summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/nouveau/nouveau_fence.c
AgeCommit message (Expand)AuthorFilesLines
3 daysdrm/nouveau: make engine subclass subdev, and noaccel a bitfieldBen Skeggs1-6/+6
3 daysdrm/nouveau: very scary looking cleanup commitBen Skeggs1-6/+6
3 daysdrm/nouveau/fence: prevent oops in various failure pathsBen Skeggs1-0/+3
3 daysdrm/nouveau/fence: fix a race where fence->channel can disappearBen Skeggs1-5/+9
3 daysdrm/nouveau/fence: convert to exec engine, and improve channel syncBen Skeggs1-412/+59
3 daysdrm/nouveau/fence: minor api changes for an upcoming reworkBen Skeggs1-130/+85
3 daysdrm/nouveau/fence: make ttm interfaces wrap ours, not the other way aroundBen Skeggs1-23/+13
3 daysdrm/nouveau: create real execution engine for software object classBen Skeggs1-5/+1
3 daysdrm/nouveau: base fence timeout on time of emissionMarcin Slusarz1-1/+4
3 daysdrm/nouveau: use the same packet header macros as userspaceBen Skeggs1-13/+13
2012-03-22drm/nouveau: move fence sequence check to start of loopBen Skeggs1-6/+5
2012-03-22drm/nouveau: remove subchannel names from places where it doesn't matterBen Skeggs1-12/+12
2011-11-09drm/nouveau: initialize chan->fence.lock before useMarcin Slusarz1-2/+0
2011-09-20drm/nouveau: workaround semaphore hw bug causing unnecessary interruptsBen Skeggs1-1/+1
2011-09-09drm/nouveau: fix oops on pre-semaphore hardwareBen Skeggs1-1/+2
2011-06-23drm/nouveau: remove 'chan' argument from nouveau_bo_newBen Skeggs1-1/+1
2011-06-23drm/nv84-nvc0: explicitly map semaphore buffer into channel vmBen Skeggs1-16/+13
2011-06-18drm/nouveau: fix assumption that semaphore dmaobj is valid in x-chan syncBen Skeggs1-34/+25
2011-03-14drm/nouveau: fix __nouveau_fence_wait performanceMarcin Slusarz1-3/+12
2011-02-24drm/nouveau: remove no_vm/mappable flags from nouveau_boBen Skeggs1-1/+1
2011-02-24drm/nv50: drop explicit yields in favour of smaller PFIFO timesliceBen Skeggs1-46/+6
2011-02-24drm/nv84: use vm offsets for semaphoresBen Skeggs1-5/+17
2011-02-24drm/nv50: 0x50 needs semaphore yields tooBen Skeggs1-4/+17
2011-02-24drm/nvc0: implement semaphores for inter-channel syncBen Skeggs1-12/+40
2011-02-24drm/nv84: switch to new-style semaphoresBen Skeggs1-39/+80
2010-12-30drm/nvc0: reserve only subc 0 for kernel useBen Skeggs1-1/+1
2010-12-21drm/nvc0: implement fencingBen Skeggs1-8/+20
2010-12-08drm/nouveau: Avoid potential race between nouveau_fence_update() and context ...Francisco Jerez1-8/+11
2010-12-08drm/nouveau: fix use of drm_mm_node in semaphore objectBen Skeggs1-3/+2
2010-12-08drm/nouveau: implicitly insert non-DMA objects into RAMHTBen Skeggs1-6/+1
2010-12-08drm/nouveau: Spin for a bit in nouveau_fence_wait() before yielding the CPU.Francisco Jerez1-1/+2
2010-12-03drm/nouveau: tidy up and extend dma object creation interfacesBen Skeggs1-2/+2
2010-12-03drm/nouveau: fix annoying nouveau_fence type issueMarcin Slusarz1-11/+11
2010-12-03drm/nouveau: use object class structs more extensivelyBen Skeggs1-1/+1
2010-12-03drm/nouveau: Take fence spinlock in nouveau_fence_channel_fini().Francisco Jerez1-0/+4
2010-12-03drm/nouveau: Avoid race in the interchannel sync code.Francisco Jerez1-9/+15
2010-12-03drm/nouveau: Make fences take a weak channel reference.Francisco Jerez1-1/+2
2010-12-03drm/nouveau: Use lazy fence waits when doing software interchannel sync.Francisco Jerez1-2/+2
2010-12-03drm/nouveau: fix thinko in channel locking in semaphore pathBen Skeggs1-3/+3
2010-12-03drm/nouveau: add more fine-grained locking to channel list + structuresBen Skeggs1-0/+10
2010-11-18drm/nouveau: Fix sleep while atomic in the semaphore code.Francisco Jerez1-1/+6
2010-10-04drm/nv50: Fix large 3D performance regression caused by the interchannel sync...Francisco Jerez1-7/+39
2010-10-04drm/nouveau: Use semaphores to handle inter-channel sync in hardware.Francisco Jerez1-1/+194
2010-10-04drm/nouveau: Provide a means to have arbitrary work run on fence completion.Francisco Jerez1-0/+30
2010-10-04drm/nouveau: Minor refactoring/cleanup of the fence code.Francisco Jerez1-17/+46
2010-09-02drm/nouveau: Take fence spinlock before reading the last sequence.Francisco Jerez1-2/+4
2010-08-02drm/nouveau: set TASK_(UN)INTERRUPTIBLE before schedule_timeout()Kulikov Vasiliy1-2/+2
2010-07-13drm/nouveau: reduce usage of fence spinlock to when absolutely necessaryBen Skeggs1-25/+6
2010-01-11drm/nouveau: Use the software object for fencing.Francisco Jerez1-1/+1
2009-12-15drm/nouveau: remove use of -ERESTARTBen Skeggs1-1/+1
[next]
generated by cgit v0.9.0.2-2-gbebe at 2012-05-24 13:35:18 (GMT)