* [PATCH 1/2] thermal/drivers/qcom/lmh: Use driver data as flags instead of bool [not found] <20230113031401.2336157-1-konrad.dybcio@linaro.org> @ 2023-01-13 3:14 ` Konrad Dybcio 2023-01-19 3:05 ` Bjorn Andersson 2023-01-13 3:14 ` [PATCH 2/2] firmware: qcom_scm: Fully implement qcom_scm_lmh_dcvsh() Konrad Dybcio 1 sibling, 1 reply; 5+ messages in thread From: Konrad Dybcio @ 2023-01-13 3:14 UTC (permalink / raw) To: linux-arm-msm, andersson, agross, krzysztof.kozlowski Cc: marijn.suijten, Konrad Dybcio, Amit Kucheria, Thara Gopinath, Rafael J. Wysocki, Daniel Lezcano, Zhang Rui, linux-pm, linux-kernel Currently the OF driver data is used to determine whether to manually enable the LMH algorithms through secure calls. In preparation for introducing support for SoCs with a configuration that differs from both SDM845 and SM8150, convert this to a set of bitflags for more flexibility. Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> --- drivers/thermal/qcom/lmh.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/thermal/qcom/lmh.c b/drivers/thermal/qcom/lmh.c index 4122a51e9874..5e8ff196c9a6 100644 --- a/drivers/thermal/qcom/lmh.c +++ b/drivers/thermal/qcom/lmh.c @@ -28,7 +28,7 @@ #define LMH_REG_DCVS_INTR_CLR 0x8 -#define LMH_ENABLE_ALGOS 1 +#define LMH_ENABLE_ALGOS BIT(0) struct lmh_hw_data { void __iomem *base; @@ -92,9 +92,11 @@ static int lmh_probe(struct platform_device *pdev) struct device_node *cpu_node; struct lmh_hw_data *lmh_data; int temp_low, temp_high, temp_arm, cpu_id, ret; - unsigned int enable_alg; + unsigned int flags; u32 node_id; + flags = (uintptr_t)of_device_get_match_data(dev); + lmh_data = devm_kzalloc(dev, sizeof(*lmh_data), GFP_KERNEL); if (!lmh_data) return -ENOMEM; @@ -144,9 +146,7 @@ static int lmh_probe(struct platform_device *pdev) if (!qcom_scm_lmh_dcvsh_available()) return -EINVAL; - enable_alg = (uintptr_t)of_device_get_match_data(dev); - - if (enable_alg) { + if (flags & LMH_ENABLE_ALGOS) { ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, LMH_NODE_DCVS, node_id, 0); if (ret) -- 2.39.0 ^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH 1/2] thermal/drivers/qcom/lmh: Use driver data as flags instead of bool 2023-01-13 3:14 ` [PATCH 1/2] thermal/drivers/qcom/lmh: Use driver data as flags instead of bool Konrad Dybcio @ 2023-01-19 3:05 ` Bjorn Andersson 0 siblings, 0 replies; 5+ messages in thread From: Bjorn Andersson @ 2023-01-19 3:05 UTC (permalink / raw) To: Konrad Dybcio Cc: linux-arm-msm, agross, krzysztof.kozlowski, marijn.suijten, Amit Kucheria, Thara Gopinath, Rafael J. Wysocki, Daniel Lezcano, Zhang Rui, linux-pm, linux-kernel On Fri, Jan 13, 2023 at 04:14:00AM +0100, Konrad Dybcio wrote: > Currently the OF driver data is used to determine whether to manually > enable the LMH algorithms through secure calls. In preparation for > introducing support for SoCs with a configuration that differs from > both SDM845 and SM8150, convert this to a set of bitflags for more > flexibility. > > Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> > --- > drivers/thermal/qcom/lmh.c | 10 +++++----- > 1 file changed, 5 insertions(+), 5 deletions(-) > > diff --git a/drivers/thermal/qcom/lmh.c b/drivers/thermal/qcom/lmh.c > index 4122a51e9874..5e8ff196c9a6 100644 > --- a/drivers/thermal/qcom/lmh.c > +++ b/drivers/thermal/qcom/lmh.c > @@ -28,7 +28,7 @@ > > #define LMH_REG_DCVS_INTR_CLR 0x8 > > -#define LMH_ENABLE_ALGOS 1 > +#define LMH_ENABLE_ALGOS BIT(0) Looks perfectly fine, but please include a patch in the same series that adds another bit. And don't forget to update the .data in the of_match_id listing. Regards, Bjorn > > struct lmh_hw_data { > void __iomem *base; > @@ -92,9 +92,11 @@ static int lmh_probe(struct platform_device *pdev) > struct device_node *cpu_node; > struct lmh_hw_data *lmh_data; > int temp_low, temp_high, temp_arm, cpu_id, ret; > - unsigned int enable_alg; > + unsigned int flags; > u32 node_id; > > + flags = (uintptr_t)of_device_get_match_data(dev); > + > lmh_data = devm_kzalloc(dev, sizeof(*lmh_data), GFP_KERNEL); > if (!lmh_data) > return -ENOMEM; > @@ -144,9 +146,7 @@ static int lmh_probe(struct platform_device *pdev) > if (!qcom_scm_lmh_dcvsh_available()) > return -EINVAL; > > - enable_alg = (uintptr_t)of_device_get_match_data(dev); > - > - if (enable_alg) { > + if (flags & LMH_ENABLE_ALGOS) { > ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, > LMH_NODE_DCVS, node_id, 0); > if (ret) > -- > 2.39.0 > ^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 2/2] firmware: qcom_scm: Fully implement qcom_scm_lmh_dcvsh() [not found] <20230113031401.2336157-1-konrad.dybcio@linaro.org> 2023-01-13 3:14 ` [PATCH 1/2] thermal/drivers/qcom/lmh: Use driver data as flags instead of bool Konrad Dybcio @ 2023-01-13 3:14 ` Konrad Dybcio 2023-01-19 3:04 ` Bjorn Andersson 1 sibling, 1 reply; 5+ messages in thread From: Konrad Dybcio @ 2023-01-13 3:14 UTC (permalink / raw) To: linux-arm-msm, andersson, agross, krzysztof.kozlowski Cc: marijn.suijten, Konrad Dybcio, Amit Kucheria, Thara Gopinath, Rafael J. Wysocki, Daniel Lezcano, Zhang Rui, AngeloGioacchino Del Regno, Loic Poulain, Stephan Gerhold, linux-kernel, linux-pm The qcom_scm_lmh_dcvsh call can actually pass two values to the secure world. The second value is used for example with the LMH_FREQ_CAP function, which limits the maximum achievable frequency directly from LMh. Add the missing arguments, handle them and update the current usages of this function. Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> --- drivers/firmware/qcom_scm.c | 13 ++++++++----- drivers/thermal/qcom/lmh.c | 28 ++++++++++++++-------------- include/linux/qcom_scm.h | 5 +++-- 3 files changed, 25 insertions(+), 21 deletions(-) diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c index cdbfe54c8146..58a19a47e442 100644 --- a/drivers/firmware/qcom_scm.c +++ b/drivers/firmware/qcom_scm.c @@ -1252,12 +1252,13 @@ int qcom_scm_lmh_profile_change(u32 profile_id) } EXPORT_SYMBOL(qcom_scm_lmh_profile_change); -int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, - u64 limit_node, u32 node_id, u64 version) +int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val0, + u32 payload_val1, u64 limit_node, u32 node_id, + u64 version, bool has_val1) { dma_addr_t payload_phys; u32 *payload_buf; - int ret, payload_size = 5 * sizeof(u32); + int ret, payload_size = (5 + has_val1) * sizeof(u32); struct qcom_scm_desc desc = { .svc = QCOM_SCM_SVC_LMH, @@ -1278,8 +1279,10 @@ int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, payload_buf[0] = payload_fn; payload_buf[1] = 0; payload_buf[2] = payload_reg; - payload_buf[3] = 1; - payload_buf[4] = payload_val; + payload_buf[3] = has_val1 ? 