Question about PCMCIA, APM and 3C3FE575CT

DISCUSSION Tom Murphy 3 participants 5 messages Aug 5, 2001 – Sep 17, 2001

lore

Share & tools

lore

Tom Murphy

author

2001-08-05 17:01 UTC #0 lore raw

Hi there,

I'm running Kernel 2.4.6 on a Debian woody system in my
IBM Thinkpad 560X (yes, I know it's ancient. :) ).

Anyhow, I got the card to configure automatically by remove the
pcmcia scripts (I uninstalled the pcmcia package because it wouldn't configure the card properly), I set up /etc/network/interfaces to do DHCP, and use 3c59x as the module. The card happily inits itself with DHCP. If I eject it, hotplug nicely unconfigures the card for me, if I pop it back in, it re-configures it properly.

The only problem I am having is if my laptop goes into suspend
mode, the network shuts off on the card. I then have to eject the card, wait 10 seconds, and plug it back in and it works.. but is there a way around having to do this every time the laptop comes back from an "apm -s"?

Thanks,

Tom

__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/

_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
http://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

David Brownell

2001-08-06 18:36 UTC #1 lore raw

Hmm, seems like a bug of some kind in how something is treating APM operations. If the suspend takes the network down, the resume is supposed to bring it back up. I'd guess that some "resume" path isn't doing something it's expected to do.

What happens if you take the card out while it's suspended, and plug it in later after the system is fully resumed? If that behaves differently, it's clear the resume path has a bug. Do you see log messages reporting that the resume path is bringin up the network interfaces? (You might need to enable debugging in the /sbin/hotplug script.)

Someone recently mentioned a problem on some distro that the networking code was sometimes brought up after the bus code (such as USB or Cardbus PCI). That sequence would cause trouble, since the request to bring the network interface "up" would likely be ignored until later, when the networking code was brought up.

- Dave

----- Original Message -----
From: "Tom Murphy" <[email protected]>
To: <[email protected]>
Sent: Sunday, August 05, 2001 10:01 AM
Subject: Question about PCMCIA, APM and 3C3FE575CT

Quoted 27 lines

Hi there,
I'm running Kernel 2.4.6 on a Debian woody system in my
IBM Thinkpad 560X (yes, I know it's ancient. :) ).
Anyhow, I got the card to configure automatically by remove the
pcmcia scripts (I uninstalled the pcmcia package because it wouldn't configure the card properly), I set up /etc/network/interfaces to do DHCP, and use 3c59x as the module. The card happily inits itself with DHCP. If I eject it, hotplug nicely unconfigures the card for me, if I pop it back in, it re-configures it properly.
The only problem I am having is if my laptop goes into suspend
mode, the network shuts off on the card. I then have to eject the card, wait 10 seconds, and plug it back in and it works.. but is there a way around having to do this every time the laptop comes back from an "apm -s"?
Thanks,
Tom
__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/

Quoted 4 lines

_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
http://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
http://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Tom Murphy

author

2001-09-16 17:56 UTC #2 lore raw

In reply to Tom Murphy

Hello David and List.. sorry it too so long to reply.

Quoted 6 lines · --- David Brownell

Hmm, seems like a bug of some kind in how something is treating APM operations. If the suspend takes the network down, the resume is supposed to bring it back up. I'd guess that some "resume" path isn't doing something it's expected to do.

Quoted 6 lines

What happens if you take the card out while it's suspended, and plug it in later after the system is fully resumed? If that behaves differently, it's clear the resume path has a bug. Do you see log messages reporting that the resume path is bringin up the network interfaces? (You might need to enable debugging in the /sbin/hotplug script.)

If I take the card out while it's suspended and put it in later after the system is fully resumed, it just comes back as it should. The laptop runs DHCP so it seems to spawn additional "pump" processes. However, if I pop it out and back in while the system is up and running, it will kill all the rest of the errant pump processes.

Quoted 6 lines

Someone recently mentioned a problem on some distro that the networking code was sometimes brought up after the bus code (such as USB or Cardbus PCI). That sequence would cause trouble, since the request to bring the network interface "up" would likely be ignored until later, when the networking code was brought up.

I was receiving messages when I suspended and resumed without touching the card:

eth0: command 0x5800 did not complete! Status = 0xffff
eth0: command 0x2804 did not complete! Status = 0xffff
eth0: command 0x3002 did not complete! Status = 0xffff
eth0: command 0x3002 did not complete! Status = 0xffff
eth0: command 0x3002 did not complete! Status = 0xffff

I will enable debugging and let you know what I find.

Thanks,
Tom

__________________________________________________
Terrorist Attacks on U.S. - How can you help?
Donate cash, emergency relief information
http://dailynews.yahoo.com/fc/US/Emergency_Information/

_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Tom Murphy

author

2001-09-16 18:12 UTC #3 lore raw

In reply to Tom Murphy

Quoted 7 lines

Quoted 6 lines · --- David Brownell
Hmm, seems like a bug of some kind in how something is treating APM operations. If the suspend takes the network down, the resume is supposed to bring it back up. I'd guess that some "resume" path isn't doing something it's expected to do.

Quoted 6 lines

Quoted 6 lines
What happens if you take the card out while it's suspended, and plug it in later after the system is fully resumed? If that behaves differently, it's clear the resume path has a bug. Do you see log messages reporting that the resume path is bringin up the network interfaces? (You might need to enable debugging in the /sbin/hotplug script.)

OK here is what appears to be happening..

I tell the laptop to suspend.
cardmgr runs ./network suspend 3c59x
cardmgr received an error: "+3c59x: error fetching interface
information: Device not found"

/etc/hotplug/net.agent doesn't even run.

Should I disable the PCMCIA card services? If I do, I can't insert other non-hotpluggable devices (like modem cards, IDE cards, etc)
and get them to work.

As far as I can tell, net.agent only runs "ifup eth0" and "ifdown eth0" and that's it. It doesn't know about suspend mode or resume mode and won't run at those times.

Any ideas?

Thanks very much!

Tom

__________________________________________________
Terrorist Attacks on U.S. - How can you help?
Donate cash, emergency relief information
http://dailynews.yahoo.com/fc/US/Emergency_Information/

Bill Nottingham

2001-09-17 03:00 UTC #4 lore raw

In reply to Tom Murphy

Quoted 3 lines

As far as I can tell, net.agent only runs "ifup eth0" and "ifdown eth0" and that's it. It doesn't know about suspend mode or resume mode and won't run at those times.

Yes, because the interfaces aren't registered or deregistered at those time unless the interface is specifically brought up or down.

It sounds more like a generic problem with the driver not handling APM suspend/resume correctly (or possibly a BIOS issue.)

Bill