| Message ID | 20210924152334.1342630-1-maxime@cerno.tech (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] drm/vc4: Unselect PM | expand |
On Fri, Sep 24, 2021 at 05:23:34PM +0200, Maxime Ripard wrote: > It turns out we can't select PM when allowing the compilation for > COMPILE_TEST. Indeed, PM might not be defined at all, or come with extra > requirements we can't meet. > > This select was initially introduced since we need to call the > vc4_hdmi_runtime_resume() at probe time to make sure our device is > properly powered at bind time, no matter whether PM is there or not, and > we needed to make sure we didn't have a defined but not used warning for > vc4_hdmi_runtime_suspend(). > > This will still happen on platforms that don't define PM though, since > SET_RUNTIME_PM_OPS will then be a nop. In order to fix both issues, > let's remove the select, and add a __maybe_unused attribute to > vc4_hdmi_runtime_resume(). > > Reported-by: Randy Dunlap <rdunlap@infradead.org> > Signed-off-by: Maxime Ripard <maxime@cerno.tech> Reviewed-by: Nathan Chancellor <nathan@kernel.org> Tested-by: Nathan Chancellor <nathan@kernel.org> # build > --- > > I'm not sure how to merge this one, since this commit has been reverted > in Linus tree, and un-reverted in linux-next. Should we wait a bit until > the reworked version of the original commit has been merged again? > > Maxime > > Changes from v1: > - remove select > - add __maybe_unused to vc4_hdmi_runtime_resume > - reword the commit log > --- > drivers/gpu/drm/vc4/Kconfig | 1 - > drivers/gpu/drm/vc4/vc4_hdmi.c | 2 +- > 2 files changed, 1 insertion(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/vc4/Kconfig b/drivers/gpu/drm/vc4/Kconfig > index 52a1c309cb4a..345a5570a3da 100644 > --- a/drivers/gpu/drm/vc4/Kconfig > +++ b/drivers/gpu/drm/vc4/Kconfig > @@ -9,7 +9,6 @@ config DRM_VC4 > select DRM_KMS_CMA_HELPER > select DRM_GEM_CMA_HELPER > select DRM_PANEL_BRIDGE > - select PM > select SND_PCM > select SND_PCM_ELD > select SND_SOC_GENERIC_DMAENGINE_PCM > diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c > index 500cdd56b335..1f2690ed8542 100644 > --- a/drivers/gpu/drm/vc4/vc4_hdmi.c > +++ b/drivers/gpu/drm/vc4/vc4_hdmi.c > @@ -2122,7 +2122,7 @@ static int vc5_hdmi_init_resources(struct vc4_hdmi *vc4_hdmi) > return 0; > } > > -static int vc4_hdmi_runtime_suspend(struct device *dev) > +static int __maybe_unused vc4_hdmi_runtime_suspend(struct device *dev) > { > struct vc4_hdmi *vc4_hdmi = dev_get_drvdata(dev); > > -- > 2.31.1 >
On 9/24/21 8:23 AM, Maxime Ripard wrote: > It turns out we can't select PM when allowing the compilation for > COMPILE_TEST. Indeed, PM might not be defined at all, or come with extra > requirements we can't meet. > > This select was initially introduced since we need to call the > vc4_hdmi_runtime_resume() at probe time to make sure our device is > properly powered at bind time, no matter whether PM is there or not, and > we needed to make sure we didn't have a defined but not used warning for > vc4_hdmi_runtime_suspend(). > > This will still happen on platforms that don't define PM though, since > SET_RUNTIME_PM_OPS will then be a nop. In order to fix both issues, > let's remove the select, and add a __maybe_unused attribute to > vc4_hdmi_runtime_resume(). > > Reported-by: Randy Dunlap <rdunlap@infradead.org> > Signed-off-by: Maxime Ripard <maxime@cerno.tech> > LGTM. Thanks. Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested > --- > > I'm not sure how to merge this one, since this commit has been reverted > in Linus tree, and un-reverted in linux-next. Should we wait a bit until > the reworked version of the original commit has been merged again? > > Maxime > > Changes from v1: > - remove select > - add __maybe_unused to vc4_hdmi_runtime_resume > - reword the commit log > --- > drivers/gpu/drm/vc4/Kconfig | 1 - > drivers/gpu/drm/vc4/vc4_hdmi.c | 2 +- > 2 files changed, 1 insertion(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/vc4/Kconfig b/drivers/gpu/drm/vc4/Kconfig > index 52a1c309cb4a..345a5570a3da 100644 > --- a/drivers/gpu/drm/vc4/Kconfig > +++ b/drivers/gpu/drm/vc4/Kconfig > @@ -9,7 +9,6 @@ config DRM_VC4 > select DRM_KMS_CMA_HELPER > select DRM_GEM_CMA_HELPER > select DRM_PANEL_BRIDGE > - select PM > select SND_PCM > select SND_PCM_ELD > select SND_SOC_GENERIC_DMAENGINE_PCM > diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c > index 500cdd56b335..1f2690ed8542 100644 > --- a/drivers/gpu/drm/vc4/vc4_hdmi.c > +++ b/drivers/gpu/drm/vc4/vc4_hdmi.c > @@ -2122,7 +2122,7 @@ static int vc5_hdmi_init_resources(struct vc4_hdmi *vc4_hdmi) > return 0; > } > > -static int vc4_hdmi_runtime_suspend(struct device *dev) > +static int __maybe_unused vc4_hdmi_runtime_suspend(struct device *dev) > { > struct vc4_hdmi *vc4_hdmi = dev_get_drvdata(dev); > >
diff --git a/drivers/gpu/drm/vc4/Kconfig b/drivers/gpu/drm/vc4/Kconfig index 52a1c309cb4a..345a5570a3da 100644 --- a/drivers/gpu/drm/vc4/Kconfig +++ b/drivers/gpu/drm/vc4/Kconfig @@ -9,7 +9,6 @@ config DRM_VC4 select DRM_KMS_CMA_HELPER select DRM_GEM_CMA_HELPER select DRM_PANEL_BRIDGE - select PM select SND_PCM select SND_PCM_ELD select SND_SOC_GENERIC_DMAENGINE_PCM diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c index 500cdd56b335..1f2690ed8542 100644 --- a/drivers/gpu/drm/vc4/vc4_hdmi.c +++ b/drivers/gpu/drm/vc4/vc4_hdmi.c @@ -2122,7 +2122,7 @@ static int vc5_hdmi_init_resources(struct vc4_hdmi *vc4_hdmi) return 0; } -static int vc4_hdmi_runtime_suspend(struct device *dev) +static int __maybe_unused vc4_hdmi_runtime_suspend(struct device *dev) { struct vc4_hdmi *vc4_hdmi = dev_get_drvdata(dev);
It turns out we can't select PM when allowing the compilation for COMPILE_TEST. Indeed, PM might not be defined at all, or come with extra requirements we can't meet. This select was initially introduced since we need to call the vc4_hdmi_runtime_resume() at probe time to make sure our device is properly powered at bind time, no matter whether PM is there or not, and we needed to make sure we didn't have a defined but not used warning for vc4_hdmi_runtime_suspend(). This will still happen on platforms that don't define PM though, since SET_RUNTIME_PM_OPS will then be a nop. In order to fix both issues, let's remove the select, and add a __maybe_unused attribute to vc4_hdmi_runtime_resume(). Reported-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> --- I'm not sure how to merge this one, since this commit has been reverted in Linus tree, and un-reverted in linux-next. Should we wait a bit until the reworked version of the original commit has been merged again? Maxime Changes from v1: - remove select - add __maybe_unused to vc4_hdmi_runtime_resume - reword the commit log --- drivers/gpu/drm/vc4/Kconfig | 1 - drivers/gpu/drm/vc4/vc4_hdmi.c | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-)