROR("invalid arg(s), enc %d\n", dpu_enc != NULL);
> return -EINVAL;
> }
>
> + drm_enc = &dpu_enc->base;
> + priv = drm_enc->dev->dev_private;
> dpu_enc->cur_master = NULL;
>
> memset(&phys_params, 0, sizeof(phys_params));
> @@ -2559,9 +2565,18 @@ static int dpu_encoder_setup_display(struct dpu_encoder_virt *dpu_enc,
> DPU_DEBUG("h_tile_instance %d = %d, split_role %d\n",
> i, controller_id, phys_params.split_role);
>
> - phys_params.hw_intf = dpu_encoder_get_intf(dpu_kms->catalog, &dpu_kms->rm,
> - disp_info->intf_type,
> - controller_id);
> + if (drm_enc_mode == DRM_MODE_ENCODER_DPMST) {
> + intf_id = msm_dp_get_mst_intf_id(priv->dp[controller_id],
> + disp_info->stream_id);
> + DPU_DEBUG("intf_id %d for disp_info->stream_id = %d\n", intf_id,
> + disp_info->stream_id);
> + phys_params.hw_intf = dpu_rm_get_intf(&dpu_kms->rm, intf_id);
> +
> + } else {
> + phys_params.hw_intf = dpu_encoder_get_intf(dpu_kms->catalog, &dpu_kms->rm,
> + disp_info->intf_type,
> + controller_id);
> + }
>
> if (disp_info->intf_type == INTF_WB && controller_id < WB_MAX)
> phys_params.hw_wb = dpu_rm_get_wb(&dpu_kms->rm, controller_id);
> @@ -2662,7 +2677,7 @@ struct drm_encoder *dpu_encoder_init(struct drm_device *dev,
> mutex_init(&dpu_enc->enc_lock);
> mutex_init(&dpu_enc->rc_lock);
>
> - ret = dpu_encoder_setup_display(dpu_enc, dpu_kms, disp_info);
> + ret = dpu_encoder_setup_display(dpu_enc, dpu_kms, disp_info, drm_enc_mode);
> if (ret) {
> DPU_ERROR("failed to setup encoder\n");
> return ERR_PTR(-ENOMEM);
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 42/45] drm/msm/dp: mark ST_DISCONNECTED only if all streams are disabled
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (40 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 41/45] drm/msm/dpu: use msm_dp_get_mst_intf_id() to get the intf id Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 11:58 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 43/45] drm/msm/dp: populate the max_streams for sa8775 mst controller Abhinav Kumar
` (4 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
HPD state machine assumes only one active stream. Fix it to account
for both while marking the state of the hpd as disconnected.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_display.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index caac0cd3ec94e7be1389d8129fbd506998cf77da..528cda1453a85f5f8dd37bb0d2366548016c88e4 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -1890,7 +1890,7 @@ void msm_dp_display_atomic_post_disable_helper(struct msm_dp *dp, struct msm_dp_
msm_dp_display_disable(msm_dp_display, msm_dp_panel);
state = msm_dp_display->hpd_state;
- if (state == ST_DISCONNECT_PENDING) {
+ if (state == ST_DISCONNECT_PENDING && !msm_dp_display->active_stream_cnt) {
/* completed disconnection */
msm_dp_display->hpd_state = ST_DISCONNECTED;
}
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 42/45] drm/msm/dp: mark ST_DISCONNECTED only if all streams are disabled
2024-12-06 4:32 ` [PATCH 42/45] drm/msm/dp: mark ST_DISCONNECTED only if all streams are disabled Abhinav Kumar
@ 2024-12-08 11:58 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 11:58 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:13PM -0800, Abhinav Kumar wrote:
> HPD state machine assumes only one active stream. Fix it to account
> for both while marking the state of the hpd as disconnected.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_display.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 43/45] drm/msm/dp: populate the max_streams for sa8775 mst controller
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (41 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 42/45] drm/msm/dp: mark ST_DISCONNECTED only if all streams are disabled Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-08 12:00 ` Dmitry Baryshkov
2024-12-06 4:32 ` [PATCH 44/45] arm64: dts: qcom: add mst support for pixel stream clk for DP0 Abhinav Kumar
` (3 subsequent siblings)
46 siblings, 1 reply; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Populate the max_streams for DP controllers to indicate MST support.
