sysfs interface for rc6 has been changed to only expose deep RC6 file to the platorms
that really support them, aka SNB and IVB.
So let's split this test and only test RC6p and RC6pp on these platforms.
Reference: https://bugs.freedesktop.org/show_bug.cgi?id=84886
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
A few weeks ago we were hitting this assertion, and the _eq macro is
more useful when you're debugging.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
warning: ignoring return value of ‘write’, declared with attribute
warn_unused_result [-Wunused-result]
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
With the current Kernel, it takes more than 10 minutes to run
modeset-stress-extra-wait on BYT with eDP, and we don't really want
IGT subtests that take more than 10 minutes to run: QA reports them as
bugs.
Since, as far as I remember, any of these RPM race condition bugs we
had could always be triggered with just 4-5 rounds of testing, let's
tune down the number of rounds we test:
Before:
rounds = (--quick option used) ? 10 : 50;
After:
rounds = (--quick option used) ? 10 : 40;
rounds = (wait_flags & WAIT_EXTRA) ? rounds / 2 : rounds;
This should make all subtests be way above the 10m limit.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84897
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
This reverts
commit d92fbc23138b1014e8574daf29dbb06b8c81aa7a.
Author: Rodrigo Vivi <rodrigo.vivi@intel.com>
Date: Tue Sep 9 12:25:47 2014 -0400
tests/kms_sink_crc_basic: Wait 2 vblanks before grabing the new crc.
Sink CRC is fixed on kernel to wait as many vblanks as needed. It was fixed b
commit ad9dc91b6e21266bfc6f466db4b95e10211f31ee
Author: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Date: Tue Sep 16 19:18:12 2014 -0400
drm/i915: Fix Sink CRC
This this fix in place we don't need this extra 2 vblanks on test case itself
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
This reverts
commit d190a1f9668680c7a429110c91f8faa27949b220.
Author: Rodrigo Vivi <rodrigo.vivi@intel.com>
Date: Wed Sep 3 21:26:29 2014 -0400
tests/kms_psr_sink_crc: Wait 2 vblanks before grabing the new crc.
Sink CRC is fixed on kernel to wait as many vblanks as needed. It was fixed by:
commit ad9dc91b6e21266bfc6f466db4b95e10211f31ee
Author: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Date: Tue Sep 16 19:18:12 2014 -0400
drm/i915: Fix Sink CRC
This this fix in place we don't need this extra 2 vblanks on test case itself.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Modifications to 'null_state_gen' so it can generate GEN9
golden context batch buffer source for SKL.
v2: - rebased on top of gen8 changes (Mika)
- fixed state base address command size (Mika)
- base address size macro as pages (Mika)
v3: - rebased on top of current master (Mika)
- removed obsolete #includes (Mika)
- added copyright (Mika)
- render and component packing added (Mika)
Cc: Damien Lespiau <damien.lespiau@intel.com>
Cc: Armin Reese <armin.c.reese@intel.com>
Cc: Volkin, Bradley D <bradley.d.volkin@intel.com>
Reviewed-by: Volkin, Bradley D <bradley.d.volkin@intel.com> (v2)
Signed-off-by: Armin Reese <armin.c.reese@intel.com> (v1)
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Previously we didn't have a clear understanding what is necessary
for a pipeline state to be properly initialized. So we had to improvise
and use a stripped out render copy.
Now we have a more clear understanding so switch out render copy based
frankenstate to state we can call golden state.
v2: - export intel_batch_state_offset
- add 3DSTATE_RASTER (Bradley Volkin)
Cc: Volkin, Bradley D <bradley.d.volkin@intel.com>
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
In null/golden context there are multiple state commands where
the actual state is always zero. For more compact batch representation
add a macro which just emits command and the rest of the state as zero.
v2: - Be more verbose about length bias (Bradley Volkin)
- strip out unrelated state_offset declaration (Bradley Volkin)
Cc: Volkin, Bradley D <bradley.d.volkin@intel.com>
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Currently our kernel side buffer object is only one page.
Limit the amount of dwords to 1024 to enforce this.
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
POSIX only requires "=" to be supported; "+=" works in bash but not in
dash.
Signed-off-by: Adam Sampson <ats@offog.org>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
The POSIX standard only requires test to support "="; "==" works in bash
but not in dash. Other comparisons in configure.ac use "=" already.
