[Bug 99842] GL_ARB_transform_feedback2 on i965 gen6

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Fri Feb 17 07:38:46 UTC 2017


Kenneth Graunke <kenneth at whitecape.org> changed:

           What    |Removed                     |Added
             Status|NEW                         |ASSIGNED
           Assignee|intel-3d-bugs at lists.freedes |kenneth at whitecape.org
                   |ktop.org                    |

--- Comment #1 from Kenneth Graunke <kenneth at whitecape.org> ---
Looking at this again, we apparently never implemented Pause and
ResumeTransformFeedback() driver hooks...which meant that we actually failed to
restore the streamout buffer writing location during [Pause A -> Bind B -> ...
-> Bind A -> Resume A] situations.

Honestly, I think we forgot that this was a part of ES 3.0...there aren't any
failing tests for this in the CTS, so it looked like everything was working. 
There is a failing Piglit test, which I'm guessing came about after we had done
ES 3.0 and moved on.

I remember looking at implementing GL_ARB_transform_feedback2 back in the day. 
I saw the lack of MI_LOAD_REGISTER_MEM and figured Pause/Resume would be
painful to implement.  Of course, I failed to notice that there's a command to
do this...but the hard part is that there isn't a register to read, making it
hard to obtain the value at Pause time.

Of course, this is part of ES 3.0, so we have to do it anyway.  I've found a
solution that seems to work, and should fix those problems.

At that point, GL_ARB_transform_feedback2 only has one feature over and above
ES 3.0 - glDrawTransformFeedback().  However...that's easy.  The mechanism for
the above bug fix makes this just work with no additional code.

So I think we can turn it on.  I have no idea how well WebGL 2.0 will work...

You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20170217/4a262426/attachment.html>

More information about the intel-3d-bugs mailing list