https://bugs.winehq.org/show_bug.cgi?id=37750
--- Comment #10 from Henri Verbeet hverbeet@gmail.com --- (In reply to Béla Gyebrószki from comment #9)
I tested your patch and it indeed fixes the issue for me, thank you. Was the debug log helpful for you or the result of the bisect gave you some hints?
Mostly the debug log. The CSMT patch doesn't seem directly related to the issue, I guess it mostly ends up changing the order in which things get called, which happens to make the bug go away. (Which would make sense for this kind of bug.) Similarly, the regression commit isn't directly related to the issue either, the wined3d_texture_bind_and_dirtify() calls were missing before that commit as well, but it cuts down on the amount of times other things end up binding the texture.