On Wednesday, 07 March 2012 11:15:36 Torsten Förtsch wrote:
> On Sunday, 04 March 2012 10:29:45 SalusaSecondus wrote:
> > Broken behavior:
> > 4 threads, 2 free interpreters
> > 1: A (1,1)
> > 2: A (2,0)
> > 3: W
> > 4: W
> > 1: P (1,1) B
> > 2: P (0,2) <-- No broadcast because there was an availab
On Sunday, 04 March 2012 10:29:45 SalusaSecondus wrote:
> Broken behavior:
> 4 threads, 2 free interpreters
> 1: A (1,1)
> 2: A (2,0)
> 3: W
> 4: W
> 1: P (1,1) B
> 2: P (0,2) <-- No broadcast because there was an available interpreter
> prior to this putback.
> 3: A (1,1)
> 3: P (0,2) <-- No broad
The bug is a race condition, so as you say, I don't really think it's
definitively testable.
Adam
On 3/6/2012 10:19 PM, SalusaSecondus wrote:
I've looked through those tests, and none seem to apply and I'm not sure
how one would even go about writing a unit test for this. We haven't
even been
I've looked through those tests, and none seem to apply and I'm not sure
how one would even go about writing a unit test for this. We haven't
even been able to reproduce the problem on all systems.
I suspect the best we can do may simply be manual load tests (such as
Max and I have done) showing
On Tue, Mar 6, 2012 at 11:12 AM, Salusa Secondus
wrote:
> Fred,
>
> I'll take a look at the tests when I get home to see what I can find.
> Do you know of any specific tests which deal with multi-threading or
> potential deadlocks?
Torsten, Steve H., and Gozer know that part of the code a lot bet
Adam,
We'd love to see this in 2.0.6. This has caused us lots of server problems.
Fred,
I'll take a look at the tests when I get home to see what I can find.
Do you know of any specific tests which deal with multi-threading or
potential deadlocks?
Greg
-
Hi Salusa,
Would you or Max be able to construct a unit test that demonstrates
this failure condition, and then success once the patch is applied?
There should be some example tests in the t/ directory which you can
draw on for inspiration.
On Sun, Mar 4, 2012 at 10:29 AM, SalusaSecondus wrote:
Seems reasonable to me. It seems like it might be a good idea to fold
this into 2.0.6 RC3 to me. Thoughts?
Adam
On 12-03-04 01:29 PM, SalusaSecondus wrote:
(Patch and system details at bottom)
-
To unsubscribe, e-mail:
(Patch and system details at bottom)
Hi all. I've root-caused and written a patch for the children stuck on
futex problem described by both Sean Thorne in 2009 and Max Barry (who I
work with) in 2011.
The core of the problem is that modperl_tipool_putback_base only
broadcasts that there are more