2 : 1; + payload_buf[4] = payload_val0; + if (has_val1) + payload_buf[5] = payload_val1; desc.args[0] = payload_phys; diff --git a/drivers/thermal/qcom/lmh.c b/drivers/thermal/qcom/lmh.c index 5e8ff196c9a6..d2b5ea8322eb 100644 --- a/drivers/thermal/qcom/lmh.c +++ b/drivers/thermal/qcom/lmh.c @@ -147,23 +147,23 @@ static int lmh_probe(struct platform_device *pdev) return -EINVAL; if (flags & LMH_ENABLE_ALGOS) { - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, - LMH_NODE_DCVS, node_id, 0); + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, 0, + LMH_NODE_DCVS, node_id, 0, false); if (ret) dev_err(dev, "Error %d enabling current subfunction\n", ret); - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_REL, LMH_ALGO_MODE_ENABLE, 1, - LMH_NODE_DCVS, node_id, 0); + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_REL, LMH_ALGO_MODE_ENABLE, 1, 0, + LMH_NODE_DCVS, node_id, 0, false); if (ret) dev_err(dev, "Error %d enabling reliability subfunction\n", ret); - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_BCL, LMH_ALGO_MODE_ENABLE, 1, - LMH_NODE_DCVS, node_id, 0); + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_BCL, LMH_ALGO_MODE_ENABLE, 1, 0, + LMH_NODE_DCVS, node_id, 0, false); if (ret) dev_err(dev, "Error %d enabling BCL subfunction\n", ret); - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_ALGO_MODE_ENABLE, 1, - LMH_NODE_DCVS, node_id, 0); + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_ALGO_MODE_ENABLE, 1, 0, + LMH_NODE_DCVS, node_id, 0, false); if (ret) { dev_err(dev, "Error %d enabling thermal subfunction\n", ret); return ret; @@ -177,22 +177,22 @@ static int lmh_probe(struct platform_device *pdev) } /* Set default thermal trips */ - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_ARM_THRESHOLD, temp_arm, - LMH_NODE_DCVS, node_id, 0); + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_ARM_THRESHOLD, temp_arm, 0, + LMH_NODE_DCVS, node_id, 0, false); if (ret) { dev_err(dev, "Error setting thermal ARM threshold%d\n", ret); return ret; } - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_HI_THRESHOLD, temp_high, - LMH_NODE_DCVS, node_id, 0); + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_HI_THRESHOLD, temp_high, 0, + LMH_NODE_DCVS, node_id, 0, false); if (ret) { dev_err(dev, "Error setting thermal HI threshold%d\n", ret); return ret; } - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_LOW_THRESHOLD, temp_low, - LMH_NODE_DCVS, node_id, 0); + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_LOW_THRESHOLD, temp_low, 0, + LMH_NODE_DCVS, node_id, 0, false); if (ret) { dev_err(dev, "Error setting thermal ARM threshold%d\n", ret); return ret; diff --git a/include/linux/qcom_scm.h b/include/linux/qcom_scm.h index 1e449a5d7f5c..9fd798d17fdd 100644 --- a/include/linux/qcom_scm.h +++ b/include/linux/qcom_scm.h @@ -117,8 +117,9 @@ extern int qcom_scm_hdcp_req(struct qcom_scm_hdcp_req *req, u32 req_cnt, extern int qcom_scm_iommu_set_pt_format(u32 sec_id, u32 ctx_num, u32 pt_fmt); extern int qcom_scm_qsmmu500_wait_safe_toggle(bool en); -extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, - u64 limit_node, u32 node_id, u64 version); +extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val0, + u32 payload_val1, u64 limit_node, u32 node_id, + u64 version, bool has_val1); extern int qcom_scm_lmh_profile_change(u32 profile_id); extern bool qcom_scm_lmh_dcvsh_available(void); -- 2.39.0 ^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] firmware: qcom_scm: Fully implement qcom_scm_lmh_dcvsh() 2023-01-13 3:14 ` [PATCH 2/2] firmware: qcom_scm: Fully implement qcom_scm_lmh_dcvsh() Konrad Dybcio @ 2023-01-19 3:04 ` Bjorn Andersson 2023-01-19 10:41 ` Konrad Dybcio 0 siblings, 1 reply; 5+ messages in thread From: Bjorn Andersson @ 2023-01-19 3:04 UTC (permalink / raw) To: Konrad Dybcio Cc: linux-arm-msm, agross, krzysztof.kozlowski, marijn.suijten, Amit Kucheria, Thara Gopinath, Rafael J. Wysocki, Daniel Lezcano, Zhang