| Message ID | 643cc0a4d12f47fd8403d42581e83b1e9c4543c7.1726680898.git.jani.nikula@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm/i915: wakeref fixes and improvements | expand |
On Wed, Sep 18, 2024 at 08:35:43PM +0300, Jani Nikula wrote: > CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND is an int, defaulting to 250. When > the wakeref is non-zero, it's either -1 or a dynamically allocated > pointer, depending on CONFIG_DRM_I915_DEBUG_RUNTIME_PM. It's likely that > the code works by coincidence with the bitwise AND, but with > CONFIG_DRM_I915_DEBUG_RUNTIME_PM=y, there's the off chance that the > condition evaluates to false, and intel_wakeref_auto() doesn't get > called. Switch to the intended logical AND. > > v2: Use != to avoid clang -Wconstant-logical-operand (Nathan) oh, this is ugly! Wouldn't it be better then to use IS_ENABLED() macro? > > Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend") > Cc: Matthew Auld <matthew.auld@intel.com> > Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> > Cc: Anshuman Gupta <anshuman.gupta@intel.com> > Cc: Andi Shyti <andi.shyti@linux.intel.com> > Cc: Nathan Chancellor <nathan@kernel.org> > Cc: <stable@vger.kernel.org> # v6.1+ > Reviewed-by: Matthew Auld <matthew.auld@intel.com> # v1 > Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> # v1 > Signed-off-by: Jani Nikula <jani.nikula@intel.com> > --- > drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c > index 5c72462d1f57..b22e2019768f 100644 > --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c > +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c > @@ -1131,7 +1131,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault *vmf) > GEM_WARN_ON(!i915_ttm_cpu_maps_iomem(bo->resource)); > } > > - if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) > + if (wakeref && CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND != 0) > intel_wakeref_auto(&to_i915(obj->base.dev)->runtime_pm.userfault_wakeref, > msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)); > > -- > 2.39.2 >
On Thu, 26 Sep 2024, Rodrigo Vivi <rodrigo.vivi@intel.com> wrote: > On Wed, Sep 18, 2024 at 08:35:43PM +0300, Jani Nikula wrote: >> CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND is an int, defaulting to 250. When >> the wakeref is non-zero, it's either -1 or a dynamically allocated >> pointer, depending on CONFIG_DRM_I915_DEBUG_RUNTIME_PM. It's likely that >> the code works by coincidence with the bitwise AND, but with >> CONFIG_DRM_I915_DEBUG_RUNTIME_PM=y, there's the off chance that the >> condition evaluates to false, and intel_wakeref_auto() doesn't get >> called. Switch to the intended logical AND. >> >> v2: Use != to avoid clang -Wconstant-logical-operand (Nathan) > > oh, this is ugly! > > Wouldn't it be better then to use IS_ENABLED() macro? It's an int config option, not a bool. (Yes, the name is misleading.) IS_ENABLED(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) would be the same as CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND == 1. We're actually checking if the int value != 0, so IMO the patch at hand is fine. BR, Jani. > >> >> Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend") >> Cc: Matthew Auld <matthew.auld@intel.com> >> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> >> Cc: Anshuman Gupta <anshuman.gupta@intel.com> >> Cc: Andi Shyti <andi.shyti@linux.intel.com> >> Cc: Nathan Chancellor <nathan@kernel.org> >> Cc: <stable@vger.kernel.org> # v6.1+ >> Reviewed-by: Matthew Auld <matthew.auld@intel.com> # v1 >> Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> # v1 >> Signed-off-by: Jani Nikula <jani.nikula@intel.com> >> --- >> drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >> index 5c72462d1f57..b22e2019768f 100644 >> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >> @@ -1131,7 +1131,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault *vmf) >> GEM_WARN_ON(!i915_ttm_cpu_maps_iomem(bo->resource)); >> } >> >> - if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) >> + if (wakeref && CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND != 0) >> intel_wakeref_auto(&to_i915(obj->base.dev)->runtime_pm.userfault_wakeref, >> msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)); >> >> -- >> 2.39.2 >>
