Hi, when I use netcfg to setup a laplink connection for device plip1 the file "/etc/sysconfig/network-scripts/ifcfg-plip1" is generated with two bugs: 1) `DEVICE=plip1' line missing. 2) `NETWORK=' line is empty, unless the user presses enter in the NETMASK field. This entry should always be generated (many people just klick on DONE after setting netmask). Example, netcfg generated the following ifcfg-plip1 file for me: BOOTPROTO=none USERCTL=yes ONBOOT=no NETWORK= NETMASK=255.255.255.0 REMIP=128.163.161.100 IPADDR=128.163.161.166 When the plip1 interface is than started, I get many script errors. After I changed the file manually to DEVICE=plip1 BOOTPROTO=none USERCTL=yes ONBOOT=no NETWORK=128.163.161.0 NETMASK=255.255.255.0 REMIP=128.163.161.100 IPADDR=128.163.161.166 ... the connection script works. BTW, Linuxconf seems to have similar problems. I append my simple `laplink' script which you may distribute. It allows to quickly manage direct serial and parallel connections. Cheers, Peter. -----------------------------------------/sbin/laplink: #!/bin/sh # Copyright GPL, P. Englmaier <ppe_at_pa.uky.edu>, 1998 PATH=/bin:/usr/bin:/usr/local/sbin:/sbin:/usr/sbin host=`hostname` SPEED=spd_normal while : ; do if [ $# = 0 ]; then cat - <<EOF Laplink - connect to a remote computer using a serial null modem or parallel laplink cable. Start a serial port connection with laplink [-no] [-hi] [-vhi] <port> <remote> -no for normal speed (38400 baud) (default) -hi for high speed (57600 baud) -vhi for very high speed (115000 baud) <port> serial device port; can be: ttyS0 ttyS1 ttyS2 ttyS3 <remote> remote ip or name (should be defined in /etc/hosts) Start a parallel port ppp connection laplink <device> <remote> <device> parallel device; can be: plip0, plip1, plip2. These correspond to parallel ports lp0, lp1, lp2. <remote> remote ip or name (should be defined in /etc/hosts) Stop a running connection laplink stop <port> or laplink stop <device> EOF exit 0 fi case $1 in -no*) SPEED=spd_normal shift ;; -hi*) SPEED=spd_hi shift ;; -vhi*) SPEED=spd_vhi shift ;; ttyS*) setserial /dev/$1 $SPEED pppd /dev/$1 38400 nocrtscts defaultroute $host:$2 local persist passive & break; ;; plip[0-9]) ifconfig $1 $host pointopoint $2 route add $2 $1 break; ;; stop) case $2 in plip[0-9]) ifconfig $2 down exit 0; ;; ttyS*) line=`ps ax | grep 'pp[p]d /dev/'$2` if [ "$line" = "" ]; then echo laplink: stop: Cannot find pppd process for $2 exit 1 else kill -9 `echo $line |awk '{print $1}'` exit 0 fi ;; *) echo "laplink: bad port or device name">&2 exit 1 ;; esac exit 0 ;; *) echo "laplink: bad parameter(s). Run $0 without arguments for help." >&2 exit 1 ;; esac done exit 0
I have verified this to be true with 5.2. When creating a plip device using netcfg, it does not properly put the DEVICE= line in the /etc/sysconfig/network-scipts/ifcfg-plip<0-2> file. Also you have to hit enter after entering the netmask before the network number is created. This should be either documented in the dialog or the app should be coded to check after hitting save to see if anything was entered into the netmask box instead if just putting a blank NETWORK= into the ifcfg file. Also something different needs to be done in regards to setting up of plip devices because you really do not know whether to use plip0, plip1, or plip2 until the module has been properly inserted and the kernel reports which device it has created. Linuxconf is supposed to address alot of these problems which will make netcfg obsolete.
Michael, have you taken any action on this?
fixed in netcfg-2.20 Will be available in RawHide and 6.0, but not as an erratum for 5.2