In this test all the subtest are doing the same exact thing but with
different tiling / on a different pipe. Tiling/pipe is in the test name
so no need to elaborate on them.

We can have just one igt_describe() on top of the group to describe all
the subtests in one go, no copy-and-paste necessary.

The description is short and explains the spirit of the test (verifying
atomicity of plane updates) without delving into the implementation
details (holes, colors, etc.)

Other changes:
 * The function used for grabbing reference CRC is now under a more
   descriptive name.
 * Comment explaining test implementation in more detail is moved to the
   top of the function actually implementing those steps.
 * Minor formatting touch ups.

Cc: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Acked-by: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Mika Kahola <mika.kahola@intel.com>

This test is quite simple which makes it perfect for the first real
world example of igt_describe usage.

v2: keep the original comment (Simon)

Cc: Simon Ser <simon.ser@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Acked-by: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

This patch adds igt_description() which attaches a description to the
following igt_subtest or igt_subtest_group block.

Descriptions are accessible via './test --describe[=pattern]'

Subtest description is its own igt_describe as well as igt_describes
of all the parenting igt_subtest_groups, starting from the outermost
scope.

Examples of code and produced outputs are included in
lib/test/igt_describe.c and as a documentation comment on igt_describe()
macro.

v2: address Simon's review

Cc: Simon Ser <simon.ser@intel.com>
Cc: Petri Latvala <petri.latvala@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Acked-by: Petri Latvala <petri.latvala@intel.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

The new EDID has been byte-by-byte checked to be exactly the same as before.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Short Video Descriptors (SVDs, not to be confused with Vendor-Specific Data
blocks) describe CEA video formats supported by the monitor.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Native Detailed Timing Descriptors follow the Data Block Collection in the CEA
extension.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

The HDMI Vendor-Specific Data Block, defined as an opaque blob in the EDID
spec, is described in the HDMI 1.4 spec.

Most of the extension fields are optional, so it doesn't translate well to a C
struct. For now callers need to manually fill the hdmi_vsdb.data field.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

"VSDB" (Vendor-Specific Data Block) is the wording used by CEA 861D.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Given an EDID, computing the size is trivial. Instead of having one size
constant per EDID and hope the callers use the right one (ie. *not* EDID_LENGTH
when there's an extension), we can make functions that take EDIDs compute the
size if they need it.

We have tests in lib/tests/igt_edid.c which assert the number of extensions
present in the EDID anyway.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

When forcing a reset, it is crucial that the kernel correctly identifies
the injected hang. Verify this is the case for reset-stress.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

The original URL used http + seems to be author's own small site and is
currently down.

Instead of waiting for it to be up again just to hammer it with our CI
we can set up our own mirror for the tarball and use it.

Cc: Petri Latvala <petri.latvala@intel.com>
Cc: Simon Ser <simon.ser@intel.com>
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

Just a little bit of feedback at the end of an otherwise quiet 20s.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>

To reduce the assumptions of RCS0 in the kernel, we want to make the
debugfs engine agnostic and so we need to adapt the igt parser for
flexibility.

If we could just adequately simulate S3/S4 in the kernel we could forgo
this test entirely...

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

This macro is defined in two tests. We can move it to common place
in ioctl_wrappers, which is included by all tests.
v2: Missed intel_device_info lib. Moved BIT definition to
intel_chipset, which is part of igt.h

Signed-off-by: Lukasz Kalamarz <lukasz.kalamarz@intel.com>
Cc: Katarzyna Dec <katarzyna.dec@intel.com>
Cc: Michal Winiarski <michal.winiarski@intel.com>
Reviewed-by: Katarzyna Dec <katarzyna.dec@intel.com>

In kernel macro BIT is used to set value 1 on a given BIT, while in
this tool it was checking value of given bit in register.
So it was renamed.

Signed-off-by: Lukasz Kalamarz <lukasz.kalamarz@intel.com>
Cc: Zhenyu Wang <zhenyu.z.wang@intel.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Katarzyna Dec <katarzyna.dec@intel.com>
Cc: Michal Winiarski <michal.winiarski@intel.com>
Reviewed-by: Katarzyna Dec <katarzyna.dec@intel.com>

Before the i915_pmu kernel interface was available, we had to rely on
doing some decidedly dodgy mmio access to registers. However, now that
we have a stable interface via perf for grabbing all the details we
need, that and its supporting infrastructure can be discarded.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Antonio Argenziano <antonio.argenziano@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Acked-by: Antonio Argenziano <antonio.argenziano@intel.com>

Matt Roper authored 6 years ago and Ville Syrjälä committed 5 years ago

When using --interactive-debug, it's sometimes desirable to ignore CRC
mismatches and let the test proceed as if they passed so that the
on-screen outcome can be inspected.  Let's add a debug option to allow
this.

Cc: igt-dev@lists.freedesktop.org
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
[vsyrjala: pimp the debug message to indicate the skip]
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>

When this happens, the logs were completely empty previously, which can be
quite confusing.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

CI is currently breaking on linus/drm-intel-fixes due to the assert that
the kernel supports context engine maps. Report the lack of maps on
older kernels gracefully!

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>

We now generate our own EDID for DP audio tests.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Martin Peres <martin.peres@linux.intel.com>

This returns the 3-letter manufacturer identifier of an EDID.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

This is a simple helper to get the size in bytes of an arbitrary EDID.

Signed-off-by: Simon Ser <simon.ser@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Commit ad129d2a
("tests/i915/gem_ctx_switch: Update with engine discovery") broke testing
of I915_EXEC_DEFAULT. Bring it back.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reported-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Andi Shyti <andi.shyti@intel.com>
Cc: Ramalingam C <ramalingam.c@intel.com>
Reviewed-by: Ramalingam C <ramalingam.c@intel.com>

We are missing PCI device ID for SKU ICLLP U GT 1.5F (0x8A54) as per BSPec.

Align with kernel commit 93d7a3b4be30 ("drm/i915/icl: Add missing device ID")

BSpec: 21141

v2: Corresponding kernel commit added in commit message (Petri)
v3: Fix error on BSPec index (Stuart)

Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Reviewed-by: Stuart Summers <stuart.summers@intel.com>

If request is to verify all engines work, use the new engine discovery API
so we can actually test all.

There is a sporadic and mysterious, possibly memory corruption issue
surronding this area so lets see if this changes the pattern.

It is easy to do in this function due it re-opening the driver and
throwing it away. So there are no concerns that the modified default
context can affect any other test.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
References: https://bugzilla.freedesktop.org/show_bug.cgi?id=110667


Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>

I converted this one as an example, or at least to drive the discussion,
how more complex tests can be converted.

I have kept the legacy execbuf abi testing prefixed with "legacy-".

New tests were added to exercise physical engines via engine discovery and
also "all" tests have been updated in the same way.

To keep things simpler and avoid having to create separate contexts legacy
tests have to be first since the __for_each_physical_engine iterator would
otherwise configure the default context and confuse the test.

So legacy tests run on the unconfigured (with engine map) context and use
a new helper gem_eb_flags_to_engine to look up the engine from the
intel_execution_engines2 static list. This is only to enable the core test
code to be shared.

Places where new contexts are created had to be updated to either equally
configure the contexts or not. Another new helper,
gem_context_has_engine_map was added to enable this.

Also beware of drive-by formatting changes.

v2:
 * Fix hyphen mess in subtest names.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ramalingam C <ramalingam.c@intel.com>
Cc: Andi Shyti <andi.shyti@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>

It was added only on the meson side of things.

Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Fixes: 3c4edeba ("tests/amdgpu: add 8bpc bypass mode test.")
Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

These just break the build on debian-mason-arm and debian-meson-mips
and aren't needed in the first place.

Cc: Dingchen Zhang <dingchen.zhang@amd.com>
Cc: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>

Guillaume Tucker authored 5 years ago and Simon Ser committed 5 years ago

Use the libatomic1:mips package only in the Debian Stretch Docker
image for MIPS and add Gitlab CI step to run tests on MIPS.

Signed-off-by: Guillaume Tucker <guillaume.tucker@collabora.com>
Fixes: 439a9f5d ("gitlab-ci: add build for MIPS")
Reviewed-by: Simon Ser <simon.ser@intel.com>

David Zhang authored 5 years ago and Nicholas Kazlauskas committed 5 years ago

To validate the AMDGPU bypass mode feature.

Generate DP test patterns, program the pipe and check if
the CRCs of framebuffer and DP receiver match.

v2: use the DPRX pipe crc source directly (Nicholas)

Cc: Harry Wentland <Harry.Wentland@amd.com>
Cc: Nick Kazlauskas <Nicholas.Kazlauskas@amd.com>
Signed-off-by: Dingchen Zhang <dingchen.zhang@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

David Zhang authored 5 years ago and Nicholas Kazlauskas committed 5 years ago

Need DPRX CRC source to validate AMDGPU 8bpc bypass mode.

Added AMDGPU DPRX pipe crc source in igt_debugfs.h

v2: dropped the CRTC and AUTO crc sources for AMDGPU (Nicholas)

Cc: Harry Wentland <Harry.Wentland@amd.com>
Cc: Nick Kazlauskas <Nicholas.Kazlauskas@amd.com>
Signed-off-by: Dingchen Zhang <dingchen.zhang@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

David Zhang authored 5 years ago and Nicholas Kazlauskas committed 5 years ago

Need framebuffer CRC to validate AMDGPU bypass mode.

For each color component less than 16-bits, padding zero bits,
loop to update CRC for each RGB compoment in the framebuffer.
The algorithm is based on DP spec v1.4.

Cc: Harry Wentland <Harry.Wentland@amd.com>
Cc: Nick Kazlauskas <Nicholas.Kazlauskas@amd.com>
Signed-off-by: Dingchen Zhang <dingchen.zhang@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

The "basic" subtests perform no verification that the read/write work,
only function as mere API exercisers and loose benchmarks. Rename them
to reflect that they are poor benchmarks instead.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Guillaume Tucker authored 5 years ago and Simon Ser committed 5 years ago

Add Docker image and Gitlab CI steps to run builds for the MIPS
architecture using Debian Stretch with backports.

Signed-off-by: Guillaume Tucker <guillaume.tucker@collabora.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

Signed-off-by: Simon Ser <simon.ser@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105363


Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

Guillaume Tucker authored 5 years ago and Simon Ser committed 5 years ago

Replace calls to the older __sync_* functions with the new atomic_*
standard ones to be consistent with other tests and improve
portability across CPU architectures.  Add dependency of sw_sync on
libatomic.

Signed-off-by: Guillaume Tucker <guillaume.tucker@collabora.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

Guillaume Tucker authored 5 years ago and Simon Ser committed 5 years ago

This fixes builds on some architectures, in particular MIPS which
doesn't have __sync_add_and_fetch_8 and __sync_val_compare_and_swap_8
for 64-bit variable handling.

* replace calls to the older __sync_* functions with the new atomic_*
  standard ones
* use the _Atomic type modifier as required with stdatomic.h functions
* add dependency for gem_create on libatomic

Signed-off-by: Guillaume Tucker <guillaume.tucker@collabora.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

Guillaume Tucker authored 5 years ago and Simon Ser committed 5 years ago

Add libatomic to the Fedora docker image so it can link binaries that
use __atomic_* functions.  Also explicitly add libatomic1 to Debian
docker images as it is needed in particular on non-x86 architectures
for run-time linkage.

Signed-off-by: Guillaume Tucker <guillaume.tucker@collabora.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

Guillaume Tucker authored 5 years ago and Simon Ser committed 5 years ago

Add conditional dependency on GCC's libatomic in order to be able to
use the __atomic_* functions instead of the older __sync_* ones.  The
libatomic library is only needed when there aren't any native support
on the current architecture, so a linker test is used for this
purpose.  This makes atomic operations available on a wider number of
architectures including MIPS.

Signed-off-by: Guillaume Tucker <guillaume.tucker@collabora.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>

Add a new knob "-A" to enable the aspect ratio client cap
and thus test modes with potentially different aspect ratios.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Simon Ser <simon.ser@intel.com>