Re: [PATCH 3/5] dma-buf: return only unsignaled fences in dma_fence_unwrap_for_each v3

[Karolina Drobnik <karolina.drobnik@intel.com>]

Hi Christian,

On 11.07.2022 14:25, Christian König wrote:
> Hi Karolina,
> 
> Am 11.07.22 um 14:17 schrieb Karolina Drobnik:
>> Hi Christian,
>>
>> On 11.07.2022 11:57, Christian König wrote:
>>> Hi Karolina,
>>>
>>> Am 11.07.22 um 11:44 schrieb Karolina Drobnik:
>>>> Hi Christian,
>>>>
>>>> I'm sorry for digging this one out so late.
>>>>
>>>> On 06.05.2022 16:10, Christian König wrote:
>>>>> dma_fence_chain containers cleanup signaled fences automatically, so
>>>>> filter those out from arrays as well.
>>>>>
>>>>> v2: fix missing walk over the array
>>>>> v3: massively simplify the patch and actually update the description.
>>>>>
>>>>> Signed-off-by: Christian König <christian.koenig@amd.com>
>>>>> ---
>>>>>   include/linux/dma-fence-unwrap.h | 6 +++++-
>>>>>   1 file changed, 5 insertions(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/include/linux/dma-fence-unwrap.h 
>>>>> b/include/linux/dma-fence-unwrap.h
>>>>> index e7c219da4ed7..a4d342fef8e0 100644
>>>>> --- a/include/linux/dma-fence-unwrap.h
>>>>> +++ b/include/linux/dma-fence-unwrap.h
>>>>> @@ -43,9 +43,13 @@ struct dma_fence *dma_fence_unwrap_next(struct 
>>>>> dma_fence_unwrap *cursor);
>>>>>    * Unwrap dma_fence_chain and dma_fence_array containers and deep 
>>>>> dive into all
>>>>>    * potential fences in them. If @head is just a normal fence only 
>>>>> that one is
>>>>>    * returned.
>>>>> + *
>>>>> + * Note that signalled fences are opportunistically filtered out, 
>>>>> which
>>>>> + * means the iteration is potentially over no fence at all.
>>>>>    */
>>>>>   #define dma_fence_unwrap_for_each(fence, cursor, head)            \
>>>>>       for (fence = dma_fence_unwrap_first(head, cursor); fence;    \
>>>>> -         fence = dma_fence_unwrap_next(cursor))
>>>>> +         fence = dma_fence_unwrap_next(cursor)) \
>>>>> +        if (!dma_fence_is_signaled(fence))
>>>>>     #endif
>>>>
>>>> It looks like this particular patch affects merging Sync Fences, 
>>>> which is reflected by failing IGT test (igt@sw_sync)[1]. The failing 
>>>> subtests are:
>>>>   - sync_merge - merging different fences on the same timeline, neither
>>>>          single nor merged fences are signaled
>>>>
>>>>   - sync_merge_same - merging the fence with itself on the same
>>>>          timeline, the fence didn't signal at all
>>>>
>>>>   - sync_multi_timeline_wait - merging different fences on different
>>>>          timelines; the subtest checks if counting fences of
>>>>          various states works. Currently, it can only see 2
>>>>          active fences, 0 signaling (should be 2 active,
>>>>          1 signaling)
>>>>
>>>> Reverting this commit on the top of drm-tip fixes the issue, but I'm 
>>>> not sure if it wouldn't impact other places in the code. Please let 
>>>> me know if I can be of any help.
>>>
>>>
>>> Thanks for letting me know. Not sure what's going on here, but I can 
>>> take a look today if time permits.
>>
>> The reproduction with IGTs should be quite easy. You'll need to 
>> clone/download the IGT code and follow instructions for Building[1] 
>> the project (make sure you have meson and ninja installed):
>>
>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.freedesktop.org%2Fdrm%2Figt-gpu-tools&amp;data=05%7C01%7Cchristian.koenig%40amd.com%7C9a9587aefd2d4ac2d86208da63375cb6%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637931386683611766%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=4WsMutcFJ2HwBqld%2BTv9N1Tx6cbFMwJJZ6kjm5rbfoI%3D&amp;reserved=0 
>>
>>
>> Once you have it up and running, go to <igt path>/build/tests, and run 
>> the subtests:
>>
>>   ./sw_sync --run sync_merge
>>   ./sw_sync --run sync_merge_same
>>   ./sw_sync --run sync_multi_timeline_wait
>>
>> You can run all the subtests with ./sw_sync, but I think these are the 
>> most relevant to you.
> 
> Thanks, I've already managed to reproduce it.
> 
> Not sure what's going on here, but could be that the test case was never 
> correct in the first place. Need to double check.

That's also a possibility, but I couldn't verify it before writing to 
you, as it's not my area of expertise.

Thanks for taking a look at this.

All the best,
Karolina

> Thanks,
> Christian.
> 
>>
>> Many thanks,
>> Karolina
>>
>> ------------------
>> [1] - 
>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.freedesktop.org%2Fdrm%2Figt-gpu-tools%23building&amp;data=05%7C01%7Cchristian.koenig%40amd.com%7C9a9587aefd2d4ac2d86208da63375cb6%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637931386683611766%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=FV0Ao6ra8EOyr4cOs4N7mCmpOEUUObTrgyOrd0tvEV8%3D&amp;reserved=0 
>>
>>
>>> Do you have a description how to easy reproduce this? E.g. how to run 
>>> just those specific igts?
>>>
>>> Thanks,
>>> Christian.
>>>
>>>>
>>>> All the best,
>>>> Karolina
>>>>
>