| Age | Commit message (Collapse) | Author |
|---|
|
After closing the perf stream the parking the GPU engines may easily
take more than 1 second: releasing the FD itself results in a new
request submission via i915_perf_release()->i915_oa_stream_destroy()->
gen8_disable_metric_set(). That means a >1sec delay for the delayed
unpark to be called due to the delay from
queue_delayed_work(retire_work, round_jiffies_up_relative(HZ))
+ the delay from
mod_delayed_work(idle_work, msecs_to_jiffies(100))
Scheduling may push this delay even further, I measured >2sec delays on
my GLK.
Fix this by calling gem_quiescent_gpu() which syncs up with the idle
work, thus making sure we'll see RC6 residency afterwards.
v2:
- Use gem_quiescent_gpu() instead of increasing the timeout. (Chris)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103179
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
|
|
1<<31 (same as 2<<30) is undefined behaviour in C. When compiling with
GCC and UBSan, it gives this error:
../tools/intel_reg_decode.c: In function ‘ivb_debug_port’:
../tools/intel_reg_decode.c:398:3: error: case label does not reduce to an integer constant
case PORT_DBG_DRRS_HW_STATE_HIGH:
^~~~
This happens because 1<<31 isn't representable as a signed int. Instead,
use an unsigned int.
Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
Clang fails to compile this file with this error:
../tests/kms_vrr.c:203:20: error: suggest braces around
initialization of subobject [-Werror,-Wmissing-braces]
drmVBlank vbl = { 0 };
^
{}
As discussed in [1], using an empty initializer list is the preferred
way to fix this.
[1]: https://lists.freedesktop.org/archives/igt-dev/2019-March/010841.html
Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
When content protection authentication is failed in kernel after all
requried retries, before declaring the test failure, set the content
protection to UNDESIRED state.
This will avoid the HDCP authentication attempts in subsequent modesets
from other tests.
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110376
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110224
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Using a filter that doesn't match any test name resulted in the runner
silently failing. Print an error message so that the user understands
why the runner fails.
Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
We have to control the cache domains, especially important before first
writing into the object.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
As debugfs_fd is used even after the teardown_drm(), it is not closed
in this function or in any other place. And then it gets worse as
every call to setup_drm() will open another file descriptor to
debugfs dir.
So lets move the opening of debugfs_fd to setup_environment() and only
open it once and close it before leave the test.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
If a subtest fail before the teardown_drm() call it will keep the
driver open as master causing the test to fail when trying to open
the driver again as master, so lets share the drm_info struct and
check if driver is already open.
Starting subtest: psr
(kms_fbcon_fbt:5270) CRITICAL: Test assertion failure function subtest, file ../tests/kms_fbcon_fbt.c:259:
(kms_fbcon_fbt:5270) CRITICAL: Failed assertion: feature->wait_until_update(drm.debugfs_fd)
Stack trace:
#0 ../lib/igt_core.c:1474 __igt_fail_assert()
#1 ../tests/kms_fbcon_fbt.c:261 subtest()
#2 ../tests/kms_fbcon_fbt.c:316 __real_main309()
#3 ../tests/kms_fbcon_fbt.c:309 main()
#4 ../csu/libc-start.c:325 __libc_start_main()
#5 [_start+0x29]
#6 [<unknown>+0x0]
Subtest psr failed.
