Merge tag 'drm-fixes-2025-07-18-1' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Dave Airlie:
 "Seems like a quiet enough week, xe/amdgpu being the usual suspects,
  then mediatek with a few fixes, and otherwise just misc other bits.

  dp:
   - aux dpcd address fix

  xe:
   - SR-IOV fixes for GT reset and TLB invalidation
   - Fix memory copy direction during migration
   - Fix alignment check on migration
   - Fix MOCS and page fault init order to correctly account
     for topology

  amdgpu:
   - Fix a DC memory leak
   - DCN 4.0.1 degamma LUT fix
   - Fix reset counter handling for soft recovery
   - GC 8 fix

  radeon:
   - Drop console locks when suspending/resuming

  nouveau:
   - ioctl validation fix

  panfrost:
   - scheduler bug fix

  mediatek:
   - Add wait_event_timeout when disabling plane
   - only announce AFBC if really supported
   - mtk_dpi: Reorder output formats on MT8195/88"

* tag 'drm-fixes-2025-07-18-1' of https://gitlab.freedesktop.org/drm/kernel:
  drm/mediatek: mtk_dpi: Reorder output formats on MT8195/88
  drm/mediatek: only announce AFBC if really supported
  drm/mediatek: Add wait_event_timeout when disabling plane
  drm/xe/pf: Resend PF provisioning after GT reset
  drm/xe/pf: Prepare to stop SR-IOV support prior GT reset
  drm/xe/migrate: Fix alignment check
  drm/xe: Move page fault init after topology init
  drm/xe/mocs: Initialize MOCS index early
  drm/xe/migrate: fix copy direction in access_memory
  drm/xe: Dont skip TLB invalidations on VF
  drm/amdgpu/gfx8: reset compute ring wptr on the GPU on resume
  drm/amdgpu: Increase reset counter only on success
  drm/radeon: Do not hold console lock during resume
  drm/radeon: Do not hold console lock while suspending clients
  drm/amd/display: Disable CRTC degamma LUT for DCN401
  drm/amd/display: Free memory allocation
  drm/dp: Change AUX DPCD probe address from LANE0_1_STATUS to TRAINING_PATTERN_SET
  drm/panfrost: Fix scheduler workqueue bug
  drm/nouveau: check ioctl command codes better