[Bug 44808] New: Regression in 3.4: Graphics show as garbage with multisampled textures enabled in The Witness
https://bugs.winehq.org/show_bug.cgi?id=44808
Bug ID: 44808 Summary: Regression in 3.4: Graphics show as garbage with multisampled textures enabled in The Witness Product: Wine Version: 3.4 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: zerooneo@01101001.net Distribution: ---
In the game The Witness, multisampling must be enabled in order for objects with transparent areas, such as trees and plants, to render correctly. These are an important aspect of the game.
Wine version 3.3, it's possible to play The Witness at full graphics quality, as long as the 64-bit version of the game is used with the 64-bit Wine, and multisampling is enabled.
However, in Wine version 3.4, all in-game graphics show as garbage data when running The Witness with multisampling enabled.
Somehow, the ability to display multisampled textures correctly has disappeared between versions 3.3 and 3.4.
This observation was made on a mobile graphics card bought in 2012, using the open source AMD driver.
https://bugs.winehq.org/show_bug.cgi?id=44808
--- Comment #1 from David Oftedal zerooneo@01101001.net --- It should also be noted that this also concerns the 64-bit versions of Wine and The Witness.
https://bugs.winehq.org/show_bug.cgi?id=44808
David Oftedal zerooneo@01101001.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|-unknown |directx-d3d
https://bugs.winehq.org/show_bug.cgi?id=44808
Nikolay Sivov bunglehead@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|major |normal
https://bugs.winehq.org/show_bug.cgi?id=44808
David Oftedal zerooneo@01101001.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1| |72de747c6a6cdd0be84a24932a4 | |452f453861dbe
https://bugs.winehq.org/show_bug.cgi?id=44808
--- Comment #2 from David Oftedal zerooneo@01101001.net --- The regression appeared in commit 72de747c6a6cdd0be84a24932a4452f453861dbe - "wined3d: Pass the depth/stencil view as wined3d_rendertarget_info structure to context_apply_fbo_state()."
It appears to replace a struct wined3d_rendertarget_info called ds_info with NULL in one or more function calls, and replaces it with a call to wined3d_rendertarget_view_get_surface() in one instance.
https://bugs.winehq.org/show_bug.cgi?id=44808
Józef Kucia joseph.kucia@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression CC| |joseph.kucia@gmail.com
https://bugs.winehq.org/show_bug.cgi?id=44808
--- Comment #3 from Henri Verbeet hverbeet@gmail.com --- Created attachment 60866 --> https://bugs.winehq.org/attachment.cgi?id=60866 Mesa patch
I believe this is a bug in Mesa. The attached patch (against current Mesa master) should fix the issue.
It appears that texture views created on multi-sample textures don't inherit the sample count from the original texture, resulting in incomplete framebuffer objects when such views are mixed with other multi-sampled textures.
https://bugs.winehq.org/show_bug.cgi?id=44808
--- Comment #4 from Henri Verbeet hverbeet@gmail.com --- Created attachment 60867 --> https://bugs.winehq.org/attachment.cgi?id=60867 Wine hack
Alternatively, the attached hack against Wine should restore the previous (not quite correct) Wine behaviour.
https://bugs.winehq.org/show_bug.cgi?id=44808
--- Comment #5 from David Oftedal zerooneo@01101001.net --- Thank you very much for that!
I don't have the source code for Mesa ready, but I was able to test the Wine patch, with the latest commit (e1c7a1f7ce03c1e69e008378e90523e85e1c6e8f / wine-3.4-178-ge1c7a1f7ce), and it most definitely resolves the graphics issues with The Witness.
https://bugs.winehq.org/show_bug.cgi?id=44808
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |NOTOURBUG Status|UNCONFIRMED |RESOLVED
--- Comment #6 from Henri Verbeet hverbeet@gmail.com --- This should be resolved by Mesa commit 0b73c86b8030a7f7cb35fc85c83eff7f2b8c24a4.
https://bugs.winehq.org/show_bug.cgi?id=44808
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #7 from Matteo Bruni matteo.mystral@gmail.com --- The fix is now in multiple Mesa releases, closing.
-
wine-bugs＠winehq.org