Alexander Coers wrote:
The timing seems to be really messed up, because with the log output
the bug disappeared.
Since delay() didn't work as expected it becomes clear that the log
output
generated a better delay than the delay() function, which returned
much too
early.
I tried to fix delay(). In the archive
ftp://ftp.netbsd.org/pub/NetBSD/misc/phx/kbdtest6.tar.gz
you will find three wscons kernels with fixed delay():
netbsd_2000: 2000ms handshake
netbsd_200: 200ms handshake
netbsd_85: 85ms handshake