Problems with pthread on ARM/PXA270

Problems with pthread on ARM/PXA270

[Juergen Schindele]

Hello all,
we tried to implement a thread based software on our PXA270 based platform but
we had trouble with pthreads. So i modyfied a tiny progam to test. This little progam creates
10 threads and each thread does print something (an index an a pointer to itself). 

On my PC with Suse 11.0 Linux (kernel 2.6.27) this works fine as supposed to.
Each thread has his turn and print his own values.

But on our pxa270 platform the thread switch does not (or really rarely) happen
and only one thread is printing always the same values. So i suppose thread
switching is not working!

Any hints or experiences are welcome !!!

we use codesourcery compiler
arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
Copyright (C) 2007 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

and here the C-code compiled with  "arm-none-linux-gnueabi-gcc -o test test.c -lpthread"

#include <errno.h>
#include <error.h>
#include <stdio.h>
#include <stdlib.h>
#include <pthread.h>
#include <time.h>

#define NUM_THREADS 10
#define NUM_ITERS   50
#define TIMEOUT_NS  100000000L

static void *thread (void *);
static int argument[NUM_THREADS];
static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_attr_t defaultattr;

static void * print_thread (void *arg)
{
    int i;
    int number = *((int*)arg);

    for (;;) {
        pthread_t self = pthread_self ();
        printf ("Thread %d is [%d]\n", number, self);
    }
}

int main (void)
{
    pthread_t th;
    int i, retvalue;

    struct sched_param param;
    pthread_attr_init(&defaultattr);

    pthread_attr_setschedpolicy(&defaultattr, SCHED_RR);

    param.sched_priority = 10;
    if ((retvalue = pthread_attr_setschedparam(&defaultattr, &param))!=0) {
        printf ("ERROR\n");
        exit (0);
    }

    for (i = 0; i < NUM_THREADS; i++)
    {
        argument[i] = i;
        if (pthread_create (&th, &defaultattr, print_thread, &argument[i]) != 0)
            error (EXIT_FAILURE, 0, "cannot create thread");
        printf ("Thread [%d]\n", i);
    }

    while (1) {
        ;
    }

    /* notreached */
    return 0;
}

--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele@nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
--------------------------------------------------------------

Problems with pthread on ARM/PXA270

[Eric Miao]

Don't remember we had such problems before. What kind of system
and libraries are you using?


On Wed, Sep 16, 2009 at 3:26 PM, Juergen Schindele <schindele@nentec.de> wrote:
> Hello all,
> we tried to implement a thread based software on our PXA270 based platform but
> we had trouble with pthreads. So i modyfied a tiny progam to test. This little progam creates
> 10 threads and each thread does print something (an index an a pointer to itself).
>
> On my PC with Suse 11.0 Linux (kernel 2.6.27) this works fine as supposed to.
> Each thread has his turn and print his own values.
>
> But on our pxa270 platform the thread switch does not (or really rarely) happen
> and only one thread is printing always the same values. So i suppose thread
> switching is not working!
>
> Any hints or experiences are welcome !!!
>
> we use codesourcery compiler
> arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
> Copyright (C) 2007 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions. ?There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
>
> and here the C-code compiled with ?"arm-none-linux-gnueabi-gcc -o test test.c -lpthread"
>
> #include <errno.h>
> #include <error.h>
> #include <stdio.h>
> #include <stdlib.h>
> #include <pthread.h>
> #include <time.h>
>
> #define NUM_THREADS 10
> #define NUM_ITERS ? 50
> #define TIMEOUT_NS ?100000000L
>
> static void *thread (void *);
> static int argument[NUM_THREADS];
> static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;
> static pthread_attr_t defaultattr;
>
> static void * print_thread (void *arg)
> {
> ? ?int i;
> ? ?int number = *((int*)arg);
>
> ? ?for (;;) {
> ? ? ? ?pthread_t self = pthread_self ();
> ? ? ? ?printf ("Thread %d is [%d]\n", number, self);
> ? ?}
> }
>
> int main (void)
> {
> ? ?pthread_t th;
> ? ?int i, retvalue;
>
> ? ?struct sched_param param;
> ? ?pthread_attr_init(&defaultattr);
>
> ? ?pthread_attr_setschedpolicy(&defaultattr, SCHED_RR);
>
> ? ?param.sched_priority = 10;
> ? ?if ((retvalue = pthread_attr_setschedparam(&defaultattr, &param))!=0) {
> ? ? ? ?printf ("ERROR\n");
> ? ? ? ?exit (0);
> ? ?}
>
> ? ?for (i = 0; i < NUM_THREADS; i++)
> ? ?{
> ? ? ? ?argument[i] = i;
> ? ? ? ?if (pthread_create (&th, &defaultattr, print_thread, &argument[i]) != 0)
> ? ? ? ? ? ?error (EXIT_FAILURE, 0, "cannot create thread");
> ? ? ? ?printf ("Thread [%d]\n", i);
> ? ?}
>
> ? ?while (1) {
> ? ? ? ?;
> ? ?}
>
> ? ?/* notreached */
> ? ?return 0;
> }
>
> --------------------------------------------------------------
> J?rgen Schindele
> Software-Entwicklung
>
> NENTEC Netzwerktechnologie GmbH
> Greschbachstr. 12
> 76229 Karlsruhe
> Deutschland
> Telefon: +49 721 94249-51
> Telefax: +49 721 94249-10
> E-Mail: ?schindele at nentec.de
> WEB: ? ? www.nentec.de
>
> Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
> Sitz der Gesellschaft: Karlsruhe
> Handelsregister: Amtsgericht Mannheim HRB 107658
> --------------------------------------------------------------
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>

Problems with pthread on ARM/PXA270

[Juergen Schindele]

Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> Don't remember we had such problems before. What kind of system
> and libraries are you using?
thanks for your reply.
System is custom PXA270 platform.
with codesourcery compiler (see below)
and glibc 2.5 with pthread.
I dont know what more information to give ?  
> 
> 
> On Wed, Sep 16, 2009 at 3:26 PM, Juergen Schindele <schindele@nentec.de> wrote:
> > Hello all,
> > we tried to implement a thread based software on our PXA270 based platform but
> > we had trouble with pthreads. So i modyfied a tiny progam to test. This little progam creates
> > 10 threads and each thread does print something (an index an a pointer to itself).
> >
> > On my PC with Suse 11.0 Linux (kernel 2.6.27) this works fine as supposed to.
> > Each thread has his turn and print his own values.
> >
> > But on our pxa270 platform the thread switch does not (or really rarely) happen
> > and only one thread is printing always the same values. So i suppose thread
> > switching is not working!
> >
> > Any hints or experiences are welcome !!!
> >
> > we use codesourcery compiler
> > arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
> > Copyright (C) 2007 Free Software Foundation, Inc.
> > This is free software; see the source for copying conditions. ?There is NO
> > warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
> >
> > and here the C-code compiled with ?"arm-none-linux-gnueabi-gcc -o test test.c -lpthread"
> >
> > #include <errno.h>
> > #include <error.h>
> > #include <stdio.h>
> > #include <stdlib.h>
> > #include <pthread.h>
> > #include <time.h>
> >
> > #define NUM_THREADS 10
> > #define NUM_ITERS ? 50
> > #define TIMEOUT_NS ?100000000L
> >
> > static void *thread (void *);
> > static int argument[NUM_THREADS];
> > static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;
> > static pthread_attr_t defaultattr;
> >
> > static void * print_thread (void *arg)
> > {
> > ? ?int i;
> > ? ?int number = *((int*)arg);
> >
> > ? ?for (;;) {
> > ? ? ? ?pthread_t self = pthread_self ();
> > ? ? ? ?printf ("Thread %d is [%d]\n", number, self);
> > ? ?}
> > }
> >
> > int main (void)
> > {
> > ? ?pthread_t th;
> > ? ?int i, retvalue;
> >
> > ? ?struct sched_param param;
> > ? ?pthread_attr_init(&defaultattr);
> >
> > ? ?pthread_attr_setschedpolicy(&defaultattr, SCHED_RR);
> >
> > ? ?param.sched_priority = 10;
> > ? ?if ((retvalue = pthread_attr_setschedparam(&defaultattr, &param))!=0) {
> > ? ? ? ?printf ("ERROR\n");
> > ? ? ? ?exit (0);
> > ? ?}
> >
> > ? ?for (i = 0; i < NUM_THREADS; i++)
> > ? ?{
> > ? ? ? ?argument[i] = i;
> > ? ? ? ?if (pthread_create (&th, &defaultattr, print_thread, &argument[i]) != 0)
> > ? ? ? ? ? ?error (EXIT_FAILURE, 0, "cannot create thread");
> > ? ? ? ?printf ("Thread [%d]\n", i);
> > ? ?}
> >
> > ? ?while (1) {
> > ? ? ? ?;
> > ? ?}
> >
> > ? ?/* notreached */
> > ? ?return 0;
> > }
> >
> > --------------------------------------------------------------
> > J?rgen Schindele
> > Software-Entwicklung
> >
> > NENTEC Netzwerktechnologie GmbH
> > Greschbachstr. 12
> > 76229 Karlsruhe
> > Deutschland
> > Telefon: +49 721 94249-51
> > Telefax: +49 721 94249-10
> > E-Mail: ?schindele at nentec.de
> > WEB: ? ? www.nentec.de
> >
> > Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
> > Sitz der Gesellschaft: Karlsruhe
> > Handelsregister: Amtsgericht Mannheim HRB 107658
> > --------------------------------------------------------------
> >
> > _______________________________________________
> > linux-arm-kernel mailing list
> > linux-arm-kernel at lists.infradead.org
> > http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> >
> 
> 



-- 
--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele at nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
 
Diese E-Mail enth?lt vertrauliche oder rechtlich gesch?tzte Informationen.
Wenn Sie nicht der vorgesehene Empf?nger sind, informieren Sie bitte sofort
den Absender und l?schen Sie diese E-Mail. Das unbefugte Kopieren dieser E-Mail
oder die unbefugte Weitergabe der enthaltenen Informationen ist nicht gestattet.
 
The information contained in this message is confidential or protected by law.
If you are not the intended recipient, please contact the sender and delete this
message. Any unauthorised copying of this message or unauthorised distribution of
the information contained herein is prohibited.
--------------------------------------------------------------

Problems with pthread on ARM/PXA270

[Eric Miao]

On Wed, Sep 16, 2009 at 11:01 PM, Juergen Schindele <schindele@nentec.de> wrote:
> Am Mittwoch, 16. September 2009 schrieb Eric Miao:
>> Don't remember we had such problems before. What kind of system
>> and libraries are you using?
> thanks for your reply.
> System is custom PXA270 platform.
> with codesourcery compiler (see below)
> and glibc 2.5 with pthread.
> I dont know what more information to give ?

Like which kernel version, where do you get the glibc-2.5 and pthread binary,
or you compiled it from scratch, and with which compiler of which version?

Problems with pthread on ARM/PXA270

[Juergen Schindele]

Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> On Wed, Sep 16, 2009 at 11:01 PM, Juergen Schindele <schindele@nentec.de> wrote:
> > Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> >> Don't remember we had such problems before. What kind of system
> >> and libraries are you using?
> > thanks for your reply.
> > System is custom PXA270 platform.
> > with codesourcery compiler (see below)
> > and glibc 2.5 with pthread.
> > I dont know what more information to give ?
> 
> Like which kernel version, where do you get the glibc-2.5 and pthread binary,
Sorry forgot that completely !
Kernel version tested is 2.6.16 and 2.6.20 (sorry no newer version available yet)
Compiler , GlibC, pthread all come from Codesourcery precompiled in a package
 "CodeSourcery Sourcery G++ Lite 2007q1-10" downloadable for free.
Compiler version is as mentioned before :
arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
 
--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele at nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
--------------------------------------------------------------

Problems with pthread on ARM/PXA270

[Beckwith, Stephen (Stephen)]

Greetings All,
  I will quickly add the following to this thread, as it seems related.
Platform:
  Custom ARM9 SoC using Montavista Linux Release 3.1, a 2.4.21 Linux Kernel.  Montavista tool chain and libraries.

Problem:
  When we try and execute a pthread_create, "when loaded as a .so", the thread will be created, but will not "return" to the main thread for further processing, we basically get "stuck" in this  new thread.  In the "real" case, this is an ISR thread, would hang on waiting for an interrupt.  For testing, I use Sleep and prints to document program flow.
   However, IF I use the "exact same code" and build the program as an application, it works just fine dandy. . .   The thread gets spawned, prints it's initial hello, then goes into the while (1) sleep/print incrementing value loop.  The thread returns to the "Main" and we spawn 2 more threads and they work just fine (based upon interleaving printouts)

NOTE:
  For our "normal" development we use an external host processor (PowerPC based) to control the SoC and the ability to create threads from a .so has worked in this setup (using same MVL 3.1 kernel/toolchain) for the past 4+ years on 10's of platforms. So we suspected some issue/problem with the MVL setup


Now, we are migrating to a new ARM11 based SoC and I am in the process of re-testing this functionality.  
This new platform will be based upon 2.6.28 kernel release and CodeSourcery G++ Lite 2008q3-72 tools/libraries.
I'm in the process of setting this up on an ARM RealView EB.  

Details to follow.

Regards,
Stephen Beckwith


-----Original Message-----
From: linux-arm-kernel-bounces@lists.infradead.org [mailto:linux-arm-kernel-bounces at lists.infradead.org] On Behalf Of Juergen Schindele
Sent: Thursday, September 17, 2009 3:36 AM
To: linux-arm-kernel at lists.infradead.org
Cc: Eric Miao
Subject: Re: Problems with pthread on ARM/PXA270

Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> On Wed, Sep 16, 2009 at 11:01 PM, Juergen Schindele <schindele@nentec.de> wrote:
> > Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> >> Don't remember we had such problems before. What kind of system
> >> and libraries are you using?
> > thanks for your reply.
> > System is custom PXA270 platform.
> > with codesourcery compiler (see below)
> > and glibc 2.5 with pthread.
> > I dont know what more information to give ?
> 
> Like which kernel version, where do you get the glibc-2.5 and pthread binary,
Sorry forgot that completely !
Kernel version tested is 2.6.16 and 2.6.20 (sorry no newer version available yet)
Compiler , GlibC, pthread all come from Codesourcery precompiled in a package
 "CodeSourcery Sourcery G++ Lite 2007q1-10" downloadable for free.
Compiler version is as mentioned before :
arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
 
--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele at nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
--------------------------------------------------------------

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Problems with pthread on ARM/PXA270

[Juergen Schindele]

Am Donnerstag, 17. September 2009 schrieb Beckwith, Stephen (Stephen):
> Greetings All,
>   I will quickly add the following to this thread, as it seems related.
> Platform:
>   Custom ARM9 SoC using Montavista Linux Release 3.1, a 2.4.21 Linux Kernel.  Montavista tool chain and libraries.
Stephen thanks for your comment.
I tested with Montavista toolchain and 2.6.12 kernel too but without success.
Perhaps this is arm/arch implementation dependant ??? 
> 
> Problem:
>   When we try and execute a pthread_create, "when loaded as a .so", the thread will be created, but will not "return" to the main thread for further processing, we basically get "stuck" in this  new thread.  In the "real" case, this is an ISR thread, would hang on waiting for an interrupt.  For testing, I use Sleep and prints to document program flow.
As you see the example i sent in first mail i build a simple application, so it should work.
Can you please test this little prog on your environment/platform ?
>    However, IF I use the "exact same code" and build the program as an application, it works just fine dandy. . .   The thread gets spawned, prints it's initial hello, then goes into the while (1) sleep/print incrementing value loop.  The thread returns to the "Main" and we spawn 2 more threads and they work just fine (based upon interleaving printouts)
This seems to be dependant of memory location or what ?
What parameters do you use for creating your threads ?
> 
> NOTE:
>   For our "normal" development we use an external host processor (PowerPC based) to control the SoC and the ability to create threads from a .so has worked in this setup (using same MVL 3.1 kernel/toolchain) for the past 4+ years on 10's of platforms. So we suspected some issue/problem with the MVL setup
> 
> 
> Now, we are migrating to a new ARM11 based SoC and I am in the process of re-testing this functionality.  
> This new platform will be based upon 2.6.28 kernel release and CodeSourcery G++ Lite 2008q3-72 tools/libraries.
> I'm in the process of setting this up on an ARM RealView EB.  
> 
> Details to follow.
> 
> Regards,
> Stephen Beckwith
> 
> 
> -----Original Message-----
> From: linux-arm-kernel-bounces at lists.infradead.org [mailto:linux-arm-kernel-bounces at lists.infradead.org] On Behalf Of Juergen Schindele
> Sent: Thursday, September 17, 2009 3:36 AM
> To: linux-arm-kernel at lists.infradead.org
> Cc: Eric Miao
> Subject: Re: Problems with pthread on ARM/PXA270
> 
> Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> > On Wed, Sep 16, 2009 at 11:01 PM, Juergen Schindele <schindele@nentec.de> wrote:
> > > Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> > >> Don't remember we had such problems before. What kind of system
> > >> and libraries are you using?
> > > thanks for your reply.
> > > System is custom PXA270 platform.
> > > with codesourcery compiler (see below)
> > > and glibc 2.5 with pthread.
> > > I dont know what more information to give ?
> > 
> > Like which kernel version, where do you get the glibc-2.5 and pthread binary,
> Sorry forgot that completely !
> Kernel version tested is 2.6.16 and 2.6.20 (sorry no newer version available yet)
> Compiler , GlibC, pthread all come from Codesourcery precompiled in a package
>  "CodeSourcery Sourcery G++ Lite 2007q1-10" downloadable for free.
> Compiler version is as mentioned before :
> arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
>  
> --------------------------------------------------------------
> J?rgen Schindele  
> Software-Entwicklung
> 
> NENTEC Netzwerktechnologie GmbH
> Greschbachstr. 12
> 76229 Karlsruhe
> Deutschland
> Telefon: +49 721 94249-51
> Telefax: +49 721 94249-10
> E-Mail:  schindele at nentec.de
> WEB:     www.nentec.de
>  
> Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
> Sitz der Gesellschaft: Karlsruhe
> Handelsregister: Amtsgericht Mannheim HRB 107658
> --------------------------------------------------------------
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> 
> 



-- 
--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele at nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
 
Diese E-Mail enth?lt vertrauliche oder rechtlich gesch?tzte Informationen.
Wenn Sie nicht der vorgesehene Empf?nger sind, informieren Sie bitte sofort
den Absender und l?schen Sie diese E-Mail. Das unbefugte Kopieren dieser E-Mail
oder die unbefugte Weitergabe der enthaltenen Informationen ist nicht gestattet.
 
The information contained in this message is confidential or protected by law.
If you are not the intended recipient, please contact the sender and delete this
message. Any unauthorised copying of this message or unauthorised distribution of
the information contained herein is prohibited.
--------------------------------------------------------------

Problems with pthread on ARM/PXA270

[Beckwith, Stephen (Stephen)]

Greetings All,
   As noted below, we experienced a problem on Linux 2.4.x kernels for ARM when running pthreads.  
   Thanks to Daniel @ CodeSourcery for assistance in getting my .so to link properly with my application.

My results are:  I do NOT see the same problem as seen on the Linux 2.4.x kernel.  When creating pthreads when calling a .so, they get created correctly and return to the "main" to continue execution.  
The test platform is:  ARM Realview EB with ARM11MPCore CoreTile using the "stock" ARM released 2.6.28 SMP built kernel.

We use real-time threads with FIFO ordering and set the priority as we define in our application so that everything works correctly.  

Regards,
Stephen Beckwith



-----Original Message-----
From: linux-arm-kernel-bounces@lists.infradead.org [mailto:linux-arm-kernel-bounces at lists.infradead.org] On Behalf Of Beckwith, Stephen (Stephen)
Sent: Thursday, September 17, 2009 8:47 AM
To: schindele at nentec.de; linux-arm-kernel at lists.infradead.org
Cc: Eric Miao
Subject: RE: Problems with pthread on ARM/PXA270

Greetings All,
  I will quickly add the following to this thread, as it seems related.
Platform:
  Custom ARM9 SoC using Montavista Linux Release 3.1, a 2.4.21 Linux Kernel.  Montavista tool chain and libraries.

Problem:
  When we try and execute a pthread_create, "when loaded as a .so", the thread will be created, but will not "return" to the main thread for further processing, we basically get "stuck" in this  new thread.  In the "real" case, this is an ISR thread, would hang on waiting for an interrupt.  For testing, I use Sleep and prints to document program flow.
   However, IF I use the "exact same code" and build the program as an application, it works just fine dandy. . .   The thread gets spawned, prints it's initial hello, then goes into the while (1) sleep/print incrementing value loop.  The thread returns to the "Main" and we spawn 2 more threads and they work just fine (based upon interleaving printouts)

NOTE:
  For our "normal" development we use an external host processor (PowerPC based) to control the SoC and the ability to create threads from a .so has worked in this setup (using same MVL 3.1 kernel/toolchain) for the past 4+ years on 10's of platforms. So we suspected some issue/problem with the MVL setup


Now, we are migrating to a new ARM11 based SoC and I am in the process of re-testing this functionality.  
This new platform will be based upon 2.6.28 kernel release and CodeSourcery G++ Lite 2008q3-72 tools/libraries.
I'm in the process of setting this up on an ARM RealView EB.  

Details to follow.

Regards,
Stephen Beckwith


-----Original Message-----
From: linux-arm-kernel-bounces@lists.infradead.org [mailto:linux-arm-kernel-bounces at lists.infradead.org] On Behalf Of Juergen Schindele
Sent: Thursday, September 17, 2009 3:36 AM
To: linux-arm-kernel at lists.infradead.org
Cc: Eric Miao
Subject: Re: Problems with pthread on ARM/PXA270

Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> On Wed, Sep 16, 2009 at 11:01 PM, Juergen Schindele <schindele@nentec.de> wrote:
> > Am Mittwoch, 16. September 2009 schrieb Eric Miao:
> >> Don't remember we had such problems before. What kind of system
> >> and libraries are you using?
> > thanks for your reply.
> > System is custom PXA270 platform.
> > with codesourcery compiler (see below)
> > and glibc 2.5 with pthread.
> > I dont know what more information to give ?
> 
> Like which kernel version, where do you get the glibc-2.5 and pthread binary,
Sorry forgot that completely !
Kernel version tested is 2.6.16 and 2.6.20 (sorry no newer version available yet)
Compiler , GlibC, pthread all come from Codesourcery precompiled in a package
 "CodeSourcery Sourcery G++ Lite 2007q1-10" downloadable for free.
Compiler version is as mentioned before :
arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
 
--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele at nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
--------------------------------------------------------------

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Problems with pthread on ARM/PXA270

[Daniel Jacobowitz]

On Wed, Sep 16, 2009 at 09:26:01AM +0200, Juergen Schindele wrote:
> Hello all,
> we tried to implement a thread based software on our PXA270 based platform but
> we had trouble with pthreads. So i modyfied a tiny progam to test. This little progam creates
> 10 threads and each thread does print something (an index an a pointer to itself). 
> 
> On my PC with Suse 11.0 Linux (kernel 2.6.27) this works fine as supposed to.
> Each thread has his turn and print his own values.
> 
> But on our pxa270 platform the thread switch does not (or really rarely) happen
> and only one thread is printing always the same values. So i suppose thread
> switching is not working!
> 
> Any hints or experiences are welcome !!!

It's just a wild guess, but check your sched_clock implementation.  I
fixed a similar problem in QEMU last week; the kernel used to function
OK with sched_clock always returning zero, but some time between
2.6.25 and 2.6.28 became dependent on it.

-- 
Daniel Jacobowitz
CodeSourcery

Problems with pthread on ARM/PXA270

[Juergen Schindele]

Am Mittwoch, 16. September 2009 schrieb Daniel Jacobowitz:
> On Wed, Sep 16, 2009 at 09:26:01AM +0200, Juergen Schindele wrote:
> > Hello all,
> > we tried to implement a thread based software on our PXA270 based platform but
> > we had trouble with pthreads. So i modyfied a tiny progam to test. This little progam creates
> > 10 threads and each thread does print something (an index an a pointer to itself). 
> > 
> > On my PC with Suse 11.0 Linux (kernel 2.6.27) this works fine as supposed to.
> > Each thread has his turn and print his own values.
> > 
> > But on our pxa270 platform the thread switch does not (or really rarely) happen
> > and only one thread is printing always the same values. So i suppose thread
> > switching is not working!
> > 
> > Any hints or experiences are welcome !!!
> 
> It's just a wild guess, but check your sched_clock implementation.  I
> fixed a similar problem in QEMU last week; the kernel used to function
> OK with sched_clock always returning zero, but some time between
> 2.6.25 and 2.6.28 became dependent on it.
> 
sched_clock is apparently not implemented the same way in all kernel versions.
What about thread implementation in codesourcery compiler ? Are they 
"linux-threads" or "Native Posix thread library" or is it possible that there 
are important bugfixes til the version i use ???

Is my example app. correct in programming or am i doing bloody beginners mistake there ??

Thanks for your help 
--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele at nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
 
Diese E-Mail enth?lt vertrauliche oder rechtlich gesch?tzte Informationen.
Wenn Sie nicht der vorgesehene Empf?nger sind, informieren Sie bitte sofort
den Absender und l?schen Sie diese E-Mail. Das unbefugte Kopieren dieser E-Mail
oder die unbefugte Weitergabe der enthaltenen Informationen ist nicht gestattet.
 
The information contained in this message is confidential or protected by law.
If you are not the intended recipient, please contact the sender and delete this
message. Any unauthorised copying of this message or unauthorised distribution of
the information contained herein is prohibited.
--------------------------------------------------------------

Problems with pthread on ARM/PXA270

[Daniel Jacobowitz]

On Fri, Sep 18, 2009 at 12:06:59PM +0200, Juergen Schindele wrote:
> sched_clock is apparently not implemented the same way in all kernel versions.
> What about thread implementation in codesourcery compiler ? Are they 
> "linux-threads" or "Native Posix thread library" or is it possible that there 
> are important bugfixes til the version i use ???

NPTL in any recent version.  I don't remember for the ancient version
you said you were using; consider an update.

-- 
Daniel Jacobowitz
CodeSourcery

Problems with pthread on ARM/PXA270

[Jamie Lokier]

Juergen Schindele wrote:
> But on our pxa270 platform the thread switch does not (or really
> rarely) happen and only one thread is printing always the same
> values. So i suppose thread switching is not working!

What time quantum does sched_rr_get_interval() return?  SCHED_RR is
only expected to switch that often.

-- Jamie

[arm-gnu] Problems with pthread on ARM/PXA270

[James]

Hi,

On Wed, 2009-09-16 at 09:26 +0200, Juergen Schindele wrote:
> we tried to implement a thread based software on our PXA270 based platform but
> we had trouble with pthreads. So i modyfied a tiny progam to test. This little progam creates
> 10 threads and each thread does print something (an index an a pointer to itself). 
> 
> On my PC with Suse 11.0 Linux (kernel 2.6.27) this works fine as supposed to.
> Each thread has his turn and print his own values.
> 
> But on our pxa270 platform the thread switch does not (or really rarely) happen
> and only one thread is printing always the same values. So i suppose thread
> switching is not working!
> 
> Any hints or experiences are welcome !!!

I'm not a very thread aware programmer so please don't shoot me, but I
was wondering, why would a thread give up control after printing its
values?  Your program has them doing a tight loop with no sleep, or
yield.

I added a usleep(1) just after the printf("Thread...."), and the results
on my Linux Ubuntu-Desktop 2.6.24-24-generic #1 SMP Tue Aug 18 17:04:53
UTC 2009 i686 GNU/Linux are much better, but still not one message per
thread.  I changed the usleep to a sched_yield() and got still better
results, but still not what I'd expect from RR.

So what does it look like on your Suse box?  Do you get one message per
thread in turn?  I.e.
Thread 0 is [...]
Thread 1 is [...]
Thread 2 is [...]
Thread 3 is [...]
etc.?

Or do you get, say, 100 messages from thread 0, then 100 from thread 1,
etc.?

Are you running with superuser privileges?  My manpage tells me that
SCHED_RR is only available if run with superuser privileges.

Out of curiosity, I got rid of the printf in the thread, and added a
global array of ints, as in int counter[NUM_THREADS];  Then made the
threads increment their counter each loop, and got the main while loop
to print the counter values.

Without a sched_yield() or usleep() in the thread loop, the values
increase at a phenomenal rate.  With a sched_yield() the values increase
at about 1/100th the speed.

I once had some interesting problems running a program with just two
threads, one doing I/O to an ISA bus, the other handling a socket
interface, on a dual core machine.  To get more predictable performance,
I ended up building a single threaded program and using select for the
whole lot, thereby implementing my own scheduling policy, effectively.

In any case, I think there's more going on with threads than meets the
eye.  IMHO, they're evil, and should be left for Bill's programmers, who
are sadly bound to using them due to the lack of better ways on that
platform ;-)

Regards,
James.

[arm-gnu] Problems with pthread on ARM/PXA270

[Uwe Kleine-König]

On Thu, Sep 17, 2009 at 08:54:44AM +1000, James wrote:
> Are you running with superuser privileges?  My manpage tells me that
> SCHED_RR is only available if run with superuser privileges.
Note that pthread_attr_setschedpolicy has a return value that tells you
if the function was successful.

Best regards
Uwe

-- 
Pengutronix e.K.                              | Uwe Kleine-K?nig            |
Industrial Linux Solutions                    | http://www.pengutronix.de/  |

[arm-gnu] Problems with pthread on ARM/PXA270

[Jamie Lokier]

Uwe Kleine-K?nig wrote:
> On Thu, Sep 17, 2009 at 08:54:44AM +1000, James wrote:
> > Are you running with superuser privileges?  My manpage tells me that
> > SCHED_RR is only available if run with superuser privileges.
> Note that pthread_attr_setschedpolicy has a return value that tells you
> if the function was successful.

And that SCHED_RR has a time quantum returned by sched_rr_get_interval
which is not guaranteed to be small.  However, if it is a small time
and it's not switching that's a bug in either libc or the kernel.

Note that printf() does locking which may complicate things.  Try
opening separate outputs for each thread, or using write().

-- Jamie

Problems with pthread on ARM/PXA270

[Juergen Schindele]

Am Mittwoch, 16. September 2009 schrieb Juergen Schindele:
Hey all,
i've done some tests on different platforms

running OK on x86 openSuse 11.0
- SCHED_OTHER OK
- SCHED_RR OK


on pxa270 (xscale) platform
compiler gcc 4.2.0
with glibc 2.5 (NPTL) from CodeSourcery
Sourcery G++ Lite 2007q1-10
on Kernel 2.6.20
- SCHED_OTHER works fine
- SCHED_RR does not work at all

on the same pxa270 (xscale) platform
compiler gcc 4.3.3
and glibc 2.8 (NPTL)  from CodeSourcery
Sourcery G++ Lite 2009q1-203
Kernel 2.6.20
- SCHED_OTHER only one thread is running all the time
  other threads are blocked.
- SCHED_RR does not work at all

on ixp425 (xscale) platform
compiler 3.3.2
with glibc 2.2.5 (not NPTL)
from uclinux
kernel 2.6.16
- SCHED_OTHER
First no thread are running. 30 minutes after starting the test all threads
are started and running. It's unbeleivable !!.
-SCHED_RR does not work at all

Why is SCHED_RR only running on x86 platform ???
Where is the difference in handling between SCHED_RR ans SCHED_OTHER ?
Any comments or helps ????

> Hello all,
> we tried to implement a thread based software on our PXA270 based platform but
> we had trouble with pthreads. So i modyfied a tiny progam to test. This little progam creates
> 10 threads and each thread does print something (an index an a pointer to itself). 
> 
> On my PC with Suse 11.0 Linux (kernel 2.6.27) this works fine as supposed to.
> Each thread has his turn and print his own values.
> 
> But on our pxa270 platform the thread switch does not (or really rarely) happen
> and only one thread is printing always the same values. So i suppose thread
> switching is not working!
> 
> Any hints or experiences are welcome !!!
> 
> we use codesourcery compiler
> arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q1-10) 4.2.0 20070413 (prerelease)
> Copyright (C) 2007 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
> 
> and here the C-code compiled with  "arm-none-linux-gnueabi-gcc -o test test.c -lpthread"
> 
> #include <errno.h>
> #include <error.h>
> #include <stdio.h>
> #include <stdlib.h>
> #include <pthread.h>
> #include <time.h>
> 
> #define NUM_THREADS 10
> #define NUM_ITERS   50
> #define TIMEOUT_NS  100000000L
> 
> static void *thread (void *);
> static int argument[NUM_THREADS];
> static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;
> static pthread_attr_t defaultattr;
> 
> static void * print_thread (void *arg)
> {
>     int i;
>     int number = *((int*)arg);
> 
>     for (;;) {
>         pthread_t self = pthread_self ();
>         printf ("Thread %d is [%d]\n", number, self);
>     }
> }
> 
> int main (void)
> {
>     pthread_t th;
>     int i, retvalue;
> 
>     struct sched_param param;
>     pthread_attr_init(&defaultattr);
> 
>     pthread_attr_setschedpolicy(&defaultattr, SCHED_RR);
> 
>     param.sched_priority = 10;
>     if ((retvalue = pthread_attr_setschedparam(&defaultattr, &param))!=0) {
>         printf ("ERROR\n");
>         exit (0);
>     }
> 
>     for (i = 0; i < NUM_THREADS; i++)
>     {
>         argument[i] = i;
>         if (pthread_create (&th, &defaultattr, print_thread, &argument[i]) != 0)
>             error (EXIT_FAILURE, 0, "cannot create thread");
>         printf ("Thread [%d]\n", i);
>     }
> 
>     while (1) {
>         ;
>     }
> 
>     /* notreached */
>     return 0;
> }
> 
> --------------------------------------------------------------
> J?rgen Schindele  
> Software-Entwicklung
> 
> NENTEC Netzwerktechnologie GmbH
> Greschbachstr. 12
> 76229 Karlsruhe
> Deutschland
> Telefon: +49 721 94249-51
> Telefax: +49 721 94249-10
> E-Mail:  schindele at nentec.de
> WEB:     www.nentec.de
>  
> Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
> Sitz der Gesellschaft: Karlsruhe
> Handelsregister: Amtsgericht Mannheim HRB 107658
> --------------------------------------------------------------
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> 
> 



-- 
--------------------------------------------------------------
J?rgen Schindele  
Software-Entwicklung

NENTEC Netzwerktechnologie GmbH
Greschbachstr. 12
76229 Karlsruhe
Deutschland
Telefon: +49 721 94249-51
Telefax: +49 721 94249-10
E-Mail:  schindele at nentec.de
WEB:     www.nentec.de
 
Gesch?ftsf?hrung: Klaus Becker, Roland Knapp
Sitz der Gesellschaft: Karlsruhe
Handelsregister: Amtsgericht Mannheim HRB 107658
--------------------------------------------------------------