July 2020 Archives by thread

• Messages sorted by: [ subject ] [ author ] [ date ]
• More info on this list...

Starting: Wed Jul 1 00:57:23 UTC 2020
Ending: Fri Jul 31 23:41:24 UTC 2020
Messages: 6015

• ✗ Fi.CI.BAT: failure for series starting with [1/4] drm/i915/gem: Avoid kmalloc under i915->mm_lock   Patchwork
  • ✗ Fi.CI.BAT: failure for series starting with [1/4] drm/i915/gem: Avoid kmalloc under i915->mm_lock   Patchwork
• ✗ Fi.CI.BAT: failure for series starting with [1/9] drm/i915/gem: Move obj->lut_list under its own lock   Patchwork
• ✗ Fi.CI.BAT: failure for drm/i915: Resolve device hotunplug issues (rev11)   Patchwork
• ✓ Fi.CI.BAT: success for drm/i915: Clamp min_cdclk to max_cdclk_freq to unblock 8K   Patchwork
• ✗ Fi.CI.IGT: failure for mm: Skip opportunistic reclaim for dma mapped pages   Patchwork
• ✓ Fi.CI.IGT: success for drm/i915: Clamp min_cdclk to max_cdclk_freq to unblock 8K   Patchwork
• [PATCH v2 0/8] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
  • [PATCH v2 1/8] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
  • [PATCH v2 2/8] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
  • [PATCH 3/8] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
  • [PATCH 4/8] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
  • [PATCH v2 5/8] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
  • [PATCH v2 6/8] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
  • [PATCH v2 7/8] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
  • [PATCH 8/8] drm/i915: Release memory regions on driver remove   Janusz Krzysztofik
  • ✗ Fi.CI.BAT: failure for drm/i915: Resolve device hotunplug issues (rev12)   Patchwork
  • [PATCH v2 0/8] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
    • [PATCH v2 1/8] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
    • [PATCH v2 2/8] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
    • [PATCH 3/8] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH 4/8] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 5/8] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
    • [PATCH v2 6/8] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
    • [PATCH v2 7/8] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
    • [PATCH 8/8] drm/i915: Resolve hotunplug related live selftest issues   Janusz Krzysztofik
    • ✗ Fi.CI.BUILD: failure for drm/i915: Resolve device hotunplug issues (rev13)   Patchwork
  • [PATCH v2 0/8] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
    • [PATCH 1/8] drm/i915: Refactor postclose driver operation   Janusz Krzysztofik
    • [PATCH v2 2/8] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
    • [PATCH v2 3/8] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
    • [PATCH 4/8] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH 5/8] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 6/8] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
    • [PATCH v2 7/8] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
    • [PATCH v2 8/8] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
    • ✓ Fi.CI.BAT: success for drm/i915: Resolve device hotunplug issues (rev17)   Patchwork
    • ✓ Fi.CI.IGT: success for drm/i915: Resolve device hotunplug issues (rev17)   Patchwork
  • [PATCH v2 0/8] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
    • [PATCH v2 1/8] drm/i915: Refactor postclose driver operation   Janusz Krzysztofik
    • [PATCH v2 2/8] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
    • [PATCH v2 3/8] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
    • [PATCH v2 4/8] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 5/8] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 6/8] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
    • [PATCH v2 7/8] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
    • [PATCH v2 8/8] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
    • ✗ Fi.CI.BAT: failure for drm/i915: Resolve device hotunplug issues (rev18)   Patchwork
• [PATCH] drm/i915/lmem: reset the lmem buffer created by fbdev   Animesh Manna
  • ✓ Fi.CI.BAT: success for drm/i915/lmem: reset the lmem buffer created by fbdev   Patchwork
  • ✓ Fi.CI.IGT: success for drm/i915/lmem: reset the lmem buffer created by fbdev   Patchwork
• [PATCH 01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Chris Wilson
  • [PATCH 02/22] drm/i915/gem: Drop forced struct_mutex from shrinker_taints_mutex   Chris Wilson
  • [PATCH 03/22] drm/i915/gem: Only revoke mmap handlers if active   Chris Wilson
  • [PATCH 04/22] drm/i915: Export ppgtt_bind_vma   Chris Wilson
  • [PATCH 05/22] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 06/22] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 07/22] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 08/22] unbound-link   Chris Wilson
  • [PATCH 09/22] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 10/22] inplace   Chris Wilson
  • [PATCH 11/22] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 12/22] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 13/22] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 14/22] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 15/22] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 16/22] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 17/22] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 18/22] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 19/22] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 20/22] copy-inplace relocs   Chris Wilson
  • [PATCH 21/22] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 22/22] eb-acquire   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Patchwork
  • [PATCH 01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Chris Wilson
    • [PATCH 02/22] drm/i915/gem: Drop forced struct_mutex from shrinker_taints_mutex   Chris Wilson
    • [PATCH 03/22] drm/i915/gem: Only revoke mmap handlers if active   Chris Wilson
    • [PATCH 04/22] drm/i915: Export ppgtt_bind_vma   Chris Wilson
    • [PATCH 05/22] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 06/22] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 07/22] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 08/22] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 09/22] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 10/22] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 11/22] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 12/22] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 13/22] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 14/22] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 15/22] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 16/22] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 17/22] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 18/22] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 19/22] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 20/22] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 21/22] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 22/22] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Patchwork
  • [PATCH 01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Chris Wilson
    • [PATCH 02/22] drm/i915/gem: Drop forced struct_mutex from shrinker_taints_mutex   Chris Wilson
    • [PATCH 03/22] drm/i915/gem: Only revoke mmap handlers if active   Chris Wilson
    • [PATCH 04/22] drm/i915: Export ppgtt_bind_vma   Chris Wilson
    • [PATCH 05/22] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 06/22] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 07/22] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 08/22] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 09/22] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 10/22] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 11/22] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 12/22] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 13/22] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 14/22] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 15/22] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 16/22] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 17/22] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 18/22] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 19/22] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 20/22] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 21/22] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 22/22] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Patchwork
  • [PATCH 01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Chris Wilson
    • [PATCH 02/22] drm/i915/gem: Drop forced struct_mutex from shrinker_taints_mutex   Chris Wilson
    • [PATCH 03/22] drm/i915/gem: Only revoke mmap handlers if active   Chris Wilson
    • [PATCH 04/22] drm/i915: Export ppgtt_bind_vma   Chris Wilson
    • [PATCH 05/22] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 06/22] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 07/22] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 08/22] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 09/22] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 10/22] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 11/22] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 12/22] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 13/22] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 14/22] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 15/22] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 16/22] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 17/22] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 18/22] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 19/22] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 20/22] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 21/22] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 22/22] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Patchwork
• [PATCH 01/23] drm/i915: Drop vm.ref for duplicate vma on construction   Chris Wilson
  • [PATCH 02/23] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Chris Wilson
    • [PATCH 02/23] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Chris Wilson
  • [PATCH 03/23] drm/i915/gem: Drop forced struct_mutex from shrinker_taints_mutex   Chris Wilson
  • [PATCH 04/23] drm/i915/gem: Only revoke mmap handlers if active   Chris Wilson
  • [PATCH 05/23] drm/i915: Export ppgtt_bind_vma   Chris Wilson
  • [PATCH 06/23] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 07/23] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 08/23] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 09/23] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 10/23] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 11/23] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 12/23] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 13/23] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 14/23] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 15/23] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 16/23] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 17/23] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 18/23] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 19/23] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 20/23] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 21/23] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 22/23] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 23/23] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/23] drm/i915: Drop vm.ref for duplicate vma on construction   Patchwork
  • [PATCH 01/23] drm/i915: Drop vm.ref for duplicate vma on construction   Chris Wilson
    • [PATCH 03/23] drm/i915/gem: Drop forced struct_mutex from shrinker_taints_mutex   Chris Wilson
    • [PATCH 04/23] drm/i915/gem: Only revoke mmap handlers if active   Chris Wilson
    • [PATCH 05/23] drm/i915: Export ppgtt_bind_vma   Chris Wilson
    • [PATCH 06/23] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 07/23] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 08/23] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 09/23] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 10/23] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 11/23] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 12/23] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 13/23] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 14/23] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 15/23] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 16/23] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 17/23] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 18/23] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 19/23] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 20/23] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 21/23] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 22/23] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 23/23] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/23] drm/i915: Drop vm.ref for duplicate vma on construction   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/23] drm/i915: Drop vm.ref for duplicate vma on construction   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/23] drm/i915: Drop vm.ref for duplicate vma on construction   Patchwork
• [PATCH v2 0/7] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
  • [PATCH v2 1/7] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
  • [PATCH v2 2/7] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
  • [PATCH 3/7] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
  • [PATCH 4/7] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
  • [PATCH v2 5/7] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
  • [PATCH v2 6/7] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
  • [PATCH v2 7/7] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
  • ✓ Fi.CI.BAT: success for drm/i915: Resolve device hotunplug issues (rev14)   Patchwork
  • ✓ Fi.CI.IGT: success for drm/i915: Resolve device hotunplug issues (rev14)   Patchwork
  • [PATCH v2 0/7] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
    • [PATCH v2 1/7] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
      • [PATCH v2 1/7] drm/i915: Drop user contexts on driver remove   Chris Wilson
      • [PATCH v2 1/7] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
    • [PATCH v2 2/7] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
    • [PATCH 3/7] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH 4/7] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 5/7] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
    • [PATCH v2 6/7] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
    • [PATCH v2 7/7] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
    • ✗ Fi.CI.BAT: failure for drm/i915: Resolve device hotunplug issues (rev15)   Patchwork
  • [PATCH v2 0/7] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
    • [PATCH v2 1/7] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
    • [PATCH v2 2/7] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
    • [PATCH 3/7] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH 4/7] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 5/7] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
    • [PATCH v2 6/7] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
    • [PATCH v2 7/7] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
    • ✓ Fi.CI.BAT: success for drm/i915: Resolve device hotunplug issues (rev16)   Patchwork
    • ✓ Fi.CI.IGT: success for drm/i915: Resolve device hotunplug issues (rev16)   Patchwork
• [PATCH 01/24] drm/i915: Kill context before taking ctx->mutex   Maarten Lankhorst
  • [PATCH 02/24] Revert "drm/i915/gem: Async GPU relocations only"   Maarten Lankhorst
  • [PATCH 03/24] drm/i915: Revert relocation chaining commits.   Maarten Lankhorst
  • [PATCH 04/24] Revert "drm/i915/gem: Drop relocation slowpath".   Maarten Lankhorst
  • [PATCH 05/24] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Maarten Lankhorst
  • [PATCH 06/24] drm/i915: Remove locking from i915_gem_object_prepare_read/write   Maarten Lankhorst
  • [PATCH 07/24] drm/i915: Parse command buffer earlier in eb_relocate(slow)   Maarten Lankhorst
  • [PATCH 08/24] Revert "drm/i915/gem: Split eb_vma into its own allocation"   Maarten Lankhorst
  • [PATCH 09/24] drm/i915: Use per object locking in execbuf, v12.   Maarten Lankhorst
  • [PATCH 10/24] drm/i915: Use ww locking in intel_renderstate.   Maarten Lankhorst
  • [PATCH 11/24] drm/i915: Add ww context handling to context_barrier_task   Maarten Lankhorst
  • [PATCH 12/24] drm/i915: Nuke arguments to eb_pin_engine   Maarten Lankhorst
  • [PATCH 13/24] drm/i915: Pin engine before pinning all objects, v4.   Maarten Lankhorst
  • [PATCH 14/24] drm/i915: Rework intel_context pinning to do everything outside of pin_mutex   Maarten Lankhorst
  • [PATCH 15/24] drm/i915: Make sure execbuffer always passes ww state to i915_vma_pin.   Maarten Lankhorst
  • [PATCH 16/24] drm/i915: Convert i915_gem_object/client_blt.c to use ww locking as well, v2.   Maarten Lankhorst
  • [PATCH 17/24] drm/i915: Kill last user of intel_context_create_request outside of selftests   Maarten Lankhorst
  • [PATCH 18/24] drm/i915: Convert i915_perf to ww locking as well   Maarten Lankhorst
  • [PATCH 19/24] drm/i915: Dirty hack to fix selftests locking inversion   Maarten Lankhorst
  • [PATCH 20/24] drm/i915/selftests: Fix locking inversion in lrc selftest.   Maarten Lankhorst
  • [PATCH 21/24] drm/i915: Use ww pinning for intel_context_create_request()   Maarten Lankhorst
  • [PATCH 22/24] drm/i915: Move i915_vma_lock in the selftests to avoid lock inversion, v2.   Maarten Lankhorst
  • [PATCH 23/24] drm/i915: Add ww locking to vm_fault_gtt   Maarten Lankhorst
  • [PATCH 24/24] drm/i915: Add ww locking to pin_to_display_plane   Maarten Lankhorst
  • ✗ Fi.CI.BAT: failure for series starting with [01/24] drm/i915: Kill context before taking ctx->mutex   Patchwork
• [PATCH 01/20] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Chris Wilson
  • [PATCH 02/20] drm/i915: Export ppgtt_bind_vma   Chris Wilson
  • [PATCH 03/20] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 04/20] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 05/20] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 06/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 07/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 08/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 09/20] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 10/20] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 11/20] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 12/20] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 13/20] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 14/20] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 15/20] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 16/20] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 17/20] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 18/20] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 19/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 20/20] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/20] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/20] drm/i915/gem: Split the context's obj:vma lut into its own mutex   Patchwork
• [PATCH i-g-t 1/2] tests/device_reset: Test device sysfs reset   Marcin Bernatowicz
  • [PATCH i-g-t 2/2] Add device_reset subtests to fast feedback list for getting results promptly   Marcin Bernatowicz
  • ✗ Fi.CI.BAT: failure for series starting with [1/2] tests/device_reset: Test device sysfs reset   Patchwork
  • [PATCH i-g-t 1/2] tests/device_reset: Test device sysfs reset   Marcin Bernatowicz
    • [PATCH i-g-t 2/2] Add device_reset subtests to fast feedback list for getting results promptly   Marcin Bernatowicz
    • ✗ Fi.CI.BAT: failure for series starting with [1/2] tests/device_reset: Test device sysfs reset   Patchwork
  • [PATCH i-g-t 1/2] tests/device_reset: Test device sysfs reset   Marcin Bernatowicz
    • [PATCH i-g-t 2/2] Add device_reset subtests to fast feedback list for getting results promptly   Marcin Bernatowicz
    • ✗ Fi.CI.BAT: failure for series starting with [1/2] tests/device_reset: Test device sysfs reset   Patchwork
• [PATCH 01/19] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/19] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/19] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 05/19] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/19] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/19] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/19] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/19] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/19] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/19] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/19] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/19] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/19] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/19] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/19] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/19] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/19] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 19/19] snoop-batch   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/19] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 04/19] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
• [PATCH 01/56] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/56] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 04/56] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 05/56] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/56] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/56] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/56] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/56] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/56] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/56] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/56] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/56] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/56] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/56] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/56] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/56] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/56] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 19/56] snoop-batch   Chris Wilson
  • [PATCH 20/56] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 21/56] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 22/56] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 23/56] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 24/56] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 25/56] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 26/56] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 27/56] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 28/56] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 29/56] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 30/56] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 31/56] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 32/56] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 33/56] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 34/56] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 35/56] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 36/56] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 37/56] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 38/56] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 39/56] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 40/56] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 41/56] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 42/56] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 43/56] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 44/56] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 45/56] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 46/56] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 47/56] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 48/56] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 49/56] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 50/56] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 51/56] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 52/56] vmscan   Chris Wilson
  • [PATCH 53/56] drm/i915/gem: Always do pin_user_pages under the mmu-notifier   Chris Wilson
  • [PATCH 54/56] ww-mutex-nested   Chris Wilson
  • [PATCH 55/56] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 56/56] nested-shrinker-ww_mutex   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/56] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 03/56] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
• [PATCH 01/24] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/24] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/24] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 03/24] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 04/24] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 05/24] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/24] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/24] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/24] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/24] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/24] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/24] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/24] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/24] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/24] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/24] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/24] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/24] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/24] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 19/24] snoop-batch   Chris Wilson
    • [PATCH 19/24] snoop-batch   Chris Wilson
  • [PATCH 20/24] vmscan   Chris Wilson
  • [PATCH 21/24] ww-mutex-nested   Chris Wilson
  • [PATCH 22/24] drm/i915/gem: Always do pin_user_pages under the mmu-notifier   Chris Wilson
  • [PATCH 23/24] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 24/24] nested-shrinker-ww_mutex   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/24] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/24] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/24] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 04/24] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/24] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/24] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/24] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/24] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/24] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/24] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/24] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/24] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/24] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/24] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/24] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/24] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/24] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/24] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 20/24] vmscan   Chris Wilson
    • [PATCH 21/24] ww-mutex-nested   Chris Wilson
    • [PATCH 22/24] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
    • [PATCH 23/24] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 24/24] nested-shrinker-ww_mutex   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/24] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/16] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/16] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/16] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 03/16] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 03/16] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 04/16] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 05/16] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 07/16] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/16] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/16] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/16] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/16] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/16] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/16] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/16] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/16] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/16] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/16] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/16] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/16] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/16] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/16] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/16] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/16] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/16] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/16] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/16] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/16] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/16] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/16] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/16] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/16] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/16] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/16] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/16] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/16] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/16] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/16] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/16] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/16] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/16] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/16] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/16] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/16] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/16] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/16] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/16] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/16] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/16] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/16] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/16] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/16] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/16] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/16] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/16] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 04/16] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/16] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/16] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/16] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/16] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/16] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/16] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/16] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/16] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/16] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/16] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/16] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/16] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/16] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/16] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/16] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 04/16] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/16] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/16] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/16] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/16] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/16] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/16] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/16] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/16] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/16] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/16] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/16] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/16] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/16] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/16] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 06/16] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
• [PATCH 01/18] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/18] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/18] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 03/18] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 03/18] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 04/18] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 05/18] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/18] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/18] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/18] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/18] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/18] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/18] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/18] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/18] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/18] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/18] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/18] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/18] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/18] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/18] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/18] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/18] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 04/18] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/18] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/18] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/18] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/18] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/18] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/18] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/18] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/18] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/18] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/18] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/18] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/18] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/18] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/18] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/18] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/18] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/18] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/18] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 04/18] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/18] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/18] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/18] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/18] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/18] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/18] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/18] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/18] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/18] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/18] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/18] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/18] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/18] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/18] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/18] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/23] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/23] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/23] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 04/23] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 05/23] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/23] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/23] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/23] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/23] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/23] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/23] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/23] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/23] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/23] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/23] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/23] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/23] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/23] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 19/23] snoop-batch   Chris Wilson
  • [PATCH 20/23] vmscan   Chris Wilson
  • [PATCH 21/23] ww-mutex-nested   Chris Wilson
  • [PATCH 22/23] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
  • [PATCH 23/23] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/23] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/23] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/23] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/23] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/23] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/23] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/23] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/23] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/23] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/23] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/23] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/23] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/23] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/23] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/23] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/23] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/23] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/23] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/23] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 19/23] snoop-batch   Chris Wilson
    • [PATCH 20/23] vmscan   Chris Wilson
    • [PATCH 21/23] ww-mutex-nested   Chris Wilson
    • [PATCH 22/23] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
    • [PATCH 23/23] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/23] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/23] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/23] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/23] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/23] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/23] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/23] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/23] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/23] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/23] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/23] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/23] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/23] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/23] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/23] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/23] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/23] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/23] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/23] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 19/23] snoop-batch   Chris Wilson
    • [PATCH 20/23] vmscan   Chris Wilson
    • [PATCH 21/23] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
    • [PATCH 22/23] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 23/23] ctx-lockref   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/23] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 1/2] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 2/2] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [1/2] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [1/2] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/22] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/22] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/22] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 04/22] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 05/22] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/22] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/22] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/22] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/22] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/22] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/22] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/22] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/22] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/22] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/22] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/22] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/22] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/22] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 19/22] snoop-batch   Chris Wilson
  • [PATCH 20/22] vmscan   Chris Wilson
  • [PATCH 21/22] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
  • [PATCH 22/22] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/21] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/21] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/21] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 04/21] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 04/21] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 05/21] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/21] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/21] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/21] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/21] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/21] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/21] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/21] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/21] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/21] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/21] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/21] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/21] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/21] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 19/21] snoop-batch   Chris Wilson
  • [PATCH 20/21] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 21/21] ctx-lockref   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/21] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/21] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/21] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/21] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/21] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/21] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/21] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/21] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/21] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/21] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/21] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/21] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/21] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/21] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/21] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/21] snoop-batch   Chris Wilson
    • [PATCH 16/21] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 17/21] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 18/21] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 19/21] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 20/21] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 21/21] smart-acquire-lock   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/21] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/21] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/21] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/21] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 05/21] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/21] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/21] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/21] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/21] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/21] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/21] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/21] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/21] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/21] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/21] snoop-batch   Chris Wilson
    • [PATCH 16/21] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 17/21] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 18/21] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 19/21] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 20/21] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 21/21] smart-acquire-lock   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/21] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/21] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/17] drm/i915/selftests: Control group_size for hexdump   Chris Wilson
  • [PATCH 02/17] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 03/17] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 04/17] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 05/17] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 06/17] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 07/17] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 08/17] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 09/17] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 10/17] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 11/17] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 12/17] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 13/17] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 14/17] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 15/17] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 16/17] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 17/17] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/17] drm/i915/selftests: Control group_size for hexdump   Patchwork
