[Bug 59555] New: Star Wars: Knights of the Old Republic - Rendering failure with active framebuffer effects
http://bugs.winehq.org/show_bug.cgi?id=59555 Bug ID: 59555 Summary: Star Wars: Knights of the Old Republic - Rendering failure with active framebuffer effects Product: Wine Version: 11.5 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@list.winehq.org Reporter: accounts.winehq@vv221.fr Distribution: --- Up to WINE 10.0, framebuffer effects could be enabled in Star Wars: Knights of the Old Republic with no adverse effect on rendering. With WINE 11.5 it leads to a failure to render the game world (including any geometry, models, etc.). Text dialogues and other 2D interface pieces can be seen, with some heavy flickering effect. When trying to take a screenshot I only get a solid black image, even when text dialogue could be seen on the actual game screen. The actual gameplay seems to work fine, this is almost certainly only a rendering error. I have not tested other WINE builds yet, but I am going to do that to get a better idea of where the regression is hiding. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 Antoine Le Gonidec <accounts.winehq@vv221.fr> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 --- Comment #1 from Antoine Le Gonidec <accounts.winehq@vv221.fr> --- Last working version: 10.5 First broken version: 10.13 It looks like the breakage is progressive, with 3D models and environment flashing in and out of visibility with WINE 10.13 each time a mouse click is registered, while they are totally invisible with WINE 11.0. There is a big gap between the last working version (10.5) and the first broken one (10.13), because the gameplay can not be reached with WINE versions 10.6 to 10.12. Instead the game loading (right after character creation) stops near the end, and never switch to the introduction video. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 --- Comment #2 from Antoine Le Gonidec <accounts.winehq@vv221.fr> --- It might be interesting to note that the sequel, Star Wars: Knights of the Old Republic II - The Sith Lords, is not affected. In that one the framebuffer effects can be enabled (with WINE 11.5) without triggering any kind of rendering failure. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 Béla Gyebrószki <gyebro69@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |gyebro69@gmail.com --- Comment #3 from Béla Gyebrószki <gyebro69@gmail.com> --- What GPU and driver version are you using? I can reproduce the problem with NVK/Mesa-26.0.3 but not with the binary Nvidia driver 595.58.03. Rendering is normal in Wine 10.5 but broken in 10.6 when using NVK/Mesa driver. Could you perform a regression test ? https://gitlab.winehq.org/wine/wine/-/wikis/Regression-Testing -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 --- Comment #4 from Antoine Le Gonidec <accounts.winehq@vv221.fr> ---
Could you perform a regression test ?
Sadly, no. Versions 10.6 to 10.12 simply fail to load gameplay on my setup, so that other bug is preventing me from looking for the rendering regression. I would probably need to first find the fix for the loading failure by running a bisect between 10.12 and 10.13, then hope that fix can apply cleanly on builds as old as 10.5 to include it in the actual bisect between 10.5 and 10.6. --- To be clear: I am not saying I can not perform a bisect at all, but that it is going to be much more time-consuming than what I initially expected. So this will have to wait until I have more time to dedicate to that bug. Hopefully in a couple days/weeks, but I can not make any promise here. --- Since you get no problem to run KotOR with WINE 10.6, unlike me, it might be worth noting that I am running amdgpu/Mesa (26.0.3). -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 --- Comment #5 from Béla Gyebrószki <gyebro69@gmail.com> --- (In reply to Antoine Le Gonidec from comment #4)
Could you perform a regression test ?
Sadly, no. Versions 10.6 to 10.12 simply fail to load gameplay on my setup, so that other bug is preventing me from looking for the rendering regression.
I would probably need to first find the fix for the loading failure by running a bisect between 10.12 and 10.13, then hope that fix can apply cleanly on builds as old as 10.5 to include it in the actual bisect between 10.5 and 10.6.
Just a wild guess but maybe adding these options in swkotor.ini may prevent the loading failure with 10.6 and newer for you? under [Graphics Options] section: FullScreen=0 AllowWindowedMode=1 under [Game Options] section: Disable Movies=1 -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 --- Comment #6 from Antoine Le Gonidec <accounts.winehq@vv221.fr> --- I was going to give a try to your suggestion… but this time I got no loading freeze with WINE 10.6 ?! Well, I’m not going to waste time investigating that heisenbug, and focus on the rendering one instead. I do get the rendering failure with WINE 10.6, so I should be able to run a bisect between 10.5 and 10.6 soon (in the coming hours/days). -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 --- Comment #7 from Antoine Le Gonidec <accounts.winehq@vv221.fr> --- I am starting to wonder if I tested another version by accident at the time of my earlier post. Now no matter if I use my fresh local build of WINE 10.6 or version 10.6~trixie-1 from WineHQ APT repository, in both cases I’m back at the failure to load the game. Disabling the movies does not help, despite the failure actually happening just before a movie is supposed to be played. Alternating between full screen and windowed mode does not help either, nor does enabling/disabling a virtual desktop. So I guess I won’t be able to skip looking for the stuck loading bug fix first, and hope I can apply it on top of 10.5. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 Antoine Le Gonidec <accounts.winehq@vv221.fr> changed: What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1| |345139cf71ab06712acf6ae4601 | |7df1e864df709 Component|-unknown |winex11.drv --- Comment #8 from Antoine Le Gonidec <accounts.winehq@vv221.fr> --- After more experiments, I noticed the loading failure only happens with framebuffer effects enabled. So it might be another symptom of the same underlying regression. With that in mind I am running a bisect where I am going to treat both symptoms (loading failure or bad rendering) as the same problem, let us see what I can spot… --- Here is the faulty commit: commit 345139cf71ab06712acf6ae46017df1e864df709 Author: Rémi Bernon <rbernon@codeweavers.com> Date: Thu Mar 6 16:28:37 2025 +0100 winex11: Use the generic context functions. 44 150 dlls/winex11.drv/opengl.c --- Reverting it on top on WINE 10.6 leads to a crash before the game even opens its first window, limiting my ability to look for a workaround. What I can confirm is that the previous commit (508f701c0e04e195804f00c889e4e5e8930f56f6) has no noticeable bug with framebuffer effects enabled. With the faulty commit on the other hand gameplay can not be reached (loading is stuck a bit before the end) if framebuffer effects are enabled. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59555 Béla Gyebrószki <gyebro69@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rbernon@codeweavers.com -- 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.
participants (1)
-
WineHQ Bugzilla