[OpenAFS-devel] possible recursive locking detected

[Troy Benjegerdes]

it seems that both locks are held by the same afsd process... 

I'm trying to get a build with inline afs_mutex* going now.

On Tue, Jul 15, 2008 at 10:52:07AM -0400, Chas Williams (CONTRACTOR) wrote:
> In message <20080714234544.GB16373@excalibur.hozed.org>,Troy Benjegerdes writes:
> >What does anyone make of this? 
> >Kernel 2.6.26-rc8, openafs-cvs HEAD
> >
> >This particular machine used to deadlock userspace without the lock
> >debugging enabled. It appears to be okay now.
> 
> i would guess that is it probably right.  there might be a recursive
> lock somewhere.  it looks like the rx layer with the afs_mutex_enter()
> pointing to one of the rx locks.  this probably highlights the need
> for the afs_mutex_enter() to be static inline so that the debugging 
> doesnt point to afs_mutex_enter().
> 
> the linux rx mutex layer has code to try to catch this so i am little
> puzzled that you dont see an osi_Panic() after this.
> 
> the debug looks a bit garbled.  i tried to correct for this below.
> it would be nice to find out which source code line corresponds to
> .rxi_ReapConnections+0x1b0.  that would tell which lock is the problem.
> 
> >[   37.850406] =============================================
> >[   37.947534] [ INFO: possible recursive locking detected ]
> >[   38.012120] 2.6.26-rc8 #10
> >[   38.044465] ---------------------------------------------
> >[   38.109049] afsd/3684 is trying to acquire lock:
> >[   38.164275]  (&l->mutex){--..}, at: [<d00000000014f0f4>]       .afs_mutex_enter+0x24/0x70 [libafs]
> >[   38.282318]
> >[   38.282319] but task is already holding lock:
> >[   38.352206]  (&l->mutex){--..}, at: [<d00000000014f0f4>]       .afs_mutex_enter+0x24/0x70 [libafs]
> >[   38.470250]
> >[   38.470251] other info that might help us debug this:
> >[   38.548459] 2 locks held by afsd/3684:
> >[   38.593282]  #0:  (afs_global_lock){--..}, at: [<d0000000001583fc>]    .osi_linux_alloc+0x17c/0x4c0 [libafs]
> >[   38.724950]  #1:  (&l->mutex){--..}, at: [<d00000000014f0f4>] .afs_mutex_enter+0x24/0x70 [libafs]
> >[   38.848298]
> >[   38.848300] stack backtrace:
> >[   38.900507] Call Trace:
> >[   38.929730] [c0000007f95330f0] [c000000000011144] .show_stack+0x64/0x210ateway: 10.1.0.2 (unreliable)
> >[   39.040389] [c0000007f95331b0] [c000000000011310] .dump_stack+0x20/0x40
> >[   39.119741] [c0000007f9533230] [c000000000098c90] .__lock_acquire+0xe20/0x1270
> >[   39.206376] [c0000007f9533330] [c0000000000991b4] .lock_acquire+0xd4/0x120
> >[   39.288851] [c0000007f95333f0] [c00000000038ce3c] .mutex_lock_nested+0xfc/0x420
> >[   39.376522] [c0000007f95334f0] [d00000000014f0f4] .afs_mutex_enter+0x24/0x70 [libafs]
> >[   39.487078] [c0000007f9533570] [d0000000001462b0] .rxi_ReapConnections+0x1b0/0x4e0 [libafs]
> >[   39.603871] [c0000007f9533690] [d00000000014acec] .rx_StartServer+0xac/0xf0 [libafs]
> >[   39.713388] [c0000007f9533740] [d000000000173c58] .afs_ResourceInit+0x1b8/0x1f0 [libafs]
> >[   39.827059] [c0000007f95337d0] [d0000000001b8024] .afs_DaemonOp+0x2f4/0x310 [libafs]
> >[   39.936573] [c0000007f9533970] [d0000000001b8b00] .afs_syscall_call+0x270/0x1ce0 [libafs]
> >[   40.051290] [c0000007f9533aa0] [d00000000013ebec] .afs_syscall+0x14c/0x6a0 [libafs]
> >[   40.159763] [c0000007f9533bc0] [d00000000015b248] .afs_unlocked_ioctl+0xc8/0x110 [libafs]
> >[   40.274476] [c0000007f9533c80] [c000000000160a40] .proc_reg_compat_ioctl+0xb0/0x100 > roo
> >[   40.382952] [c0000007f9533d30] [c00000000014a5d0] .compat_sys_ioctl+0xe0/0x500
> >[   40.469585] [c0000007f9533e30] [c0000000000086d4] syscall_exit+0x0/0x40

-- 
--------------------------------------------------------------------------
Troy Benjegerdes                'da hozer'                hozer@hozed.org  

Somone asked me why I work on this free (http://www.gnu.org/philosophy/)
software stuff and not get a real job. Charles Shultz had the best answer:

"Why do musicians compose symphonies and poets write poems? They do it
because life wouldn't have any meaning for them if they didn't. That's why
I draw cartoons. It's my life." -- Charles Shultz