• [PATCH 01/20] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/20] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/20] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 06/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 07/20] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 08/20] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 09/20] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 10/20] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 11/20] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 12/20] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 13/20] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 14/20] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 15/20] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 16/20] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 17/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 18/20] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 19/20] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 20/20] drm/i915: Track i915_vma with its own reference counter   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/20] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/20] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/20] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/20] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/20] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/20] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/20] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/20] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/20] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/20] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/20] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/20] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/20] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/20] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 19/20] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 20/20] drm/i915: Track i915_vma with its own reference counter   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/20] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/20] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/20] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/20] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/20] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/20] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/20] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/20] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/20] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/20] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/20] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/20] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/20] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/20] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 19/20] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 20/20] drm/i915: Track i915_vma with its own reference counter   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/20] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/20] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/20] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/20] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/20] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/20] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/20] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/20] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/20] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/20] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/20] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/20] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/20] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/20] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 19/20] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 20/20] drm/i915: Track i915_vma with its own reference counter   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/20] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/20] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/20] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 04/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/20] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/20] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/20] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/20] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/20] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/20] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/20] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/20] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/20] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/20] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/20] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 19/20] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 20/20] drm/i915: Track i915_vma with its own reference counter   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/20] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/20] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/20] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 05/20] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 06/20] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 07/20] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 08/20] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 09/20] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 10/20] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 11/20] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 12/20] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 13/20] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 14/20] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 15/20] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 16/20] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 17/20] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 18/20] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 19/20] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 20/20] drm/i915: Track i915_vma with its own reference counter   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/20] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 04/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 04/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 04/20] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
• [PATCH] weird loadavg on idle machine post 5.7   Chris Wilson
  • ✓ Fi.CI.BAT: success for weird loadavg on idle machine post 5.7   Patchwork
  • ✓ Fi.CI.IGT: success for weird loadavg on idle machine post 5.7   Patchwork
• [PATCH 01/23] Revert "drm/i915/gem: Async GPU relocations only"   Maarten Lankhorst
  • [PATCH 02/23] drm/i915: Revert relocation chaining commits.   Maarten Lankhorst
  • [PATCH 03/23] Revert "drm/i915/gem: Drop relocation slowpath".   Maarten Lankhorst
  • [PATCH 04/23] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Maarten Lankhorst
  • [PATCH 05/23] drm/i915: Remove locking from i915_gem_object_prepare_read/write   Maarten Lankhorst
  • [PATCH 06/23] drm/i915: Parse command buffer earlier in eb_relocate(slow)   Maarten Lankhorst
  • [PATCH 07/23] Revert "drm/i915/gem: Split eb_vma into its own allocation"   Maarten Lankhorst
  • [PATCH 08/23] drm/i915: Use per object locking in execbuf, v12.   Maarten Lankhorst
  • [PATCH 09/23] drm/i915: Use ww locking in intel_renderstate.   Maarten Lankhorst
  • [PATCH 10/23] drm/i915: Add ww context handling to context_barrier_task   Maarten Lankhorst
  • [PATCH 11/23] drm/i915: Nuke arguments to eb_pin_engine   Maarten Lankhorst
  • [PATCH 12/23] drm/i915: Pin engine before pinning all objects, v5.   Maarten Lankhorst
  • [PATCH 13/23] drm/i915: Rework intel_context pinning to do everything outside of pin_mutex   Maarten Lankhorst
  • [PATCH 14/23] drm/i915: Make sure execbuffer always passes ww state to i915_vma_pin.   Maarten Lankhorst
  • [PATCH 15/23] drm/i915: Convert i915_gem_object/client_blt.c to use ww locking as well, v2.   Maarten Lankhorst
  • [PATCH 16/23] drm/i915: Kill last user of intel_context_create_request outside of selftests   Maarten Lankhorst
  • [PATCH 17/23] drm/i915: Convert i915_perf to ww locking as well   Maarten Lankhorst
  • [PATCH 18/23] drm/i915: Dirty hack to fix selftests locking inversion   Maarten Lankhorst
  • [PATCH 19/23] drm/i915/selftests: Fix locking inversion in lrc selftest.   Maarten Lankhorst
  • [PATCH 20/23] drm/i915: Use ww pinning for intel_context_create_request()   Maarten Lankhorst
  • [PATCH 21/23] drm/i915: Move i915_vma_lock in the selftests to avoid lock inversion, v2.   Maarten Lankhorst
  • [PATCH 22/23] drm/i915: Add ww locking to vm_fault_gtt   Maarten Lankhorst
  • [PATCH 23/23] drm/i915: Add ww locking to pin_to_display_plane   Maarten Lankhorst
  • [PATCH 01/23] Revert "drm/i915/gem: Async GPU relocations only"   Maarten Lankhorst
    • [PATCH 02/23] drm/i915: Revert relocation chaining commits.   Maarten Lankhorst
    • [PATCH 03/23] Revert "drm/i915/gem: Drop relocation slowpath".   Maarten Lankhorst
    • [PATCH 04/23] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Maarten Lankhorst
    • [PATCH 05/23] drm/i915: Remove locking from i915_gem_object_prepare_read/write   Maarten Lankhorst
    • [PATCH 06/23] drm/i915: Parse command buffer earlier in eb_relocate(slow)   Maarten Lankhorst
    • [PATCH 07/23] Revert "drm/i915/gem: Split eb_vma into its own allocation"   Maarten Lankhorst
    • [PATCH 08/23] drm/i915: Use per object locking in execbuf, v12.   Maarten Lankhorst
    • [PATCH 09/23] drm/i915: Use ww locking in intel_renderstate.   Maarten Lankhorst
    • [PATCH 10/23] drm/i915: Add ww context handling to context_barrier_task   Maarten Lankhorst
    • [PATCH 11/23] drm/i915: Nuke arguments to eb_pin_engine   Maarten Lankhorst
    • [PATCH 12/23] drm/i915: Pin engine before pinning all objects, v5.   Maarten Lankhorst
    • [PATCH 13/23] drm/i915: Rework intel_context pinning to do everything outside of pin_mutex   Maarten Lankhorst
    • [PATCH 14/23] drm/i915: Make sure execbuffer always passes ww state to i915_vma_pin.   Maarten Lankhorst
    • [PATCH 15/23] drm/i915: Convert i915_gem_object/client_blt.c to use ww locking as well, v2.   Maarten Lankhorst
    • [PATCH 16/23] drm/i915: Kill last user of intel_context_create_request outside of selftests   Maarten Lankhorst
    • [PATCH 17/23] drm/i915: Convert i915_perf to ww locking as well   Maarten Lankhorst
    • [PATCH 18/23] drm/i915: Dirty hack to fix selftests locking inversion   Maarten Lankhorst
    • [PATCH 19/23] drm/i915/selftests: Fix locking inversion in lrc selftest.   Maarten Lankhorst
    • [PATCH 20/23] drm/i915: Use ww pinning for intel_context_create_request()   Maarten Lankhorst
    • [PATCH 21/23] drm/i915: Move i915_vma_lock in the selftests to avoid lock inversion, v2.   Maarten Lankhorst
    • [PATCH 22/23] drm/i915: Add ww locking to vm_fault_gtt   Maarten Lankhorst
    • [PATCH 23/23] drm/i915: Add ww locking to pin_to_display_plane   Maarten Lankhorst
    • ✗ Fi.CI.BAT: failure for series starting with [01/23] Revert "drm/i915/gem: Async GPU relocations only"   Patchwork
  • ✗ Fi.CI.BUILD: failure for series starting with [01/23] Revert "drm/i915/gem: Async GPU relocations only"   Patchwork
• [PATCH] drm/vgem: Do not allocate backing shmemfs file for an import dmabuf object   Chris Wilson
  • ✗ Fi.CI.BUILD: failure for drm/vgem: Do not allocate backing shmemfs file for an import dmabuf object   Patchwork
  • [PATCH] drm/vgem: Do not allocate backing shmemfs file for an import dmabuf object   Chris Wilson
  • [PATCH] drm/vgem: Do not allocate backing shmemfs file for an import dmabuf object   Chris Wilson
    • ✗ Fi.CI.BAT: failure for drm/vgem: Do not allocate backing shmemfs file for an import dmabuf object (rev3)   Patchwork
    • ✓ Fi.CI.BAT: success for drm/vgem: Do not allocate backing shmemfs file for an import dmabuf object (rev4)   Patchwork
    • ✗ Fi.CI.IGT: failure for drm/vgem: Do not allocate backing shmemfs file for an import dmabuf object (rev4)   Patchwork
• [PATCH] drm/i915/gem: Unpin idle contexts from kswapd reclaim   Chris Wilson
  • ✓ Fi.CI.BAT: success for drm/i915/gem: Unpin idle contexts from kswapd reclaim   Patchwork
  • ✗ Fi.CI.IGT: failure for drm/i915/gem: Unpin idle contexts from kswapd reclaim   Patchwork
• [PATCH i-g-t] tests/device_reset: Test device sysfs reset   Marcin Bernatowicz
  • [PATCH i-g-t 2/2] Add device_reset subtests to fast feedback list for getting results promptly   Marcin Bernatowicz
  • ✗ Fi.CI.BAT: failure for tests/device_reset: Test device sysfs reset   Patchwork
  • ✗ Fi.CI.BAT: failure for tests/device_reset: Test device sysfs reset   Patchwork
• [PATCH 01/27] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Chris Wilson
  • [PATCH 02/27] drm/i915: Release shortlived maps of longlived objects   Chris Wilson
  • [PATCH 04/27] drm/i915/display: Fix initial fb to use resource_size_t   Chris Wilson
  • [PATCH 05/27] drm/i915/gem: Unpin idle contexts from kswapd reclaim   Chris Wilson
  • [PATCH 06/27] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 07/27] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 08/27] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 09/27] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 10/27] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 11/27] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 12/27] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 13/27] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 14/27] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 15/27] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 16/27] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 17/27] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 18/27] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 19/27] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 20/27] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 21/27] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 22/27] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 23/27] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 24/27] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 25/27] ggtt-locked   Chris Wilson
  • [PATCH 26/27] ctx-acq   Chris Wilson
  • [PATCH 27/27] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/27] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
  • [PATCH 01/27] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Chris Wilson
    • [PATCH 02/27] drm/i915: Release shortlived maps of longlived objects   Chris Wilson
    • [PATCH 03/27] drm/i915: Remove i915_gem_object_get_dirty_page()   Chris Wilson
    • [PATCH 04/27] drm/i915/display: Fix initial fb to use resource_size_t   Chris Wilson
    • [PATCH 05/27] drm/i915/gem: Unpin idle contexts from kswapd reclaim   Chris Wilson
    • [PATCH 06/27] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 07/27] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 08/27] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 09/27] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 10/27] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 11/27] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 12/27] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 13/27] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 14/27] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 15/27] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 16/27] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 17/27] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 18/27] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 19/27] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 20/27] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 21/27] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 22/27] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 23/27] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 24/27] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 25/27] ggtt-locked   Chris Wilson
    • [PATCH 26/27] ctx-acq   Chris Wilson
    • [PATCH 27/27] wait-context   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/27] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
• [PATCH 03/27] drm/i915: Remove i915_gem_object_get_dirty_page()   Chris Wilson
• [PATCH 01/28] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Chris Wilson
  • [PATCH 02/28] drm/i915: Release shortlived maps of longlived objects   Chris Wilson
  • [PATCH 03/28] drm/i915: Remove i915_gem_object_get_dirty_page()   Chris Wilson
  • [PATCH 04/28] drm/i915/display: Fix initial fb to use resource_size_t   Chris Wilson
  • [PATCH 05/28] drm/i915/gem: Unpin idle contexts from kswapd reclaim   Chris Wilson
  • [PATCH 06/28] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 07/28] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 08/28] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 09/28] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 10/28] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 11/28] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 12/28] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 13/28] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 14/28] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 15/28] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 16/28] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 17/28] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 18/28] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 19/28] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 20/28] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 21/28] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 22/28] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 23/28] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 24/28] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 25/28] ggtt-locked   Chris Wilson
  • [PATCH 26/28] ctx-acq   Chris Wilson
  • [PATCH 27/28] wait-context   Chris Wilson
  • [PATCH 28/28] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/28] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
  • [PATCH 01/28] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Chris Wilson
    • [PATCH 02/28] drm/i915: Release shortlived maps of longlived objects   Chris Wilson
    • [PATCH 03/28] drm/i915: Remove i915_gem_object_get_dirty_page()   Chris Wilson
    • [PATCH 04/28] drm/i915/display: Fix initial fb to use resource_size_t   Chris Wilson
    • [PATCH 05/28] drm/i915/gem: Unpin idle contexts from kswapd reclaim   Chris Wilson
    • [PATCH 06/28] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 07/28] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 08/28] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 09/28] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 10/28] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 11/28] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 12/28] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 13/28] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 14/28] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 15/28] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 16/28] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 17/28] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 18/28] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 19/28] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 20/28] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 21/28] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 22/28] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 23/28] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 24/28] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 25/28] ggtt-locked   Chris Wilson
    • [PATCH 26/28] ctx-acq   Chris Wilson
    • [PATCH 27/28] wait-context   Chris Wilson
    • [PATCH 28/28] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/28] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
• [PATCH 1/5] drm/i915/display: Replace drm_i915_private in voltage swing functions by intel_encoder   José Roberto de Souza
  • [PATCH 3/5] drm/i915/bios: Parse HOBL parameter   José Roberto de Souza
  • [PATCH 5/5] Debug hobl   José Roberto de Souza
  • ✗ Fi.CI.BAT: failure for series starting with [1/5] drm/i915/display: Replace drm_i915_private in voltage swing functions by intel_encoder   Patchwork
  • [PATCH 1/5] drm/i915/display: Replace drm_i915_private in voltage swing functions by intel_encoder   José Roberto de Souza
    • [PATCH 3/5] drm/i915/bios: Parse HOBL parameter   José Roberto de Souza
    • [PATCH 4/5] drm/i915/display: Implement HOBL   José Roberto de Souza
    • [PATCH 5/5] Debug hobl   José Roberto de Souza
    • ✗ Fi.CI.BAT: failure for series starting with [1/5] drm/i915/display: Replace drm_i915_private in voltage swing functions by intel_encoder   Patchwork
    • ✓ Fi.CI.BAT: success for series starting with [1/5] drm/i915/display: Replace drm_i915_private in voltage swing functions by intel_encoder (rev2)   Patchwork
    • ✓ Fi.CI.IGT: success for series starting with [1/5] drm/i915/display: Replace drm_i915_private in voltage swing functions by intel_encoder (rev2)   Patchwork
• [PATCH 2/5] drm/i915/display: Remove port and phy from voltage swing functions   José Roberto de Souza
  • [PATCH 2/5] drm/i915/display: Remove port and phy from voltage swing functions   José Roberto de Souza
• [PATCH 4/5] drm/i915/display: Implement HOBL   José Roberto de Souza
• [PATCH 01/26] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Chris Wilson
  • [PATCH 02/26] drm/i915: Release shortlived maps of longlived objects   Chris Wilson
  • [PATCH 03/26] drm/i915: Remove i915_gem_object_get_dirty_page()   Chris Wilson
  • [PATCH 04/26] drm/i915/display: Fix initial fb to use resource_size_t   Chris Wilson
  • [PATCH 05/26] drm/i915/gem: Unpin idle contexts from kswapd reclaim   Chris Wilson
  • [PATCH 06/26] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 07/26] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 08/26] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 09/26] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 10/26] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 11/26] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 12/26] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 13/26] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 14/26] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 15/26] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 16/26] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 17/26] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 18/26] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 19/26] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 20/26] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 21/26] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 22/26] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 23/26] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 24/26] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 25/26] ggtt-locked   Chris Wilson
  • [PATCH 26/26] ctx-acq   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/26] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/26] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
• [PATCH 01/29] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Chris Wilson
  • [PATCH 02/29] drm/i915: Release shortlived maps of longlived objects   Chris Wilson
  • [PATCH 03/29] drm/i915: Remove i915_gem_object_get_dirty_page()   Chris Wilson
  • [PATCH 04/29] drm/i915/display: Fix initial fb to use resource_size_t   Chris Wilson
  • [PATCH 05/29] drm/i915/gem: Unpin idle contexts from kswapd reclaim   Chris Wilson
  • [PATCH 06/29] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 07/29] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 08/29] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
  • [PATCH 09/29] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 10/29] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 11/29] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 12/29] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 13/29] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 14/29] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 15/29] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 16/29] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 17/29] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 18/29] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 19/29] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 20/29] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 21/29] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 22/29] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 23/29] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 24/29] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 25/29] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 26/29] ggtt-locked   Chris Wilson
  • [PATCH 27/29] ctx-acq   Chris Wilson
  • [PATCH 28/29] wait-context   Chris Wilson
  • [PATCH 29/29] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/29] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/29] drm/i915/gt: Replace opencoded i915_gem_object_pin_map()   Patchwork
• [PATCH 01/28] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 02/28] act-reuse-cache   Chris Wilson
  • [PATCH 03/28] act   Chris Wilson
  • [PATCH 04/28] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 05/28] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 06/28] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
  • [PATCH 07/28] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 08/28] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 09/28] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 10/28] rearrange-lookup   Chris Wilson
  • [PATCH 11/28] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 12/28] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 13/28] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 14/28] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 15/28] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 16/28] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 17/28] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 18/28] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 19/28] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 20/28] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 21/28] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 22/28] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 23/28] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 24/28] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 25/28] ggtt-locked   Chris Wilson
  • [PATCH 26/28] ctx-acq   Chris Wilson
  • [PATCH 27/28] wait-context   Chris Wilson
  • [PATCH 28/28] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/28] drm/i915: Soften the tasklet flush frequency before waits   Patchwork
• [PATCH 01/29] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 02/29] drm/i915: Rename struct i915_active_node   Chris Wilson
  • [PATCH 03/29] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 04/29] drm/i915: Skip taking acquire mutex if !ref->active   Chris Wilson
  • [PATCH 05/29] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 06/29] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 07/29] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
  • [PATCH 08/29] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 09/29] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 10/29] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 11/29] rearrange-lookup   Chris Wilson
  • [PATCH 12/29] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 13/29] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 14/29] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 15/29] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 16/29] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 17/29] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 18/29] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 19/29] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 20/29] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 21/29] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 22/29] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 23/29] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 24/29] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 25/29] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 26/29] ggtt-locked   Chris Wilson
  • [PATCH 27/29] ctx-acq   Chris Wilson
  • [PATCH 28/29] wait-context   Chris Wilson
  • [PATCH 29/29] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/29] drm/i915: Soften the tasklet flush frequency before waits   Patchwork
• [PATCH 01/30] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 02/30] drm/i915: Rename struct i915_active_node   Chris Wilson
  • [PATCH 03/30] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 04/30] drm/i915: Skip taking acquire mutex if !ref->active   Chris Wilson
  • [PATCH 05/30] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 06/30] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 07/30] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
  • [PATCH 08/30] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 09/30] tl-parse   Chris Wilson
  • [PATCH 10/30] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 11/30] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 12/30] rearrange-lookup   Chris Wilson
  • [PATCH 13/30] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 14/30] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 15/30] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 16/30] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 17/30] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 18/30] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 19/30] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 20/30] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 21/30] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 22/30] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 23/30] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 24/30] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 25/30] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 26/30] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 27/30] ggtt-locked   Chris Wilson
  • [PATCH 28/30] ctx-acq   Chris Wilson
  • [PATCH 29/30] wait-context   Chris Wilson
  • [PATCH 30/30] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/30] drm/i915: Soften the tasklet flush frequency before waits   Patchwork
  • ✗ Fi.CI.BUILD: warning for series starting with [01/30] drm/i915: Soften the tasklet flush frequency before waits   Patchwork
• [PATCH] drm/i915/dp: Try the bspec value for CLKTOP2_CORECLKCTL   Manasi Navare
  • ✓ Fi.CI.BAT: success for drm/i915/dp: Try the bspec value for CLKTOP2_CORECLKCTL   Patchwork
  • ✓ Fi.CI.IGT: success for drm/i915/dp: Try the bspec value for CLKTOP2_CORECLKCTL   Patchwork
• [PATCH 1/6] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 2/6] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 3/6] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 4/6] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 5/6] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 6/6] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 1/6] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 2/6] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 3/6] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 4/6] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 5/6] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 6/6] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • ✗ Fi.CI.BUILD: failure for series starting with [1/6] drm/i915: Soften the tasklet flush frequency before waits   Patchwork
  • ✗ Fi.CI.BUILD: failure for series starting with [1/6] drm/i915: Soften the tasklet flush frequency before waits   Patchwork
• [PATCH 1/6] drm/i915: Skip taking acquire mutex if !ref->active   Chris Wilson
  • [PATCH 2/6] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 3/6] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 4/6] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 5/6] drm/i915: Rename struct i915_active_node   Chris Wilson
  • [PATCH 6/6] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/6] drm/i915: Skip taking acquire mutex if !ref->active   Patchwork
• [PATCH 1/4] drm/i915: Skip taking acquire mutex if !ref->active   Chris Wilson
  • [PATCH 2/4] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 3/4] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 4/4] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/4] drm/i915: Skip taking acquire mutex if !ref->active   Patchwork
• [PATCH 1/3] drm/i915: Skip taking acquire mutex if !ref->active   Chris Wilson
  • [PATCH 2/3] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 3/3] hmm   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [1/3] drm/i915: Skip taking acquire mutex if !ref->active   Patchwork
  • ✓ Fi.CI.IGT: success for series starting with [1/3] drm/i915: Skip taking acquire mutex if !ref->active   Patchwork
• [PATCH 01/30] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 02/30] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 03/30] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 04/30] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 05/30] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 06/30] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 07/30] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 08/30] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
  • [PATCH 09/30] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 10/30] tl-parse   Chris Wilson
  • [PATCH 11/30] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 12/30] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 13/30] rearrange-lookup   Chris Wilson
  • [PATCH 14/30] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 15/30] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 16/30] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 17/30] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 18/30] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 19/30] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 20/30] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 21/30] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 22/30] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 23/30] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 24/30] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 25/30] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 26/30] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 27/30] ggtt-locked   Chris Wilson
  • [PATCH 28/30] ctx-acq   Chris Wilson
  • [PATCH 29/30] wait-context   Chris Wilson
  • [PATCH 30/30] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/30] drm/i915: Skip taking acquire mutex for no ref->active callback   Patchwork
• [PATCH 01/32] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 02/32] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 03/32] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 04/32] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 05/32] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 06/32] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 07/32] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 08/32] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 09/32] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 10/32] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
  • [PATCH 11/32] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 12/32] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 13/32] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 14/32] rearrange-lookup   Chris Wilson
  • [PATCH 15/32] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 16/32] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 17/32] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 18/32] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 19/32] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 20/32] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 21/32] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 22/32] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 23/32] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 24/32] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 25/32] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 26/32] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 27/32] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 28/32] no-pin-pages   Chris Wilson
  • [PATCH 29/32] ggtt-locked   Chris Wilson
  • [PATCH 30/32] ctx-acq   Chris Wilson
  • [PATCH 31/32] wait-context   Chris Wilson
  • [PATCH 32/32] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/32] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
  • [PATCH 01/32] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 02/32] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 03/32] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 04/32] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 05/32] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 06/32] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 07/32] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 08/32] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 09/32] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 10/32] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
    • [PATCH 11/32] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 12/32] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 13/32] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 14/32] rearrange-lookup   Chris Wilson
    • [PATCH 15/32] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 16/32] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 17/32] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 18/32] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 19/32] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 20/32] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 21/32] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 22/32] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 23/32] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 24/32] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 25/32] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 26/32] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 27/32] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 28/32] no-pin-pages   Chris Wilson
    • [PATCH 29/32] ggtt-locked   Chris Wilson
    • [PATCH 30/32] ctx-acq   Chris Wilson
    • [PATCH 31/32] wait-context   Chris Wilson
    • [PATCH 32/32] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/32] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
• [PATCH 01/31] drm/i915/gt: Optimise aliasing-ppgtt allocations   Chris Wilson
  • [PATCH 03/31] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 04/31] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 06/31] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 07/31] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 08/31] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 10/31] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 13/31] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 14/31] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 15/31] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 20/31] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 22/31] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 23/31] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 24/31] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 25/31] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 26/31] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/31] drm/i915/gt: Optimise aliasing-ppgtt allocations   Patchwork
• [PATCH 02/31] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
• [PATCH 05/31] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
• [PATCH 09/31] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
• [PATCH 11/31] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
• [PATCH 12/31] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
• [PATCH 16/31] drm/i915: Always defer fenced work to the worker   Chris Wilson
• [PATCH 17/31] drm/i915/gem: Assign context id for async work   Chris Wilson
• [PATCH 18/31] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
• [PATCH 19/31] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
• [PATCH 21/31] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
• [PATCH 27/31] no-pin-pages   Chris Wilson
• [PATCH 28/31] ggtt-locked   Chris Wilson
• [PATCH 29/31] ctx-acq   Chris Wilson
• [PATCH 30/31] wait-context   Chris Wilson
• [PATCH 31/31] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
• [PATCH 01/25] drm/i915/selftest: Check that GPR are restored across noa_wait   Chris Wilson
  • [PATCH 02/25] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 03/25] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 04/25] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 05/25] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 06/25] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 07/25] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 08/25] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 09/25] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 10/25] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 11/25] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 12/25] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 13/25] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 14/25] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 15/25] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 16/25] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 17/25] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 18/25] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 19/25] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 20/25] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 21/25] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 22/25] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 23/25] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 24/25] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 25/25] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/25] drm/i915/selftest: Check that GPR are restored across noa_wait   Patchwork
• [PATCH 01/30] drm/i915/selftest: Check that GPR are restored across noa_wait   Chris Wilson
  • [PATCH 02/30] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 03/30] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 04/30] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 05/30] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 06/30] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 07/30] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 08/30] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 09/30] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 10/30] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 11/30] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 12/30] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 13/30] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 14/30] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 15/30] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 16/30] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 17/30] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 18/30] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 19/30] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 20/30] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 21/30] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 22/30] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 23/30] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 24/30] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 25/30] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 26/30] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 27/30] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 28/30] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 29/30] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 30/30] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/30] drm/i915/selftest: Check that GPR are restored across noa_wait   Patchwork
• [PATCH 01/31] drm/i915/perf: Use GTT when saving/restoring engine GPR   Chris Wilson
  • [PATCH 02/31] drm/i915/selftest: Check that GPR are restored across noa_wait   Chris Wilson
  • [PATCH 03/31] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 04/31] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 05/31] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 06/31] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 07/31] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 08/31] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 09/31] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 10/31] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 11/31] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 12/31] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 13/31] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 14/31] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 15/31] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 16/31] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 17/31] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 18/31] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 19/31] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 20/31] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 21/31] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 22/31] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 23/31] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 24/31] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 25/31] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 26/31] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 27/31] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 28/31] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 29/31] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 30/31] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 31/31] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/31] drm/i915/perf: Use GTT when saving/restoring engine GPR   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/31] drm/i915/perf: Use GTT when saving/restoring engine GPR   Patchwork
• [PATCH 01/30] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 02/30] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 03/30] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 04/30] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 05/30] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 06/30] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 07/30] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 08/30] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 09/30] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 10/30] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 11/30] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 12/30] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 13/30] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 14/30] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 15/30] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 16/30] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 17/30] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 18/30] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 19/30] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 20/30] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 21/30] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 22/30] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 23/30] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 24/30] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 25/30] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 26/30] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 27/30] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 28/30] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 29/30] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 30/30] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/30] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/30] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
  • [PATCH 01/30] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 02/30] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 03/30] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 04/30] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 05/30] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 06/30] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 07/30] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 08/30] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 09/30] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 10/30] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 11/30] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 12/30] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 13/30] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 14/30] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 15/30] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 16/30] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 17/30] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 18/30] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 19/30] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 20/30] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 21/30] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 22/30] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 23/30] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 24/30] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 25/30] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 26/30] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 27/30] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 28/30] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 29/30] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 30/30] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/30] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/30] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
