drm/xe: Make bind queues operate cross-tile

Browse source

Since bind jobs execute on the CPU rather than the GPU, maintaining a
per-tile bind queue no longer provides value. Convert the driver to use
a single bind queue shared across tiles. The primary change is routing
all GT TLB invalidations through this unified bind queue.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Shuicheng Lin <shuicheng.lin@intel.com>
Reviewed-by: Maarten Lankhorst <dev@lankhorst.se>
Link: https://patch.msgid.link/20260925201526.1355279-16-matthew.brost@intel.com

...

This commit is contained in:

Matthew Brost 2026-09-25 13:15:17 -07:00

parent 9144a7c68e

commit 366365b748

9 changed files with 124 additions and 177 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

140

drivers/gpu/drm/xe/xe_exec_queue.c

View file

14

drivers/gpu/drm/xe/xe_exec_queue.h

View file

20

drivers/gpu/drm/xe/xe_exec_queue_types.h

View file

23

drivers/gpu/drm/xe/xe_pt.c

View file

20

drivers/gpu/drm/xe/xe_sync.c

View file

15

drivers/gpu/drm/xe/xe_tlb_inval_job.c

View file

2

drivers/gpu/drm/xe/xe_tlb_inval_job.h

View file

65

drivers/gpu/drm/xe/xe_vm.c

View file

2

drivers/gpu/drm/xe/xe_vm_types.h

View file