On Sat, Oct 20, 2001 at 08:18:35PM -0700, Matthew Dillon wrote:
Try turning off vmiodirenable:
sysctl -w vfs.vmiodirenable=0
And see if that makes a difference.
-Matt
Nope, it crashes just as fast.
I am now turning on crash dumps, so I can at least give a backtrace
This is the crash message
IdlePTD 5226496
initial pcb at 3b3500
panicstr: bwrite: buffer is not busy???
panic messages:
---
panic: lockmgr: pid 556, not exclusive lock holder 536 unlocking
syncing disks... panic: bwrite: buffer is not busy???
Uptime: 8m48s
pfs_vncache_unload(): 1 entries remaining
This is the backtrace:
#0 0xc021013e in dumpsys ()
#1 0xc020ff2b in boot ()
#2 0xc0210338 in poweroff_wait ()
#3 0xc023c6cf in bwrite ()
#4 0xc023d958 in vfs_bio_awrite ()
#5 0xc01e9588 in spec_fsync ()
#6 0xc01e9171 in spec_vnoperate ()
#7 0xc02b95d8 in ffs_sync ()
#8 0xc0248ce9 in sync ()
#9 0xc020fb30 in boot ()
#10 0xc0210338 in poweroff_wait ()
#11 0xc0206b59 in lockmgr ()
#12 0xc02423f2 in vop_stdunlock ()
#13 0xc046bcbc in ?? ()
#14 0xc02461e3 in vput ()
#15 0xc024aba1 in lstat ()
#16 0xc02f19f4 in syscall ()
#17 0xc02e603d in syscall_with_err_pushed ()
#18 0x804b27a in ?? ()
#19 0x804a58d in ?? ()
#20 0x8052d19 in ?? ()
#21 0x8048135 in ?? ()
Unfortunately I've already done a cvsup since that build, so it will
take a buildworld cycle to get a kernel with more symbol information. I
will start that work now
Mark
:2 seconds of testing showed that 'tar -czf /dev/null /nt' was enough for
:a kernel panick within 2 seconds.
:
:A few upgrades later the problem still exists.
:
--
Nice testing in little China...
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message