RE: IRQ5 and PPP network activity
On Fri, 8 Jan 1999, Olson, Greger J. <gjolson@bpa.gov> wrote:
> To: 'Debian-68k' <debian-68k@lists.debian.org>
>
> I had also mentioned getting "unexpected IRQ" messages.
> These are not as benign as I had imagined.
> The IRQ 3s on booting don't seem to hurt anything
That's been my experience too. A dozen or so IRQ 3's in rapid succession
early in the boot process, and then I never see them again. Harmless, I
think.
> however, once I had PPP up and running I was getting IRQ 5s. After a
> few of those My Mac would reboot, with the bootup chord & no warning.
> This happened both during FTP & using Lynx (I don't have X up yet).
> I using a 2.0.33 kernel dated 12/8/98. Any suggested alternative?
This has been the real killer for me as well, although I'm not sure
whether the couple IRQ 5's I get now and then are the problem or merely
coincidental. IRQ 5 is definitely related to network activity though, as
they don't start til I open a PPP connection and cease when I close it.
Curiouser and curiouser, telnet seems to work so far without a hitch.
(case in point, I'm running pine just fine remotely at the moment)
Only FTP (haven't tried lynx yet) brings on the unclean reboot, and only
when I actually start to transfer a file; login and maneuvering around
the remote site works fine. As such, it seems to me that something in
binary transfers that's triggering the behavior.
Some special character not being escaped properly? I say that because
something keeps gnawing at me that it has to do with the fact the
control-command-power reboot combo. Curious: why in the default keymap
is control-SHIFT-power (I assume character 127 is the power button)
assigned as reboot then?
I currently use the suggested "escape FF" option, and have tried "asyncmap"
set to 0 (default in /etc/ppp/options), 20A0000 (from /usr/doc/ppp/*),
200A0000 (typo of previous item? from pppd'd man page), with the same
results. Using "default-asyncmap" to escape all control characters in
both directions (just to be safe) a few minutess ago caused the spontaneous
reboot mere moments after the PPP connection was established. Weird.
I still need to go back and test again with daemon.notice or .info
dumped. So far pppd's debug and kdebug options aren't showing anything
useful in the logs just prior to the restart... logging just suddenly
halts mid-line, then a restart message, with no indication of a panic
(would that show up in the logs?) or anything else unusual.
To answer your question on the LC475 having the same problem, Michael,
my Quadra605 (effectively the same machine) is getting the exact same
behavior and it doesn't have any unusual hardware, at least not add-on.
Doesn't rule out something inherently weird in the model itself though.
I noticed a patch someone put out on the linux-ppc the other day to
ignore certain unexpected signals. Might that have use here, barring
discovery of where those signals are coming from?
On the X front...
I got it working, with surprisingly little trouble. Nice surprise.
Will have to take a look at the colormap stuff in with Penguin-16. Ummm...
where should I start looking in the X source regarding that? Glanced
through linux/fb.h earlier today. At least some of it made sense. =)
-jope
Reply to: