search for: execbuf

On Mon, 5 Nov 2018 at 11:42, Robert Foss <robert.foss at collabora.com> wrote: > > When the execbuf call receives an in-fence it will get the dma_fence > related to that fence fd and wait on it before submitting the draw call. > > On the out-fence side we get fence returned by the submitted draw call > and attach it to a sync_file and send the sync_file fd to userspace. On > error...

On Mon, 5 Nov 2018 at 11:42, Robert Foss <robert.foss at collabora.com> wrote: > > When the execbuf call receives an in-fence it will get the dma_fence > related to that fence fd and wait on it before submitting the draw call. > > On the out-fence side we get fence returned by the submitted draw call > and attach it to a sync_file and send the sync_file fd to userspace. On > error...

...igt/gem_exec_nop: Relax assertion for parallel execution igt/gem_exec_nop: Check submitting nops to each engine in parallel igt/gem_exec_nop: Refine the target calculation igt/gem_ctx_switch: Increase execution weight igt/gem_exec_latency: Initial sketch for measuring execbuf cost in GPU cycles igt/gem_exec_nop: Show more timing details igt/gem_exec_whisper: Add competing child processes igt/gem_busy: Replace extended busy-ioctl testing in bat with simpler tests igt/gem_busy: Actually flag the hang tests to cause a GPU hang igt/gem_busy: Pr...

...tests/drv_missed_irq_hang: Wait harder lib: Show the command that fails when suspending lib/igt_fb: Fix domain tracking for GTT cairo surfaces igt/gem_exec_suspend: Exercise simple exec-and-read across suspend igt/gem_exec_basic: Check that we don't write into the const execbuf arg igt/gem_softpin: Check norelocations hold versus suspend/resume lib: Tidy testing for rtcwake lib/igt_gt: Replace asm clflush/mfence with __builtin_ia32 variants igt/pm_rpm: Wait for PCI D3 igt/gem_reloc_overflow: Fix errno tests for "overflow" lib:...

...ss wrote: > Hey Gerd, > > On 2018-11-09 11:13, Gerd Hoffmann wrote: > > On Mon, Nov 05, 2018 at 05:25:05PM +0000, Emil Velikov wrote: > > > On Mon, 5 Nov 2018 at 11:42, Robert Foss <robert.foss at collabora.com> wrote: > > > > > > > > When the execbuf call receives an in-fence it will get the dma_fence > > > > related to that fence fd and wait on it before submitting the draw call. > > > > > > > > On the out-fence side we get fence returned by the submitted draw call > > > > and attach it to a syn...

On Mon, Nov 12, 2018 at 05:51:53PM +0100, Robert Foss wrote: > > This series implements fence support for drm/virtio and > has been tested using qemu, kmscube and the below branches. > > Rob Herring solved a reference counting issue and > suggested a context check for the execbuf ioctl, his > changes have been included in the below commits to > keep the tree working at all commits. Patches added to qemu queue, should land in drm-misc-next soon. thanks, Gerd

...non-staging stuff really shouldn't depend upon code from > staging. Fully agree. I thought the best way towards that would be to show some driver code that _would_ use it. :) > I'm all for adding explicit syncing. Our plans are roughly. - Add both an in > and and out fence to execbuf to sync with other rendering and give userspace > a fence back. Needs to different flags probably. > > - Maybe add an ioctl to dma-bufs to get at the current implicit fences > attached to them (both an exclusive and non-exclusive version). This > should help with making explicit...

...ode Chad Versace (2): Bump version for 2.4.75 release Bump version for 2.4.75 release Chris Wilson (5): Import uapi/i915_drm.h from v4.10-rc5-950-g152d5750dda9 intel: Allow the client to control implicit synchronisation intel: Support passing of explicit fencing from execbuf intel: Clear execobject flags before preserving object in reuse cache intel: Export a function to re-enable implicit synchronisation Dave Airlie (2): Revert "Bump version for 2.4.75 release" intel: fix make distcheck Dongwon Kim (1): intel: update global_na...

...exec_nop: Include the impact of signaling a fence igt/gem_exec_nop: Restore check on available signal rings igt/amd_cs_nop: Spread submissions across multiple processes amdgpu/amd_cs_nop: Use a counter not seqno igt/gem_exec_store: Add welcome screen wsim: Only require execbuf wr ioctl for FENCE_OUT wsim: Fix reporting of workload/s for slaves wsim: Allow assigning priorities to each workload igt/gem_spin_batch: Avoid interleave throttle within open spin batches lib: Refactor testing for ability to use MI_STORE_DATA_IMM igt/gem_exec_fence: S...

...abilize measurements igt/gem_exec_blt: Repeat measurement and average igt/gem_concurrent_blit: Separate out the combinatorial explosion igt/gem_exec_(nop|blt): Repeat measurements at min/max GPU frequencies igt/gem_exec_(blt|nop): IGT magic igt/gem_cs_prefetch: Use gem_execbuf for pretty printing of failures igt/gem_exec_parse: Use a macro for pretty printing the test failure igt/gem_evict_everything: Use mlock to reduce available memory Revert "tests/gem_exec_lut_handle" igt/gem_evict_everything: Simplify mlocked_evictions igt/kms...

