This reverts commit d6e020819612a4a06207af858e0978be4d3e3140.
The IS_DGFX check was put in place because error capture of buffer objects is expected to be broken on devices with VRAM.
We seem to have already submitted the userspace fix to remove that flag, so lets just rely on that for DG1.
Cc: stable@vger.kernel.org # v6.0+ Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Andi Shyti andi.shyti@linux.intel.com Cc: Matthew Auld matthew.auld@intel.com Cc: Thomas Hellström thomas.hellstrom@linux.intel.com Cc: Tvrtko Ursulin tursulin@ursulin.net Signed-off-by: Joonas Lahtinen joonas.lahtinen@linux.intel.com --- drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c index 7d44aadcd5a5..02c59808cbe4 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c @@ -2013,7 +2013,7 @@ static int eb_capture_stage(struct i915_execbuffer *eb) continue;
if (i915_gem_context_is_recoverable(eb->gem_context) && - GRAPHICS_VER_FULL(eb->i915) > IP_VER(12, 10)) + (IS_DGFX(eb->i915) || GRAPHICS_VER_FULL(eb->i915) > IP_VER(12, 0))) return -EINVAL;
for_each_batch_create_order(eb, j) {
On 22/05/2025 07:41, Joonas Lahtinen wrote:
This reverts commit d6e020819612a4a06207af858e0978be4d3e3140.
The IS_DGFX check was put in place because error capture of buffer objects is expected to be broken on devices with VRAM.
I don't quite remember the history and it is a pity 71b1669ea9bd ("drm/i915/uapi: tweak error capture on recoverable contexts") did not spell that out but almost made it sound like uapi disablement for some reason. But if there is an userspace fix, and you say VRAM capture is broken then that's fine by me.
Acked-by: Tvrtko Ursulin tvrtko.ursulin@igalia.com
Regards,
Tvrtko
We seem to have already submitted the userspace fix to remove that flag, so lets just rely on that for DG1.
Cc: stable@vger.kernel.org # v6.0+ Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Andi Shyti andi.shyti@linux.intel.com Cc: Matthew Auld matthew.auld@intel.com Cc: Thomas Hellström thomas.hellstrom@linux.intel.com Cc: Tvrtko Ursulin tursulin@ursulin.net Signed-off-by: Joonas Lahtinen joonas.lahtinen@linux.intel.com
drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c index 7d44aadcd5a5..02c59808cbe4 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c @@ -2013,7 +2013,7 @@ static int eb_capture_stage(struct i915_execbuffer *eb) continue; if (i915_gem_context_is_recoverable(eb->gem_context) &&
GRAPHICS_VER_FULL(eb->i915) > IP_VER(12, 10))
(IS_DGFX(eb->i915) || GRAPHICS_VER_FULL(eb->i915) > IP_VER(12, 0))) return -EINVAL;for_each_batch_create_order(eb, j) {
Quoting Joonas Lahtinen (2025-05-22 09:41:27)
This reverts commit d6e020819612a4a06207af858e0978be4d3e3140.
The IS_DGFX check was put in place because error capture of buffer objects is expected to be broken on devices with VRAM.
We seem to have already submitted the userspace fix to remove that flag, so lets just rely on that for DG1.
Further, it seems that the userspace fix[1] has already been pushed and released as part of media-driver 25.2.3, so really no reason to unblock the broken codepath.
[1] https://github.com/intel/media-driver/commit/93c07d9b4b96a78bab21f6acd4eb863...
Regards, Joonas
Cc: stable@vger.kernel.org # v6.0+ Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Andi Shyti andi.shyti@linux.intel.com Cc: Matthew Auld matthew.auld@intel.com Cc: Thomas Hellström thomas.hellstrom@linux.intel.com Cc: Tvrtko Ursulin tursulin@ursulin.net Signed-off-by: Joonas Lahtinen joonas.lahtinen@linux.intel.com
drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c index 7d44aadcd5a5..02c59808cbe4 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c @@ -2013,7 +2013,7 @@ static int eb_capture_stage(struct i915_execbuffer *eb) continue; if (i915_gem_context_is_recoverable(eb->gem_context) &&
GRAPHICS_VER_FULL(eb->i915) > IP_VER(12, 10))
(IS_DGFX(eb->i915) || GRAPHICS_VER_FULL(eb->i915) > IP_VER(12, 0))) return -EINVAL;for_each_batch_create_order(eb, j) { -- 2.49.0
Hi Joonas,
On Thu, May 22, 2025 at 09:41:27AM +0300, Joonas Lahtinen wrote:
This reverts commit d6e020819612a4a06207af858e0978be4d3e3140.
The IS_DGFX check was put in place because error capture of buffer objects is expected to be broken on devices with VRAM.
We seem to have already submitted the userspace fix to remove that flag, so lets just rely on that for DG1.
yes, sorry, I should have checked git bisect before accepting the patch. I still think that a comment would help. In any case:
Reviewed-by: Andi Shyti andi.shyti@linux.intel.com
Andi
Applied this as there were no further objections now that the userspace fix is merged.
Thanks for the reviews.
Regards, Joonas
linux-stable-mirror@lists.linaro.org
-
Andi Shyti -
Joonas Lahtinen -
Tvrtko Ursulin