**** DEBUG ****
(kms_fbcon_fbt:5270) drmtest-DEBUG: Test requirement passed: !(fd<0)
(kms_fbcon_fbt:5270) igt_debugfs-DEBUG: Opening debugfs directory '/sys/kernel/debug/dri/1'
(kms_fbcon_fbt:5270) DEBUG: Test requirement passed: drm->res
(kms_fbcon_fbt:5270) igt_kms-DEBUG: VT: graphics mode set (mode was 0x0)
(kms_fbcon_fbt:5270) DEBUG: Test requirement passed: feature->supported_on_chipset(drm.debugfs_fd)
(kms_fbcon_fbt:5270) CRITICAL: Test assertion failure function subtest, file ../tests/kms_fbcon_fbt.c:259:
(kms_fbcon_fbt:5270) CRITICAL: Failed assertion: feature->wait_until_update(drm.debugfs_fd)
(kms_fbcon_fbt:5270) igt_core-INFO: Stack trace:
(kms_fbcon_fbt:5270) igt_core-INFO: #0 ../lib/igt_core.c:1474 __igt_fail_assert()
(kms_fbcon_fbt:5270) igt_core-INFO: #1 ../tests/kms_fbcon_fbt.c:261 subtest()
(kms_fbcon_fbt:5270) igt_core-INFO: #2 ../tests/kms_fbcon_fbt.c:316 __real_main309()
(kms_fbcon_fbt:5270) igt_core-INFO: #3 ../tests/kms_fbcon_fbt.c:309 main()
(kms_fbcon_fbt:5270) igt_core-INFO: #4 ../csu/libc-start.c:325 __libc_start_main()
(kms_fbcon_fbt:5270) igt_core-INFO: #5 [_start+0x29]
(kms_fbcon_fbt:5270) igt_core-INFO: #6 [<unknown>+0x0]
**** END ****
Subtest psr: FAIL (0.845s)
Starting subtest: fbc-suspend
Test requirement not met in function igt_device_set_master, file ../lib/igt_device.c:55:
Test requirement: __igt_device_set_master(fd) == 0
Can't become DRM master, please check if no other DRM client is running.
Subtest fbc-suspend: SKIP (0.001s)
Starting subtest: psr-suspend
Test requirement not met in function igt_device_set_master, file ../lib/igt_device.c:55:
Test requirement: __igt_device_set_master(fd) == 0
Can't become DRM master, please check if no other DRM client is running.
Subtest psr-suspend: SKIP (0.000s)
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
When fbcon is enabled, PSR will be active between cursor blinks so
what it should really use to test PSR is psr_wait_entry(), so a new
feature callback was added.
But the fbcon cursor blinks at 5hz what give us 200ms between each
screen update what make psr_wait_update() prone to fail the test
because it timed out before a blink could happen, so here adding and
using psr_long_wait_update() that have a longer timeout.
v3:
- 3 previous patches squashed in this one (Maarten)
- Back to !feature->wait_until_enabled() to test feature state when
all CRTCS are disabled(Dhinakaran)
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Dhinkaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
If cursor blinking is disabled no screen updates will happen and
fbcon_fbt subtests will fail, so lets enable cursor blink while
running this test and restore to the previous value when exiting it.
v4:
- renaming restore fd (Dhinakaran)
- saving previous value as char (Dhinakaran)
- skipping test if not able to open cursor blink file (Dhinakaran)
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
9c4114ec5d87 (lib: Always unbind the fbcon around igt) broke
fbcon_fbt test as fbcon is not allowed to bind when executing any IGT
test, so lets allow it again just for this test.
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
kms_fbcon_fbt was doing its own handling to wait for PSR to get
enabled while it is already available in lib.
v2: splitted previous patch into this one and the previous one(Dhinakaran)
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
|
|
Tests various cursor plane interactions with primary and overlay planes.
Correctness is verified by comparing CRC values to reference buffers
drawn on the primary plane.
There are existing generic multi-plane interaction tests but these
don't test positional output on overlapping planes or extensively
cover edge and corner cases for offseting and positioning planes based
on DRM parameters.
On hardware without dedicated cursor planes (such as AMDGPU) these tests
are helpful for verifying software calculations done for positioning the
cursor plane.
Cc: Leo Li <sunpeng.li@amd.com>
Cc: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Leo Li <sunpeng.li@amd.com>
|
|
Failed to configure intel-gpu-tools with error:
meson.build:151:0: ERROR: Could not generate cargs for libdw:
libdw depends on liblzma.
Installation of liblzma-dev fixes issue.
At least it has to be pointed in Readme.
v2: just readme updated
Cc: Petri Latvala <petri.latvala@intel.com>
CC: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Sergii Romantsov <sergii.romantsov@globallogic.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
Tests that eat all of the RAM and then some to invoke the oom-killer
deliberately sometimes cause extra casualties. Make sure the runner
stays alive.
Signed-off-by: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
The driver is supposed to update the size when it's trying to outsmart
userspace. Let's test it.
v2: Rework to use local ioctl wrapper (Chris)
v3: s/16/15, move test a bit earlier (Chris)
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Janusz Krzysztofik <janusz.krzysztofik@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
For fast-feedback, we need to sanitycheck that CRC reads work, otherwise
the display validation tests are flawed. However, we don't need to
extensively test each pipe as the HW is fundamentally the same. Instead,
do the full suite of tests on one pipe, and just verify that the interface
works for the other pipes. This should speed BAT up by reducing
NUM_PIPES suspend and resume cycles to just 1.
