(qa0401) If I suspend my laptop (beta hw list #12) and resume, my pcmcia network interface eth0 is gone. Card will work again if I say "service pcmcia restart" on the command line. If I use "cardclt suspend/resume" from the command line everything works fine. Also had no problems under RH7.0 (without PCMCIARESTART="yes"). I tried to use PCMCIARESTART="yes" PCMCIABIOSBUG="yes" PCMCIAWAIT="no" in /etc/sysconfig/apmd but it makes no difference. For the records: ------------------------------------ [root@soliton /root]# cat /etc/sysconfig/pcmcia PCMCIA=yes #PCIC=i82365 PCIC=yenta_socket PCIC_OPTS= CORE_OPTS= CARDMGR_OPTS="-q -f" [root@soliton /root]# cardctl ident Socket 0: product info: "TAMARACK", "Ethernet", "A", "004743118001" function: 6 (network) --------------------------------------
Same on my Toshiba Libretto 50CT with all three network cards I tried -- two Linksys ones and a Cisco Aeronet card. With the linksys cards, popping them out and putting them back in again causes them to be recognized. (With the cisco one, that causes a kernel panic -- that's probably a separate issue, though -- I need to investigate that further.)
Still present in 7.1. It even suffices to say "ifup eth0" to bring the if up. However, the NET_RESTART="yes" option doesn't seem to help. Pretty annoying!
Fixed in -31
Nope. Not fixed. In fact, now it is even worse since I cannot use ifup anymore [root@soliton gerald]# ifup eth0 SIOCSIFFLAGS: No such device SIOCADDRT: Network is down SIOCADDRT: Network is unreachable and I have to do a full "service pcmcia restart".
Please ignor the above "ifup" problem, it works now. I just did some further investigations and located the problem: I once suggested a clean solution for the "failed" message when a pcmcia network interface is started before pcmcia is started (which is the case during boot). However, redhat turned my patch down since it is "much better" to set ONBOOT="no" for such devices. But if you do this, the command "service network start" in the apm script has no effect on this interface and it is not brought up. So much about "quick and easy" solutions. So if I set ONBOOT="yes" everything is fine again.
See bug #22979
Unfortunately the [root@soliton gerald]# ifup eth0 SIOCSIFFLAGS: No such device SIOCADDRT: Network is down SIOCADDRT: Network is unreachable problem is still here. However, it only happens sometimes. If I do a "cardctl reset 0" it does not help. But if I say "cardctl eject 0; cardctl insert 0" it works. Not sure if this is an apmd or pcmcia problem.
I just did some further investigations and observed that the card is "ejected" correctly by the apm script upon suspend. However, upon resume the card is already active when the apm scripts "inserts" the card again. As a temorary work around I use --- apmscript.orig Sat May 12 19:19:09 2001 +++ apmscript Sat Jun 16 09:13:34 2001 @@ -217,6 +217,7 @@ if [ "$PCMCIARESTART" = "yes" ] ; then if [ "$PCMCIABIOSBUG" = "yes" ]; then + /sbin/cardctl eject /sbin/cardctl insert else /sbin/cardctl resume
Still broken in beta3
Please check if this is still a problem in 3.0.2-6 - my PCMCIA cards (and chipsets) are apparnetly much more tolerant to being inserted a couple of times in a row than yours.
Seems to work in 7.2. Sometimes the interface does not come up after wakeup from sleep