The current upstream Kernel code stops and then restarts FBC at every
page flip. I am working on patches to keep FBC enabled all the time,
so this brings the possibility of having the hardware tracking
mechanism looking at the old buffer instead of the new one. This test
should catch this problem.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
There's no reason to do a mode unset since we don't care about whether
FBC or PSR are enabled or disabled. This should save about a second
for each time you invoke the program.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
It's not testing a lot since we lack the Kernel Stolen Memory Checker,
but some day this will be useful.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
... instead of just setting it for the fullscreen_plane subtest and
then unsetting it. We want to use it more, so just make sure all the
code knows how to deal with multiple planes per CRTC.
This also allows us to do some upfront caching of the plane
information, so the code for the tests gets simpler.
There's a new test that's going to use all these changes.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Exercises the new blob-creation ioctl, testing lifetimes and behaviour
of user-created blobs, as well as exercising all the invariant
conditions we guarantee from modes exposed as blob properties.
v2: Renamed to core_prop_blob, skip test if blob not available.
v3: No changes.
v4: Consistently return 0/errno.
v5: Use do_ioctl_err and igt_assert_fd.
Use igt_assert_*() helper macros rather than direct igt_assert().
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Daniel Vetter <daniel.vetter@intel.com>
The code is confused about the units of CODE_TIME. The comment
says 50 microsseconds, but the actual code makes it 50
milliseconds. Avoid the whole mess by measuring the sleep
duration ourselves. Since the time measurement is taken around
the whole operation it obviously includes a bit of extra, but
at least it's much less than the fixed 50 ms.
For instance on one VLV board I now get something like this:
- Residency in rc6 or deeper state: 3002 ms (ratio to expected duration: 0.98)
+ Residency in rc6 or deeper state: 3001 ms (sleep duration 3003 ms) (ratio to expected duration: 1.00)
so the reported ratio is now much closer to reality.
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
When testing all modes on a connector with a single mode, if the modeset
fails, the code attempts to remove fb_info[-1], because old_fb still has
the inital value of -1.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90625
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Test that VMAs associated with a context are cleaned up when
contexts are destroyed.
In practice this emulates the leak seen between fbcon and X server.
Every time the X server exits we gain one VMA on the fbcon frame
buffer object as externally visible via for example
/sys/kernel/debug/dri/0/i915_gem_gtt.
v2: Use igt_debugfs_open, getline and strstr instead of home-brewed
string matching. (Thomas Wood)
v3: Rebase for drm_open_driver.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Thomas Wood <thomas.wood@intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Commit fb66cd5 (tests/gem_storedw_loop: Fix use after free for bufmgr)
introduced a segmentation fault when listing subtests because
drm_intel_bufmgr_destroy is called with NULL. Move this and the call to
the close function inside an igt_fixture block to prevent them being
called when listing subtests.
Cc: Robert Beckett <robert.beckett@intel.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Move bufmgr destruction to end of tests. This avoids use after free on
successive tests.
This could be squashed with the following patch to aid bisectability:
commit 0679702150157706a6def66b893b29c16345f4db
Author: Jesse Barnes <jbarnes@virtuousgeek.org>
Date: Wed Aug 5 16:06:31 2015 -0700
tests/gem_storedw_loop: add new store_dword test to unify per-ring ones v2
Signed-off-by: Robert Beckett <robert.beckett@intel.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The test expects an ioctl failure when it tries to set
CONTEXT_PARAM_NO_ZEROMAP from a non-root process. However, there is no
requirement in the kernel for the user to be root to set this parameter,
so the test is failing (it never passed as far as I'm aware of).
Fix the test by making it expect a successful ioctl completion.
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Apply the new API to all call sites within the test suite using the following
semantic patch:
// Semantic patch for replacing drm_open_any* with arch-specific drm_open_driver* calls
@@
identifier i =~ "\bdrm_open_any\b";
@@
- i()
+ drm_open_driver(DRIVER_INTEL)
@@
identifier i =~ "\bdrm_open_any_master\b";
@@
- i()
+ drm_open_driver_master(DRIVER_INTEL)
@@
identifier i =~ "\bdrm_open_any_render\b";
@@
- i()
+ drm_open_driver_render(DRIVER_INTEL)
@@
identifier i =~ "\b__drm_open_any\b";
@@
- i()
+ __drm_open_driver(DRIVER_INTEL)
Signed-off-by: Micah Fedke <micah.fedke@collabora.co.uk>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
The drm core doesn't check unused fields of ADDFB2 for pre-FB_MODIFIERS
userspace, so use igt_require_fb_modifiers(). Also, the size of the
ioctl changed with the addition of the modifiers, so it is necessary to
use the LOCAL_ version of it, otherwise some data may get truncated.
v2: Improve commit message. (Thomas)
Remove one spurious change to use LOCAL_DRM_IOCTL_ADDFB2. (Thomas)
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
2 subparts of gem_bad_reloc check that the reloc address is below the
global gtt boundary. However, when executing from ppgtt the reloc
address can be greater than that and still be a valid address.
