Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 29 Nov 2006 12:47:48 -0800
From:      Peter Grehan <grehan@freebsd.org>
To:        Dario Freni <saturnero@freesbie.org>
Cc:        ppc@freebsd.org
Subject:   Re: Instant panic on a recent current
Message-ID:  <456DF1F4.6060404@freebsd.org>
In-Reply-To: <20061122095458.GA86970@builder.freesbie.org>
References:  <20061122095458.GA86970@builder.freesbie.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Hi Dario,

> I just upgraded to a -CURRENT from yesterday (previously it was the
> 200608 snapshot) and I've got a panic right after the kernel is
> booted:
> 
> panic: mutex (null) not owned at /usr/src/sys/vm/vm_page.c:368
> [thread pid 0 tid 0 ]
> Stopped at  0x29fb0c:	lwz	r11, r1, 0x0

  Back from holidays and yes, instant panic :( I have a few hacks to 
try, though maybe it's time for a pmap rewrite :(

> Then the ddb prompt, which is unusable with the usb keyboard.

  And that brings up an interesting point. Console input for devices 
that have large s/w stacks such as USB is difficult this early in the 
boot sequence. However, accessing the keyboard through openfirmware (or 
a PROM i/f such as EFI on other systems) is possible. An experiment I 
might try is to have an optional 'fallback' for syscons's cngetc, where 
a PROM call could be done until keyboard devices are probed. This would 
allow entry to the debugger at boot, which is a worthwhile feature.

later,

Peter.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?456DF1F4.6060404>