https://bugs.winehq.org/show_bug.cgi?id=46148
--- Comment #6 from Henri Verbeet hverbeet@gmail.com --- (In reply to Zebediah Figura from comment #5)
If I did, I would have shared them. As things stand, we haven't really made any modifications to d3d-related patches (or, IIRC, any patches?) that would have caused this regression. On the other hand, we may have messed up a rebase somewhere.
Alistair mentioned the rcp thing for SpinTires in the 3.20 announcement, which I'd expect to have a performance impact, but not as severe as described here.
I think it's probably more likely that something was broken upstream, but I have no intuition as to what that might be, as this seems to be a D3D bug and I have not worked on that area. Hence my request for comment from the D3D developers.
Sorry, nothing comes to mind. We did have a decent number of D3D patches in wine-3.20, but they're more architectural than changing application visible behaviour.