2.6.19-rc2mm2 as well as 2.6.19-rc3 panics with the following OOPS when I
run kernbench continuously and execute cpu offline and online in parallel. 
This is a 2 socket 4 core tyan opteron system with 8Gs of RAM.  
The kernel has lockdep, and other debug flags enabled.  Also, we have a 
lockdep warning with XFS.  Attaching the dmesg and the .config used as well.  
Here's the snapshot of lockdep warning + oops.


[  469.336719] 
[  469.336721] =============================================
[  469.336728] [ INFO: possible recursive locking detected ]
[  469.336731] 2.6.19-rc3 #2
[  469.336737] ---------------------------------------------
[  469.336745] rm/4538 is trying to acquire lock:
[  469.336753]  (&(&ip->i_lock)->mr_lock){----}, at: [<ffffffff80320050>] xfs_ilock+0x52/0x77
[  469.336782] 
[  469.336784] but task is already holding lock:
[  469.336791]  (&(&ip->i_lock)->mr_lock){----}, at: [<ffffffff80320050>] xfs_ilock+0x52/0x77
[  469.336816] 
[  469.336817] other info that might help us debug this:
[  469.336825] 3 locks held by rm/4538:
[  469.336832]  #0:  (&inode->i_mutex/1){--..}, at: [<ffffffff80286055>] do_unlinkat+0x75/0x136
[  469.336866]  #1:  (&inode->i_mutex){--..}, at: [<ffffffff80285f98>] vfs_unlink+0x45/0x8d
[  469.336897]  #2:  (&(&ip->i_lock)->mr_lock){----}, at: [<ffffffff80320050>] xfs_ilock+0x52/0x77
[  469.336928] 
[  469.336929] stack backtrace:
[  469.336936] 
[  469.336938] Call Trace:
[  469.336947]  [<ffffffff8024b4cc>] __lock_acquire+0x381/0xb8c
[  469.336957]  [<ffffffff8024bf80>] lock_acquire+0x4b/0x66
[  469.336966]  [<ffffffff80320050>] xfs_ilock+0x52/0x77
[  469.336976]  [<ffffffff80247c97>] down_write+0x1e/0x27
[  469.336985]  [<ffffffff80320050>] xfs_ilock+0x52/0x77
[  469.336993]  [<ffffffff8033cad1>] xfs_lock_dir_and_entry+0x92/0xce
[  469.337002]  [<ffffffff8033cd23>] xfs_remove+0x216/0x44a
[  469.337013]  [<ffffffff80346c96>] xfs_vn_unlink+0x21/0x4f
[  469.337024]  [<ffffffff804a72b9>] __mutex_lock_slowpath+0x249/0x279
[  469.337034]  [<ffffffff8024aaaf>] mark_held_locks+0x65/0x84
[  469.337043]  [<ffffffff804a72b9>] __mutex_lock_slowpath+0x249/0x279
[  469.337053]  [<ffffffff8024ac7c>] trace_hardirqs_on+0x101/0x128
[  469.337062]  [<ffffffff804a72dc>] __mutex_lock_slowpath+0x26c/0x279
[  469.337072]  [<ffffffff802832d9>] permission+0xa0/0xa9
[  469.337081]  [<ffffffff80285fb4>] vfs_unlink+0x61/0x8d
[  469.337090]  [<ffffffff80286098>] do_unlinkat+0xb8/0x136
[  469.337100]  [<ffffffff804a8649>] trace_hardirqs_on_thunk+0x35/0x37
[  469.337109]  [<ffffffff8024ac7c>] trace_hardirqs_on+0x101/0x128
[  469.337118]  [<ffffffff804a8649>] trace_hardirqs_on_thunk+0x35/0x37
[  469.337129]  [<ffffffff80209a4e>] system_call+0x7e/0x83
[  469.337137] 


[  488.264267] Unable to handle kernel NULL pointer dereference at 0000000000000000 RIP: 
[  488.264271]  [<ffffffff8025d255>] __drain_pages+0x33/0x7d
[  488.264288] PGD 2753ae067 PUD 275f66067 PMD 0 
[  488.264302] Oops: 0000 [1] PREEMPT SMP 
[  488.264315] CPU 2 
[  488.264323] Modules linked in:
[  488.264332] Pid: 7617, comm: sh Not tainted 2.6.19-rc3 #2
[  488.264339] RIP: 0010:[<ffffffff8025d255>]  [<ffffffff8025d255>] __drain_pages+0x33/0x7d
[  488.264353] RSP: 0018:ffff81016dd3fdf8  EFLAGS: 00010046
[  488.264358] RAX: 0000000000000000 RBX: 0000000000000086 RCX: 0000000000000000
[  488.264366] RDX: 0000000000000010 RSI: 0000000000000000 RDI: ffff810180000000
[  488.264373] RBP: 0000000000000000 R08: 0000000000000002 R09: 0000000000000001
[  488.264381] R10: ffffffff8025cb12 R11: 0000000000000000 R12: ffff810180000000
[  488.264389] R13: 0000000000000001 R14: 0000000000000001 R15: ffff81027a7e5998
[  488.264397] FS:  00002af4d80a5ae0(0000) GS:ffff81018001ad18(0000) knlGS:0000000000000000
[  488.264405] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[  488.264412] CR2: 0000000000000000 CR3: 00000002767bf000 CR4: 00000000000006a0
[  488.264420] Process sh (pid: 7617, threadinfo ffff81016dd3e000, task ffff81016e03d080)
[  488.264427] Stack:  0000000000000001 0000000000000001 0000000000000007 0000000000000001
[  488.264448]  000000000000000f ffffffff8025e771 ffffffff80538860 ffffffff8023e3c2
[  488.264466]  0000000000000001 0000000000000001 ffff81016e11a0c0 ffffffff8024ff50
[  488.264481] Call Trace:
[  488.264490]  [<ffffffff8025e771>] page_alloc_cpu_notify+0x18/0x33
[  488.264500]  [<ffffffff8023e3c2>] notifier_call_chain+0x23/0x32
[  488.264509]  [<ffffffff8024ff50>] _cpu_down+0x184/0x245
[  488.264516]  [<ffffffff8025003c>] cpu_down+0x2b/0x42
[  488.264525]  [<ffffffff803b25b8>] store_online+0x27/0x71
[  488.264534]  [<ffffffff802b7b4a>] sysfs_write_file+0xb6/0xe5
[  488.264544]  [<ffffffff8027c6e5>] vfs_write+0xb2/0x155
[  488.264551]  [<ffffffff8027c83d>] sys_write+0x45/0x70
[  488.264560]  [<ffffffff80209a4e>] system_call+0x7e/0x83
[  488.264566] 
[  488.264570] 
[  488.264572] Code: 8b 75 00 4c 89 e7 e8 82 f8 ff ff f6 c7 02 c7 45 00 00 00 00 
[  488.264634] RIP  [<ffffffff8025d255>] __drain_pages+0x33/0x7d
[  488.264644]  RSP <ffff81016dd3fdf8>
[  488.264650] CR2: 0000000000000000
[  488.264655]  

Thanks,
Kiran