Rui, AngeloGioacchino Del Regno, Loic Poulain, Stephan Gerhold, linux-kernel, linux-pm On Fri, Jan 13, 2023 at 04:14:01AM +0100, Konrad Dybcio wrote: > The qcom_scm_lmh_dcvsh call can actually pass two values to the > secure world. The second value is used for example with the > LMH_FREQ_CAP function, which limits the maximum achievable frequency > directly from LMh. Add the missing arguments, handle them and update > the current usages of this function. > > Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> > --- > drivers/firmware/qcom_scm.c | 13 ++++++++----- > drivers/thermal/qcom/lmh.c | 28 ++++++++++++++-------------- > include/linux/qcom_scm.h | 5 +++-- > 3 files changed, 25 insertions(+), 21 deletions(-) > > diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c > index cdbfe54c8146..58a19a47e442 100644 > --- a/drivers/firmware/qcom_scm.c > +++ b/drivers/firmware/qcom_scm.c > @@ -1252,12 +1252,13 @@ int qcom_scm_lmh_profile_change(u32 profile_id) > } > EXPORT_SYMBOL(qcom_scm_lmh_profile_change); > > -int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, > - u64 limit_node, u32 node_id, u64 version) > +int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val0, > + u32 payload_val1, u64 limit_node, u32 node_id, > + u64 version, bool has_val1) Rather than always passing a dummy payload_val1 and then having has_val1 to indicate if it should be considered or not... how about moving the payload last in the call, as a va_list with a "count" before that? I.e: int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u64 limit_node, u32 node_id, u64 version, unsigned int payload_count, ...) > { > dma_addr_t payload_phys; > u32 *payload_buf; > - int ret, payload_size = 5 * sizeof(u32); > + int ret, payload_size = (5 + has_val1) * sizeof(u32); allocate 4 + payload_count > > struct qcom_scm_desc desc = { > .svc = QCOM_SCM_SVC_LMH, > @@ -1278,8 +1279,10 @@ int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, > payload_buf[0] = payload_fn; > payload_buf[1] = 0; > payload_buf[2] = payload_reg; > - payload_buf[3] = 1; > - payload_buf[4] = payload_val; > + payload_buf[3] = has_val1 ? 2 : 1; > + payload_buf[4] = payload_val0; > + if (has_val1) > + payload_buf[5] = payload_val1; Something like: payload_buf[3] = payload_count; va_start(ap, payload_count); for (i = 0; i < payload_count; i++) payload_buf[4 + i] = va_arg(ap, uint32_t); va_end(ap); That said, I don't see a single "true" below. Perhaps I'm missing it? I would expect some code in the same series use the newly introduced logic. Thanks, Bjorn > > desc.args[0] = payload_phys; > > diff --git a/drivers/thermal/qcom/lmh.c b/drivers/thermal/qcom/lmh.c > index 5e8ff196c9a6..d2b5ea8322eb 100644 > --- a/drivers/thermal/qcom/lmh.c > +++ b/drivers/thermal/qcom/lmh.c > @@ -147,23 +147,23 @@ static int lmh_probe(struct platform_device *pdev) > return -EINVAL; > > if (flags & LMH_ENABLE_ALGOS) { > - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, > - LMH_NODE_DCVS, node_id, 0); > + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, 0, > + LMH_NODE_DCVS, node_id, 0, false); > if (ret) > dev_err(dev, "Error %d enabling current subfunction\n", ret); > > - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_REL, LMH_ALGO_MODE_ENABLE, 1, > - LMH_NODE_DCVS, node_id, 0); > + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_REL, LMH_ALGO_MODE_ENABLE, 1, 0, > + LMH_NODE_DCVS, node_id, 0, false); > if (ret) > dev_err(dev, "Error %d enabling reliability subfunction\n", ret); > > - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_BCL, LMH_ALGO_MODE_ENABLE, 