I've left the asserts as-is where they check for tests that use the
infrastructure in drmtest.c incorrectly. That way piglit's
differentiation between a failed testcase and a crashing testcase
should be more useful.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The aim is that we keep on running subtests even when a not-too-lethal
assert failed in a subtest. To make that useful also print per-subtest
test results from igt_skip|fail|success functions.
If required we can always go googletest-nuts with different types of
asserts later on, but I think for now we're good with what we have
here.
v2: Also print out proper SKIP message when skipping all subsequent
tests since a global (i.e. outside of a subtest) init step failed and
resulted in an igt_skip call.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Upstream broke our dynamic creation of the testlist, and I think
adding stupid .tests suffixes everywhere just to appease upstream
autohell tools isn't that great. So scrap it, we can use piglit
instead.
References: https://lists.gnu.org/archive/html/help-debbugs/2013-06/msg00000.html
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Requested-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The _block postfix meant to convey that a C statement/block must
follow can be misread as the verb to block. So drop it.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Now that callers must setup the longjmp target (we can't do that in
run_subtest itself since that callframe won't survive) tests shouldn't
call this function directly any more.
Make this clear by adding a __ prefix.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
To simplify things add a set of gem_check_<ring> functions which take
care of this. Since I've opted for static inlines drmtest.h grew a few
more header includes which was a neat opportunity to dump a few redundant
#defines.
This kills all the skipped_all hand-rolled logic we have.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This way we can rip out all the skip handling from the test control flow,
and additionally (by using drmtest_retval()) even get correct exit codes.
The only tricky part is that when we only want ot skip parts of a test
(like for gem_pread and gem_pwrite) we need to split out those parts as
subtests. But no addition of control-flow is required, the set/longjmp
magic in the helpers all makes it happen.
Also we make extensive use of the behaviour of drmtest_skip to skip
all subsequent subtests if it is called outside of a subtest. This allows
us to re-flatten the control flow a lot.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
If we skip a test and and fail somewhere the parent might die before the
child. So add some cleanup to handle this case. Also make sure that the
parent indeed waits for the child to die.
This is required to make the latest version of the piglit runner happy,
it tends to wait for all zombies to disappear.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Exitcode handling in igt testcases has too ugly facets in combination
with subtests:
- When individual subtest needed to be skipped for different reasons
we need to painstakingly thread this information through the test to
make sure that we still succeed if just one testcase worked (for
running it with the simple automake test runner in igt). But it also
needs to correctly report the skip exit value if only this one
skipped testcase has been run (e.g. when run with piglit or in QA's
test infrastructure).
- We'd prefer to easily skip (and also fail) subtests without
hampering other subtests. Again threading the return value of each
subtest through the code is cumbersome.
To simplify test case writing use longjmps to be able to get out of
subcases easily. But since longjmps are funny things thug it all away
into the newly added drmtest_subtest_block macro.
Note that if drmtest_skip is called outside of a subtest, but in a
testcase with subtests all subsequent subtests will be automatically
skipped.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Doesn't do more than an if (drmtest_run_test(name)) right now, but
as soon as we get a bit of infrastructure to handle test failures and
skipping, this will get more interesting.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
As the sysfs is almost always mounted and readable, we have a higher
success rate checking for our error state there than in debugfs.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This testcase mixes correctnes tests with performance tests, so it's
good to track the different correctness test separate for QA.
Together with pread_after_blt the pwrite->blt tests here exercise the
full cache coherency lifecycle of both snooped and uncached objects.
/me is happy
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
New testcase to check that pwrite/pread correctly synchronize with
oustanding rendering. Just to catch regressions when we change
something.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Similar to how we test flink races. Note that on unfixed kernels this
oopses, and with my current set of patches it still leaks like mad.
v2: Only close the prime fd if we've successfully created it.
v3: Add a reimport test to check whether we don't race when reaping
the obj->dma_buf link.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
intel_copy_bo assumes a 32bpp bo, whereas we passed it bos with
arbitrary bpp values. This resulted in thrashing GPU memory following the
destination bo.
Fix this by using a blit helper that can handle other bpps too.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Needed by an upcoming patch fixing kms_render's blits for fbs that have
other than 32bpp formats.
Based on the corresponding SNA function.
v2:
- fix random ordering of src, dst parameters (Chris)
- pass pitch in bytes rather than pixels (Chris)
Signed-off-by: Imre Deak <imre.deak@intel.com>
When I moved code around here:
commit 57bc7631e161a7bdcf7f68e1a46049dcb49c2256
Author: Ben Widawsky <ben@bwidawsk.net>
Date: Thu May 23 11:09:55 2013 -0700
lib/intel_mmio: Fix the fix
I accidently dropped the intel_gen() for the PCI id.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67289
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
- Enable subtest support.
- Add a check for the same flink name in the racing threads, which is
an issue one of my recent patches actually fixes.
- Add the test I've actually wanted to write which races an flink
against gem close (with no open in between). That one does indeed
leak.
- Readd the leak check, but note that this needs a fixed kernel.
Otherwise the leak counter will be utter garbage.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This exercises a race in the flink name descruction of the current drm
gem core. When racing a gem close with a gem open the open can sneak
in and cause the kernel to leak the flink name and its reference.
This results in leaked gem objects that won't get reaped even at drm
file close time. On my 2 core/4 threads snb machine this leaks on the
order of 1k objects per second.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