If the pixel clock handle for the second stream fails, treat it as
SST case rather than a complete failure.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
drivers/gpu/drm/msm/dp/dp_ctrl.c | 9 +++++++++
drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 ++
drivers/gpu/drm/msm/dp/dp_display.c | 13 ++++++++++---
3 files changed, 21 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
index d4915a962f97b3d2a347456e197265a5dc043eb0..0e5ebcdab00936cf0ce6e35514f2932158530cc0 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
@@ -2438,6 +2438,15 @@ static int msm_dp_ctrl_clk_init(struct msm_dp_ctrl *msm_dp_ctrl)
return 0;
}
+int msm_dp_ctrl_get_stream_cnt(struct msm_dp_ctrl *msm_dp_ctrl)
+{
+ struct msm_dp_ctrl_private *ctrl;
+
+ ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
+
+ return ctrl->num_pixel_clks;
+}
+
struct msm_dp_ctrl *msm_dp_ctrl_get(struct device *dev, struct msm_dp_link *link,
struct msm_dp_panel *panel, struct drm_dp_aux *aux,
struct msm_dp_catalog *catalog,
diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
index 9ad7022d6217572395d69294c3cc4d4dbaddf0ac..f72a499809ec01a9cc5bb79762bc2604820426a1 100644
--- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
+++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
@@ -54,4 +54,6 @@ void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
u32 start_slot, u32 tot_slots);
void msm_dp_ctrl_push_vcpf(struct msm_dp_ctrl *dp_ctrl, struct msm_dp_panel *msm_dp_panel);
+int msm_dp_ctrl_get_stream_cnt(struct msm_dp_ctrl *dp_ctrl);
+
#endif /* _DP_CTRL_H_ */
diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
index 528cda1453a85f5f8dd37bb0d2366548016c88e4..db563af19461089dbcaccce0d03fd03d5aa567aa 100644
--- a/drivers/gpu/drm/msm/dp/dp_display.c
+++ b/drivers/gpu/drm/msm/dp/dp_display.c
@@ -126,6 +126,7 @@ struct msm_dp_desc {
unsigned int id;
bool wide_bus_supported;
const unsigned int *intf_map;
+ unsigned int max_streams;
};
/* to be kept in sync with enum dpu_intf of dpu_hw_mdss.h */
@@ -149,10 +150,10 @@ static const unsigned int stream_intf_map_sa_8775p[][DP_STREAM_MAX] = {
};
static const struct msm_dp_desc msm_dp_desc_sa8775p[] = {
- { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true,
+ { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true, .max_streams = 2,
.intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_0],
},
- { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true,
+ { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true, .max_streams = 2,
.intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_1],
},
{ .io_start = 0x22154000, .id = MSM_DP_CONTROLLER_2, .wide_bus_supported = true },
@@ -1481,7 +1482,10 @@ int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display)
dp_priv = container_of(dp_display, struct msm_dp_display_private, msm_dp_display);
- return dp_priv->max_stream;
+ if (dp_priv->max_stream == msm_dp_ctrl_get_stream_cnt(dp_priv->ctrl))
+ return dp_priv->max_stream;
+ else
+ return DEFAULT_STREAM_COUNT;
}
int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
@@ -1517,6 +1521,9 @@ static int msm_dp_display_probe(struct platform_device *pdev)
dp->max_stream = DEFAULT_STREAM_COUNT;
+ if (desc->max_streams > DEFAULT_STREAM_COUNT)
+ dp->max_stream = desc->max_streams;
+
dp->intf_map = desc->intf_map;
rc = msm_dp_init_sub_modules(dp);
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 43/45] drm/msm/dp: populate the max_streams for sa8775 mst controller
2024-12-06 4:32 ` [PATCH 43/45] drm/msm/dp: populate the max_streams for sa8775 mst controller Abhinav Kumar
@ 2024-12-08 12:00 ` Dmitry Baryshkov
0 siblings, 0 replies; 111+ messages in thread
From: Dmitry Baryshkov @ 2024-12-08 12:00 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Sean Paul, Marijn Suijten, David Airlie, Simona Vetter,
Stephen Boyd, Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh,
Bjorn Andersson, Konrad Dybcio, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm,
dri-devel, freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart
On Thu, Dec 05, 2024 at 08:32:14PM -0800, Abhinav Kumar wrote:
> Populate the max_streams for DP controllers to indicate MST support.
>
> If the pixel clock handle for the second stream fails, treat it as
> SST case rather than a complete failure.
Could you possibly clarify:
- Are there MSM DP controllers which do not support MST?
- Are there MSM DP controller which support more than 2 streams?
It might be easier to to set max_streams statically rather than setting
it via the descs.
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 9 +++++++++
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 2 ++
> drivers/gpu/drm/msm/dp/dp_display.c | 13 ++++++++++---
> 3 files changed, 21 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> index d4915a962f97b3d2a347456e197265a5dc043eb0..0e5ebcdab00936cf0ce6e35514f2932158530cc0 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c
> @@ -2438,6 +2438,15 @@ static int msm_dp_ctrl_clk_init(struct msm_dp_ctrl *msm_dp_ctrl)
> return 0;
> }
>
> +int msm_dp_ctrl_get_stream_cnt(struct msm_dp_ctrl *msm_dp_ctrl)
> +{
> + struct msm_dp_ctrl_private *ctrl;
> +
> + ctrl = container_of(msm_dp_ctrl, struct msm_dp_ctrl_private, msm_dp_ctrl);
> +
> + return ctrl->num_pixel_clks;
> +}
> +
> struct msm_dp_ctrl *msm_dp_ctrl_get(struct device *dev, struct msm_dp_link *link,
> struct msm_dp_panel *panel, struct drm_dp_aux *aux,
> struct msm_dp_catalog *catalog,
> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.h b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> index 9ad7022d6217572395d69294c3cc4d4dbaddf0ac..f72a499809ec01a9cc5bb79762bc2604820426a1 100644
> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.h
> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.h
> @@ -54,4 +54,6 @@ void msm_dp_ctrl_set_mst_channel_info(struct msm_dp_ctrl *msm_dp_ctrl,
> u32 start_slot, u32 tot_slots);
> void msm_dp_ctrl_push_vcpf(struct msm_dp_ctrl *dp_ctrl, struct msm_dp_panel *msm_dp_panel);
>
> +int msm_dp_ctrl_get_stream_cnt(struct msm_dp_ctrl *dp_ctrl);
> +
> #endif /* _DP_CTRL_H_ */
> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> index 528cda1453a85f5f8dd37bb0d2366548016c88e4..db563af19461089dbcaccce0d03fd03d5aa567aa 100644
> --- a/drivers/gpu/drm/msm/dp/dp_display.c
> +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> @@ -126,6 +126,7 @@ struct msm_dp_desc {
> unsigned int id;
> bool wide_bus_supported;
> const unsigned int *intf_map;
> + unsigned int max_streams;
> };
>
> /* to be kept in sync with enum dpu_intf of dpu_hw_mdss.h */
> @@ -149,10 +150,10 @@ static const unsigned int stream_intf_map_sa_8775p[][DP_STREAM_MAX] = {
> };
>
> static const struct msm_dp_desc msm_dp_desc_sa8775p[] = {
> - { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true,
> + { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true, .max_streams = 2,
> .intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_0],
> },
> - { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true,
> + { .io_start = 0x0af5c000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true, .max_streams = 2,
> .intf_map = stream_intf_map_sa_8775p[MSM_DP_CONTROLLER_1],
> },
> { .io_start = 0x22154000, .id = MSM_DP_CONTROLLER_2, .wide_bus_supported = true },
> @@ -1481,7 +1482,10 @@ int msm_dp_get_mst_max_stream(const struct msm_dp *dp_display)
>
> dp_priv = container_of(dp_display, struct msm_dp_display_private, msm_dp_display);
>
> - return dp_priv->max_stream;
> + if (dp_priv->max_stream == msm_dp_ctrl_get_stream_cnt(dp_priv->ctrl))
> + return dp_priv->max_stream;
> + else
> + return DEFAULT_STREAM_COUNT;
> }
>
> int msm_dp_mst_bridge_init(struct msm_dp *dp_display, struct drm_encoder *encoder)
> @@ -1517,6 +1521,9 @@ static int msm_dp_display_probe(struct platform_device *pdev)
>
> dp->max_stream = DEFAULT_STREAM_COUNT;
>
> + if (desc->max_streams > DEFAULT_STREAM_COUNT)
> + dp->max_stream = desc->max_streams;
> +
> dp->intf_map = desc->intf_map;
>
> rc = msm_dp_init_sub_modules(dp);
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 44/45] arm64: dts: qcom: add mst support for pixel stream clk for DP0
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (42 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 43/45] drm/msm/dp: populate the max_streams for sa8775 mst controller Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2024-12-06 12:20 ` Konrad Dybcio
2025-01-07 0:52 ` Bjorn Andersson
2024-12-06 4:32 ` [PATCH 45/45] arm64: dts: qcom: add mst support for pixel 1 stream clk for DP1 Abhinav Kumar
` (2 subsequent siblings)
46 siblings, 2 replies; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar, Yongxing Mou
From: Yongxing Mou <quic_yongmou@quicinc.com>
Populate the pixel clock for stream 1 for DP0 for sa8775p DP controller.
Signed-off-by: Yongxing Mou <quic_yongmou@quicinc.com>
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
arch/arm64/boot/dts/qcom/sa8775p.dtsi | 12 ++++++++----
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/boot/dts/qcom/sa8775p.dtsi b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
index 0dbaa17e5e3f06c61b2aa777e45b73a48e50e66b..0150ce27b98e9894fa9ee6cccd020528d716f543 100644
--- a/arch/arm64/boot/dts/qcom/sa8775p.dtsi
+++ b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
@@ -3944,16 +3944,20 @@ mdss0_dp0: displayport-controller@af54000 {
<&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_AUX_CLK>,
<&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_CLK>,
<&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_INTF_CLK>,
- <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK>;
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK>,
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL1_CLK>;
clock-names = "core_iface",
"core_aux",
"ctrl_link",
"ctrl_link_iface",
- "stream_pixel";
+ "stream_pixel",
+ "stream_1_pixel";
assigned-clocks = <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_CLK_SRC>,
- <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK_SRC>;
- assigned-clock-parents = <&mdss0_dp0_phy 0>, <&mdss0_dp0_phy 1>;
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK_SRC>,
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL1_CLK_SRC>;
+ assigned-clock-parents = <&mdss0_dp0_phy 0>, <&mdss0_dp0_phy 1>, <&mdss0_dp0_phy 1>;
phys = <&mdss0_dp0_phy>;
+
phy-names = "dp";
operating-points-v2 = <&dp_opp_table>;
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 44/45] arm64: dts: qcom: add mst support for pixel stream clk for DP0
2024-12-06 4:32 ` [PATCH 44/45] arm64: dts: qcom: add mst support for pixel stream clk for DP0 Abhinav Kumar
@ 2024-12-06 12:20 ` Konrad Dybcio
2025-01-07 0:52 ` Bjorn Andersson
1 sibling, 0 replies; 111+ messages in thread
From: Konrad Dybcio @ 2024-12-06 12:20 UTC (permalink / raw)
To: Abhinav Kumar, Rob Clark, Dmitry Baryshkov, Sean Paul,
Marijn Suijten, David Airlie, Simona Vetter, Stephen Boyd,
Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh, Bjorn Andersson,
Konrad Dybcio, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Yongxing Mou
On 6.12.2024 5:32 AM, Abhinav Kumar wrote:
> From: Yongxing Mou <quic_yongmou@quicinc.com>
>
> Populate the pixel clock for stream 1 for DP0 for sa8775p DP controller.
>
> Signed-off-by: Yongxing Mou <quic_yongmou@quicinc.com>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> arch/arm64/boot/dts/qcom/sa8775p.dtsi | 12 ++++++++----
> 1 file changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/arch/arm64/boot/dts/qcom/sa8775p.dtsi b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
> index 0dbaa17e5e3f06c61b2aa777e45b73a48e50e66b..0150ce27b98e9894fa9ee6cccd020528d716f543 100644
> --- a/arch/arm64/boot/dts/qcom/sa8775p.dtsi
> +++ b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
> @@ -3944,16 +3944,20 @@ mdss0_dp0: displayport-controller@af54000 {
> <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_AUX_CLK>,
> <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_CLK>,
> <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_INTF_CLK>,
> - <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK>;
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK>,
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL1_CLK>;
dispcc also defines PIXEL2/3 clocks.
> clock-names = "core_iface",
> "core_aux",
> "ctrl_link",
> "ctrl_link_iface",
> - "stream_pixel";
> + "stream_pixel",
> + "stream_1_pixel";
> assigned-clocks = <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_CLK_SRC>,
> - <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK_SRC>;
> - assigned-clock-parents = <&mdss0_dp0_phy 0>, <&mdss0_dp0_phy 1>;
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK_SRC>,
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL1_CLK_SRC>;
> + assigned-clock-parents = <&mdss0_dp0_phy 0>, <&mdss0_dp0_phy 1>, <&mdss0_dp0_phy 1>;
Please turn this into a vertical list
Konrad
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 44/45] arm64: dts: qcom: add mst support for pixel stream clk for DP0
2024-12-06 4:32 ` [PATCH 44/45] arm64: dts: qcom: add mst support for pixel stream clk for DP0 Abhinav Kumar
2024-12-06 12:20 ` Konrad Dybcio
@ 2025-01-07 0:52 ` Bjorn Andersson
1 sibling, 0 replies; 111+ messages in thread
From: Bjorn Andersson @ 2025-01-07 0:52 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart, Yongxing Mou
On Thu, Dec 05, 2024 at 08:32:15PM -0800, Abhinav Kumar wrote:
> From: Yongxing Mou <quic_yongmou@quicinc.com>
>
I'd expect "sa8775p" in the subject prefix.
> Populate the pixel clock for stream 1 for DP0 for sa8775p DP controller.
Please write your commit messages in the style expressed in
https://docs.kernel.org/process/submitting-patches.html#describe-your-changes
Use the commit message to document why the code/dt looks like it does.
Describe the problem your solving and why.
>
> Signed-off-by: Yongxing Mou <quic_yongmou@quicinc.com>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> arch/arm64/boot/dts/qcom/sa8775p.dtsi | 12 ++++++++----
> 1 file changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/arch/arm64/boot/dts/qcom/sa8775p.dtsi b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
> index 0dbaa17e5e3f06c61b2aa777e45b73a48e50e66b..0150ce27b98e9894fa9ee6cccd020528d716f543 100644
> --- a/arch/arm64/boot/dts/qcom/sa8775p.dtsi
> +++ b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
> @@ -3944,16 +3944,20 @@ mdss0_dp0: displayport-controller@af54000 {
> <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_AUX_CLK>,
> <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_CLK>,
> <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_INTF_CLK>,
> - <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK>;
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK>,
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL1_CLK>;
> clock-names = "core_iface",
> "core_aux",
> "ctrl_link",
> "ctrl_link_iface",
> - "stream_pixel";
> + "stream_pixel",
> + "stream_1_pixel";
I don't see this being a valid clock-names in the DT binding, does this
pass dtbs_check?
Regards,
Bjorn
> assigned-clocks = <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_LINK_CLK_SRC>,
> - <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK_SRC>;
> - assigned-clock-parents = <&mdss0_dp0_phy 0>, <&mdss0_dp0_phy 1>;
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL0_CLK_SRC>,
> + <&dispcc0 MDSS_DISP_CC_MDSS_DPTX0_PIXEL1_CLK_SRC>;
> + assigned-clock-parents = <&mdss0_dp0_phy 0>, <&mdss0_dp0_phy 1>, <&mdss0_dp0_phy 1>;
> phys = <&mdss0_dp0_phy>;
> +
> phy-names = "dp";
>
> operating-points-v2 = <&dp_opp_table>;
>
> --
> 2.34.1
>
^ permalink raw reply [flat|nested] 111+ messages in thread
* [PATCH 45/45] arm64: dts: qcom: add mst support for pixel 1 stream clk for DP1
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (43 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 44/45] arm64: dts: qcom: add mst support for pixel stream clk for DP0 Abhinav Kumar
@ 2024-12-06 4:32 ` Abhinav Kumar
2025-01-07 1:06 ` [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Bjorn Andersson
2025-05-10 12:12 ` Jens Glathe
46 siblings, 0 replies; 111+ messages in thread
From: Abhinav Kumar @ 2024-12-06 4:32 UTC (permalink / raw)
To: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Bjorn Andersson, Konrad Dybcio,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Abhinav Kumar
Populate the pixel clock for stream 1 for DP1 for sa8775p DP controller.
Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
---
arch/arm64/boot/dts/qcom/sa8775p.dtsi | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/boot/dts/qcom/sa8775p.dtsi b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
index 0150ce27b98e9894fa9ee6cccd020528d716f543..91149f8b3adb93ece159f30bfea39f9725b6c9e8 100644
--- a/arch/arm64/boot/dts/qcom/sa8775p.dtsi
+++ b/arch/arm64/boot/dts/qcom/sa8775p.dtsi
@@ -4027,15 +4027,18 @@ mdss0_dp1: displayport-controller@af5c000 {
<&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_AUX_CLK>,
<&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_LINK_CLK>,
<&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_LINK_INTF_CLK>,
- <&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_PIXEL0_CLK>;
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_PIXEL0_CLK>,
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_PIXEL1_CLK>;
clock-names = "core_iface",
"core_aux",
"ctrl_link",
"ctrl_link_iface",
- "stream_pixel";
+ "stream_pixel",
+ "stream_1_pixel";
assigned-clocks = <&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_LINK_CLK_SRC>,
- <&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_PIXEL0_CLK_SRC>;
- assigned-clock-parents = <&mdss0_dp1_phy 0>, <&mdss0_dp1_phy 1>;
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_PIXEL0_CLK_SRC>,
+ <&dispcc0 MDSS_DISP_CC_MDSS_DPTX1_PIXEL1_CLK_SRC>;
+ assigned-clock-parents = <&mdss0_dp1_phy 0>, <&mdss0_dp1_phy 1>, <&mdss0_dp1_phy 1>;
phys = <&mdss0_dp1_phy>;
phy-names = "dp";
--
2.34.1
^ permalink raw reply related [flat|nested] 111+ messages in thread* Re: [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (44 preceding siblings ...)
2024-12-06 4:32 ` [PATCH 45/45] arm64: dts: qcom: add mst support for pixel 1 stream clk for DP1 Abhinav Kumar
@ 2025-01-07 1:06 ` Bjorn Andersson
2025-01-07 1:10 ` Abhinav Kumar
2025-05-10 12:12 ` Jens Glathe
46 siblings, 1 reply; 111+ messages in thread
From: Bjorn Andersson @ 2025-01-07 1:06 UTC (permalink / raw)
To: Abhinav Kumar
Cc: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart, Yongxing Mou
On Thu, Dec 05, 2024 at 08:31:31PM -0800, Abhinav Kumar wrote:
Please discuss with and reply to Dmitry's questions/feedback on the
list, so that it's possible for others to join the discussion.
Regards,
Bjorn
> Add support for Multi-stream transport for MSM chipsets that allow
> a single instance of DP controller to send multiple streams.
>
> This series has been validated on sa8775p ride platform using multiple
> MST dongles and also daisy chain method on both DP0 and DP1 upto 1080P.
>
> With 4x4K monitors, due to lack of layer mixers that combination will not
> work but this can be supported as well after some rework on the DPU side.
>
> In addition, SST was re-validated with all these changes to ensure there
> were no regressions.
>
> This patch series was made on top of:
>
> [1] : https://patchwork.freedesktop.org/patch/622243/ (to avoid a log spam)
> [2] : https://patchwork.freedesktop.org/series/142010/ (to fix up HPD)
> [3] : https://patchwork.freedesktop.org/patch/612740/ (to avoid blank screens)
> [4] : https://patchwork.freedesktop.org/series/140216/ (MDSS DT for sa8775p)
> [5] : https://patchwork.kernel.org/project/linux-arm-msm/list/?series=912200
> (Display Port DT changes for sa8775p)
>
> Bindings for the pixel clock for additional stream is available at :
>
> [6] : https://patchwork.freedesktop.org/series/142016/
>
> Overall, the patch series has been organized in the following way:
>
> 1) First set are a couple of fixes made while debugging MST but applicable
> to SST as well so go ahead of everything else
> 2) Prepare the DP driver to get ready to handle multiple streams. This is the bulk
> of the work as current DP driver design had to be adjusted to make this happen.
> 3) Finally, new files to handle MST related operations
>
> Validation was done on the latest linux-next on top of above changes and
> both FB console and weston compositors were validated with these changes.
>
> To: Rob Clark <robdclark@gmail.com>
> To: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> To: Sean Paul <sean@poorly.run>
> To: Marijn Suijten <marijn.suijten@somainline.org>
> To: David Airlie <airlied@gmail.com>
> To: Simona Vetter <simona@ffwll.ch>
> To: Stephen Boyd <swboyd@chromium.org>
> To: Chandan Uddaraju <chandanu@codeaurora.org>
> To: Guenter Roeck <groeck@chromium.org>
> To: Kuogee Hsieh <quic_khsieh@quicinc.com>
> To: Bjorn Andersson <andersson@kernel.org>
> To: Konrad Dybcio <konradybcio@kernel.org>
> To: Rob Herring <robh@kernel.org>
> To: Krzysztof Kozlowski <krzk+dt@kernel.org>
> To: Conor Dooley <conor+dt@kernel.org>
> Cc: Vara Reddy <quic_varar@quicinc.com>
> Cc: Rob Clark <robdclark@chromium.org>
> Cc: Tanmay Shah <tanmay@codeaurora.org>
> Cc: linux-arm-msm@vger.kernel.org
> Cc: dri-devel@lists.freedesktop.org
> Cc: freedreno@lists.freedesktop.org
> Cc: linux-kernel@vger.kernel.org
> Cc: devicetree@vger.kernel.org
> Cc: Jessica Zhang <quic_jesszhan@quicinc.com>
> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
>
> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
> ---
> Abhinav Kumar (43):
> drm/msm/dp: dont call dp_catalog_ctrl_mainlink_ctrl in dp_ctrl_configure_source_params()
> drm/msm/dp: disable the opp table request even for dp_ctrl_off_link()
> drm/msm/dp: fix the intf_type of MST interfaces
> drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts
> drm/msm/dp: add a helper to read mst caps for dp_panel
> drm/msm/dp: remove dp_display's dp_mode and use dp_panel's instead
> drm/msm/dp: break up dp_display_enable into two parts
> drm/msm/dp: re-arrange dp_display_disable() into functional parts
> drm/msm/dp: allow dp_ctrl stream APIs to use any panel passed to it
> drm/msm/dp: move the pixel clock control to its own API
> drm/msm/dp: split dp_ctrl_off() into stream and link parts
> drm/msm/dp: make bridge helpers use dp_display to allow re-use
> drm/msm/dp: separate dp_display_prepare() into its own API
> drm/msm/dp: introduce stream_id for each DP panel
> drm/msm/dp: convert dp_display_set_mode() to use dp_panel argument
> drm/msm/dp: add support for programming p1 register block
> drm/msm/dp: use stream_id to change offsets in dp_catalog