...em_softpin exercise the userptr ioctl to create shared buffers between CPU and GPU (Vinay Belgaumkar) - New tests: prime_mmap_coherency/kms_mmap_write_crc cache coherency tests (Tiago Vignatti) - New test: prime_mmap mmap() on dma-buf fds (Rob Bradford) - New test: gem_exec_reloc sanity check of execbuf-ioctl relocations (Chris Wilson) - New test: gem_exec_basic sanity check of execbuf-ioctl rings (Chris Wilson) - improved igt_hang_ring() infrastructure for generic hang injection support in the core library (Chris Wilson) - new igt_pm library to collect power management testing helpers (David W...

...grab ww_mutexes, grab a reference to the current fence > object, release ww_mutex, wait for fence, release reference to the fence > object. Yeah, if the only thing you want to do is wait for fences, then the rcu-protected fence ref grabbing + lockless waiting is all you need. But e.g. in an execbuf you also need to update fences and maybe deep down in the reservation code you notice that you need to evict some stuff and so need to wait on some other guy to finish, and it's too complicated to drop and reacquire all the locks. Or you simply need to do a blocking wait on other gpus (because...

...g ahead". :) I'm still secretly hoping that the Android guys at Google would jump in to help, now that we seem to agree that we could de-stage sync_fence. > > > I'm all for adding explicit syncing. Our plans are roughly. - Add both an in > > > and and out fence to execbuf to sync with other rendering and give userspace > > > a fence back. Needs to different flags probably. > > > > > > - Maybe add an ioctl to dma-bufs to get at the current implicit fences > > > attached to them (both an exclusive and non-exclusive version). This...

...you think we could use the android sync_fence for > > passing fences between user space? Or did you have something else in mind for > > explicit sync in the drm world? > > I'm all for adding explicit syncing. Our plans are roughly. > - Add both an in and and out fence to execbuf to sync with other rendering > and give userspace a fence back. Needs to different flags probably. > > - Maybe add an ioctl to dma-bufs to get at the current implicit fences > attached to them (both an exclusive and non-exclusive version). This > should help with making explic...

...t given the ridiculous backwards compat guarantees Linus expects us to keep up totally justified. Mistakes are really expensive. So I'm happy to see you charge ahead here. > > I'm all for adding explicit syncing. Our plans are roughly. - Add both an in > > and and out fence to execbuf to sync with other rendering and give userspace > > a fence back. Needs to different flags probably. > > > > - Maybe add an ioctl to dma-bufs to get at the current implicit fences > > attached to them (both an exclusive and non-exclusive version). This > > should...

...lib: Disable MI_STORE_DATA_IMM for gen3 (i915g and i915gm) lib: Avoid actually throttling from igt_require_gem() igt/pm_rpm: Use libc 'ftw' rather than opencoding our own filetree walk igt/gem_userptr_blits: Errors from gup are permanent igt/gem_ringfill: Prime execbuf before measuring ring size igt/gem_exec_schedule: Do not overcommit batches to the ring igt/gem_exec_schedule: Exercise reordering with many priority levels igt/gem_exec_schedule: Basic tests for preemption igt/gem_exec_scheduler: Demonstrate priority inversion between full...

...ynchronization model? Do you think we could use the android sync_fence for > passing fences between user space? Or did you have something else in mind for > explicit sync in the drm world? I'm all for adding explicit syncing. Our plans are roughly. - Add both an in and and out fence to execbuf to sync with other rendering and give userspace a fence back. Needs to different flags probably. - Maybe add an ioctl to dma-bufs to get at the current implicit fences attached to them (both an exclusive and non-exclusive version). This should help with making explicit and implicit sync work...

...a reference to the current fence >> object, release ww_mutex, wait for fence, release reference to the fence >> object. > Yeah, if the only thing you want to do is wait for fences, then the > rcu-protected fence ref grabbing + lockless waiting is all you need. > But e.g. in an execbuf you also need to update fences and maybe deep > down in the reservation code you notice that you need to evict some > stuff and so need to wait on some other guy to finish, and it's too > complicated to drop and reacquire all the locks. Or you simply need to > do a blocking wait on...

...ld use the android sync_fence for >>> passing fences between user space? Or did you have something else in mind for >>> explicit sync in the drm world? >> >> I'm all for adding explicit syncing. Our plans are roughly. >> - Add both an in and and out fence to execbuf to sync with other rendering >> and give userspace a fence back. Needs to different flags probably. >> >> - Maybe add an ioctl to dma-bufs to get at the current implicit fences >> attached to them (both an exclusive and non-exclusive version). This >> should h...

Am 22.07.2014 17:42, schrieb Daniel Vetter: > On Tue, Jul 22, 2014 at 5:35 PM, Christian K?nig > <christian.koenig at amd.com> wrote: >> Drivers exporting fences need to provide a fence->signaled and a fence->wait >> function, everything else like fence->enable_signaling or calling >> fence_signaled() from the driver is optional. >> >> Drivers