Re: [Threads-devel] showstopper bug for pthread backend

Subject: Re: [Threads-devel] showstopper bug for pthread backend
From: Anthony Williams (anthony_at_[hidden])
Date: 2010-02-08 11:49:19

On 08/02/10 16:02, Paul Pogonyshev wrote:
> Can you please look at bug in tickets 2330 and 3735 (both reference
> the same bug)?  It's a showstopper as it can potentially lock down any
> multithreaded application on pthread backend.  I verified that the bug
> is reproducible with Boost 1.42.
>
> Ticket 3735 has an attached program that demonstrates the bug.  (I
> just disregarded all extra files and compiled the only .cpp file in
> that archive and it worked).  Note that here (2 CPU cores) the program
> freezes after 1000--4000 iterations, so the bug is extremelly
> difficult to reproduce in real programs, but when it does happen,
> result can be just full program lockdown and likely data loss.  On
> machines with more cores the bug is probably reproduced more readily.
>
> Ticket 2330 has a proposed patch, but it doesn't compile when applied
> to Boost 1.42 and I don't understand it enough to try to upgrade.

The patch is insufficient. I have an idea how to fix it, but have not 
yet found the time to implement it. If my fix works it should be in 
boost 1.43

Anthony
-- 
Author of C++ Concurrency in Action     http://www.stdthread.co.uk/book/
just::thread C++0x thread library             http://www.stdthread.co.uk
Just Software Solutions Ltd       http://www.justsoftwaresolutions.co.uk
15 Carrallack Mews, St Just, Cornwall, TR19 7UL, UK. Company No. 5478976