VM bug Mantis 4709 (was Re: [Seaside] Squeak 3.9?)

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

VM bug Mantis 4709 (was Re: [Seaside] Squeak 3.9?)

Adrian Lienhard
On Feb 15, 2007, at 07:53 , Keith Hodges wrote:

> I found that using vm 3.6-3 solved my instability freezing issues  
> with 3.9

Interesting..

I thought your image freezing might be related to the VM bug reported  
here: http://bugs.impara.de/view.php?id=4709. However, some tests I  
did recently show that the bug does not depend on the VM version but  
on the Linux kernel version 2.6.

1 - Squeak 3.9-9 VM, Linux 2.6 BUG
2 - Squeak 3.6-3 VM, Linux 2.6 BUG
3 - Squeak 3.7b-5 VM, Linux 2.4 OK
4 - Squeak 3.9-9 VM, Linux 2.4 OK

I updated the Mantis report with more details.

John, I tried what you suggest in this mail http://
lists.squeakfoundation.org/pipermail/seaside/2006-September/
008981.html. I attached the gc stats files to the Mantis report.

Keith, could you provide a gdb backtrace of a frozen image? Do you  
also see the problem when running with -mmap?

Cheers,
Adrian

PS: cross posted to the vm-dev list

Reply | Threaded
Open this post in threaded view
|

Re: VM bug Mantis 4709 (was Re: [Seaside] Squeak 3.9?)

Adrian Lienhard

Keith,

The workaround for [4709] which works well for us here is to use -
mmap. Though, from your backtrace below I suspect that you have  
another problem.

Adrian

BTW, cc-ed to vm-dev list.


On Mar 20, 2007, at 15:59 , Keith Hodges wrote:

> Another data point on the frozen images:
>
> #uname -a
>
> Linux Knoppix 2.6.19 #7 SMP PREEMPT Sun Dec 17 22:01:07 CET 2006  
> i686 GNU/Linux
>
>
> My frozen image give me this backtrace:
>
> (gdb) backtrace
> #0 0xb7e5d8e5 in nanosleep () from /lib/libc.so.6
> #1 0x08070b58 in aioSleep (microSeconds=-1080448704) at /home/
> piumarta/squeak/platforms/unix/vm/aio.c:281
> #2 0xb7f50466 in display_ioRelinquishProcessorForMicroseconds  
> (microSeconds=1000)
> at /home/piumarta/squeak/platforms/unix/vm-display-X11/sqUnixX11.c:
> 2317
> #3 0x0805921d in ioRelinquishProcessorForMicroseconds (us=-516) at /
> home/piumarta/squeak/platforms/unix/vm/sqUnixMain.c:471
> #4 0x0805f334 in primitiveRelinquishProcessor () at /home/piumarta/
> squeak/platforms/unix/src/vm/interp.c:17833
> #5 0x0806dc25 in interpret () at /home/piumarta/squeak/platforms/
> unix/src/vm/interp.c:7756
> #6 0x0805aae2 in main (argc=2, argv=0xbf99bc94, envp=0xfffffdfc)
> at /home/piumarta/squeak/platforms/unix/vm/sqUnixMain.c:1388
>
>> On Feb 15, 2007, at 07:53 , Keith Hodges wrote:
>>
>>> I found that using vm 3.6-3 solved my instability freezing issues  
>>> with 3.9
>>
>> Interesting..
>>
>> I thought your image freezing might be related to the VM bug  
>> reported here: http://bugs.impara.de/view.php?id=4709. However,  
>> some tests I did recently show that the bug does not depend on the  
>> VM version but on the Linux kernel version 2.6.
>>
>> 1 - Squeak 3.9-9 VM, Linux 2.6 BUG
>> 2 - Squeak 3.6-3 VM, Linux 2.6 BUG
>> 3 - Squeak 3.7b-5 VM, Linux 2.4 OK
>> 4 - Squeak 3.9-9 VM, Linux 2.4 OK
>>
>> I updated the Mantis report with more details.
>>
>> John, I tried what you suggest in this mail http://
>> lists.squeakfoundation.org/pipermail/seaside/2006-September/
>> 008981.html. I attached the gc stats files to the Mantis report.
>>
>> Keith, could you provide a gdb backtrace of a frozen image? Do you  
>> also see the problem when running with -mmap?
>>
>> Cheers,
>> Adrian
>>
>
>
>
> ___________________________________________________________ All New  
> Yahoo! Mail – Tired of Vi@gr@! come-ons? Let our SpamGuard protect  
> you. http://uk.docs.yahoo.com/nowyoucan.html