Bug 20849 – baselayout-1.8.6.7: cardmgr/runscript.sh errors on init 6

• Product: Gentoo Linux
• Status: RESOLVED DUPLICATE

[Description Jason Waldhelm 2003-05-12 10:22:12 UTC]

This issue occurs only when I have my Cisco Aironet card in one of my pcmcia slots.  
The card is brought up properly on boot, with no errors whatsoever.  On init 6 (or 
anything involving bringing down services), cardmgr logs the following: 
 
May 11 20:39:33 frozen cardmgr[852]: watching 2 sockets 
May 11 20:39:33 frozen cardmgr[852]: starting, version is 3.2.4 
May 11 20:39:33 frozen cardmgr[852]: socket 1: 350 Series Wireless LAN Adapter 
May 11 20:39:34 frozen cardmgr[852]: executing: './network start eth1' 
May 11 20:39:34 frozen cardmgr[852]: + /sbin/runscript.sh: line 27: /softlevel: No 
such file or directory 
May 11 20:39:34 frozen cardmgr[852]: + /sbin/runscript.sh: line 384: wrap_rcscript: 
command not found 
May 11 20:39:34 frozen cardmgr[852]: + /sbin/runscript.sh: line 385: eerror: 
command not found 
May 11 20:39:34 frozen cardmgr[852]: start cmd exited with status 1 
 
The remaining shutdown follows through cleanly, but I don't think it cleanly releases the 
dhcp address (dhcpcd shutting down is never seen).  Further, this, for hopefully obvious 
reasons, never occurs when the pcmcia slots are empty.  I have a feeling this may 
happen with any card in the slots, but cannot confirm this as I only have the Aironet. 

Reproducible: Always
Steps to Reproduce:
1. emerge -u baselayout 
2. shutdown with Cisco Aironet in slot 
3. 



Portage 2.0.47-r10 (default-x86-1.4, gcc-3.2.2, glibc-2.3.2-r1) 
================================================================= 
System uname: 2.4.20-pfeifer-r1_pre11 i686 Intel(R) Pentium(R) 4 Mobile CPU 
2.00GHz 
GENTOO_MIRRORS="http://gentoo.oregonstate.edu/ 
http://www.ibiblio.org/pub/Linux/distributions/gentoo" 
CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config 
/usr/kde/3/share/config /usr/X11R6/lib/X11/xkb /usr/kde/3.1/share/config 
/usr/share/config" 
CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" 
PORTDIR="/usr/portage" 
DISTDIR="/images/portage/distfiles" 
PKGDIR="/usr/portage/packages" 
PORTAGE_TMPDIR="/var/tmp" 
PORTDIR_OVERLAY="/images/portage/overlay" 
USE="x86 oss 3dnow apm avi encode gif jpeg libg++ mikmod mmx mpeg ncurses nls 
pdflib png quicktime spell truetype xml2 xmms xv zlib gdbm berkdb slang readline arts 
svga tcltk java guile sdl gpm tcpd pam libwww perl python esd imlib oggvorbis qt motif 
opengl mozilla cdr scanner X gtk cups alsa crypt acpi apache2 kde postgres ssl 
-gnome" 
COMPILER="gcc3" 
CHOST="i686-pc-linux-gnu" 
CFLAGS="-march=pentium3 -O3 -pipe -fomit-frame-pointer" 
CXXFLAGS="-march=pentium3 -O3 -pipe -fomit-frame-pointer" 
ACCEPT_KEYWORDS="x86 ~x86" 
MAKEOPTS="-j2" 
AUTOCLEAN="yes" 
SYNC="rsync://rsync7.us.gentoo.org/gentoo-portage" 
FEATURES="sandbox ccache"

[Comment 1 Martin Schlemmer (RETIRED) 2003-05-12 17:40:56 UTC]

Do you also get the errors with:

  # /etc/init.d/net.eth1 stop|start

being run manually ?

[Comment 2 Jason Waldhelm 2003-05-12 20:42:46 UTC]

output of /etc/init.d/net.eth1 stop|start: 
 
root@frozen /var/log 
 # /etc/init.d/net.eth1 stop 
 * Bringing eth1 down... 
 *   Waiting for dhcpcd to shutdown.....done       [ ok ] 
root@frozen /var/log 
# /etc/init.d/net.eth1 start 
 * Bringing eth1 up...     [ ok ] 
 
what escaped my attention before was it seems that the services start a bit out of order.  
When a link isn't detected on eth0 (the on board NIC), cupsd, netmount and ntpd try to 
start before pcmcia (I can't seem to find the exact logging so I'll "paraphrase": 
 
Network link on Eth0 not detected... network not initialized... 
 
Error: problem starting needed services 
	"cupsd" was not started. 
Error: problem starting needed services 
	"netmount" was not started. 
Starting syslogd, klogd... [ ok ] 
Error: problem starting needed services 
	"ntpd" was not started. 
Starting pcmcia... [ ok ] 
cardmgr[874]: watching 2 sockets... 
 
then everything else continues as normal. 
 
The link detection is a bit of a hack to net.eth0: 
    else 
        if [ ! -n "$(which mii-diag 2> /dev/null)" -o -n "$(mii-diag --status ${IFACE} | grep "link 
beat" | grep "OK")" ] 
        then 
            /sbin/dhcpcd ${dhcp_args} ${IFACE} >/dev/null || { 
            retval=$? 
            eend ${retval} "dhcpcd failed to bring up ${IFACE}" 
            return ${retval} 
            } 
        else 
            eend 2 "No ethernet link on ${IFACE} detected - Network Not Initialized" 
            return 2 
        fi 
 
    fi 
 

[Comment 3 Martin Schlemmer (RETIRED) 2003-05-13 13:35:24 UTC]

Hmm, well in theory if it do not get it right with net.eth0 failing, it should
also not with it starting, as pcmcia *should* have been started already ...

[Comment 4 Jörg Gollnick 2003-05-14 05:27:17 UTC]

Please have a look on 20851, the patches should fix your problem.
cardmgr is calling ./network start eth1

[Comment 5 Martin Schlemmer (RETIRED) 2003-05-14 14:43:09 UTC]

Same issue.

*** This bug has been marked as a duplicate of 20851 ***

[Comment 6 Martin Schlemmer (RETIRED) 2003-05-14 16:08:50 UTC]

I forgot to add .. the issues with pcmcia not starting properly was a
bug in the depend code, fixed on cvs:

-----------------------------------
13 May 2003; Martin Schlemmer <[e-mail]>:

    For the network functions, the dependency code tried to create a funcion
    with a '.' in the name, causing the network rc-scripts to not have any
    dependencies.  This should fix bug #20849 's order issue, thanks to
    YAMAKURA Makoto <[e-mail]> for noticing.