[Intel-gfx] ✗ Fi.CI.BAT: failure for drm/i915/guc: Update to GuC v49

[John Harrison]

On 9/21/2020 13:36, Daniele Ceraolo Spurio wrote:
> On 9/21/2020 12:22 PM, John Harrison wrote:
>> On 9/16/2020 23:48, Petri Latvala wrote:
>>> On Wed, Sep 16, 2020 at 06:22:45PM -0700, John Harrison wrote:
>>>> Hello,
>>>>
>>>> The failures below all appear to be because the new GuC firmware 
>>>> was not
>>>> found on the test system.
>>>>
>>>> My understanding is that all we need to do to get the CI system to 
>>>> update
>>>> with new firmwares is to push the firmware to a branch on the FDO
>>>> drm-firmware repo and then send a pull request to this mailing 
>>>> list. That
>>>> was done yesterday.
>>> That pull request used an ssh:// url though. Can you send it again
>>> with a git:// url? I suppose that's a plausible reason why I don't see
>>> the binaries in CI's deploy dir.
>>>
>>
>> Hello,
>>
>> We reset the pull request with a git:// URL as requested. I even gave 
>> it a full weekend to propagate through. However, I am still getting 
>> missing firmware failures after posting a new patch set.
>>
>> John.
>>
>
> You sure you're looking at the correct logs? AFAICS BAT on the new 
> patches passed (https://patchwork.freedesktop.org/series/81906/) and I 
> see the correct GuC being loaded in the logs.
> e.g. 
> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18542/fi-icl-u2/boot0.txt:
>
> <6>[    9.283866] i915 0000:00:02.0: [drm] GuC firmware 
> i915/icl_guc_49.0.1.bin version 49.0 submission:disabled
>
> Daniele

Grrr. When I downloaded the dmesg.txt it saved is as dmesg.txt.1. So 
yes, I was looking at a stale log :(.

I guess the other question is are we supposed to be supporting HuC by 
default on KBL? It is running the gem_huc_copy test but we have only 
enabled GuC/HuC loading on Gen11+ platforms.

John.