Re: kern/37881: repeatable crash in filesystem when running MP

[dieter roelants <dieter.NetBSD%pandora.be@localhost>]

The following reply was made to PR kern/37881; it has been noted by GNATS.

From: dieter roelants <dieter.NetBSD%pandora.be@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/37881: repeatable crash in filesystem when running MP
Date: Sun, 27 Jan 2008 14:46:26 +0100

 On Sun, 27 Jan 2008 13:10:03 +0000 (UTC)
 Antti Kantee <pooka%cs.hut.fi@localhost> wrote:
 
 >  Hmm.  How are you copying the files?  Many at a time?  fd 16 looks
 >  semi-suspicious.
 
 In this case it were only 2 files, and nautilus copies them one after
 the other. Nautilus always uses a bunch of file descriptors. Even now,
 when I have no file manager windows open, it has 20 files open (most of
 them pipes).
 
 >  > #25 0xc02c6727 in knote (list=0xcefe5be4, hint=6)
 >  >     at /usr/src/sys/kern/kern_event.c:1301
 >  
 >  print *list
 $1 = {slh_first = 0x0}
 
 >  print *kn
 $2 = {kn_link = {sle_next = 0xc304aea4}, kn_selnext = {sle_next = 0x0}, 
   kn_tqe = {tqe_next = 0xc055996c, tqe_prev = 0x0}, kn_kq = 0x0, kn_kevent = {
     ident = 0, filter = 1048576, flags = 1, fflags = 4096, data = 4096, 
     udata = -842719232}, kn_status = 3487384, kn_sfflags = 0, 
   kn_sdata = 3487447, kn_ptr = {p_fp = 0x0, p_proc = 0x0, p_opaque = 0x0}, 
   kn_fop = 0x0, kn_hook = 0x0}
  
 >  > #26 0xc025b9ed in ffs_write (v=0xcdccbc04)
 >  >     at /usr/src/sys/ufs/ufs/ufs_readwrite.c:507
 >  
 >  print *uio
 $3 = {uio_iov = 0xcdccbc98, uio_iovcnt = 1, uio_offset = 348717056, 
   uio_resid = 0, uio_rw = UIO_WRITE, uio_vmspace = 0xcd8b1004}
 
 >  print *vp
 $4 = {v_uobj = {vmobjlock = {u = {mtxa_owner = 4}}, pgops = 0xc0480460, 
     memq = {tqh_first = 0xc24f9938, tqh_last = 0xc1ede580}, uo_npages = 85136, 
     uo_refs = 2}, v_cv = {cv_wmesg = 0xc051326d "vnode", cv_waiters = 0}, 
   v_size = 348717056, v_writesize = 348717056, v_iflag = 2113536, 
   v_vflag = 48, v_uflag = 0, v_numoutput = 2, v_writecount = 1, v_holdcnt = 2, 
   v_synclist_slot = 26, v_mount = 0xc300f000, v_op = 0xc2c7d400, v_freelist = {
     tqe_next = 0x0, tqe_prev = 0x0}, v_freelisthd = 0x0, v_mntvnodes = {
     tqe_next = 0x0, tqe_prev = 0xcea7f830}, v_cleanblkhd = {lh_first = 0x0}, 
   v_dirtyblkhd = {lh_first = 0xc314a3a8}, v_synclist = {tqe_next = 0x0, 
     tqe_prev = 0xc292a6d0}, v_dnclist = {lh_first = 0x0}, v_nclist = {
     lh_first = 0x0}, v_un = {vu_mountedhere = 0x0, vu_socket = 0x0, 
     vu_specnode = 0x0, vu_fifoinfo = 0x0, vu_ractx = 0x0}, v_type = VREG, 
   v_tag = VT_UFS, v_lock = {lk_flags = 66560, lk_sharecount = 0, 
     lk_interlock = {u = {mtxa_owner = 4}}, lk_exclusivecount = 1, 
     lk_recurselevel = 0, lk_waitcount = 0, lk_wmesg = 0xc05127d2 "vnlock", 
     lk_lockholder = 815, lk_locklwp = 12, lk_prio = 20, lk_timo = 0, 
     lk_lock_addr = 3223693237, lk_unlock_addr = 3223693104}, 
   v_vnlock = 0xcefe5bac, v_data = 0xcefe6e78, v_klist = {slh_first = 0x0}}
 
 >  > #29 0xc0302cc5 in dofilewrite (fd=22, fp=0xcea71414, buf=0xba226000, 
 >  >     nbyte=65536, offset=0xcea71440, flags=1, retval=0xcdccbd28)
 >  >     at /usr/src/sys/kern/sys_generic.c:392
 >  
 >  print *offset
 $5 = 348651520
 
 >  print numvnodes
 $6 = 5364
 
 >  print desiredvnodes
 $7 = 107334
 
 >  >   Copy some (large?) files around with nautilus, possibly
 >  >   from an smb share on a system with more than one CPU active?
 >  
 >  Does it happen when you start copying or when you've been copying for
 >  a while or what?  Be as specific as you can.
 
 Sometimes it only takes a few seconds, sometimes it takes longer. While
 trying if it was still there with a new kernel yesterday, I copied
 about 8Gb without crashing. I have the impression it crashes sooner if
 nothing else is going on my system. Does thamek any sense at all? :)
 
 >  .. actually, if it's easy to repeat, try this patch, but do the above
 
 I will, thanks.
 
 >  also.  I'm mostly guessing here, as I can't quite wrap my head around
 >  why this happens now and not before.
 
 Kind regards
 dieter