I've recently been having problems with my Nokia 770 going into a reboot loop. After some research (along with personal past experience as a Linux distro creator and UNIX sys admin), I've become convinced it has to do with the watchdogs not being "tickled" in time.

That is, the watchdog thinks things are hung when in fact they are just taking a long time. I've begun investigating the startup processes and running daemons in order to remove some that perhaps take too long - I've removed clickc and no longer have privoxy run automatically on boot.

I also have samba-common et al (but not samba, just smbfs), as well as rsync and others. But they don't seem to have daemons running (or do they?) sshd (openssh) is another daemon starting up at start.

Is there a way to set the time interval of the watchdog timers? Having used the Linux /dev/watchdog timer, there should be... right?

Or should I take this question to the Developer's Forum?

PS: After removing privoxy (at startup), things seem to be running better...

PPS: I loaded syslog in order to see what was happening - didn't seem to see anything of use there. I might remove it later.

Which version? Mine (3.0.6) or previous?
--
Cezary

Privoxy 3.0.6-2. However, it works fine - it's just that (apparently) with all of the daemons running on startup, the watchdog timer expires before bootup is complete - leading to a reboot loop. Removing privoxy from the startup process provides enough time to activate the watchdog properly.

I just started up Privoxy this afternoon; as expected, everything works fine. I just don't run it on startup.

There's a lifeguard reset watchdog that can be disabled with the flasher tool.

Have you tried disabling this watchdog to see if your device stays up? That might help narrow it down or confirm your suspicion.

No, I haven't - I've been reluctant to use any tool that "flashes" the device (or is the word "brick" better?). In any case, I don't want to lose the watchdog capability. On top of that, there seems to be two underdocumented watchdogs, and I wasn't sure which to manipulate.

I removed as many daemons as I could, and now have no problems. The problem I was experiencing was only on bootup; after it was running, everything was okay - including whatever daemons were running.

Privoxy was previously mentioned - I started it up (separately) and have had no problems with it running normally. It was just the last daemon to be removed. I also removed clinkc and others.

I came upon another much simpler solution - though, again, this requires a bit of typing at the command line. This solution utilizes the fact that the Nokia 770 comes up in run level 2.

Using the xterm (osso-xterm) - or better yet, ssh in using a ssh daemon - and perform these commands as root (without the prompts shown):

Now, if you know what you want to remove, then you can use the rm command. For example, consider the ssh daemon:

This will take care of it. Level 3 and up still have the automatic startup, and the script remains available in /etc/init.d. To start the ssh daemon manually, use:

...and off it goes.

This lets one keep commands available and let the system kick in right away. If you really want everything running (after cleaning out the excessive startups) just do:

...and it should start things up right. Can't say as any unknown items (included on the Nokia) would start up, but there you go.