Age | Commit message (Collapse) | Author |
|
kms_frontbuffer_tracking should test PSR the same way that kms_psr does.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
kms_frontbuffer_tracking and kms_psr test PSR in different ways, let'
fix that by creating common library functions.
v2: Include the new file in meson.build
v3: Leave --no-psr intact (Rodrigo)
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
Change the function arguments to not rely on test specific data as
the following patches change kms_frontbuffer_tracking to reuse PSR
functions.
v2: Leave --no-psr intact (Rodrigo)
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
"drm/i915: Kill sink_crc for good" removes the kernel support for sink
crc.
References: https://patchwork.freedesktop.org/series/46039/
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
This was always a placeholder for GVT stakeholders to provide some
better tests. 2 years later and none have been put forward so stop
wasting CI's time running a placeholder.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106989
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Zhi Wang <zhi.a.wang@intel.com>
Acked-by: Martin Peres <martin.peres@linux.intel.com>
|
|
Before running a hang test, check if we can inject a hang and expect
recover to work. If we can't control a hang, skip the subtest.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
As we try to blit into the buffer to check CPU<->GPU snooping, we have
to check we have a functional GPU first.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
|
|
We don't use sink CRC anymore in this test.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
Wait for PSR HW status to show active before starting an PSR exit action.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
eDP sink crc reads use vblank interrupts that cause PSR exit and
therefore makes them unsuitable for PSR testing. Besides that, reading
sink CRC via the AUX channel for testing when the HW also is most likely
is going to be using AUX channel is a recipe for inconsistent test
results. Thirdly, CRC's have been seen to be noisy/inconsistent across
sinks. We tradeoff the ability to validate what the sink is displaying
for correctness.
We also make use of source PSR status register to check whether HW tracking
triggered PSR exit upon an exit event.
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
eDP sink crc reads use vblank interrupts that cause PSR exit and
therefore makes them unsuitable for PSR testing. Besides that, reading
sink CRC via the AUX channel for testing when the HW also is most likely
is going to be using AUX channel is a recipe for inconsistent test
results. Thirdly, CRC's have been seen to be noisy or vary across sinks for
the same driver and test code. We tradeoff the ability to validate what the
sink is displaying for correctness.
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
|
|
An hanging batch is nothing more than a spinning batch that never gets
stopped, so re-use the routines implemented in dummyload.c.
v2: Let caller decide spin loop size
v3: Only use loose loops for hangs (Chris)
v4: No requires
v5: Free the spinner
v6: Chamelium exists.
Signed-off-by: Antonio Argenziano <antonio.argenziano@intel.com> #v3
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Antonio Argenziano <antonio.argenziano@intel.com>
|
|
Setup a userptr object that only has a read-only mapping back to a file
store (memfd). Then attempt to write into that mapping using the GPU and
assert that those writes do not land (while also writing via a writable
userptr mapping into the same memfd to verify that the GPU is working!)
v2: Pull the random batch construction into a routine to avoid
duplication.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Time intervals as produced by timersub() are normalized to have
the tv_usec in the range 0-999999. That leads to very confusing
looking debug output for negative interval. Eg. an interval
of -0.1 seconds would be represented as tv_sec=-1, tv_usec=900000.
Let's just convert the thing to a float seconds value and print
that so that we'll get less confusing debug output.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
As we want to make the buffers active on the GPU before removing their
fence, an operational GPU (not wedged!) is required.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
As render copy wants to use the GPU, we should make sure it is not
wedged first.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
As we need GEM and the GPU to do a GPGPU fill, we should check that it
is operable before using -- skipping rather than failing when the device
is wedged.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
We use options!=NULL to determine if we should require the module to be
reloaded and fail if we find it already loaded. In pm_rpm, we are only
ensuring the MSR module is loaded, and only want default options.
Fixes: 4dc2ce0e ("lib/kmod: Fail if the module is already loaded")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
|
|
pthread_create() expects a void *(*start_routine) (void *). Fix warning
the following warning on gcc 8:
../tests/sw_sync.c:773:37: warning: cast between incompatible function types from ‘int (*)(void *)’ to ‘void * (*)(void *)’ [-Wcast-function-type]
pthread_create(&threads[i], NULL, (void * (*)(void *))
^
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
gem_render_copy requires a working GPU so check first.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
Mark up gem_pwrite_pread's dependence on a functioning GPU, by calling
igt_require_gem in its setup fixture.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Distinguish between the latency required to switch away from the
pollable spinner into the target nops from the client wakeup of
synchronisation on the last nop.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
|
|
To further defeat any contemplated spin-optimisations to avoid the irq
latency for synchronous wakeups, increase the queue length.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
|
|
Apply a different sort of stress by timing how long it takes to sync a
second nop batch in the pipeline. We first start a spinner on the
engine, then when we know the GPU is active, we submit the second nop;
start timing as we then release the spinner and wait for the nop to
complete.
As with every other gem_sync test, it serves two roles. The first is
that it checks that we do not miss a wakeup under common stressful
conditions (the more conditions we check, the happier we will be that
they do not occur in practice). And the second role it fulfils, is that
it provides a very crude estimate for how long it takes for a nop to
execute from a running start (we already have a complimentary estimate
for an idle start).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
|
|
When using the pollable spinner, we often want to use it as a means of
ensuring the task is running on the GPU before switching to something
else. In which case we don't want to add extra delay inside the spinner,
but the current 1000 NOPs add on order of 5us, which is often larger
than the target latency.
v2: Don't change perf_pmu as that is sensitive to the extra CPU latency
from a tight GPU spinner.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Antonio Argenziano <antonio.argenziano@intel.com> #v1
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> #v1
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
In order to make adding more options easier, expose the full set of
options to the caller.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Add a new subtest that does renders the test pattern into a
compressed buffer. And we'll follow it up with another copy
back to an uncompressed buffer so that we also test the
capability to sampled from compressed buffers, and also so
that we can actually compare the results against the reference
image.
We'll also do a quick check of the aux surface to check that
it actually indicates that at least some parts of the buffer
were in fact compressed. Further visual verification can be
done via the dumped png.
v2: Test various tiling formats with CCS as well
Combine the ccs test into the same function as
the rest
Pass the correct thing to intel_gen()
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Make sure our rendercopy implementations do the right thing with
tiled buffers.
For now we'll just do linear->linear, x-tiled->x-tiled, and
y-tiled->y-tiled. Not sure there's much point in adding tests
for different src vs. dst tiling modes?
v2: Test all tiling combos (Chris)
Allocate with drm_intel_bo_alloc_tiled() (Chris)
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Store a bit of aux surface state in igt_buf. This will be needed
for rendercopy AUX_CCS_E color compression.
We also have to sprinkle memset()s and whatnot all over to make
sure the current igt_buf users don't leave the aux stuff full
of stack garbage.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
As we want to compare a templated tiling pattern against the target_bo,
we need to know that the swizzling is compatible. Or else the two
tiling pattern may differ due to underlying page address that we cannot
know, and so the test may sporadically fail.
References: https://bugs.freedesktop.org/show_bug.cgi?id=102575
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>
|
|
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
If the GPU is not usable, we will not be able to submit workloads to be
measured and so observing them will fail.
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>
|
|
Checking whether we can copy solid rectangles isn't particularly
robust. Eg. errors in texture coordinates/interpolation wouldn't
necessarily show up at all because all texels are identical.
Let's switch to a more elaborate pattern that should catch such
errors. And we'll also change the test to not start the copy
from position 0,0 in the texture.
We'll generate the reference image (against which the rendercopy
results are compared) by peforming an identical copy using the
cpu.
v2: Use gtt mmap instead of pread/pwrite (Chris)
Offset the dst coordinates by -1,-1 to make sure
the copy doesn't go past the intended region
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
If we allow 5s (+20% systematic error) to construct the fences, we may
reasonably assume that it will take equally as long to consume them. As
we only have 10s before the vgem fence times out, there is no margin of
safety. Err on the side of safety and reduce it down to 2s, we won't test
importing as many fences simultaneously, but after the first full pages
we should be good at spotting errors!
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
The goal of gem_exec_gttfill is to exercise execbuf under heavy GTT
pressure (by trying to execute more objects than may fit into the GTT).
We spread the same set of handles across different processes, with the
result that each would occasionally stall waiting for execution of an
unrelated batch, limiting the pressure we were applying. If we using a
steaming write via a WC pointer, we can avoid the serialisation penalty
and so submit faster.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
These tests exercise an obscure piece of libdrm_intel API to keep the
number of VMA opened in a process under a certain limit (only ever used
by UXA as once upon a time we ran into the limit with many, many 1x1
pixmaps or something like that). The tests are not exercising kernel API
(or ABI) and need to check their resource requirements first. However,
since they are only testing libdrm_intel, remove them from igt as they
would better belong in a libdrm_intel test suite.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106010
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Martin Peres <martin.peres@linux.intel.com>
Reviewed-by: Antonio Argenziano <antonio.argenziano@intel.com>
|
|
Distributions want explicit control over optional parts so they can
state runtime dependencies before building. Let's restore the
functionality autotools used to provide.
Where possible, the selection is done by choosing whether to build a
particular item and the option name is build_$item. Example:
build_overlay. Where not possible, the option name is
with_$item. Example: with_valgrind.
Array options require a bump of required meson version to 0.44. Debian
stable has meson 0.37 which is already too old, stable-backports has
0.45, CI uses 0.45. Mesa's meson requirement is 0.44.1, for a
perspective.
Note, the old hack for not building docs when cross-compiling is
gone, as doc building can be explicitly controlled now.
v2: glib not optional
v3: bump meson version to 0.44
Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
This make perf tests to run in Icelake.
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
|
|
This patch fix the following gcc warnings:
warning: ISO C90 forbids mixed declarations and code
[-Wdeclaration-after-statement] [..]
igt_color_encoding.c:45:2: warning: ISO C90 forbids mixed declarations
and code [-Wdeclaration-after-statement] [..]
igt_color_encoding.c: In function ‘ycbcr_to_rgb_matrix’:
igt_color_encoding.c:72:2: warning: ISO C90 forbids mixed declarations
and code [-Wdeclaration-after-statement] [..]
Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
|
|
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Katarzyna Dec <katarzyna.dec@intel.com>
|
|
We want to make sure RT tasks which use a lot of CPU times can submit
batch buffers with roughly the same latency (and certainly not worse)
compared to normal tasks.
v2: Add tests to run across all engines simultaneously to encourage
ksoftirqd to kick in even more often.
v3: More passes to improve measurement stability.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> #v1
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Wait until the previous nop batch is running before submitting the next.
This prevents the kernel from batching up sequential requests into a
a ringfull, more strenuous exercising the "lite-restore" execution path.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Katarzyna Dec <katarzyna.dec@intel.com>
|
|
As we hang ctx0 quite frequently, it needs to be harden against being
banned.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Antonio Argenziano <antonio.argenziano@intel.com>
|
|
During a review came across a line of commented code. No specific reason
for the line is given so remove it.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Antonio Argenziano <antonio.argenziano@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
sw_sync/sync_multi_consumer_producer was communicating between threads
using the sw_sync ioctl and manipulating a shared volatile counter.
However, the ioctl itself does not imply a memory barrier, and so
different CPUs may see different states of the counter (the volatile
making GCC perform the operation in stages making the race even more
likely). Instead of using volatile, use locked operations to make the
counter manipulation thread-safe.
References: https://bugs.freedesktop.org/show_bug.cgi?id=106344
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
|
|
We're little bit too enthusiastic in our initial attempt to lock all
available memory. Let's use the mlock probe from lib rather than trying
to lock everything that sysinfo.freeram has to offer.
Note that we're only tweaking the initial step - it's still possible
that we're going to get killed later on.
v2: Just increment lock instead of modifying addr passed to mlock
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ewelina Musial <ewelina.musial@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ewelina Musial <ewelina.musial@intel.com>
|
|
We already have the routine we need in drv_suspend. Let's move it to lib
and use it in the mlocking tests. We can also make it a bit faster if we
tweak the initial step and initial amount.
(I think it's safe to assume that we should be able to lock 3/4
of RAM, this cuts the probe time on my 32G SKL - from ~530s to ~180s)
v2: Use available mem, amend step, also lock outside of fork,
early exit if the assumption is wrong (Chris)
Update the function name in doc (Ewelina)
v3: Total for pin, available for initial lock (Chris)
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ewelina Musial <ewelina.musial@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ewelina Musial <ewelina.musial@intel.com>
|
|
All subtests send a workload to the engines and then return without
waiting on it, while this is not a problem because the test targets the
API, it makes the hang detector pointless since the driver will declare
an hang long after the test has completed.
v2:
- Use common functions to create/terminate a batch. (Chris)
Signed-off-by: Antonio Argenziano <antonio.argenziano@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
As the swizzling is baked into the tiling pattern, the swizzling has to
be consistent across the entire GTT mmap for our tests to work. However,
under L-shaped memory configurations on older architectures, the
swizzling varied depending on which region the page found itself in --
invalidating our assumptions and ability to predict the tiling pattern.
Reported-by: Adric Blake <promarbler14@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106848
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Recently we discovered that we have a race between swapping and
suspend in our resume path (we might be trying to page in an object
after disabling the block devices). Let's try to exercise that by
exhausting all of system memory before suspend.
v2: Explicitly share the large memory area on forking to avoid running
out of memory inside the suspend helpers (for they fork!)
References: https://bugs.freedesktop.org/show_bug.cgi?id=106640
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tomi Sarvela <tomi.p.sarvela@intel.com>
Reviewed-by: Antonio Argenziano <antonio.argenziano@intel.com>
|