suspend/resume problems: NetworkManager and iwl3945/ipw3945

[Paul Johnson]

On 6/3/07, George Avrunin <avrunin at math.umass.edu> wrote:
> As I mentioned in an earlier message, I'm running Fedora 7 (x86_64) on a
> Lenovo X60 with an Intel 3945 wireless card.
>
> I started with the iwl3945 module that the installation set up, running
> NetworkManager. The wireless network seems to work fine (with WPA2-PSK), at
> least in the one location I've tried so far, but suspend (and hibernation)
> don't.  I found various suggestions on the web (including the quirks, which
> don't seem to be necessary for this machine) and it seems clear that the
> iwl3945 module must be removed before suspending.  But that seems to be
> impossible with NetworkManager running.  I've tried various ways of
> removing the module (manually, in the SUSPEND_MODULES line
> of /usr/lib/pm-utils/defaults, etc.), and it doesn't seem to get removed,
> causing the suspend to hang.
>

I don't know for sure if you are right in thinking that the network is
the only trouble here.

I have a Dell D820 with the same wireless device.  I've run the FC7
kernel as well as the David Jones update testing version
kernel-2.6.21-1.3232.fc7 from here:
http://people.redhat.com/davej/kernels/Fedora/fc7/RPMS.kernel/i386/

 I can't get the laptop to resume, but I can get it to suspend to RAM.
 I mean, it turns off and the power light slowly flashes.  But on
powerup, the system shows some life and dies.  It is unresponsive to
keyboard and one cannot log in from remote sources, indicating either
the network never wakes up or that the whole system is dead.

Is your system the same?

One thing I learned while working on this is that the pm-utils package
is lacking the "vbetool" program in F7, and there is an updated
version you can get, but not from the F7 updates repository.  The one
I'm using is: pm-utils-0.99.3-8.fc7.i386.rpm from here
http://koji.fedoraproject.org/koji/buildinfo?buildID=8378
(its been reported
https://bugzilhttps://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=246211).

Another thing I have learned is that there are new hal rules available
all the time, and you can see if they match your system.  The hal
rules are in /usr/share/hal/fdi/ and so forth, and they have specific
information on your drivers.

The RPM for ipw3945 from freshrpms drops a file into /etc/pm/hooks
that is supposed to remove ipw3945d and the module, but I've heard in
the net that there is a flaw in pm-suspend so that the hooks are
ignored.  Perhaps that traces back to the trouble you see.

In my system, I see no trouble on suspend anymore, just no evidence of resume.

# cat /var/log/pm-suspend.log
Mon Jul  2 04:18:22 CDT 2007: running suspend hooks.
===== Mon Jul  2 04:18:22 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/00clear =====
===== Mon Jul  2 04:18:22 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/01grub =====
===== Mon Jul  2 04:18:22 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/05led =====
===== Mon Jul  2 04:18:22 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/10NetworkManager =====
===== Mon Jul  2 04:18:23 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/20video =====
kernel.acpi_video_flags = 0
===== Mon Jul  2 04:18:23 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/49bluetooth =====
===== Mon Jul  2 04:18:23 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/50modules =====
===== Mon Jul  2 04:18:23 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/55battery =====
===== Mon Jul  2 04:18:23 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/60sysfont =====
===== Mon Jul  2 04:18:23 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/65alsa =====
===== Mon Jul  2 04:18:23 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/90clock =====
===== Mon Jul  2 04:18:24 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/94cpufreq =====
===== Mon Jul  2 04:18:24 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/95led =====
===== Mon Jul  2 04:18:24 CDT 2007: running hook:
/usr/lib/pm-utils/sleep.d/99video =====
Mon Jul  2 04:18:24 CDT 2007: done running suspend hooks.

-- 
Paul E. Johnson
Professor, Political Science
1541 Lilac Lane, Room 504
University of Kansas