[Intel-gfx] [PATCH] drm/i915: Drain the freed state from the tail of the next commit

[Maarten Lankhorst]

Op 02-02-17 om 21:47 schreef Chris Wilson:
> If we have any residual freed atomic state from earlier commits, flush
> the freed list after performing the current modeset. This prevents the
> freed list from ever-growing if userspace manages to starve the kernel
> threads (i.e. we are never able to run our free state worker and
> eventually the system may even oom).
>
> Fixes: eb955eee27d9 ("drm/i915: Move atomic state free from out of fence release")
> Testcase: igt/kms_cursor/legacy/all-pipes-single-bo
kms_cursor_legacy
> Reported-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
> Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
> Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
> ---
>
> Though the earlier patch is cc'ed for v4.10, I don't this fix to the fix
> merits applying to v4.10 this late. The impact should be small, it takes
> the deliberate abuse of RT tasks to starve the kernel so badly as that
> the delayed free of the atomic state will cause system-wide consequences
> (e.g. oom)
Fixes the kms_cursor_legacy testcase, and the patch makes sense.

Reviewed-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>