June 2024 - Wine-gitlab - list.winehq.org

[PATCH v5 0/3] MR4826: Mach vm operations improvements
by Marc-Aurel Zent (＠mzent) 19 Jun '24

19 Jun '24

For all mach vm operations this removes the task suspend and resume, which are not needed. This uses `mach_vm_read_overwrite` to read into a caller-specified buffer, saving the `mach_vm_deallocate` call (bringing all read operations down to 1 syscall from 4). The only alignment restriction on `mach_vm_write` according to the original CMU documentation is that data is > [pointer to page aligned in array of bytes] An array of data to be written. (In practice it also works with arbitrary addresses on macOS, but it probably doesn't hurt to follow the original specifications here). The only other reference that these read/writes should be page-aligned is from the GNU Hurd documentation > The current implementation requires that address, data and data_count all be page-aligned. Otherwise, KERN_INVALID_ARGUMENT is returned. which I assume was the reason why this was originally done (plus it sounds to me like they will fix that in the future and 4fe19777 already broke GNU Hurd support anyways, if that was supposed to be working). Also this includes the missing mach part of 5b1f3b14, which was only applied to the ptrace backend, and together with the `write_process_memory` rework, this gets rid of all fixmes in mach.c -- v5: server: Simplify mach write_process_memory. server: Use mach_vm_read_overwrite in get_selector_entry. https://gitlab.winehq.org/wine/wine/-/merge_requests/4826

2 3

[PATCH v5 0/3] MR5584: mf: Fixes the missing MESessionClosed event in various scenarios.
by Brendan McGrath (＠redmcg) 19 Jun '24

19 Jun '24

Killsquad sends a MESessionClose at various times depending on user input. It always expects a MESessionClosed event and will wait indefinitely if not received. There are currently a number of scenarios where we don't report the MESessionClose event. This MR attempts to fix that. -- v5: mf: Handle an error during Media Session Close. mf: Handle MediaSession Close when state is SESSION_STATE_RESTARTING_SOURCES. mf/tests: Add additional tests for MESessionClosed event. https://gitlab.winehq.org/wine/wine/-/merge_requests/5584

4 7

[PATCH v15 0/4] MR4551: Server-side thread priorities implementation (Part 1)
by Marc-Aurel Zent (＠mzent) 19 Jun '24

19 Jun '24

This is the current proton thread priority implementation by @rbernon rebased for upstream with a few `#ifdef`s added since AFAIK Linux is the only operating system where threads have a unique PID which can be used to set niceness on. ~~I also ran `./tools/make_requests` on https://gitlab.winehq.org/mzent/wine/-/commit/6705d3481be0409f7e971c1d2c7a3… as well and `autoconf` on https://gitlab.winehq.org/mzent/wine/-/commit/d7bafe40c411753662b2ad97148a6… (which does blow up the line count a bit).~~ A few tiny changes ~~(with the ready variable for example)~~ are in anticipation for Part 2, which also adds Mach thread priorities and recalculates thread priorities on process priority change. Since this is a rather large MR, I hope the split here is appropriate ~~(with the second part being slightly smaller)~~, but I think logically it makes the most sense here. -- v15: server: Check wineserver privileges on init with -20 niceness. server: Use setpriority to update thread niceness when safe. ntdll: Set RLIMIT_NICE to its hard limit. server: Introduce new set_thread_priority helper. https://gitlab.winehq.org/wine/wine/-/merge_requests/4551

4 6

[PATCH v3 0/5] MR5844: opengl32: Add default implementations for wglGetPixelFormatAttrib{i,f}vARB.
by Alexandros Frantzis (＠afrantzis) 19 Jun '24

19 Jun '24

This MR extends the `wgl_pixel_format` struct with extra fields required to implement `wglGetPixelFormatAttribivARB` in opengl32.dll. I have added only fields that are relevant to the current needs of the drivers. The default implementation will be used automatically if the driver populates the extra fields, which this MR does for the winex11 driver. The new winex11 implementation tries to remain faithful to the previous behavior in terms of error handling and fallbacks. To achieve this the `wgl_pixel_format` extra fields can be marked as invalid in order for the default implementation to return errors accordingly. Note, however, that error handling around `glXGetFBConfigAttrib` is not very consistent (e.g., the parts of `describe_pixel_format` dealing with PIXELFORMATDESCRIPTOR generally disregard errors). The MR also adds a default implementation for `wglGetPixelFormatAttribfvARB` which piggy-backs on the integer variant of this function, and removes the need for the corresponding code in the drivers. Note that we could potentially remove this function completely from `struct opengl_funcs` (i.e., place it in `manual_win_functions` in `make_opengl`), but it seems the pattern is to maintain a complete list of extension functions (and wined3d also populates the field internally, although it doesn't seem to do anything with it). Finally, the `winex11.drv/opengl.c` file contains a mix of coding styles. I chose to be consistent with the style of the closest context (e.g., the function) I was making changes/additions in. -- v3: winex11: Use default wglGetPixelFormatAttribivARB implementation. winex11: Pass wgl_pixel_format to describe_pixel_format. winex11: Update describe_pixel_format coding style. opengl32: Add default implementation for wglGetPixelFormatAttribfvARB. opengl32: Add default implementation for wglGetPixelFormatAttribivARB. https://gitlab.winehq.org/wine/wine/-/merge_requests/5844

