Subject: Re: panic: old sleep...
To: Chris G Demetriou <Chris_G_Demetriou@LAGAVULIN.PDL.CS.CMU.EDU>
From: Darren Reed <firstname.lastname@example.org>
Date: 03/03/1995 11:40:40
In some email I received from Chris G Demetriou, they wrote:
> > p.s. has anyone thought that perhaps panic messages should go through log()
> > rather than printf() in the kernel ?
> umm, think about this for a second... so, you write it it with log(),
> which puts it to the message buffer, but _NOT_ the console...
> (printf() sends to both the message buffer _and_ the console).
> so, the next time syslog gets run, it'll print out the message...
> but since you've _panic'd_ that's _NOT_ going to happen until after a
> reboot, at least... On boxes which clean memory when rebooting
> (e.g. most i386en), you'll never see the panic message. "not a feature."
Could it do both, perhaps ? It (may) mean having duplicates on the console,
but being able to get at them after reboot (ie in /var/log/messages) would
be nice. Or would this require a bit of work in how the system shuts down
and reboots after panic'ing ?