drm-intel-fixes
Hi Dave & Sima,
Here's still a drm-intel-fixes PR towards 7.0. Took a few more hours as Linus seems to prefer the EXECBUF fix via -fixes rather than -next.
One display fix for SU area WARN in PSR code, for GT refcount underflow fix in hangcheck and removal condition on execbuf BO processing which gives a false alert of a bug in mainline.
Regards, Joonas
***
drm-intel-fixes-2026-04-09:
- Drop check for changed VM in EXECBUF
- Fix refcount underflow race in intel_engine_park_heartbeat
- Do not use pipe_src as borders for SU area in PSR
The following changes since commit 591cd656a1bf5ea94a222af5ef2ee76df029c1d2:
Linux 7.0-rc7 (2026-04-05 15:26:23 -0700)
are available in the Git repository at:
https://gitlab.freedesktop.org/drm/i915/kernel.git tags/drm-intel-fixes-2026-04-09
for you to fetch changes up to a13edf9b92fc4700b3020d7ea547a3d64dd33b63:
drm/i915/gem: Drop check for changed VM in EXECBUF (2026-04-09 12:57:15 +0300)
----------------------------------------------------------------
- Drop check for changed VM in EXECBUF
- Fix refcount underflow race in intel_engine_park_heartbeat
- Do not use pipe_src as borders for SU area in PSR
----------------------------------------------------------------
Joonas Lahtinen (1):
drm/i915/gem: Drop check for changed VM in EXECBUF
Jouni Högander (1):
drm/i915/psr: Do not use pipe_src as borders for SU area
Sebastian Brzezinka (1):
drm/i915/gt: fix refcount underflow in intel_engine_park_heartbeat
drivers/gpu/drm/i915/display/intel_psr.c | 30 +++++++++++++++--------- drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 2 +- drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c | 26 +++++++++++++------- 3 files changed, 38 insertions(+), 20 deletions(-)
Quoting Joonas Lahtinen (2026-04-09 13:07:56)
Quoted 8 lines
Hi Dave & Sima,
Here's still a drm-intel-fixes PR towards 7.0. Took a few more hours as Linus seems to prefer the EXECBUF fix via -fixes rather than -next.
One display fix for SU area WARN in PSR code, for GT refcount underflow fix in hangcheck and removal condition on execbuf BO processing which gives a false alert of a bug in mainline.
Also, if you haven't been following all the discussion, the execbuf cleanup here removes the need for the fix that Linus applied himself to v7.0-rc7. This just removes the condition check that is always true, avoiding false alarm of a bug in current mainline.
Regards, Joonas