[PATCH v2] media: venus: pm_helpers: use opp-table for the frequency

[PATCH v2] media: venus: pm_helpers: use opp-table for the frequency

[Renjiang Han]

Some platforms (such as qcs615 and sc7180) use the same core but have
different frequency tables. Using the opp-table allows us to separate
the core description from the frequency data and supports the use of
fallback compatibles.

Reviewed-by: Vikash Garodia <quic_vgarodia@quicinc.com>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Renjiang Han <quic_renjiang@quicinc.com>
---
video core in QCS615 is same as that in SC7180, while the clock frequency
corners are different. In such cases, the clock frequencies are better
represented in DT than static table in video driver. The patch make use
of OPP table in DT to choose the frequency corner to scale.

Note:
Earlier discussion was concluded in [1] where it was agreed to rely on
opp-table for QCS615.

[1] https://lore.kernel.org/linux-arm-msm/c9b83c8b-68d1-43bc-99d6-d2d2b9e445f4@oss.qualcomm.com/
---
Changes in v2:
- 1. Update message based on the review comment.
- Link to v1: https://lore.kernel.org/r/20250610-use_freq_with_opp_table-v1-1-7339d690fd1a@quicinc.com
---
 drivers/media/platform/qcom/venus/pm_helpers.c | 58 +++++++++++++++-----------
 1 file changed, 34 insertions(+), 24 deletions(-)

diff --git a/drivers/media/platform/qcom/venus/pm_helpers.c b/drivers/media/platform/qcom/venus/pm_helpers.c
index 409aa9bd0b5d099c993eedb03177ec5ed918b4a0..8dd5a9b0d060cddfeafd4da477ade0c7aeb6c390 100644
--- a/drivers/media/platform/qcom/venus/pm_helpers.c
+++ b/drivers/media/platform/qcom/venus/pm_helpers.c
@@ -41,16 +41,14 @@ static int core_clks_get(struct venus_core *core)
 static int core_clks_enable(struct venus_core *core)
 {
 	const struct venus_resources *res = core->res;
-	const struct freq_tbl *freq_tbl = core->res->freq_tbl;
-	unsigned int freq_tbl_size = core->res->freq_tbl_size;
-	unsigned long freq;
+	struct device *dev = core->dev;
+	unsigned long freq = 0;
+	struct dev_pm_opp *opp;
 	unsigned int i;
 	int ret;
 
-	if (!freq_tbl)
-		return -EINVAL;
-
-	freq = freq_tbl[freq_tbl_size - 1].freq;
+	opp = dev_pm_opp_find_freq_ceil(dev, &freq);
+	dev_pm_opp_put(opp);
 
 	for (i = 0; i < res->clks_num; i++) {
 		if (IS_V6(core)) {
@@ -636,7 +634,9 @@ static int decide_core(struct venus_inst *inst)
 	u32 min_coreid, min_load, cur_inst_load;
 	u32 min_lp_coreid, min_lp_load, cur_inst_lp_load;
 	struct hfi_videocores_usage_type cu;
-	unsigned long max_freq;
+	unsigned long max_freq = ULONG_MAX;
+	struct device *dev = core->dev;
+	struct dev_pm_opp *opp;
 	int ret = 0;
 
 	if (legacy_binding) {
@@ -659,7 +659,8 @@ static int decide_core(struct venus_inst *inst)
 	cur_inst_lp_load *= inst->clk_data.low_power_freq;
 	/*TODO : divide this inst->load by work_route */
 
-	max_freq = core->res->freq_tbl[0].freq;
+	opp = dev_pm_opp_find_freq_floor(dev, &max_freq);
+	dev_pm_opp_put(opp);
 
 	min_loaded_core(inst, &min_coreid, &min_load, false);
 	min_loaded_core(inst, &min_lp_coreid, &min_lp_load, true);
@@ -949,7 +950,10 @@ static int core_resets_get(struct venus_core *core)
 static int core_get_v4(struct venus_core *core)
 {
 	struct device *dev = core->dev;
+	const struct freq_tbl *freq_tbl = core->res->freq_tbl;
+	unsigned int num_rows = core->res->freq_tbl_size;
 	const struct venus_resources *res = core->res;
+	unsigned int i;
 	int ret;
 
 	ret = core_clks_get(core);
@@ -986,9 +990,17 @@ static int core_get_v4(struct venus_core *core)
 
 	if (core->res->opp_pmdomain) {
 		ret = devm_pm_opp_of_add_table(dev);
-		if (ret && ret != -ENODEV) {
-			dev_err(dev, "invalid OPP table in device tree\n");
-			return ret;
+		if (ret) {
+			if (ret == -ENODEV) {
+				for (i = 0; i < num_rows; i++) {
+					ret = dev_pm_opp_add(dev, freq_tbl[i].freq, 0);
+					if (ret)
+						return ret;
+				}
+			} else {
+				dev_err(dev, "invalid OPP table in device tree\n");
+				return ret;
+			}
 		}
 	}
 
