[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: pmud 0.7-1



> > do you mean 'ifconfig eth0 up' fails in this special case or everytime?
> > do you got a line like: 'alias eth0 tulip'?
> 
> I mean that it fails on wakeup only. And the module (mine is de4x5.o) loads at startup,
> although I didn't add that line (I just added 'de4x5' in /etc/modules).
> 
> Aha, I tried again, and it's not really my eth0 that's not coming back up, but my routing
> table...

That's what I suspected. Will be fixed (for the trivial case of just the
interface route) with 2.4 kernels. 
 
> > I'm using the 'G3'-part of the pwrctl script, 'cause nap mode + 'hdparm -d0' is not
> > usable regarding dialup connections (+ random 'lost interrupts'); pb3400 too.
> 
> heh, I remember you telling me to try that... Never did get around to it though ;)
> Might as well try now :)
> 
> Works a little better.. of course, the G3 function doesn't deactivate the network interfaces
> on sleep or minimum power, so...

With a dialup connection, shutting down eth0 doesn't help you a bit. You'd
have to shut down ppp0 instead. And there's no need for that unless you
have serial DMA in your 3400 (don't think so). The only reason nap mode
with active eth0 is bad on the 3400 is the lack of cache consistency in
nap mode. 

The hdparm -d0 is another issue - what exactly is the problem there? The
Lombard wakes up from sleep fine with -d0 set, and setting -d1 later
doesn't cause any lost interrupts either. Powering down the drive with -Y
will cause lost interrupts and a drive reset after which everything's fine
again. 

> I just start pmud with /etc/init.d/pmud start (or it does it on startup...). Never saw
> anything about options.

See /etc/default/power (in general, see /usr/share/doc/pmud for clues). 

	Michael



Reply to: