Hi,

On Thu, May 31, 2012 at 11:31 AM, Lauri Kasanen <[email protected]> wrote:
> Hi
>
> There are a number of races in monkey right now. This patchset fixes all but 
> two, clock runtime and signal exit. The clock runtime one is a severe one, as 
> at any time a read may happen in the middle of a write, and as many of the 
> clock buffers are pointers, a half-written pointer would point absolutely 
> nowhere.
>
> Atomic instructions are really the only solution that makes sense for the 
> clock runtime issue.

   Any exclusive mechanism would work..

>
> The signal one is not too severe, at worst it could cause a crash at exit, 
> but we're exiting anyway.
>
> --
>
> Helgrind logs for all these patches are available at 
> http://monkey-project.com/~cand/helgrind_logs/ - you'll see how each patch 
> cuts down a specific warning. The amount of clock warnings fluctuates between 
> runs, but they still form the majority in all runs.
>
> --
>
> This first patch solves clock init. If the memory were allocated in a thread, 
> it may happen too late - a worker thread may try to read the memory before 
> it's allocated.
>

thanks for catching this, applied!

>
> - Lauri
>
> _______________________________________________
> Monkey mailing list
> [email protected]
> http://lists.monkey-project.com/listinfo/monkey
>



-- 
Eduardo Silva
http://edsiper.linuxchile.cl
http://www.monkey-project.com
_______________________________________________
Monkey mailing list
[email protected]
http://lists.monkey-project.com/listinfo/monkey

Reply via email to