On 12 Mar, Dag-Erling Smorgrav wrote:
= Mikhail Teterin <[EMAIL PROTECTED]> writes:
= > > If you can, please reproduce the panic on a kernel compiled with the
= > > INVARIANTS, INVARIANT_SUPPORT and WITNESS options.
= > Well, with this options on, the machine does not crash, but the
= > program segfaults on startup:
=
= The trace you're showing looks like it's from a shell script that
= starts civctp. I need to see the trace from the civctp binary itself.
No, that trace was obtained from a simple
ktrace civctp
There is no shell-wrapper around the binary:
file /opt/bin/civctp
/opt/bin/civctp: ELF 32-bit LSB executable, Intel 80386, version 1, statically
linked, stripped
It is just one big executable. You are welcome to download it from:
http://aldan.algebra.com:8015/~mi/civctp-crash/civctp.bz2
uncompress it and try to run it (just 43Kb compressed).
May be, it is because it is a _staticly_ linked Linux executable (the
_dynamicly_ linked Netscape works fine).
= > lock order reversal
= > 1st lockmgr interlock last acquired @ ../../kern/kern_lock.c:239
= > 2nd 0xcefa0520 process lock @ ../../kern/kern_sig.c:183
= > 3rd 0xc1029f80 lockmgr interlock @ ../../kern/kern_lock.c:560
=
= Haven't seen this one before... If it's reproducible, could you do the
= following:
No... This the only machine I have at home. No serial console or
cable... It is reproduceable -- happens now at boot time...
-mi
= 1) recompile your kernel with WITNESS_DDB
=
= 2) hook up a serial console and boot with '-h' in /boot.config
=
= 3) provoke the reversal, then get the output from 'trace', 'show
= mutex' and 'show witness' at the DDB prompt
=
= 4) type 'continue' to exit DDB and continue running normally.
= > lock order reversal
= > 1st vnode interlock last acquired @ ../../kern/vfs_vnops.c:625
= > 2nd 0xc0419680 mntvnode @ ../../ufs/ffs/ffs_vfsops.c:939
= > 3rd 0xcefb986c vnode interlock @ ../../ufs/ffs/ffs_vfsops.c:948
=
= This is a known (and probably benign) bug.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
