 Chris Wilson
|
6ffb7d0756
drm/i915: Construct a request even if the GPU is currently hung
|
8 年之前 |
|
Chris Wilson
|
c781c978e7
drm/i915: Add a sanity check that no request is submitted in the middle
|
8 年之前 |
 Daniel Vetter
|
a402eae64d
Merge tag 'v4.10-rc2' into drm-intel-next-queued
|
8 年之前 |
|
Chris Wilson
|
f73e73999d
drm/i915: Swap if(enable_execlists) in i915_gem_request_alloc for a vfunc
|
8 年之前 |
|
Chris Wilson
|
e8a9c58fcd
drm/i915: Unify active context tracking between legacy/execlists/guc
|
8 年之前 |
 Linus Torvalds
|
9439b3710d
Merge tag 'drm-for-v4.10' of git://people.freedesktop.org/~airlied/linux
|
8 年之前 |
|
Chris Wilson
|
0e932c080c
drm/i915: Hold a reference on the request for its fence chain
|
8 年之前 |
|
Chris Wilson
|
fc1584059d
drm/i915: Integrate i915_sw_fence with debugobjects
|
8 年之前 |
|
Chris Wilson
|
48bc2a4a42
drm/i915: Hold a reference on the request for its fence chain
|
8 年之前 |
|
Chris Wilson
|
1618bdb89b
drm/i915: Assert no external observers when unwind a failed request alloc
|
8 年之前 |
|
Chris Wilson
|
4ffd6e0cfe
drm/i915: Add is-completed assert to request retire entrypoint
|
8 年之前 |
 Joonas Lahtinen
|
4c266edb4c
drm/i915: Rename i915_gem_timeline.next_seqno to .seqno
|
8 年之前 |
 Mika Kuoppala
|
bc1d53c647
drm/i915: Wipe hang stats as an embedded struct
|
8 年之前 |
|
Mika Kuoppala
|
e5e1fc47ea
drm/i915: Use request retirement as context progress
|
8 年之前 |
|
Chris Wilson
|
786d290cae
drm/i915: Check that each request phase is completed before retiring
|
8 年之前 |
|
Chris Wilson
|
4302055b29
drm/i915: Be more careful to drop the GT wakeref
|
8 年之前 |
 Christian Borntraeger
|
f2f09a4cee
locking/core: Remove cpu_relax_lowlatency() users
|
8 年之前 |
|
Chris Wilson
|
9f792ebafe
drm/i915: Store the execution priority on the context
|
8 年之前 |
|
Chris Wilson
|
20311bd350
drm/i915/scheduler: Execute requests in order of priorities
|
8 年之前 |
|
Chris Wilson
|
52e5420907
drm/i915/scheduler: Record all dependencies upon request construction
|
8 年之前 |
|
Chris Wilson
|
0de9136dbb
drm/i915/scheduler: Signal the arrival of a new request
|
8 年之前 |
|
Chris Wilson
|
d55ac5bf97
drm/i915: Defer transfer onto execution timeline to actual hw submission
|
8 年之前 |
|
Chris Wilson
|
23902e49c9
drm/i915: Split request submit/execute phase into two
|
8 年之前 |
|
Chris Wilson
|
bb89485e99
drm/i915: Create distinct lockclasses for execution vs user timelines
|
8 年之前 |
|
Chris Wilson
|
6a5d1db98e
drm/i915: Spin until breadcrumb threads are complete
|
8 年之前 |
 Imre Deak
|
5bd11a34e4
drm/i915: Avoid early GPU idling due to already pending idle work
|
8 年之前 |
|
Chris Wilson
|
80b204bce8
drm/i915: Enable multiple timelines
|
8 年之前 |
|
Chris Wilson
|
f2d13290e3
drm/i915: Defer setting of global seqno on request to submission
|
8 年之前 |
|
Chris Wilson
|
28176ef4cf
drm/i915: Reserve space in the global seqno during request allocation
|
8 年之前 |
|
Chris Wilson
|
85e17f5974
drm/i915: Move the global sync optimisation to the timeline
|
8 年之前 |
