Re: [Intel-gfx] â Fi.CI.BAT: failure for drm/i915/display: Check GGTT to determine phys_base (rev2)

[Andrzej Hajda]

On 07.12.2023 11:10, Andrzej Hajda wrote:
> On 07.12.2023 01:18, Patchwork wrote:
>> *Patch Details*
>> *Series:*    drm/i915/display: Check GGTT to determine phys_base (rev2)
>> *URL:*    https://patchwork.freedesktop.org/series/127130/ 
>> <https://patchwork.freedesktop.org/series/127130/>
>> *State:*    failure
>> *Details:* 
>> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/index.html 
>> <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/index.html>
>>
>>
>>   CI Bug Log - changes from CI_DRM_13990 -> Patchwork_127130v2
>>
>>
>>     Summary
>>
>> *FAILURE*
>>
>> Serious unknown changes coming with Patchwork_127130v2 absolutely need 
>> to be
>> verified manually.
>>
>> If you think the reported changes have nothing to do with the changes
>> introduced in Patchwork_127130v2, please notify your bug team 
>> (I915-ci-infra at lists.freedesktop.org) to allow them
>> to document this new failure mode, which will reduce false positives 
>> in CI.
>>
>> External URL: 
>> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/index.html
>>
>>
>>     Participating hosts (37 -> 34)
>>
>> Missing (3): fi-pnv-d510 fi-snb-2520m bat-dg1-5
>>
>>
>>     Possible new issues
>>
>> Here are the unknown changes that may have been introduced in 
>> Patchwork_127130v2:
>>
>>
>>       IGT changes
>>
>>
>>         Possible regressions
>>
>>   * igt at i915_module_load@load:
>>       o bat-mtlp-8: PASS
>>         
>> <https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_13990/bat-mtlp-8/igt@i915_module_load@load.html> -> INCOMPLETE <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/bat-mtlp-8/igt@i915_module_load@load.html>
> 
> 
> It seems related. I think the patch is correct but it just unveils other 
> display take-over issues.
> Ie with this patch initial_plane_vma returns valid buffer, but 
> subsequent display code fails miserably with kernel panic.
> 
> So until this is not solved, we shouldn't merge the patch, IMO.
> 
> CC: i915 maintainers and display developers


After taking a look on panic log [1], I have found:
[drm:i915_init_ggtt [i915]] Failed to reserve top of GGTT for GuC

I don't know why it is only debug level? It seems serious failure, as a 
result i915_init_ggtt fails and probe fails.

The cause is that initial framebuffer is located at the end of GGTT and 
it overlaps with reserved area (see ggtt_reserve_guc_top).

I am not sure how it can be properly fixed, I guess dirty fix could be
just relocation of vma (hopefully into free area), sth like:
new_gte = gsm + (ggtt->vm.total - GUC_TOP_RESERVE_SIZE - size) / 
I915_GTT_PAGE_SIZE;
memmove(new_gte, gte, size / I915_GTT_PAGE_SIZE);

but I have no idea of possible side effects :)

[1]: 
https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/bat-mtlp-8/pstore0-2849851684_Panic_1.txt

Regards
Andrzej


> 
> Regards
> Andrzej
> 
>>
>>
>>     Known issues
>>
>> Here are the changes found in Patchwork_127130v2 that come from known 
>> issues:
>>
>>
>>       IGT changes
>>
>>
>>         Issues hit
>>
>>   * igt at kms_pm_backlight@basic-brightness at edp-1:
>>       o bat-rplp-1: NOTRUN -> ABORT
>>         
>> <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/bat-rplp-1/igt@kms_pm_backlight@basic-brightness@edp-1.html> (i915#8668 <https://gitlab.freedesktop.org/drm/intel/issues/8668>)
>>
>>
>>         Possible fixes
>>
>>   *
>>
>>     igt at gem_exec_suspend@basic-s0 at lmem0:
>>
>>       o bat-dg2-9: INCOMPLETE
>>         
>> <https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_13990/bat-dg2-9/igt@gem_exec_suspend@basic-s0@lmem0.html> (i915#9275 <https://gitlab.freedesktop.org/drm/intel/issues/9275>) -> PASS <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/bat-dg2-9/igt@gem_exec_suspend@basic-s0@lmem0.html>
>>   *
>>
>>     igt at kms_flip@basic-flip-vs-dpms at d-dp6:
>>
>>       o bat-adlp-11: DMESG-FAIL
>>         
>> <https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_13990/bat-adlp-11/igt@kms_flip@basic-flip-vs-dpms@d-dp6.html> (i915#6868 <https://gitlab.freedesktop.org/drm/intel/issues/6868>) -> PASS <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/bat-adlp-11/igt@kms_flip@basic-flip-vs-dpms@d-dp6.html>
>>   *
>>
>>     igt at kms_flip@basic-flip-vs-modeset at d-dp6:
>>
>>       o bat-adlp-11: DMESG-WARN
>>         
>> <https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_13990/bat-adlp-11/igt@kms_flip@basic-flip-vs-modeset@d-dp6.html> -> PASS <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/bat-adlp-11/igt@kms_flip@basic-flip-vs-modeset@d-dp6.html>
>>   *
>>
>>     igt at kms_pipe_crc_basic@read-crc-frame-sequence at pipe-d-edp-1:
>>
>>       o bat-rplp-1: ABORT
>>         
>> <https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_13990/bat-rplp-1/igt@kms_pipe_crc_basic@read-crc-frame-sequence@pipe-d-edp-1.html> (i915#8668 <https://gitlab.freedesktop.org/drm/intel/issues/8668>) -> PASS <https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_127130v2/bat-rplp-1/igt@kms_pipe_crc_basic@read-crc-frame-sequence@pipe-d-edp-1.html>
>>
>>
>>     Build changes
>>
>>   * Linux: CI_DRM_13990 -> Patchwork_127130v2
>>
>> CI-20190529: 20190529
>> CI_DRM_13990: 85d33d0ad82a5c1a71492f14a5ceb67ada6a22d8 @ 
>> git://anongit.freedesktop.org/gfx-ci/linux
>> IGT_7626: 154b7288552cd7ed3033f8ef396e88d0bd1b7646 @ 
>> https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
>> Patchwork_127130v2: 85d33d0ad82a5c1a71492f14a5ceb67ada6a22d8 @ 
>> git://anongit.freedesktop.org/gfx-ci/linux
>>
>>
>>       Linux commits
>>
>> 43f210e851cd drm/i915/display: Check GGTT to determine phys_base
>>
>