Re: kern/46377 (Use of Coda file system triggers an immediate kernel panic)

[u-bsdmwle%aetey.se@localhost]

The following reply was made to PR kern/46377; it has been noted by GNATS.

From: u-bsdmwle%aetey.se@localhost
To: gnats-bugs%NetBSD.org@localhost
Cc: kern-bug-people%netbsd.org@localhost, netbsd-bugs%netbsd.org@localhost,
        gnats-admin%netbsd.org@localhost, rmind%NetBSD.org@localhost
Subject: Re: kern/46377 (Use of Coda file system triggers an immediate kernel
 panic)
Date: Wed, 2 May 2012 14:23:49 +0000

 On Sat, Apr 28, 2012 at 08:35:03PM +0000, rmind%NetBSD.org@localhost wrote:
 > Synopsis: Use of Coda file system triggers an immediate kernel panic
 > 
 > State-Changed-From-To: open->feedback
 > State-Changed-By: rmind%NetBSD.org@localhost
 > State-Changed-When: Sat, 28 Apr 2012 20:35:03 +0000
 > State-Changed-Why:
 > Should be fixed.  Can you try latest -current?
 
 Testing with
 "NetBSD xxx 6.99.6 NetBSD 6.99.6 (GENERIC) #0: Tue May  1 16:11:22 UTC 2012  
builds%b6.netbsd.org@localhost:/home/builds/ab/HEAD/i386/201205010910Z-obj/home/builds/ab/HEAD/src/sys/arch/i386/compile/GENERIC
 i386"
 
 $ cp /coda/coda.cs.cmu.edu/WELCOME /tmp/xxx
 
 seems now working, but copying another (and bigger, 640160 bytes) file leads
 to a different system crash, several seconds after the copy operation finishes
 (apparently at sync?) :
 
 ------------------------------------
 $ sync
 $ sync
 $ date; cp /coda/coda.cs.cmu.edu/WELCOME /tmp/xxx; while date; do sleep 1; done
  ... the date loop runs for several minutes without problems
  (was it because of "sync" above or because the file is small - 879 bytes ?)
 $ date; cp /coda/coda.cs.cmu.edu/playground/do-not-trust-me/busybox 
/tmp/busybox; while date; do sleep 1; done
 Wed May  2 16:03:53 CEST 2012
 Wed May  2 16:04:05 CEST 2012    <=== ~ 12 seconds to fetch the file
 Wed May  2 16:04:06 CEST 2012
 Wed May  2 16:04:07 CEST 2012
 Wed May  2 16:04:08 CEST 2012
 Wed May  2 16:04:09 CEST 2012
 Wed May  2 16:04:10 CEST 2012
 Wed May  2 16:04:11 CEST 2012
 Wed May  2 16:04:12 CEST 2012
 Wed May  2 16:04:13 CEST 2012
 Wed May  2 16:04:14 CEST 2012
 Wed May  2 16:04:15 CEST 2012
 Wed May  2 16:04:16 CEST 2012
 Wed May  2 16:04:17 CEST 2012
 Wed May  2 16:04:18 CEST 2012
 Wed May  2 16:04:19 CEST 2012
 Wed May  2 16:04:20 CEST 2012
 Wed May  2 16:04:21 CEST 2012
 Wed May  2 16:04:22 CEST 2012
 Wed May  2 16:04:23 CEST 2012
 Wed May  2 16:04:24 CEST 2012
 [then nothing happens any more, as the system crashes]
 ------------------------------------
 
 on the console:
 ------------------------------------
 panic: kernel diagnostic assertion "bp->b_objlock == vp->v_interlock" fa
 iled: file "/home/builds/ab/HEAD/src/sys/kern/vfs_bio.c", line 1109 
 fatal breakpoint trap in supervisor mode
 trap type 1 code 0 eip c0269d74 cs 8 eflags 282 cr2 bbb70bf0 ilevel 0
 Stopped in pid 0.61 (system) at netbsd:breakpoint+0x4:  popl    %ebp
 db{1}> bt
 breakpoint(c0c3ffbd,c0cfbee0,c0b9472c,dcdc1618,5,0,0,0,0,0) at 
netbsd:breakpoint
 +0x4
 
vpanic(c0b9472c,dcdc1618,c0cd7280,c07d1efa,6,0,dcdc165c,c08e14b0,c0b9472c,c0b948
 dd) at netbsd:vpanic+0x1e2
 
kern_assert(c0b9472c,c0b948dd,c0c3f0f0,c0c41090,455,c0cd72f8,c0cd72f4,c4d23f88,c
 4c2010c,c4c20000) at netbsd:kern_assert+0x23
 incore(c52dd6f0,fffffff4,ffffffff,c4d2188c,fffffff4,ffffffff,dcdc16dc,3,0,25) 
at
  netbsd:incore+0xad
 
getblk(c52dd6f0,fffffff4,ffffffff,4000,0,0,dcdc172c,c08cc01d,c52c57a0,c45f5f30) 
 a
 t netbsd:getblk+0x39
 ufs_bmaparray(c52dd6f0,25,0,dcdc188c,0,0,dcdc18a0,c0867d40,c52dd6f0,0) at 
netbsd
 :ufs_bmaparray+0x214
 
ufs_bmap(dcdc17f8,c52f5ec8,dcdc17fc,c08e1865,c52f5f68,c0c40f81,c52c57a0,c52dd6f0
 ,0,c0b8427c) at netbsd:ufs_bmap+0x84
 VOP_BMAP(c52dd6f0,25,0,dcdc189c,dcdc188c,dcdc18a0,0,4040000,1000000,dad80000) 
at
  netbsd:VOP_BMAP+0x47
 genfs_do_io(dae14000,c000,8011,1,c08d861d,3,7,dcdc1a7c,6,0) at 
netbsd:genfs_do_i
 o+0x2b7
 
genfs_gop_write(c52dd6f0,dcdc1a4c,c,8011,dcdc1a70,13,dcdc0000,c01c643b,c4c2010c,
 c4d23f88) at netbsd:genfs_gop_write+0x5f
 genfs_do_putpages(c52dd6f0,0,0,0,0,8011,0,0,c4d44000,c4d869e0) at 
netbsd:genfs_d
 o_putpages+0xa9e
 
genfs_putpages(dcdc1b08,c0b842a0,dcdc1b1c,c52d9300,c4b41924,0,dcdc1b1c,c05e290f,
 c0d1cb68,c0b8439c) at netbsd:genfs_putpages+0x3f
 VOP_PUTPAGES(c52dd6f0,0,0,0,0,8011,dcdc1b6c,c07cf41c,c460b160,c4ce71e8) at 
netbs
 d:VOP_PUTPAGES+0x47
 vflushbuf(c52dd6f0,8,dcdc1bbc,0,c4d44000,2,dcdc1bcc,c08f546e,c4d44000,1) at 
netb
 sd:vflushbuf+0x75
 ffs_full_fsync(c52dd6f0,8,1,c08e139b,0,c52d9300,c4d9dbb4,c52dd6f0,0,12) at 
netbs
 d:ffs_full_fsync+0x162
 
ffs_fsync(dcdc1cb4,1,dcdc1cdc,c08f83a0,c52dd6f0,12,4,c46e9c00,c0b84048,c52dd6f0)
  at netbsd:ffs_fsync+0x55
 VOP_FSYNC(c52dd6f0,c46e5f00,8,0,0,0,0,c4ce7020,c4d4ad20,0) at 
netbsd:VOP_FSYNC+0
 x4a
 sched_sync(c4ce7020,e69000,e72000,0,c0100307,0,0,0,0,0) at 
netbsd:sched_sync+0x1
 09
 db{1}>
 ------------------------------------
 
 The crash seems to be sligtly non-deterministic, sometimes it does not
 go into the debugger, just hangs:
 console:
 ------------------------------------
 # sysctl -w ddb.onpanic=1
 ddb.onpanic: 0 -> 1
 # [some activity (starting/restarting Coda client), then the test and]
 panic: kernel diagnostic assertion "bp->b_objlock == vp->v_interlock" failed: 
fi
 le "/home/builds/ab/HEAD/src/sys/kern/vfs_bio.c", line 1109 
 fatal breakpoint trap in supervisor mode
 [nothing happens, no db-prompt either]
 ------------------------------------
 
 Regards,
 Rune