On 07.09.2014 12:06, Christian König wrote:
> From: Christian König <christian.koenig@xxxxxxx>
>
> Semaphore values have 64 bits, not 32. This fixes a very subtle bug
> that disables synchronization when the upper 32bits wasn't zero.
>
So essentially, half the semaphore values were never properly
initialized and some loads with a lot of semaphore synchronization going
on tried to use the uninitialized semaphores. I think the description in
the commit could be improved according to that.
I didn't get any DMA L2T copy hangs with semaphores disabled completely,
but unfortunately this doesn't fix the hangs.
Anyway,
Reviewed-By: Grigori Goronzy <greg@xxxxxxxxxxxx>
Grigori
> Signed-off-by: Christian König <christian.koenig@xxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx
> ---
> drivers/gpu/drm/radeon/radeon_semaphore.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c
> index 56d9fd6..abd6753 100644
> --- a/drivers/gpu/drm/radeon/radeon_semaphore.c
> +++ b/drivers/gpu/drm/radeon/radeon_semaphore.c
> @@ -34,7 +34,7 @@
> int radeon_semaphore_create(struct radeon_device *rdev,
> struct radeon_semaphore **semaphore)
> {
> - uint32_t *cpu_addr;
> + uint64_t *cpu_addr;
> int i, r;
>
> *semaphore = kmalloc(sizeof(struct radeon_semaphore), GFP_KERNEL);
>
Attachment:
signature.asc
Description: OpenPGP digital signature
_______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel
