drm/amdgpu: Convert to common fdinfo format v5
authorChristian König <christian.koenig@amd.com>
Wed, 11 May 2022 09:06:26 +0000 (11:06 +0200)
committerAlex Deucher <alexander.deucher@amd.com>
Thu, 26 May 2022 18:56:34 +0000 (14:56 -0400)
commitaf0b541670090e87996e0894bd0e457edf617541
treeaf10dd4b7fe4f156c35394aa3db4a455d42e954a
parent08cffb3eb731fefd0dea12424cedbfa63c356ee0
drm/amdgpu: Convert to common fdinfo format v5

Convert fdinfo format to one documented in drm-usage-stats.rst.

It turned out that the existing implementation was actually completely
nonsense. The calculated percentages indeed represented the usage of the
engine, but with varying time slices.

So 10% usage for application A could mean something completely different
than 10% usage for application B.

Completely nuke that and just use the now standardized nanosecond
interface.

v2: drop the documentation change for now, nuke percentage calculation
v3: only account for each hw_ip, move the time_spend to the ctx mgr.
v4: move general ctx changes into separate patch, rework the fdinfo to
    ctx_mgr interface so that all usages are calculated at once, drop
    some unecessary and dangerous refcount dance.
v5: add one more comment how we calculate the time spend

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Shashank Sharma <shashank.sharma@amd.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h
drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c