> drm/msm/dp: add support to send ACT packets for MST
> drm/msm/dp: add support to program mst support in mainlink
> drm/msm/dp: no need to update tu calculation for mst
> drm/msm/dp: add support for mst channel slot allocation
> drm/msm/dp: add support to send vcpf packets in dp controller
> drm/msm/dp: always program MST_FIFO_CONSTANT_FILL for MST
> drm/msm/dp: abstract out the dp_display stream helpers to accept a panel
> drm/msm/dp: move link related operations to dp_display_unprepare()
> drm/msm/dp: replace power_on with active_stream_cnt for dp_display
> drm/msm/dp: make the SST bridge disconnected when mst is active
> drm/msm/dp: add an API to initialize MST on sink side
> drm/msm/dp: skip reading the EDID for MST cases
> drm/msm/dp: add dp_display_get_panel() to initialize DP panel
> drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations
> drm/msm/dp: add connector abstraction for DP MST
> drm/msm/dp: add irq hpd callback for dp mst
> drm/msm/dp: add support to re-use and clear the panel edid
> drm/msm/dp: add a mst session mutex to protect bridge ops
> drm/msm: add support for non-blocking commits
> drm/msm: initialize DRM MST encoders for DP controllers
> drm/msm/dp: initialize dp_mst module for each DP MST controller
> drm/msm: add a stream to intf map for DP controller
> drm/msm/dpu: use msm_dp_get_mst_intf_id() to get the intf id
> drm/msm/dp: mark ST_DISCONNECTED only if all streams are disabled
> drm/msm/dp: populate the max_streams for sa8775 mst controller
> arm64: dts: qcom: add mst support for pixel 1 stream clk for DP1
>
> Yongxing Mou (2):
> drm/msm/dp: propagate hpd state changes to dp mst module
> arm64: dts: qcom: add mst support for pixel stream clk for DP0
>
> arch/arm64/boot/dts/qcom/sa8775p.dtsi | 23 +-
> drivers/gpu/drm/msm/Makefile | 3 +-
> .../drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h | 8 +-
> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 25 +-
> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h | 2 +
> drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 29 +-
> drivers/gpu/drm/msm/dp/dp_audio.c | 2 +-
> drivers/gpu/drm/msm/dp/dp_aux.h | 1 +
> drivers/gpu/drm/msm/dp/dp_catalog.c | 292 ++++-
> drivers/gpu/drm/msm/dp/dp_catalog.h | 27 +
> drivers/gpu/drm/msm/dp/dp_ctrl.c | 454 ++++++--
> drivers/gpu/drm/msm/dp/dp_ctrl.h | 21 +-
> drivers/gpu/drm/msm/dp/dp_display.c | 559 +++++++---
> drivers/gpu/drm/msm/dp/dp_display.h | 33 +-
> drivers/gpu/drm/msm/dp/dp_drm.c | 53 +-
> drivers/gpu/drm/msm/dp/dp_drm.h | 12 -
> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 1118 ++++++++++++++++++++
> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 110 ++
> drivers/gpu/drm/msm/dp/dp_panel.c | 41 +-
> drivers/gpu/drm/msm/dp/dp_panel.h | 15 +-
> drivers/gpu/drm/msm/dp/dp_reg.h | 25 +-
> drivers/gpu/drm/msm/msm_atomic.c | 2 +
> drivers/gpu/drm/msm/msm_drv.h | 29 +
> drivers/gpu/drm/msm/msm_kms.c | 1 +
> 24 files changed, 2589 insertions(+), 296 deletions(-)
> ---
> base-commit: b166256c1e6ce356fa1404d4c8531830e6f100a8
> change-id: 20241205-dp_mst-d62fa86257ed
>
> Best regards,
> --
> Abhinav Kumar <quic_abhinavk@quicinc.com>
>
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets
2025-01-07 1:06 ` [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Bjorn Andersson
@ 2025-01-07 1:10 ` Abhinav Kumar
0 siblings, 0 replies; 111+ messages in thread
From: Abhinav Kumar @ 2025-01-07 1:10 UTC (permalink / raw)
To: Bjorn Andersson
Cc: Rob Clark, Dmitry Baryshkov, Sean Paul, Marijn Suijten,
David Airlie, Simona Vetter, Stephen Boyd, Chandan Uddaraju,
Guenter Roeck, Kuogee Hsieh, Konrad Dybcio, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Vara Reddy, Rob Clark,
Tanmay Shah, linux-arm-msm, dri-devel, freedreno, linux-kernel,
devicetree, Jessica Zhang, Laurent Pinchart, Yongxing Mou
Hi Bjorn / Dmitry
Happy New Year !
On 1/6/2025 5:06 PM, Bjorn Andersson wrote:
> On Thu, Dec 05, 2024 at 08:31:31PM -0800, Abhinav Kumar wrote:
>
> Please discuss with and reply to Dmitry's questions/feedback on the
> list, so that it's possible for others to join the discussion.
>
> Regards,
> Bjorn
Sorry for the delay. I got back from vacation only today, I will respond
to this feature by next week starting with the bindings as it got
delayed due to my vacation.
Thanks
Abhinav
>
>> Add support for Multi-stream transport for MSM chipsets that allow
>> a single instance of DP controller to send multiple streams.
>>
>> This series has been validated on sa8775p ride platform using multiple
>> MST dongles and also daisy chain method on both DP0 and DP1 upto 1080P.