1, > - LMH_NODE_DCVS, node_id, 0); > + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_BCL, LMH_ALGO_MODE_ENABLE, 1, 0, > + LMH_NODE_DCVS, node_id, 0, false); > if (ret) > dev_err(dev, "Error %d enabling BCL subfunction\n", ret); > > - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_ALGO_MODE_ENABLE, 1, > - LMH_NODE_DCVS, node_id, 0); > + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_ALGO_MODE_ENABLE, 1, 0, > + LMH_NODE_DCVS, node_id, 0, false); > if (ret) { > dev_err(dev, "Error %d enabling thermal subfunction\n", ret); > return ret; > @@ -177,22 +177,22 @@ static int lmh_probe(struct platform_device *pdev) > } > > /* Set default thermal trips */ > - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_ARM_THRESHOLD, temp_arm, > - LMH_NODE_DCVS, node_id, 0); > + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_ARM_THRESHOLD, temp_arm, 0, > + LMH_NODE_DCVS, node_id, 0, false); > if (ret) { > dev_err(dev, "Error setting thermal ARM threshold%d\n", ret); > return ret; > } > > - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_HI_THRESHOLD, temp_high, > - LMH_NODE_DCVS, node_id, 0); > + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_HI_THRESHOLD, temp_high, 0, > + LMH_NODE_DCVS, node_id, 0, false); > if (ret) { > dev_err(dev, "Error setting thermal HI threshold%d\n", ret); > return ret; > } > > - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_LOW_THRESHOLD, temp_low, > - LMH_NODE_DCVS, node_id, 0); > + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_LOW_THRESHOLD, temp_low, 0, > + LMH_NODE_DCVS, node_id, 0, false); > if (ret) { > dev_err(dev, "Error setting thermal ARM threshold%d\n", ret); > return ret; > diff --git a/include/linux/qcom_scm.h b/include/linux/qcom_scm.h > index 1e449a5d7f5c..9fd798d17fdd 100644 > --- a/include/linux/qcom_scm.h > +++ b/include/linux/qcom_scm.h > @@ -117,8 +117,9 @@ extern int qcom_scm_hdcp_req(struct qcom_scm_hdcp_req *req, u32 req_cnt, > extern int qcom_scm_iommu_set_pt_format(u32 sec_id, u32 ctx_num, u32 pt_fmt); > extern int qcom_scm_qsmmu500_wait_safe_toggle(bool en); > > -extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, > - u64 limit_node, u32 node_id, u64 version); > +extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val0, > + u32 payload_val1, u64 limit_node, u32 node_id, > + u64 version, bool has_val1); > extern int qcom_scm_lmh_profile_change(u32 profile_id); > extern bool qcom_scm_lmh_dcvsh_available(void); > > -- > 2.39.0 > ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] firmware: qcom_scm: Fully implement qcom_scm_lmh_dcvsh() 2023-01-19 3:04 ` Bjorn Andersson @ 2023-01-19 10:41 ` Konrad Dybcio 0 siblings, 0 replies; 5+ messages in thread From: Konrad Dybcio @ 2023-01-19 10:41 UTC (permalink / raw) To: Bjorn Andersson Cc: linux-arm-msm, agross, krzysztof.kozlowski, marijn.suijten, Amit Kucheria, Thara Gopinath, Rafael J. Wysocki, Daniel Lezcano, Zhang Rui, AngeloGioacchino Del Regno, Loic Poulain, Stephan Gerhold, linux-kernel, linux-pm On 19.01.2023 04:04, Bjorn Andersson wrote: > On Fri, Jan 13, 2023 at 04:14:01AM +0100, Konrad Dybcio wrote: >> The qcom_scm_lmh_dcvsh call can actually pass two values to the >> secure world. The second value is used for example with the >> LMH_FREQ_CAP function, which limits the maximum achievable frequency >> directly from LMh. Add the missing arguments, handle them and update >> the current usages of this function. >> >> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> >> --- >> drivers/firmware/qcom_scm.c | 13 ++++++++----- >> drivers/thermal/qcom/lmh.c | 28 ++++++++++++++-------------- >> include/linux/qcom_scm.h | 5 +++-- >> 3 files changed, 25 insertions(+), 21 deletions(-) >> >> diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c >> index cdbfe54c8146..58a19a47e442 100644 >> --- a/drivers/firmware/qcom_scm.c >> +++ b/drivers/firmware/qcom_scm.c >> @@ -1252,12 +1252,13 @@ int qcom_scm_lmh_profile_change(u32 profile_id) >> } >> EXPORT_SYMBOL(qcom_scm_lmh_profile_change); >> >> -int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, >> - u64 limit_node, u32 node_id, u64 version) >> +int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val0, >> + u32 payload_val1, u64 limit_node, u32 node_id, >> + u64 version, bool has_val1) > > Rather than always passing a dummy payload_val1 and then having has_val1 > to indicate if it should be considered or not... how about moving the > payload last in the call, as a va_list with a "count" before that? Sounds neat, but.. > > I.e: > > int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u64 limit_node, u32 node_id, > u64 version, unsigned int payload_count, ...) > >> { >> dma_addr_t payload_phys; >> u32 *payload_buf; >> - int ret, payload_size = 5 * sizeof(u32); >> + int ret, payload_size = (5 + has_val1) * sizeof(u32); > > allocate 4 + payload_count > >> >> struct qcom_scm_desc desc = { >> .svc = QCOM_SCM_SVC_LMH, >> @@ -1278,8 +1279,10 @@ int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, >> payload_buf[0] = payload_fn; >> payload_buf[1] = 0; >> payload_buf[2] = payload_reg; >> - payload_buf[3] = 1; >> - payload_buf[4] = payload_val; >> + payload_buf[3] = has_val1 ? 2 : 1; >> + payload_buf[4] = payload_val0; >> + if (has_val1) >> + payload_buf[5] = payload_val1; > > Something like: > > payload_buf[3] = payload_count; > va_start(ap, payload_count); > for (i = 0; i < payload_count; i++) > payload_buf[4 + i] = va_arg(ap, uint32_t); > va_end(ap); ..can the call accept more arguments? And will they be interpreted in any way? Otherwise I may add also add WARN_ON() or something like this to prevent user error. > > > > That said, I don't see a single "true" below. Perhaps I'm missing it? I > would expect some code in the same series use the newly introduced > logic. Yeah there's no "true"s, this patch only refactored the code in preparation for 8998/660, but adding them as-is makes little sense before LMh_lite is also supported (AFAIUU this LMh_normal part is just an interface for OSM and the actual limits programming is either done on an internal-consensus-between-all-3-blocks basis OR just by LMh_lite. I can delay resending this series until the changes are actually required if you prefer. On newer SoCs LMh wakes up as part of OSM_secure/EPSS programming and needs little to no configuration externally (as you can see in this driver) and there's no external _lite block. Konrad > > Thanks, > Bjorn > >> >> desc.args[0] = payload_phys; >> >> diff --git a/drivers/thermal/qcom/lmh.c b/drivers/thermal/qcom/lmh.c >> index 5e8ff196c9a6..d2b5ea8322eb 100644 >> --- a/drivers/thermal/qcom/lmh.c >> +++ b/drivers/thermal/qcom/lmh.c >> @@ -147,23 +147,23 @@ static int lmh_probe(struct platform_device *pdev) >> return -EINVAL; >> >> if (flags & LMH_ENABLE_ALGOS) { >> - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, >> - LMH_NODE_DCVS, node_id, 0); >> + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_CRNT, LMH_ALGO_MODE_ENABLE, 1, 0, >> + LMH_NODE_DCVS, node_id, 0, false); >> if (ret) >> dev_err(dev, "Error %d enabling