3 10

[PATCH v7 0/2] MR5047: server, ntdll: Use 64 bit for unix thread id.
by Marc-Aurel Zent (＠mzent) 19 Jun '24

19 Jun '24

This is in preparation for the mach thread id rework, which requires 64 bit there (and incidentally also stops truncating the FreeBSD one). There wasn't any direct usage of `__int64` in the sever protocol before, so I added a new typedef and a new `dump_int64()` function (not sure if needed per se though). -- v7: ntdll: Remove unneeded FreeBSD codepath in get_unix_tid(). server, ntdll: Use 64 bit for unix thread id. https://gitlab.winehq.org/wine/wine/-/merge_requests/5047

2 3

[PATCH 0/6] MR5830: dxgi: Fix some d3d12 swapchain issues.
by Conor McCarthy (＠cmccarthy) 19 Jun '24

19 Jun '24

The last commit fixes glitches in Horizon Zero Dawn with vsync on, for reasons stated in the comment there. The wined3d implementation already blocks the client's call to `Present()` when `vkAcquireNextImageKHR()` blocks, so is roughly equivalent to this when frame latency and buffer count are equal. It has a frame latency implementation which unblocks after calling `vkQueuePresentKHR()`, rather than waiting for its execution to complete. I think the d3d12 implementation is supposed to do the latter, though using `KHR_present_wait` may be unnecessary overkill. Waiting on queue execution comes with a performance cost either way, so without a compelling reason for a strict implementation, something similar to wined3d makes sense. Another potential issue is after the call to `ID3D12CommandQueue_Signal()` in `d3d12_swapchain_present()`, the client is free to enqueue more commands when `Present()` returns, so by the time the signal is executed and the swapchain worker blits the image and calls `vkQueuePresentKHR()`, these commands are not necessarily enqueued immediately after the `Signal()`. This may not always be harmless. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5830

4 17

Re: [PATCH v5 0/1] MR5846: quartz: Set biSizeImage for WMV media type.
by Alfred Agrell (＠Alcaro) 19 Jun '24

19 Jun '24

Whoops, pushed something to wrong branch... pushed it back. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5846#note_73717

1 0

[PATCH v4 0/2] MR5222: kernelbase: Allow WriteProcessMemory to succeed on PAGE_EXECUTE_READ.
by Marc-Aurel Zent (＠mzent) 19 Jun '24

19 Jun '24

More details here: https://devblogs.microsoft.com/oldnewthing/20181206-00/?p=100415 However it does not mention that `PAGE_NOACCESS` and `PAGE_READONLY` still result in an error on Windows, which is properly implemented in this MR. Only `WriteProcessMemory` offers this "service", `NtWriteVirtualMemory` will fail on non-writeable and executable regions (and already does so, except for the the mach server backend, which needs https://gitlab.winehq.org/wine/wine/-/merge_requests/4826 to also behave correctly here). -- v4: kernelbase: Flush instruction cache in WriteProcessMemory. kernelbase: Allow WriteProcessMemory to succeed on PAGE_EXECUTE and PAGE_EXECUTE_READ. https://gitlab.winehq.org/wine/wine/-/merge_requests/5222

3 6

[PATCH 0/1] MR5880: msado15: Use the correct version when loading the typelib
by Alistair Leslie-Hughes (＠alesliehughes) 19 Jun '24

19 Jun '24

Since 1,0 was passed it never found the Typelib in the register. Changing to the actual version allows the typelib to be loaded. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5880

2 1

[PATCH v4 0/3] MR4826: Mach vm operations improvements
by Marc-Aurel Zent (＠mzent) 19 Jun '24

19 Jun '24

For all mach vm operations this removes the task suspend and resume, which are not needed. This uses `mach_vm_read_overwrite` to read into a caller-specified buffer, saving the `mach_vm_deallocate` call (bringing all read operations down to 1 syscall from 4). The only alignment restriction on `mach_vm_write` according to the original CMU documentation is that data is > [pointer to page aligned in array of bytes] An array of data to be written. (In practice it also works with arbitrary addresses on macOS, but it probably doesn't hurt to follow the original specifications here). The only other reference that these read/writes should be page-aligned is from the GNU Hurd documentation > The current implementation requires that address, data and data_count all be page-aligned. Otherwise, KERN_INVALID_ARGUMENT is returned. which I assume was the reason why this was originally done (plus it sounds to me like they will fix that in the future and 4fe19777 already broke GNU Hurd support anyways, if that was supposed to be working). Also this includes the missing mach part of 5b1f3b14, which was only applied to the ptrace backend, and together with the `write_process_memory` rework, this gets rid of all fixmes in mach.c -- v4: server: Simplify mach write_process_memory. server: Use mach_vm_read_overwrite in get_selector_entry. server: Use mach_vm_read_overwrite in read_process_memory. https://gitlab.winehq.org/wine/wine/-/merge_requests/4826

2 4