  • [PATCH 01/30] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 02/30] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 03/30] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 04/30] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 05/30] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 06/30] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 07/30] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 08/30] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 09/30] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 10/30] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 11/30] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 12/30] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 13/30] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 14/30] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 15/30] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 16/30] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 17/30] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 18/30] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 19/30] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 20/30] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 21/30] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 22/30] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 23/30] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 24/30] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 25/30] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 26/30] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 27/30] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 28/30] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 29/30] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 30/30] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/30] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
• [PATCH 01/63] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 02/63] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 02/63] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 03/63] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 04/63] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 05/63] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 06/63] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 07/63] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 08/63] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 09/63] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 10/63] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 11/63] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 12/63] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 13/63] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 14/63] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 15/63] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 16/63] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 17/63] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 18/63] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 19/63] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 20/63] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 21/63] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 22/63] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 23/63] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 24/63] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 25/63] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 26/63] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 27/63] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 28/63] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 29/63] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 30/63] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 31/63] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 32/63] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 33/63] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 34/63] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 35/63] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 36/63] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 37/63] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 38/63] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 39/63] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 40/63] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 41/63] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 42/63] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 43/63] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 44/63] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 45/63] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 46/63] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 47/63] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 48/63] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 49/63] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 50/63] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 51/63] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 52/63] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 53/63] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 54/63] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 55/63] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 56/63] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 57/63] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 58/63] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 59/63] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 60/63] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 61/63] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 62/63] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 63/63] vmscan   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/63] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
  • [PATCH 01/63] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 03/63] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 04/63] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 05/63] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 06/63] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 07/63] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 08/63] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 09/63] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 10/63] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 11/63] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 12/63] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 13/63] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 14/63] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 15/63] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 16/63] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 17/63] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 18/63] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 19/63] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 20/63] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 21/63] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 22/63] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 23/63] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 24/63] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 25/63] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 26/63] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 27/63] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 28/63] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 29/63] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 30/63] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 31/63] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 32/63] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 33/63] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 34/63] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 35/63] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 36/63] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 37/63] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 38/63] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 39/63] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 40/63] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 41/63] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 42/63] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 43/63] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 44/63] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 45/63] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 46/63] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 47/63] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 48/63] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 49/63] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 50/63] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 51/63] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 52/63] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 53/63] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 54/63] drm/i915: Move saturated workload detection to the GT   Chris Wilson
    • [PATCH 55/63] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 56/63] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 57/63] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 58/63] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 59/63] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 60/63] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 61/63] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 62/63] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 63/63] vmscan   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/63] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
• [PATCH 01/31] drm/i915/gt: Be defensive in the face of false CS events   Chris Wilson
  • [PATCH 02/31] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 03/31] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 04/31] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 05/31] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 06/31] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 07/31] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 08/31] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 09/31] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 10/31] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 11/31] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 12/31] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 13/31] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 14/31] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 15/31] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 16/31] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 17/31] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 18/31] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 19/31] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 20/31] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 21/31] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 22/31] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 23/31] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 24/31] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 25/31] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 26/31] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 27/31] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 28/31] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 29/31] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 30/31] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 31/31] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/31] drm/i915/gt: Be defensive in the face of false CS events   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/31] drm/i915/gt: Be defensive in the face of false CS events   Patchwork
• [PATCH 01/65] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
  • [PATCH 02/65] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 03/65] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 04/65] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 05/65] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 06/65] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 07/65] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 08/65] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 09/65] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 10/65] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 11/65] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 12/65] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 13/65] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 14/65] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 15/65] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 16/65] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 17/65] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 18/65] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 19/65] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 20/65] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 21/65] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 22/65] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 23/65] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 24/65] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 25/65] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 26/65] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 27/65] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 28/65] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 29/65] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 30/65] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 31/65] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 32/65] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 33/65] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 34/65] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 35/65] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 36/65] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 37/65] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 38/65] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 39/65] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 40/65] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 41/65] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 42/65] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 43/65] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 44/65] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 45/65] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 46/65] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 47/65] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 48/65] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 49/65] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 50/65] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 51/65] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 52/65] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 53/65] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 54/65] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 55/65] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 56/65] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 57/65] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 58/65] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 59/65] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 60/65] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 61/65] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 62/65] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 63/65] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 64/65] vmscan   Chris Wilson
  • [PATCH 65/65] kcsan-lock   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/65] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Patchwork
  • [PATCH 01/65] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
    • [PATCH 02/65] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 03/65] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 04/65] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 05/65] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 06/65] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 07/65] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 08/65] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 09/65] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 10/65] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 11/65] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 12/65] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 13/65] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 14/65] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 15/65] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 16/65] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 17/65] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 18/65] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 19/65] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 20/65] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 21/65] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 22/65] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 23/65] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 24/65] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 25/65] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 26/65] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 27/65] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 28/65] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 29/65] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 30/65] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 31/65] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 32/65] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 33/65] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 34/65] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 35/65] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 36/65] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 37/65] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 38/65] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 39/65] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 40/65] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 41/65] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 42/65] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 43/65] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 44/65] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 45/65] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 46/65] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 47/65] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 48/65] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 49/65] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 50/65] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 51/65] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 52/65] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 53/65] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 54/65] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 55/65] drm/i915: Move saturated workload detection to the GT   Chris Wilson
    • [PATCH 56/65] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 57/65] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 58/65] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 59/65] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 60/65] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 61/65] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 62/65] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 63/65] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 64/65] vmscan   Chris Wilson
    • [PATCH 65/65] kcsan-lock   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/65] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/65] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Patchwork
• [PATCH 01/66] drm/i915/gt: Ignore irq enabling on the virtual engines   Chris Wilson
  • [PATCH 02/66] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
  • [PATCH 03/66] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 04/66] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 05/66] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 06/66] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 07/66] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 08/66] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 09/66] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 10/66] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 11/66] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 12/66] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 13/66] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 14/66] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 15/66] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 16/66] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 17/66] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 18/66] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 19/66] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 20/66] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 21/66] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 22/66] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 23/66] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 24/66] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 25/66] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 26/66] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 27/66] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 28/66] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 29/66] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 30/66] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 31/66] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 32/66] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 33/66] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 34/66] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 35/66] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 36/66] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 37/66] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 38/66] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 39/66] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 40/66] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 41/66] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 42/66] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 43/66] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 44/66] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 45/66] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 46/66] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 47/66] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 48/66] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 49/66] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 50/66] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 51/66] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 52/66] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 53/66] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 54/66] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 55/66] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 56/66] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 57/66] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 58/66] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 59/66] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 60/66] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 61/66] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 62/66] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 63/66] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 64/66] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 65/66] vmscan   Chris Wilson
  • [PATCH 66/66] kcsan-lock   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/66] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/66] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
• [PATCH 01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Chris Wilson
  • [PATCH 02/67] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
    • [PATCH 02/67] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
  • [PATCH 03/67] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 05/67] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 06/67] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 07/67] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 08/67] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 09/67] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 10/67] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 11/67] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 12/67] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 13/67] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 14/67] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 15/67] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 16/67] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 17/67] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 18/67] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 19/67] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 20/67] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 21/67] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 22/67] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 23/67] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 24/67] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 25/67] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 26/67] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 27/67] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 28/67] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 29/67] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 30/67] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 31/67] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 32/67] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 33/67] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 34/67] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 35/67] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 36/67] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 37/67] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 38/67] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 39/67] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 40/67] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 41/67] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 42/67] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 43/67] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 44/67] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 45/67] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 46/67] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 47/67] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 48/67] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 49/67] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 50/67] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 51/67] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 52/67] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 53/67] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 54/67] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 55/67] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 56/67] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 57/67] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 58/67] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 59/67] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 60/67] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 61/67] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 62/67] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 63/67] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 64/67] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 65/67] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 66/67] vmscan   Chris Wilson
  • [PATCH 67/67] kcsan-lock   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
  • [PATCH 01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Chris Wilson
    • [PATCH 02/67] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
    • [PATCH 03/67] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 04/67] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 05/67] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 06/67] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 07/67] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 08/67] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 09/67] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 10/67] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 11/67] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 12/67] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 13/67] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 14/67] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 15/67] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 16/67] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 17/67] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 18/67] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 19/67] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 20/67] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 21/67] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 22/67] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 23/67] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 24/67] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 25/67] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 26/67] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 27/67] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 28/67] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 29/67] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 30/67] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 31/67] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 32/67] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 33/67] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 34/67] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 35/67] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 36/67] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 37/67] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
    • [PATCH 38/67] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 39/67] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 40/67] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 41/67] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 42/67] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 43/67] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 44/67] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 45/67] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 46/67] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 47/67] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 48/67] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 49/67] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 50/67] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 51/67] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 52/67] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 53/67] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 54/67] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 55/67] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 56/67] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 57/67] drm/i915: Move saturated workload detection to the GT   Chris Wilson
    • [PATCH 58/67] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 59/67] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 60/67] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 61/67] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 62/67] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 63/67] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 64/67] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 65/67] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 66/67] vmscan   Chris Wilson
    • [PATCH 67/67] kcsan-lock   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
  • [PATCH 01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Chris Wilson
    • [PATCH 03/67] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 04/67] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 05/67] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 06/67] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 07/67] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 08/67] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 09/67] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 10/67] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 11/67] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 12/67] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 13/67] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 14/67] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 15/67] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 16/67] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 17/67] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 18/67] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 19/67] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 20/67] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 21/67] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 22/67] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 23/67] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 24/67] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 25/67] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 26/67] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 27/67] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 28/67] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 29/67] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 30/67] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 31/67] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 32/67] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 33/67] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 34/67] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 35/67] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 36/67] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 37/67] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
    • [PATCH 38/67] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 39/67] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 40/67] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 41/67] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 42/67] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 43/67] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 44/67] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 45/67] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 46/67] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 47/67] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 48/67] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 49/67] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 50/67] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 51/67] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 52/67] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 53/67] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 54/67] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 55/67] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 56/67] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 57/67] drm/i915: Move saturated workload detection to the GT   Chris Wilson
    • [PATCH 58/67] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 59/67] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 60/67] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 61/67] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 62/67] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 63/67] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 64/67] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 65/67] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 66/67] vmscan   Chris Wilson
    • [PATCH 67/67] kcsan-lock   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
  • [PATCH 01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Chris Wilson
    • [PATCH 02/67] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
    • [PATCH 03/67] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 04/67] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 05/67] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 06/67] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 07/67] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 08/67] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 09/67] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
    • [PATCH 10/67] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 11/67] drm/i915: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 12/67] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 13/67] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 14/67] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 15/67] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 16/67] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 17/67] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 18/67] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 19/67] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 20/67] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 21/67] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 22/67] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 23/67] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 24/67] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 25/67] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 26/67] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 27/67] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 28/67] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 29/67] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 30/67] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 31/67] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 32/67] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 33/67] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 34/67] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 35/67] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 36/67] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 37/67] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
    • [PATCH 38/67] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 39/67] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 40/67] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 41/67] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 42/67] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 43/67] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 44/67] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 45/67] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 46/67] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 47/67] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 48/67] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 49/67] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 50/67] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 51/67] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 52/67] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 53/67] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 54/67] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 55/67] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 56/67] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 57/67] drm/i915: Move saturated workload detection to the GT   Chris Wilson
    • [PATCH 58/67] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 59/67] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 60/67] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 61/67] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 62/67] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 63/67] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 64/67] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 65/67] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 66/67] vmscan   Chris Wilson
    • [PATCH 67/67] kcsan-lock   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/67] drm/i915/gt: Ignore irq enabling on the virtual engines   Patchwork
• [PATCH 04/67] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
• [PATCH i-g-t] lib/i915: Set NO_RELOC flag for spinners   Chris Wilson
  • ✓ Fi.CI.BAT: success for lib/i915: Set NO_RELOC flag for spinners   Patchwork
  • ✓ Fi.CI.IGT: success for lib/i915: Set NO_RELOC flag for spinners   Patchwork
• [PATCH 01/68] drm/i915: Skip signaling a signaled request   Chris Wilson
  • [PATCH 02/68] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 03/68] drm/i915/gt: Ignore irq enabling on the virtual engines   Chris Wilson
  • [PATCH 04/68] drm/i915/gt: Always reset the engine, even if inactive, on execlists failure   Chris Wilson
  • [PATCH 05/68] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 06/68] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 07/68] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 08/68] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 09/68] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 10/68] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 11/68] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 12/68] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 13/68] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 14/68] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 15/68] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 16/68] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 17/68] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 18/68] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 19/68] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 20/68] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 21/68] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 22/68] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 23/68] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 24/68] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 25/68] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 26/68] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 27/68] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 28/68] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 29/68] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 30/68] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 31/68] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 32/68] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 33/68] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 34/68] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 35/68] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 36/68] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 37/68] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 38/68] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 39/68] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 40/68] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 41/68] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 42/68] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 43/68] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 44/68] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 45/68] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 46/68] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 47/68] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 48/68] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 49/68] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 50/68] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 51/68] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 52/68] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 53/68] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 54/68] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 55/68] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 56/68] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 57/68] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 58/68] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 59/68] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 60/68] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 61/68] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 62/68] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 63/68] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 64/68] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 65/68] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 66/68] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 67/68] vmscan   Chris Wilson
  • [PATCH 68/68] kcsan-lock   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/68] drm/i915: Skip signaling a signaled request   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/68] drm/i915: Skip signaling a signaled request   Patchwork
• [PATCH 1/2] drm/i915: Remove i915_request.lock requirement for execution callbacks   Chris Wilson
  • [PATCH 2/2] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 1/2] drm/i915: Remove i915_request.lock requirement for execution callbacks   Chris Wilson
    • [PATCH 2/2] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/2] drm/i915: Remove i915_request.lock requirement for execution callbacks   Patchwork
    • ✓ Fi.CI.IGT: success for series starting with [1/2] drm/i915: Remove i915_request.lock requirement for execution callbacks   Patchwork
  • [PATCH 1/2] drm/i915: Remove i915_request.lock requirement for execution callbacks   Chris Wilson
    • [PATCH 2/2] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [1/2] drm/i915: Remove i915_request.lock requirement for execution callbacks   Patchwork
  • ✗ Fi.CI.BAT: failure for series starting with [1/2] drm/i915: Remove i915_request.lock requirement for execution callbacks   Patchwork
• [PATCH 01/67] drm/i915: Remove i915_request.lock requirement for execution callbacks   Chris Wilson
  • [PATCH 02/67] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 03/67] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 04/67] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 05/67] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 06/67] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 07/67] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 08/67] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 09/67] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 10/67] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 11/67] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 12/67] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 13/67] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 14/67] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 15/67] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 16/67] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 17/67] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 18/67] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 19/67] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 20/67] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 21/67] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 22/67] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 23/67] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 24/67] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 25/67] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 26/67] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 27/67] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 28/67] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 29/67] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 30/67] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 31/67] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 32/67] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 33/67] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 34/67] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 35/67] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 36/67] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 37/67] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 38/67] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 39/67] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 40/67] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 41/67] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 42/67] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 43/67] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 44/67] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 45/67] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 46/67] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 47/67] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 48/67] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 49/67] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 50/67] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 51/67] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 52/67] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 53/67] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 54/67] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 55/67] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 56/67] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 57/67] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 58/67] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 59/67] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 60/67] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 61/67] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 62/67] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 63/67] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 64/67] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 65/67] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 66/67] vmscan   Chris Wilson
