The following reply was made to PR kern/31902; it has been noted by GNATS.

From: Chuck Silvers <chuq@chuq.com>
To: gnats-bugs@netbsd.org
Cc: kern-bug-people@netbsd.org, gnats-admin@netbsd.org,
	netbsd-bugs@netbsd.org
Subject: Re: kern/31902: kern uvm_sched crash
Date: Mon, 24 Oct 2005 06:10:22 -0700

 hi,
 
 On Mon, Oct 24, 2005 at 12:53:00AM +0000, lamer0@lamer0.com wrote:
 > 	When swap gets low on space the kernel panics.
 ...
 > cpu_Debugger(4,0,0,ab517e98,800) at netbsd:cpu_Debugger+0x4
 > comintr(c0855c00,0,c0510010,c0260030,c0490010) at netbsd:comintr+0xcd
 > Xintr_legacy4() at netbsd:Xintr_legacy4+0xa9
 > --- interrupt ---
 > cpu_switch(c0496640,0,1,0,0) at netbsd:cpu_switch+0x9f
 > ltsleep(c0496480,4,c03f2330,0,0) at netbsd:ltsleep+0x207
 > uvm_scheduler(c0496460,1,c051300c,513000,51b000) at netbsd:uvm_scheduler+0x76
 > setrootfstime(0,0,0,0,0) at netbsd:setrootfstime
 
 this stack trace does not indicate that the machine has paniced,
 but rather that the serial console has received the input (a serial BREAK
 by default) that tells it to enter the debugger.  perhaps the cable
 is flakey and generating spurious BREAKs?  power-cycling whatever is at
 the other end of the cable will often do this as well.
 
 you can change the input sequence that enters the debugger by using (eg.)
 	sysctl -w hw.cnmagic=+++
 
 or you can disable entering the debugger from the console entirely with
 	sysctl -w ddb.fromconsole=0
 
 -Chuck