>>
>> With 4x4K monitors, due to lack of layer mixers that combination will not
>> work but this can be supported as well after some rework on the DPU side.
>>
>> In addition, SST was re-validated with all these changes to ensure there
>> were no regressions.
>>
>> This patch series was made on top of:
>>
>> [1] : https://patchwork.freedesktop.org/patch/622243/ (to avoid a log spam)
>> [2] : https://patchwork.freedesktop.org/series/142010/ (to fix up HPD)
>> [3] : https://patchwork.freedesktop.org/patch/612740/ (to avoid blank screens)
>> [4] : https://patchwork.freedesktop.org/series/140216/ (MDSS DT for sa8775p)
>> [5] : https://patchwork.kernel.org/project/linux-arm-msm/list/?series=912200
>> (Display Port DT changes for sa8775p)
>>
>> Bindings for the pixel clock for additional stream is available at :
>>
>> [6] : https://patchwork.freedesktop.org/series/142016/
>>
>> Overall, the patch series has been organized in the following way:
>>
>> 1) First set are a couple of fixes made while debugging MST but applicable
>> to SST as well so go ahead of everything else
>> 2) Prepare the DP driver to get ready to handle multiple streams. This is the bulk
>> of the work as current DP driver design had to be adjusted to make this happen.
>> 3) Finally, new files to handle MST related operations
>>
>> Validation was done on the latest linux-next on top of above changes and
>> both FB console and weston compositors were validated with these changes.
>>
>> To: Rob Clark <robdclark@gmail.com>
>> To: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
>> To: Sean Paul <sean@poorly.run>
>> To: Marijn Suijten <marijn.suijten@somainline.org>
>> To: David Airlie <airlied@gmail.com>
>> To: Simona Vetter <simona@ffwll.ch>
>> To: Stephen Boyd <swboyd@chromium.org>
>> To: Chandan Uddaraju <chandanu@codeaurora.org>
>> To: Guenter Roeck <groeck@chromium.org>
>> To: Kuogee Hsieh <quic_khsieh@quicinc.com>
>> To: Bjorn Andersson <andersson@kernel.org>
>> To: Konrad Dybcio <konradybcio@kernel.org>
>> To: Rob Herring <robh@kernel.org>
>> To: Krzysztof Kozlowski <krzk+dt@kernel.org>
>> To: Conor Dooley <conor+dt@kernel.org>
>> Cc: Vara Reddy <quic_varar@quicinc.com>
>> Cc: Rob Clark <robdclark@chromium.org>
>> Cc: Tanmay Shah <tanmay@codeaurora.org>
>> Cc: linux-arm-msm@vger.kernel.org
>> Cc: dri-devel@lists.freedesktop.org
>> Cc: freedreno@lists.freedesktop.org
>> Cc: linux-kernel@vger.kernel.org
>> Cc: devicetree@vger.kernel.org
>> Cc: Jessica Zhang <quic_jesszhan@quicinc.com>
>> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
>>
>> Signed-off-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
>> ---
>> Abhinav Kumar (43):
>> drm/msm/dp: dont call dp_catalog_ctrl_mainlink_ctrl in dp_ctrl_configure_source_params()
>> drm/msm/dp: disable the opp table request even for dp_ctrl_off_link()
>> drm/msm/dp: fix the intf_type of MST interfaces
>> drm/msm/dp: split msm_dp_panel_read_sink_caps() into two parts
>> drm/msm/dp: add a helper to read mst caps for dp_panel
>> drm/msm/dp: remove dp_display's dp_mode and use dp_panel's instead
>> drm/msm/dp: break up dp_display_enable into two parts
>> drm/msm/dp: re-arrange dp_display_disable() into functional parts
>> drm/msm/dp: allow dp_ctrl stream APIs to use any panel passed to it
>> drm/msm/dp: move the pixel clock control to its own API
>> drm/msm/dp: split dp_ctrl_off() into stream and link parts
>> drm/msm/dp: make bridge helpers use dp_display to allow re-use
>> drm/msm/dp: separate dp_display_prepare() into its own API
>> drm/msm/dp: introduce stream_id for each DP panel
>> drm/msm/dp: convert dp_display_set_mode() to use dp_panel argument
>> drm/msm/dp: add support for programming p1 register block
>> drm/msm/dp: use stream_id to change offsets in dp_catalog
>> drm/msm/dp: add support to send ACT packets for MST
>> drm/msm/dp: add support to program mst support in mainlink
>> drm/msm/dp: no need to update tu calculation for mst
>> drm/msm/dp: add support for mst channel slot allocation
>> drm/msm/dp: add support to send vcpf packets in dp controller
>> drm/msm/dp: always program MST_FIFO_CONSTANT_FILL for MST
>> drm/msm/dp: abstract out the dp_display stream helpers to accept a panel
>> drm/msm/dp: move link related operations to dp_display_unprepare()
>> drm/msm/dp: replace power_on with active_stream_cnt for dp_display
>> drm/msm/dp: make the SST bridge disconnected when mst is active
>> drm/msm/dp: add an API to initialize MST on sink side
>> drm/msm/dp: skip reading the EDID for MST cases
>> drm/msm/dp: add dp_display_get_panel() to initialize DP panel
>> drm/msm/dp: add dp_mst_drm to manage DP MST bridge operations