  • [PATCH 67/67] kcsan-lock   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/67] drm/i915: Remove i915_request.lock requirement for execution callbacks   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/67] drm/i915: Remove i915_request.lock requirement for execution callbacks   Patchwork
• [PATCH 1/2] drm/i915/gt: Assert the kernel context is using the HWSP   Chris Wilson
  • [PATCH 2/2] ban-scratch-overwrite   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/2] drm/i915/gt: Assert the kernel context is using the HWSP   Patchwork
• [PATCH 1/4] drm/i915/gt: Assert the kernel context is using the HWSP   Chris Wilson
  • [PATCH 2/4] drm/i915: Remove i915_request.lock requirement for execution callbacks   Chris Wilson
  • [PATCH 3/4] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 4/4] breadcrumbsref   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/4] drm/i915/gt: Assert the kernel context is using the HWSP   Patchwork
• [PATCH 01/66] drm/i915/gt: Assert the kernel context is using the HWSP   Chris Wilson
  • [PATCH 02/66] drm/i915: Remove i915_request.lock requirement for execution callbacks   Chris Wilson
  • [PATCH 03/66] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 04/66] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 05/66] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 06/66] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 07/66] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 08/66] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 09/66] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 10/66] drm/i915: Soften the tasklet flush frequency before waits   Chris Wilson
  • [PATCH 11/66] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 12/66] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 13/66] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 14/66] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 15/66] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 16/66] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 17/66] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 18/66] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 19/66] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 20/66] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 21/66] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 22/66] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 23/66] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 24/66] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 25/66] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 26/66] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 27/66] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 28/66] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 29/66] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 30/66] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 31/66] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 32/66] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 33/66] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 34/66] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 35/66] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 36/66] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 37/66] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 38/66] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 39/66] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 40/66] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 41/66] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 42/66] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 43/66] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 44/66] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 45/66] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 46/66] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 47/66] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 48/66] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 49/66] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 50/66] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 51/66] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 52/66] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 53/66] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 54/66] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 55/66] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 56/66] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 57/66] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 58/66] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 59/66] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 60/66] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 61/66] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 62/66] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 63/66] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 64/66] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 65/66] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 66/66] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/66] drm/i915/gt: Assert the kernel context is using the HWSP   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/66] drm/i915/gt: Assert the kernel context is using the HWSP   Patchwork
• [PATCH 1/4] drm/i915/display: Implement HOBL   José Roberto de Souza
  • [PATCH 2/4] DO_NOT_MERGE_IT: drm/i915/display: Enable HOBL by default   José Roberto de Souza
  • [PATCH 3/4] squash: remove second hobl bool   José Roberto de Souza
  • [PATCH 4/4] debug   José Roberto de Souza
  • ✓ Fi.CI.BAT: success for series starting with [1/4] drm/i915/display: Implement HOBL   Patchwork
  • ✓ Fi.CI.IGT: success for series starting with [1/4] drm/i915/display: Implement HOBL   Patchwork
• [PATCH i-g-t] hmm   Chris Wilson
  • ✓ Fi.CI.BAT: success for hmm   Patchwork
  • ✗ Fi.CI.IGT: failure for hmm   Patchwork
• [PATCH 1/7] drm/i915: Reduce i915_request.lock contention for i915_request_wait   Chris Wilson
  • [PATCH 2/7] drm/i915/selftests: Mock the status_page.vma for the kernel_context   Chris Wilson
  • [PATCH 3/7] drm/i915: Be wary of data races when reading the active execlists   Chris Wilson
  • [PATCH 4/7] drm/i915: Remove i915_request.lock requirement for execution callbacks   Chris Wilson
  • [PATCH 5/7] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 6/7] drm/i915/gt: Drop intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 7/7] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [1/7] drm/i915: Reduce i915_request.lock contention for i915_request_wait   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [1/7] drm/i915: Reduce i915_request.lock contention for i915_request_wait   Patchwork
• [PATCH] drm/i915: Add debugs to catch delays during updates   Stanislav Lisovskiy
  • ✗ Fi.CI.BAT: failure for drm/i915: Add debugs to catch delays during updates   Patchwork
• [PATCH v2] drm/i915: Add debugs to catch delays during updates   Stanislav Lisovskiy
  • [PATCH v2] drm/i915: Add debugs to catch delays during updates   Stanislav Lisovskiy
    • ✗ Fi.CI.BAT: failure for drm/i915: Add debugs to catch delays during updates (rev3)   Patchwork
  • ✗ Fi.CI.BAT: failure for drm/i915: Add debugs to catch delays during updates (rev2)   Patchwork
• [PATCH] Add new IGT test to validate crtc Dithering   Bhanuprakash Modem
  • [PATCH] Add new IGT test to validate crtc Dithering   Bhanuprakash Modem
    • ✗ Fi.CI.BUILD: failure for Add new IGT test to validate crtc Dithering (rev2)   Patchwork
    • ✗ Fi.CI.BUILD: failure for Add new IGT test to validate crtc Dithering (rev3)   Patchwork
• [PATCH 1/8] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 2/8] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 3/8] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 4/8] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 5/8] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 6/8] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 7/8] llist   Chris Wilson
  • [PATCH 8/8] signlock   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/8] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/8] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 2/8] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 3/8] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/8] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/8] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/8] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 7/8] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 8/8] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/8] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/8] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
• [PATCH 1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 2/7] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 2/7] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 3/7] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 4/7] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 5/7] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 6/7] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 7/7] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 2/7] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 3/7] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/7] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/7] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/7] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 7/7] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 3/7] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/7] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/7] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/7] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 7/7] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/7] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
• [PATCH 1/6] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 2/6] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 3/6] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 4/6] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 5/6] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 6/6] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/6] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/6] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 2/6] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 3/6] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/6] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/6] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/6] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/6] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/6] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
• [PATCH 1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 2/9] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 2/9] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 3/9] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 4/9] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 5/9] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 6/9] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 7/9] drm/i915/gt: Remove breadcrumb interrupt shadow   Chris Wilson
  • [PATCH 8/9] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 9/9] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 2/9] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 3/9] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/9] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/9] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/9] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 7/9] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 8/9] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 9/9] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 3/9] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/9] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/9] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/9] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 7/9] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 8/9] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 9/9] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 2/9] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 3/9] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/9] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/9] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/9] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 7/9] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 8/9] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 9/9] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 2/9] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 3/9] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/9] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/9] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/9] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 7/9] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 8/9] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 9/9] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
  • [PATCH 1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 2/9] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 3/9] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 4/9] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 5/9] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 6/9] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 7/9] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 8/9] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 9/9] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/9] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
• [PATCH v3] drm/i915: Add debugs to catch delays during updates   Stanislav Lisovskiy
  • [PATCH v3] drm/i915: Add debugs to catch delays during updates   Stanislav Lisovskiy
    • ✗ Fi.CI.BAT: failure for drm/i915: Add debugs to catch delays during updates (rev5)   Patchwork
  • [PATCH v3] drm/i915: Add debugs to catch delays during updates   Stanislav Lisovskiy
    • ✗ Fi.CI.BAT: failure for drm/i915: Add debugs to catch delays during updates (rev6)   Patchwork
• [PATCH i-g-t 00/17] tests/core_hotunplug: Fixes and enhancements   Janusz Krzysztofik
  • [PATCH i-g-t 01/17] tests/core_hotunplug: Use igt_assert_fd()   Janusz Krzysztofik
  • [PATCH i-g-t 02/17] tests/core_hotunplug: Constify dev_bus_addr string   Janusz Krzysztofik
  • [PATCH i-g-t 03/17] tests/core_hotunplug: Consolidate duplicated debug messages   Janusz Krzysztofik
  • [PATCH i-g-t 04/17] tests/core_hotunplug: Assert successful device filter application   Janusz Krzysztofik
  • [PATCH i-g-t 05/17] tests/core_hotunplug: Fix missing newline   Janusz Krzysztofik
  • [PATCH i-g-t 06/17] tests/core_hotunplug: Maintain a single data structure instance   Janusz Krzysztofik
  • [PATCH i-g-t 07/17] tests/core_hotunplug: Pass errors via a data structure field   Janusz Krzysztofik
  • [PATCH i-g-t 08/17] tests/core_hotunplug: Handle device close errors   Janusz Krzysztofik
  • [PATCH i-g-t 09/17] tests/core_hotunplug: Prepare invariant data once per test run   Janusz Krzysztofik
  • [PATCH i-g-t 10/17] tests/core_hotunplug: Skip selectively on sysfs close errors   Janusz Krzysztofik
  • [PATCH i-g-t 11/17] tests/core_hotunplug: Follow failed subtests with health checks   Janusz Krzysztofik
  • [PATCH i-g-t 12/17] tests/core_hotunplug: Fail subtests on device close errors   Janusz Krzysztofik
  • [PATCH i-g-t 13/17] tests/core_hotunplug: Process return values of sysfs operations   Janusz Krzysztofik
  • [PATCH i-g-t 14/17] tests/core_hotunplug: Assert expected device presence/absence   Janusz Krzysztofik
  • [PATCH i-g-t 15/17] tests/core_hotunplug: Explicitly ignore unused return values   Janusz Krzysztofik
  • [PATCH i-g-t 16/17] tests/core_hotunplug: Un-blacklist the test   Janusz Krzysztofik
  • [PATCH i-g-t 17/17] tests/core_hotunplug: Add the test to BAT scope   Janusz Krzysztofik
  • ✗ Fi.CI.BAT: failure for tests/core_hotunplug: Fixes and enhancements   Patchwork
  • [PATCH i-g-t 00/17] tests/core_hotunplug: Fixes and enhancements   Janusz Krzysztofik
    • [PATCH i-g-t 01/17] tests/core_hotunplug: Use igt_assert_fd()   Janusz Krzysztofik
    • [PATCH i-g-t 02/17] tests/core_hotunplug: Constify dev_bus_addr string   Janusz Krzysztofik
    • [PATCH i-g-t 03/17] tests/core_hotunplug: Consolidate duplicated debug messages   Janusz Krzysztofik
    • [PATCH i-g-t 04/17] tests/core_hotunplug: Assert successful device filter application   Janusz Krzysztofik
    • [PATCH i-g-t 05/17] tests/core_hotunplug: Fix missing newline   Janusz Krzysztofik
    • [PATCH i-g-t 06/17] tests/core_hotunplug: Maintain a single data structure instance   Janusz Krzysztofik
    • [PATCH i-g-t 07/17] tests/core_hotunplug: Pass errors via a data structure field   Janusz Krzysztofik
    • [PATCH i-g-t 08/17] tests/core_hotunplug: Handle device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 09/17] tests/core_hotunplug: Prepare invariant data once per test run   Janusz Krzysztofik
    • [PATCH i-g-t 10/17] tests/core_hotunplug: Skip selectively on sysfs close errors   Janusz Krzysztofik
    • [PATCH i-g-t 11/17] tests/core_hotunplug: Follow failed subtests with health checks   Janusz Krzysztofik
    • [PATCH i-g-t 12/17] tests/core_hotunplug: Fail subtests on device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 13/17] tests/core_hotunplug: Process return values of sysfs operations   Janusz Krzysztofik
    • [PATCH i-g-t 14/17] tests/core_hotunplug: Assert expected device presence/absence   Janusz Krzysztofik
    • [PATCH i-g-t 15/17] tests/core_hotunplug: Explicitly ignore unused return values   Janusz Krzysztofik
    • [PATCH i-g-t 16/17] tests/core_hotunplug: Un-blacklist the test   Janusz Krzysztofik
    • [PATCH i-g-t 17/17] tests/core_hotunplug: Add selected subtests to BAT scope   Janusz Krzysztofik
    • ✗ Fi.CI.BAT: failure for tests/core_hotunplug: Fixes and enhancements (rev2)   Patchwork
• [PATCH i-g-t] Add new IGT test to validate crtc Dithering   Bhanuprakash Modem
  • ✓ Fi.CI.BAT: success for Add new IGT test to validate crtc Dithering   Patchwork
  • ✗ Fi.CI.IGT: failure for Add new IGT test to validate crtc Dithering   Patchwork
• [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
  • ✗ Fi.CI.BAT: failure for Expose crtc dither state and connector max bpc via debugfs   Patchwork
  • ✗ Fi.CI.BAT: failure for Expose crtc dither state and connector max bpc via debugfs (rev2)   Patchwork
  • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
  • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev3)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev3)   Patchwork
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev4)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev4)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev5)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev5)   Patchwork
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev6)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✗ Fi.CI.BAT: failure for Expose crtc dither state and connector max bpc via debugfs (rev8)   Patchwork
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev9)   Patchwork
    • ✓ Fi.CI.IGT: success for Expose crtc dither state and connector max bpc via debugfs (rev9)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev10)   Patchwork
    • ✓ Fi.CI.IGT: success for Expose crtc dither state and connector max bpc via debugfs (rev10)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev11)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev11)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev12)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev12)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev13)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev13)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev14)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev14)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev15)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev15)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
  • [PATCH 0/2] Expose crtc dither state and connector max bpc via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev16)   Patchwork
    • [PATCH 1/2] i915/debug: Expose crtc dither state via debugfs   Bhanuprakash Modem
    • [PATCH 2/2] i915/debug: Expose Max BPC info via debugfs   Bhanuprakash Modem
    • ✓ Fi.CI.BAT: success for Expose crtc dither state and connector max bpc via debugfs (rev17)   Patchwork
    • ✗ Fi.CI.IGT: failure for Expose crtc dither state and connector max bpc via debugfs (rev17)   Patchwork
• [PATCH 1/8] HAX sched/core: Paper over the ttwu() race   Chris Wilson
  • [PATCH 2/8] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 3/8] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 4/8] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 5/8] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 6/8] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 7/8] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 8/8] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [1/8] HAX sched/core: Paper over the ttwu() race   Patchwork
• [PATCH v2 0/9] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
  • [PATCH v2 1/9] drm/i915: Refactor postclose driver operation   Janusz Krzysztofik
  • [PATCH v2 2/9] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
  • [PATCH v2 3/9] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
  • [PATCH v2 4/9] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
  • [PATCH v2 5/9] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
  • [PATCH v2 6/9] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
  • [PATCH v2 7/9] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
  • [PATCH v2 8/9] drm/i915: Release memory regions on driver remove   Janusz Krzysztofik
  • [PATCH v2 9/9] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
  • ✗ Fi.CI.BAT: failure for drm/i915: Resolve device hotunplug issues (rev19)   Patchwork
  • [PATCH v2 0/9] drm/i915: Resolve device hotunplug issues   Janusz Krzysztofik
    • [PATCH v2 1/9] drm/i915: Refactor postclose driver operation   Janusz Krzysztofik
    • [PATCH v2 2/9] drm/i915: Drop user contexts on driver remove   Janusz Krzysztofik
    • [PATCH v2 3/9] drm/i915: Drop user requested address spaces on driver remove   Janusz Krzysztofik
    • [PATCH v2 4/9] drm/i915: Drop user created GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 5/9] drm/i915: Drop PRIME exported GEM objects on driver remove   Janusz Krzysztofik
    • [PATCH v2 6/9] drm/i915: Release GT resources on driver remove   Janusz Krzysztofik
    • [PATCH v2 7/9] drm/i915: Move UC firmware cleanup from driver_release to _remove   Janusz Krzysztofik
    • [PATCH v2 8/9] drm/i915: Release memory regions on driver remove   Janusz Krzysztofik
    • [PATCH v2 9/9] drm/i915: Move GGTT cleanup from driver_release to _remove   Janusz Krzysztofik
    • ✗ Fi.CI.BAT: failure for drm/i915: Resolve device hotunplug issues (rev20)   Patchwork
• [PATCH] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank   José Roberto de Souza
  • ✗ Fi.CI.BAT: failure for WIP: drm/i915/psr: Sync frontbuffer modifications with vblank (rev3)   Patchwork
• [PATCH 1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank   José Roberto de Souza
  • ✓ Fi.CI.BAT: success for series starting with [1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank (rev3)   Patchwork
  • [PATCH 1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank   José Roberto de Souza
    • ✓ Fi.CI.BAT: success for series starting with [1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank   Patchwork
    • ✓ Fi.CI.IGT: success for series starting with [1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank (rev3)   Patchwork
  • [PATCH 1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank   José Roberto de Souza
    • [PATCH 2/2] squash   José Roberto de Souza
    • ✗ Fi.CI.BAT: failure for series starting with [1/2] WIP: drm/i915/psr: Sync frontbuffer modifications with vblank   Patchwork
• [PATCH 2/2] wip: drm/i915/psr: Only take display from DC5/DC6   José Roberto de Souza
• [PATCH 2/2] squash   José Roberto de Souza
• [PATCH 1/4] drm/i915/perf: Ensure observation logic is not clock gated   Umesh Nerlige Ramappa
  • ✗ Fi.CI.BAT: failure for series starting with [1/4] drm/i915/perf: Ensure observation logic is not clock gated   Patchwork
  • [PATCH 1/4] drm/i915/perf: Ensure observation logic is not clock gated   Umesh Nerlige Ramappa
    • [PATCH 2/4] drm/i915/perf: Whitelist OA report trigger registers   Umesh Nerlige Ramappa
    • [PATCH 3/4] drm/i915/perf: Whitelist OA counter and buffer registers   Umesh Nerlige Ramappa
    • [PATCH 4/4] drm/i915/perf: Map OA buffer to user space for gen12 performance query   Umesh Nerlige Ramappa
    • ✓ Fi.CI.BAT: success for series starting with [1/4] drm/i915/perf: Ensure observation logic is not clock gated   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/4] drm/i915/perf: Ensure observation logic is not clock gated   Patchwork
  • [PATCH 1/4] drm/i915/perf: Ensure observation logic is not clock gated   Umesh Nerlige Ramappa
    • ✓ Fi.CI.BAT: success for series starting with [1/4] drm/i915/perf: Ensure observation logic is not clock gated   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [1/4] drm/i915/perf: Ensure observation logic is not clock gated   Patchwork
  • [PATCH 1/4] drm/i915/perf: Ensure observation logic is not clock gated   Umesh Nerlige Ramappa
    • [PATCH 2/4] drm/i915/perf: Whitelist OA report trigger registers   Umesh Nerlige Ramappa
    • [PATCH 3/4] drm/i915/perf: Whitelist OA counter and buffer registers   Umesh Nerlige Ramappa
    • [PATCH 4/4] drm/i915/perf: Map OA buffer to user space for gen12 performance query   Umesh Nerlige Ramappa
• [PATCH 2/4] drm/i915/perf: Whitelist OA report trigger registers   Umesh Nerlige Ramappa
  • [PATCH 2/4] drm/i915/perf: Whitelist OA report trigger registers   Umesh Nerlige Ramappa
• [PATCH 3/4] drm/i915/perf: Whitelist OA counter and buffer registers   Umesh Nerlige Ramappa
  • [PATCH 3/4] drm/i915/perf: Whitelist OA counter and buffer registers   Umesh Nerlige Ramappa
• [PATCH 4/4] drm/i915/perf: Map OA buffer to user space for gen12 performance query   Umesh Nerlige Ramappa
  • [PATCH 4/4] drm/i915/perf: Map OA buffer to user space for gen12 performance query   Umesh Nerlige Ramappa
• [PATCH 01/78] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 02/78] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 03/78] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 04/78] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 05/78] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 06/78] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 07/78] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 08/78] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 09/78] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 10/78] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 11/78] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 12/78] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 13/78] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 14/78] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 15/78] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 16/78] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 17/78] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 18/78] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 19/78] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 20/78] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 21/78] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 22/78] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 23/78] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 24/78] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 25/78] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 26/78] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 27/78] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 28/78] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 29/78] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 30/78] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 31/78] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 32/78] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 33/78] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 34/78] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 35/78] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 36/78] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 37/78] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 38/78] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 39/78] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 40/78] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 41/78] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 42/78] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 43/78] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 44/78] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 45/78] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 46/78] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 47/78] heart   Chris Wilson
  • [PATCH 48/78] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 49/78] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 50/78] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 51/78] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 52/78] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 53/78] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 54/78] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 55/78] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 56/78] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 57/78] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 58/78] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 59/78] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 60/78] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 61/78] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 62/78] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 63/78] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 64/78] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 65/78] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 66/78] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 67/78] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 68/78] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 69/78] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 70/78] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 71/78] pin-ve-irq   Chris Wilson
  • [PATCH 72/78] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 73/78] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 74/78] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 75/78] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 76/78] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 77/78] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 78/78] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/78] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/78] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Patchwork
• [PATCH 01/13] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 02/13] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 03/13] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 04/13] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 05/13] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 06/13] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 07/13] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 08/13] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 09/13] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 10/13] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 11/13] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 12/13] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 13/13] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/13] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Patchwork
  • [PATCH 01/13] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 02/13] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 03/13] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 04/13] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 05/13] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 06/13] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 07/13] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 08/13] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 09/13] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 10/13] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 11/13] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 12/13] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 13/13] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/13] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/13] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Patchwork
• [PATCH i-g-t 1/3] i915/gem_exec_hostile:   Chris Wilson
  • [PATCH i-g-t 2/3] hostile-int   Chris Wilson
  • [PATCH i-g-t 3/3] HAX:hostile   Chris Wilson
  • ✗ GitLab.Pipeline: warning for series starting with [1/3] i915/gem_exec_hostile:   Patchwork
  • ✗ Fi.CI.BAT: failure for series starting with [1/3] i915/gem_exec_hostile:   Patchwork
• [PATCH 01/43] drm/i915/gem: Remove disordered per-file request list for throttling   Daniel Vetter
  • [PATCH 02/43] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Daniel Vetter
  • [PATCH 03/43] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Daniel Vetter
    • [PATCH 03/43] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Daniel Vetter
  • [PATCH 04/43] drm/i915/gt: Only transfer the virtual context to the new engine if active   Daniel Vetter
  • [PATCH 05/43] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Daniel Vetter
    • [PATCH 05/43] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Daniel Vetter
  • [PATCH 06/43] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Daniel Vetter
    • [PATCH 06/43] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Daniel Vetter
  • [PATCH 07/43] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Daniel Vetter
  • [PATCH 08/43] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Daniel Vetter
  • [PATCH 09/43] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Daniel Vetter
  • [PATCH 10/43] drm/i915: Drop i915_request.lock serialisation around await_start   Daniel Vetter
  • [PATCH 11/43] drm/i915: Add a couple of missing i915_active_fini()   Daniel Vetter
  • [PATCH 12/43] drm/i915: Skip taking acquire mutex for no ref->active callback   Daniel Vetter
  • [PATCH 13/43] drm/i915: Export a preallocate variant of i915_active_acquire()   Daniel Vetter
  • [PATCH 14/43] drm/i915: Keep the most recently used active-fence upon discard   Daniel Vetter
  • [PATCH 15/43] drm/i915: Make the stale cached active node available for any timeline   Daniel Vetter
  • [PATCH 16/43] drm/i915: Provide a fastpath for waiting on vma bindings   Daniel Vetter
  • [PATCH 17/43] drm/i915: Preallocate stashes for vma page-directories   Daniel Vetter
  • [PATCH 18/43] drm/i915: Switch to object allocations for page directories   Daniel Vetter
  • [PATCH 19/43] drm/i915/gem: Don't drop the timeline lock during execbuf   Daniel Vetter
  • [PATCH 20/43] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Daniel Vetter
  • [PATCH 21/43] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Daniel Vetter
  • [PATCH 22/43] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Daniel Vetter
  • [PATCH 23/43] drm/i915: Add list_for_each_entry_safe_continue_reverse   Daniel Vetter
  • [PATCH 24/43] drm/i915: Always defer fenced work to the worker   Daniel Vetter
  • [PATCH 25/43] drm/i915/gem: Assign context id for async work   Daniel Vetter
  • [PATCH 26/43] drm/i915/gem: Separate the ww_mutex walker into its own list   Daniel Vetter
  • [PATCH 27/43] drm/i915/gem: Asynchronous GTT unbinding   Daniel Vetter
  • [PATCH 01/43] drm/i915/gem: Remove disordered per-file request list for throttling   Daniel Vetter
    • [PATCH 02/43] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Daniel Vetter
    • [PATCH 04/43] drm/i915/gt: Only transfer the virtual context to the new engine if active   Daniel Vetter
    • [PATCH 07/43] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Daniel Vetter
    • [PATCH 08/43] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Daniel Vetter
    • [PATCH 09/43] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Daniel Vetter
    • [PATCH 10/43] drm/i915: Drop i915_request.lock serialisation around await_start   Daniel Vetter
    • [PATCH 11/43] drm/i915: Add a couple of missing i915_active_fini()   Daniel Vetter
    • [PATCH 12/43] drm/i915: Skip taking acquire mutex for no ref->active callback   Daniel Vetter
    • [PATCH 13/43] drm/i915: Export a preallocate variant of i915_active_acquire()   Daniel Vetter
    • [PATCH 14/43] drm/i915: Keep the most recently used active-fence upon discard   Daniel Vetter
    • [PATCH 15/43] drm/i915: Make the stale cached active node available for any timeline   Daniel Vetter
    • [PATCH 16/43] drm/i915: Provide a fastpath for waiting on vma bindings   Daniel Vetter
    • [PATCH 17/43] drm/i915: Preallocate stashes for vma page-directories   Daniel Vetter
    • [PATCH 18/43] drm/i915: Switch to object allocations for page directories   Daniel Vetter
    • [PATCH 19/43] drm/i915/gem: Don't drop the timeline lock during execbuf   Daniel Vetter
    • [PATCH 20/43] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Daniel Vetter
    • [PATCH 21/43] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Daniel Vetter
    • [PATCH 22/43] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Daniel Vetter
    • [PATCH 23/43] drm/i915: Add list_for_each_entry_safe_continue_reverse   Daniel Vetter
    • [PATCH 24/43] drm/i915: Always defer fenced work to the worker   Daniel Vetter
    • [PATCH 25/43] drm/i915/gem: Assign context id for async work   Daniel Vetter
    • [PATCH 26/43] drm/i915/gem: Separate the ww_mutex walker into its own list   Daniel Vetter
    • [PATCH 27/43] drm/i915/gem: Asynchronous GTT unbinding   Daniel Vetter
    • [PATCH 28/43] drm/i915/gem: Bind the fence async for execbuf   Daniel Vetter
    • [PATCH 29/43] drm/i915/gem: Include cmdparser in common execbuf pinning   Daniel Vetter
    • [PATCH 30/43] drm/i915/gem: Include secure batch in common execbuf pinning   Daniel Vetter
    • [PATCH 31/43] drm/i915/gem: Reintroduce multiple passes for reloc processing   Daniel Vetter
    • [PATCH 32/43] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Daniel Vetter
    • [PATCH 33/43] drm/i915/gem: Pull execbuf dma resv under a single critical section   Daniel Vetter
    • [PATCH 34/43] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Daniel Vetter
    • [PATCH 35/43] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Daniel Vetter
    • [PATCH 36/43] drm/i915: Specialise GGTT binding   Daniel Vetter
    • [PATCH 37/43] drm/i915/gt: Acquire backing storage for the context   Daniel Vetter
    • [PATCH 38/43] drm/i915/gt: Push the wait for the context to bound to the request   Daniel Vetter
    • [PATCH 39/43] drm/i915: Remove unused i915_gem_evict_vm()   Daniel Vetter
    • [PATCH 40/43] drm/i915/gt: Decouple completed requests on unwind   Daniel Vetter
    • [PATCH 41/43] drm/i915/gt: Check for a completed last request once   Daniel Vetter
    • [PATCH 42/43] drm/i915/gt: Replace direct submit with direct call to tasklet   Daniel Vetter
    • [PATCH 43/43] annotate new execbuf async work   Daniel Vetter
    • ✗ Fi.CI.BAT: failure for series starting with [01/43] drm/i915/gem: Remove disordered per-file request list for throttling   Patchwork
• [PATCH 01/43] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 02/43] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 03/43] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 04/43] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 05/43] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 06/43] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 07/43] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 08/43] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 09/43] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 10/43] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 11/43] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 12/43] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 13/43] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 14/43] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 15/43] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 16/43] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 17/43] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 18/43] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 19/43] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 20/43] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 21/43] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 22/43] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 23/43] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 24/43] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 25/43] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 26/43] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 27/43] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 28/43] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 29/43] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 30/43] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 31/43] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 32/43] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 33/43] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 34/43] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 35/43] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 36/43] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 37/43] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 38/43] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 39/43] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 40/43] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 41/43] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 42/43] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 43/43] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/43] drm/i915: Remove gen check before calling intel_rps_boost   Patchwork
• [PATCH i-g-t 00/20] tests/core_hotunplug: Fixes and enhancements   Janusz Krzysztofik
  • [PATCH i-g-t 01/20] tests/core_hotunplug: Use igt_assert_fd()   Janusz Krzysztofik
  • [PATCH i-g-t 02/20] tests/core_hotunplug: Constify dev_bus_addr string   Janusz Krzysztofik
    • [PATCH i-g-t 02/20] tests/core_hotunplug: Constify dev_bus_addr string   Janusz Krzysztofik
  • [PATCH i-g-t 03/20] tests/core_hotunplug: Consolidate duplicated debug messages   Janusz Krzysztofik
  • [PATCH i-g-t 04/20] tests/core_hotunplug: Assert successful device filter application   Janusz Krzysztofik
  • [PATCH i-g-t 05/20] tests/core_hotunplug: Fix missing newline   Janusz Krzysztofik
  • [PATCH i-g-t 06/20] tests/core_hotunplug: Maintain a single data structure instance   Janusz Krzysztofik
  • [PATCH i-g-t 07/20] tests/core_hotunplug: Pass errors via a data structure field   Janusz Krzysztofik
  • [PATCH i-g-t 08/20] tests/core_hotunplug: Handle device close errors   Janusz Krzysztofik
  • [PATCH i-g-t 09/20] tests/core_hotunplug: Prepare invariant data once per test run   Janusz Krzysztofik
  • [PATCH i-g-t 10/20] tests/core_hotunplug: Skip selectively on sysfs close errors   Janusz Krzysztofik
  • [PATCH i-g-t 11/20] tests/core_hotunplug: Follow failed subtests with health checks   Janusz Krzysztofik
  • [PATCH i-g-t 12/20] tests/core_hotunplug: Fail subtests on device close errors   Janusz Krzysztofik
  • [PATCH i-g-t 13/20] tests/core_hotunplug: Process return values of sysfs operations   Janusz Krzysztofik
  • [PATCH i-g-t 14/20] tests/core_hotunplug: Assert expected device presence/absence   Janusz Krzysztofik
  • [PATCH i-g-t 15/20] tests/core_hotunplug: Explicitly ignore unused return values   Janusz Krzysztofik
  • [PATCH i-g-t 16/20] tests/core_hotunplug: Turn off autoprobe on bus rescan   Janusz Krzysztofik
  • [PATCH i-g-t 17/20] tests/core_hotunplug: Add unbind-unplug-restore variant   Janusz Krzysztofik
  • [PATCH i-g-t 18/20] tests/core_hotunplug: Add 'lateclose before recover' variants   Janusz Krzysztofik
  • [PATCH i-g-t 19/20] tests/core_hotunplug: Un-blacklist the test   Janusz Krzysztofik
  • [PATCH i-g-t 20/20] tests/core_hotunplug: Add the test to BAT scope   Janusz Krzysztofik
  • ✗ GitLab.Pipeline: warning for tests/core_hotunplug: Fixes and enhancements (rev3)   Patchwork
  • ✗ Fi.CI.BAT: failure for tests/core_hotunplug: Fixes and enhancements (rev3)   Patchwork
  • [PATCH i-g-t 00/20] tests/core_hotunplug: Fixes and enhancements   Janusz Krzysztofik
    • [PATCH i-g-t 01/20] tests/core_hotunplug: Use igt_assert_fd()   Janusz Krzysztofik
    • [PATCH i-g-t 02/20] tests/core_hotunplug: Constify dev_bus_addr string   Janusz Krzysztofik
    • [PATCH i-g-t 03/20] tests/core_hotunplug: Consolidate duplicated debug messages   Janusz Krzysztofik
    • [PATCH i-g-t 04/20] tests/core_hotunplug: Assert successful device filter application   Janusz Krzysztofik
    • [PATCH i-g-t 05/20] tests/core_hotunplug: Fix missing newline   Janusz Krzysztofik
    • [PATCH i-g-t 06/20] tests/core_hotunplug: Maintain a single data structure instance   Janusz Krzysztofik
    • [PATCH i-g-t 07/20] tests/core_hotunplug: Pass errors via a data structure field   Janusz Krzysztofik
    • [PATCH i-g-t 08/20] tests/core_hotunplug: Handle device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 09/20] tests/core_hotunplug: Prepare invariant data once per test run   Janusz Krzysztofik
    • [PATCH i-g-t 10/20] tests/core_hotunplug: Skip selectively on sysfs close errors   Janusz Krzysztofik
    • [PATCH i-g-t 11/20] tests/core_hotunplug: Follow failed subtests with health checks   Janusz Krzysztofik
    • [PATCH i-g-t 12/20] tests/core_hotunplug: Fail subtests on device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 13/20] tests/core_hotunplug: Process return values of sysfs operations   Janusz Krzysztofik
    • [PATCH i-g-t 14/20] tests/core_hotunplug: Assert expected device presence/absence   Janusz Krzysztofik
    • [PATCH i-g-t 15/20] tests/core_hotunplug: Explicitly ignore unused return values   Janusz Krzysztofik
    • [PATCH i-g-t 16/20] tests/core_hotunplug: Reload driver module on device restore   Janusz Krzysztofik
    • [PATCH i-g-t 17/20] tests/core_hotunplug: Add unbind-unplug-restore variant   Janusz Krzysztofik
    • [PATCH i-g-t 18/20] tests/core_hotunplug: Add 'lateclose before recover' variants   Janusz Krzysztofik
    • [PATCH i-g-t 19/20] tests/core_hotunplug: Un-blacklist the test   Janusz Krzysztofik
    • [PATCH i-g-t 20/20] tests/core_hotunplug: Add the test to BAT scope   Janusz Krzysztofik
    • ✓ Fi.CI.BAT: success for tests/core_hotunplug: Fixes and enhancements (rev4)   Patchwork
    • ✗ Fi.CI.IGT: failure for tests/core_hotunplug: Fixes and enhancements (rev4)   Patchwork
  • [PATCH i-g-t 00/20] tests/core_hotunplug: Fixes and enhancements   Janusz Krzysztofik
    • [PATCH i-g-t 01/20] tests/core_hotunplug: Use igt_assert_fd()   Janusz Krzysztofik
    • [PATCH i-g-t 02/20] tests/core_hotunplug: Constify dev_bus_addr string   Janusz Krzysztofik
    • [PATCH i-g-t 03/20] tests/core_hotunplug: Consolidate duplicated debug messages   Janusz Krzysztofik