The shards should cover the extended test suite on all pipes for
complete validation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Martin Peres <martin.peres@free.fr>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Acked-by: Martin Peres <martin.peres@free.fr>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
With some upcoming changes i915 might not allow
all sprite planes enabled, depending on available
bandwidth limitation. Thus the test need to decrement
amount of planes and try again, instead of panicking.
v2: Removed excessive nested conditions, making code a bit
more readable(hopefully).
v3: Stopped using global n_planes variable as it might cause
resources being unreleased.
Using now parms[i].mask as a way to determine if plane
has to be included into commit.
v4: Removed unneeded n_overlays initialization.
v5: Randomize which of sprite planes to remove if hitting
resource limits.
v6: Replace igt_warn with igt_info, to make IGT tests happier.
v7: Removed unneeded retry logic, made plane random removal simplier.
Great thanks to Maarten Lankhorst for suggestions.
v8: Fail if we have less than 3 planes
v9: Remove unneeded diffs
Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
While fixing used amount of planes, discovered that if
wm_setup_plane is called with specific parameter that
gives parms[i].mask & mask == 0 for all used planes,
then subsequent wait_transition fails in assertion on fd_completed.
So added return value to wm_setup_plane, which would allow to
determine, if we really need to wait for any transitions.
v2: Fixed commit message, to properly describe a reasoning for
wm_setup_plane changes.
Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
Using igt_plane_has_format_mod allows to save
time not doing redundant retry, knowing if
plane supports alpha beforehand.
v2: Remove unneeded assertion also.
Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
This patch saves off the original value of force_dsc_en and
restores it back after each test and in the igt exit handler
so that it gets restored on any failure/assertion.
Suggested-by: Imre Deak <imre.deak@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Anusha Srivatsa <anusha.srivatsa@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
DSC enable gets configured during compute_config and needs
a full modeset to force DSC.
Sometimes in between the tests, if the initial output is same as the
mode being set for DSC then it will not do a full modeset.
So we disable the output before forcing a mode with DSC enable.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110270
Fixes: db19bccc1c22 ("test/kms_dp_dsc: Basic KMS test to validate VESA DSC on DP/eDP")
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Anusha Srivatsa <anusha.srivatsa@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
If we unexpectedly pass, we get a cryptic
(kms_flip:935) CRITICAL: Test assertion failure function set_flag, file kms_flip.c:271:
(kms_flip:935) CRITICAL: Failed assertion: !(*v & flag)
(kms_flip:935) CRITICAL: Last errno: 25, Inappropriate ioctl for device
Instead of a more descriptive assertion error, clear the request for
event to fix this.
This will change the error reported in bug #103257.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
References: https://bugs.freedesktop.org/show_bug.cgi?id=103257
[mlankhorst: Replace igt_assert with igt_assert_eq based on ickle's suggestion]
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
We don't actually care about running this for 30 seconds, all we
care is that we get a -EBUSY if a flip is already queued.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Remove the -interruptible test, the test only tests that we get an
-EBUSY after doing a pageflip. Doing this interruptibly adds the
possibility the test will take too long from retrying.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Missing \n at the end of log message caused some hard to read logs.
Cc: Martin Peres <martin.peres@linux.intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
Compilation support was dropped in January 2019, since then it's just a
dead code in the repo and no one seems to be missing it.
Anyway, it can always be reintroduced from the chasms of git history.
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Acked-by: Petri Latvala <petri.latvala@intel.com>
|
|
There is no real need for testing on each output, as they do not affect
CRC coming out from the pipe. Let's use first viable one.
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Martin Peres <martin.peres@linux.intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
Since igt_assert family of functions logs last errno we get a lot of
those: "Last errno: 25, Inappropriate ioctl for device"
isatty() seems to be the biggest offender in that area, so this patch
should limit amount of confusing messages significantly.
Cc: Martin Peres <martin.peres@linux.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>
|
|
Due to Gamma/Degamma limitation wrt representation of intermediate
values between 0 and 1.0 causing rounding issues and inaccuracies,
applying linear gamma affects crc. This patch fixes the same by
making ctm max test independent of gamma/degamma.
v2: Disable degamma/gamma programming for ctm max test as it
leads to crc mimsmatch. Limiting it to this test case alone as
other tests need it to be enabled, hence not touching those
scenarios.
v3: Fixed a fumble with compilation.
v4: Disabling degamma and gamma for ctm max tests, after the logic
in kernel has been updated by Ville's series.
v5: Disabled gamma/degamma for all ctm tests as suggested by Ville.
v6: Restricting disabling of linear gamma luts for ctm max test.
Updated the commit message and comment as suggested by Daniel.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108147
Signed-off-by: Uma Shankar <uma.shankar@intel.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Lot of tests fail, when assertion checking
how many vblanks should have passed during
suspend/resume is compared to magic value 150.
At the same time even for failed cases, sometimes
it is clearly visible that there were no issue -
simply suspend took longer on that machine.
If suspend took around 10 s and we get roughly
60 vblanks per second the value then should be
around 600 and not 150.
This change removes 150 magic value and starts
to use calculation of what it is expected to
be instead of being hardcoded.
v2: Add possible error delta interval, where expected
vblanks must lie, i.e:
[estimated_vblanks - err, estimated_vblanks + err]
v3: Made an error range for estimated vblanks a bit less
strict.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104894
Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
The kmstest_dumb_create API isn't suitable for creating multi-planar
buffers since it tries to calculate the size based on the first plane's
pitch only.
AMDGPU requires that the luma pitch be aligned to 256 for YUV buffers
which results in crashes on kms_plane@pixel-format-pipe-*-planes tests
when using kmstest_dumb_create since the buffer returned is smaller than
needed (16384 size returned, 24576 size required).
Create and map the buffer with the correct size by using the AMD helpers
introduced by this patch: igt_amd_create_bo and igt_amd_mmap_bo.
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Leo Li <sunpeng.li@amd.com>
|
|
Let's add a check for supported pixel format. Otherwise, we fail the test,
for example, with the following error message
"[drm:intel_framebuffer_init [i915]] unsupported pixel format Y210 little-endian
(0x30313259) / modifier 0x100000000000003"
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110369
Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
If we have two tasks running on xcs0 and xcs1 independently, but who
queue subsequent work onto rcs, we may insert semaphores before the rcs
work and pick unwisely which task to run first. To maximise throughput,
we want to run on rcs whichever task is ready first. Conversely, if we
pick wrongly that can be used to trigger a GPU hang with unaware
userspace.
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 test uses the new library helper so it doesn't need any local
helper.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Antonio Argenziano <antonio.argenziano@intel.com>
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
|
|
Petri pointed out that the maximum allowed number of files per process
is nr_open, not the system cap of file-max.
Suggested-by: Petri Latvala <petri.latvala@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110351
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
Let's check if 90/270 rotation is supported for the format that we are
trying to test.
Currently, if we try to test unsupported format the kernel complains with
the following message.
"[drm:skl_plane_check [i915]] Unsupported pixel format Y210
little-endian (0x30313259) for 90/270!"
v2: Use igt_plane_has_format_mod() function to select formats that are
capable of 90/270 rotation (Ville)
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
|
|
I was looking into a failure in which I had
libkmod: ERROR ../libkmod/libkmod-config.c:790 conf_files_list: opendir(/etc/modprobe.d): Too many open files
libkmod: ERROR ../libkmod/libkmod-config.c:790 conf_files_list: opendir(/lib/modprobe.d): Too many open files
(gem_exec_parallel:1315) igt_kmod-WARNING: Could not load i915
(gem_exec_parallel:1315) igt_kmod-WARNING: Could not load i915
I got curious because libkmod doesn't open more than one config file at
a time. What's happening is that libkmod is not the culprit, it's just
the one that failed because we open /dev/dri/card0 -ETOOMANYTIMES.
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Share the implementation to tweak the maximum number of open files.
The version in tests/i915/gem_exec_reuse.c was a little bit different,
but I don't think it needs to be because it would still return a
failure if any of the calls to setrlimit() fail. So I'm using the other
one.
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
With the new engine query method engines are reachable through
an index and context they are combined with.
The 'gem_has_ring()' becomes 'gem_context_has_engine()' that
requires the index that the engine is mapped within the driver.
The function has been moved from lib/ioctl_wappers to
lib/i915/gem_context where it is more appropriate.
The previous 'gem_has_ring()' function becomes a wrapper to the
new 'gem_context_has_engine()'.
