[PATCH i-g-t, v3 0/6] Prepare lib/intel_compute for SVM/system allocator

[Francois Dugast]

This series provides a series of changes to give more control to the
user of lib/intel_compute when executing a compute kernel. This will
help to test corner cases requiring a specific execution environment
as well as to test the system allocator from the compute kernel
context.

v2: Add patch 1/6, rebase others on top of it, improve documentation

v3: Rework patch 6/6: introduce input/output array address override
    in all pipelines

Francois Dugast (6):
  lib/intel_compute: Make loop count a parameter
  lib/intel_compute: Rename variables for input and output data
  lib/intel_compute: Allow the user to provide a vm
  lib/intel_compute: Allow the user to provide a custom compute kernel
  lib/intel_compute: Give option to skip results check
  lib/intel_compute: Allow the user to override input and output
    addresses

 lib/intel_compute.c       | 353 ++++++++++++++++++++++++--------------
 lib/intel_compute.h       |  33 +++-
 tests/intel/gem_compute.c |   2 +-
 tests/intel/xe_compute.c  |   4 +-
 4 files changed, 262 insertions(+), 130 deletions(-)

-- 
2.43.0