2022-10-26 08:42:57

by Erico Nunes

Subject: [PATCH] drm/lima: Fix dev_pm_opp_set_config in case of missing regulator

Commit d8c32d3971e4 ("drm/lima: Migrate to dev_pm_opp_set_config()")
introduced a regression as it may undo the clk_names setting in case
the optional regulator is missing. This resulted in test and performance
regressions with lima.

Restore the old behavior where clk_names is set separately so it is not
undone in case of a missing optional regulator.

Fixes: d8c32d3971e4 ("drm/lima: Migrate to dev_pm_opp_set_config()")
Signed-off-by: Erico Nunes <[email protected]>
---
drivers/gpu/drm/lima/lima_devfreq.c | 16 +++++++++++++---
1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/lima/lima_devfreq.c b/drivers/gpu/drm/lima/lima_devfreq.c
index 011be7ff51e1..9c8654934fea 100644
--- a/drivers/gpu/drm/lima/lima_devfreq.c
+++ b/drivers/gpu/drm/lima/lima_devfreq.c
@@ -113,10 +113,12 @@ int lima_devfreq_init(struct lima_device *ldev)
int ret;
const char *regulator_names[] = { "mali", NULL };
const char *clk_names[] = { "core", NULL };
- struct dev_pm_opp_config config = {
- .regulator_names = regulator_names,
+ struct dev_pm_opp_config config_clk_names = {
.clk_names = clk_names,
};
+ struct dev_pm_opp_config config_regulator_names = {
+ .regulator_names = regulator_names,
+ };

if (!device_property_present(dev, "operating-points-v2"))
/* Optional, continue without devfreq */
@@ -124,7 +126,15 @@ int lima_devfreq_init(struct lima_device *ldev)

spin_lock_init(&ldevfreq->lock);

- ret = devm_pm_opp_set_config(dev, &config);
+ /*
+ * Set clk_names separately so it is not undone in case of
+ * a missing optional regulator.
+ */
+ ret = devm_pm_opp_set_config(dev, &config_clk_names);
+ if (ret)
+ return ret;
+
+ ret = devm_pm_opp_set_config(dev, &config_regulator_names);
if (ret) {
/* Continue if the optional regulator is missing */
if (ret != -ENODEV)
--
2.37.3

2022-10-26 09:38:15

by Viresh Kumar

Subject: Re: [PATCH] drm/lima: Fix dev_pm_opp_set_config in case of missing regulator

On 26-10-22, 10:39, Erico Nunes wrote:
> Commit d8c32d3971e4 ("drm/lima: Migrate to dev_pm_opp_set_config()")
> introduced a regression as it may undo the clk_names setting in case
> the optional regulator is missing. This resulted in test and performance
> regressions with lima.
>
> Restore the old behavior where clk_names is set separately so it is not
> undone in case of a missing optional regulator.
>
> Fixes: d8c32d3971e4 ("drm/lima: Migrate to dev_pm_opp_set_config()")
> Signed-off-by: Erico Nunes <[email protected]>
> ---
> drivers/gpu/drm/lima/lima_devfreq.c | 16 +++++++++++++---
> 1 file changed, 13 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/lima/lima_devfreq.c b/drivers/gpu/drm/lima/lima_devfreq.c
> index 011be7ff51e1..9c8654934fea 100644
> --- a/drivers/gpu/drm/lima/lima_devfreq.c
> +++ b/drivers/gpu/drm/lima/lima_devfreq.c
> @@ -113,10 +113,12 @@ int lima_devfreq_init(struct lima_device *ldev)
> int ret;
> const char *regulator_names[] = { "mali", NULL };
> const char *clk_names[] = { "core", NULL };
> - struct dev_pm_opp_config config = {
> - .regulator_names = regulator_names,
> + struct dev_pm_opp_config config_clk_names = {
> .clk_names = clk_names,
> };
> + struct dev_pm_opp_config config_regulator_names = {
> + .regulator_names = regulator_names,
> + };
>
> if (!device_property_present(dev, "operating-points-v2"))
> /* Optional, continue without devfreq */
> @@ -124,7 +126,15 @@ int lima_devfreq_init(struct lima_device *ldev)
>
> spin_lock_init(&ldevfreq->lock);
>
> - ret = devm_pm_opp_set_config(dev, &config);
> + /*
> + * Set clk_names separately so it is not undone in case of
> + * a missing optional regulator.
> + */
> + ret = devm_pm_opp_set_config(dev, &config_clk_names);
> + if (ret)
> + return ret;
> +
> + ret = devm_pm_opp_set_config(dev, &config_regulator_names);

You can directly call devm_pm_opp_set_clkname() and
devm_pm_opp_set_regulators(), like it was done before my patch, for
single configurations. So basically a simple revert of my commit, with
additional comments you added above.

> if (ret) {
> /* Continue if the optional regulator is missing */
> if (ret != -ENODEV)
> --
> 2.37.3

--
viresh

2022-10-26 13:09:53

by Erico Nunes

Subject: Re: [PATCH] drm/lima: Fix dev_pm_opp_set_config in case of missing regulator

Hello,

On Wed, Oct 26, 2022 at 11:36 AM Viresh Kumar <[email protected]> wrote:
> You can directly call devm_pm_opp_set_clkname() and
> devm_pm_opp_set_regulators(), like it was done before my patch, for
> single configurations. So basically a simple revert of my commit, with
> additional comments you added above.

Ok, I'll send a v2 with that.

Thanks

Erico