Panic with chromium and 8.1-STABLE (Thu Sep 16 09:52:17 BRT 2010)

[jhell]

On 09/22/2010 16:37, Kostik Belousov wrote:
> On Wed, Sep 22, 2010 at 03:58:12PM -0400, jhell wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 09/22/2010 09:28, Kostik Belousov wrote:
>>> On Sun, Sep 19, 2010 at 07:28:13PM -0300, Mario Sergio Fujikawa Ferreira wrote:
>>>> Hi,
>>>>
>>>> 	I've just began trying chrome web browser from
>>>> http://chromium.hybridsource.org/ but it triggered 2 panics on my
>>>> 8.1-STABLE system.
>>>>
>>>> $ uname -a
>>>> FreeBSD exxodus.fedaykin.here 8.1-STABLE FreeBSD 8.1-STABLE #26: Thu Sep 16 09:52:17 BRT 2010     lioux at exxodus:/usr/obj/usr/src/sys/LIOUX  amd64
>>>>
>>>> 	The panic information is:
>>>>
>>>> ------------
>>>> panic: vm_page_unwire: invalid wire count: 0
>>>> cpuid = 0
>>>> KDB: enter: panic
>>>>
>>>> 0xffffff006ecce000: tag ufs, type VREG
>>>>     usecount 1, writecount 1, refcount 4 mountedhere 0
>>>>     flags ()
>>>>     v_object 0xffffff0151489870 ref 0 pages 8
>>>>     lock type ufs: EXCL by thread 0xffffff00200947c0 (pid 25025)
>>>>         ino 119526591, on dev ufs/fsusr
>>>>
>>>> 0xffffff011107f938: tag ufs, type VREG
>>>>     usecount 0, writecount 0, refcount 4 mountedhere 0
>>>>     flags (VV_NOSYNC|VI_DOINGINACT)
>>>>     v_object 0xffffff0151f7f870 ref 0 pages 1284
>>>>     lock type ufs: EXCL by thread 0xffffff01882cc7c0 (pid 26689)
>>>>         ino 263, on dev md0
>>>> ------------
>>>>
>>>> 	I've made available 2 ddb textdumps at:
>>>>
>>>> http://people.freebsd.org/~lioux/panic/2010091900/textdump.tar.0
>>>> http://people.freebsd.org/~lioux/panic/2010091900/textdump.tar.1
>>>>
>>>> 	I was able to use chrome prior to this latest kernel update.
>>>> Now, I can reproduce a kernel panic even browsing www.google.com
>>>>
>>>> 	Please, let me know if I can provide any further information.
>>>
>>> Does it panic if you remove ZERO_COPY_SOCKETS option from the kernel
>>> config ?
>>
>> This is triggered as well on a system without ZERO_COPY_SOCKETS just to
>> clear that bit up.
> I do not know what did prompted you to decide that the issue is the same.
> There is nothing common except the word "panic" in the report by
> lioux and your backtraces.
> 
> You could have better luck showing your traces on the fs@ or asking zfs
> porters directly.

Well taking a stab at it I thought you were involved with that.
Considering it is being triggered on the FS level there is somthing in
common but that won't always mean the traces will be exact for two
different file-systems. It may just as well be two different cases so
Ill just forward them on to the correct people. Thanks anyway.


Regards,

-- 

 jhell,v