[PATCH i-g-t v2 0/4] Simplify compute preemption tests

[Zbigniew Kempczyński]

I planned to rewrite threadgroup preemption test to use similar to wmtp
pattern, but at the moment I haven't succeeded. To make my further work
on this easier I simplified the code and logic of verification wmtp or
threadgroup preemption occurred. So threadgroup preemption still rely
on increment loop shader which is executed on large x-dimension. I hope
I'll find a way to rewrite it to use wmtp pattern where I release part
of threads triggering preemption when threadgroup will end with EOT.

v2: use ARRAY_SIZE() and simplify input/output data assignment

Cc: Zbigniew Kempczyński <zbigniew.kempczynski at intel.com>
Cc: Francois Dugast <francois.dugast at intel.com>

Zbigniew Kempczyński (4):
  lib/intel_compute: move user-fence to allow async exec call
  lib/intel_compute: simplify compute-preempt subtests
  lib/intel_compute: calculate array size at compile time
  lib/intel_compute: refactor to remove code duplication

 lib/intel_compute.c | 395 ++++++++++++++++++--------------------------
 1 file changed, 161 insertions(+), 234 deletions(-)

-- 
2.43.0