[RFC PATCH 0/5] Add memory page offlining support

[Tejas Upadhyay <tejas.upadhyay@intel.com>]

This functionality represents a significant step in making
the xe driver gracefully handle hardware memory degradation.
By integrating with the DRM Buddy allocator, the driver
can permanently "carve out" faulty memory so it isn't reused
by subsequent allocations.

This series adds memory page offlining support with following:
1. Link and track ttm BO's with physical addresses
2. Handle the generated physical address error by reserving addresses 4K page
3. Adds supporting debugfs to inject manual physcal address error
4. Add buddy block allocation dump for debuggin buddy related issues
5. Sysfs entry to provide statistics of bad gpu vram pages for user info

Opens:
1. mm->avail under drm_buddy throwing WARN_ON(mm->avail != mm->size) with no leaks 
in memory, mutliple bind/ubind works fine. Debug in progress.
2. dump_allocated_blocks() and xe_ttm_vram_addr_to_tbo() API will move under drm_buddy, 
right now just to showcase concept its part of xe code

Tejas Upadhyay (5):
  drm/xe: Implement VRAM object tracking ability using physical address
  drm/xe: Handle physical memory address error
  [DO_NOT_REVIEW]drm/xe/cri: Add debugfs to inject faulty vram address
  drm/xe: Add routine to dump allocated VRAM blocks
  [DO NOT REVIEW]drm/xe/cri: Add sysfs interface for bad gpu vram pages

 drivers/gpu/drm/xe/xe_debugfs.c            |  49 +++
 drivers/gpu/drm/xe/xe_device_sysfs.c       |   2 +
 drivers/gpu/drm/xe/xe_tile_sysfs.c         |   1 +
 drivers/gpu/drm/xe/xe_ttm_vram_mgr.c       | 366 +++++++++++++++++++++
 drivers/gpu/drm/xe/xe_ttm_vram_mgr.h       |   6 +-
 drivers/gpu/drm/xe/xe_ttm_vram_mgr_types.h |  23 ++
 6 files changed, 446 insertions(+), 1 deletion(-)

-- 
2.52.0