| Age | Commit message (Collapse) | Author |
|---|
|
Noticed while running some tests that adding any kind of spaces into the
name of a test run would stop igt_resume from working for said test run.
Turns out that when we parse test metadata, we accidentally use the
'%ms' specifier with fscanf() which finishes parsing strings when any
kind of whitespace is encountered.
So, fix this by using the proper %m[^\n] specifier, which dynamically
allocates it's result and doesn't stop reading the string until a
newline is encountered. Additionally, add a test for this.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
_open_pmu() was calling perf_igfx_open_group() which in turn was
calling igt_perf_type_id("i915") each time.
By making _open_pmu() take type as a parameter and call
igt_perf_open_group() we achieve two things:
* we use the correct type for engines->device instead of the
hardcoded "i915"
* optimization - we call igt_perf_type_id() once - it's reading
sysfs
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Signed-off-by: Ayaz A Siddiqui <ayaz.siddiqui@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
__cancel_work_at_exit writes -1 to param "reset", with the intention
that it's "any available method". Hex values to numeric params should
be prefixed with 0x to be parsed as a number. Use the convention of %u
with -1 as is used elsewhere with the "reset" param.
Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Sometimes creating the string representation fails. This usually happens
when we have a huge logs (e.g.: something was spamming the dmesg) or the
result generation was run on a very low-end system (e.g. embedded board
with 256 megs of RAM).
Sadly json-c call returns us NULL and provides no explanation
whatsoever. Let's fix a NULL pointer dereference in such cases and print
a mesage that should help people make sense out of what have just
happened.
Cc: Swati Sharma <swati2.sharma@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>
|
|
In commit b943d64f95bc ("kms_atomic_transition: Make the available
pipe discovery dynamic") the subtests of kms_atomic_transition were
split into dynamic subtests. What avoided scrutiny however was that
the looping uses for_each_pipe_with_valid_output, which gives you
every valid pipe/output combo, so just using the pipe name as the
dynamic subtest name causes duplicates, resulting in weird CI results.
As the nature of the original test was a stress test for watermarks,
keep the full combo rather than switching to
for_each_pipe_with_single_output, but use the output name as well in
the dynamic subtest name.
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>
|
|
This allows the test to work on setups where the primary plane cannot
be disabled while the CRTC is enabled.
It should make no difference for the output.
Signed-off-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
It got removed in Python 3.9 and ever since Python 3.3 it's:
from xml.etree.ElementTree import *
Link: https://docs.python.org/3.9/whatsnew/3.9.html#removed
Issue: https://gitlab.freedesktop.org/drm/igt-gpu-tools/-/issues/86#note_520308
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@linux.intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
In intel_gpu_top device path was hard coded for integrated GPU.
With this patch we:
* use igt_device_scan library for device scanning,
* make discrete GPU the default one,
* provided options for card selection.
v2:
* explicitly set ret to EXIT_SUCCESS after all the other uses
* fix use after free of opt_device (Tvrtko)
* use EXIT_FAILURE instead of "1" (Tvrtko)
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Signed-off-by: Ayaz A Siddiqui <ayaz.siddiqui@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Will be used in the next patch.
1. set_pci_slot_name(): stores PCI_SLOT_NAME from prop to device
2. igt_device_find_first_discrete_card(): try to find first discrete GPU
3. igt_devices_free(): Free device buffers created during scan
Signed-off-by: Ayaz A Siddiqui <ayaz.siddiqui@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
igt_device_scan can now be used as a separate library which only depends
glib and libudev - some IGT internals are being stubbed in this case.
v2: (mostly) sort includes (Lucas)
Cc: Lucas De Marchi <lucas.de.marchi@gmail.com>
Signed-off-by: Ayaz A Siddiqui <ayaz.siddiqui@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
Return value of udev_enumerate_add_match_property() was not being checked
in scan_drm_devices().
Signed-off-by: Ayaz A Siddiqui <ayaz.siddiqui@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Since preempt-reset is 640ms we have to wait at least that long to
recover from a hostile client.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Remove all "localized" functions in media spin and libdrm usage
in the test.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
For some scenarios we want to keep previous objects and their offsets
and recreate only batchbuffer object. To allow user do that add
bb reset function which can or not purge collected objects from
previous run.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
To avoid relocations when intel_buf is used we need to keep previous
offset and context. Add addr structure with offset and ctx fields.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
After replugging a connector and the kernel reports the connector as
connected the cached modes returned by drmModeGetConnectorCurrent() (the
kernel's cached list of modes) may be stale. The modes will be only
updated by a full reprobe. Ensure this by setting output->force_reprobe
which will result in calling drmModeGetConnector().
