- 14 Feb, 2017 5 commits
-
-
Tvrtko Ursulin authored
For some reason my compiler (and CI as well) failed to spot the uninitialized ret in mi_set_context. Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Fixes: 73dec95e ("drm/i915: Emit to ringbuffer directly") Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by:
Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/20170214152901.20361-1-tvrtko.ursulin@linux.intel.com
-
Tvrtko Ursulin authored
This removes the usage of intel_ring_emit in favour of directly writing to the ring buffer. intel_ring_emit was preventing the compiler for optimising fetch and increment of the current ring buffer pointer and therefore generating very verbose code for every write. It had no useful purpose since all ringbuffer operations are started and ended with intel_ring_begin and intel_ring_advance respectively, with no bail out in the middle possible, so it is fine to increment the tail in intel_ring_begin and let the code manage the pointer itself. Useless instruction removal amounts to approximately two and half kilobytes of saved text on my build. Not sure if this has any measurable performance implications but executing a ton of useless instructions on fast paths cannot be good. v2: * Change return from intel_ring_begin to error pointer by popular demand. * Move tail increment to intel_ring_advance to enable some error checking. v3: * Move tail advance back into intel_ring_begin. * Rebase and tidy. v4: * Complete rebase after a few months since v3. v5: * Remove unecessary cast and fix !debug compile. (Chris Wilson) v6: * Make intel_ring_offset take request as well. * Fix recording of request postfix plus a sprinkle of asserts. (Chris Wilson) v7: * Use intel_ring_offset to get the postfix. (Chris Wilson) * Convert GVT code as well. v8: * Rename *out++ to *cs++. v9: * Fix GVT out to cs conversion in GVT. v10: * Rebase for new intel_ring_begin in selftests. Signed-off-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170214113242.29241-1-tvrtko.ursulin@linux.intel.com
-
Ander Conselvan de Oliveira authored
I screwed up the rebase of commit d8fc70b7 ("drm/i915: Make power domain masks 64 bit long") before sending v2, causing a couple of conversions from 32 to 64 bit masks to be lost. Fixes: d8fc70b7 ("drm/i915: Make power domain masks 64 bit long") Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: intel-gfx@lists.freedesktop.org Signed-off-by:
Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> Reviewed-by:
-
Chris Wilson authored
The i915_gem_object_wait_fence() uses an incoming timeout=0 to query whether the current fence is busy or idle, without waiting. This can be used by the wait-ioctl to implement a busy query. Fixes: e95433c7 ("drm/i915: Rearrange i915_wait_request() accounting with callers") Testcase: igt/gem_wait/basic-busy-write-all Signed-off-by:
-
Chris Wilson authored
Explicitly disable stolen memory when running as a guest in a virtual machine, since the memory is not mediated between clients and reserved entirely for the host. The actual size should be reported as zero, but like every other quirk we want to tell the user what is happening. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99028Signed-off-by:
Zhenyu Wang <zhenyuw@linux.intel.com> Cc: stable@vger.kernel.org
-
- 13 Feb, 2017 35 commits
-
-
Chris Wilson authored
Check that we can reset the GPU and continue executing from the next request. Signed-off-by:
Mika Kuoppala <mika.kuoppala@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170213171558.20942-47-chris@chris-wilson.co.uk
-
Chris Wilson authored
As the page-table trees within the GTT are naturally aligned to power-of-two boundaries, by inserting an object that crosses a power-of-two (and the power-of-two intervals) we can quickly check the code for errors in switching between levels in the tree. Signed-off-by:
Matthew Auld <matthew.auld@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170213171558.20942-46-chris@chris-wilson.co.uk
-
Chris Wilson authored
Move a single page of an object around within the GGTT and check coherency of writes and reads. Signed-off-by:
-
Chris Wilson authored
Use the live tests against the mock ppgtt for quick testing on all platforms of the VMA layer. Signed-off-by:
-
Chris Wilson authored
i915_gem_gtt_insert should allocate from the available free space in the GTT, evicting as necessary to create space. Signed-off-by:
-
Chris Wilson authored
i915_gem_gtt_reserve should put the node exactly as requested in the GTT, evicting as required. Signed-off-by:
-
Chris Wilson authored
Very simple tests to just ask eviction to find some free space in a full GTT and one with some available space. Signed-off-by:
-
Chris Wilson authored
Ensure that we minimally exercise the aliasing_ppgtt, even on a full-ppgtt, by allocating one and similarly creating a context to use it. Signed-off-by:
-
Chris Wilson authored
In order to force testing of the aliasing ppgtt, extract its initialisation function. v2: Also extract the cleanup function for symmetry. Signed-off-by:
