Skip to contentSkip to conversation

media: qcom: iris: rollback OPP vote on PM resume failure

PATCH In review Hungyu Lin 2 participants 4 messages Jun 4 – Jul 29

Subsystems (4)
lore
Share & tools
lore

Hungyu Lin

author
#0 lore raw

PATCH 1/1 media: qcom: iris: rollback OPP vote on PM resume failure

Rollback the maximum OPP vote when
pm_runtime_resume_and_get() fails in
iris_enable_power_domains().

This ensures the performance state vote is released when power domain enablement does not complete successfully.

Fixes bb8a95aa038e ("media: iris: implement power management")
---
 drivers/media/platform/qcom/iris/iris_resources.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/media/platform/qcom/iris/iris_resources.c +3 -1
index f5e3341e1430..a7d25f7a201a 100644
@@ -79,8 +79,10 @@ int iris_enable_power_domains(struct iris_core *core, struct device *pd_dev)
 		return ret;
 
 	ret = pm_runtime_resume_and_get(pd_dev);
-	if (ret)
+	if (ret) {
+		iris_opp_set_rate(core->dev, 0);
 		return ret;
+	}
 
 	return 0;
 }
Signature · 1 line
2.34.1
#1 lore raw
Quoted 18 lines · Hungyu Lin

Rollback the maximum OPP vote when
pm_runtime_resume_and_get() fails in
iris_enable_power_domains().

This ensures the performance state vote is released when power domain enablement does not complete successfully.

Fixes bb8a95aa038e ("media: iris: implement power management")
---
 drivers/media/platform/qcom/iris/iris_resources.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/media/platform/qcom/iris/iris_resources.c +0 -0
index f5e3341e1430..a7d25f7a201a 100644

drivers/media/platform/qcom/iris/iris_resources.c int iris_enable_power_domains(struct iris_core *core, struct device *pd_dev) · line 85

Quoted 8 lines
drivers/media/platform/qcom/iris/iris_resources.c +3 -1
@@ -79,8 +79,10 @@ int iris_enable_power_domains(struct iris_core *core, struct device *pd_dev)
 		return ret;
 
 	ret = pm_runtime_resume_and_get(pd_dev);
-	if (ret)
+	if (ret) {
+		iris_opp_set_rate(core->dev, 0);
 		return ret;
+	}

opp_set_rate(0) would only drop the performance vote but keep the clock ticking (without ensuring power). If the device's power state remains disabled, the OPP vote will be kept disabled too. I don't think we need it at all. The clock may possibly need to be disabled instead. Or maybe not, if it's hw-controlled (which I'm not sure if it is)

Konrad

author
#2 lore raw
Quoted 6 lines

opp_set_rate(0) would only drop the performance vote but keep the clock ticking (without ensuring power). If the device's power state remains disabled, the OPP vote will be kept disabled too. I don't think we need it at all. The clock may possibly need to be disabled instead. Or maybe not, if it's hw-controlled (which I'm not sure if it is)

Thanks for the review.

After tracing the call flow further, I found that the maximum OPP belongs to the complete VPU power-on transaction:

    iris_vpu_power_on()
    |-- iris_set_icc_bw()
    |-- iris_opp_set_rate(..., ULONG_MAX)
    |-- power_on_controller()
    |   `-- iris_enable_power_domains(...)
    `-- power_on_hw()
        `-- iris_enable_power_domains(...)

The power-domain helper only manages one functional domain, so it should not own the transaction-wide OPP state. I will move the OPP lifetime and error cleanup to iris_vpu_power_on() and send the update as v7 after validation.

Also, sorry for the versioning confusion. The previous unversioned posting was intended to be v6.

Thanks,
Hungyu Lin

author
#3 lore raw
Quoted 5 lines

as v7 after validation.

Also, sorry for the versioning confusion. The previous unversioned posting was intended to be v6.

Correction: I got the version number wrong. The unversioned patch was sent before v5, so the next version will be v6, not v7.

Sorry for the confusion.

Thanks,
Hungyu Lin