[PATCH] amdgpu: fix integer overflow on 32-bit architectures

[PATCH] amdgpu: fix integer overflow on 32-bit architectures

[Arnd Bergmann]

Multiplying 1000000000 by four overruns a 'long' variable, as clang
points out:

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4160:53: error: overflow in expression; result is -294967296 with type 'long' [-Werror,-Winteger-overflow]
                expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4L;
                                                                  ^
Make this a 'long long' constant instead.

Fixes: 3f12acc8d6d4 ("drm/amdgpu: put the audio codec into suspend state before gpu reset V3")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
I'm not sure the ktime_get_mono_fast_ns() call is necessary here
either. Is it intentional because ktime_get_ns() doesn't work
during a driver suspend, or just a mistake?
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 6f93af972b0a..2e07e3e6b036 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4157,7 +4157,7 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
 		 * the audio controller default autosuspend delay setting.
 		 * 4S used here is guaranteed to cover that.
 		 */
-		expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4L;
+		expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4LL;
 
 	while (!pm_runtime_status_suspended(&(p->dev))) {
 		if (!pm_runtime_suspend(&(p->dev)))
-- 
2.26.0

Re: [PATCH] amdgpu: fix integer overflow on 32-bit architectures

[Christian König]

Am 05.05.20 um 16:15 schrieb Arnd Bergmann:
> Multiplying 1000000000 by four overruns a 'long' variable, as clang
> points out:
>
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4160:53: error: overflow in expression; result is -294967296 with type 'long' [-Werror,-Winteger-overflow]
>                  expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4L;
>                                                                    ^
> Make this a 'long long' constant instead.
>
> Fixes: 3f12acc8d6d4 ("drm/amdgpu: put the audio codec into suspend state before gpu reset V3")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Reviewed-by: Christian König <christian.koenig@amd.com>

> ---
> I'm not sure the ktime_get_mono_fast_ns() call is necessary here
> either. Is it intentional because ktime_get_ns() doesn't work
> during a driver suspend, or just a mistake?
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 6f93af972b0a..2e07e3e6b036 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -4157,7 +4157,7 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
>   		 * the audio controller default autosuspend delay setting.
>   		 * 4S used here is guaranteed to cover that.
>   		 */
> -		expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4L;
> +		expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4LL;
>   
>   	while (!pm_runtime_status_suspended(&(p->dev))) {
>   		if (!pm_runtime_suspend(&(p->dev)))

Re: [PATCH] amdgpu: fix integer overflow on 32-bit architectures

[Nick Desaulniers]

See also https://lore.kernel.org/lkml/CADnq5_NdtZh5_RGDWKJ9c_42XLvrncCs5DDU1YSptfZP94KXkQ@mail.gmail.com/T/#me707e09e92c6e487285e8bb382a607e4e782c249

On Tue, May 5, 2020 at 7:17 AM Christian König <christian.koenig@amd.com> wrote:
>
> Am 05.05.20 um 16:15 schrieb Arnd Bergmann:
> > Multiplying 1000000000 by four overruns a 'long' variable, as clang
> > points out:
> >
> > drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4160:53: error: overflow in expression; result is -294967296 with type 'long' [-Werror,-Winteger-overflow]
> >                  expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4L;
> >                                                                    ^
> > Make this a 'long long' constant instead.
> >
> > Fixes: 3f12acc8d6d4 ("drm/amdgpu: put the audio codec into suspend state before gpu reset V3")
> > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
>
> Reviewed-by: Christian König <christian.koenig@amd.com>
>
> > ---
> > I'm not sure the ktime_get_mono_fast_ns() call is necessary here
> > either. Is it intentional because ktime_get_ns() doesn't work
> > during a driver suspend, or just a mistake?
> > ---
> >   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> > index 6f93af972b0a..2e07e3e6b036 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> > @@ -4157,7 +4157,7 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
> >                * the audio controller default autosuspend delay setting.
> >                * 4S used here is guaranteed to cover that.
> >                */
> > -             expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4L;
> > +             expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4LL;
> >
> >       while (!pm_runtime_status_suspended(&(p->dev))) {
> >               if (!pm_runtime_suspend(&(p->dev)))
>
> --
> You received this message because you are subscribed to the Google Groups "Clang Built Linux" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to clang-built-linux+unsubscribe@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/clang-built-linux/e4a852b2-807b-bc73-7328-bcc399341085%40amd.com.



-- 
Thanks,
~Nick Desaulniers