No problems known or showed up and there were further bugfixes, closing this 
issue therefore.

regards,
-mika-

Fixed in grml-autoconfig (0.9.12) and grml-etc (1.2.2) - both available in 
testing.

Setting status to testing until we tested all possible use-cases.

thanks && regards,
-mika-

patch against grml-etc

patch against grml-autoconfig

I would suggest patching the original inittab and autoconfig.functions to only
enable the serial console on the device that was actually given on the command
line. That way, there would only be a getty on ttyS0 if "serial" were given on
the grml command line, and on ttySx if "serial console=ttySx,..." were given.

Actually, the getty on the "other" serial port may cause unintended behavior

The (only) line in inittab mentioning the serial console would be:

#grmlserial#T0:23:respawn:/sbin/getty -L #grmlserialdevice# -l /usr/bin/zsh-
login 115200,57600,38400,19200,9600,4800,2400,1200 vt100

and the autoconfig function would do something along the lines of

sed -i '/^#grmlserial#/{s/#grmlserialdevice#/$TTYDEVICE/;s/^#grmlserial#//;}'
/etc/inittab

Please indicate if such a patch would be acceptable to you.

Greetings
Marc

Hi,

the IPMI card that I currently fight with offers a redirected virtual serial
port as ttyS2. The natural serial console=ttyS2,115200n8 does give me the
console on ttyS2, but doesn't start a getty on that device, which leaves me with
a useless console.

Please consider generating the appropriate getty config that fits the
console=ttys number.

Greetings
Marc