current subfunction\n", ret); >> >> - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_REL, LMH_ALGO_MODE_ENABLE, 1, >> - LMH_NODE_DCVS, node_id, 0); >> + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_REL, LMH_ALGO_MODE_ENABLE, 1, 0, >> + LMH_NODE_DCVS, node_id, 0, false); >> if (ret) >> dev_err(dev, "Error %d enabling reliability subfunction\n", ret); >> >> - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_BCL, LMH_ALGO_MODE_ENABLE, 1, >> - LMH_NODE_DCVS, node_id, 0); >> + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_BCL, LMH_ALGO_MODE_ENABLE, 1, 0, >> + LMH_NODE_DCVS, node_id, 0, false); >> if (ret) >> dev_err(dev, "Error %d enabling BCL subfunction\n", ret); >> >> - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_ALGO_MODE_ENABLE, 1, >> - LMH_NODE_DCVS, node_id, 0); >> + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_ALGO_MODE_ENABLE, 1, 0, >> + LMH_NODE_DCVS, node_id, 0, false); >> if (ret) { >> dev_err(dev, "Error %d enabling thermal subfunction\n", ret); >> return ret; >> @@ -177,22 +177,22 @@ static int lmh_probe(struct platform_device *pdev) >> } >> >> /* Set default thermal trips */ >> - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_ARM_THRESHOLD, temp_arm, >> - LMH_NODE_DCVS, node_id, 0); >> + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_ARM_THRESHOLD, temp_arm, 0, >> + LMH_NODE_DCVS, node_id, 0, false); >> if (ret) { >> dev_err(dev, "Error setting thermal ARM threshold%d\n", ret); >> return ret; >> } >> >> - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_HI_THRESHOLD, temp_high, >> - LMH_NODE_DCVS, node_id, 0); >> + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_HI_THRESHOLD, temp_high, 0, >> + LMH_NODE_DCVS, node_id, 0, false); >> if (ret) { >> dev_err(dev, "Error setting thermal HI threshold%d\n", ret); >> return ret; >> } >> >> - ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_LOW_THRESHOLD, temp_low, >> - LMH_NODE_DCVS, node_id, 0); >> + ret = qcom_scm_lmh_dcvsh(LMH_SUB_FN_THERMAL, LMH_TH_LOW_THRESHOLD, temp_low, 0, >> + LMH_NODE_DCVS, node_id, 0, false); >> if (ret) { >> dev_err(dev, "Error setting thermal ARM threshold%d\n", ret); >> return ret; >> diff --git a/include/linux/qcom_scm.h b/include/linux/qcom_scm.h >> index 1e449a5d7f5c..9fd798d17fdd 100644 >> --- a/include/linux/qcom_scm.h >> +++ b/include/linux/qcom_scm.h >> @@ -117,8 +117,9 @@ extern int qcom_scm_hdcp_req(struct qcom_scm_hdcp_req *req, u32 req_cnt, >> extern int qcom_scm_iommu_set_pt_format(u32 sec_id, u32 ctx_num, u32 pt_fmt); >> extern int qcom_scm_qsmmu500_wait_safe_toggle(bool en); >> >> -extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val, >> - u64 limit_node, u32 node_id, u64 version); >> +extern int qcom_scm_lmh_dcvsh(u32 payload_fn, u32 payload_reg, u32 payload_val0, >> + u32 payload_val1, u64 limit_node, u32 node_id, >> + u64 version, bool has_val1); >> extern int qcom_scm_lmh_profile_change(u32 profile_id); >> extern bool qcom_scm_lmh_dcvsh_available(void); >> >> -- >> 2.39.0 >> ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2023-01-19 10:42 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <20230113031401.2336157-1-konrad.dybcio@linaro.org>
2023-01-13 3:14 ` [PATCH 1/2] thermal/drivers/qcom/lmh: Use driver data as flags instead of bool Konrad Dybcio
2023-01-19 3:05 ` Bjorn Andersson
2023-01-13 3:14 ` [PATCH 2/2] firmware: qcom_scm: Fully implement qcom_scm_lmh_dcvsh() Konrad Dybcio
2023-01-19 3:04 ` Bjorn Andersson
2023-01-19 10:41 ` Konrad Dybcio
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).