Re: deadlock in -current

On Sun, 25 Sep 2005, Kris Kennaway wrote:

> On Fri, Sep 23, 2005 at 11:43:22PM -0400, Kris Kennaway wrote:
>> Since updating my (UP i386) machines yesterday they're all hanging
>> after a few hours of load with all processes in the following states:
>>
>> db> ps
>>   pid   proc     uid  ppid  pgrp  flag   stat  wmesg    wchan  cmd
>> 28397 c5267a3c    0 28396 28397 0000110 [SLPQ vmmaps 0xc1027178][SLP] cron
>> 28396 c5267c48    0   386   386 0000000 [SLPQ ppwait 0xc5267c48][SLP] cron
>
> [...]
>
> This is still broken.  Someone must have broken this in the past few 
> days, but I don't have time right now to do a binary search.  Please, 
> everyone review your recent commits to see if you could be the 
> originator.

You could try backing this out locally:

   1.127     +16 -3     src/sys/fs/fifofs/fifo_vnops.c

But I'm pretty sure it's not the source of the problem you've been seeing.

If you were using process accounting, it could be the recent kern_acct.c 
change to convert NULL's to curthread's (even more unlikely), but I'm 
pretty sure you're not running process accounting.

Robert N M Watson