| Age | Commit message (Collapse) | Author |
|---|
|
On dg2+, MI_BATCH_BUFFER_START | MI_USE_PREDICATE switches to using the
MI_SET_PREDICATE_RESULT register and requires bit 0 to be set to skip
the looping MI_BATCH_BUFFER_START.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/5463
Reviewed-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Signed-off-by: Chris Wilson <chris.p.wilson@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
|
|
Add test description for below subtests:
basic-busy-all, basic-wait-all, busy-hang-all, wait-hang-all,
basic-busy, basic-wait, basic-await, nb-await.
v2 : Modified subtests description.
Cc: Kamil Konieczny <kamil.konieczny@linux.intel.com>
Signed-off-by: Janga Rahul Kumar <janga.rahul.kumar@intel.com>
Reviewed-by: Kamil Konieczny <kamil.konieczny@linux.intel.com>
|
|
In few tests we got some MI_* duplicates (MI_MATH for example).
Add common definitions in intel_reg.h and remove local definitions
in the tests.
v2: Definitions MI_LOAD_REGISTER_MEM_GEN8 was removed so from now
on user will need to encode length on it own. :
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
Three subtests:
- syncobj-timeline-chain-engines
- syncobj-stationary-timeline-chain-engines
- syncobj-backward-timeline-chain-engines
were not previously rewritten to use no-reloc.
Using allocator is not necessary in this case, we need to softpin
only counter object. Offsets for all batches are chosen by the
kernel as their location within gpu vm doesn't matter.
Some explanation is required regarding batchbuffer updates for each
iteration. Before introducing softpin all batchbuffers were touched by
the relocations what introduces stalls between them during execution.
These stalls could be removed as batchbuffers don't change their
contents for each iteration. But I decided to keep this behavior
intact for relocations changing it only for no-reloc mode. With
softpinning batchbuffer for each engine is written once (for first
iteration) so next execbuf reuses same batch. This removes stalls
on subsequent iterations as batchbuffer for each engine is ready
immediate after completion.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
|
|
The update to use intel_ctx_t missed a line that configures the
context to allow hanging. Fix that.
Fixes: 09c36188b ("tests/i915/gem_exec_fence: Convert to intel_ctx_t (v2)")
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
|
|
A lot of tests use almost identical code for creating a batch buffer
which does a single write to memory and another is about to be added.
Instead, move the most generic version into a common helper function.
Unfortunately, the other instances are all subtly different enough to
make it not so trivial to try to use the helper. It could be done but
it is unclear if it is worth the effort at this point. This patch
proves the concept, if people like it enough then it can be extended.
v2: Fix up object address vs store offset confusion (with help from
Zbigniew K).
v3: Cope with >32bit store_offset (review feedback from Matthew Brost).
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
|
|
From Haswell up until gen12, only rcs supports all the registers and MI
commands required to build the wait batch; before Haswell MI does not
support the required MATH operations nor the predication. However, after
gen12 we may not always have rcs available and so require to run the
delay on other command streamers, which requires us to use the
appropriate mmio base for the alternate engine.
Signed-off-by: Chris Wilson <chris.p.wilson@intel.com>
Reviewed-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
|
|
For newer gens we're not able to rely on relocations. Adopt to use
offsets acquired from the allocator.
Three subtests are not covered here, there are:
- syncobj-timeline-chain-engines
- syncobj-stationary-timeline-chain-engines
- syncobj-backward-timeline-chain-engines
Due to sophisticated manner of the three tests mentioned above they
will be subject of separate patch. At the moment CI depends on reloc
version only so this doesn't introduce regression there and decreases
coverage gap on no-reloc discrete runs.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Signed-off-by: Andrzej Turko <andrzej.turko@linux.intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
|
|
We lost explicit WC mmaps on discrete, where we now only support FIXED,
however such mappings should be device coherent. In gem_exec_fence it
looks like we can just use mmap__device_coherent, which should also work
on discrete platforms, while still using an explicit WC mmap on
integrated platforms.
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Ramalingam C <ramalingam.c@intel.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
v2 (Jason Ekstrand):
- Pass the context config to gem_submission_measure()
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
This will make iteration easier when we switch to intel_ctx_t.
v2 (Zbigniew Kempczyński):
- Require nengines > 0 in setup_timeline_chain_engines()
v3 (Zbigniew Kempczyński):
- Should have been nengines > 1 in setup_timeline_chain_engines()
v3 (Jason Ekstrand):
- Delete more intel_init_engine_list() in this patch
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
If provided, the engine (or ALL_ENGINES) is relative to the given
context config. This is intended to be transitional. We'll get rid of
all the __for_each_physical_engine stuff later.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
Allows building cleanly with clang again. It complains about using
invalid initialisers like char[][] = { 0 }.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
Add a wrapper for gem_create_ext ioctl (a version of gem_create that
accepts extensions). In preparation for the driver change implementing it,
a local definition of its id and necessary structs have been added,
which are to be erased as soon as those definitions
appear in the i915_drm.h file.
