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

Re: buildd macs, was Re: [buildd] Etch?




On Mon, 14 Aug 2006, Wouter Verhelst wrote:

> On Mon, Aug 14, 2006 at 06:18:14PM +0200, Michael Schmitz wrote:
> > > > Not sure the RTC ever worked, but the serial drivers used to.
[snip]
> > >
> > > It appears that there is no way to collect a crash dump on this arch :-(
> > 
> > Not at this early stage in booting. You'd really need to delay ADB 
> > init until after syslogd starting :-) Dumping state information from 
> > the ADB in handler is what I did to figure out what happened. Have fun
> > :-)
> 
> Isn't it possible to get the serial port to work before ADB and then to 
> use a serial console or so? You could at least catch a panic that way, 
> plus whatever other debugging printk() output you want to send out.
> 
> Or am I missing some crucial bit here?

You are right that serial output works from very early in the boot, even 
before the framebuffer initialisation. So, if you pass "debug=ser" (or 
sera/serb) you can get all the printk's etc which is very useful. 
(However, /dev/ttyS* are broken, so there's no serial console 
unfortunately.)

Anyway, I have the logs, what I was after was a kernel core dump, like 
kdump, netdump or lkcd provides on other architectures, so as to do a 
postmortem in gdb.

-f



Reply to: