- Mar 24, 2023
-
-
Mauro Carvalho Chehab authored
Use the more portable join_paths() macro. Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
Mauro Carvalho Chehab authored
Simplify the checks inside testplan/meson.build and ensure that it will only build if option build_testplan is selected, by moving such check to docs/meson.build. Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
Mauro Carvalho Chehab authored
Currently, this holds just a variable used by testplan. Drop it, cleaning up the building system. This change allows building the testplan documentation without needing to build tests. Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
Mauro Carvalho Chehab authored
A warning is generated with newer meson versions: NOTICE: Future-deprecated features used: * 0.56.0: {'meson.source_root'} While there is a new macro to get it, that would rise the dependency chain to meson 0.56. So, instead, just store it on a variable. Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
Mauro Carvalho Chehab authored
The requirement for having gtk-doc is specific to docs/reference/igt-gpu-tools. Moved the code to be there. This lets build testplan docs without needing gtk-doc (nor having the IGT executables). It should be noticed that testplan indirectly depends on build_tests, for one reason only: the tests subdir is only included if build_tests is true. Without it, the xe_test_config variable will be undefined, as it is declared inside a tests subdir. To ensure that the variable is declared we need to add an extra test for the testplan subdir. Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
Mauro Carvalho Chehab authored
PDF files are easier to share, and it costs almost nothing to produce them with Sphinx, via rst2pdf. Added support for it if rst2pdf is installed. Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
Mauro Carvalho Chehab authored
Add the remaining changes at meson for it to build Xe documentation. Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
Apply the OA timestamp shift to the raw timestamp as well. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
In some OA formats, gpu_ts is a 64 bit value and the right shift can result in bit[31] being set. This throws off the correlation and the timelines. Apply the mask on gpu_ts as well. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
MTL OAM xml files have been updated to specify the OA format used to interpret the counter. The format supported for OAM is I915_OAM_FORMAT_MPEC8u32_B8_C8, hence add support in perf library for I915_OAM_FORMAT_MPEC8u32_B8_C8 format. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
MTL has configuration similar to DG2 for OA timestamp and threads per EU. Update the same for MTL as per Wa_14015846243:mtl. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Add class instance awareness to perf library and enable GPUvis to select specific GT in perf recorder. v2: Use gt 0 for gem_barrier_race test since the race is independent of gt used. (Umesh) v3: Use i915_drm_local.h in perf tools Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Make the asserts in the sanity check descriptive. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Enable non-zero-reason test to run on media engines as well. v2: Check if new params are available before passing (CI failure) Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Metrics config is different for media vs render engines. Choose the metrics configurations as per the engine used for the test. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Add OAM metrics configurations and update MTL GT3 metrics. v2: Fix commit description Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Acked-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Add OAM metrics configurations and update MTL GT2 metrics. v2: Fix commit description Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Acked-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Include OAM SAMedia (Standalone Media) format for MTL. v2: Improve commit message (Kamil) Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Acked-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Add MTL to supprted HW in oa guid registry v2: Improve commit message (Kamil) Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Some counters do not have units specified, so use 'number' as the unit. v2: Drop " from commit description (Kamil) Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Acked-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
If the metric set has OAM format defined, update the generated metrics accordingly. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Acked-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Make chipsets aware of oa formats to allow addition of different formats per chipset. v2: Improve commit message (Kamil) Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Acked-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Add OAM formats and support for media engines in perf tests Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
With multiple OA buffers, verify that the perf interface allows concurrent access to the OA buffers in different groups. v2: (Ashutosh) - Rename concurrency functions to use group instead of gt - Fix comment v3: (CI issues) - Run group-exclusive-stream tests for >= gen12 - Fix exponent for group-exclusive-stream tests Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
With multiple OA buffers, we need a way to group engines per OA buffer. Add support for oa perf groups for tests v2: When populating oa unit id in engine info, check for perf revision Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>#1>
-
Refactor and use a helper for OA format Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
The OA format size will change based on the specific format used in some of the tests. In preparation for that, allocated the space required for the format on heap. v2: Allocate the correct bytes for OA format (CI failure) Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
Similar to OAG (Global) and OAR (Render), OAM (Observability Architecture for Media) adds some new OA formats that require different metrics configurations for B and C counters. Don't do a sanity check for B and C counters for OAM formats until such metrics are available. v2: Improve commit message (Kamil) Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
New OA formats in MTL have 64 bit OA report header. Update the timestamp in the tests to be default 64 bit. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Kamil Konieczny <kamil.konieczny@linux.intel.com>
-
To support 64 bit OA report headers, treat all ticks as 64 bits. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
New perf interface enabled passing engine class instance to the perf interface. Enable some tests to utlize this interface to run the test on supported engines. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Add test to verify class:instance interface to perf OA. v2: Temporarily move uapi to lib/i915/i915_drm_local.h Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
A prior patch changed the way size of properties is calculated. It missed one test. Replace the sizeof with ARRAY_SIZE for buffer-fill test. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Make sure rcs0 is present for gen12-mi-rpc. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Explicitly pass the default engine for the sseu tests. v2: Pass context id in sseu tests to fix CI failure Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Use a helper to set the default engine for use with OA. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
Newer gen12 platforms introduced 64 bit OA formats that are used by MTL. Add support to parse reports for 64 bit counters. Signed-off-by:
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com> Reviewed-by:
Ashutosh Dixit <ashutosh.dixit@intel.com>
-
KunalJoshi95 authored
It is known issue that we don't get hotplug with enabled mode on TypeC connector, issue a disabling modeset for every plug/unplug cycle Signed-off-by:
Kunal Joshi <kunal1.joshi@intel.com> Reviewed-by:
Karthik B S <karthik.b.s@intel.com>
-
KunalJoshi95 authored
In some subtests of kms_chamelium_edid we were trying to plug a port that could already be plugged, so reset chamelium state at the start Signed-off-by:
Kunal Joshi <kunal1.joshi@intel.com> Reviewed-by:
Karthik B S <karthik.b.s@intel.com>
-
- Mar 23, 2023
-
-
Imre Deak authored
The Nouveau/gm200 and the i915 drivers' AUX handlers return -ENXIO if there was no sink detected. The other drivers don't use this error code in the same handler, so assume this error always means a sink detect error. Reviewed-by:
Radhakrishna Sripada <radhakrishna.sripada@intel.com> Signed-off-by:
Imre Deak <imre.deak@intel.com>
-