>> drm/msm/dp: add connector abstraction for DP MST
>> drm/msm/dp: add irq hpd callback for dp mst
>> drm/msm/dp: add support to re-use and clear the panel edid
>> drm/msm/dp: add a mst session mutex to protect bridge ops
>> drm/msm: add support for non-blocking commits
>> drm/msm: initialize DRM MST encoders for DP controllers
>> drm/msm/dp: initialize dp_mst module for each DP MST controller
>> drm/msm: add a stream to intf map for DP controller
>> drm/msm/dpu: use msm_dp_get_mst_intf_id() to get the intf id
>> drm/msm/dp: mark ST_DISCONNECTED only if all streams are disabled
>> drm/msm/dp: populate the max_streams for sa8775 mst controller
>> arm64: dts: qcom: add mst support for pixel 1 stream clk for DP1
>>
>> Yongxing Mou (2):
>> drm/msm/dp: propagate hpd state changes to dp mst module
>> arm64: dts: qcom: add mst support for pixel stream clk for DP0
>>
>> arch/arm64/boot/dts/qcom/sa8775p.dtsi | 23 +-
>> drivers/gpu/drm/msm/Makefile | 3 +-
>> .../drm/msm/disp/dpu1/catalog/dpu_8_4_sa8775p.h | 8 +-
>> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 25 +-
>> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h | 2 +
>> drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 29 +-
>> drivers/gpu/drm/msm/dp/dp_audio.c | 2 +-
>> drivers/gpu/drm/msm/dp/dp_aux.h | 1 +
>> drivers/gpu/drm/msm/dp/dp_catalog.c | 292 ++++-
>> drivers/gpu/drm/msm/dp/dp_catalog.h | 27 +
>> drivers/gpu/drm/msm/dp/dp_ctrl.c | 454 ++++++--
>> drivers/gpu/drm/msm/dp/dp_ctrl.h | 21 +-
>> drivers/gpu/drm/msm/dp/dp_display.c | 559 +++++++---
>> drivers/gpu/drm/msm/dp/dp_display.h | 33 +-
>> drivers/gpu/drm/msm/dp/dp_drm.c | 53 +-
>> drivers/gpu/drm/msm/dp/dp_drm.h | 12 -
>> drivers/gpu/drm/msm/dp/dp_mst_drm.c | 1118 ++++++++++++++++++++
>> drivers/gpu/drm/msm/dp/dp_mst_drm.h | 110 ++
>> drivers/gpu/drm/msm/dp/dp_panel.c | 41 +-
>> drivers/gpu/drm/msm/dp/dp_panel.h | 15 +-
>> drivers/gpu/drm/msm/dp/dp_reg.h | 25 +-
>> drivers/gpu/drm/msm/msm_atomic.c | 2 +
>> drivers/gpu/drm/msm/msm_drv.h | 29 +
>> drivers/gpu/drm/msm/msm_kms.c | 1 +
>> 24 files changed, 2589 insertions(+), 296 deletions(-)
>> ---
>> base-commit: b166256c1e6ce356fa1404d4c8531830e6f100a8
>> change-id: 20241205-dp_mst-d62fa86257ed
>>
>> Best regards,
>> --
>> Abhinav Kumar <quic_abhinavk@quicinc.com>
>>
^ permalink raw reply [flat|nested] 111+ messages in thread
* Re: [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets
2024-12-06 4:31 [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Abhinav Kumar
` (45 preceding siblings ...)
2025-01-07 1:06 ` [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets Bjorn Andersson
@ 2025-05-10 12:12 ` Jens Glathe
2025-05-22 18:19 ` Abhinav Kumar
46 siblings, 1 reply; 111+ messages in thread
From: Jens Glathe @ 2025-05-10 12:12 UTC (permalink / raw)
To: Abhinav Kumar, Rob Clark, Dmitry Baryshkov, Sean Paul,
Marijn Suijten, David Airlie, Simona Vetter, Stephen Boyd,
Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh, Bjorn Andersson,
Konrad Dybcio, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Yongxing Mou
On 06.12.24 05:31, Abhinav Kumar wrote:
> base-commit: b166256c1e6ce356fa1404d4c8531830e6f100a8
Hi Abhinav,
I would like to test / play around with this patchset, unfortunately
this base commit is not easy to find. Trying to apply without gives lots
of conflicts. Can you please rebase?
with best regards
Jens
^ permalink raw reply [flat|nested] 111+ messages in thread* Re: [PATCH 00/45] drm/msm/dp: Add MST support for MSM chipsets
2025-05-10 12:12 ` Jens Glathe
@ 2025-05-22 18:19 ` Abhinav Kumar
0 siblings, 0 replies; 111+ messages in thread
From: Abhinav Kumar @ 2025-05-22 18:19 UTC (permalink / raw)
To: Jens Glathe, Rob Clark, Dmitry Baryshkov, Sean Paul,
Marijn Suijten, David Airlie, Simona Vetter, Stephen Boyd,
Chandan Uddaraju, Guenter Roeck, Kuogee Hsieh, Bjorn Andersson,
Konrad Dybcio, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: Vara Reddy, Rob Clark, Tanmay Shah, linux-arm-msm, dri-devel,
freedreno, linux-kernel, devicetree, Jessica Zhang,
Laurent Pinchart, Yongxing Mou
Hi Jens
On 5/10/2025 5:12 AM, Jens Glathe wrote:
> On 06.12.24 05:31, Abhinav Kumar wrote:
>> base-commit: b166256c1e6ce356fa1404d4c8531830e6f100a8
>
> Hi Abhinav,
>
> I would like to test / play around with this patchset, unfortunately
> this base commit is not easy to find. Trying to apply without gives lots
> of conflicts. Can you please rebase?
>
> with best regards
>
> Jens
>
>
We will post a rebased version of this series with the review comments
addressed within the next 3-4 weeks.
Thanks
Abhinav
^ permalink raw reply [flat|nested] 111+ messages in thread