The new ioctl wrapper is added to a separate file.
For consistency the wrapper of the old ioctl, gem_create
is moved from ioctl_wrappers to gem_create.
Signed-off-by: Andrzej Turko <andrzej.turko@linux.intel.com>
Cc: Zbigniew Kempczynski <zbigniew.kempczynski@intel.com>
Cc: Dominik Grzegorzek <dominik.grzegorzek@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Chris P Wilson <chris.p.wilson@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Acked-by: Petri Latvala <petri.latvala@intel.com>
|
|
Fix the use of the legacy ring selection after the default context had
an engine map installed.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Improve the differentiation between the legacy ring selector ABI and the
more recent engine selection API.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
|
|
Report we cannot run the timeline tests (SKIP) if the kernel doesn't
support the API.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Submit a chain of spinners across all the engines, using the submit
fence to launch them in parallel.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
As a stepping stone, increase the assumed 16 engines is enough for
everyone, to cover the current RING_MASK, the maximum number of engines
that can currently be selected during execbuf.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
|
|
We want to recognise future devices (gen = -1u) and treat them as an
extension of the latest known device, which is typically true.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
|
|
Use PRI[ux]64 for printing 64bit values in a 32bit build.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
syncobj-timeline-wait is only exercising waits on the syncobj along a
single timeline, for which it can use the universally available default
context, and so not try and fail to create a HW context on very old
platforms.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2308
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
|
|
Those tests are meant to verify the ordering while messing around with
the timeline points.
v2: Add munmap() (Chris)
v3: Rebase
Drop gem_context_set_all_engines()
Fix subtest require on number of engines
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
We can now give a tuple (handle, point_value) for timeline semaphores.
v2: Fix up syncobj-timeline-repeat test (Lionel)
Update following kernel rename
v3: Add more ordering tests for syncobj/buffer handle state (Chris)
v4: Rebase
Drop gem_context_set_all_engines()
v5: Fix syncobj-timeline-wait, ensure cloned context remains blocked
on the spinner.
v6: Ensure syncobj-timeline-wait can run on all generations, make the
spinner use the same GEM context and spin on all engines.
v7: Add test for multiple execbuffer extension nodes
(syncobj-timeline-multiple-ext-nodes)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v4)
|
|
Non-preemptible spinners should be avoided as often as possible.
gem_exec_fence@[basic-await, nb-await] has no reason to use non-preemptible
workload, only hanging subtests need it.
Signed-off-by: Dominik Grzegorzek <dominik.grzegorzek@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
gem_ring_measure_inflight() uses the legacy ring idx, whereas
gem_exec_schedule is using the new dynamic ctx->engine[] indices.
Mismatch and hanging tests ensue.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Since we are hiding our write hazards, gem_read() may not know if it
needs to flush the CPU caches prior to reading back. As we lie to the
kernel, use an explicit WC mmap for result readback.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/1875
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Like the parallel, but restricted to a single engine -- assumes we have
timeslicing.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
igt_require_gem() is a pecularity of i915/, move it out of the core.
Similar opportunistic move of gem_reopen_driver() and
gem_quiescent_gpu().
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Submit more coupled batches than can possibly fit into the ELSP and
verify that we can slice between them all, while fulfilling the
submit-fence.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Replace the open-coded spinner with igt_spin_t so that it is cmdparser
friendly.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
I never finished writing this test. Its goal was to try and verify that
a flip cannot happen before a fence is signaled; currently that is done
by kms_busy.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
The basic-await has been converted to use an igt_spin_t and so does not
need the mutable check anymore.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Move inline upper_32_bits() and lower_32_bits() function from tests to
library. Functiom will be usefull for elsewhere use.
v2: fix function description
Cc: Katarzyna Dec <katarzyna.dec@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Daniel Mrzyglod <daniel.t.mrzyglod@intel.com>
|
|
Put the test inside the igt_dynamic{} scope.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
engines.
Replaced the legacy for_each_engine* defines with the ones implemented
in the gem_engine_topology library.
