Damien Lespiau [Wed, 25 Jul 2012 15:04:32 +0000 (16:04 +0100)]
build: Make autoreconf honour ACLOCAL_FLAGS
When running autoreconf, it's possible to give flags to the underlying
aclocal by declaring a ACLOCAL_AMFLAGS variable in the top level
Makefile.am.
Putting ${ACLOCAL_FLAGS} there allows the user to set an environment
variable up before running autogen.sh and pull in the right directories
to look for m4 macros, say an up-to-date version of the xorg-util macros.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Daniel Vetter [Tue, 22 May 2012 17:55:21 +0000 (19:55 +0200)]
test/flip_test: don't fail if we can't find a crtc
some connectors are forced to certain crcs ...
Daniel Vetter [Mon, 23 Jul 2012 09:25:47 +0000 (11:25 +0200)]
tests: fixup prime test paranoia
Daniel Vetter [Mon, 23 Jul 2012 08:06:25 +0000 (10:06 +0200)]
tests/prime: more paranoia
- check the old contents first in check_bo
- drop all GEM handles for the re-import test
Daniel Vetter [Mon, 23 Jul 2012 07:51:12 +0000 (09:51 +0200)]
tests: add missing stuff to .gitignore
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Mon, 23 Jul 2012 07:49:06 +0000 (09:49 +0200)]
tests: add basic re-importing prime test
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Mon, 23 Jul 2012 04:41:10 +0000 (06:41 +0200)]
tests: s/drm_reg_read/gem_reg_read
gem_ is the prefix for low-level kernel tests, drm_ tests
libdrm_intel specific stuff.
Daniel Vetter [Sat, 21 Jul 2012 15:36:36 +0000 (17:36 +0200)]
shut up the compiler a bit
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Sun, 22 Jul 2012 23:30:04 +0000 (16:30 -0700)]
reg_read: use good offset, and errno
The test was bonghits.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Wed, 18 Jul 2012 18:47:42 +0000 (11:47 -0700)]
reg_read: skip older kernels
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Thu, 12 Jul 2012 00:06:01 +0000 (17:06 -0700)]
reg_read: basic register read ioctl test
This will need to get modified when the ioctl expands, and so is only
here for reference/to make Daniel happy.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Daniel Vetter [Fri, 13 Jul 2012 08:35:11 +0000 (10:35 +0200)]
tests: improve concurrent blit tests
By adding another testcase that follows up with a gpu read. This
checks whether the kernel properly tracks the pending write and
doesn't lose it (or sync up to the wrong seqno).
For some odd reason only the cpu mmap variant blows up, the gtt one
works here. I need to look into that some more.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 3 Jul 2012 18:04:05 +0000 (11:04 -0700)]
gem_ctx_basic: fixes
I must have checked in the wrong version the first time.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Tue, 3 Jul 2012 17:40:32 +0000 (10:40 -0700)]
ctx_basic: we stopped using rendercopy
so get rid of the remants. perhaps we'll use it again someday.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Sun, 1 Jul 2012 02:26:59 +0000 (19:26 -0700)]
rendercopy_gen7: add support for haswell
Only compile tested.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Thu, 14 Jun 2012 18:04:01 +0000 (11:04 -0700)]
context_basic: the most stressful context test to date.
this context test requiring libdrm updates
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Sun, 15 Jan 2012 22:52:33 +0000 (14:52 -0800)]
context: libdrm wrappers
This wraps libdrm functionality to exec with contexts. This patch
shouldn't be applied until libdrm for contexts is updated.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Fri, 29 Jun 2012 18:23:34 +0000 (11:23 -0700)]
intel: update libdrm requirement
We want context APIs in order to use our super context test.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Fri, 29 Jun 2012 05:43:47 +0000 (22:43 -0700)]
gem_stress: use the gen7 rendercopy
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Fri, 29 Jun 2012 05:42:58 +0000 (22:42 -0700)]
lib: add a gen7 rendercopy
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Wed, 27 Jun 2012 18:13:23 +0000 (11:13 -0700)]
ctx_bad_destroy: danvet changed to ENOENT
In all fairness if the test worked, he would have found this earlier
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Wed, 27 Jun 2012 18:13:06 +0000 (11:13 -0700)]
ctx_bad_destroy - ioctl is 2e, not 23
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Wed, 27 Jun 2012 18:07:09 +0000 (11:07 -0700)]
ctx_bad_destroy: a little more description is nice
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Daniel Vetter [Tue, 26 Jun 2012 17:47:30 +0000 (19:47 +0200)]
tests/gem_wait_rendering_timeout: needs more fudge
Due to the way we calculate the workload by doubling it each time we
might end up with almost twice as much as we want. Hence increase our
fudge by 1.5 to account for that.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50666
Eugeni Dodonov [Thu, 21 Jun 2012 17:31:34 +0000 (14:31 -0300)]
intel_reg_dumper: use intel_register_access_init/fini
We need to hold forcewake lock in order to be able to read GT registers.
