ghc-6.8.1, +RTS -N2

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

ghc-6.8.1, +RTS -N2

Johannes Waldmann
what could be the reason for my program
(compiled with -threaded) crashing with +RTS -N2

it seems to work fine with "-N1".
I compiled with ghc-6.8.1 (installed from binary package).
I am using forkIO, some MVars, some IORefs, some Channels.

It seems  -prof -threaded does not work:
ghc-6.8.1: combination not supported: Threaded/Profiling

When I run the executable under gdb I get

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6afeb90 (LWP 7801)]
0x08138d8c in checkBlackHoles ()

This behaviour occurs both on a dual core PC
and on an single core hyperthreaded PC,
both running openSuse-10.3

I cannot tell whether this is an regression from 6.6
(I cannot easily backport my code).

Any hints appreciated, Johannes Waldmann.
_______________________________________________
Glasgow-haskell-users mailing list
[hidden email]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Reply | Threaded
Open this post in threaded view
|

Re[2]: ghc-6.8.1, +RTS -N2

Bulat Ziganshin-2
Hello Johannes,

Tuesday, November 13, 2007, 12:08:20 PM, you wrote:

>> i suggest you to test it on one-threaded CPU too

> I just did. The behaviour is identical
> (works with -N1, segfaults with -N2)

> Any ideas? - Best regards, J.W.

no. just publish this into maillist, it's important to identify the
problem

well, in my practice the only way to make GHC crash (except for
errors in GHC itself) is to use unsafePerformIO, in particular for
creation global variables

--
Best regards,
 Bulat                            mailto:[hidden email]

_______________________________________________
Glasgow-haskell-users mailing list
[hidden email]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Reply | Threaded
Open this post in threaded view
|

Re: ghc-6.8.1, +RTS -N2

Simon Marlow-5
Bulat Ziganshin wrote:

> Hello Johannes,
>
> Tuesday, November 13, 2007, 12:08:20 PM, you wrote:
>
>>> i suggest you to test it on one-threaded CPU too
>
>> I just did. The behaviour is identical
>> (works with -N1, segfaults with -N2)
>
>> Any ideas? - Best regards, J.W.
>
> no. just publish this into maillist, it's important to identify the
> problem
>
> well, in my practice the only way to make GHC crash (except for
> errors in GHC itself) is to use unsafePerformIO, in particular for
> creation global variables

Yes, it could be a real bug.  Can you reduce the example as much as
possible and submit a bug report?

Cheers,
        Simon
_______________________________________________
Glasgow-haskell-users mailing list
[hidden email]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users