Source drm_lib.sh before skipping the test to ensure that subtest
enumeration is always handled correctly.
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
These tests require DRM master right, so make sure they have it from the
beginning. This gives an early indication if another DRM master is running
and makes the given test skip (with a proper explanation of the reason)
instead of exiting with error.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Currently tests that use igt_simple_main will simply call
"exit()" if they pass, making it difficult to ensure that
any required cleanup is done. At present this is not an
issue, but it will be when I submit a patch to turn off the
lowmemorykiller for all tests.
Signed-off-by: Tim Gore <tim.gore@intel.com>
[danvet: Also update api docs.]
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
In order to exercise https://bugs.freedesktop.org/show_bug.cgi?id=84207
we need to interrupt the mmu_notifier_register with a signal. This is
likely to be quite difficult, but let's just try running the
create-destroy test in an interruptible loop for 5s.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Due to the nature of the test, we can be reasonably sure that it is
either all copied or not. So we can sacrificing testing the entire
buffer for the expected value in order to speed up the test by only
testing along the diagonal.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84354
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Env variables are a bit more annoying since much harder to discover.
With options you can just see what they do with --help.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Since the last patched value does not match the presumed relocation
value we submit to the kernel, it occasionally makes mistakes. Also note
that the libdrm interface makes the same presumptive mistake, and would
easily be broken by a threaded environment.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=77793
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Multithreaded test to validate the (lack of) locking in the flink/open
code in libdrm-intel. Based on a testcase from Rafel Sapala.
Cc: Rafal Sapala <rafal.a.sapala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
To minimally verify that non-square cursors work on the platforms where
they're supported perform the tests first with WxH cursor and then
repeat with WxH/3 cursor.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
gen8 actually changed the command layout, not just extended the
relocation value. Oh well.
References: https://bugs.freedesktop.org/show_bug.cgi?id=83915
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
When testing each format, iterate through the formats and not randomly
through the modes.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=83788
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
kms_psr_sink_crc check psr activity, residency and exit for screen updates.
So this test is useless.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Black screen is forbidden on this test. So let's fail if sink crc shows
it is back.
Also there are many cases where we know for shure it should be all green,
so let's check for them.
Instead of checking colors we could print with sw using cairo and check if we
have identical crc like cursor testcases. However with PSR the chance of artifacts
is low and chance of getting blank screen or unchanged screen is high. So even drawing
on sw and comparing both CRCs we can have the same result. However the risk is that
screen never changed. So the safest way is to compare if screen changed and check
the green color when we know it should be green or not green.
v2: Hardcoded green was simply wrong because green CRC can change depending
on display.Split R, G and B on CRC and bitewise them with mask to verify
it is green.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
In order to get all test cases fixed and the matrix planes-operations working
it was needed to use the common new igt kms functions for all cases.
Previously only sprite testcase was using it.
Fixed the fb colors in a way to make tests more clear and be impossible to see
black screen during the tests.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
This will allow us to test input/write oprations on any kind of plane.
At this point PLANE_ONOF is just the new name of TEST_SPRITE and
PLANE_MOVE is the one for TEST_CURSOR_MOVE. They will be extended and fixed
on the following patche(s).
v2: fix conflict after changing previous patch
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
This is needed to be able to split tests in a matrix that tests different
input/write methods and operations for different type of planes.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Just to make life easier and be eable to easily test with
PSR disabled to know exactly what to expect when running it
for real
v3: Use igt_debug helpers and add env option for running with psr disabled
on this test without have to recompile like v1 or changing igt infrastructure like v2.
I tried to add --disable-psr or local --dry-run but than it fails to print subtests
so this was the safest way.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Tests were broken on platforms that doesn't have psr or on new kernel that contains new interface.
A lot more need to be done to get these tests really useful, but for now lets avoid it breaking
tests framework.
v2: Doesn't duplicate kernel's HAS_PSR. skip based on debugfs output.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
v2: sink CRC R, G and B might change depending on display. So let's split the
colors and bitwise them.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
gen2 will read the batch up until it reaches the end as defined by the
MI_BATCH_BUFFER_END command. Play nice and make sure that it does end on a
MI_BB_END.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Only attempt to test the number of planes that exist on the device so as
not to trigger spurious failures.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=82233
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
The test relies on having to wait upon the GPU at some point and so
simulating a missed interrupt. If there is insufficient load on the GPU,
we can complete the task before we even wait.
References: https://bugs.freedesktop.org/show_bug.cgi?id=71334
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Move create_stereo_fb from testdisplay to igt_create_stereo_fb in igt_fb
so that it can be used in other tests.
v2: update for new igt_create_fb API
add parameters for format and tiling
remove some old debug code
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Use the existing allocation, saves having to make fresh allocations in
the innermost loop - trimming code and potential failure paths.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Since relocations are variable size, depending upon generation, it is
easier to handle the resizing of the batch request inside the
BEGIN_BATCH macro. This still leaves us with having to resize commands
in a few places - which still need adaption for gen8+.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This reveal that quite a few locations were writing relocation offsets
but only allowing for 32 bit addresses. To reveal such places in active
tests, we also now double check that we do not use more batch space than
declared.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
As before, we also want to race against access through the fence
registers. This overlaps slightly with gem_set_tiling_vs_blt, but the
different access pattern should make it useful.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