@@ -1078,11 +1090,11 @@ static unsigned long calculate_inst_freq(struct venus_inst *inst,
 static int load_scale_v4(struct venus_inst *inst)
 {
 	struct venus_core *core = inst->core;
-	const struct freq_tbl *table = core->res->freq_tbl;
-	unsigned int num_rows = core->res->freq_tbl_size;
 	struct device *dev = core->dev;
 	unsigned long freq = 0, freq_core1 = 0, freq_core2 = 0;
+	unsigned long max_freq = ULONG_MAX;
 	unsigned long filled_len = 0;
+	struct dev_pm_opp *opp;
 	int i, ret = 0;
 
 	for (i = 0; i < inst->num_input_bufs; i++)
@@ -1108,20 +1120,18 @@ static int load_scale_v4(struct venus_inst *inst)
 
 	freq = max(freq_core1, freq_core2);
 
-	if (freq > table[0].freq) {
-		dev_dbg(dev, VDBGL "requested clock rate: %lu scaling clock rate : %lu\n",
-			freq, table[0].freq);
+	opp = dev_pm_opp_find_freq_floor(dev, &max_freq);
+	dev_pm_opp_put(opp);
 
-		freq = table[0].freq;
+	if (freq > max_freq) {
+		dev_dbg(dev, VDBGL "requested clock rate: %lu scaling clock rate : %lu\n",
+			freq, max_freq);
+		freq = max_freq;
 		goto set_freq;
 	}
 
-	for (i = num_rows - 1 ; i >= 0; i--) {
-		if (freq <= table[i].freq) {
-			freq = table[i].freq;
-			break;
-		}
-	}
+	opp = dev_pm_opp_find_freq_ceil(dev, &freq);
+	dev_pm_opp_put(opp);
 
 set_freq:
 

---
base-commit: b27cc623e01be9de1580eaa913508b237a7a9673
change-id: 20250610-use_freq_with_opp_table-b81162cfecba

Best regards,
-- 
Renjiang Han <quic_renjiang@quicinc.com>

Re: [PATCH v2] media: venus: pm_helpers: use opp-table for the frequency

[Renjiang Han]

On 6/12/2025 10:23 AM, Renjiang Han wrote:
> Some platforms (such as qcs615 and sc7180) use the same core but have
> different frequency tables. Using the opp-table allows us to separate
> the core description from the frequency data and supports the use of
> fallback compatibles.
>
> Reviewed-by: Vikash Garodia<quic_vgarodia@quicinc.com>
> Reviewed-by: Bryan O'Donoghue<bryan.odonoghue@linaro.org>
> Signed-off-by: Renjiang Han<quic_renjiang@quicinc.com>
> ---
> video core in QCS615 is same as that in SC7180, while the clock frequency
> corners are different. In such cases, the clock frequencies are better
> represented in DT than static table in video driver. The patch make use
> of OPP table in DT to choose the frequency corner to scale.
>
> Note:
> Earlier discussion was concluded in [1] where it was agreed to rely on
> opp-table for QCS615.
>
> [1]https://lore.kernel.org/linux-arm-msm/c9b83c8b-68d1-43bc-99d6-d2d2b9e445f4@oss.qualcomm.com/
> ---
> Changes in v2:
> - 1. Update message based on the review comment.
> - Link to v1:https://lore.kernel.org/r/20250610-use_freq_with_opp_table-v1-1-7339d690fd1a@quicinc.com
> ---
>   drivers/media/platform/qcom/venus/pm_helpers.c | 58 +++++++++++++++-----------
>   1 file changed, 34 insertions(+), 24 deletions(-)
I’d appreciate any thoughts you might have on this.

-- 
Best Regards,
Renjiang

Re: [PATCH v2] media: venus: pm_helpers: use opp-table for the frequency

[Renjiang Han]

On 7/4/2025 3:21 PM, Renjiang Han wrote:
>
> On 6/12/2025 10:23 AM, Renjiang Han wrote:
>> Some platforms (such as qcs615 and sc7180) use the same core but have
>> different frequency tables. Using the opp-table allows us to separate
>> the core description from the frequency data and supports the use of
>> fallback compatibles.
>>
>> Reviewed-by: Vikash Garodia<quic_vgarodia@quicinc.com>
>> Reviewed-by: Bryan O'Donoghue<bryan.odonoghue@linaro.org>
>> Signed-off-by: Renjiang Han<quic_renjiang@quicinc.com>
>> ---
>> video core in QCS615 is same as that in SC7180, while the clock 
>> frequency
>> corners are different. In such cases, the clock frequencies are better
>> represented in DT than static table in video driver. The patch make use
>> of OPP table in DT to choose the frequency corner to scale.
>>
>> Note:
>> Earlier discussion was concluded in [1] where it was agreed to rely on
>> opp-table for QCS615.
>>
>> [1]https://lore.kernel.org/linux-arm-msm/c9b83c8b-68d1-43bc-99d6-d2d2b9e445f4@oss.qualcomm.com/ 
>>
>> ---
>> Changes in v2:
>> - 1. Update message based on the review comment.
>> - Link to 
>> v1:https://lore.kernel.org/r/20250610-use_freq_with_opp_table-v1-1-7339d690fd1a@quicinc.com
>> ---
>>   drivers/media/platform/qcom/venus/pm_helpers.c | 58 
>> +++++++++++++++-----------
>>   1 file changed, 34 insertions(+), 24 deletions(-)
> I’d appreciate any thoughts you might have on this.
>
This patch has already been picked up, please ignore last email.

-- 
Best Regards,
Renjiang