Re: [Valgrind-users] helgrind/drd annotations for one statement

[Godmar B. <go...@gm...>]

On Wed, Dec 5, 2012 at 12:29 PM, Julian Seward <js...@ac...> wrote:
>
> On Tuesday, December 04, 2012, Patrick J. LoPresti wrote:
>> There is no such thing as a benign data race.  Ever.
>
> <soapbox>
>
> Well said.  I couldn't have put any of this better myself.
>

For the interested audience: Adve/Boehm's very readable and
informative 2010 CACM paper [1] reflects, I believe, Julien's
sentiment as well. If my reading is correct, then the new C and C++
standards (C11 and C11++) effectively adopt the same approach by
leaving the semantics of programs with data races completely
undefined.

> Having spent a considerable amount of time working on Helgrind and then
> using it to chase races in some big hairy C++ codes, I became very
> skeptical of the "oh it's only a harmless race" arguments.  However,
> I gave up shouting about it after a while since it just made me look
> like a tiresome pedant hellbent on criticising people's clever go-faster-
> by-avoiding-locking schemes.
>

Julien, while you're still on your soapbox, let me make an attempt to
request an answer to a question I sent Oct 1 to this mailing list -
out of academic and practical curiosity, I'm really interested in
learning why Helgrind works the way it does currently, i.e., why did
it abandon Eraser-style locksets [2] in recent versions?

Thanks.

 - Godmar

[1] Memory Models: A Case for Rethinking Parallel Languages and Hardware
By Sarita V. Adve, Hans-J. Boehm
Communications of the ACM, Vol. 53 No. 8, Pages 90-101
10.1145/1787234.1787255
http://cacm.acm.org/magazines/2010/8/96610-memory-models-a-case-for-rethinking-parallel-languages-and-hardware/fulltext

[2] http://markmail.org/thread/odrhlvarckfgzvnk
"[Valgrind-users] Q.: Why did Helgrind abandon Eraser-style locksets?"