Bug 123457 - FC2 Final yenta_socket not loaded so PCMCIA broken
Summary: FC2 Final yenta_socket not loaded so PCMCIA broken
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel   
(Show other bugs)
Version: 2
Hardware: i686
OS: Linux
Target Milestone: ---
Assignee: Dave Jones
QA Contact:
: 123713 125015 126172 (view as bug list)
Depends On:
TreeView+ depends on / blocked
Reported: 2004-05-18 19:40 UTC by Njål Karevoll
Modified: 2015-01-04 22:06 UTC (History)
7 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2005-04-16 04:25:23 UTC
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)
Contents of dmesg during boot (2.95 KB, text/plain)
2004-05-18 19:42 UTC, Njål Karevoll
no flags Details

Description Njål Karevoll 2004-05-18 19:40:27 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040510

Description of problem:
When I try to load the module orinoco_cs it complains about an unknown
symbol in the module or unknown parameter. Further investigation says
'irq' is an unknown parameter.

The card worked perfectly in FC1 using the same module.

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
1. Install FC2
2. Try to bring up Wireless connection

Actual Results:  "orinoco_cs device eth1 does not seem to be present,
initialization" at first, but now I can't even get the module to load.

Expected Results:  Working WiFi-network

Additional info:

System: Dell Inspiron 8200 
Card: Dell Truemobile 1150 (orinoco_cs should work)

When I first booted FC2 after a clean install, I got an error during
startup ( orinoco_cs device eth1 does not seem to be present, delaying
initialization ), so I went into system-config-network and removed it
from the list to try to reconfigure it, but I can't find my card in
the list (should the list of Wireless and Ethernet-cards really be the
same, or is this another bug? I can't find my card or any compliant
card (which uses the same orinoco_cs-module) on the list)..

Running modprobe with -v gives me:

[root@mobile root]# modprobe -v orinoco_cs
/lib/modules/2.6.5-1.358/kernel/drivers/net/wireless/orinoco_cs.ko irq=10
FATAL: Error inserting orinoco_cs
Unknown symbol in module, or unknown parameter (see dmesg)

In /var/log/messages I find "May 18 21:25:30 mobile kernel:
orinoco_cs: Unknown parameter `irq'"

Comment 1 Njål Karevoll 2004-05-18 19:42:14 UTC
Created attachment 100310 [details]
Contents of dmesg during boot

Comment 2 Njål Karevoll 2004-05-18 19:43:51 UTC
This happens if I try to insmod it by hand:

[root@mobile root]# insmod
[root@mobile root]# lsmod | grep orinoco
orinoco_cs              6152  0
orinoco                31500  1 orinoco_cs
hermes                  6528  2 orinoco_cs,orinoco
ds                     12292  7 orinoco_cs
pcmcia_core            46792  3 orinoco_cs,ds,yenta_socket

It looks ok to me, but the card doesn't show when I issue "ifconfig -a"

Comment 3 Penelope Fudd 2004-05-21 06:51:35 UTC
I have this problem too, Dell Inspiron 7500, FC2.
'cardctl status' says open_sock(): No such device

If I type 'modprobe yenta_socket' then 'cardctl status' says good
things about my wireless card.

After that, if I type 'service pcmcia start' then 'ifconfig eth0' says
good things about eth0 being up and having an ip address (after dhcp
has done its work).

I'm hoping for an automatic solution, though.

Comment 4 Penelope Fudd 2004-05-21 07:10:53 UTC
Aha... I found the problem.

In /etc/rc.d/init.d/pcmcia, there is a section that says 'if the
string pcmcia isn't found in /proc/devices, modprobe $PCIC'.

For some strange reason, 'pcmcia' is in /proc/devices before it gets
to this stage.

The right test would probably be to see if 'cardctl status' works or
not.  If not, then modprobe $PCIC.

Comment 5 Penelope Fudd 2004-05-21 07:52:32 UTC
And for the final coup-de-grace, getting the boot process to print no
errors (like orinoco_cs not found, delaying initialization):

Edit /etc/sysconfig/network-scripts/ifcfg-eth0 and set 'ONBOOT=no'.

Eth0 still comes up when pcmcia comes up, a little later.

The line I changed in /etc/rc.d/init.d/pcmcia is 95:
        #if ! grep -q pcmcia /proc/devices ; then
        if ! cardctl status 2>&1 > /dev/null ; then


Comment 6 Alan Cox 2004-05-22 00:08:01 UTC
*** Bug 123713 has been marked as a duplicate of this bug. ***

Comment 7 Clive Long 2004-06-22 09:53:31 UTC
*** Bug 125015 has been marked as a duplicate of this bug. ***

Comment 8 holmes wilson 2004-06-24 03:18:58 UTC
Hey everybody.  I'm a pretty inexperienced user, and I was delighted
to find the fix proposed in comment #3.  However, the automatic
solution given in comment #5 simply didn't work for me.  I edited both
files, and while there was no error message the next time I booted,
the internet didn't work.

running "ifconfig" only showed me the loopback interface.. no eth0

I don't want to have to type in the series of commands above everyday.
 What's the best solution?

Also, since it seems that this problem will prevent quite a lot of
people from using the internet on their laptops (those with pcmcia
network cards), isn't this a rather serious bug?

