From: Charles Ballowe (hangman@steelballs.org)
Date: Mon Sep 30 2002 - 11:56:38 EDT
Thanks to Johan Brusche of HP for his answer -- this is a known
bug and is fixed in PK4 or PK5
-Charlie Ballowe
----- Forwarded message from "Brusche, Johan" <Johan.Brusche@hp.com> -----
Delivered-To: hangman@steelballs.org
Subject: RE: mcs_lock: time limit exceeded?
Date: Mon, 30 Sep 2002 17:39:08 +0200
From: "Brusche, Johan" <Johan.Brusche@hp.com>
To: "Charles Ballowe" <hangman@steelballs.org>
Charlie,
The "mcs_lock:time limit exceeded" from routine nshmget() is a known
problem, solved by patch OSF510-064 in patchkit#4 or #5.
14 panic(s = (unallocated - symbol optimized away))
["../../../../src/kernel/bsd/subr_prf.c":1334, 0xffffffff00064b2c]
15 simple_lock_fault(slp = (unallocated - symbol optimized away),
state = (unallocated - symbol optimized away), caller = (unallocated -
symbol optimized away), arg = (unallocated - symbol optimized away), fmt
= (unallocated - symbol optimized away), error = (unallocated - symbol
optimized away)) ["../../../../src/kernel/kern/lock.c":2805,
0xffffffff00092f04]
16 mcs_lock_time_violation(slp = (unallocated - symbol optimized
away), state = (unallocated - symbol optimized away), caller =
(unallocated - symbol optimized away))
["../../../../src/kernel/kern/lock.c":3073, 0xffffffff000935ec]
17 nshmget(p = (unallocated - symbol optimized away), key =
(unallocated - symbol optimized away), size = (unallocated - symbol
optimized away), shmflg = (unallocated - symbol optimized away), attr =
(unallocated - symbol optimized away), retval = (unallocated - symbol
optimized away)) ["../../../../src/kernel/bsd/svipc_shm.c":2033,
0xffffffff000705f8]
18 syscall(0x1e000000, 0x7a0, 0x1, 0x0, 0xfffffc007d5c7500)
["../../../../src/kernel/arch/alpha/syscall_trap.c":713,
0xffffffff003b0550]
19 _Xsyscall(0x8, 0x3ff80115088, 0x1401130f8, 0x0, 0x1e000000)
["../../../../src/kernel/arch/alpha/locore.s":1785, 0xffffffff003b3d9c]
Cheers,
Johan.
-----Original Message-----
From: Charles Ballowe [mailto:hangman@steelballs.org]
Sent: maandag 30 september 2002 16:57
To: Brusche, Johan
Subject: Re: mcs_lock: time limit exceeded?
On Mon, Sep 30, 2002 at 09:03:25AM +0200, Brusche, Johan wrote:
> Charles,
>
> There isn't enough data in your entry, we need at least the CPU stack
> traces to see what this panic is about (if it's a known softw
> problem).
>
> The advise will anyway come to : install PK5 for V5.1
>
> Johan.
i'm attaching a full copy of the crash-data file from this weekend's
crash.
Is it possible to confirm that this is a known bug - I'll be opening a
ticket with compaq support shortly because that's how management wants
it played,
but i generally find that answers come faster from list member.
thanks
-charlie
----- End forwarded message -----
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:48:54 EDT