[PATCH v10 0/3] drm/i915: Move system memory to TTM for discrete

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Early implementation of moving system memory for discrete cards over to
TTM. We first add the notion of objects being migratable under the object
lock to i915 gem, and add some asserts to verify that objects are either
locked or pinned when the placement is checked by the gem code.

Patch 2 deals with updating the i915 gem bookkeeping after a TTM move,
Patch 3 moves system over from shmem to TTM for discrete

Note that the mock device doesn't consider itself discrete so the TTM
system path is not checked by the mock selftests.

v2:
- Style fixes (reported by Matthew Auld)
- Drop the last patch (migration) It needs selftests and some additional work.
- Unconditionally add VM_IO at mmap time.

v3:
- More style fixes (reported by Matthew Auld)
- Don't overfill the busy placement vector (reported by Matthew Auld)

v4:
- Remove confusion around shrinkable objects (reported by Matthew Auld)

v5:
- Remove confusion around shrinkable objects again, but this time in the
  correct patch. (reported by Matthew Auld)

v6:
- One patch already committed.
- Introduce a __i915_gem_object_is_lmem() to be used in situations where we
  know that a fence that can't currently signal keeps the object from being
  migrated or evicted.
- Rebase on accelerated TTM moves
- Fix TODO:s for supporting system memory with TTM.
- Update the object GEM region after a TTM move if compatible.
- Move a couple of warnings for shmem on DGFX.

v7:
- Just updated a commit message with version history under dashes.

v8:
- Reinstate alignment at ttm_bo_init_reserved() time. (Reported by
  Matthew Auld).
- When changing regions, also move the object to the new region list
  and break early. (Reported by Matthew Auld).
- Don't flag the object as contiguous based on the current region min
  pages size.
v9:
- Remove a DGFX warning in __i915_gem_object_release_shmem since it
  is called from the userptr code.
v10:
- Add some comments around the gem object "mem_flags" field, and
  use a full unsigned int for it. (Suggested by Daniel Vetter).
- Add an object locked section while checking mem_flags in the live
  mman selftest.

Thomas Hellström (3):
  drm/i915: Update object placement flags to be mutable
  drm/i915/ttm: Adjust gem flags and caching settings after a move
  drm/i915/ttm: Use TTM for system memory

 drivers/gpu/drm/i915/gem/i915_gem_internal.c  |   4 +-
 drivers/gpu/drm/i915/gem/i915_gem_lmem.c      |  22 ++
 drivers/gpu/drm/i915/gem/i915_gem_lmem.h      |   2 +
 drivers/gpu/drm/i915/gem/i915_gem_mman.c      |  12 +-
 drivers/gpu/drm/i915/gem/i915_gem_object.c    |  38 ++++
 drivers/gpu/drm/i915/gem/i915_gem_object.h    |  14 +-
 .../gpu/drm/i915/gem/i915_gem_object_types.h  |  27 ++-
 drivers/gpu/drm/i915/gem/i915_gem_pages.c     |   2 +-
 drivers/gpu/drm/i915/gem/i915_gem_phys.c      |   2 +-
 drivers/gpu/drm/i915/gem/i915_gem_region.c    |   4 -
 drivers/gpu/drm/i915/gem/i915_gem_shmem.c     |   9 +-
 drivers/gpu/drm/i915/gem/i915_gem_ttm.c       | 196 ++++++++++++++----
 drivers/gpu/drm/i915/gem/i915_gem_userptr.c   |   4 +-
 .../drm/i915/gem/selftests/huge_gem_object.c  |   4 +-
 .../gpu/drm/i915/gem/selftests/huge_pages.c   |   5 +-
 .../drm/i915/gem/selftests/i915_gem_mman.c    |  25 ++-
 .../drm/i915/gem/selftests/i915_gem_phys.c    |   3 +-
 drivers/gpu/drm/i915/i915_drv.h               |   3 -
 drivers/gpu/drm/i915/i915_gpu_error.c         |   2 +-
 drivers/gpu/drm/i915/intel_memory_region.c    |   7 +-
 drivers/gpu/drm/i915/intel_memory_region.h    |   8 +
 21 files changed, 288 insertions(+), 105 deletions(-)

-- 
2.31.1




[Index of Archives]     [Linux DRI Users]     [Linux Intel Graphics]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux