[PATCH 0/6] Ground foundation for VCE 1 implementation

[Alexandre Demers]

Oh and patch 3 "Small precision when failing to load legacy firmware"
should have been a standalone patch, it is related to the radeon driver.
I'll send it as so with drm/radeon in the title.

Alexandre Demers

On 2017-09-08 07:55, Alexandre Demers wrote:
> Hi Christian!
>
> On 2017-09-08 04:24, Christian König wrote:
>> First of all please prefix all subject lines for amdgpu kernel patches
>> with drm/amdgpu.
> You're right, I forgot.
>> Additional to that I think we won't accept loading firmware files in
>> the old format. Instead please just add the header to the files. If
>> that turns out to be problematic I can start looking into providing
>> those.
> Noted, I'll drop the code and I'll create a temporary firmware with the
> needed header. However, it would be convenient to see newer firmwares
> for both VCE and UVD (cheers Trevor) coming from an official source
> (AMD). Let's remove the old firmware loading patch and propose a glued
> firmware.
>> Apart from that thanks a lot for this work, this is really welcome.
>> But does it actually work? Have you tested it with mesa?
> This work doesn't activate any VCE functionality. On the other hand, it
> shouldn't break anything either. It needs the next series of patches.
> While the firmware loading and initialization seem to go through
> correctly, I'm encountering hangs at boot. That's where I'll need some help.
>
> I'll post the next series in the next few days as a RFC so others can
> help me go through. I just need to split the code in a few patches.
>
> Cheers,
> Alex
>> Thanks,
>> Christian.
>>
>> Am 08.09.2017 um 04:48 schrieb Alexandre Demers:
>>> This is the foundation for VCE 1 implementation, mostly ported from
>>> radeon to
>>> amdgpu. This should have no impact for now on the usability of the
>>> driver. This
>>> is the pretty straight-forward part. The heavier part is coming as a
>>> RFC.
>>>
>>> I'd like to have this first part available to all, since I don't have
>>> as much
>>> time as I'd like to work on VCE 1 for now, but this could help others
>>> to help me
>>> on the task.
>>>
>>> If you are not satisfied with this first series, this can also be
>>> considered as
>>> a RFC. Please, comment as needed.
>>>
>>> Alexandre Demers (6):
>>>    Add initial VCE_V1_0 files.
>>>    Reorganize, rename, move and split some VCE 1 defines.
>>>    Small precision when failing to load legacy firmware.
>>>    Moving read and write pointer functions from radeon to amdgpu
>>>    Add support for SI's VCE firmwares (with and without header)
>>>    Fix indentation.
>>>
>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c            | 100 ++++--
>>>   drivers/gpu/drm/amd/amdgpu/si.c                    |  17 +-
>>>   drivers/gpu/drm/amd/amdgpu/si_dma.c                |   5 +-
>>>   drivers/gpu/drm/amd/amdgpu/si_ih.c                 |  21 +-
>>>   drivers/gpu/drm/amd/amdgpu/sid.h                   |  70 ----
>>>   drivers/gpu/drm/amd/amdgpu/vce_v1_0.c              | 394
>>> +++++++++++++++++++++
>>>   drivers/gpu/drm/amd/amdgpu/vce_v1_0.h              |  29 ++
>>>   .../gpu/drm/amd/include/asic_reg/vce/vce_1_0_d.h   |  80 +++--
>>>   .../drm/amd/include/asic_reg/vce/vce_1_0_sh_mask.h |  13 +-
>>>   drivers/gpu/drm/radeon/radeon_uvd.c                |   2 +-
>>>   10 files changed, 589 insertions(+), 142 deletions(-)
>>>   create mode 100644 drivers/gpu/drm/amd/amdgpu/vce_v1_0.c
>>>   create mode 100644 drivers/gpu/drm/amd/amdgpu/vce_v1_0.h
>>>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20170908/9ff8371c/attachment.html>