libva Add description of the return status for vaSyncSurface()

this will indicate APP behavior when meet VA_STATUS_ERROR_DECODING_ERROR from vaSyncSurface

Nov 27 '23 10:11 pengxin99

Hi, @XinfengZhang @Jexu, have add more detail description, could you pls help to review this PR again?

Dec 15 '23 07:12 pengxin99

I agree with the direction, provide hint to application , then application could handle correlated errors , ignore , or reset the context etc. .
but I suggest to clarify the difference. currently: VA_STATUS_ERROR_DECODING_ERROR may caused by a bistream corruption. related with syntax VA_STATUS_ERROR_HW_BUSY is about the HW reset , but looks could not distinguish , media reset or gpu reset. if media reset , it could be ignored directly. it also related with #714?

but operation failed is unclear.

Jan 09 '24 08:01 XinfengZhang

@XinfengZhang Thanks for the comments. and I will refine this PR for below two point:

VA_STATUS_ERROR_HW_BUSY is about the HW reset , but looks could not distinguish , media reset or gpu reset. if media reset , it could be ignored directly. it also related with https://github.com/intel/libva/pull/714?

The VA_STATUS_ERROR_HW_BUSY is only for HW reset, so cannot ignore this error, vaSyncSurface will return success if media reset happended.

but operation failed is unclear.

can see the below comment: https://github.com/intel/libva/pull/774#discussion_r1449800161

Jan 12 '24 04:01 pengxin99

Also, please, squash commits. For this change there should be just 1 commit in the PR.

Jan 25 '24 00:01 dvrogozh

Also, please, squash commits. For this change there should be just 1 commit in the PR.

@dvrogozh Thanks for your review, and have refined this PR, could pls have a look again?

Jan 26 '24 02:01 pengxin99

Hi, @dvrogozh @XinfengZhang , Can you take another look at this PR if you have time?

Feb 20 '24 02:02 pengxin99

Hi, @XinfengZhang , Could you pls take another look at this PR when you have time?

Mar 06 '24 08:03 pengxin99