March 2026 - Wine-GitLab - list.winehq.org

[PATCH v14 0/6] MR10419: Draft: ntdll: Fix thread stops inside __wine_syscall_dispatcher
by Hoshino Lina (＠hoshinolina) March 23, 2026

March 23, 2026

When SIGUSR1 is called while running inside `__wine_syscall_dispatcher`, the CPU context is not in a consistent state. To ensure that the captured context is always consistent: * If the SIGUSR1 arrives while saving the context, set a flag and defer handling it until after the context is fully saved. * If the SIGUSR1 arrives while restoring the context, rewind and retry restoring the context from the beginning. The latter relies on context restore being idempotent. The only case where that is not true is when the instrumentation callback is enabled. To address this, first make that path idempotent by saving the user %rip into %r10 in paths that enable the instrumentation callback. This fixes a regression introduced by 0a5f7a710365 which started leaking bogus SP values, but also generally fixes all other register value leaks due to SIGUSR1 racing with the syscall dispatcher. Supersedes: !10232 (FYI, the threadctx test introduced in this MR is stricter and also checks %rip, which will fail with that previous MR). Fixes: 0a5f7a710365 ("ntdll: Switch to the user stack before restoring the %fs register.") \ Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=59333 -- v14: gitlab: Run ntdll tests with win64 architecture. ntdll/tests: Abort wow64 tests early if wow64 is unavailable ntdll/tests: Add a test for interrupted syscall handling https://gitlab.winehq.org/wine/wine/-/merge_requests/10419

2 5

[PATCH v13 0/6] MR10419: Draft: ntdll: Fix thread stops inside __wine_syscall_dispatcher
by Hoshino Lina (＠hoshinolina) March 23, 2026

March 23, 2026

When SIGUSR1 is called while running inside `__wine_syscall_dispatcher`, the CPU context is not in a consistent state. To ensure that the captured context is always consistent: * If the SIGUSR1 arrives while saving the context, set a flag and defer handling it until after the context is fully saved. * If the SIGUSR1 arrives while restoring the context, rewind and retry restoring the context from the beginning. The latter relies on context restore being idempotent. The only case where that is not true is when the instrumentation callback is enabled. To address this, first make that path idempotent by saving the user %rip into %r10 in paths that enable the instrumentation callback. This fixes a regression introduced by 0a5f7a710365 which started leaking bogus SP values, but also generally fixes all other register value leaks due to SIGUSR1 racing with the syscall dispatcher. Supersedes: !10232 (FYI, the threadctx test introduced in this MR is stricter and also checks %rip, which will fail with that previous MR). Fixes: 0a5f7a710365 ("ntdll: Switch to the user stack before restoring the %fs register.") \ Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=59333 -- v13: gitlab: Run ntdll tests with win64 architecture. ntdll/tests: Abort wow64 tests early if wow64 is unavailable ntdll/tests: Add a test for interrupted syscall handling https://gitlab.winehq.org/wine/wine/-/merge_requests/10419

2 6

[PATCH v3 0/1] MR10326: mmdevapi: Avoid copying more memory than reserved in validate_fmt (ASan).
by Bernhard Übelacker March 23, 2026

March 23, 2026

The function `validate_fmt` may receive a pointer which got just reserved memory for the `WAVEFORMATEX`. Unfortunately the assignment `fmt2 = *fmt` always copies the size of `WAVEFORMATEXTENSIBLE`. CC: @giomasce (I currently cannot point to a specific commit, but this may also be related to the recent changes?) <details> <summary>ASan details [gitlab CI](https://gitlab.winehq.org/bernhardu/wine/-/jobs/237050#L2464)</summary> ``` ==qcap_test.exe==728==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x0022c122 at pc 0x785f4342 bp 0x0022bf4c sp 0x0022bb1c READ of size 40 at 0x0022c122 thread T0 #0 0x785f4341 in __asan_memcpy /home/runner/work/llvm-mingw/llvm-mingw/llvm-project/compiler-rt/lib/asan/asan_interceptors_memintrinsics.cpp:65:3 #1 0x778957a5 in validate_fmt /builds/bernhardu/wine/build64/../dlls/mmdevapi/client.c:363:33 #2 0x77896d1e in client_IsFormatSupported /builds/bernhardu/wine/build64/../dlls/mmdevapi/client.c:881:10 #3 0x783e3624 in IAudioClient_IsFormatSupported /builds/bernhardu/wine/build64/include/audioclient.h:406:12 #4 0x783e5cd7 in WINMM_TestFormat /builds/bernhardu/wine/build64/../dlls/winmm/waveform.c:454:10 #5 0x783e5928 in WINMM_GetSupportedFormats /builds/bernhardu/wine/build64/../dlls/winmm/waveform.c:504:14 #6 0x783e553b in WINMM_InitMMDevice /builds/bernhardu/wine/build64/../dlls/winmm/waveform.c:524:35 #7 0x783e4e9b in WINMM_EnumDevices /builds/bernhardu/wine/build64/../dlls/winmm/waveform.c:610:17 #8 0x783d30e8 in WINMM_InitMMDevices /builds/bernhardu/wine/build64/../dlls/winmm/waveform.c:856:10 #9 0x7be4a996 in RtlRunOnceExecuteOnce /builds/bernhardu/wine/build64/../dlls/ntdll/sync.c:462:10 ... Address 0x0022c122 is located in stack of thread T0 at offset 34 in frame #0 0x783e5c0f in WINMM_TestFormat /builds/bernhardu/wine/build64/../dlls/winmm/waveform.c:442 This frame has 2 object(s): [16, 34) 'fmt' (line 443) <== Memory access at offset 34 overflows this variable [80, 84) 'junk' (line 443) HINT: this may be a false positive if your program uses some custom stack unwind mechanism, swapcontext or vfork (longjmp, SEH and C++ exceptions *are* supported) SUMMARY: AddressSanitizer: stack-buffer-overflow /builds/bernhardu/wine/build64/../dlls/mmdevapi/client.c:363:33 in validate_fmt ``` </details> -- v3: mmdevapi: Avoid copying more memory than reserved in validate_fmt (ASan). https://gitlab.winehq.org/wine/wine/-/merge_requests/10326

