linux_dsm_epyc7002/drivers/gpu/drm/i915/selftests
Chris Wilson a679f58d05 drm/i915: Flush pages on acquisition
When we return pages to the system, we ensure that they are marked as
being in the CPU domain since any external access is uncontrolled and we
must assume the worst. This means that we need to always flush the pages
on acquisition if we need to use them on the GPU, and from the beginning
have used set-domain. Set-domain is overkill for the purpose as it is a
general synchronisation barrier, but our intent is to only flush the
pages being swapped in. If we move that flush into the pages acquisition
phase, we know then that when we have obj->mm.pages, they are coherent
with the GPU and need only maintain that status without resorting to
heavy handed use of set-domain.

The principle knock-on effect for userspace is through mmap-gtt
pagefaulting. Our uAPI has always implied that the GTT mmap was async
(especially as when any pagefault occurs is unpredicatable to userspace)
and so userspace had to apply explicit domain control itself
(set-domain). However, swapping is transparent to the kernel, and so on
first fault we need to acquire the pages and make them coherent for
access through the GTT. Our use of set-domain here leaks into the uABI
that the first pagefault was synchronous. This is unintentional and
baring a few igt should be unoticed, nevertheless we bump the uABI
version for mmap-gtt to reflect the change in behaviour.

Another implication of the change is that gem_create() is presumed to
create an object that is coherent with the CPU and is in the CPU write
domain, so a set-domain(CPU) following a gem_create() would be a minor
operation that merely checked whether we could allocate all pages for
the object. On applying this change, a set-domain(CPU) causes a clflush
as we acquire the pages. This will have a small impact on mesa as we move
the clflush here on !llc from execbuf time to create, but that should
have minimal performance impact as the same clflush exists but is now
done early and because of the clflush issue, userspace recycles bo and
so should resist allocating fresh objects.

Internally, the presumption that objects are created in the CPU
write-domain and remain so through writes to obj->mm.mapping is more
prevalent than I expected; but easy enough to catch and apply a manual
flush.

For the future, we should push the page flush from the central
set_pages() into the callers so that we can more finely control when it
is applied, but for now doing it one location is easier to validate, at
the cost of sometimes flushing when there is no need.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Antonio Argenziano <antonio.argenziano@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.william.auld@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190321161908.8007-1-chris@chris-wilson.co.uk
2019-03-21 17:28:12 +00:00
..
huge_gem_object.c drm/i915: Make object/vma allocation caches global 2019-02-28 11:08:02 +00:00
huge_gem_object.h
huge_pages.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
i915_active.c drm/i915: Beware temporary wedging when determining -EIO 2019-02-20 16:31:08 +00:00
i915_gem_coherency.c drm/i915: Store the BIT(engine->id) as the engine's mask 2019-03-05 18:19:50 +00:00
i915_gem_context.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
i915_gem_dmabuf.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
i915_gem_evict.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
i915_gem_gtt.c drm/i915: Separate GEM context construction and registration to userspace 2019-03-21 15:59:25 +00:00
i915_gem_object.c drm/i915: Store the BIT(engine->id) as the engine's mask 2019-03-05 18:19:50 +00:00
i915_gem.c drm/i915: Do a synchronous switch-to-kernel-context on idling 2019-03-08 10:57:05 +00:00
i915_live_selftests.h drm/i915: Generalise GPU activity tracking 2019-02-05 17:12:00 +00:00
i915_mock_selftests.h drm/i915: Replace global breadcrumbs with per-context interrupt tracking 2019-01-29 21:45:22 +00:00
i915_random.c drm/i915: Share per-timeline HWSP using a slab suballocator 2019-01-28 19:07:06 +00:00
i915_random.h drm/i915: Share per-timeline HWSP using a slab suballocator 2019-01-28 19:07:06 +00:00
i915_request.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
i915_selftest.c drm/i915/selftests: Upgrade printing test/subtest name to pr_info 2019-03-06 11:36:36 +00:00
i915_sw_fence.c drm/i915/selftests: Disable preemption while setting up fence-timers 2019-03-14 11:47:06 +00:00
i915_syncmap.c
i915_timeline.c drm/i915: Stop storing the context name as the timeline name 2019-03-21 15:59:31 +00:00
i915_vma.c drm/i915/selftests: Move local mock_ggtt allocations to the heap 2019-02-17 21:07:46 +00:00
igt_flush_test.c drm/i915: Reduce presumption of request ordering for barriers 2019-03-08 10:57:08 +00:00
igt_flush_test.h
igt_live_test.c drm/i915: Drop fake breadcrumb irq 2019-01-29 21:45:30 +00:00
igt_live_test.h drm/i915/selftests: Refactor common live_test framework 2019-01-22 13:01:20 +00:00
igt_reset.c drm/i915/selftests: Add tests for GT and engine workaround verification 2018-12-04 12:23:19 +00:00
igt_reset.h drm/i915/selftests: Add tests for GT and engine workaround verification 2018-12-04 12:23:19 +00:00
igt_spinner.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
igt_spinner.h drm/i915/selftests: Extract spinner code 2018-11-30 15:09:00 +00:00
igt_wedge_me.h
intel_engine_cs.c
intel_guc.c drm/i915: Store the BIT(engine->id) as the engine's mask 2019-03-05 18:19:50 +00:00
intel_hangcheck.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
intel_lrc.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
intel_uncore.c drm/i915: move regs pointer inside the uncore structure 2019-03-20 21:12:50 +00:00
intel_workarounds.c drm/i915: Flush pages on acquisition 2019-03-21 17:28:12 +00:00
lib_sw_fence.c drm/i915: Replace global breadcrumbs with per-context interrupt tracking 2019-01-29 21:45:22 +00:00
lib_sw_fence.h drm/i915: Replace global breadcrumbs with per-context interrupt tracking 2019-01-29 21:45:22 +00:00
mock_context.c drm/i915: Stop storing ctx->user_handle 2019-03-21 15:59:29 +00:00
mock_context.h
mock_dmabuf.c
mock_dmabuf.h
mock_drm.c
mock_drm.h
mock_engine.c drm/i915: Stop storing the context name as the timeline name 2019-03-21 15:59:31 +00:00
mock_engine.h drm/i915: Replace global breadcrumbs with per-context interrupt tracking 2019-01-29 21:45:22 +00:00
mock_gem_device.c drm/i915: make more uncore function work on intel_uncore 2019-03-20 21:12:42 +00:00
mock_gem_device.h
mock_gem_object.h
mock_gtt.c drm/i915/selftests: Create a clean GGTT for vma/gtt selftesting 2019-01-22 12:50:39 +00:00
mock_gtt.h drm/i915/selftests: Create a clean GGTT for vma/gtt selftesting 2019-01-22 12:50:39 +00:00
mock_request.c drm/i915: Make request allocation caches global 2019-02-28 11:07:56 +00:00
mock_request.h drm/i915: Make request allocation caches global 2019-02-28 11:07:56 +00:00
mock_timeline.c drm/i915: Introduce i915_timeline.mutex 2019-03-01 14:54:46 +00:00
mock_timeline.h
mock_uncore.c drm/i915: make more uncore function work on intel_uncore 2019-03-20 21:12:42 +00:00
mock_uncore.h drm/i915: make more uncore function work on intel_uncore 2019-03-20 21:12:42 +00:00
scatterlist.c