    • [PATCH i-g-t 04/20] tests/core_hotunplug: Assert successful device filter application   Janusz Krzysztofik
    • [PATCH i-g-t 05/20] tests/core_hotunplug: Fix missing newline   Janusz Krzysztofik
    • [PATCH i-g-t 06/20] tests/core_hotunplug: Maintain a single data structure instance   Janusz Krzysztofik
    • [PATCH i-g-t 07/20] tests/core_hotunplug: Pass errors via a data structure field   Janusz Krzysztofik
    • [PATCH i-g-t 08/20] tests/core_hotunplug: Handle device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 09/20] tests/core_hotunplug: Prepare invariant data once per test run   Janusz Krzysztofik
    • [PATCH i-g-t 10/20] tests/core_hotunplug: Skip selectively on sysfs close errors   Janusz Krzysztofik
    • [PATCH i-g-t 11/20] tests/core_hotunplug: Follow failed subtests with health checks   Janusz Krzysztofik
    • [PATCH i-g-t 12/20] tests/core_hotunplug: Fail subtests on device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 13/20] tests/core_hotunplug: Process return values of sysfs operations   Janusz Krzysztofik
    • [PATCH i-g-t 14/20] tests/core_hotunplug: Assert expected device presence/absence   Janusz Krzysztofik
    • [PATCH i-g-t 15/20] tests/core_hotunplug: Explicitly ignore unused return values   Janusz Krzysztofik
    • [PATCH i-g-t 16/20] tests/core_hotunplug: Reload driver module on device restore   Janusz Krzysztofik
    • [PATCH i-g-t 17/20] tests/core_hotunplug: Add unbind-unplug-restore variant   Janusz Krzysztofik
    • [PATCH i-g-t 18/20] tests/core_hotunplug: Add 'lateclose before recover' variants   Janusz Krzysztofik
    • [PATCH i-g-t 19/20] tests/core_hotunplug: Un-blacklist the test   Janusz Krzysztofik
    • [PATCH i-g-t 20/20] tests/core_hotunplug: Add the test to BAT scope   Janusz Krzysztofik
    • ✗ Fi.CI.BAT: failure for tests/core_hotunplug: Fixes and enhancements (rev5)   Patchwork
  • [PATCH i-g-t 00/20] tests/core_hotunplug: Fixes and enhancements   Janusz Krzysztofik
    • [PATCH i-g-t 01/20] tests/core_hotunplug: Use igt_assert_fd()   Janusz Krzysztofik
    • [PATCH i-g-t 03/20] tests/core_hotunplug: Consolidate duplicated debug messages   Janusz Krzysztofik
    • [PATCH i-g-t 04/20] tests/core_hotunplug: Assert successful device filter application   Janusz Krzysztofik
    • [PATCH i-g-t 05/20] tests/core_hotunplug: Fix missing newline   Janusz Krzysztofik
    • [PATCH i-g-t 06/20] tests/core_hotunplug: Maintain a single data structure instance   Janusz Krzysztofik
    • [PATCH i-g-t 07/20] tests/core_hotunplug: Pass errors via a data structure field   Janusz Krzysztofik
    • [PATCH i-g-t 08/20] tests/core_hotunplug: Handle device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 09/20] tests/core_hotunplug: Prepare invariant data once per test run   Janusz Krzysztofik
    • [PATCH i-g-t 10/20] tests/core_hotunplug: Skip selectively on sysfs close errors   Janusz Krzysztofik
    • [PATCH i-g-t 11/20] tests/core_hotunplug: Follow failed subtests with health checks   Janusz Krzysztofik
    • [PATCH i-g-t 12/20] tests/core_hotunplug: Fail subtests on device close errors   Janusz Krzysztofik
    • [PATCH i-g-t 13/20] tests/core_hotunplug: Process return values of sysfs operations   Janusz Krzysztofik
    • [PATCH i-g-t 14/20] tests/core_hotunplug: Assert expected device presence/absence   Janusz Krzysztofik
    • [PATCH i-g-t 15/20] tests/core_hotunplug: Explicitly ignore unused return values   Janusz Krzysztofik
    • [PATCH i-g-t 16/20] tests/core_hotunplug: Reload driver module on device restore   Janusz Krzysztofik
    • [PATCH i-g-t 17/20] tests/core_hotunplug: Add unbind-unplug-restore variant   Janusz Krzysztofik
    • [PATCH i-g-t 18/20] tests/core_hotunplug: Add 'lateclose before recover' variants   Janusz Krzysztofik
    • [PATCH i-g-t 19/20] tests/core_hotunplug: Un-blacklist the test   Janusz Krzysztofik
    • [PATCH i-g-t 20/20] tests/core_hotunplug: Add the test to BAT scope   Janusz Krzysztofik
    • ✗ Fi.CI.BAT: failure for tests/core_hotunplug: Fixes and enhancements (rev6)   Patchwork
• [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
  • ✗ Fi.CI.BAT: failure for tests/kms_dither: New IGT to validate crtc Dithering   Patchwork
  • ✓ Fi.CI.BAT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev2)   Patchwork
  • ✗ Fi.CI.IGT: failure for tests/kms_dither: New IGT to validate crtc Dithering (rev2)   Patchwork
  • ✓ Fi.CI.BAT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev3)   Patchwork
  • ✓ Fi.CI.IGT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev3)   Patchwork
  • [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
    • [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
      • [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
      • [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
      • [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
      • [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
      • [PATCH i-g-t] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
  • ✓ Fi.CI.BAT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev4)   Patchwork
  • ✓ Fi.CI.BAT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev5)   Patchwork
  • ✗ Fi.CI.IGT: failure for tests/kms_dither: New IGT to validate crtc Dithering (rev4)   Patchwork
  • ✗ Fi.CI.IGT: failure for tests/kms_dither: New IGT to validate crtc Dithering (rev5)   Patchwork
  • ✓ Fi.CI.BAT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev6)   Patchwork
  • ✗ Fi.CI.IGT: failure for tests/kms_dither: New IGT to validate crtc Dithering (rev6)   Patchwork
  • ✓ Fi.CI.BAT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev7)   Patchwork
  • ✓ Fi.CI.IGT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev7)   Patchwork
  • ✓ Fi.CI.BAT: success for tests/kms_dither: New IGT to validate crtc Dithering (rev8)   Patchwork
  • ✗ Fi.CI.IGT: failure for tests/kms_dither: New IGT to validate crtc Dithering (rev8)   Patchwork
• [PATCH 01/44] drm/i915: Remove gen check before calling intel_rps_boost   Daniel Vetter
  • [PATCH 02/44] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Daniel Vetter
  • [PATCH 03/44] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Daniel Vetter
  • [PATCH 04/44] drm/i915: Copy default modparams to mock i915_device   Daniel Vetter
  • [PATCH 05/44] drm/i915/gem: Remove disordered per-file request list for throttling   Daniel Vetter
  • [PATCH 06/44] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Daniel Vetter
  • [PATCH 07/44] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Daniel Vetter
  • [PATCH 08/44] drm/i915/gt: Only transfer the virtual context to the new engine if active   Daniel Vetter
  • [PATCH 09/44] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Daniel Vetter
  • [PATCH 10/44] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Daniel Vetter
  • [PATCH 11/44] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Daniel Vetter
  • [PATCH 12/44] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Daniel Vetter
  • [PATCH 13/44] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Daniel Vetter
  • [PATCH 14/44] drm/i915: Drop i915_request.lock serialisation around await_start   Daniel Vetter
  • [PATCH 15/44] drm/i915: Add a couple of missing i915_active_fini()   Daniel Vetter
  • [PATCH 16/44] drm/i915: Skip taking acquire mutex for no ref->active callback   Daniel Vetter
  • [PATCH 17/44] drm/i915: Export a preallocate variant of i915_active_acquire()   Daniel Vetter
  • [PATCH 18/44] drm/i915: Keep the most recently used active-fence upon discard   Daniel Vetter
  • [PATCH 19/44] drm/i915: Make the stale cached active node available for any timeline   Daniel Vetter
  • [PATCH 20/44] drm/i915: Provide a fastpath for waiting on vma bindings   Daniel Vetter
  • [PATCH 21/44] drm/i915: Preallocate stashes for vma page-directories   Daniel Vetter
  • [PATCH 22/44] drm/i915: Switch to object allocations for page directories   Daniel Vetter
  • [PATCH 23/44] drm/i915/gem: Don't drop the timeline lock during execbuf   Daniel Vetter
  • [PATCH 24/44] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Daniel Vetter
  • [PATCH 25/44] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Daniel Vetter
  • [PATCH 26/44] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Daniel Vetter
  • [PATCH 27/44] drm/i915: Add list_for_each_entry_safe_continue_reverse   Daniel Vetter
  • [PATCH 28/44] drm/i915: Always defer fenced work to the worker   Daniel Vetter
  • [PATCH 29/44] drm/i915/gem: Assign context id for async work   Daniel Vetter
  • [PATCH 30/44] drm/i915/gem: Separate the ww_mutex walker into its own list   Daniel Vetter
  • [PATCH 31/44] drm/i915/gem: Asynchronous GTT unbinding   Daniel Vetter
  • [PATCH 32/44] drm/i915/gem: Bind the fence async for execbuf   Daniel Vetter
  • [PATCH 33/44] drm/i915/gem: Include cmdparser in common execbuf pinning   Daniel Vetter
  • [PATCH 34/44] drm/i915/gem: Include secure batch in common execbuf pinning   Daniel Vetter
  • [PATCH 35/44] drm/i915/gem: Reintroduce multiple passes for reloc processing   Daniel Vetter
  • [PATCH 36/44] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Daniel Vetter
  • [PATCH 37/44] drm/i915/gem: Pull execbuf dma resv under a single critical section   Daniel Vetter
  • [PATCH 38/44] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Daniel Vetter
  • [PATCH 39/44] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Daniel Vetter
  • [PATCH 40/44] drm/i915: Specialise GGTT binding   Daniel Vetter
  • [PATCH 41/44] drm/i915/gt: Acquire backing storage for the context   Daniel Vetter
  • [PATCH 42/44] drm/i915/gt: Push the wait for the context to bound to the request   Daniel Vetter
  • [PATCH 43/44] drm/i915: Remove unused i915_gem_evict_vm()   Daniel Vetter
  • [PATCH 44/44] annotate new execbuf async work   Daniel Vetter
  • [PATCH 01/44] drm/i915: Remove gen check before calling intel_rps_boost   Daniel Vetter
    • [PATCH 02/44] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Daniel Vetter
    • [PATCH 03/44] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Daniel Vetter
    • [PATCH 04/44] drm/i915: Copy default modparams to mock i915_device   Daniel Vetter
    • [PATCH 05/44] drm/i915/gem: Remove disordered per-file request list for throttling   Daniel Vetter
    • [PATCH 06/44] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Daniel Vetter
    • [PATCH 07/44] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Daniel Vetter
    • [PATCH 08/44] drm/i915/gt: Only transfer the virtual context to the new engine if active   Daniel Vetter
    • [PATCH 09/44] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Daniel Vetter
    • [PATCH 10/44] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Daniel Vetter
    • [PATCH 11/44] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Daniel Vetter
    • [PATCH 12/44] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Daniel Vetter
    • [PATCH 13/44] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Daniel Vetter
    • [PATCH 14/44] drm/i915: Drop i915_request.lock serialisation around await_start   Daniel Vetter
    • [PATCH 15/44] drm/i915: Add a couple of missing i915_active_fini()   Daniel Vetter
    • [PATCH 16/44] drm/i915: Skip taking acquire mutex for no ref->active callback   Daniel Vetter
    • [PATCH 17/44] drm/i915: Export a preallocate variant of i915_active_acquire()   Daniel Vetter
    • [PATCH 18/44] drm/i915: Keep the most recently used active-fence upon discard   Daniel Vetter
    • [PATCH 19/44] drm/i915: Make the stale cached active node available for any timeline   Daniel Vetter
    • [PATCH 20/44] drm/i915: Provide a fastpath for waiting on vma bindings   Daniel Vetter
    • [PATCH 21/44] drm/i915: Preallocate stashes for vma page-directories   Daniel Vetter
    • [PATCH 22/44] drm/i915: Switch to object allocations for page directories   Daniel Vetter
    • [PATCH 23/44] drm/i915/gem: Don't drop the timeline lock during execbuf   Daniel Vetter
    • [PATCH 24/44] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Daniel Vetter
    • [PATCH 25/44] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Daniel Vetter
    • [PATCH 26/44] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Daniel Vetter
    • [PATCH 27/44] drm/i915: Add list_for_each_entry_safe_continue_reverse   Daniel Vetter
    • [PATCH 28/44] drm/i915: Always defer fenced work to the worker   Daniel Vetter
    • [PATCH 29/44] drm/i915/gem: Assign context id for async work   Daniel Vetter
    • [PATCH 30/44] drm/i915/gem: Separate the ww_mutex walker into its own list   Daniel Vetter
    • [PATCH 31/44] drm/i915/gem: Asynchronous GTT unbinding   Daniel Vetter
    • [PATCH 32/44] drm/i915/gem: Bind the fence async for execbuf   Daniel Vetter
    • [PATCH 33/44] drm/i915/gem: Include cmdparser in common execbuf pinning   Daniel Vetter
    • [PATCH 34/44] drm/i915/gem: Include secure batch in common execbuf pinning   Daniel Vetter
    • [PATCH 35/44] drm/i915/gem: Reintroduce multiple passes for reloc processing   Daniel Vetter
    • [PATCH 36/44] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Daniel Vetter
    • [PATCH 37/44] drm/i915/gem: Pull execbuf dma resv under a single critical section   Daniel Vetter
    • [PATCH 38/44] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Daniel Vetter
    • [PATCH 39/44] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Daniel Vetter
    • [PATCH 40/44] drm/i915: Specialise GGTT binding   Daniel Vetter
    • [PATCH 41/44] drm/i915/gt: Acquire backing storage for the context   Daniel Vetter
    • [PATCH 42/44] drm/i915/gt: Push the wait for the context to bound to the request   Daniel Vetter
    • [PATCH 43/44] drm/i915: Remove unused i915_gem_evict_vm()   Daniel Vetter
    • [PATCH 44/44] annotate new execbuf async work   Daniel Vetter
    • ✓ Fi.CI.BAT: success for series starting with [01/44] drm/i915: Remove gen check before calling intel_rps_boost   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/44] drm/i915: Remove gen check before calling intel_rps_boost   Patchwork
• [PATCH] findreg   Chris Wilson
  • ✓ Fi.CI.BAT: success for findreg   Patchwork
  • ✗ Fi.CI.IGT: failure for findreg   Patchwork
• [PATCH 01/89] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/89] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 04/89] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 05/89] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 06/89] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 07/89] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 08/89] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 09/89] drm/i915/selftests: Downgrade severity of CS/SRM frequency scaling tests   Chris Wilson
  • [PATCH 10/89] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/89] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/89] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/89] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/89] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/89] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/89] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/89] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/89] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/89] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/89] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/89] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/89] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/89] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/89] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 25/89] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 26/89] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 27/89] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 28/89] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 29/89] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 30/89] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 31/89] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 32/89] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 33/89] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 34/89] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 35/89] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 36/89] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 37/89] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 38/89] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 39/89] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 40/89] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 41/89] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 42/89] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 43/89] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 44/89] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 45/89] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 46/89] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 47/89] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 48/89] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 49/89] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 50/89] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 51/89] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 52/89] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 53/89] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 54/89] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 55/89] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 56/89] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 57/89] heart   Chris Wilson
  • [PATCH 58/89] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 59/89] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 60/89] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 61/89] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 62/89] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 63/89] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 64/89] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 65/89] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 66/89] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 67/89] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 68/89] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 69/89] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 70/89] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 71/89] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 72/89] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 73/89] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 74/89] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 75/89] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 76/89] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 77/89] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 78/89] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 79/89] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 80/89] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 81/89] pin-ve-irq   Chris Wilson
  • [PATCH 82/89] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 83/89] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 84/89] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 85/89] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 86/89] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 87/89] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 88/89] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 89/89] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/89] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Patchwork
• [PATCH 02/89] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
• [PATCH 01/26] Revert "drm/i915/gem: Async GPU relocations only"   Maarten Lankhorst
  • [PATCH 02/26] drm/i915: Revert relocation chaining commits.   Maarten Lankhorst
  • [PATCH 03/26] Revert "drm/i915/gem: Drop relocation slowpath".   Maarten Lankhorst
  • [PATCH 04/26] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Maarten Lankhorst
  • [PATCH 05/26] drm/i915: Remove locking from i915_gem_object_prepare_read/write   Maarten Lankhorst
  • [PATCH 06/26] drm/i915: Parse command buffer earlier in eb_relocate(slow)   Maarten Lankhorst
  • [PATCH 07/26] Revert "drm/i915/gem: Split eb_vma into its own allocation"   Maarten Lankhorst
  • [PATCH 08/26] drm/i915: Use per object locking in execbuf, v12.   Maarten Lankhorst
  • [PATCH 09/26] drm/i915: Use ww locking in intel_renderstate.   Maarten Lankhorst
  • [PATCH 10/26] drm/i915: Add ww context handling to context_barrier_task   Maarten Lankhorst
  • [PATCH 11/26] drm/i915: Nuke arguments to eb_pin_engine   Maarten Lankhorst
  • [PATCH 12/26] drm/i915: Pin engine before pinning all objects, v5.   Maarten Lankhorst
  • [PATCH 13/26] drm/i915: Rework intel_context pinning to do everything outside of pin_mutex   Maarten Lankhorst
  • [PATCH 14/26] drm/i915: Make sure execbuffer always passes ww state to i915_vma_pin.   Maarten Lankhorst
  • [PATCH 15/26] drm/i915: Convert i915_gem_object/client_blt.c to use ww locking as well, v2.   Maarten Lankhorst
  • [PATCH 16/26] drm/i915: Kill last user of intel_context_create_request outside of selftests   Maarten Lankhorst
  • [PATCH 17/26] drm/i915: Convert i915_perf to ww locking as well   Maarten Lankhorst
  • [PATCH 18/26] drm/i915: Dirty hack to fix selftests locking inversion   Maarten Lankhorst
  • [PATCH 19/26] drm/i915/selftests: Fix locking inversion in lrc selftest.   Maarten Lankhorst
  • [PATCH 20/26] drm/i915: Use ww pinning for intel_context_create_request()   Maarten Lankhorst
  • [PATCH 21/26] drm/i915: Move i915_vma_lock in the selftests to avoid lock inversion, v3.   Maarten Lankhorst
  • [PATCH 22/26] drm/i915: Add ww locking to vm_fault_gtt   Maarten Lankhorst
  • [PATCH 23/26] drm/i915: Add ww locking to pin_to_display_plane   Maarten Lankhorst
  • [PATCH 24/26] drm/i915: Ensure we hold the object mutex in pin correctly.   Maarten Lankhorst
  • [PATCH 25/26] drm/i915: Add gem object locking to madvise.   Maarten Lankhorst
  • [PATCH 26/26] drm/i915: Convert i915_gem_object_attach_phys() to ww locking   Maarten Lankhorst
  • ✓ Fi.CI.BAT: success for series starting with [01/26] Revert "drm/i915/gem: Async GPU relocations only"   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/26] Revert "drm/i915/gem: Async GPU relocations only"   Patchwork
• [PATCH 01/90] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/90] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/90] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/90] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/90] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/90] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/90] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/90] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/90] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/90] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/90] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/90] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/90] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/90] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/90] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/90] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/90] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/90] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/90] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/90] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/90] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/90] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/90] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/90] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 25/90] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 26/90] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 27/90] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 28/90] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 29/90] risque-active   Chris Wilson
  • [PATCH 30/90] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 31/90] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 32/90] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 33/90] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 34/90] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 35/90] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 36/90] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 37/90] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 38/90] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 39/90] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 40/90] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 41/90] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 42/90] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 43/90] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 44/90] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 45/90] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 46/90] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 47/90] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 48/90] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 49/90] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 50/90] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 51/90] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 52/90] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 53/90] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 54/90] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 55/90] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 56/90] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 57/90] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 58/90] heart   Chris Wilson
  • [PATCH 59/90] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 60/90] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 61/90] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 62/90] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 63/90] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 64/90] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 65/90] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 66/90] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 67/90] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 68/90] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 69/90] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 70/90] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 71/90] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 72/90] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 73/90] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 74/90] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 75/90] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 76/90] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 77/90] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 78/90] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 79/90] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 80/90] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 81/90] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 82/90] pin-ve-irq   Chris Wilson
  • [PATCH 83/90] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 84/90] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 85/90] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 86/90] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 87/90] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 88/90] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 89/90] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 90/90] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/90] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/91] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/91] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/91] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/91] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/91] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/91] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/91] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/91] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/91] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/91] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/91] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/91] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/91] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/91] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/91] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/91] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/91] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/91] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/91] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/91] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/91] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/91] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/91] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/91] rq-split-boost   Chris Wilson
  • [PATCH 25/91] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 26/91] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 27/91] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 28/91] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 29/91] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 30/91] risque-active   Chris Wilson
  • [PATCH 31/91] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 32/91] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 33/91] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 34/91] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 35/91] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 36/91] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 37/91] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 38/91] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 39/91] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 40/91] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 41/91] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 42/91] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 43/91] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 44/91] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 45/91] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 46/91] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 47/91] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 48/91] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 49/91] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 50/91] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 51/91] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 52/91] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 53/91] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 54/91] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 55/91] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 56/91] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 57/91] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 58/91] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 59/91] heart   Chris Wilson
  • [PATCH 60/91] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 61/91] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 62/91] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 63/91] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 64/91] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 65/91] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 66/91] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 67/91] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 68/91] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 69/91] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 70/91] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 71/91] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 72/91] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 73/91] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 74/91] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 75/91] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 76/91] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 77/91] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 78/91] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 79/91] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 80/91] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 81/91] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 82/91] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 83/91] pin-ve-irq   Chris Wilson
  • [PATCH 84/91] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 85/91] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 86/91] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 87/91] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 88/91] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 89/91] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 90/91] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 91/91] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/91] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/24] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/24] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/24] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/24] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/24] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/24] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/24] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/24] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/24] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/24] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/24] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/24] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/24] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/24] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/24] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/24] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/24] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/24] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/24] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/24] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/24] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/24] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/24] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/24] rq-split-boost   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/24] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/25] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/25] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/25] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/25] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/25] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/25] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/25] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/25] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/25] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/25] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/25] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/25] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/25] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/25] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/25] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/25] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/25] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/25] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/25] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/25] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/25] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/25] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 23/25] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 24/25] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 25/25] rq-split-boost   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/25] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • [PATCH 01/25] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/25] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 03/25] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/25] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/25] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/25] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/25] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/25] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/25] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/25] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/25] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/25] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/25] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/25] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/25] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 16/25] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 17/25] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 18/25] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 19/25] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 20/25] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 21/25] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 22/25] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 23/25] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 24/25] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 25/25] rq-split-boost   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/25] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/22] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/22] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/22] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/22] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/22] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/22] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/22] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/22] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/22] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/22] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/22] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/22] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/22] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/22] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/22] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/22] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/22] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/22] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/22] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/22] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/22] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 03/22] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/22] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/22] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/22] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/22] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/22] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/22] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/22] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/22] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/22] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/22] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/22] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/22] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 16/22] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 17/22] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 18/22] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 19/22] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 20/22] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 21/22] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 22/22] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • [PATCH 01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/22] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 03/22] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/22] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/22] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/22] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/22] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/22] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/22] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/22] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/22] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/22] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/22] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/22] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/22] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 16/22] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 17/22] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 18/22] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 19/22] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 20/22] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 21/22] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 22/22] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • [PATCH 01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/22] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 03/22] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/22] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/22] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/22] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/22] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/22] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/22] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/22] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/22] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/22] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/22] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/22] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/22] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 16/22] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 17/22] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 18/22] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 19/22] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 20/22] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 21/22] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 22/22] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • [PATCH 01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/22] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 03/22] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/22] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/22] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/22] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/22] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/22] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/22] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/22] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/22] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/22] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/22] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/22] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/22] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 16/22] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 17/22] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 18/22] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 19/22] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 20/22] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 21/22] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 22/22] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/22] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 03/22] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
• [PATCH 01/29] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/29] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/29] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/29] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/29] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/29] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/29] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/29] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/29] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/29] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/29] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/29] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/29] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/29] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/29] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/29] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/29] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/29] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/29] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/29] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/29] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/29] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 23/29] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 24/29] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 25/29] rq-split-boost   Chris Wilson
  • [PATCH 26/29] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 27/29] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 28/29] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 29/29] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/29] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/31] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/31] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/31] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/31] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/31] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/31] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/31] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/31] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/31] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/31] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/31] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/31] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/31] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/31] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/31] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/31] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/31] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/31] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/31] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/31] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/31] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/31] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 23/31] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 24/31] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 25/31] rq-split-boost   Chris Wilson
  • [PATCH 26/31] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 27/31] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 28/31] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 29/31] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 30/31] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 31/31] risque-active   Chris Wilson
  • [PATCH 01/31] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/31] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 03/31] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/31] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/31] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/31] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/31] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/31] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/31] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/31] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/31] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/31] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/31] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/31] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/31] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 16/31] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 17/31] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 18/31] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 19/31] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 20/31] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 21/31] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 22/31] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 23/31] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 24/31] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 25/31] rq-split-boost   Chris Wilson