This fixes sporadic test failures due to seeing no modes unexpectedly.
So far things happened to work, because some in-kernel user of the
output - like fbdev - does a full reprobe sometime after the hotplug
event.
Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
|
|
If we fail to fork (e.g. due to restrictive limits) -1 gets written as
PID of our child and we assert out.
The cleanup that happens afterwards tries to kill all our children,
which ends up in kill(-1, SIGKILL) which is not good.
This patch makes sure of two things:
* -1 doesn't get written down as our child
* when we are killing children we make sure that pid > 0
Reported-by: Fei Yang <fei.yang@intel.com>
Cc: Michael Hebenstreit <michael.hebenstreit@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Tested-by: Michael Hebenstreit <michael.hebenstreit@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
The gem_create_v2 uapi was never merged, which would have been a nice
addition to allow userspace to utilise stolen memory. Since it can only
get in the way at this point, let's just remove it.
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Disable the pre-parser on Tigerlake as we are dynamically rewriting the
batches and so need to disable the parser cache.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
DRM logging in the kernel has switched to device based logging,
changing the pattern of logging from
[drm:function_name] Message
to
i915 0000:00:02.0: [drm] Message
Change the pattern we use with --piglit-style-dmesg to match the new
style logging, catching all warnings that contain [drm]. We're no
longer result-wise exactly piglit compatible but then again that ship
sailed long time ago with the change of introducing dmesg-level
parameter. Piglit compatibility result-wise is not a goal regardless.
Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
|
|
This tests is being sporadically skipped in CI as it is not due
"PSR sink not reliable: yes" lets print the i915_edp_psr_status to
find out the reason.
This can be reverted afterwards.
Reference: https://gitlab.freedesktop.org/drm/intel/-/issues/1911
Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
Not sure why but in recent kernels + IGT when PSR status debugfs is
read the frame already passed, so the su blocks is set in frame 1
and causing the test to fail for page flips.
So here reading from frame 1 if frame 0 has 0 blocks, as this test
always changes screen with the same number of su blocks it is not a
issue.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/608
Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
If by some reason or tests, this tests is executed and VT mode is
already in KD_GRAPHICS the call to kmstest_set_vt_graphics_mode() will
set orig_vt_mode as KD_GRAPHICS and when it was calling
kmstest_restore_vt_mode() it would set KD_GRAPHICS again not returning
to fbcon and causing the test to fail.
As it can be seen here:
(kms_fbcon_fbt:11004) igt_kms-DEBUG: VT: graphics mode set (mode was 0x1)"
https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_499/fi-ehl-1/igt@kms_fbcon_fbt@fbc.html
So here adding a new function to alaways set mode the KD_TEXT when we
want to restore to fbcon.
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
Remove libdrm functions and replace them with new ones after removing
_v2 suffix introduced for transition state.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
To allow remove "_v2" suffix in functions add this not-fully-clean
step and localize names. Some additional requirement is needed to
fulfill its needs (passing batchbuffer from the caller side).
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
After Chamelium does an FSM signaling by deasserting its HPD, in
response to which the test's FSM handler disables the output with a
DPMS-off, we have to make sure that Chamelium has reasserted its HPD
before we re-enable the output with DPMS-on (for instance to avoid link
training errors, or enable the output in the wrong TypeC mode). To
ensure this wait for the connector state to become asserted.
On TypeC connectors with an enabled mode the IOM firmware will signal a
connected state (via a connect hotplug interrupt delivered to the
driver) in a deferred way only after the mode is disabled. So wait for
the connected state after DPMS-off.
Reported-and-tested-by: Kunal Joshi <kunal1.joshi@intel.com>
Cc: Hiler Arkadiusz <arkadiusz.hiler@intel.com>
Cc: Kunal Joshi <kunal1.joshi@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
|
|
Be careful and check the engine support store-dw as we are using
__igt_spin_new() which skips the requirement checks.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
The test works fine if the scheduler has semaphores, strangely only
being skipped for old devices adding semaphore support.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Can't test queues, if we don't have queues.
The queue ordering tests checked for scheduler support, assuming that
would mean we had full-ppgtt. Check for queues in case we solve the
scheduler problem on gen6/gen7.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Add a pass that imposes a delay between submitting the master and its
bonded pair. This should make it more likely that another thread is able
to submit in the interval.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Since we create a few GPU objects in the locals, if we decided to skip
this bonded-sync phase, we need to jump to the cleanup.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Add another bonded-runner variant for a pair of inter-locked
non-preemptible waiters (each batch waits to be cancelled by the other),
as we are trying to replicate the HW feature where a bonded pair is
locked on HW and cannot be preempted until all members of the bond have
executed (i.e. all must execute in lockstep).
The key difference with this variant is that both halves contain a
non-preemptible spin until the other half has executed.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Randomly submit a paired spinner and its cancellation as a bonded
(submit fence) pair. Apply congestion to the engine with more bonded
pairs to see if the execution order fails. If we prevent a cancellation
from running, then the spinner will remain spinning forever.
v2: Test both immediate submission and fenced submission
v3: Copy-n-paste a single context variant
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>
|
|
For determining whether or not we flip too fast, we don't have to wait
long, just long enough. A ballpark measure of how long is reasonable is
given by how fast we perform the warmup, as that is 3 flips/modesets.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
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>
|
|
If we are using GPU relocations, then the batch may legitimately be marked
as having been written to by the GPU. As this is expected, remove the
write marker, which may be a different engine to that we execute on, and
the corresponding read engine if different.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Actually remember to unplug to the busy buffer after waiting for the
flip to complete early. Otherwise we wait for the hangcheck to kick in,
which while effective at testing the flip recovery for hanging buffers
is not the intent of the test and wastes 40s in BAT.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
The hangcheck/reset modparam has far reaching effects and disables
functionality if switch off. This can surprise a few tests causing them
to skip.
References: https://gitlab.freedesktop.org/drm/intel/-/issues/1929
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
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>
|
|
The cmd-parser causes a disturbance in our priority scheme. It uses a
CPU worker that does not inherit the context priority and so may use a
simple FIFO that causes the requests to be ready (and so executed) in
parser order rather than our expected inherited priority order. If we do
not need to wait for relocations, the cmdparser is run inline, i.e.
before the test unleashes the fence holding the requests back.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Randomise the position of the spinner to reduce the number of
relocations we might require.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
If we use GPU relocations, then we will mark the batch as being written
to by the GPU. This can be any engine, and so may perturb the reported
busyness. Take this into account before reporting an error.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
The kernel no longer uses semaphores between engines, unless it can do
so by preempting them with timeslices. Update the semaphore-busy to only
run when we expect semaphore usage, i.e. not on bdw/bsw.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/1939
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Function is now libdrm-free, but mixes calls to replaced and "_v2"
versions. This cleanup and old code removal will be addressed
in the another commit.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
As gem_media_vme uses mostly gen11 there's no need to add _v2
version and migrate from intel_batchbuffer to intel_bb to remove
libdrm dependency.
Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
While we may chide userspace if they try to use the same batches from
multiple threads (the order of operations is undetermined), we do try to
ensure that each ioctl appears to be atomic from the perspective of
userspace.
In particular, relocations within execbuf are expected to be consistent
for the executing batch. That is we want the relocations applied by
this execbuf to be visible for the associated batch, and we especially
do not want to execute the batch with conflicting relocations from
another thread.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
kms_flip's exit handler was wrapped in an igt_fixture some 7 years ago
because "it can fail". That's not a problem anymore, as igt_fail()
will explicitly handle failing in an exit handler.
Rather, using igt_fixture in an exit handler is harmful because on a
subtest failure, we don't always keep track of the currently running
context (whether we're inside a subtest anymore) and we forcefully
assert correct nesting of IGT magic control blocks now.
We already know the state is cleanable because we only install the
handler on a successful init. Unconditional cleanup in an exit handler
is the correct thing to do in general.
Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
If our CPU client is very slow to notice that the GPU spinner has
started, we may consume the full heartbeat interval without noticing.
This is bad if we are trying to test that a client that yield within the
heartbeat interval is not selected for termination.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/1879
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Start our preparations for guaranteeing endless execution.
First, we just want to estimate the direct userspace dispatch overhead
of running an endless chain of batch buffers. The legacy binding process
here will be replaced by async VM_BIND, but for the moment this
suffices to construct the GTT as required for arbitrary indirect
execution.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
|
|
Set NO_PREEMPTION on the rcs0 hanger so that we are guaranteed
(hopefully, if nothing else breaks!) to timeout.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