Corrected subtest name in fast-feedback.testlist from,
basic-busy-default --> basic-busy
basic-wait-default --> basic-wait
basi-await-default --> basic-await
nb-await-default --> nb-await
Signed-off-by: saigowth <sai.gowtham.ch@intel.com>
Reviewed-by: Katarzyna Dec <katarzyna.dec@intel.com>
|
|
Some spinners are used with the intent of never ending and being
declared hung by the kernel. In some cases, these are being used to
simulate invalid payloads and so we can use an invalid command to
trigger a GPU hang. (Other cases, they are simulating infinite workloads
that truly never end, but we still need to be able to curtail to provide
multi-tasking). This patch adds IGT_SPIN_INVALID_CS to request the
injection of 0xdeadbeef into the command stream that should trigger a
GPU hang.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Fixes: 9e57f8a51d59 ("lib/i915: Add query to detect if engine accepts only ro batches")
References: https://gitlab.freedesktop.org/drm/intel/issues/997
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
|
|
Removing igt_skip_on_simulation() from all the gem tests
since this feature is not supported anymore.
v2: Rebase
v3: Corrected build failure, because of deletion of
gem_exec_blt.c
Signed-off-by: Swati Sharma <swati2.sharma@intel.com>
Signed-off-by: Karthik B S <karthik.b.s@intel.com>
Reviewed-by: Martin Peres <martin.peres@linux.intel.com>
|
|
v2: Fix mistake in syncobj_busy() (Lionel)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Drop local defines etc..
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
If cmd parser is mandatory, batch can't be modified post execbuf.
Some tests rely on modifying batch post execbuf. Give those
tests a method to query if those modifications ever reach
the actual engine command stream.
v2: pull in the test changes, doh
v3: class based query
v4: rebase
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
The kernel is now enforcing that clients are not allowed to block higher
priority contexts from accessing the GPU; one is no longer allowed to
sleep for a second hogging the GPU. Reduce the sleep down to 50ms, short
enough not to anger the preempt-off checks while long enough for any
ordinary GPU work to complete.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
The pair of *-hang-all will generate sufficient hangs for the kernel to
ban the client. This is unfortunate as it means all further tests are
skipped. Ask nicely not to be banned.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Within this set of fence execution tests, we never once try to modeset;
being KMS master is not required.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Provide the iterator name as an explicit macro parameter so that it is
known to the caller, and allows for them to properly nest loops over all
engines.
Fixes:
../tests/i915/gem_exec_schedule.c: In function ‘semaphore_noskip’:
../lib/igt_gt.h:84:44: warning: declaration of ‘e__’ shadows a previous local [-Wshadow]
for (const struct intel_execution_engine *e__ = intel_execution_engines;\
^~~
../tests/i915/gem_exec_schedule.c:653:2: note: in expansion of macro ‘for_each_physical_engine’
for_each_physical_engine(i915, other) {
^~~~~~~~~~~~~~~~~~~~~~~~
../lib/igt_gt.h:84:44: note: shadowed declaration is here
for (const struct intel_execution_engine *e__ = intel_execution_engines;\
^~~
../tests/i915/gem_exec_schedule.c:652:2: note: in expansion of macro ‘for_each_physical_engine’
for_each_physical_engine(i915, engine) {
^~~~~~~~~~~~~~~~~~~~~~~~
../tests/i915/gem_exec_schedule.c: In function ‘measure_semaphore_power’:
../lib/igt_gt.h:84:44: warning: declaration of ‘e__’ shadows a previous local [-Wshadow]
for (const struct intel_execution_engine *e__ = intel_execution_engines;\
^~~
../tests/i915/gem_exec_schedule.c:1740:3: note: in expansion of macro ‘for_each_physical_engine’
for_each_physical_engine(i915, engine) {
^~~~~~~~~~~~~~~~~~~~~~~~
../lib/igt_gt.h:84:44: note: shadowed declaration is here
for (const struct intel_execution_engine *e__ = intel_execution_engines;\
^~~
../tests/i915/gem_exec_schedule.c:1719:2: note: in expansion of macro ‘for_each_physical_engine’
for_each_physical_engine(i915, signaler) {
^~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Andi Shyti <andi.shyti@intel.com>
|
|
When using a spinner to trigger a hang, make it unpreemptable so that it
appears like a true hang.
References: https://bugs.freedesktop.org/show_bug.cgi?id=109661
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
|
|
Continue instead of breaking on the first engine.
Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
The only detail of note here was that we were creating a fence from the
recursive batch, now supported by igt_spin_t (thanks Tvrtko).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
|
