[Bug 47873] New: kernel32:comm fails randomly

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47873

Bug ID: 47873
           Summary: kernel32:comm fails randomly
           Product: Wine
           Version: 4.17
          Hardware: x86
                OS: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: kernel32
          Assignee: [e-mail]
          Reporter: [e-mail]

kernel32:comm fails randomly on multiple Windows versions:

comm.c:708: Found Com port COM1. Connected devices may disturb results
comm.c:859: Test failed: WaitCommEvent failed with a timeout
comm.c:870: recovering after WAIT_TIMEOUT...
comm.c:880: Test failed: WaitCommEvent error 0
comm.c:881: Test failed: WaitCommEvent: expected EV_TXEMPTY, got 0
comm.c:885: WaitCommEvent for EV_TXEMPTY took 1500 ms (timeout 1500)
comm.c:886: Test failed: WaitCommEvent used 1500 ms for waiting

https://test.winehq.org/data/tests/kernel32:comm.html

These failures are random and have been seen on the following TestBot VM4
configurations:
wxppro
w2003std
w7pro64-32
w1064v1507-32
w1064v1607-32
w1064v1607-64
w1064v1709-64
w1064v1607-64
w1064v1809-32
w1064v1809-32
w1064v1809-64
w1064v1809-2scr-32
w1064v1809-fr-64
w1064v1809-ja-32
w1064v1809-ja-64

So it does not depend on the Windows version, bitness of the test, or locale.
However it only happens on QEmu VMs: it never happened on the fg-win7u64 VMware
VMs or on the real hardware machines, cw-rx460 and cw2-gtx560.

Yet the VMware VMs have two serial COM ports, and, according to dmesg | grep
tty, the real hardware machines do have a serial port too.

QEmu bug?