LOR on head ...

Dan Mack mack at macktronics.com
Tue Aug 6 00:36:55 UTC 2013


On Mon, 5 Aug 2013, Davide Italiano wrote:

> The LOR is a false positive.
> See the comment in sys/ufs/ufs/ufs_dirhash.c
> Also, switching motherboards is not related to this in any way. You'll
> eventually hit that LOR report, unless you disabled WITNESS.

Ah, thank you Davide; sorry for the noise ... I've been using only ZFS for 
so long that I haven't run across this yet.  I'm also getting these too 
which appear to be the same sort of thing, no?

lock order reversal:
  1st 0xfffffe010157d418 ufs (ufs) @ /usr/src/sys/kern/vfs_subr.c:2099
  2nd 0xffffff80ec37fa78 bufwait (bufwait) @ /usr/src/sys/ufs/ffs/ffs_vnops.c:262
  3rd 0xfffffe010157b9a0 ufs (ufs) @ /usr/src/sys/kern/vfs_subr.c:2099
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xffffff81166cecb0
kdb_backtrace() at kdb_backtrace+0x39/frame 0xffffff81166ced60
witness_checkorder() at witness_checkorder+0xd4f/frame 0xffffff81166cedf0
__lockmgr_args() at __lockmgr_args+0x6f2/frame 0xffffff81166cef20
ffs_lock() at ffs_lock+0x84/frame 0xffffff81166cef70
VOP_LOCK1_APV() at VOP_LOCK1_APV+0xf5/frame 0xffffff81166cefa0
_vn_lock() at _vn_lock+0xab/frame 0xffffff81166cf010
vget() at vget+0x70/frame 0xffffff81166cf060
vfs_hash_get() at vfs_hash_get+0xf5/frame 0xffffff81166cf0b0
ffs_vgetf() at ffs_vgetf+0x41/frame 0xffffff81166cf140
softdep_sync_buf() at softdep_sync_buf+0x8fa/frame 0xffffff81166cf1f0
ffs_syncvnode() at ffs_syncvnode+0x258/frame 0xffffff81166cf270
ffs_truncate() at ffs_truncate+0x5ca/frame 0xffffff81166cf450
ufs_direnter() at ufs_direnter+0x891/frame 0xffffff81166cf510
ufs_makeinode() at ufs_makeinode+0x573/frame 0xffffff81166cf6d0
VOP_CREATE_APV() at VOP_CREATE_APV+0xea/frame 0xffffff81166cf700
vn_open_cred() at vn_open_cred+0x300/frame 0xffffff81166cf850
kern_openat() at kern_openat+0x1f5/frame 0xffffff81166cf9a0
amd64_syscall() at amd64_syscall+0x265/frame 0xffffff81166cfab0
Xfast_syscall() at Xfast_syscall+0xfb/frame 0xffffff81166cfab0
--- syscall (5, FreeBSD ELF64, sys_open), rip = 0x80093d3ea, rsp = 0x7fffffffd758, rbp = 0x7fffffffd7b0 ---

dan
--
Dan Mack



More information about the freebsd-current mailing list