Thanks for the help!

Comment 9 Clive Long 2004-06-25 09:02:10 UTC
I found yesterday that if I completely delete both the Ethernet and
Hardware definition for the PCMCIA NIC from the "Network
Configuration" screens provided in Fedora / Redhat, then reboot, the
card is not visible in the screens but is visible in ifconfi and the
card SEEMS to work correctly at the moment.

Note I have made the change to the config file described earlier so
that may also be necessary. 

I agree completely this problem needs a permanent fix to be fitted
back into the appropriate component. Who does that , I don't know.


Comment 10 Bill Nottingham 2004-07-12 20:46:06 UTC
*** Bug 126172 has been marked as a duplicate of this bug. ***

Comment 11 Gabriel Somlo 2004-07-13 06:06:51 UTC
I have chkconfig-ed pcmcia to off, so it won't start automatically.
However, for some reason, after the system is completely booted up,
there are a bunch of modules (orinoco_cs, orinoco, hermes, ds, pcmcia)
already loaded into the kernel. This is why cat /proc/devices already
contains a pcmcia line, which then in turn prevents /etc/init.d/pcmcia
start from loading yenta_socket.

Comment 12 Scott Farwell 2004-11-02 18:12:04 UTC
On a Toshiba 4300 Series laptop I have the same problem.  I currently 
let the system boot normally with the errors, then use the following 
commands "service network stop", "service pcmcia stop", "modprobe 
yenta_socket", "service pcmcia start", "service network start" in a 
shell scrip to start my network.  I'm going to put the commands in 
the /etc/rc.d/rc.local file so they will automatically run at start 
time.  This should be a temporary fix until the /etc/init.d/pcmcia 
file is fixed.  Does anybody else have a better suggestion for 
automating this until it's fixed ?  

Comment 13 Clive Long 2004-11-02 22:10:16 UTC
Some time ago I found a fix for this

As I suspected, the problem on some laptop machines is the network is 
attempted to be started BEFORE pcmcia support

No unfortunately I can't find the posting that explained the fix but 
it is somthing along the lines of the following ...

Look in directory


where 5 is the "run-level" for the "environment" you are running in. 
I think 5 is used for X-systems

Now there will be entries named 

Knn and Snn

where nn are integers between 0 and 9

I think the rule is some process at start of the run-level reads the 
rcx.d directory and starts all the applications via these parameter 
files or "scripts" in turn - high numbers first (K99) to low numbers 
last (K00).

The point of all this is I renumbered the pcmcia "script" K24pcmcia 
to "above" the network script: K20network 

The result of this is the PCMCIA support is active before an attempt 
is made to start the network through the PCMCIA network card. Subtle, 

(Maybe the order of start is the other way round and maybe the 
network "script" is called something else - but this is the broad 

If someone could post the correct "rules for start up" or provide a 
link to where this is described - I feel confident (it hasn't failed 
yet) that this is a permanent fix to the intermittent problems of 
PCMCIA network support on Linux laptop

Good luck


Comment 14 Scott Farwell 2004-11-04 20:58:25 UTC
I ran an up2date -u" last night on a freshly loaded FDC2 machine and 
noticed that the /etc/init.d/pcmcia file has been totally rewritten.

Unfortunately, even though the yenta_socket now loads automatically, 
the network startup is unable to see the device pcnet_cs during the 
machines bootup sequence.  I still have to do a "service network 
stop; service pcmcia restart; service network start" to get the 
network working.

During bootup when the network loads I get the error "Bringing up 
interface eth0:  pcnet_cs device eth0 does not seem to be present, 
delaying initialization.    [FAILED]"

When I restart the network using the commands I supplied aboe, this 
is the output I get on the screen :

>>>service network stop (I didn't capture this output)
>>>service pcmcia restart
Shutting down PCMCIA services: ERROR: Module ds is in use by pcnet_cs
ERROR: Module pcmcia_core is in use by yenta_socket,pcnet_cs,ds 
Starting PCMCIA services: service network cardmgr[2558]: open_sock
(socket 2) failed: Bad file descriptor
cardmgr[2558]: watching 2 sockets
ip_tables: (C) 2000-2002 Netfilter core team
>>>service network restart
Shutting down interface eth0:                              [  OK  ]
Shutting down loopback interface:                          [  OK  ]
Setting network parameters:                                [  OK  ]
Bringing up loopback interface:  ip_tables: (C) 2000-2002 Netfilter 
core team    [  OK  ]
Bringing up interface eth0:  ip_tables: (C) 2000-2002 Netfilter core 
team           [  OK  ]

My network now works after the manual restart.  

It looks like the pcmcia file needs some more tinkering, but at least 
the yenta_socket driver now loads automatically !!!

Comment 15 Dave Jones 2005-04-16 04:25:23 UTC
Fedora Core 2 has now reached end of life, and no further updates will be
provided by Red Hat.  The Fedora legacy project will be producing further kernel
updates for security problems only.

If this bug has not been fixed in the latest Fedora Core 2 update kernel, please
try to reproduce it under Fedora Core 3, and reopen if necessary, changing the
product version accordingly.

Thank you.

Note You need to log in before you can comment on or make changes to this bug.