[nptl] Re: SuSe SLES 9 vs. test_schedRR

["Loic Domaigne" <loic-dev@gmx.net>]

Hi, 

... Sorry, I am over-busy right now. So I have to restrict myselves 
to the bare minimun... 


> Ok, I just tested with SuSe Linux Entreprise Server (SLES) 9.
> 
> Fresh install: (kernel 2.6.5-97smp ; NPTL 0.61)
> $ ./test 200 3
> will return PASSED instantly about 9 runs out of 10, and will hang the
> 10th run. After the timeout expires, we can see that a lot of threads
> have counter = 1.

Ok. 

 
> Modified kernel: kernel 2.6.8.1 from kernel.org (is it what is called
> "vanilla" ?), same libc (NPTL 0.61).
> -> same behavior.

Ok.

Yeah, people used to say "vanilla kernel" for a kernel free from any 
patch to constrat to kernel found in distros that usually are (more 
or less heavily) patched.

 
> Modified kernel + libc (more recent libc as well)
> -> same behavior as with FC2 -- the error appears more frequently and
> even when the test passes there is a short (but noticeable) delay. 

Wait, wait... 

The problem occurs more frequently with a vanilla kernel +  a glibc
closer to RH's glibc (and different from the one shipped by default 
with SuSE)?

Is it correct? 


> I think the delay and change in bug frequency could be related to a
> different behavior of the loader ld-linux. 

I do not see why ld would play a role here ? 


> Actually, to "change" the libc, the test sample is linked against an
> alternative libc which does not stand in a standard directory (/lib,
> ...). This might explain this behavior difference.

I don't get it :-\ 


So Long,
Loic. 

-- 
--
// Sender address goes to /dev/null (!!) 
// Use my 32/64 bits, ANSI C89, compliant email-address instead:   

unsigned y[]=
{0,34432,26811,16721,41866,63119,61007,48155,26147,10986};
void x(z){putchar(z);}; unsigned t; 
main(i){if(i<10){t=(y[i]*47560)%65521;x(t>>8);x(t&255);main(++i);}}

GMX ProMail mit bestem Virenschutz http://www.gmx.net/de/go/mail
+++ Empfehlung der Redaktion +++ Internet Professionell 10/04 +++

Liste de diffusion nptl
Pour se désinscrire : mailto:nptl_request@bullopensource.org?subject=unsubscribe