Otherwise, when the GPU is in RC6 mode, we'll read all zeros.
Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Ben Widawsky [Thu, 14 Jun 2012 17:00:57 +0000 (10:00 -0700)]
context tests not requiring libdrm
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Mon, 19 Mar 2012 01:42:44 +0000 (18:42 -0700)]
context: update for new execbuf2 element
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Eugeni Dodonov [Wed, 13 Jun 2012 18:34:40 +0000 (15:34 -0300)]
intel_reg_dumper: dump more PM registers
Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Jesse Barnes [Mon, 11 Jun 2012 19:08:32 +0000 (12:08 -0700)]
add VLV PCI ID
This allows the tests to run on the prototype boards.
Daniel Vetter [Fri, 8 Jun 2012 17:35:19 +0000 (19:35 +0200)]
tests/debgufs_emon_crash: make it quiet
Daniel Vetter [Thu, 7 Jun 2012 14:37:02 +0000 (16:37 +0200)]
intel_reg_dumper: dump pre-pch-split blc regs
Chris Wilson [Mon, 4 Jun 2012 16:29:20 +0000 (17:29 +0100)]
test/gem_gtt_speed: Add a baseline test for the performance of a CPU mmap
When looking at the pwrite/pread/wc performance, it is useful to judge
that against the performance of an ordinary CPU mmap.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Thu, 31 May 2012 13:45:17 +0000 (15:45 +0200)]
test/sysfs_l3_parity: skip test if the sysfs interface doesn't exist
Daniel Vetter [Thu, 31 May 2012 13:45:17 +0000 (15:45 +0200)]
test/sysfs_l3_parity: fixup logic
It didn't fail properly.
Daniel Vetter [Thu, 31 May 2012 11:49:32 +0000 (13:49 +0200)]
cleanups after l3_parity merge
- call it sysfs_l3_parity for more consistency
- add .gitignore for intel_l3_parity
Ben Widawsky [Fri, 25 May 2012 23:56:27 +0000 (16:56 -0700)]
tests/dpf: simple dpf test
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Fri, 25 May 2012 23:56:26 +0000 (16:56 -0700)]
tools/dpf: Tool to read and write l3 remap registers.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Daniel Vetter [Wed, 30 May 2012 18:16:24 +0000 (20:16 +0200)]
tests/gem_wait_rendering_timeout: check polling behaviour
... with timeout=0. We have special code in the kernel to optimize
that one.
Daniel Vetter [Tue, 29 May 2012 20:23:14 +0000 (22:23 +0200)]
lib/gem_wait_rendering_timeout: sprinkle busy_ioctl checks
While we have that testcase, let's cross-check the busy ioctl.
Daniel Vetter [Tue, 29 May 2012 20:14:06 +0000 (22:14 +0200)]
lib: fix gen5 workaround emission
Using BEGIN_BATCH can lead to a nice inf recursion through require_space
-> flush_batch -> BEGIN_BATCH.
Also fix things up to always require BATCH_RESERVED. We need 2 dwords
for the gen5 workaround and 2 dwords for MI_BB_END.
Daniel Vetter [Tue, 29 May 2012 14:52:23 +0000 (16:52 +0200)]
tests/gem_wait_rendering_timeout: test the timeout case
When I've first tried this it somehow failed, but I can't reproduce
this any more. I guess I've hunted a ghost.
