Hi,
  I have a looong time opened a bugreport on XFS at
http://bugzilla.kernel.org/show_bug.cgi?id=7287 and I see it still
appear in my kernel output during bootup. I guess this is one of the
relatively new kernel self-testing features introduced recently. I
just wanted to let you know about that.


=============================================
[ INFO: possible recursive locking detected ]
2.6.19-rc6-git10 #1
---------------------------------------------
mount/3439 is trying to acquire lock:
 (&(&ip->i_lock)->mr_lock){----}, at: [<c11136fc>] xfs_ilock+0x4a/0x68

but task is already holding lock:
 (&(&ip->i_lock)->mr_lock){----}, at: [<c11136fc>] xfs_ilock+0x4a/0x68

other info that might help us debug this:
2 locks held by mount/3439:
 #0:  (&inode->i_mutex){--..}, at: [<c12f138d>] mutex_lock+0x8/0xa
 #1:  (&(&ip->i_lock)->mr_lock){----}, at: [<c11136fc>]
xfs_ilock+0x4a/0x68

stack backtrace:
 [<c1003107>] show_trace_log_lvl+0x1a/0x2f
 [<c1003202>] show_trace+0x12/0x14
 [<c10039c5>] dump_stack+0x19/0x1b
 [<c10307da>] __lock_acquire+0x106/0x94e
 [<c10315d7>] lock_acquire+0x5c/0x79
 [<c102d5da>] down_write+0x2b/0x44
 [<c11136fc>] xfs_ilock+0x4a/0x68
 [<c1113fd2>] xfs_iget+0x2a0/0x5de
 [<c112a239>] xfs_trans_iget+0xd6/0x135
 [<c1117e95>] xfs_ialloc+0xa7/0x41f
 [<c112abde>] xfs_dir_ialloc+0x6d/0x267
 [<c11314d8>] xfs_create+0x2f4/0x5ae
 [<c113981d>] xfs_vn_mknod+0x127/0x242
 [<c1139961>] xfs_vn_create+0x12/0x14
 [<c105d615>] vfs_create+0x6a/0xb4
 [<c105fbe1>] open_namei+0x179/0x57a
 [<c1055b79>] do_filp_open+0x26/0x3b
 [<c1055bd1>] do_sys_open+0x43/0xc7
 [<c1055c8d>] sys_open+0x1c/0x1e
 [<c1002cc5>] sysenter_past_esp+0x56/0x8d
 =======================

I can provide more details upon request. Please Cc: me in replies.
Thanks.
Martin
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Reply via email to