    • [PATCH 26/31] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 27/31] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 28/31] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 29/31] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 30/31] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 31/31] risque-active   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/31] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • ✗ Fi.CI.BAT: failure for series starting with [01/31] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/21] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/21] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/21] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/21] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/21] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/21] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/21] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/21] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/21] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/21] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/21] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/21] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/21] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/21] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/21] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/21] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/21] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/21] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/21] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/21] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/21] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/21] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/93] locking/lockdep: Fix overflow in presentation of average lock-time   Chris Wilson
  • [PATCH 02/93] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 03/93] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/93] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 05/93] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 06/93] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 07/93] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 08/93] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 09/93] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 10/93] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 11/93] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 12/93] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/93] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/93] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 15/93] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 16/93] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 17/93] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 18/93] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 19/93] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 20/93] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 21/93] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 22/93] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 23/93] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 24/93] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 25/93] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 26/93] rq-split-boost   Chris Wilson
  • [PATCH 27/93] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 28/93] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 29/93] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 30/93] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 31/93] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 32/93] risque-active   Chris Wilson
  • [PATCH 33/93] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 34/93] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 35/93] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 36/93] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 37/93] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 38/93] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 39/93] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 40/93] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 41/93] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 42/93] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 43/93] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 44/93] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 45/93] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 46/93] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 47/93] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 48/93] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 49/93] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 50/93] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 51/93] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 52/93] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 53/93] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 54/93] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 55/93] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 56/93] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 57/93] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 58/93] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 59/93] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 60/93] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 61/93] heart   Chris Wilson
  • [PATCH 62/93] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 63/93] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 64/93] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 65/93] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 66/93] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 67/93] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 68/93] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 69/93] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 70/93] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 71/93] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 72/93] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 73/93] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 74/93] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 75/93] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 76/93] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 77/93] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 78/93] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 79/93] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 80/93] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 81/93] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 82/93] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 83/93] drm/i915: Move saturated workload detection to the GT   Chris Wilson
  • [PATCH 84/93] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 85/93] pin-ve-irq   Chris Wilson
  • [PATCH 86/93] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 87/93] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 88/93] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 89/93] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 90/93] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 91/93] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 92/93] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 93/93] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/93] locking/lockdep: Fix overflow in presentation of average lock-time   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/93] locking/lockdep: Fix overflow in presentation of average lock-time   Patchwork
• [PATCH 01/23] locking/lockdep: Fix overflow in presentation of average lock-time   Chris Wilson
  • [PATCH 02/23] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 03/23] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/23] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 05/23] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 06/23] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 07/23] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 08/23] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 09/23] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 10/23] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 11/23] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 12/23] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/23] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/23] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 15/23] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 16/23] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 17/23] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 18/23] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 19/23] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 20/23] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 21/23] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 22/23] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 23/23] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/23] locking/lockdep: Fix overflow in presentation of average lock-time   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/23] locking/lockdep: Fix overflow in presentation of average lock-time   Patchwork
• [PATCH 1/3] Revert "drm/i915/gem: Async GPU relocations only"   Daniel Vetter
  • [PATCH 2/3] drm/i915: Revert relocation chaining commits.   Daniel Vetter
  • [PATCH 3/3] Revert "drm/i915/gem: Drop relocation slowpath".   Daniel Vetter
  • ✓ Fi.CI.BAT: success for series starting with [1/3] Revert "drm/i915/gem: Async GPU relocations only"   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [1/3] Revert "drm/i915/gem: Async GPU relocations only"   Patchwork
• [PATCH 01/36] locking/lockdep: Fix overflow in presentation of average lock-time   Chris Wilson
  • [PATCH 02/36] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 03/36] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/36] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 05/36] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 06/36] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 07/36] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 08/36] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 09/36] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 10/36] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 11/36] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 12/36] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/36] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/36] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 15/36] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 16/36] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 17/36] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 18/36] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 19/36] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 20/36] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 21/36] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 22/36] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 23/36] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 24/36] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 25/36] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 26/36] rq-split-boost   Chris Wilson
  • [PATCH 27/36] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 28/36] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 29/36] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 30/36] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 31/36] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 32/36] risque-active   Chris Wilson
  • [PATCH 33/36] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 34/36] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 35/36] drm/i915: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 36/36] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/36] locking/lockdep: Fix overflow in presentation of average lock-time   Patchwork
• [PATCH 01/95] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/95] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/95] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/95] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/95] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/95] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/95] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/95] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/95] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/95] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/95] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/95] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/95] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/95] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/95] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/95] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/95] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/95] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/95] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/95] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/95] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/95] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 23/95] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 24/95] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 25/95] rq-split-boost   Chris Wilson
  • [PATCH 26/95] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 27/95] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 28/95] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 29/95] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 30/95] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 31/95] risque-active   Chris Wilson
  • [PATCH 32/95] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 33/95] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 34/95] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 35/95] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 36/95] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 37/95] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 38/95] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 39/95] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 40/95] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 41/95] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 42/95] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 43/95] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 44/95] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 45/95] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 46/95] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 47/95] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 48/95] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 49/95] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 50/95] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 51/95] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 52/95] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 53/95] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 54/95] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 55/95] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 56/95] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 57/95] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 58/95] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 59/95] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 60/95] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 61/95] heart   Chris Wilson
  • [PATCH 62/95] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 63/95] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 64/95] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 65/95] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 66/95] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 67/95] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 68/95] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 69/95] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 70/95] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 71/95] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 72/95] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 73/95] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 74/95] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 75/95] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 76/95] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 77/95] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 78/95] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 79/95] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 80/95] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 81/95] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 82/95] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 83/95] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 84/95] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 85/95] pin-ve-irq   Chris Wilson
  • [PATCH 86/95] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 87/95] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 88/95] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 89/95] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 90/95] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 91/95] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 92/95] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 93/95] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • [PATCH 94/95] perf   Chris Wilson
  • [PATCH 95/95] locking/lockdep: Fix overflow in presentation of average lock-time   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/95] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/93] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/93] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/93] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/93] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/93] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/93] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/93] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/93] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/93] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/93] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/93] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/93] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/93] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/93] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/93] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 16/93] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 17/93] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 18/93] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/93] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 20/93] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 21/93] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 22/93] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 23/93] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 24/93] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 25/93] rq-split-boost   Chris Wilson
  • [PATCH 26/93] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 27/93] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 28/93] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 29/93] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 30/93] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 31/93] risque-active   Chris Wilson
  • [PATCH 32/93] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 33/93] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 34/93] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 35/93] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 36/93] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 37/93] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 38/93] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 39/93] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 40/93] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 41/93] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 42/93] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 43/93] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 44/93] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 45/93] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 46/93] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 47/93] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 48/93] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 49/93] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 50/93] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 51/93] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 52/93] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 53/93] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 54/93] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 55/93] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 56/93] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 57/93] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 58/93] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 59/93] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 60/93] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 61/93] heart   Chris Wilson
  • [PATCH 62/93] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 63/93] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 64/93] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 65/93] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 66/93] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 67/93] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 68/93] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 69/93] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 70/93] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 71/93] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 72/93] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 73/93] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 74/93] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 75/93] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 76/93] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 77/93] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 78/93] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 79/93] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 80/93] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 81/93] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 82/93] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 83/93] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 84/93] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 85/93] pin-ve-irq   Chris Wilson
  • [PATCH 86/93] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 87/93] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 88/93] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 89/93] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 90/93] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 91/93] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 92/93] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 93/93] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/93] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/93] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH 01/94] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 02/94] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 02/94] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 03/94] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/94] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 05/94] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 06/94] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 07/94] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 08/94] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 09/94] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 10/94] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 11/94] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 12/94] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 13/94] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 14/94] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 15/94] st-active-flush   Chris Wilson
  • [PATCH 16/94] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 17/94] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 18/94] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 19/94] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 20/94] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 21/94] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 22/94] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 23/94] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 24/94] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 25/94] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 26/94] rq-split-boost   Chris Wilson
  • [PATCH 27/94] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 28/94] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 29/94] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 30/94] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 31/94] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 32/94] risque-active   Chris Wilson
  • [PATCH 33/94] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 34/94] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 35/94] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 36/94] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 37/94] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 38/94] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 39/94] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 40/94] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 41/94] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 42/94] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 43/94] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 44/94] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 45/94] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 46/94] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 47/94] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 48/94] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 49/94] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 50/94] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 51/94] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 52/94] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 53/94] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 54/94] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 55/94] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 56/94] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 57/94] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 58/94] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 59/94] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 60/94] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 61/94] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 62/94] heart   Chris Wilson
  • [PATCH 63/94] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 64/94] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 65/94] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 66/94] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 67/94] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 68/94] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 69/94] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 70/94] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 71/94] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 72/94] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 73/94] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 74/94] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 75/94] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 76/94] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 77/94] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 78/94] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 79/94] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 80/94] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 81/94] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 82/94] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 83/94] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 84/94] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 85/94] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 86/94] pin-ve-irq   Chris Wilson
  • [PATCH 87/94] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 88/94] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 89/94] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 90/94] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 91/94] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 92/94] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 93/94] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 94/94] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/94] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • [PATCH 01/94] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/94] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 03/94] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/94] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/94] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/94] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/94] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/94] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/94] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/94] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/94] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/94] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/94] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/94] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/94] st-active-flush   Chris Wilson
    • [PATCH 16/94] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 17/94] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 18/94] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 19/94] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 20/94] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 21/94] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 22/94] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 23/94] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 24/94] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 25/94] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 26/94] rq-split-boost   Chris Wilson
    • [PATCH 27/94] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 28/94] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 29/94] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 30/94] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 31/94] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 32/94] risque-active   Chris Wilson
    • [PATCH 33/94] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 34/94] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 35/94] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 36/94] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 37/94] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 38/94] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 39/94] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 40/94] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 41/94] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 42/94] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 43/94] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 44/94] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 45/94] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 46/94] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 47/94] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 48/94] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 49/94] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 50/94] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 51/94] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 52/94] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 53/94] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 54/94] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 55/94] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 56/94] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 57/94] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 58/94] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 59/94] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 60/94] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
    • [PATCH 61/94] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 62/94] heart   Chris Wilson
    • [PATCH 63/94] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
    • [PATCH 64/94] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 65/94] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 66/94] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 67/94] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 68/94] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 69/94] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 70/94] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 71/94] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 72/94] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 73/94] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 74/94] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 75/94] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 76/94] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 77/94] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 78/94] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 79/94] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 80/94] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 81/94] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 82/94] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 83/94] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 84/94] drm/i915: Move saturated workload detection back to the context   Chris Wilson
    • [PATCH 85/94] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 86/94] pin-ve-irq   Chris Wilson
    • [PATCH 87/94] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 88/94] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 89/94] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 90/94] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 91/94] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 92/94] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 93/94] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 94/94] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/94] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/94] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
  • [PATCH 01/94] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 03/94] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 04/94] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 05/94] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 06/94] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 07/94] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 08/94] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 09/94] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 10/94] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 11/94] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/94] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/94] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 14/94] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 15/94] st-active-flush   Chris Wilson
    • [PATCH 16/94] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 17/94] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 18/94] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 19/94] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 20/94] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 21/94] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 22/94] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 23/94] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 24/94] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 25/94] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 26/94] rq-split-boost   Chris Wilson
    • [PATCH 27/94] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 28/94] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 29/94] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 30/94] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 31/94] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 32/94] risque-active   Chris Wilson
    • [PATCH 33/94] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 34/94] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 35/94] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 36/94] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 37/94] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 38/94] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 39/94] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 40/94] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 41/94] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 42/94] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 43/94] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 44/94] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 45/94] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 46/94] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 47/94] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 48/94] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 49/94] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 50/94] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 51/94] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 52/94] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 53/94] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 54/94] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 55/94] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 56/94] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 57/94] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 58/94] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 59/94] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 60/94] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
    • [PATCH 61/94] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 62/94] heart   Chris Wilson
    • [PATCH 63/94] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
    • [PATCH 64/94] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 65/94] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 66/94] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 67/94] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 68/94] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 69/94] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 70/94] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 71/94] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 72/94] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 73/94] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 74/94] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 75/94] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 76/94] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 77/94] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 78/94] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 79/94] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 80/94] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 81/94] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 82/94] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 83/94] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 84/94] drm/i915: Move saturated workload detection back to the context   Chris Wilson
    • [PATCH 85/94] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 86/94] pin-ve-irq   Chris Wilson
    • [PATCH 87/94] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 88/94] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 89/94] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 90/94] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 91/94] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 92/94] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 93/94] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 94/94] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/94] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Patchwork
• [PATCH] ilk-irq   Chris Wilson
  • ✓ Fi.CI.BAT: success for ilk-irq   Patchwork
  • ✗ Fi.CI.IGT: failure for ilk-irq   Patchwork
• [PATCH 01/95] ilk-irq   Chris Wilson
  • [PATCH 02/95] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 02/95] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 03/95] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/95] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 05/95] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 06/95] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 07/95] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 08/95] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 09/95] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 10/95] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 11/95] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 12/95] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/95] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/95] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 15/95] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 16/95] st-active-flush   Chris Wilson
  • [PATCH 17/95] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 18/95] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 19/95] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 20/95] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 21/95] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 22/95] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 23/95] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 24/95] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 25/95] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 26/95] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 27/95] rq-split-boost   Chris Wilson
  • [PATCH 28/95] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 29/95] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 30/95] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 31/95] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 32/95] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 33/95] risque-active   Chris Wilson
  • [PATCH 34/95] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 35/95] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 36/95] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 37/95] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 38/95] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 39/95] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 40/95] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 41/95] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 42/95] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 43/95] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 44/95] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 45/95] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 46/95] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 47/95] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 48/95] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 49/95] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 50/95] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 51/95] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 52/95] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 53/95] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 54/95] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 55/95] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 56/95] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 57/95] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 58/95] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 59/95] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 60/95] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 61/95] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 62/95] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 63/95] heart   Chris Wilson
  • [PATCH 64/95] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 65/95] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 66/95] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 67/95] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 68/95] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 69/95] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 70/95] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 71/95] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 72/95] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 73/95] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 74/95] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 75/95] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 76/95] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 77/95] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 78/95] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 79/95] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 80/95] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 81/95] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 82/95] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 83/95] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 84/95] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 85/95] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 86/95] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 87/95] pin-ve-irq   Chris Wilson
  • [PATCH 88/95] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 89/95] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 90/95] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 91/95] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 92/95] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 93/95] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 94/95] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 95/95] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/95] ilk-irq   Patchwork
• [PATCH 01/95] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 03/95] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/95] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 05/95] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 06/95] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 07/95] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 08/95] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 09/95] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 10/95] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 11/95] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 12/95] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/95] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/95] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 15/95] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 16/95] st-active-flush   Chris Wilson
  • [PATCH 17/95] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 18/95] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 19/95] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 20/95] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 21/95] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 22/95] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 23/95] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 24/95] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 25/95] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 26/95] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 27/95] rq-split-boost   Chris Wilson
  • [PATCH 28/95] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 29/95] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 30/95] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 31/95] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 32/95] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 33/95] risque-active   Chris Wilson
