drivers/media/platform/qcom/venus/pm_helpers.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)
Since the device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX
do not include an opp-table and have not configured opp-pmdomain, they
still need to use the frequencies defined in the driver's freq_tbl.
Both core_power_v1 and core_power_v4 functions require core_clks_enable
function during POWER_ON. Therefore, in the core_clks_enable function,
if calling dev_pm_opp_find_freq_ceil to obtain the frequency fails,
it needs to fall back to the freq_tbl to retrieve the frequency.
Signed-off-by: Renjiang Han <quic_renjiang@quicinc.com>
---
Since device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX do not
contain an opp-table and have not configured opp-pmdomain, they still
need to use the frequencies defined in the driver's freq_tbl.
Therefore, if calling dev_pm_opp_find_freq_ceil to obtain the frequency
fails in the core_clks_enable, it needs to fall back to the freq_tbl to
retrieve the frequency.
Validated this series on QCS615 and msm8916.
---
drivers/media/platform/qcom/venus/pm_helpers.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/drivers/media/platform/qcom/venus/pm_helpers.c b/drivers/media/platform/qcom/venus/pm_helpers.c
index 8dd5a9b0d060cddfeafd4da477ade0c7aeb6c390..27b033a3a26e0474c37353a0fe7627586e6416e2 100644
--- a/drivers/media/platform/qcom/venus/pm_helpers.c
+++ b/drivers/media/platform/qcom/venus/pm_helpers.c
@@ -40,6 +40,8 @@ static int core_clks_get(struct venus_core *core)
static int core_clks_enable(struct venus_core *core)
{
+ const struct freq_tbl *freq_tbl = core->res->freq_tbl;
+ unsigned int freq_tbl_size = core->res->freq_tbl_size;
const struct venus_resources *res = core->res;
struct device *dev = core->dev;
unsigned long freq = 0;
@@ -48,7 +50,14 @@ static int core_clks_enable(struct venus_core *core)
int ret;
opp = dev_pm_opp_find_freq_ceil(dev, &freq);
- dev_pm_opp_put(opp);
+ if (IS_ERR(opp)) {
+ if (!freq_tbl)
+ return -EINVAL;
+ freq = freq_tbl[freq_tbl_size - 1].freq;
+ } else {
+ dev_pm_opp_put(opp);
+ }
+
for (i = 0; i < res->clks_num; i++) {
if (IS_V6(core)) {
---
base-commit: d086c886ceb9f59dea6c3a9dae7eb89e780a20c9
change-id: 20250721-fallback_of_opp_table-4ea39376f617
Best regards,
--
Renjiang Han <quic_renjiang@quicinc.com>
On 7/23/25 2:56 PM, Renjiang Han wrote: > Since the device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX > do not include an opp-table and have not configured opp-pmdomain, they > still need to use the frequencies defined in the driver's freq_tbl. > > Both core_power_v1 and core_power_v4 functions require core_clks_enable > function during POWER_ON. Therefore, in the core_clks_enable function, > if calling dev_pm_opp_find_freq_ceil to obtain the frequency fails, > it needs to fall back to the freq_tbl to retrieve the frequency. > > Signed-off-by: Renjiang Han <quic_renjiang@quicinc.com> > --- > Since device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX do not > contain an opp-table and have not configured opp-pmdomain, they still > need to use the frequencies defined in the driver's freq_tbl. > > Therefore, if calling dev_pm_opp_find_freq_ceil to obtain the frequency > fails in the core_clks_enable, it needs to fall back to the freq_tbl to > retrieve the frequency. > > Validated this series on QCS615 and msm8916. > --- This is not a proper fix, the logic you added in core_get_v4() should be moved to a common handler Konrad
On 7/23/25 3:30 PM, Konrad Dybcio wrote: > On 7/23/25 2:56 PM, Renjiang Han wrote: >> Since the device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX >> do not include an opp-table and have not configured opp-pmdomain, they >> still need to use the frequencies defined in the driver's freq_tbl. >> >> Both core_power_v1 and core_power_v4 functions require core_clks_enable >> function during POWER_ON. Therefore, in the core_clks_enable function, >> if calling dev_pm_opp_find_freq_ceil to obtain the frequency fails, >> it needs to fall back to the freq_tbl to retrieve the frequency. >> >> Signed-off-by: Renjiang Han <quic_renjiang@quicinc.com> >> --- >> Since device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX do not >> contain an opp-table and have not configured opp-pmdomain, they still >> need to use the frequencies defined in the driver's freq_tbl. >> >> Therefore, if calling dev_pm_opp_find_freq_ceil to obtain the frequency >> fails in the core_clks_enable, it needs to fall back to the freq_tbl to >> retrieve the frequency. >> >> Validated this series on QCS615 and msm8916. >> --- > > This is not a proper fix, the logic you added in core_get_v4() should > be moved to a common handler Well, it is a proper fix in the sense that it makes things work, but you're expanding the delta between the power handling across generations, even though it's almost the same.. I submitted a series to clean it up once, but it didn't go anywhere https://lore.kernel.org/linux-arm-msm/20230911-topic-mars-v3-0-79f23b81c261@linaro.org/ Konrad