2 6

[PATCH v8 0/1] MR10386: vbscript: Return proper error codes for compile errors.
by Francis De Brabandere (＠francisdb) March 23, 2026

March 23, 2026

Add error_code field to parse error tests, using negative values for codes that Wine does not yet return correctly (todo_wine), matching the existing pattern for error_char. Return MAKE_VBSERROR with proper error codes instead of E_FAIL for name redefinition, invalid exit statements, and invalid parenthesized call statements. -- v8: vbscript: Return proper error codes for compile errors. https://gitlab.winehq.org/wine/wine/-/merge_requests/10386

2 1

[PATCH v11 0/6] MR10419: ntdll: Fix thread stops inside __wine_syscall_dispatcher
by Hoshino Lina (＠hoshinolina) March 23, 2026

March 23, 2026

When SIGUSR1 is called while running inside `__wine_syscall_dispatcher`, the CPU context is not in a consistent state. To ensure that the captured context is always consistent: * If the SIGUSR1 arrives while saving the context, set a flag and defer handling it until after the context is fully saved. * If the SIGUSR1 arrives while restoring the context, rewind and retry restoring the context from the beginning. The latter relies on context restore being idempotent. The only case where that is not true is when the instrumentation callback is enabled. To address this, first make that path idempotent by saving the user %rip into %r10 in paths that enable the instrumentation callback. This fixes a regression introduced by 0a5f7a710365 which started leaking bogus SP values, but also generally fixes all other register value leaks due to SIGUSR1 racing with the syscall dispatcher. Supersedes: !10232 (FYI, the threadctx test introduced in this MR is stricter and also checks %rip, which will fail with that previous MR). Fixes: 0a5f7a710365 ("ntdll: Switch to the user stack before restoring the %fs register.") \ Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=59333 -- v11: gitlab: Run ntdll tests with win64 architecture. ntdll/tests: Abort wow64 tests early if wow64 is unavailable ntdll/tests: Add a test for interrupted syscall handling https://gitlab.winehq.org/wine/wine/-/merge_requests/10419

2 6

[PATCH v7 0/1] MR10386: Draft: vbscript: Add tests for specific compile error codes.
by Francis De Brabandere (＠francisdb) March 23, 2026

March 23, 2026

Add error code definitions and string resources for compile errors (1002-1057) and todo_wine tests to verify the error codes returned via OnScriptError. Currently Wine returns error code 0 (generic syntax error) for most cases instead of the specific codes Windows returns. -- v7: vbscript: Return proper error codes for compile errors. https://gitlab.winehq.org/wine/wine/-/merge_requests/10386

2 1

[PATCH v6 0/1] MR10386: vbscript: Add tests for specific compile error codes.
by Francis De Brabandere (＠francisdb) March 23, 2026

March 23, 2026

Add error code definitions and string resources for compile errors (1002-1057) and todo_wine tests to verify the error codes returned via OnScriptError. Currently Wine returns error code 0 (generic syntax error) for most cases instead of the specific codes Windows returns. -- v6: vbscript: Add tests for specific compile error codes. https://gitlab.winehq.org/wine/wine/-/merge_requests/10386

2 3

[PATCH v2 0/1] MR10408: mf: Add stub for MFCreateDeviceSource().
by eric pouech (＠epo) March 23, 2026

March 23, 2026

Signed-off-by: Eric Pouech <epouech(a)codeweavers.com> -- v2: mf: Add stub for MFCreateDeviceSource(). https://gitlab.winehq.org/wine/wine/-/merge_requests/10408

3 4

[PATCH v10 0/6] MR10419: ntdll: Fix thread stops inside __wine_syscall_dispatcher
by Hoshino Lina (＠hoshinolina) March 23, 2026

March 23, 2026

When SIGUSR1 is called while running inside `__wine_syscall_dispatcher`, the CPU context is not in a consistent state. To ensure that the captured context is always consistent: * If the SIGUSR1 arrives while saving the context, set a flag and defer handling it until after the context is fully saved. * If the SIGUSR1 arrives while restoring the context, rewind and retry restoring the context from the beginning. The latter relies on context restore being idempotent. The only case where that is not true is when the instrumentation callback is enabled. To address this, first make that path idempotent by saving the user %rip into %r10 in paths that enable the instrumentation callback. This fixes a regression introduced by 0a5f7a710365 which started leaking bogus SP values, but also generally fixes all other register value leaks due to SIGUSR1 racing with the syscall dispatcher. Supersedes: !10232 (FYI, the threadctx test introduced in this MR is stricter and also checks %rip, which will fail with that previous MR). Fixes: 0a5f7a710365 ("ntdll: Switch to the user stack before restoring the %fs register.") \ Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=59333 -- v10: gitlab: Run ntdll tests with win64 architecture. ntdll/tests: Abort wow64 tests early if wow64 is unavailable ntdll/tests: Add a test for interrupted syscall handling https://gitlab.winehq.org/wine/wine/-/merge_requests/10419

2 5

[PATCH 0/1] MR10427: windows.ui.core.textinput: Add CoreTextServicesManager stubbed interface
by Alistair Leslie-Hughes (＠alesliehughes) March 23, 2026

March 23, 2026

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/10427

2 1