  • [PATCH 34/95] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 35/95] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 36/95] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 37/95] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 38/95] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 39/95] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 40/95] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 41/95] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 42/95] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 43/95] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 44/95] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 45/95] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 46/95] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 47/95] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 48/95] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 49/95] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 50/95] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 51/95] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 52/95] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 53/95] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 54/95] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 55/95] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 56/95] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 57/95] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 58/95] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 59/95] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 60/95] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 61/95] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 62/95] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 63/95] heart   Chris Wilson
  • [PATCH 64/95] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 65/95] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 66/95] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 67/95] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 68/95] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 69/95] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 70/95] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 71/95] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 72/95] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 73/95] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 74/95] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 75/95] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 76/95] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 77/95] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 78/95] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 79/95] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 80/95] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 81/95] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 82/95] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 83/95] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 84/95] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 85/95] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 86/95] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 87/95] pin-ve-irq   Chris Wilson
  • [PATCH 88/95] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 89/95] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 90/95] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 91/95] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 92/95] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 93/95] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 94/95] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 95/95] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/95] drm/i915: Reduce register reads around GT interrupts   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/95] drm/i915: Reduce register reads around GT interrupts   Patchwork
• [PATCH 1/5] drm/i915: Preallocate stashes for vma page-directories   Matthew Auld
  • [PATCH 2/5] drm/i915: Switch to object allocations for page directories   Matthew Auld
  • [PATCH 3/5] flat   Matthew Auld
  • [PATCH 4/5] vaddr   Matthew Auld
  • [PATCH 5/5] lmem-pd   Matthew Auld
  • ✗ Fi.CI.BAT: failure for series starting with [1/5] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 001/101] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 002/101] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 003/101] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 004/101] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 005/101] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 006/101] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 007/101] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 008/101] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 009/101] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 010/101] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 011/101] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 012/101] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 013/101] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 014/101] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 015/101] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 016/101] drm/i915/selftests: Flush the active barriers before asserting   Chris Wilson
  • [PATCH 017/101] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 018/101] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 019/101] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 020/101] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 021/101] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 022/101] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 023/101] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 024/101] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 025/101] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 026/101] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 027/101] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 028/101] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 029/101] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 030/101] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 031/101] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 032/101] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 033/101] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 034/101] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 035/101] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 036/101] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 037/101] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 038/101] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 039/101] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 040/101] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 041/101] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 042/101] pin-inplace   Chris Wilson
  • [PATCH 043/101] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 044/101] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 045/101] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 046/101] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 047/101] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 048/101] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 049/101] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 050/101] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 051/101] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 052/101] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 053/101] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 054/101] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 055/101] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 056/101] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 057/101] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 058/101] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 059/101] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 060/101] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 061/101] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()   Chris Wilson
  • [PATCH 062/101] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 063/101] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 064/101] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 065/101] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 066/101] heart   Chris Wilson
  • [PATCH 067/101] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 068/101] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 069/101] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 070/101] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 071/101] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 072/101] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 073/101] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 074/101] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 075/101] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 076/101] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 077/101] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 078/101] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 079/101] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 080/101] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 081/101] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 082/101] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 083/101] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 084/101] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 085/101] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 086/101] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 087/101] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 088/101] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 089/101] drm/i915/gt: Defer the kmem_cache_free() until after the HW submit   Chris Wilson
  • [PATCH 090/101] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 091/101] drm/i915/gt: Pin all the siblings' breadcrumb irq on the virtual engine   Chris Wilson
  • [PATCH 092/101] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 093/101] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 094/101] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 095/101] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 096/101] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 097/101] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 098/101] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 099/101] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • [PATCH 100/101] drm/i915/gt: Name the individual i915_address_space classes   Chris Wilson
  • [PATCH 101/101] vma-sync-before-lock   Chris Wilson
  • [PATCH 001/101] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
    • [PATCH 002/101] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
    • [PATCH 003/101] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
    • [PATCH 004/101] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
    • [PATCH 005/101] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
    • [PATCH 006/101] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
    • [PATCH 007/101] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 008/101] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 009/101] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 010/101] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 011/101] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
    • [PATCH 012/101] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 013/101] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 014/101] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
    • [PATCH 015/101] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
    • [PATCH 016/101] drm/i915/selftests: Flush the active barriers before asserting   Chris Wilson
    • [PATCH 017/101] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 018/101] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 019/101] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 020/101] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 021/101] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 022/101] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 023/101] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 024/101] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 025/101] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 026/101] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 027/101] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
    • [PATCH 028/101] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 029/101] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 030/101] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 031/101] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 032/101] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 033/101] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
    • [PATCH 034/101] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 035/101] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 036/101] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 037/101] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 038/101] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 039/101] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 040/101] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 041/101] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 042/101] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
    • [PATCH 043/101] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 044/101] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 045/101] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 046/101] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 047/101] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 048/101] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 049/101] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 050/101] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 051/101] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 052/101] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 053/101] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 054/101] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 055/101] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 056/101] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 057/101] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 058/101] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 059/101] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 060/101] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
    • [PATCH 061/101] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()   Chris Wilson
    • [PATCH 062/101] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 063/101] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 064/101] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
    • [PATCH 065/101] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 066/101] heart   Chris Wilson
    • [PATCH 067/101] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
    • [PATCH 068/101] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 069/101] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 070/101] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 071/101] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 072/101] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 073/101] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 074/101] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 075/101] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 076/101] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 077/101] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 078/101] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 079/101] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 080/101] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 081/101] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 082/101] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 083/101] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 084/101] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 085/101] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 086/101] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 087/101] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 088/101] drm/i915: Move saturated workload detection back to the context   Chris Wilson
    • [PATCH 089/101] drm/i915/gt: Defer the kmem_cache_free() until after the HW submit   Chris Wilson
    • [PATCH 090/101] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
    • [PATCH 091/101] drm/i915/gt: Pin all the siblings' breadcrumb irq on the virtual engine   Chris Wilson
    • [PATCH 092/101] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
    • [PATCH 093/101] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
    • [PATCH 094/101] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
    • [PATCH 095/101] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
    • [PATCH 096/101] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
    • [PATCH 097/101] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
    • [PATCH 098/101] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
    • [PATCH 099/101] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
    • [PATCH 100/101] drm/i915/gt: Name the individual i915_address_space classes   Chris Wilson
    • [PATCH 101/101] drm/i915: Sync before taking vm->mutex on i915_vma_release   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [001/101] drm/i915: Reduce register reads around GT interrupts   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [001/101] drm/i915: Reduce register reads around GT interrupts   Patchwork
• [PATCH] drm/i915/display: Ensure that ret is always initialized in icl_combo_phy_verify_state   José Roberto de Souza
  • ✗ Fi.CI.BAT: failure for drm/i915/display: Ensure that ret is always initialized in icl_combo_phy_verify_state   Patchwork
• [PATCH 01/89] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 02/89] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 03/89] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 04/89] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • [PATCH 05/89] drm/i915/selftests: Add compiler paranoia for checking HWSP values   Chris Wilson
  • [PATCH 06/89] drm/i915: Filter wake_flags passed to default_wake_function   Chris Wilson
  • [PATCH 07/89] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 08/89] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 09/89] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 10/89] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 11/89] drm/i915: Remove gen check before calling intel_rps_boost   Chris Wilson
  • [PATCH 12/89] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/89] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/89] drm/i915: Copy default modparams to mock i915_device   Chris Wilson
  • [PATCH 15/89] drm/i915/gem: Remove disordered per-file request list for throttling   Chris Wilson
  • [PATCH 16/89] drm/i915/selftests: Flush the active barriers before asserting   Chris Wilson
  • [PATCH 17/89] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 18/89] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 19/89] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 20/89] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 21/89] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 22/89] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 23/89] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 24/89] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 25/89] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 26/89] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 27/89] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 28/89] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 29/89] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 30/89] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 31/89] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 32/89] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 33/89] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 34/89] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 35/89] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 36/89] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 37/89] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 38/89] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 39/89] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 40/89] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 41/89] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 42/89] pin-inplace   Chris Wilson
  • [PATCH 43/89] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 44/89] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 45/89] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 46/89] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 47/89] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 48/89] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 49/89] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 50/89] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 51/89] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 52/89] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 53/89] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 54/89] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 55/89] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 56/89] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 57/89] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 58/89] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 59/89] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 60/89] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 61/89] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()   Chris Wilson
  • [PATCH 62/89] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 63/89] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 64/89] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 65/89] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 66/89] heart   Chris Wilson
  • [PATCH 67/89] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 68/89] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 69/89] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 70/89] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 71/89] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 72/89] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 73/89] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 74/89] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 75/89] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 76/89] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 77/89] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 78/89] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 79/89] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 80/89] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 81/89] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 82/89] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 83/89] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 84/89] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 85/89] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 86/89] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 87/89] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 88/89] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 89/89] drm/i915/gt: Defer the kmem_cache_free() until after the HW submit   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/89] drm/i915: Reduce register reads around GT interrupts   Patchwork
• [PATCH 1/4] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 2/4] drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool   Chris Wilson
  • [PATCH 3/4] drm/i915/gt: Disable preparser around xcs invalidations on tgl   Chris Wilson
  • [PATCH 4/4] drm/i915/gt: Stall around xcs invalidations on tgl   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [1/4] drm/i915: Reduce register reads around GT interrupts   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [1/4] drm/i915: Reduce register reads around GT interrupts   Patchwork
• [PATCH 00/27] drm/i915/gem: Rework i915_gem_object locking to use ww locks.   Maarten Lankhorst
  • [PATCH 01/27] Revert "drm/i915/gem: Async GPU relocations only"   Maarten Lankhorst
  • [PATCH 03/27] Revert "drm/i915/gem: Drop relocation slowpath".   Maarten Lankhorst
  • [PATCH 04/27] Revert "drm/i915/gem: Split eb_vma into its own allocation"   Maarten Lankhorst
  • [PATCH 05/27] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Maarten Lankhorst
  • [PATCH 06/27] drm/i915: Remove locking from i915_gem_object_prepare_read/write   Maarten Lankhorst
  • [PATCH 07/27] drm/i915: Parse command buffer earlier in eb_relocate(slow)   Maarten Lankhorst
  • [PATCH 08/27] drm/i915: Use per object locking in execbuf, v12.   Maarten Lankhorst
  • [PATCH 09/27] drm/i915: Use ww locking in intel_renderstate.   Maarten Lankhorst
  • [PATCH 10/27] drm/i915: Add ww context handling to context_barrier_task   Maarten Lankhorst
  • [PATCH 11/27] drm/i915: Nuke arguments to eb_pin_engine   Maarten Lankhorst
  • [PATCH 12/27] drm/i915: Pin engine before pinning all objects, v5.   Maarten Lankhorst
  • [PATCH 13/27] drm/i915: Rework intel_context pinning to do everything outside of pin_mutex   Maarten Lankhorst
  • [PATCH 14/27] drm/i915: Make sure execbuffer always passes ww state to i915_vma_pin.   Maarten Lankhorst
  • [PATCH 15/27] drm/i915: Convert i915_gem_object/client_blt.c to use ww locking as well, v2.   Maarten Lankhorst
  • [PATCH 16/27] drm/i915: Kill last user of intel_context_create_request outside of selftests   Maarten Lankhorst
  • [PATCH 17/27] drm/i915: Convert i915_perf to ww locking as well   Maarten Lankhorst
  • [PATCH 18/27] drm/i915: Dirty hack to fix selftests locking inversion   Maarten Lankhorst
  • [PATCH 19/27] drm/i915/selftests: Fix locking inversion in lrc selftest.   Maarten Lankhorst
  • [PATCH 20/27] drm/i915: Use ww pinning for intel_context_create_request()   Maarten Lankhorst
  • [PATCH 21/27] drm/i915: Move i915_vma_lock in the selftests to avoid lock inversion, v3.   Maarten Lankhorst
  • [PATCH 22/27] drm/i915: Add ww locking to vm_fault_gtt   Maarten Lankhorst
  • [PATCH 23/27] drm/i915: Add ww locking to pin_to_display_plane   Maarten Lankhorst
  • [PATCH 24/27] drm/i915: Ensure we hold the object mutex in pin correctly.   Maarten Lankhorst
  • [PATCH 25/27] drm/i915: Add gem object locking to madvise.   Maarten Lankhorst
  • [PATCH 26/27] drm/i915: Convert i915_gem_object_attach_phys() to ww locking   Maarten Lankhorst
  • [RFC PATCH 27/27] drm/i915: Fix userptr so we do not have to worry about obj->mm.lock   Maarten Lankhorst
  • ✓ Fi.CI.BAT: success for drm/i915/gem: Rework i915_gem_object locking to use ww locks.   Patchwork
  • ✗ Fi.CI.IGT: failure for drm/i915/gem: Rework i915_gem_object locking to use ww locks.   Patchwork
• [PATCH 02/27] drm/i915: Revert relocation chaining commits.   Maarten Lankhorst
• [PATCH 1/8] drm/i915: Preallocate stashes for vma page-directories   Matthew Auld
  • [PATCH 2/8] drm/i915: Switch to object allocations for page directories   Matthew Auld
  • [PATCH 3/8] flat   Matthew Auld
  • [PATCH 4/8] vaddr   Matthew Auld
  • [PATCH 5/8] lmem-pd   Matthew Auld
  • [PATCH 6/8] flush   Matthew Auld
  • [PATCH 7/8] min   Matthew Auld
  • [PATCH 8/8] scratch   Matthew Auld
  • ✓ Fi.CI.BAT: success for series starting with [1/8] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • ✓ Fi.CI.IGT: success for series starting with [1/8] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/45] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
  • [PATCH 02/45] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 03/45] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 04/45] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 05/45] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 06/45] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 07/45] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 08/45] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 09/45] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 10/45] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 11/45] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 12/45] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/45] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/45] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 15/45] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 16/45] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 17/45] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 18/45] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 19/45] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 20/45] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 21/45] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 22/45] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/45] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/45] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 25/45] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 26/45] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 27/45] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 28/45] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 29/45] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 30/45] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 31/45] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 32/45] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 33/45] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 34/45] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 35/45] fence-inplace   Chris Wilson
  • [PATCH 36/45] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 37/45] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 38/45] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 39/45] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 40/45] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 41/45] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 42/45] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 43/45] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 44/45] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 45/45] cache1   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/45] drm/i915/display: Check for an LPSP encoder before dereferencing   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/45] drm/i915/display: Check for an LPSP encoder before dereferencing   Patchwork
• [PATCH 1/3] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 2/3] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 3/3] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [1/3] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [1/3] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH i-g-t] DO_NOT_MERGE: Sanity check for CRC mismatch   Bhanuprakash Modem
  • ✓ Fi.CI.BAT: success for DO_NOT_MERGE: Sanity check for CRC mismatch   Patchwork
  • ✗ Fi.CI.IGT: failure for DO_NOT_MERGE: Sanity check for CRC mismatch   Patchwork
• [PATCH 01/84] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/84] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/84] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 04/84] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
  • [PATCH 05/84] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 06/84] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 07/84] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 08/84] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 09/84] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 10/84] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 11/84] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 12/84] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/84] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/84] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 15/84] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 16/84] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 17/84] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 18/84] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 19/84] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 20/84] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 21/84] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 22/84] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/84] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/84] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 25/84] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 26/84] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 27/84] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 28/84] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 29/84] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 30/84] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 31/84] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 32/84] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 33/84] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 34/84] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 35/84] fence-inplace   Chris Wilson
  • [PATCH 36/84] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 37/84] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 38/84] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 39/84] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 40/84] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 41/84] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 42/84] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 43/84] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 44/84] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 45/84] cache1   Chris Wilson
  • [PATCH 46/84] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 47/84] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 48/84] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 49/84] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 50/84] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 51/84] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 52/84] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 53/84] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 54/84] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 55/84] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()   Chris Wilson
  • [PATCH 56/84] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 57/84] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 58/84] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 59/84] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 60/84] heart   Chris Wilson
  • [PATCH 61/84] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 62/84] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 63/84] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 64/84] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 65/84] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 66/84] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 67/84] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 68/84] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 69/84] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 70/84] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 71/84] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 72/84] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 73/84] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 74/84] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 75/84] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 76/84] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 77/84] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 78/84] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 79/84] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 80/84] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 81/84] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 82/84] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 83/84] drm/i915/gt: Defer the kmem_cache_free() until after the HW submit   Chris Wilson
  • [PATCH 84/84] drm/i915: Prune empty priolists   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/84] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/84] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/84] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/84] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 04/84] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
    • [PATCH 05/84] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 06/84] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 07/84] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 08/84] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 09/84] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 10/84] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
    • [PATCH 11/84] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 12/84] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 13/84] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 14/84] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 15/84] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 16/84] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 17/84] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 18/84] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 19/84] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 20/84] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 21/84] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 22/84] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 23/84] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 24/84] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
    • [PATCH 25/84] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
    • [PATCH 26/84] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 27/84] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 28/84] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 29/84] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 30/84] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 31/84] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 32/84] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 33/84] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 34/84] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
    • [PATCH 35/84] fence-inplace   Chris Wilson
    • [PATCH 36/84] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 37/84] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 38/84] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 39/84] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 40/84] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 41/84] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 42/84] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 43/84] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 44/84] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 45/84] cache1   Chris Wilson
    • [PATCH 46/84] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 47/84] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 48/84] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 49/84] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 50/84] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 51/84] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 52/84] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 53/84] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • [PATCH 54/84] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
    • [PATCH 55/84] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()   Chris Wilson
    • [PATCH 56/84] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
    • [PATCH 57/84] drm/i915/gt: Check for a completed last request once   Chris Wilson
    • [PATCH 58/84] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
    • [PATCH 59/84] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
    • [PATCH 60/84] heart   Chris Wilson
    • [PATCH 61/84] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
    • [PATCH 62/84] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
    • [PATCH 63/84] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
    • [PATCH 64/84] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
    • [PATCH 65/84] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
    • [PATCH 66/84] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
    • [PATCH 67/84] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
    • [PATCH 68/84] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
    • [PATCH 69/84] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
    • [PATCH 70/84] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
    • [PATCH 71/84] drm/i915: Lift waiter/signaler iterators   Chris Wilson
    • [PATCH 72/84] drm/i915: Strip out internal priorities   Chris Wilson
    • [PATCH 73/84] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
    • [PATCH 74/84] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
    • [PATCH 75/84] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
    • [PATCH 76/84] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
    • [PATCH 77/84] drm/i915: Restructure priority inheritance   Chris Wilson
    • [PATCH 78/84] drm/i915/gt: Remove timeslice suppression   Chris Wilson
    • [PATCH 79/84] drm/i915: Fair low-latency scheduling   Chris Wilson
    • [PATCH 80/84] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
    • [PATCH 81/84] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
    • [PATCH 82/84] drm/i915: Move saturated workload detection back to the context   Chris Wilson
    • [PATCH 83/84] drm/i915/gt: Defer the kmem_cache_free() until after the HW submit   Chris Wilson
    • [PATCH 84/84] drm/i915: Prune empty priolists   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/84] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/45] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/45] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/45] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 04/45] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
  • [PATCH 05/45] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 06/45] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 07/45] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 08/45] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 09/45] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 10/45] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 11/45] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 12/45] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/45] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/45] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 15/45] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 16/45] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 17/45] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 18/45] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 19/45] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 20/45] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 21/45] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 22/45] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/45] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/45] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 25/45] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 26/45] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 27/45] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 28/45] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 29/45] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 30/45] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 31/45] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 32/45] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 33/45] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 34/45] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 35/45] fence-inplace   Chris Wilson
  • [PATCH 36/45] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 37/45] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 38/45] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 39/45] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 40/45] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 41/45] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 42/45] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 43/45] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 44/45] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 45/45] cache1   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/45] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/45] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/45] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/45] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/45] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 04/45] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
    • [PATCH 05/45] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 06/45] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 07/45] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 08/45] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 09/45] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 10/45] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
    • [PATCH 11/45] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 12/45] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 13/45] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 14/45] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 15/45] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 16/45] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 17/45] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 18/45] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 19/45] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 20/45] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 21/45] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 22/45] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 23/45] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 24/45] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
    • [PATCH 25/45] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
    • [PATCH 26/45] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 27/45] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 28/45] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 29/45] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 30/45] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 31/45] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 32/45] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 33/45] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 34/45] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
    • [PATCH 35/45] fence-inplace   Chris Wilson
    • [PATCH 36/45] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 37/45] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 38/45] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 39/45] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 40/45] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 41/45] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 42/45] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 43/45] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 44/45] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 45/45] cache1   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/45] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 01/73] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/73] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 02/73] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
• [PATCH 03/73] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
• [PATCH 01/53] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/53] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/53] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 03/53] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 05/53] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 06/53] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 07/53] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 08/53] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 09/53] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 10/53] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 11/53] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 12/53] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/53] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/53] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 15/53] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 16/53] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 17/53] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 18/53] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 19/53] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 20/53] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 20/53] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 21/53] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 22/53] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/53] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/53] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
