[Bug 35749] New: Screen update problems on Counter Strike: Global Offensive

[wine-bugs at winehq.org]

http://bugs.winehq.org/show_bug.cgi?id=35749

            Bug ID: 35749
           Summary: Screen update problems on Counter Strike: Global
                    Offensive
           Product: Wine
           Version: 1.7.14
          Hardware: x86
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: directx-d3d
          Assignee: wine-bugs at winehq.org
          Reporter: gediminas at varciai.lt
                CC: ken at codeweavers.com, stefan at codeweavers.com
   Regression SHA1: 4c4552c5a1910a9d5adf8eccff0ac62d89ffe376

A new regression:
While playing Counter Strike: Global Offensive (other source engine games might
also be affected; did not test yet), it runs fine at first, but if I alt-tab
out of it and back several times, the game starts stuttering a lot and the
*apparent* framerate drops several times.
Even though it looks as if the framerate plummeted, the in-game FPS counter
keeps showing a normal framerate, i.e. there is no framerate drop from the
game's point of view. GPU load also remains the same, too. Thus, this leads to
believe that rendering is still done normally, but Wine doesn't flip / update
the screen properly, leading to stuttering and apparent framerate drop.

After some testing, I found that reverting this patch makes the problem go
away:

commit 4c4552c5a1910a9d5adf8eccff0ac62d89ffe376
Author: Ken Thomases <ken at codeweavers.com>
Date:   Wed Feb 19 16:14:53 2014 -0600

    wined3d: Restore the pixel format of the window whose pixel format was
actually changed.

I also tried it with the patch from bug #35705 - It did not help at all, so it
is possibly a different bug.

Judging from its description, it does not seem that bug #35655 would the same
bug, either.

With the bug triggered, I did not see any errors / warnings / fixmes that were
not present with the bug NOT triggered, even with +d3d on. So, I am not sure
whether posting debug output would be of any use.

Although it in itself is unsupported, to shed more light on the matter, I
should also mention about the behavior of this bug when encountered on Wine
with Stefan Dösinger's CSMT patchset:
Everything is the same, except the game is stuck on showing one frame that was
drawn right before triggering the bug. Which *visually* makes it seem as the
game has hanged. Leaving game like this for longer amounts of time, I can see
that it then flips / updates the screen approximately once every one minute
[leading to 0.016 frames per second, effectively]. Sometimes, after waiting a
minute or two, graphics start updating / working normally again, but that lasts
for fifteen seconds tops and it gets stuck on showing one frame again.

-- 
Do not reply to this email, post in Bugzilla using the
above URL to reply.
You are receiving this mail because:
You are watching all bug changes.