Signed-off-by: Andi Shyti <andi.shyti@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
__for_each_engine_class_instance(fd, e) doesn't need and doesn't
use the fd argument. Remove it.
Signed-off-by: Andi Shyti <andi.shyti@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
First, we set errno to 0 before doing select() to avoid random pollution
of the assert message with things like:
"Last errno: 25, Inappropriate ioctl for device"
Second, we log explicitly if we exceeded the timeout (ret == 0).
Third, if we fail the select() we log that with some explanation.
Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Petri Latvala <petri.latvala@intel.com>
|
|
backlight fade with suspend test turns off dpms which turns off the edp
backlight. Then it does a system suspend and resume. After resume,
the edp backlight would still be off, but the test sets the brightness
value and reads it back. Since the edp backlight is off, the brightness
values written and read are different causing the test to fail.
Do not turn off the DPMS before suspend so that after system resume,
the edp backlight would be on and setting the brightness value would
be successful.
v2: Remove "DPMS off" before system suspend instead of adding
"DPMS on" after system resume.
Cc: Jyoti Yadav <jyoti.r.yadav@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Clinton Taylor <clinton.a.taylor@intel.com>
Signed-off-by: Harish Chegondi <harish.chegondi@intel.com>
References: https://bugs.freedesktop.org/show_bug.cgi?id=107820
Fixes: 377752242995 ("Brightness test with DPMS and System suspend.")
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
|
|
We only need the warning once, not for the several thousand relocations
we try. The current execbuf implementation will set all presumed_offset
to -1 so this loop should quit on the first entry if we hit the
pagefault, but for the sake of completeness check all.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110269
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
|
|
Sprinkle some asserts into rendercopy to make sure we don't try
to exceed the render engine surface size/stride limitations.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Not displaying the flip on the next vblank is bad, but not the end of
the world -- so long as that is only a temporary glitch. Give the vblank
a few more frames to complete, and warn instead of failing if it takes
more than one vblank interval to flip.
v2: Bump the warning to >1 missed flip, to spare us the noise.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Just like we try to sanitize all properties in igt_plane_reset, we
should do the same for pipe properties. Loading a wrong lut could
affect passing and failing tests, so we need to be careful and set
sane defaults for everything.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
|
|
i915_ring_missed_irq was removed from debugfs in kernel patch
789659f4307a ("drm/i915: Drop fake breadcrumb irq") and it was the
base of which i915_missed_irq was written, so removing this test for
good.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
stridechange subtest
As explained in c1edee186d18 ("tests/frontbuffer_tracking: Do not
assert FBC state after a page flip changing stride") after changing
the plane stride there is the possibility that CFB will not be big
enough to keep FBC enabled, that is why do_assertions() is called
with DONT_ASSERT_FEATURE_STATUS but DONT_ASSERT_FEATURE_STATUS is
overkill and will not check the status of the other features like PSR
and DRRS when running combined feature tests and possibly hiding
bugs.
So lets add a new flag that will only not assert FBC.
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
|
|
changing stride
When the stridechange subtest was introduced f23ea58f1fbb
("kms_frontbuffer_tracking: expand badstride and stridechange")
atomic was not around and change the stride using drmModePageFlip()
was not allowed so it was expected that it would return -EINVAL and
kernel would keep the old framebuffer with the smaller plane that is
know to fit on CFB(if it don't fit the test will skip on the first
full-modeset because "not enough stolen memory" is set).
But after the introduction of atomic the subtest was updated by
f63e070b469d ("kms_frontbuffer_tracking: Fix tests with the new
atomic reality.") to accept a no error return from drmModePageFlip()
but the do_assertions() that follows it was not updated.
As the subtest function comment states, kernel will do fastsets in
this scenario and the allocated CFB could not be enough to keep FBC
enabled over the new framebuffer, so here adding the missing
DONT_ASSERT_FEATURE_STATUS to ignore the FBC state and just test if
CRC match and if kernel do not misbehave.
Other way to solve this issue would be make the kernel do a
full-modeset when CFB is not enough for the new plane so FBC is
disabled with the CRC freeing the actual CFB and then after enable
CRTC again it will try to enable FBC again if it can allocate the
required CFB but by the subtest comment this is not intended.
v2: Assert features when drmModePageFlip() fails aka non-atomic driver(Dhinakaran)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105683
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
|