To be sure that we're using the right upper limit, select it based on
the ppgtt mode.
Cc: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Disable -Wcast-qual temporarily to allow memchr_inv to return non-const
data (similar to memchr), without causing a compiler warning.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
These have been replaced by subtests in gem_storedw_loop.
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Add --as-needed to the linker flags to reduce the number of shared
library dependencies, since not all the tests and tools use all the
libraries required by the helper library (for example, many tests do not
use cairo). This helps portability of the binaries and also makes a
very small improvement to the execution speed and memory consumption.
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
The VEBOX ring is not available in generations before Haswell, so make
tests that use it skip instead of fail in previous gens.
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
In the reset-pin-leak test we were calling
igt_set_stop_rings(STOP_RING_DEFAULTS) which sets the
stop_rings bits for all gpu engines. But we only submit
work to the render engine. When TDR is enabled (as it is
in Android currently) only the render engine gets reset,
which clears the stop_rings bit for the render engine but
not for the other engines. This causes the test to fail on
the second iteration because stop_rings is not clear.
So just set the stop_rings bit for the engine we are going
to hang, namely the render engine.
Signed-off-by: Tim Gore <tim.gore@intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
I'm using this to debug some aspects of the GTT tracking.
While at it, do a small rename and fix the ASCII art.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Add a header that includes all the headers for the library. This allows
reorganisation of the library without affecting programs using it and
also simplifies the headers that need to be included to use the library.
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Just like we did for PSR, let's do it for FBC. FBC gets reenabled in
just 50ms, so the 5000ms timeout is huge. On the other hand, we only
pay the 5000ms timeout full price 9 times when running
kms_frontbuffer_tracking --fbc-only, so this change shouldn't save too
much time.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
PSR only takes up to 100ms to be enabled, but we were using 5000ms
timeouts. The problem with PSR is that the MMAP_GTT tests have to
assert that PSR is disabled and stays disabled during the whole
timeout, so that 5s cost is a little to high when we consider that we
do the full 5s wait 74 times when running kms_frontbuffer_tracking
--psr-only.
By reducing to 2s we still make sure we're safe, since we're waiting
20x the time PSR needs to get reenabled, and we reduce the runtime of
running kms_frontbuffer_tracking --psr-only from 14m10s to 10m29s.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
If we already tested the combination of m1+m2, don't test m2+m1 later:
the drawing pattern used already has 4 squares, so we're already
testing the "m2 followed by m1" case.
This should reduce the test time from about 60s to about 30s.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Instead of having a single pair of methods per subtest. Having this in
pairs is not very useful since we end doing all the drawing methods
per subtest anyway.
This saves a few modesets, which makes eDP slightly faster - about 22s
on my local machine.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
We now have per-pixel-format FBs and CRC values. Also t->format to
indicate the format we're using on the subtest. We still only use the
default RGB888 format (AKA FORMAT_DEFAULT) for all tests since
multiplying the current number of tests by the number of new pixel
formats doesn't sound like a great idea.
With this, it should be really easy to add new tests that use
different pixel formats.
Also, if you want a full IGT run on a specific pixel format, all you
need to do is to change FORMAT_DEFAULT.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
We're not using those formats yet, but a simple change to create_fb()
allows us to use these formats without problems now.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
We want to add support for different pixel formats and the current
hardcoded pixel values won't work with that. So add enum color and
its auxiliary functions so we can decide the pixel values based on the
chosen color enum.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
It is create_fb() who is going to decide the buffer format based on
the arguments provided by the caller. This is another step for the
non-XRGB8888 support.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
The wrapper will do a few additional things when we add support for
formats different than XRGB8888.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Now that the MMAP_WC operations call the dirty ioctl, PSR gets
reenabled after some time. So we have to adjust op_disables_psr() to
take that into account.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
We need to test those pixel formats on the FBC code, so let's make
sure the drawing library works on them first.
v2: Update the gtkdoc (Daniel).
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
