Re: Stuck kernel while cleaning up the object tree

On Sat, 4 Apr 2009 11:51:17 +0200 Thierry Herbelot wrote:
> Le Saturday 04 April 2009, Thierry Herbelot a écrit :
> > Hello,
> >
> > On recent -current machines, I have seen a common pattern, with the machine
> > being frozen (still responsive to pings, though) in the initial phases of
> > the buildworld procedure :
> >
> > example freeze :
> > --------------------------------------------------------------
> >
> > >>> stage 2.1: cleaning up the object tree
> >
> > --------------------------------------------------------------
> > cd /usr/src; MAKEOBJDIRPREFIX=/usr/obj  MACHINE_ARCH=i386  MACHINE=i386
> > CPUTYPE=  GROFF_BIN_PATH=/usr/obj/usr/src/tmp/legacy/usr/bin
> > GROFF_FONT_PATH=/usr/obj/usr/src/tmp/legacy/usr/share/groff_font
> > GROFF_TMAC_PATH=/usr/obj/usr/src/tmp/legacy/usr/share/tmac
> > _SHLIBDIRPREFIX=/usr/obj/usr/src/tmp  VERSION="FreeBSD 8.0-CURRENT i386
> > 800074"  INSTALL="sh /usr/src/tools/install.sh"
> > PATH=/usr/obj/usr/src/tmp/legacy/usr/sbin:/usr/obj/usr/src/tmp/legacy/usr/b
> >in:/usr/obj/usr/src/tmp/legacy/usr/games:/usr/obj/usr/src/tmp/usr/sbin:/usr/
> >obj/usr/src/tmp/usr/bin:/usr/obj/usr/src/tmp/usr/games:/sbin:/bin:/usr/sbin:
> >/usr/bin NO_CTF=1 make -f Makefile.inc1 DESTDIR=/usr/obj/usr/src/tmp
> > par-cleandir ===> share/info (cleandir)
> > ===> lib (cleandir)
> > ===> lib/csu/i386-elf (cleandir)
> > [type ^T in the console]
> > load: 0.00  cmd: sh 24587 [*Name Cache] 0.01u 0.00s 0% 1584k
> >
> > The other machines also froze while "cleaning up the object tree".
> >
> > The machines are configured with serial consoles : I have no kernel stack
> > backtrace to aid in pinpointing the cause of this freeze.
> >
> > 	Cheers
> >
> > 	TfH

> With a bit more investigation :

> on a separate ssh session, top is still live and shows processes stuck as :
> 24523 root             1  76    0  1888K   764K *Name   1   0:00  0.00% make

> on still another machine, running Witnesses (all other machines run with a 
> lean GENERIC, with most of the debuging features commented out) :
> System call __getcwd returning with the following locks held:
> shared rw Name Cache (Name Cache) r = 0 (0xc0ee7e1c) locked 
> _at_ /usr/src/sys/kerne/vfs_cache.c:974

This is definitely related to:
SVN rev 190655 on 2009-04-02 21:16:20Z by peter
(peter_at_ CCed)

> panic: witness_warn
> cpuid = 0
> KDB: enter: panic


WBR
-- 
Boris Samorodov (bsam)
Research Engineer, http://www.ipt.ru Telephone & Internet SP
FreeBSD Committer, http://www.FreeBSD.org The Power To Serve