Signed-off-by: Adam Sampson <ats@offog.org>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Those registers are big core registers and weren't really relevant for
VLV/CHV. Drop them from the dump then.
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
SKL will have a whole separate display regs file, so merge
base_display.txt into each platform file.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Because quick-dump was only selecting a few files in lib/ and we move
stuff around and/or add new dependencies we were failing to provide the
necessary symbols to the shim library providing python bindings. And so
we had a run-time error:
Traceback (most recent call last):
File "./tools/quick_dump/quick_dump.py", line 17, in <module>
import chipset
File "/home/damien/gfx/sources/intel-gpu-tools/tools/quick_dump/chipset.py", line 28, in <module>
_chipset = swig_import_helper()
File "/home/damien/gfx/sources/intel-gpu-tools/tools/quick_dump/chipset.py", line 24, in swig_import_helper
_mod = imp.load_module('_chipset', fp, pathname, description)
File "/usr/lib64/python3.3/imp.py", line 183, in load_module
return load_dynamic(name, filename, file)
ImportError: /home/damien/gfx/sources/intel-gpu-tools/tools/quick_dump/_chipset.so: undefined symbol: kmstest_pipe_name
So, let's simplify maintainance and just link against the library we're
building and using elsewhere.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
../../tests/gem_wait.c: In function ‘render_timeout’:
../../tests/gem_wait.c:182:3: warning: format ‘%llu’ expects argument of type
‘long long unsigned int’, but argument 3 has type ‘uint64_t’ [-Wformat=]
igt_info("Finished with %llu time remaining\n", timeout);
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
I didn't notice this on the machine I developed it since the original
wait testcase fails there. Oops.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Since the igt_create_stereo_fb function moved into igt_fb.c
we need to add the definition of IGT_DATADIR in the Android
makefile in order to build the cairo dependent tests.
Signed-off-by: Tim Gore <tim.gore@intel.com>
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Yay for breaking piles of tests.
This regression has been introduced with
commit 5782eca1e19e85a04ad402fa4094aa1b5f9c53ce
Author: Tim Gore <tim.gore@intel.com>
Date: Wed Oct 1 13:25:20 2014 +0100
lib/igt_core.c: disable lowmemorykiller during tests
Cc: Tim Gore <tim.gore@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Several IGT tests cycle through a lot of GEM memory and
when running these tests on Android they tend to get
killed by the lowmemorykiller. The lowmemorykiller really
is not usefull in this context and is just preventing the
test from doing its job. This commit adds a function to
disable the lowmemorykiller by writing "9999" to its
oom adj parameter, which means it will never "select"
any process to kill. The normal linux oom killer is still
there to protect the kernel.
The low memory killer is disabled during the common
init function and then re-enabled by the exit handler.
To make this work for single tests the exit handler is now
installed in the common init function also so that it is
invoked for all tests.
This is just a hack to get round the fact that the i915
driver uses the concept of purgeable memory which is not
understood by the lowmemorykiller. If this ever gets
fixed then this patch can be removed.
Signed-off-by: Tim Gore <tim.gore@intel.com>
[danvet: A bit of polish (coding style, static and rename
check_igt_exit to common_exit_handler, drop the is_disabled check
since we only call this once now).]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
With the structured logging it makes more sense to tune this down a
bit. Also, this way it is consistent with Thomas Wood's new activity
indicator helper.
Spotted while discussing Thomas' patch with him.
v2: Thomas noticed that I've forgotten the fflush. Extract
igt_interactive_info for both igt_progress and igt_print_activity.
v3: Interactive output should go to stderr. Also extract the "is this
a terminal" check.
Cc: Thomas Wood <thomas.wood@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
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>
The FORCE_MEDIA_AWAKE bit is added for the PIPELINE_SELECTION command and
some instructions requires that the media enginee is awake.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
The send instruction on gen9 uses the 32bit immediate instead of 6bit immediate
for the extended message descriptors. And some bits of SEND instruction are defined
as the extdesc field.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
READ(...) is used for Render Target read and Media Block read. But there is no
sampler cache agent on gen9. At the same time two message types don't
share the same cache agent any more. So a parameter is needed for cache agent.
The 2th parameter of read(...) is not used for gen6/gen7/gen8. Hence it is
reused as cache agent for SKL as that on ILK.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
The different cache agent type is defined for SKL although it still uses
the same function ID as the previous generations.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
A value less than 4 might result in GPU hang on simulation
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@linux.intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
