Re: kern/43884: subr_log.c: Mutex error: lockdebug_wantlock: locking against myself

[David Holland <dholland-bugs%netbsd.org@localhost>]

The following reply was made to PR kern/43884; it has been noted by GNATS.

From: David Holland <dholland-bugs%netbsd.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/43884: subr_log.c: Mutex error: lockdebug_wantlock:
 locking against myself
Date: Sun, 10 Oct 2010 02:31:44 +0000

 On Tue, Sep 28, 2010 at 10:00:08AM +0000, Christoph Egger wrote:
  >  >   > breakpoint() at netbsd:breakpoint+0x5
  >  >   > panic() at netbsd:panic+0x28f
  >  >   > lockdebug_more() at netbsd:lockdebug_more
  >  >   > mutex_enter() at netbsd:mutex_enter+0x263
  >  >   > logputchar() at netbsd:logputchar+0x68
  >  >   > putchar() at netbsd:putchar+0x95
  >  >   > kprintf() at netbsd:kprintf+0x1d2
  >  >   > printf() at netbsd:printf+0xb7
  >  >  
  >  >  Where did the printf come from?
  >  
  >  From sys/arch/xen/xen/evtchn.c:evtchn_do_event()
  >  line 276
 
 Yes, and before that? The question is why it's printing while already
 holding the message buffer lock, and the answer appears to be that
 it's printing in an interrupt handler above IPL_VM. That or interrupt
 masking isn't working right, but if that were the case you'd be
 getting a lot more fireworks.
 
 -- 
 David A. Holland
 dholland%netbsd.org@localhost