On Wed, Jul 23, 2025 at 06:26:02PM +0530, Renjiang Han wrote: > Since the device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX > do not include an opp-table and have not configured opp-pmdomain, they > still need to use the frequencies defined in the driver's freq_tbl. > > Both core_power_v1 and core_power_v4 functions require core_clks_enable > function during POWER_ON. Therefore, in the core_clks_enable function, > if calling dev_pm_opp_find_freq_ceil to obtain the frequency fails, > it needs to fall back to the freq_tbl to retrieve the frequency. > > Signed-off-by: Renjiang Han <quic_renjiang@quicinc.com> > --- > Since device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX do not > contain an opp-table and have not configured opp-pmdomain, they still > need to use the frequencies defined in the driver's freq_tbl. > > Therefore, if calling dev_pm_opp_find_freq_ceil to obtain the frequency > fails in the core_clks_enable, it needs to fall back to the freq_tbl to > retrieve the frequency. > > Validated this series on QCS615 and msm8916. > --- > drivers/media/platform/qcom/venus/pm_helpers.c | 11 ++++++++++- > 1 file changed, 10 insertions(+), 1 deletion(-) Fixes: b179234b5e59 ("media: venus: pm_helpers: use opp-table for the frequency") Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com> -- With best wishes Dmitry
On 23/07/2025 13:56, Renjiang Han wrote: > Since the device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX > do not include an opp-table and have not configured opp-pmdomain, they > still need to use the frequencies defined in the driver's freq_tbl. > > Both core_power_v1 and core_power_v4 functions require core_clks_enable > function during POWER_ON. Therefore, in the core_clks_enable function, > if calling dev_pm_opp_find_freq_ceil to obtain the frequency fails, > it needs to fall back to the freq_tbl to retrieve the frequency. > > Signed-off-by: Renjiang Han <quic_renjiang@quicinc.com> > --- > Since device trees for both HFI_VERSION_1XX and HFI_VERSION_3XX do not > contain an opp-table and have not configured opp-pmdomain, they still > need to use the frequencies defined in the driver's freq_tbl. > > Therefore, if calling dev_pm_opp_find_freq_ceil to obtain the frequency > fails in the core_clks_enable, it needs to fall back to the freq_tbl to > retrieve the frequency. > > Validated this series on QCS615 and msm8916. > --- > drivers/media/platform/qcom/venus/pm_helpers.c | 11 ++++++++++- > 1 file changed, 10 insertions(+), 1 deletion(-) > > diff --git a/drivers/media/platform/qcom/venus/pm_helpers.c b/drivers/media/platform/qcom/venus/pm_helpers.c > index 8dd5a9b0d060cddfeafd4da477ade0c7aeb6c390..27b033a3a26e0474c37353a0fe7627586e6416e2 100644 > --- a/drivers/media/platform/qcom/venus/pm_helpers.c > +++ b/drivers/media/platform/qcom/venus/pm_helpers.c > @@ -40,6 +40,8 @@ static int core_clks_get(struct venus_core *core) > > static int core_clks_enable(struct venus_core *core) > { > + const struct freq_tbl *freq_tbl = core->res->freq_tbl; > + unsigned int freq_tbl_size = core->res->freq_tbl_size; > const struct venus_resources *res = core->res; > struct device *dev = core->dev; > unsigned long freq = 0; > @@ -48,7 +50,14 @@ static int core_clks_enable(struct venus_core *core) > int ret; > > opp = dev_pm_opp_find_freq_ceil(dev, &freq); > - dev_pm_opp_put(opp); > + if (IS_ERR(opp)) { > + if (!freq_tbl) > + return -EINVAL; -ENODEV > + freq = freq_tbl[freq_tbl_size - 1].freq; > + } else { > + dev_pm_opp_put(opp); > + } > + > > for (i = 0; i < res->clks_num; i++) { > if (IS_V6(core)) { > > --- > base-commit: d086c886ceb9f59dea6c3a9dae7eb89e780a20c9 > change-id: 20250721-fallback_of_opp_table-4ea39376f617 > > Best regards, With that changed. Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
© 2016 - 2025 Red Hat, Inc.