June 2025 - wine-gitlab

Re: [PATCH 0/1] MR8288: msvcrt: Don't release io memory in msvcrt_free_io.
by Yuxuan Shui (＠yshui) 11 Jun '25

11 Jun '25

Maybe `_fcloseall` needs to be removed too? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8288#note_106243

1 0

[PATCH 0/1] MR8273: msvcrt: Make ioinfo for stdio fds static.
by Yuxuan Shui (＠yshui) 11 Jun '25

11 Jun '25

In kernel32/tests/loader.c, child_process will try to write to stdout after calling LdrShutdownProcess. LdrShutdownProcess calls DLL_PROCESS_DETACH on msvcrt, which calls msvcrt_free_io, which frees the ioinfo blocks. So to prevent use after free in this case, we keep the ioinfos for stdin/out/err static. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8273

3 5

[PATCH v2 0/1] MR8228: mshtml: Fix misuse of IWinInetHttpInfo_QueryInfo.
by Yuxuan Shui (＠yshui) 11 Jun '25

11 Jun '25

IWinInetHttpInfo_QueryInfo returns a multibyte string, not a wide string. We were also wrongly expecting it to have a NUL terminator. * * * this one doesn't feel good. IWinInetHttpInfo_QueryInfo calls HttpInfo_QueryInfo (urlmon), which calls HttpQueryInfoA (wininet), so it returns char*. but inside, HttpQueryInfoA is calling HttpQueryInfoW and does charset conversion. so we are converting the string to multibyte and back. maybe there's an API that returns wchar* i don't know about. -- v2: mshtml: Fix misuse of IWinInetHttpInfo_QueryInfo. https://gitlab.winehq.org/wine/wine/-/merge_requests/8228

4 8

[PATCH 0/1] MR8287: include: Add GetCurrentThreadStackLimits() prototype.
by Nikolay Sivov (＠nsivov) 11 Jun '25

11 Jun '25

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8287

2 1

[PATCH 0/1] MR8270: mf: Prevent double submission of session commands.
by Yuxuan Shui (＠yshui) 11 Jun '25

11 Jun '25

This is not the right fix, and I need some help. Let me explain. `session_submit_command` checks that: 1. `session->commands` is empty, and 2. `SESSION_FLAG_PENDING_COMMAND` is NOT set before submitting the `session_op`. The assumption is, if there are queued commands, or if there is a command currently running, that pending command will submit subsequent commands when it completes. However, in this call chain: ``` #0 0x00006fffe1d27ad7 in create_async_result (object=object@entry=0x0, callback=callback@entry=0x7e62489f1c08, state=state@entry=0x7e61cb5322b0, out=out@entry=0x7e6242bac7a0) at ../dlls/rtworkq/queue.c:1229 #1 0x00006fffe1d222d5 in RtwqCreateAsyncResult (object=object@entry=0x0, callback=callback@entry=0x7e62489f1c08, state=state@entry=0x7e61cb5322b0, out=out@entry=0x7e6242bac7a0) at ../dlls/rtworkq/queue.c:1245 #2 0x00006fffd45e2096 in MFPutWorkItem (queue=queue@entry=1, callback=callback@entry=0x7e62489f1c08, state=state@entry=0x7e61cb5322b0) at ../dlls/mfplat/queue.c:46 #3 0x00006fffe41244ee in session_command_complete (session=0x7e62489f1be0) at ../dlls/mf/session.c:975 #4 0x00006fffe412d00f in session_reset (session=session@entry=0x7e62489f1be0) at ../dlls/mf/session.c:1051 #5 0x00006fffe41299fc in session_handle_source_shutdown (session=session@entry=0x7e62489f1be0) at ../dlls/mf/session.c:2957 #6 0x00006fffe4126e90 in session_events_callback_Invoke (iface=0x7e62489f1c20, result=<optimized out>) at ../dlls/mf/session.c:4445 ``` nowhere was the `SESSION_FLAG_PENDING_COMMAND` flag set. In addition to that, `session_events_callback_Invoke` is not a command. So this could happend: ``` Thread 1 | Thread 2 call session_events_callback_Invoke | | call session_submit_command | <critical section> | list_empty(&session->commands) is true, and PENDING_COMMAND not set | MFPutWorkItem(op) | list_add(&session->commands, op) <- (1) | </critical_section> call session_handle_source_shutdown | <critical_section> | call session_reset | call session_command_complete | list_head(&session->commands) | op is returned because of (1) | MFPutWorkItem(op) again | </critical_section> | ``` The result is this op will be executed twice, and also double freed, because `session_commands_callback_Invoke` releases the `op`. * * * There is even another way an `op` can be submitted twice, `session_handle_start_error` calls `session_reset`, which already calls `session_command_complete`, and later it calls `session_command_complete_with_event` again. I think a missing return? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8270

2 2

[PATCH 0/1] MR8278: dxgi: Do not print a FIXME for DXGI_PRESENT_ALLOW_TEARING.
by Elizabeth Figura (＠zfigura) 11 Jun '25

11 Jun '25

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/8278

3 2

[PATCH v2 0/1] MR8277: d3d11: Implement CreateRasterizerState2().
by Elizabeth Figura (＠zfigura) 11 Jun '25

11 Jun '25

-- v2: d3d11: Implement CreateRasterizerState2(). https://gitlab.winehq.org/wine/wine/-/merge_requests/8277

3 2

[PATCH 0/2] MR8286: win32u: Use a dedicated struct for window destroy entries.
by Rémi Bernon (＠rbernon) 11 Jun '25

11 Jun '25

So wIDmenu and userdata can be removed from the WND struct and read from shared memory, before the window handle gets destroyed. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8286

2 2

[PATCH 0/5] MR8280: ir50_32: Various, part 1/2.
by Elizabeth Figura (＠zfigura) 11 Jun '25

11 Jun '25

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/8280

3 6

[PATCH 0/2] MR8285: kernelbase: Support extended paths in GetFinalPathNameByHandleW
by Michael Stopa (＠stomy) 11 Jun '25

11 Jun '25

All unicode API functions are supposed to support paths longer than `MAX_PATH` if prepended with the extended-path prefix `\\?\`. As of writing this function in particular doesn't. This is causing problems with Unreal Engine shader preprocessing when running under Wine. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/8285

2 2

2025

2024

2023

2022

wine-gitlab June 2025