    • [PATCH 24/53] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 25/53] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 26/53] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 27/53] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 28/53] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 29/53] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 30/53] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 30/53] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 31/53] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 32/53] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 33/53] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 34/53] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 35/53] fence-inplace   Chris Wilson
  • [PATCH 36/53] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 37/53] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 38/53] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 39/53] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 39/53] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 40/53] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 41/53] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 42/53] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 43/53] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 44/53] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 45/53] cache1   Chris Wilson
  • [PATCH 46/53] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 47/53] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • [PATCH 47/53] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 48/53] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 49/53] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 50/53] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 51/53] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 52/53] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 53/53] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 01/53] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/53] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 04/53] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
    • [PATCH 05/53] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 06/53] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 07/53] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 08/53] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 09/53] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 10/53] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
    • [PATCH 11/53] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 12/53] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 13/53] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 14/53] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 15/53] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 16/53] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 17/53] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 18/53] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 19/53] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 21/53] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 22/53] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 23/53] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 25/53] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
    • [PATCH 26/53] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 27/53] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 28/53] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 29/53] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 31/53] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 32/53] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 33/53] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 34/53] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
    • [PATCH 35/53] fence-inplace   Chris Wilson
    • [PATCH 36/53] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 37/53] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 38/53] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 40/53] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 41/53] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 42/53] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 43/53] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 44/53] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 45/53] cache1   Chris Wilson
    • [PATCH 46/53] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 48/53] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
    • [PATCH 49/53] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
    • [PATCH 50/53] drm/i915: Specialise GGTT binding   Chris Wilson
    • [PATCH 51/53] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
    • [PATCH 52/53] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
    • [PATCH 53/53] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/53] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • ✗ Fi.CI.BAT: failure for series starting with [01/53] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 04/73] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
• [PATCH 05/73] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
• [PATCH 06/73] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
• [PATCH 04/53] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
• [PATCH 07/73] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
• [PATCH 08/73] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
• [PATCH 09/73] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
• [PATCH 10/73] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
• [PATCH 11/73] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
• [PATCH 12/73] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
• [PATCH 13/73] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
• [PATCH 14/73] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
• [PATCH 15/73] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
• [PATCH 16/73] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
• [PATCH 17/73] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
• [PATCH 18/73] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
• [PATCH 19/73] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
• [PATCH 20/73] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
• [PATCH 21/73] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
• [PATCH 22/73] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
• [PATCH 23/73] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
• [PATCH 24/73] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
• [PATCH 25/73] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
• [PATCH 26/73] drm: Restore driver.preclose() for all to use   Chris Wilson
• [PATCH 27/73] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
• [PATCH 28/73] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
• [PATCH 29/73] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
• [PATCH 30/73] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
• [PATCH 31/73] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
• [PATCH 32/73] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
• [PATCH 33/73] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
• [PATCH 34/73] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
• [PATCH 35/73] fence-inplace   Chris Wilson
• [PATCH 36/73] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
• [PATCH 37/73] drm/i915: Always defer fenced work to the worker   Chris Wilson
• [PATCH 38/73] drm/i915/gem: Assign context id for async work   Chris Wilson
• [PATCH 39/73] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
• [PATCH 40/73] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
• [PATCH 41/73] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
• [PATCH 42/73] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
• [PATCH 43/73] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
• [PATCH 44/73] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
• [PATCH 45/73] cache1   Chris Wilson
• [PATCH 46/73] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
• [PATCH 47/73] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
• [PATCH 48/73] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
• [PATCH 49/73] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
• [PATCH 50/73] drm/i915: Specialise GGTT binding   Chris Wilson
• [PATCH 51/73] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
• [PATCH 52/73] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
• [PATCH 53/73] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
• [PATCH 54/73] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
• [PATCH 55/73] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()   Chris Wilson
• [PATCH 56/73] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
• [PATCH 57/73] drm/i915/gt: Check for a completed last request once   Chris Wilson
• [PATCH 58/73] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
• [PATCH 59/73] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
• [PATCH 60/73] heart   Chris Wilson
• [PATCH 61/73] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
• [PATCH 62/73] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
• [PATCH 63/73] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
• [PATCH 64/73] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
• [PATCH 65/73] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
• [PATCH 66/73] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
• [PATCH 67/73] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
• [PATCH 68/73] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
• [PATCH 69/73] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
• [PATCH 70/73] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
• [PATCH 71/73] drm/i915: Lift waiter/signaler iterators   Chris Wilson
• [PATCH 72/73] drm/i915: Strip out internal priorities   Chris Wilson
• [PATCH 73/73] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
• [PATCH 01/47] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 07/47] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 08/47] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 12/47] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/47] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 17/47] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 19/47] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 20/47] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 21/47] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 22/47] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/47] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/47] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 26/47] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 28/47] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 29/47] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 30/47] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 31/47] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 32/47] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 34/47] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 35/47] fence-inplace   Chris Wilson
  • [PATCH 36/47] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 37/47] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 38/47] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 39/47] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 41/47] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 42/47] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 43/47] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 44/47] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 45/47] cache1   Chris Wilson
  • [PATCH 46/47] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 47/47] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • ✗ Fi.CI.BAT: failure for series starting with [01/47] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/47] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/47] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/47] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 04/47] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
    • [PATCH 05/47] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
    • [PATCH 06/47] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 07/47] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 08/47] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 09/47] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 10/47] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
    • [PATCH 11/47] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 12/47] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 13/47] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 14/47] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 15/47] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 16/47] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 17/47] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 18/47] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 19/47] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 20/47] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 21/47] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 22/47] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 23/47] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 24/47] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
    • [PATCH 25/47] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
    • [PATCH 26/47] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 27/47] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 28/47] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 29/47] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 30/47] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 31/47] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 32/47] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 33/47] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 34/47] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
    • [PATCH 35/47] fence-inplace   Chris Wilson
    • [PATCH 36/47] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 37/47] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 38/47] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 39/47] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 40/47] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 41/47] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 42/47] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 43/47] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 44/47] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 45/47] cache1   Chris Wilson
    • [PATCH 46/47] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 47/47] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✗ Fi.CI.BAT: failure for series starting with [01/47] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 02/47] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
• [PATCH 03/47] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
• [PATCH 04/47] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
• [PATCH 05/47] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
• [PATCH 06/47] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
• [PATCH 09/47] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
• [PATCH 10/47] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
• [PATCH 11/47] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
• [PATCH 14/47] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
• [PATCH 15/47] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
• [PATCH 16/47] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
• [PATCH 18/47] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
• [PATCH 25/47] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
• [PATCH 27/47] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
• [PATCH 33/47] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
• [PATCH 40/47] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
• [PATCH 1/6] drm/i915: Allow removal of whitelist register and refactor   Umesh Nerlige Ramappa
  • [PATCH 2/6] drm/i915/selftests: Clear flags when using wa->reg for comparison   Umesh Nerlige Ramappa
  • [PATCH 3/6] drm/i915/perf: Ensure observation logic is not clock gated   Umesh Nerlige Ramappa
  • [PATCH 4/6] drm/i915/perf: Whitelist OA report trigger registers   Umesh Nerlige Ramappa
  • [PATCH 5/6] drm/i915/perf: Whitelist OA counter and buffer registers   Umesh Nerlige Ramappa
  • [PATCH 6/6] drm/i915/perf: Map OA buffer to user space for gen12 performance query   Umesh Nerlige Ramappa
  • ✓ Fi.CI.BAT: success for series starting with [1/6] drm/i915: Allow removal of whitelist register and refactor   Patchwork
  • ✓ Fi.CI.IGT: success for series starting with [1/6] drm/i915: Allow removal of whitelist register and refactor   Patchwork
• [PATCH 01/46] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 02/46] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 03/46] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 04/46] drm/i915/display: Check for an LPSP encoder before dereferencing   Chris Wilson
  • [PATCH 05/46] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 06/46] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 07/46] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 08/46] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 09/46] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 10/46] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 11/46] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 12/46] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 13/46] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 14/46] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 15/46] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 16/46] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 17/46] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 18/46] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 19/46] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 20/46] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 21/46] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 22/46] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 23/46] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 24/46] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 25/46] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 26/46] drm: Restore driver.preclose() for all to use   Chris Wilson
  • [PATCH 27/46] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
  • [PATCH 28/46] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
  • [PATCH 29/46] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
  • [PATCH 30/46] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 31/46] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 32/46] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 33/46] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 34/46] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 35/46] fence-inplace   Chris Wilson
  • [PATCH 36/46] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 37/46] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 38/46] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 39/46] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 40/46] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 41/46] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 42/46] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 43/46] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 44/46] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 45/46] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 46/46] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/46] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/46] drm/i915: Preallocate stashes for vma page-directories   Patchwork
  • [PATCH 01/46] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
    • [PATCH 02/46] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
    • [PATCH 03/46] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
    • [PATCH 05/46] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
    • [PATCH 06/46] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
    • [PATCH 07/46] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
    • [PATCH 08/46] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
    • [PATCH 09/46] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
    • [PATCH 10/46] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
    • [PATCH 11/46] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
    • [PATCH 12/46] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
    • [PATCH 13/46] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
    • [PATCH 14/46] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
    • [PATCH 15/46] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
    • [PATCH 16/46] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
    • [PATCH 17/46] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
    • [PATCH 18/46] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
    • [PATCH 19/46] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
    • [PATCH 20/46] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
    • [PATCH 21/46] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
    • [PATCH 22/46] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
    • [PATCH 23/46] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
    • [PATCH 24/46] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
    • [PATCH 25/46] drm: Restore driver.preclose() for all to use   Chris Wilson
    • [PATCH 26/46] drm/i915/gem: Delay tracking the GEM context until it is registered   Chris Wilson
    • [PATCH 27/46] drm/i915/gem: Move context decoupling from postclose to preclose   Chris Wilson
    • [PATCH 28/46] drm/i915/gem: Serialise debugfs i915_gem_objects with ctx->mutex   Chris Wilson
    • [PATCH 29/46] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
    • [PATCH 30/46] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
    • [PATCH 31/46] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
    • [PATCH 32/46] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
    • [PATCH 33/46] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
    • [PATCH 34/46] fence-inplace   Chris Wilson
    • [PATCH 35/46] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
    • [PATCH 36/46] drm/i915: Always defer fenced work to the worker   Chris Wilson
    • [PATCH 37/46] drm/i915/gem: Assign context id for async work   Chris Wilson
    • [PATCH 38/46] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
    • [PATCH 39/46] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
    • [PATCH 40/46] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
    • [PATCH 41/46] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
    • [PATCH 42/46] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
    • [PATCH 43/46] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
    • [PATCH 44/46] cache1   Chris Wilson
    • [PATCH 45/46] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
    • [PATCH 46/46] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
    • ✓ Fi.CI.BAT: success for series starting with [01/46] drm/i915: Preallocate stashes for vma page-directories   Patchwork
    • ✗ Fi.CI.IGT: failure for series starting with [01/46] drm/i915: Preallocate stashes for vma page-directories   Patchwork
• [PATCH 04/46] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
• [PATCH 01/49] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 02/49] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 03/49] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 04/49] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 05/49] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 06/49] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 07/49] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 08/49] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 09/49] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 10/49] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 11/49] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 12/49] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 13/49] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 14/49] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 15/49] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 16/49] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 17/49] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 18/49] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 19/49] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 20/49] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 21/49] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 22/49] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 23/49] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 24/49] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 25/49] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 26/49] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 27/49] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 28/49] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 29/49] fence-inplace   Chris Wilson
  • [PATCH 30/49] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 31/49] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 32/49] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 33/49] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 34/49] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 35/49] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 36/49] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 37/49] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 38/49] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 39/49] cache1   Chris Wilson
  • [PATCH 40/49] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 41/49] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 42/49] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 43/49] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 44/49] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 45/49] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 46/49] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 47/49] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 48/49] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 49/49] drm/i915/gem: Ignore domain flushing for an unwritten object   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/49] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/49] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
• [PATCH 01/50] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 02/50] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 03/50] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
  • [PATCH 04/50] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 05/50] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 06/50] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 07/50] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 08/50] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 09/50] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 10/50] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 11/50] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 12/50] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 13/50] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 14/50] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 15/50] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 16/50] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 17/50] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 18/50] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 19/50] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 20/50] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 21/50] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 22/50] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 23/50] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 24/50] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 25/50] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 26/50] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 27/50] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 28/50] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 29/50] fence-inplace   Chris Wilson
  • [PATCH 30/50] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 31/50] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 32/50] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 33/50] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 34/50] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 35/50] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 36/50] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 37/50] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 38/50] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 39/50] cache1   Chris Wilson
  • [PATCH 40/50] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 41/50] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 42/50] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 43/50] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 44/50] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 45/50] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 46/50] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 47/50] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 48/50] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 49/50] drm/i915/gem: Ignore domain flushing for an unwritten object   Chris Wilson
  • [PATCH 50/50] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/50] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/50] drm/i915: Add a couple of missing i915_active_fini()   Patchwork
• [PATCH 01/94] drm/i915/gt: Support multiple pinned timelines   Chris Wilson
  • [PATCH 02/94] drm/i915/gt: Pull release of node->age under the spinlock   Chris Wilson
  • [PATCH 03/94] drm/i915: Add a couple of missing i915_active_fini()   Chris Wilson
  • [PATCH 04/94] drm/i915: Skip taking acquire mutex for no ref->active callback   Chris Wilson
  • [PATCH 06/94] drm/i915: Keep the most recently used active-fence upon discard   Chris Wilson
  • [PATCH 07/94] drm/i915: Make the stale cached active node available for any timeline   Chris Wilson
  • [PATCH 08/94] drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()   Chris Wilson
  • [PATCH 09/94] drm/i915: Provide a fastpath for waiting on vma bindings   Chris Wilson
  • [PATCH 10/94] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 11/94] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 12/94] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 13/94] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 14/94] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 15/94] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 16/94] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 17/94] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 18/94] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 19/94] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 20/94] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 21/94] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 22/94] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 23/94] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 24/94] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 25/94] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 26/94] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 27/94] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 28/94] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 29/94] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 30/94] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 31/94] fence-inplace   Chris Wilson
  • [PATCH 32/94] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 33/94] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 34/94] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 35/94] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 36/94] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 37/94] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 38/94] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 39/94] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 40/94] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 41/94] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 42/94] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 43/94] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 44/94] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 45/94] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 46/94] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 47/94] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 48/94] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 49/94] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 50/94] drm/i915/gem: Ignore domain flushing for an unwritten object   Chris Wilson
  • [PATCH 51/94] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
  • [PATCH 52/94] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()   Chris Wilson
  • [PATCH 53/94] drm/i915/gt: Decouple completed requests on unwind   Chris Wilson
  • [PATCH 54/94] drm/i915/gt: Check for a completed last request once   Chris Wilson
  • [PATCH 55/94] drm/i915/gt: Refactor heartbeat request construction and submission   Chris Wilson
  • [PATCH 56/94] drm/i915/gt: Replace direct submit with direct call to tasklet   Chris Wilson
  • [PATCH 57/94] heart   Chris Wilson
  • [PATCH 58/94] drm/i915/gt: Free stale request on destroying the virtual engine   Chris Wilson
  • [PATCH 59/94] drm/i915/gt: Use virtual_engine during execlists_dequeue   Chris Wilson
  • [PATCH 60/94] drm/i915/gt: Decouple inflight virtual engines   Chris Wilson
  • [PATCH 61/94] drm/i915/gt: Defer schedule_out until after the next dequeue   Chris Wilson
  • [PATCH 62/94] drm/i915/gt: Resubmit the virtual engine on schedule-out   Chris Wilson
  • [PATCH 63/94] drm/i915/gt: Simplify virtual engine handling for execlists_hold()   Chris Wilson
  • [PATCH 64/94] drm/i915/gt: ce->inflight updates are now serialised   Chris Wilson
  • [PATCH 65/94] drm/i915/gt: Drop atomic for engine->fw_active tracking   Chris Wilson
  • [PATCH 66/94] drm/i915/gt: Extract busy-stats for ring-scheduler   Chris Wilson
  • [PATCH 67/94] drm/i915/gt: Convert stats.active to plain unsigned int   Chris Wilson
  • [PATCH 68/94] drm/i915: Lift waiter/signaler iterators   Chris Wilson
  • [PATCH 69/94] drm/i915: Strip out internal priorities   Chris Wilson
  • [PATCH 70/94] drm/i915: Remove I915_USER_PRIORITY_SHIFT   Chris Wilson
  • [PATCH 71/94] drm/i915: Replace engine->schedule() with a known request operation   Chris Wilson
  • [PATCH 72/94] drm/i915/gt: Do not suspend bonded requests if one hangs   Chris Wilson
  • [PATCH 73/94] drm/i915: Teach the i915_dependency to use a double-lock   Chris Wilson
  • [PATCH 74/94] drm/i915: Restructure priority inheritance   Chris Wilson
  • [PATCH 75/94] drm/i915/gt: Remove timeslice suppression   Chris Wilson
  • [PATCH 76/94] drm/i915: Fair low-latency scheduling   Chris Wilson
  • [PATCH 77/94] drm/i915/gt: Specify a deadline for the heartbeat   Chris Wilson
  • [PATCH 78/94] drm/i915: Replace the priority boosting for the display with a deadline   Chris Wilson
  • [PATCH 79/94] drm/i915: Move saturated workload detection back to the context   Chris Wilson
  • [PATCH 80/94] drm/i915/gt: Defer the kmem_cache_free() until after the HW submit   Chris Wilson
  • [PATCH 81/94] drm/i915: Prune empty priolists   Chris Wilson
  • [PATCH 82/94] drm/i915: Reduce register reads around GT interrupts   Chris Wilson
  • [PATCH 83/94] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq"   Chris Wilson
  • [PATCH 84/94] drm/i915/gt: Pin all the siblings' breadcrumb irq on the virtual engine   Chris Wilson
  • [PATCH 85/94] drm/i915/gt: Couple tasklet scheduling for all CS interrupts   Chris Wilson
  • [PATCH 86/94] drm/i915/gt: Support creation of 'internal' rings   Chris Wilson
  • [PATCH 87/94] drm/i915/gt: Use client timeline address for seqno writes   Chris Wilson
  • [PATCH 88/94] drm/i915/gt: Infrastructure for ring scheduling   Chris Wilson
  • [PATCH 89/94] drm/i915/gt: Implement ring scheduler for gen6/7   Chris Wilson
  • [PATCH 90/94] drm/i915/gt: Enable ring scheduling for gen6/7   Chris Wilson
  • [PATCH 91/94] drm/i915/gem: Remove timeline nesting from snb relocs   Chris Wilson
  • [PATCH 92/94] drm/i915/gt: Another tweak for flushing the tasklets   Chris Wilson
  • [PATCH 93/94] drm/i915/gt: Name the individual i915_address_space classes   Chris Wilson
  • [PATCH 94/94] drm/i915: Sync before taking vm->mutex on i915_vma_release   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/94] drm/i915/gt: Support multiple pinned timelines   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/94] drm/i915/gt: Support multiple pinned timelines   Patchwork
• [PATCH 05/94] drm/i915: Export a preallocate variant of i915_active_acquire()   Chris Wilson
• [PATCH 01/44] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 03/44] drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs   Chris Wilson
  • [PATCH 04/44] drm/i915/gt: Replace intel_engine_transfer_stale_breadcrumbs   Chris Wilson
  • [PATCH 05/44] drm/i915/gt: Only transfer the virtual context to the new engine if active   Chris Wilson
  • [PATCH 06/44] drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs   Chris Wilson
  • [PATCH 07/44] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 08/44] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 09/44] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 10/44] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 11/44] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 12/44] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 13/44] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 14/44] drm/i915: Preallocate stashes for vma page-directories   Chris Wilson
  • [PATCH 15/44] drm/i915/gt: Switch to object allocations for page directories   Chris Wilson
  • [PATCH 16/44] drm/i915/gt: Shrink i915_page_directory's slab bucket   Chris Wilson
  • [PATCH 17/44] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 18/44] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 19/44] drm/i915/gem: Rename the list of relocations to reloc_list   Chris Wilson
  • [PATCH 20/44] drm/i915/gem: Move the 'cached' info to i915_execbuffer   Chris Wilson
  • [PATCH 21/44] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 22/44] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 23/44] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 24/44] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 25/44] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 26/44] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 27/44] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 28/44] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 29/44] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 30/44] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 31/44] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 32/44] drm/i915/gem: Manage ppGTT placement bias (starting offset) explicitly   Chris Wilson
  • [PATCH 33/44] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 34/44] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 35/44] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 36/44] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 37/44] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 38/44] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 39/44] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 40/44] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 41/44] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 42/44] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 43/44] drm/i915/gem: Ignore domain flushing for an unwritten object   Chris Wilson
  • [PATCH 44/44] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/44] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Patchwork
  • ✗ Fi.CI.IGT: failure for series starting with [01/44] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Patchwork
• [PATCH 02/44] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
• [PATCH 01/28] Revert "drm/i915/gem: Async GPU relocations only"   Maarten Lankhorst
  • [PATCH 02/28] drm/i915: Revert relocation chaining commits.   Maarten Lankhorst
  • [PATCH 04/28] Revert "drm/i915/gem: Split eb_vma into its own allocation"   Maarten Lankhorst
  • [PATCH 05/28] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Maarten Lankhorst
  • [PATCH 06/28] drm/i915: Remove locking from i915_gem_object_prepare_read/write   Maarten Lankhorst
  • [PATCH 07/28] drm/i915: Parse command buffer earlier in eb_relocate(slow)   Maarten Lankhorst
  • [PATCH 08/28] drm/i915: Use per object locking in execbuf, v12.   Maarten Lankhorst
  • [PATCH 09/28] drm/i915: Use ww locking in intel_renderstate.   Maarten Lankhorst
  • [PATCH 10/28] drm/i915: Add ww context handling to context_barrier_task   Maarten Lankhorst
  • [PATCH 11/28] drm/i915: Nuke arguments to eb_pin_engine   Maarten Lankhorst
  • [PATCH 12/28] drm/i915: Pin engine before pinning all objects, v5.   Maarten Lankhorst
  • [PATCH 13/28] drm/i915: Rework intel_context pinning to do everything outside of pin_mutex   Maarten Lankhorst
  • [PATCH 14/28] drm/i915: Make sure execbuffer always passes ww state to i915_vma_pin.   Maarten Lankhorst
  • [PATCH 15/28] drm/i915: Convert i915_gem_object/client_blt.c to use ww locking as well, v2.   Maarten Lankhorst
  • [PATCH 16/28] drm/i915: Kill last user of intel_context_create_request outside of selftests   Maarten Lankhorst
  • [PATCH 17/28] drm/i915: Convert i915_perf to ww locking as well   Maarten Lankhorst
  • [PATCH 18/28] drm/i915: Dirty hack to fix selftests locking inversion   Maarten Lankhorst
  • [PATCH 19/28] drm/i915/selftests: Fix locking inversion in lrc selftest.   Maarten Lankhorst
  • [PATCH 20/28] drm/i915: Use ww pinning for intel_context_create_request()   Maarten Lankhorst
  • [PATCH 21/28] drm/i915: Move i915_vma_lock in the selftests to avoid lock inversion, v3.   Maarten Lankhorst
  • [PATCH 22/28] drm/i915: Add ww locking to vm_fault_gtt   Maarten Lankhorst
  • [PATCH 23/28] drm/i915: Add ww locking to pin_to_display_plane   Maarten Lankhorst
  • [PATCH 24/28] drm/i915: Ensure we hold the object mutex in pin correctly.   Maarten Lankhorst
  • [PATCH 25/28] drm/i915: Add gem object locking to madvise.   Maarten Lankhorst
  • [PATCH 26/28] drm/i915: Convert i915_gem_object_attach_phys() to ww locking   Maarten Lankhorst
  • [PATCH 27/28] drm/i915: Fix userptr so we do not have to worry about obj->mm.lock   Maarten Lankhorst
  • [PATCH 28/28] fixup userptr implementation   Maarten Lankhorst
  • ✗ Fi.CI.BUILD: failure for series starting with [01/28] Revert "drm/i915/gem: Async GPU relocations only"   Patchwork
• [PATCH 03/28] Revert "drm/i915/gem: Drop relocation slowpath".   Maarten Lankhorst
• [PATCH i-g-t 0/2] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
  • ✗ Fi.CI.BUILD: failure for tests/kms_dither: New IGT to validate crtc Dithering (rev9)   Patchwork
  • [PATCH i-g-t 1/2] tests/kms_dither: New IGT to validate crtc Dithering   Bhanuprakash Modem
  • [PATCH i-g-t 2/2] tests/kms_dither: Validate newly added dither after CC blocks   Bhanuprakash Modem
• [PATCH 01/27] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 02/27] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 03/27] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 04/27] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 05/27] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 06/27] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 07/27] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 08/27] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 09/27] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 10/27] drm/i915/gt: Decouple obj<->fence reference cycles on freeing the GT pool   Chris Wilson
  • [PATCH 11/27] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 12/27] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 13/27] drm/i915/gem: Rename the list of relocations to reloc_list   Chris Wilson
  • [PATCH 14/27] drm/i915/gem: Move the 'cached' info to i915_execbuffer   Chris Wilson
  • [PATCH 15/27] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 16/27] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 17/27] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 18/27] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 19/27] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 20/27] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 21/27] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 22/27] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 23/27] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 24/27] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 25/27] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 26/27] drm/i915/gem: Manage ppGTT placement bias (starting offset) explicitly   Chris Wilson
  • [PATCH 27/27] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/27] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Patchwork
• [PATCH 01/37] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Chris Wilson
  • [PATCH 02/37] drm/i915/gt: Hold context/request reference while breadcrumbs are active   Chris Wilson
  • [PATCH 03/37] drm/i915/gt: Track signaled breadcrumbs outside of the breadcrumb spinlock   Chris Wilson
  • [PATCH 04/37] drm/i915/gt: Protect context lifetime with RCU   Chris Wilson
  • [PATCH 05/37] drm/i915/gt: Split the breadcrumb spinlock between global and contexts   Chris Wilson
  • [PATCH 06/37] drm/i915: Drop i915_request.lock serialisation around await_start   Chris Wilson
  • [PATCH 07/37] drm/i915: Drop i915_request.lock requirement for intel_rps_boost()   Chris Wilson
  • [PATCH 08/37] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source   Chris Wilson
  • [PATCH 09/37] drm/i915/gem: Reduce ctx->engines_mutex for get_engines()   Chris Wilson
  • [PATCH 10/37] drm/i915/gt: Decouple obj<->fence reference cycles on freeing the GT pool   Chris Wilson
  • [PATCH 11/37] drm/i915/gem: Don't drop the timeline lock during execbuf   Chris Wilson
  • [PATCH 12/37] drm/i915/gem: Rename execbuf.bind_link to unbound_link   Chris Wilson
  • [PATCH 13/37] drm/i915/gem: Rename the list of relocations to reloc_list   Chris Wilson
  • [PATCH 14/37] drm/i915/gem: Move the 'cached' info to i915_execbuffer   Chris Wilson
  • [PATCH 15/37] drm/i915/gem: Break apart the early i915_vma_pin from execbuf object lookup   Chris Wilson
  • [PATCH 16/37] drm/i915/gem: Remove the call for no-evict i915_vma_pin   Chris Wilson
  • [PATCH 17/37] drm/i915: Serialise i915_vma_pin_inplace() with i915_vma_unbind()   Chris Wilson
  • [PATCH 18/37] drm/i915: Add list_for_each_entry_safe_continue_reverse   Chris Wilson
  • [PATCH 19/37] drm/i915: Always defer fenced work to the worker   Chris Wilson
  • [PATCH 20/37] drm/i915/gem: Assign context id for async work   Chris Wilson
  • [PATCH 21/37] drm/i915/gem: Separate the ww_mutex walker into its own list   Chris Wilson
  • [PATCH 22/37] drm/i915/gem: Asynchronous GTT unbinding   Chris Wilson
  • [PATCH 23/37] drm/i915/gem: Bind the fence async for execbuf   Chris Wilson
  • [PATCH 24/37] drm/i915/gem: Include cmdparser in common execbuf pinning   Chris Wilson
  • [PATCH 25/37] drm/i915/gem: Include secure batch in common execbuf pinning   Chris Wilson
  • [PATCH 26/37] drm/i915/gem: Manage ppGTT placement bias (starting offset) explicitly   Chris Wilson
  • [PATCH 27/37] drm/i915/gem: Reintroduce multiple passes for reloc processing   Chris Wilson
  • [PATCH 28/37] drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.   Chris Wilson
  • [PATCH 29/37] drm/i915/gem: Pull execbuf dma resv under a single critical section   Chris Wilson
  • [PATCH 30/37] drm/i915/gem: Replace i915_gem_object.mm.mutex with reservation_ww_class   Chris Wilson
  • [PATCH 31/37] drm/i915: Hold wakeref for the duration of the vma GGTT binding   Chris Wilson
  • [PATCH 32/37] drm/i915: Specialise GGTT binding   Chris Wilson
  • [PATCH 33/37] drm/i915/gt: Acquire backing storage for the context   Chris Wilson
  • [PATCH 34/37] drm/i915/gt: Push the wait for the context to bound to the request   Chris Wilson
  • [PATCH 35/37] drm/i915: Remove unused i915_gem_evict_vm()   Chris Wilson
  • [PATCH 36/37] drm/i915/display: Drop object lock from intel_unpin_fb_vma   Chris Wilson
  • [PATCH 37/37] drm/i915/gem: Delay attach mmu-notifier until we acquire the pinned userptr   Chris Wilson
  • ✓ Fi.CI.BAT: success for series starting with [01/37] drm/i915/gt: Move intel_breadcrumbs_arm_irq earlier   Patchwork