igt-gpu-tools.git/benchmarks, branch vm-bind

benchmarks/gem_blt: fix baseline estimation

2022-06-15T06:33:47+00:00

This test is expected to run under a certain time defined by a command line parameter (by default, 2s). In order to achive the specified time, the baseline() function tries to estimate the value of a counter that will get the minimal amount of interaction to wait for > 0.1s. However, currently, the baseline estimation is broken, returning a too big number, as it is measuring the memcpy() time, instead of actually gem_execbuf(). Due to that, a default test without passing any command line parameter would take more than 1.5 years to output the first benchmark data! Fix it by using the same logic that it is inside run() at the baseline time estimation. Before this patch, baseline could return 399242693. After it, it now return 20. That means an interval of about 120 ms at the code which runs a gem_execbuf() loop, which should be good enough to ensure that the tests won't take too long, and will approximately match the time specified by the excecution parameter '-t'. Reviewed-by: Kamil Konieczny Signed-off-by: Mauro Carvalho Chehab

lib/intel_reg: Add common MI_* macros to remove duplicates

2022-06-13T15:28:24+00:00

In few tests we got some MI_* duplicates (MI_MATH for example). Add common definitions in intel_reg.h and remove local definitions in the tests. v2: Definitions MI_LOAD_REGISTER_MEM_GEN8 was removed so from now on user will need to encode length on it own. : Signed-off-by: Zbigniew Kempczyński Reviewed-by: Petri Latvala

benchmarks/gem_exec_fault: Add softpin mode to support gens with ppgtt

2021-10-15T03:54:36+00:00

Alignment trick doesn't work properly for ppgtt gens - kernel is able to keep previous offset and doesn't call unbind/bind. With softpin on ppgtt we're able to enforce rebind and benchmark should behave correctly on such gens. To avoid inaccurate results kernel CONFIG_PROVE_LOCKING should be set to N, otherwise kernel can call unbind/bind for same offset more than one (backoff is not visible from userspace). v2: rename to gem_allows_obj_alignment() Signed-off-by: Zbigniew Kempczyński Cc: Petri Latvala Cc: Ashutosh Dixit Reviewed-by: Ashutosh Dixit

benchmarks/gem_exec_fault: Add timeout argument

2021-10-15T03:54:36+00:00

Add timeout argument and change elapsed time to inner loop to be more precise in timeout processing. Signed-off-by: Zbigniew Kempczyński Cc: Ashutosh Dixit Cc: Petri Latvala Reviewed-by: Ashutosh Dixit

benchmarks/gem_busy: nuke LOCAL_IOCTL_SYNCOBJ_WAIT

2021-07-15T19:01:57+00:00

This is DRM_IOCTL_SYNCOBJ_WAIT. Signed-off-by: Lucas De Marchi Acked-by: Daniel Vetter Reviewed-by: Ashutosh Dixit

benchmarks/gem_busy: nuke LOCAL_IOCTL_SYNCOBJ_CREATE

2021-07-15T19:01:57+00:00

This is DRM_IOCTL_SYNCOBJ_CREATE. Signed-off-by: Lucas De Marchi Acked-by: Daniel Vetter Reviewed-by: Ashutosh Dixit

benchmarks/gem_exec_tracer: nuke LOCAL_IOCTL_I915_GEM_EXECBUFFER2_WR

2021-07-15T19:01:57+00:00

Use DRM_IOCTL_I915_GEM_EXECBUFFER2_WR from kernel header. Signed-off-by: Lucas De Marchi Acked-by: Daniel Vetter Reviewed-by: Ashutosh Dixit

lib/i915/gem_create: Add gem_create_ext

2021-05-27T10:36:21+00:00

Add a wrapper for gem_create_ext ioctl (a version of gem_create that accepts extensions). In preparation for the driver change implementing it, a local definition of its id and necessary structs have been added, which are to be erased as soon as those definitions appear in the i915_drm.h file. The new ioctl wrapper is added to a separate file. For consistency the wrapper of the old ioctl, gem_create is moved from ioctl_wrappers to gem_create. Signed-off-by: Andrzej Turko Cc: Zbigniew Kempczynski Cc: Dominik Grzegorzek Cc: Petri Latvala Cc: Chris P Wilson Signed-off-by: Matthew Auld Acked-by: Petri Latvala

benchmarks/gem_userptr_benchmark: Remove tests with unsynchronized flag

2021-05-26T09:10:16+00:00

As flag I915_USERPTR_UNSYNCHRONIZED is not supported by the i915 anymore remove tests which tries to use it in the benchmark. Remove benchmark from meson libdrm depending code as it is a little bit confusing. Signed-off-by: Zbigniew Kempczyński Cc: Tvrtko Ursulin Cc: Petri Latvala Reviewed-by: Tvrtko Ursulin

gitignore: Remove various .gitignore files

2021-04-23T10:27:44+00:00

Now that autotools is gone and you always build to a dedicated build directory with meson, we can remove .gitignore files that only had the purpose of ignoring built binaries in the source directories. Signed-off-by: Petri Latvala Cc: Arkadiusz Hiler Reviewed-by: Arkadiusz Hiler