Sometimes we end up in /bin and I am not able to explain that just by
looking at the machine or the IGT tarball that was deployed there so
let's get some more breadcrumbs.

Cc: Petri Latvala <petri.latvala@intel.com>
Issue: #12


Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>

instead of 0.5 use 1.0 for white color, this is due to bit replication
effect. Also set cairo to use same blending mode as what HW uses.

Signed-off-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Mika Kahola <mika.kahola@intel.com>

casting unsigned char pointer to int pointer causes gcc to be unhappy with
comment: "warning: dereferencing type-punned pointer will break
strict-aliasing rules"

Signed-off-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>

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>

If a test is only targeting the GGTT API and its corner cases, it can
only run if we have a mappable aperture.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Antonio Argenziano <antonio.argenziano@intel.com>
Acked-by: Antonio Argenziano <antonio.argenziano@intel.com>

Our list was something between Wayland and Linux Kernel list
implementations, right in the uncanny valley.

On top of that it falsely claimed that it's a straight copy from the
Wayland project.

Let's make our impl more akin to the kernel one to ease the cognitive
dissonance for the developers working on all those projects.

This patch:
 * mimics the current kernel list interface
 * separates IGT helpers in the source files
 * adds brief explanation and code example for igt-doc
 * introduces igt_list.c as static inlines are not visible in the docs

v2: mimic the kernel instead of wayland (Chris)
    - _head suffix for the sentinel/link struct
    - _entry_ in iterator names that go over the elements
v3: I forgot to merge this in time and there was another call site
    that had to be converted in gem_spin_batch.c

Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>

Register a userspace fault handler for a memory region that we also pass
to the GPU via userptr, and make sure the pagefault is properly serviced
before we execute.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

The following error looks quite mysterious:

   $ ../../tests/igt_command_line.sh kms_ccs
   ./kms_ccs:
     Checking invalid option handling...
     Checking valid option handling...
     Checking subtest enumeration...
   FAIL: ./kms_ccs

Let's add some explanation when we fail the subtests enumeration checks
so people are not completely baffled when they get failures from the CI.

Cc: Mika Kahola <mika.kahola@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>

Commit f56ed6b1 ("kms_busy: Rename tests so pipe-specific subtests
are blacklistable") renamed the subtests but fast-feedback.testlist
change was forgotten. Oopsie!

Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Cc: Martin Peres <martin.peres@linux.intel.com>
Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

We cannot control how long RCU takes to find a quiescent point as that
depends upon the background load and so may take an arbitrary time.
Instead, let's try to avoid that impacting our measurements by inserting
an rcu_barrier() before our critical timing sections and hope that hides
the issue, letting us always perform a fast reset. Fwiw, we do the
expedited RCU synchronize, but that is not always enough.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

igt_aux.h already provides the optimal igt_fls(), so use that in
preference to open coding the brute force version.

Reported-by: Stuart Summers <stuart.summers@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Stuart Summers <stuart.summers@intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

Blacklist pipes E and F usage even when the subtest name ends in
pipe-x.

Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Cc: Martin Peres <martin.peres@linux.intel.com>

The blacklist for pipes E and F are not capturing kms_busy's usage of
them, because of the differing naming convention. Make that more
uniform.

Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Cc: Martin Peres <martin.peres@linux.intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

So when we start sending emails about those failures including
`| tail -n 20` will actually make sense.

v2: make -j $(nproc) by default (Petri)

Cc: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>

Each call to igt_debugfs_read() opens a file descriptor to debugfs,
calls igt_debugfs_simple_read() and then closes it.
So lets stop this waste of time and reuse the debugfs file descriptor
that we already have.

Cc: Jeevan B <jeevan.b@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

igt_debugfs_simple_read() requires the debugfs file descriptor not
the drm device file descriptor, so the requirement check was complete
broken.

Cc: Jeevan B <jeevan.b@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Fixes: 99f04b14 ("igt: Another combinatorial exercise for blits")
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

The aim of this test is to combine gem_linear_blits, gem_tiled_blits etc
into one test runner that covers investigation into HW alignment issues
as well as driver boundaries (relocs, access, thrashing). This here is
but a start!

See also gem_concurrent_blits which looks for coherency issues between
parallel execution.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Katarzyna Dec <katarzyna.dec@intel.com>
Cc: Vanshidhar Konda <vanshidhar.r.konda@intel.com>
Reviewed-by: Vanshidhar Konda <vanshidhar.r.konda@intel.com>

If the device does not expose an aperture for indirect access with
detiling functionality, the kernel rejects an attempt to create such a
mapping with -ENODEV. If the ioctl is not supported, we can skip all of
our mmap_gtt specific tests.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Abdiel Janulgue <abdiel.janulgue@linux.intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

As the scratch buf is shared between the two requests on both engines,
we need to wait for both to finish using the buffer before we reset it.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

Add a subtest for DC3CO video playback case
to generate selective frame update and validate
that system stays in DC3CO state during execution.

v2: Changed PSR2 idle check to sleep check and addressed cosmetic changes.
v3: Renamed a function and restructured code.
v4: Cosmetic changes.
v5: Removed DC5 check, Platform check and a function parameter.
    Renamed a function name as per Arek and Imre's Comments.
v6: Added a new function require_dc_counter as per Arek's Comments.
    The test is now running based on time instead of the number of frames
    increased frame delay from 1 to 1.5 as per Imre's Comments.
v7: Removed psr2_active_sleep_check and corrected the switch indentation.
v8: Changed skip to assert in read_dc_counter function and changed
    IGT description

Signed-off-by: Jeevan B <jeevan.b@intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Reviewed-by: Anshuman Gupta <anshuman.gupta@intel.com>
Acked-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Link: https://patchwork.freedesktop.org/patch/338131/?series=66648&rev=8

Add DC3CO IGT validation prerequisites stuff
so we can enable DC3CO IGT test.

v2: Removed psr2_idle_wait_entry and get_psr2_status function.
v3: Changed macro definition and removed extra line.
v4: Simplified edp_psr2_enabled function as per Arek's comments.
v5: Simplified read_dc_counter function as per Arek's comments.

Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Signed-off-by: Jeevan B <jeevan.b@intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Reviewed-by: Animesh Manna <animesh.manna@intel.com>
Acked-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Link: https://patchwork.freedesktop.org/patch/338130/?series=66648&rev=8

Allow the subtests that want to limit the test to a single engine, to
limit the test to a single engine.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

After any process termination, use an rcu-barrier to be sure that any
deferred struct file cleanup has been performed. By being consistent in
our paranoia here means that we can rule out more false positives and so
focus on what remains.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Andi Shyti <andi.shyti@intel.com>

Find all the perf* selftests and expose them in our igt. In CI we expect
these not to be run in BAT (since they are just performance tests rather
than conformance), but expect the results to be available from the
shards.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Tomi Sarvela <tomi.p.sarvela@intel.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

The gem_exec_reloc basic subtests cover a couple dozen basic ABI
sanitychecks, aiming to prove the ABI works. While relocations used to
be essential, they are no longer the basis of current clients who all
softpin. It is a stagnant portion of the ABI that does not contribute
much to overall driver health (the complicated portions of relocs are
covered in the smoketests like gem_exec_parallel and gem_exec_gttfill).
However, even though each of the basic subtest is trivial and runs in
very little time, since CI is running each individually the time between
tests mounts up (especially on eMMC devices like fi-kbl-soraka).

By moving these tests to the shards we should speed up BAT on the
ratelimiting fi-kbl-soraka, while losing no coverage overall -- and
hopefully without losing any of the predictive failure coverage in BAT.
This stagnant bit of uAPI/uABI will remain checked by the irregular idle
runs.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tomi Sarvela <tomi.p.sarvela@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Martin Peres <martin.peres@linux.intel.com>
Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Acked-by: Petri Latvala <petri.latvala@intel.com>

Vanshidhar Konda authored 5 years ago and Chris Wilson committed 5 years ago

Add a test that show how to submit parallel spinners to all engines,
either from one context or a different context per engine; and verifies
that only one spinner is on each.

Signed-off-by: Vanshidhar Konda <vanshidhar.r.konda@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

The vblank tests try to make sure a relative vblank wait of
2 counts completes in ~1-2 frames (depends on how far along the
first frame we are when initiating the wait). The actual
limits we check are 0.9 to 2.1 frames to account for inaccuracies
in the reported dotclock. Unfortunately it seems we either
have even more inaccurate dotclocks or scheduling latencies
and whatnot are pushing that out quite a bit further. So let's
extend the end limit to ~2.6 frames to cut down on the noise a bit.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111609


Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>

Only test one plane of each type (HDR, SDR UV, SDR Y) on icl+.
For the purposes of validating pixel formats this should be sufficient
as the programming of each plane of the same type is identical and
as such it's unlikely we'd have different bugs between them.

As before we'll leave the full test set available to be run at
the user's discretion via --extended.

v2: Pimp comment (Petri)

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>

It seems the first rcu barrier may race with the addition of the file to
the rcu task list; so wait again.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Andi Shyti <andi.shyti@intel.com>

And initialise fence to -1 to avoid closing stdin (fd:0)!

The delayed fput is first queued with schedule (task_work) before being
rcu freed.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Andi Shyti <andi.shyti@intel.com>

Some time ago amdgpu changed their ABI to reject unknown compute rings,
so we should query the available set prior to execution.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

Update to the newest version.

Python 3 is now the default one, so we can drop the suffix from pip
invocation and we have to install it explicitly.

Cc: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>

For this test, we need a laptop running on battery power so that we can
read the battery charge level before and after suspend. And then wait
long enough for a reliable measure.

References: https://bugs.freedesktop.org/show_bug.cgi?id=111909


Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

Sanity test existing persistence and new exciting non-persistent context
behaviour.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Jon Bloomfield <jon.bloomfield@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Andi Shyti <andi.shyti@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>

Expose a new context parameters to opting out of persistent behaviour.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>

Update to commit a0e047156cde ("drm/i915/gem: Make context persistence
optional") for I915_CONTEXT_PARAM_PERSISTENCE

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Andi Shyti <andi.shyti@intel.com>

Janusz Krzysztofik authored 5 years ago and Chris Wilson committed 5 years ago

Name of the i915 module parameter providing fault injection function is
changing for consistency with a new convention of naming i915 driver
internal functions called from the driver PCI .probe entry point.  Adjust
the test to use the new name.

[ickle: keep the old parameter name around for stable testing]

Suggested-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Cc: Michał Wajdeczko <michal.wajdeczko@intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Piotr Piórkowski <piotr.piorkowski@intel.com>
Cc: Tomasz Lis <tomasz.lis@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>