On Fri, 27 Sep 2024, Jani Nikula <jani.nikula@intel.com> wrote: > On Thu, 26 Sep 2024, Rodrigo Vivi <rodrigo.vivi@intel.com> wrote: >> On Wed, Sep 18, 2024 at 08:35:43PM +0300, Jani Nikula wrote: >>> CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND is an int, defaulting to 250. When >>> the wakeref is non-zero, it's either -1 or a dynamically allocated >>> pointer, depending on CONFIG_DRM_I915_DEBUG_RUNTIME_PM. It's likely that >>> the code works by coincidence with the bitwise AND, but with >>> CONFIG_DRM_I915_DEBUG_RUNTIME_PM=y, there's the off chance that the >>> condition evaluates to false, and intel_wakeref_auto() doesn't get >>> called. Switch to the intended logical AND. >>> >>> v2: Use != to avoid clang -Wconstant-logical-operand (Nathan) >> >> oh, this is ugly! >> >> Wouldn't it be better then to use IS_ENABLED() macro? > > It's an int config option, not a bool. (Yes, the name is misleading.) > > IS_ENABLED(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) would be the same as > CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND == 1. > > We're actually checking if the int value != 0, so IMO the patch at hand > is fine. Ping, r-b on this one too? BR, Jani. > > BR, > Jani. > > >> >>> >>> Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend") >>> Cc: Matthew Auld <matthew.auld@intel.com> >>> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> >>> Cc: Anshuman Gupta <anshuman.gupta@intel.com> >>> Cc: Andi Shyti <andi.shyti@linux.intel.com> >>> Cc: Nathan Chancellor <nathan@kernel.org> >>> Cc: <stable@vger.kernel.org> # v6.1+ >>> Reviewed-by: Matthew Auld <matthew.auld@intel.com> # v1 >>> Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> # v1 >>> Signed-off-by: Jani Nikula <jani.nikula@intel.com> >>> --- >>> drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>> index 5c72462d1f57..b22e2019768f 100644 >>> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>> @@ -1131,7 +1131,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault *vmf) >>> GEM_WARN_ON(!i915_ttm_cpu_maps_iomem(bo->resource)); >>> } >>> >>> - if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) >>> + if (wakeref && CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND != 0) >>> intel_wakeref_auto(&to_i915(obj->base.dev)->runtime_pm.userfault_wakeref, >>> msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)); >>> >>> -- >>> 2.39.2 >>>
On 30/09/2024 14:54, Jani Nikula wrote: > On Fri, 27 Sep 2024, Jani Nikula <jani.nikula@intel.com> wrote: >> On Thu, 26 Sep 2024, Rodrigo Vivi <rodrigo.vivi@intel.com> wrote: >>> On Wed, Sep 18, 2024 at 08:35:43PM +0300, Jani Nikula wrote: >>>> CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND is an int, defaulting to 250. When >>>> the wakeref is non-zero, it's either -1 or a dynamically allocated >>>> pointer, depending on CONFIG_DRM_I915_DEBUG_RUNTIME_PM. It's likely that >>>> the code works by coincidence with the bitwise AND, but with >>>> CONFIG_DRM_I915_DEBUG_RUNTIME_PM=y, there's the off chance that the >>>> condition evaluates to false, and intel_wakeref_auto() doesn't get >>>> called. Switch to the intended logical AND. >>>> >>>> v2: Use != to avoid clang -Wconstant-logical-operand (Nathan) >>> >>> oh, this is ugly! >>> >>> Wouldn't it be better then to use IS_ENABLED() macro? >> >> It's an int config option, not a bool. (Yes, the name is misleading.) >> >> IS_ENABLED(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) would be the same as >> CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND == 1. >> >> We're actually checking if the int value != 0, so IMO the patch at hand >> is fine. > > Ping, r-b on this one too? r-b still holds for v2, fwiw. > > BR, > Jani. > >> >> BR, >> Jani. >> >> >>> >>>> >>>> Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend") >>>> Cc: Matthew Auld <matthew.auld@intel.com> >>>> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> >>>> Cc: Anshuman Gupta <anshuman.gupta@intel.com> >>>> Cc: Andi Shyti <andi.shyti@linux.intel.com> >>>> Cc: Nathan Chancellor <nathan@kernel.org> >>>> Cc: <stable@vger.kernel.org> # v6.1+ >>>> Reviewed-by: Matthew Auld <matthew.auld@intel.com> # v1 >>>> Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> # v1 >>>> Signed-off-by: Jani Nikula <jani.nikula@intel.com> >>>> --- >>>> drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +- >>>> 1 file changed, 1 insertion(+), 1 deletion(-) >>>> >>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>>> index 5c72462d1f57..b22e2019768f 100644 >>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>>> @@ -1131,7 +1131,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault *vmf) >>>> GEM_WARN_ON(!i915_ttm_cpu_maps_iomem(bo->resource)); >>>> } >>>> >>>> - if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) >>>> + if (wakeref && CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND != 0) >>>> intel_wakeref_auto(&to_i915(obj->base.dev)->runtime_pm.userfault_wakeref, >>>> msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)); >>>> >>>> -- >>>> 2.39.2 >>>> >