Daniel Vetter [Tue, 29 May 2012 14:49:41 +0000 (16:49 +0200)]
tests/gem_wait_rendering_timeout: fix up the autotuner
duuh
Daniel Vetter [Tue, 29 May 2012 14:35:35 +0000 (16:35 +0200)]
tests/gem_wait_rendering_timout: adjust .gitignore
Daniel Vetter [Tue, 29 May 2012 14:35:11 +0000 (16:35 +0200)]
tests/get_wait_rendering_timeout: fixup the auto-tuner
- start at 1 iterations
- actually adjust the number of iterations like the comment claims
Daniel Vetter [Tue, 29 May 2012 14:34:32 +0000 (16:34 +0200)]
tests/gem_wait_rendering_timeout: some fixes for the blt function
- widht/height are in pixel, and the cmd uses 32bit, so we need to
adjust this.
- we write to the buffer, set the right reloc domain.
Daniel Vetter [Tue, 29 May 2012 14:23:41 +0000 (16:23 +0200)]
tests/gem_wait_rendering_timeout: properly skip test
also assign the timeout value unconditionally, we want to know it even
when we've timeout out.
Daniel Vetter [Tue, 29 May 2012 14:14:31 +0000 (16:14 +0200)]
tests/gem_wait_rendering_timeout: make it compile
... we need to do this libdrm depency dance. Or just copy&paste it.
Ben Widawsky [Fri, 11 May 2012 20:54:12 +0000 (13:54 -0700)]
tests/wait render timeout test
Assures that signals interrupting the wait works properly. Because of
the scheduling around signals, interrupted waits will *seem* faster as
the GPU continues to work while all the CPU scheduling stuff happens.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Daniel Vetter [Fri, 25 May 2012 09:29:28 +0000 (11:29 +0200)]
tests: fixup type in module_reload
Oops, I've frobbed this to test the check but committed the wrong
version.
root [Fri, 25 May 2012 09:26:53 +0000 (11:26 +0200)]
tests: add some help to module_reload
... so that it's easier to get the kernel configuration right.
root [Fri, 25 May 2012 09:02:09 +0000 (11:02 +0200)]
tests: add module reload testcase
We've broken this way too often in the past.
Daniel Vetter [Wed, 23 May 2012 09:37:04 +0000 (11:37 +0200)]
tests/hangman: fix when run with absolute path
QA's test rig does that.
Daniel Vetter [Tue, 22 May 2012 17:54:22 +0000 (19:54 +0200)]
test/flip_test: fail harder
We want testcase to return non-zero if they fail ...
Daniel Vetter [Tue, 22 May 2012 14:15:15 +0000 (16:15 +0200)]
lib: extract kmstest_dump_mode
Daniel Vetter [Tue, 22 May 2012 14:01:32 +0000 (16:01 +0200)]
test/flip_test: use kmstest_create_fb
And bring back some frame indication that actually works.
Daniel Vetter [Tue, 22 May 2012 12:37:19 +0000 (14:37 +0200)]
lib: extract kmstest_create_fb
We should get more kms tests soon, and not needing to copy-paste a
nice test pattern should be useful.
That establishes a firm depency of i-g-t on cairo over everything, but
I don't care so much about that.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Tue, 22 May 2012 10:01:36 +0000 (12:01 +0200)]
tests: use drm_open_any more
And fix it up to not leak open fds, which kills all the master only
stuff.
Daniel Vetter [Tue, 22 May 2012 08:57:17 +0000 (10:57 +0200)]
tests/flip_test: rip out usage
Totally useless if the only option is "-h"
Daniel Vetter [Tue, 22 May 2012 08:53:38 +0000 (10:53 +0200)]
tests: add flip_test to .gitignore
Daniel Vetter [Tue, 22 May 2012 08:51:29 +0000 (10:51 +0200)]
tests/flip_test: rip out funny pictures
The path isn't properly autodetected and it yells stupid stuff into
stderr. Which causes spurious failures.
And really, these are automated tests. If you need to look at the output,
something is wrong.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Fri, 27 Apr 2012 19:14:34 +0000 (16:14 -0300)]
intel_infoframes: options -p and -P require an argument
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Eugeni Dodonov [Wed, 16 May 2012 13:14:55 +0000 (10:14 -0300)]
intel_gpu_top: allow to work on Gen7
Somehow nobody noticed this before, but we were missing blt and bsd6
initialization on Gen7.
