Hello Jakub,

I tested the fix on the posixtest code I sent you and it seems to work. I
ran the test more than hundred times and no problem manifested. Also I
discovered a deadlock when using FIBRIL_DEFIBRILATE, that confirms that
something that should have been locked haven't been before the fix.

I still get some page faults running  PCUT tests for pthread library. They
manifest occasionally - race condition. I can't rule out possibility of a
race condition in my code though. Will do more investigating and get back
to you afterwards.

2015-06-16 16:24 GMT+02:00 Jakub Jermar <[email protected]>:

> Hi Jan,
>
> On 5.6.2015 9:38, Jan Mareš wrote:
> > Any input on that piece of code I sent in the previous message?
>
> Can you verify with the fix in mainline,2351? The trouble was that
> async_futex was not being held during FIBRIL_FROM_DEAD. So the counter
> in the async_futex got out of sync and remained unlocked.
>
> Jakub
>
>
> _______________________________________________
> HelenOS-devel mailing list
> [email protected]
> http://lists.modry.cz/listinfo/helenos-devel
>
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/listinfo/helenos-devel

Reply via email to