Udev monitor monitors the whole input subsystem, but not all devices
in the input subsystem have devnodes associated to them.
This MR makes the event processing ignore such devices.
All device handling assumes devices have devnodes, so here we just
simply ignore all devices which do not have one. They are irrelevant.
Previously, udev bus thread aborted when an event for a device without
a devnode was processed:
```
10111.330:0068:0084:trace:hid:process_monitor_event Received action "remove" for udev device (null)
10111.330:0068:0084:warn:hid:bus_main_thread L"UDEV" bus wait returned status 0xc0000005
```
Just plugging in and out a normal mouse was enough cause this.
This was because root input devices (which do not have devnodes) were
handled too and `find_device_from_devnode()` choked on NULL argument.
--
v4: winebus: change debugging class of an error case from WARN to ERR.
winebus: change debugging class of an error case from FIXME to ERR
winebus: group local variable declarations
https://gitlab.winehq.org/wine/wine/-/merge_requests/5385
--
v8: explorer: Restore display settings on process exit.
winex11.drv: Process RRNotify events in xrandr14_get_id.
user32/tests: Test that display settings are restored on process exit.
https://gitlab.winehq.org/wine/wine/-/merge_requests/5060
With some intermediate refactoring to make the code simpler.
I believe the failures that happened previously were coming from how NtGdiDdDDIOpenAdapterFromLuid initialized desc->hAdapter in win32u while winex11 was then relying on it. I missed this detail before and it should be working fine now that the vulkan path is only used to retrieve the physical device.
--
v4: win32u: Move D3DKMT vulkan implementation out of winex11.
winex11: Introduce a new find_adapter_from_handle helper.
winex11: Introduce a new get_vulkan_physical_device helper.
winex11: Initialize D3DKMT vulkan instance only once.
win32u: Open adapters in NtGdiDdDDIEnumAdapters2 outside of the display devices lock.
https://gitlab.winehq.org/wine/wine/-/merge_requests/5306
Also: Add new timeouts to NETCON calls impacted by change
Timeout values are able to be set on all hInternet handles, so they should be able to set in a unified way.
Submitting what I have so far for review.
Known TODO: Tests, verify timeouts are set where NETCON is used, verify ULONG vs DWORD usage rules.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/3518
This intentionally leaves quartz_transform aside as we don't support MF media type conversion for MPEG1 video yet. I have plans to implement it later, but through some media type conversion factorization.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/5558