PCMCIA only works if cancelled?

[Paul Morgan]

On Fri, 2003-10-24 at 21:11, sean darcy wrote:
> After a number of succesful boots, it  froze at pcmcia. Went to interactive 
> startup. Once up, I tried ./pcmcia start. It froze again.
> 
> Here's the wierd part: I used ^C and it completed successfully.  I had tail 
> -f /var/log/messages open in aother window. As soon as I hit ^C the rest of 
> the messages appeared.
> 
> Oct 24 16:31:35 gateway kernel: Linux Kernel Card Services 3.1.22
> Oct 24 16:31:35 gateway kernel:   options:  [pci] [cardbus] [pm]
> Oct 24 16:31:36 gateway kernel: PCI: Enabling device 00:0a.0 (0000 -> 0002)
> Oct 24 16:31:36 gateway kernel: PCI: Enabling device 00:0a.1 (0000 -> 0002)
> Oct 24 16:31:36 gateway kernel: Yenta IRQ list 0c98, PCI irq9
> Oct 24 16:31:36 gateway kernel: Socket status: 30000010
> Oct 24 16:31:36 gateway kernel: Yenta IRQ list 0c98, PCI irq9
> Oct 24 16:31:36 gateway kernel: Socket status: 30000010
> 
> ******** this is where it hung **************
> 
> Oct 24 16:32:33 gateway cardmgr[6903]: starting, version is 3.1.31
> Oct 24 16:32:33 gateway cardmgr[6903]: watching 2 sockets
> Oct 24 16:32:33 gateway cardmgr[6903]: Card Services release does not match
> Oct 24 16:32:33 gateway kernel: cs: IO port probe 0x0c00-0x0cff: clean.
> Oct 24 16:32:33 gateway kernel: cs: IO port probe 0x0100-0x04ff: excluding 
> 0x170-0x177 0x200-0x207 0x240-0x24f 0x330-0x337 0x370-0x37f 0x388-0x38f 
> 0x398-0x39f 0x4d0-0x4d7
> Oct 24 16:32:33 gateway kernel: cs: IO port probe 0x0a00-0x0aff: clean.
> Oct 24 16:32:33 gateway kernel: cs: memory probe 0xa0000000-0xa0ffffff: 
> clean.
> Oct 24 16:32:33 gateway cardmgr[6903]: socket 0: 3Com 589 Ethernet
> Oct 24 16:32:33 gateway cardmgr[6903]: executing: 'modprobe 3c589_cs'
> Oct 24 16:32:33 gateway kernel: eth0: 3Com 3c589, io 0x300, irq 3, hw_addr 
> 00:60:97:8F:5D:D1
> Oct 24 16:32:33 gateway kernel:   8K FIFO split 5:3 Rx:Tx, auto xcvr
> Oct 24 16:32:33 gateway cardmgr[6903]: executing: './network start eth0'
> Oct 24 16:32:34 gateway cardmgr[6903]: socket 1: Xircom IIps Ethernet
> Oct 24 16:32:34 gateway cardmgr[6903]: executing: 'modprobe xirc2ps_cs'
> Oct 24 16:32:34 gateway kernel: xirc2ps_cs.c 1.31 1998/12/09 19:32:55 
> (dd9jn+kvh)
> Oct 24 16:32:34 gateway kernel: eth1: Xircom: port 0x310, irq 10, hwaddr 
> 00:80:C7:37:6C:9E
> Oct 24 16:32:34 gateway cardmgr[6903]: executing: './network start eth1'

I've experienced this two or three times: under RH9,
severn/rawhide/fedora, and (I think) RH8, and "solved" it the same way
you did (^C during pcmcia start). This happened under both an IBM
Thinkpad and my current Sony Vaio. 

At the time my Vaio was dual-bootable between rawhide and winXP, so I
tried booting into XP (in which the card worked) and then into RH (in
which it continued to hang) several times to prove reproducibility
before trying the ^C. -paul