| Message ID | 20211108123637.929617-1-thomas.hellstrom@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] drm/i915/gem: Fix gem_madvise for ttm+shmem objects | expand |
On 08/11/2021 12:36, Thomas Hellström wrote: > Gem-TTM objects that are backed by shmem might have populated > page-vectors without having the GEM pages set. Those objects > aren't moved to the correct shrinker / purge list by gem_madvise. > > For such objects, identified by having the > _SELF_MANAGED_SHRINK_LIST set, make sure they end up on the > correct list. > > v2: > - Revert a change that made swapped-out objects inaccessible for > truncating. (Matthew Auld) > > Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com> > --- > drivers/gpu/drm/i915/i915_gem.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > index d0e642c82064..527228d4da7e 100644 > --- a/drivers/gpu/drm/i915/i915_gem.c > +++ b/drivers/gpu/drm/i915/i915_gem.c > @@ -1005,7 +1005,8 @@ i915_gem_madvise_ioctl(struct drm_device *dev, void *data, > obj->ops->adjust_lru(obj); > } > > - if (i915_gem_object_has_pages(obj)) { > + if (i915_gem_object_has_pages(obj) || > + i915_gem_object_has_self_managed_shrink_list(obj)) { > unsigned long flags; > > spin_lock_irqsave(&i915->mm.obj_lock, flags); >
On 11/8/21 15:44, Patchwork wrote: > Project List - Patchwork *Patch Details* > *Series:* drm/i915/gem: Fix gem_madvise for ttm+shmem objects (rev2) > *URL:* https://patchwork.freedesktop.org/series/96611/ > *State:* failure > *Details:* > https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21533/index.html > > > CI Bug Log - changes from CI_DRM_10853_full -> Patchwork_21533_full > > > Summary > > *FAILURE* > > Serious unknown changes coming with Patchwork_21533_full absolutely > need to be > verified manually. > > If you think the reported changes have nothing to do with the changes > introduced in Patchwork_21533_full, please notify your bug team to > allow them > to document this new failure mode, which will reduce false positives > in CI. > > > Participating hosts (11 -> 11) > > No changes in participating hosts > > > Possible new issues > > Here are the unknown changes that may have been introduced in > Patchwork_21533_full: > > > IGT changes > > > Possible regressions > > * igt@kms_flip@flip-vs-suspend-interruptible@c-hdmi-a2: > o shard-glk: PASS > <https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10853/shard-glk1/igt@kms_flip@flip-vs-suspend-interruptible@c-hdmi-a2.html> > -> INCOMPLETE > <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21533/shard-glk6/igt@kms_flip@flip-vs-suspend-interruptible@c-hdmi-a2.html> > Lakshmi, This failure is unrelated. Thanks, Thomas
Filed new issue and re-reported.
https://gitlab.freedesktop.org/drm/intel/-/issues/4456
igt@kms_flip@flip-vs-suspend-interruptible@c-hdmi-a2 - incomplete - PM: suspend entry (deep)
Lakshmi.
From: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Sent: Monday, November 8, 2021 6:48 AM
To: intel-gfx@lists.freedesktop.org; Vudum, Lakshminarayana <lakshminarayana.vudum@intel.com>
Subject: Re: ✗ Fi.CI.IGT: failure for drm/i915/gem: Fix gem_madvise for ttm+shmem objects (rev2)
On 11/8/21 15:44, Patchwork wrote:
Patch Details
Series:
drm/i915/gem: Fix gem_madvise for ttm+shmem objects (rev2)
URL:
https://patchwork.freedesktop.org/series/96611/
State:
failure
Details:
https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21533/index.html
CI Bug Log - changes from CI_DRM_10853_full -> Patchwork_21533_full
Summary
FAILURE
Serious unknown changes coming with Patchwork_21533_full absolutely need to be
verified manually.
If you think the reported changes have nothing to do with the changes
introduced in Patchwork_21533_full, please notify your bug team to allow them
to document this new failure mode, which will reduce false positives in CI.
Participating hosts (11 -> 11)
No changes in participating hosts
Possible new issues
Here are the unknown changes that may have been introduced in Patchwork_21533_full:
IGT changes
Possible regressions
* igt@kms_flip@flip-vs-suspend-interruptible@c-hdmi-a2:
* shard-glk: PASS<https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10853/shard-glk1/igt@kms_flip@flip-vs-suspend-interruptible@c-hdmi-a2.html> -> INCOMPLETE<https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21533/shard-glk6/igt@kms_flip@flip-vs-suspend-interruptible@c-hdmi-a2.html>
Lakshmi, This failure is unrelated.
Thanks,
Thomas
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index d0e642c82064..527228d4da7e 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1005,7 +1005,8 @@ i915_gem_madvise_ioctl(struct drm_device *dev, void *data, obj->ops->adjust_lru(obj); } - if (i915_gem_object_has_pages(obj)) { + if (i915_gem_object_has_pages(obj) || + i915_gem_object_has_self_managed_shrink_list(obj)) { unsigned long flags; spin_lock_irqsave(&i915->mm.obj_lock, flags);
Gem-TTM objects that are backed by shmem might have populated page-vectors without having the GEM pages set. Those objects aren't moved to the correct shrinker / purge list by gem_madvise. For such objects, identified by having the _SELF_MANAGED_SHRINK_LIST set, make sure they end up on the correct list. v2: - Revert a change that made swapped-out objects inaccessible for truncating. (Matthew Auld) Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> --- drivers/gpu/drm/i915/i915_gem.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)