On Mon, 30 Sep 2024, Matthew Auld <matthew.auld@intel.com> wrote: > On 30/09/2024 14:54, Jani Nikula wrote: >> On Fri, 27 Sep 2024, Jani Nikula <jani.nikula@intel.com> wrote: >>> On Thu, 26 Sep 2024, Rodrigo Vivi <rodrigo.vivi@intel.com> wrote: >>>> On Wed, Sep 18, 2024 at 08:35:43PM +0300, Jani Nikula wrote: >>>>> CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND is an int, defaulting to 250. When >>>>> the wakeref is non-zero, it's either -1 or a dynamically allocated >>>>> pointer, depending on CONFIG_DRM_I915_DEBUG_RUNTIME_PM. It's likely that >>>>> the code works by coincidence with the bitwise AND, but with >>>>> CONFIG_DRM_I915_DEBUG_RUNTIME_PM=y, there's the off chance that the >>>>> condition evaluates to false, and intel_wakeref_auto() doesn't get >>>>> called. Switch to the intended logical AND. >>>>> >>>>> v2: Use != to avoid clang -Wconstant-logical-operand (Nathan) >>>> >>>> oh, this is ugly! >>>> >>>> Wouldn't it be better then to use IS_ENABLED() macro? >>> >>> It's an int config option, not a bool. (Yes, the name is misleading.) >>> >>> IS_ENABLED(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) would be the same as >>> CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND == 1. >>> >>> We're actually checking if the int value != 0, so IMO the patch at hand >>> is fine. >> >> Ping, r-b on this one too? > > r-b still holds for v2, fwiw. Thanks for the reviews, pushed the lot to drm-intel-next. BR, Jani. > >> >> BR, >> Jani. >> >>> >>> BR, >>> Jani. >>> >>> >>>> >>>>> >>>>> Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend") >>>>> Cc: Matthew Auld <matthew.auld@intel.com> >>>>> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> >>>>> Cc: Anshuman Gupta <anshuman.gupta@intel.com> >>>>> Cc: Andi Shyti <andi.shyti@linux.intel.com> >>>>> Cc: Nathan Chancellor <nathan@kernel.org> >>>>> Cc: <stable@vger.kernel.org> # v6.1+ >>>>> Reviewed-by: Matthew Auld <matthew.auld@intel.com> # v1 >>>>> Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> # v1 >>>>> Signed-off-by: Jani Nikula <jani.nikula@intel.com> >>>>> --- >>>>> drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +- >>>>> 1 file changed, 1 insertion(+), 1 deletion(-) >>>>> >>>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>>>> index 5c72462d1f57..b22e2019768f 100644 >>>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c >>>>> @@ -1131,7 +1131,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault *vmf) >>>>> GEM_WARN_ON(!i915_ttm_cpu_maps_iomem(bo->resource)); >>>>> } >>>>> >>>>> - if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) >>>>> + if (wakeref && CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND != 0) >>>>> intel_wakeref_auto(&to_i915(obj->base.dev)->runtime_pm.userfault_wakeref, >>>>> msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)); >>>>> >>>>> -- >>>>> 2.39.2 >>>>> >>
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c index 5c72462d1f57..b22e2019768f 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c @@ -1131,7 +1131,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault *vmf) GEM_WARN_ON(!i915_ttm_cpu_maps_iomem(bo->resource)); } - if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND) + if (wakeref && CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND != 0) intel_wakeref_auto(&to_i915(obj->base.dev)->runtime_pm.userfault_wakeref, msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND));
CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND is an int, defaulting to 250. When the wakeref is non-zero, it's either -1 or a dynamically allocated pointer, depending on CONFIG_DRM_I915_DEBUG_RUNTIME_PM. It's likely that the code works by coincidence with the bitwise AND, but with CONFIG_DRM_I915_DEBUG_RUNTIME_PM=y, there's the off chance that the condition evaluates to false, and intel_wakeref_auto() doesn't get called. Switch to the intended logical AND. v2: Use != to avoid clang -Wconstant-logical-operand (Nathan) Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend") Cc: Matthew Auld <matthew.auld@intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Anshuman Gupta <anshuman.gupta@intel.com> Cc: Andi Shyti <andi.shyti@linux.intel.com> Cc: Nathan Chancellor <nathan@kernel.org> Cc: <stable@vger.kernel.org> # v6.1+ Reviewed-by: Matthew Auld <matthew.auld@intel.com> # v1 Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> # v1 Signed-off-by: Jani Nikula <jani.nikula@intel.com> --- drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)