-
Chris Wilson authored
Check we can create and execution within a context. v2: Write one set of dwords through each context/engine to exercise more contexts within the same time period. Signed-off-by:
-
Chris Wilson authored
Mock testing to ensure we can create and lookup partial VMA. v2: Named phases Signed-off-by:
-
Chris Wilson authored
Exercise creating rotated VMA and checking the page order within. v2..v3: Be more creative in rotated params Signed-off-by:
-
Chris Wilson authored
High-level testing of the struct drm_mm by verifying our handling of weird requests to i915_vma_pin. Signed-off-by:
-
Chris Wilson authored
Simple test to exercise creation and lookup of VMA within an object. Signed-off-by:
-
Chris Wilson authored
It is possible whilst allocating the page-directory tree for a ppgtt bind that the shrinker may run and reap unused parts of the tree. If the shrinker happens to remove a chunk of the tree that the allocate_va_range has already processed, we may then try to insert into the dangling tree. This test uses the fault-injection framework to force the shrinker to be invoked before we allocate new pages, i.e. new chunks of the PD tree. References: https://bugs.freedesktop.org/show_bug.cgi?id=99295Signed-off-by:
-
Chris Wilson authored
Directly test allocating the va range and clearing it, this bypasses the use of i915_vma_bind() and inserting the pages to focus on testing of the pagetables. Signed-off-by:
-
Chris Wilson authored
Test the low-level i915_address_space interfaces to sanity check the live insertion/removal of address ranges. v2: Split out the low-level operations to a new test Signed-off-by:
-
Chris Wilson authored
Exercise filling different pages of the GTT v2: Walk all holes until we timeout Signed-off-by:
-
Chris Wilson authored
Same test as previously for the per-process GTT instead applied to the global GTT. Signed-off-by:
-
Chris Wilson authored
Allocate objects with varying number of pages (which should hopefully consist of a mixture of contiguous page chunks and so coalesced sg lists) and check that the sg walkers in insert_pages cope. v2: Check both small <-> large Signed-off-by:
-
Chris Wilson authored
Simple starting point for adding selftests for i915_gem_gtt, first try creating a ppGTT and filling it. Signed-off-by:
-
Chris Wilson authored
Though we have good coverage of our dmabuf interface through the mock tests, we also want to check the heavy module unload paths of the live i915 driver. Signed-off-by:
-
Chris Wilson authored
Check that we can create both dmabuf and objects from dmabuf. v2: Cleanups, correct include, fix unpin on dead path and prevent explosion on dmabuf init failure Signed-off-by:
-
Chris Wilson authored
Add a late selftest that walks over all forcewake registers (those below 0x40000) and uses the mmio debug register to check to see if any are unclaimed. This is possible if we fail to wake the appropriate powerwells for the register. Signed-off-by:
-
Chris Wilson authored
In addition to just testing the fw table we load, during the initial mock testing we can test that all tables are valid (so the testing is not limited to just the platforms that load that particular table). Signed-off-by:
-
Chris Wilson authored
Now that the kselftest infrastructure exists, put it to use and add to it the existing consistency checks on the fw register lookup tables. v2: s/tabke/table/ Signed-off-by:
-
Chris Wilson authored
Write into an object using WB, WC, GTT, and GPU paths and make sure that our internal API is sufficient to ensure coherent reads and writes. v2: Avoid invalid free upon allocation error Signed-off-by:
-
Chris Wilson authored
An unlikely error condition that we can simulate by stealing most of the range before trying to insert new objects. Signed-off-by:
-
Chris Wilson authored
Create partial mappings to cover a large object, investigating tiling (fenced regions) and VMA reuse. Signed-off-by:
-
Chris Wilson authored
Starting with a placeholder test just to reassure that we can create a test object, Signed-off-by:
-
Chris Wilson authored
The phys object is a rarely used device (only very old machines require a chunk of physically contiguous pages for a few hardware interactions). As such, it is not exercised by CI and to combat that we want to add a test that exercises the phys object on all platforms. v2: Always set err on error paths and not rely on inheriting the err. Signed-off-by:
-
Chris Wilson authored
Primarily to emphasize the difference between just advancing the breadcrumb using a bare request and the overhead of dispatching an execbuffer. Signed-off-by:
-
Chris Wilson authored
A request on one engine with a dependency on a request on another engine must wait for completion of the first request before starting. Signed-off-by:
-
Chris Wilson authored
Use a recursive-batch to busy spin on each to ensure that each is being run simultaneously. Signed-off-by:
-
Chris Wilson authored
Just create several batches of requests and expect it to not fall over! Signed-off-by:
-