Reported-by: Philippe Lecluse <philippe.lecluse@intel.com>
Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Daniel Vetter [Thu, 10 May 2012 18:25:33 +0000 (20:25 +0200)]
tests/gtt_speed: try to outwit gcc
It optimized away the read loops ;-)
Also fix up the size to not be 1/4th of the one we want for tiled
buffer tests.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Wed, 9 May 2012 13:06:28 +0000 (14:06 +0100)]
tests/gem_tiled_pread_pwrite fix progname for pretty print progress
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Wed, 9 May 2012 12:50:53 +0000 (13:50 +0100)]
intel_error_decode: Be more lax for whitespace around parsing PCI-ID
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Eugeni Dodonov [Tue, 8 May 2012 18:55:05 +0000 (15:55 -0300)]
tools: add Haswell registers into intel_reg_dumper
For now, only print their content for diffing, but also add the necessary
bits that can be used for more verbose output in the fugure.
Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Jesse Barnes [Mon, 7 May 2012 19:57:05 +0000 (12:57 -0700)]
tests/flip_test: test on each CRTC for each detected output
This allows us to test on 3 pipe configs for example.
Jesse Barnes [Mon, 7 May 2012 19:36:28 +0000 (12:36 -0700)]
tests: add simple flip test
Add a simple flip test to test flipping on a single output with the
preferred mode.
TODO: test each crtc for 3s and print pass/fail results
Ben Widawsky [Sun, 29 Apr 2012 03:10:09 +0000 (20:10 -0700)]
intel_infoframe: fix assertion off by 1
this makes my compiler very unhappy
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Sun, 29 Apr 2012 03:07:40 +0000 (20:07 -0700)]
chipset: accidentally left the old IS_GEN7 macro
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Daniel Vetter [Fri, 27 Apr 2012 12:08:34 +0000 (14:08 +0200)]
test/ZZ_hangman: adjust to new error_state reset semantics
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Tue, 24 Apr 2012 21:10:25 +0000 (18:10 -0300)]
intel_infoframes: add support for Gen 4
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Tue, 24 Apr 2012 21:10:24 +0000 (18:10 -0300)]
intel_infoframes: AVI InfoFrame version should be 2
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Wed, 25 Apr 2012 20:33:17 +0000 (13:33 -0700)]
chipset updates
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Daniel Vetter [Wed, 25 Apr 2012 11:21:38 +0000 (13:21 +0200)]
test: fix up hangman script and enable by default
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Sun, 22 Apr 2012 10:35:23 +0000 (12:35 +0200)]
tests: add gem_unfence_active_buffers
Unfortunately this requires slab poisoning to catch anything :(
Also add a new helper to drmtest to get the available fence count.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Thu, 19 Apr 2012 13:11:51 +0000 (14:11 +0100)]
lib: Fix render copy on gen2
Daniel Vetter [Thu, 19 Apr 2012 10:34:16 +0000 (12:34 +0200)]
tests/gem_ringfill: add progress indicator
This thing takes ages on older chips.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Tue, 17 Apr 2012 20:38:44 +0000 (22:38 +0200)]
tests: re-enable commented-out testcase
Meh, I suck.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Tue, 17 Apr 2012 20:02:33 +0000 (22:02 +0200)]
tests: add gem_set_tiling_vs_pwrite
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Tue, 17 Apr 2012 16:01:48 +0000 (18:01 +0200)]
tests/gem_set_tiling_vs_gtt: update copyright header
Meh, I've forgotten these.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Tue, 17 Apr 2012 15:25:24 +0000 (17:25 +0200)]
tests: add gem_set_tiling_vs_gtt
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Tue, 17 Apr 2012 11:25:04 +0000 (13:25 +0200)]
tests/gem_set_tiling_vs_blt: fix on gen4+
Oops, the new checks need correctly tiled blts to work.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Sun, 15 Apr 2012 21:59:12 +0000 (23:59 +0200)]
tests: rename gem_tiled_after_untiled_blt into gem_set_tiling_vs_blt
... it checks much more now.
Also update the comments a bit in the testcase.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Sun, 15 Apr 2012 21:40:20 +0000 (23:40 +0200)]
tests: pimp tiled_after_untiled_blt
Now with more tiling transition tests. Changes:
- refactoring to reuse the tiling transition functions.
- add a tiled->untiled transition
- add a tiled->tiled transition with different stride
- add a check at the end to ensure that the tiling has indeed changed
on the test_bo
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Fri, 13 Apr 2012 13:27:01 +0000 (10:27 -0300)]
Add tools/intel_infoframes
This is a command-line tool that allows us to display and modify the
InfoFrames we send.
v2: use argv instead of stdin
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Thu, 12 Apr 2012 21:12:24 +0000 (23:12 +0200)]
tests: skip rc6_residency test if not supported
Fixes spurious failures.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Thu, 12 Apr 2012 14:47:00 +0000 (15:47 +0100)]
Use mmap64() for performing the GTT mapping
Although the address space for GTT mappings may only be 32-bits, we need
to use the explicit 64-bit mmap interface so that on a 32-bit platform
the offset we pass is not truncated to 31-bits.
Fixes gem_mmap_offset_exhaustion on 32-bit platforms.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Wed, 11 Apr 2012 13:28:26 +0000 (15:28 +0200)]
tests: add gem_gtt_concurrent_blit
Same test as Chris Wilson's gem_cpu_concurrent_blit, but for
gtt mmaps.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Wed, 11 Apr 2012 13:02:32 +0000 (15:02 +0200)]
tests: add sysfs_rc6_residency to .gitignore
Daniel Vetter [Wed, 11 Apr 2012 12:55:30 +0000 (14:55 +0200)]
tests/sysfs_rc6_residency: add some sleep before testing
That way I can run i-g-t on my ivb without spurious failures.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Sun, 25 Mar 2012 02:04:10 +0000 (19:04 -0700)]
tests: rc6 residency test
This is meant to test the sysfs entry for showing rc6 residency in
milliseconds. Remember, sysfs is a permanent interface.
v2: use new get_card interface to try "all" devices
check rc6p and rc6pp in addition to rc6
v3: rename rc6_residency.c to sysfs_rc6_residency.c
print better error messages
skip test if rc6 isn't enabled
v4: update to new sysfs names
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Ben Widawsky [Sun, 25 Mar 2012 23:04:07 +0000 (16:04 -0700)]
drm/i915: extract card getting
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Ben Widawsky [Sun, 25 Mar 2012 02:03:38 +0000 (19:03 -0700)]
build: make sure we have asprintf
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Chris Wilson [Tue, 10 Apr 2012 18:41:43 +0000 (19:41 +0100)]
test/gem_cpu_concurrent_blit: Limit memory usage to half aperture
So that we don't cause needless thrashing on older devices and spoil the
test.
Chris Wilson [Tue, 10 Apr 2012 17:21:33 +0000 (18:21 +0100)]
test: Exercise concurrent GPU read/write with CPU domain access
Designed to exercise this patch to i915.ko:
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index fbf1118..57ae1f2 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3181,9 +3181,11 @@ i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_obj
if (ret)
return ret;
- ret = i915_gem_object_wait_rendering(obj);
- if (ret)
- return ret;
+ if (write || obj->pending_gpu_write) {
+ ret = i915_gem_object_wait_rendering(obj);
+ if (ret)
+ return ret;
+ }
i915_gem_object_flush_gtt_write_domain(obj);
By exercising the conditions whereby should either of the checks be missed
an error is detected.
Daniel Vetter [Tue, 10 Apr 2012 14:23:18 +0000 (16:23 +0200)]
tests: explicit use bash
We use bashisms, which debian's default sh dash doesn't like.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Wed, 4 Apr 2012 06:02:59 +0000 (23:02 -0700)]
tests: add wedging tests
Used to check that i915_error_state works. Unfortunately this
kills the gpu because wedging is permanent.
Base on a patch by Ben Widawsky.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Wed, 4 Apr 2012 09:07:44 +0000 (11:07 +0200)]
tests: fixup drm_lib.sh sourcing
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Wed, 4 Apr 2012 08:56:50 +0000 (10:56 +0200)]
tests: use / as path separator instead of _
Oops. Originally by me, noticed while reviewing Ben Widawsky's patch.
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>