https://bugs.winehq.org/show_bug.cgi?id=57808
Bug ID: 57808 Summary: crash when switching linux virtual consoles: XI_BadDevice Product: Wine-staging Version: 9.4 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: forestix@nom.one CC: leslie_alistair@hotmail.com, z.figura12@gmail.com Distribution: ---
Whenever I switch away from the current virtual console on linux (using Control+Alt+F2 etc.) wine processes crash with output like this:
X Error of failed request: XI_BadDevice (invalid Device parameter) Major opcode of failed request: 131 (XInputExtension) Minor opcode of failed request: 3 (X_OpenDevice) Device id in failed request: 0xXXXXXXX Serial number of failed request: 251 Current serial number in output stream: 251 X Error of failed request: XI_BadDevice (invalid Device parameter) Major opcode of failed request: 131 (XInputExtension) Minor opcode of failed request: 3 (X_OpenDevice) Device id in failed request: 0xXXX Serial number of failed request: 222 Current serial number in output stream: 222
(I have replaced the actual device IDs with X characters above.)
As far as I have seen, it affects all applications. Even winecfg crashes on VC switch.
The problem occurs only in wine-staging; not in wine-devel or wine-stable. The problem is present in versions 9.4, 9.8, 10.0, and presumably all versions in between. The problem is not present in version 9.3. Tested with official debian builds from dl.winehq.org.
This looks similar to bug 57123. I opened a new report instead of commenting in that one because in my case, no devices are unplugged or plugged in to cause the crash, and because I have no easy way to test that scenario.
https://bugs.winehq.org/show_bug.cgi?id=57808
--- Comment #1 from Forest forestix@nom.one --- In case it's not obvious from the output I quoted above, I'm using an X11 desktop session.
https://bugs.winehq.org/show_bug.cgi?id=57808
--- Comment #2 from Forest forestix@nom.one --- All testing was done on Debian Bookworm amd64.
https://bugs.winehq.org/show_bug.cgi?id=57808
--- Comment #3 from Forest forestix@nom.one --- Kernel version: 6.12.12-amd64
https://bugs.winehq.org/show_bug.cgi?id=57808
--- Comment #4 from Forest forestix@nom.one --- In case it helps track down the problem, wine-staging 9.3 prints this line 11 times instead of crashing when I switch to another VC and back again:
01e4:fixme:keyboard:X11DRV_ActivateKeyboardLayout 0x4090409, 0000: semi-stub!
https://bugs.winehq.org/show_bug.cgi?id=57808
--- Comment #5 from Forest forestix@nom.one --- Maybe introduced by this commit?
aa8b27c396f926212689d18da91778f5f8821c6d
Author: Alistair Leslie-Hughes leslie_alistair@hotmail.com 2024-03-08 15:46:02 Committer: Alistair Leslie-Hughes leslie_alistair@hotmail.com 2024-03-08 16:53:27 Parent: daf7cb4cb9c8dbdc8850c8189eacd4302e1a76b6 (ntdll-WRITECOPY: Trigger write watches on the "info" pointer in SystemInterruptInformation.) Child: 9a4fd0b9b0c1a7c77887b7769874ce5b2e5c94dd (Updated vkd3d-latest patchset) Branches: master, remotes/origin/master Follows: v9.3 Precedes: v9.4
Updated user32-rawinput-mouse patchset
This update supersede user32-rawinput-mouse-experimental, removing.
https://bugs.winehq.org/show_bug.cgi?id=57808
--- Comment #6 from Forest forestix@nom.one --- I was able to test a build with the user32-rawinput-mouse patchset excluded, and the bug was gone. Since I don't see any other related patches introduced in wine-staging 9.4, I conclude that aa8b27c396f926212689d18da91778f5f8821c6d is indeed the cause of the regression.
https://bugs.winehq.org/show_bug.cgi?id=57808
Forest forestix@nom.one changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
https://bugs.winehq.org/show_bug.cgi?id=57808
Forest forestix@nom.one changed:
What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1| |aa8b27c396f926212689d18da91 | |778f5f8821c6d
https://bugs.winehq.org/show_bug.cgi?id=57808
Forest forestix@nom.one changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|crash when switching linux |new user32-rawinput-mouse |virtual consoles: |patchset causes crash when |XI_BadDevice |switching linux virtual | |consoles: XI_BadDevice
https://bugs.winehq.org/show_bug.cgi?id=57808
Alistair Leslie-Hughes leslie_alistair@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |rbernon@codeweavers.com
http://bugs.winehq.org/show_bug.cgi?id=57808
--- Comment #7 from Forest forestix@nom.one --- Bug 56789 looks like it has the same cause as this one.
http://bugs.winehq.org/show_bug.cgi?id=57808
mittorn@disroot.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |mittorn@disroot.org
--- Comment #8 from mittorn@disroot.org --- (In reply to Forest from comment #6)
I was able to test a build with the user32-rawinput-mouse patchset excluded, and the bug was gone. Since I don't see any other related patches introduced in wine-staging 9.4, I conclude that aa8b27c396f926212689d18da91778f5f8821c6d is indeed the cause of the regression.
I cannot build wine-staging with user32-rawinput-mouse patchset excluded now (wine-staging-10.9), some xaudio code fails due to missing WINBOOL. It seems to introduce conflicts with mingw-runtime
-
WineHQ Bugzilla