From realluca76 at yahoo.it  Thu Mar  8 09:16:18 2007
From: realluca76 at yahoo.it (realLuca76 bahbihboh)
Date: Thu, 8 Mar 2007 10:16:18 +0100 (CET)
Subject: threads sync with heavy CPU load
Message-ID: <96191.81416.qm@web23406.mail.ird.yahoo.com>

Hi.

I have to synchronize two threads (T1 and T2) in this
way: T1 does a job and meanwhile T2 waits, then T2
does a job and meanwhile T1 waits, then T1 does a job
and meanwhile T1 waits and so on... (loop).
I have used two approaches:

1) pthread_mutex_lock + pthread_cond_wait +
pthread_cond_signal + pthread_mutex_unlock (in both
threads)
Here's the code for the example:

T1 increments a counter and does some additional CPU
heavy work while T2 is blocked. then it (u)sleeps. T2
does the same.
Well: if both threads sleep for 1 second after their
operations all seems to work. If they (usleep) for
10000 us, for example (or if they don't sleep at all),
the synchronization doesn't work properly. The value
of pthread_cond_wait is uncorrect.

Should i have to add real time extensions to my code
in order to solve the problem?
Any help is GREATLY appreciated,

Luca

---------------------------------------------------------------------------------------------------------------


#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
#include <iostream>
#include <pthread.h>

pthread_mutex_t my_mutex;
pthread_cond_t my_condition;
bool condizione=false;
int shared_counter = 0;

void* routineT1(void* obj)
{
	while(1)
 	{
		int a = pthread_mutex_lock(&my_mutex);
		int b = 145; //dummy default		
		while(condizione) b =
pthread_cond_wait(&my_condition, &my_mutex);	
		if(shared_counter < 10) shared_counter ++;
		else shared_counter = 1;
		std::cerr<< "\n T1: I do my job ( I increment the
counter: "<<shared_counter<<" ) and T2 waits";
		//some dummy and HEAVY work
		for(int q = 0; q< 10000000; q++) 
		{
			q = q + 1;
			q--;
			int f = q-2;
			f++;
		}
		usleep(1000000);
		condizione = true;
		int c = pthread_cond_signal(&my_condition);	
		int d = pthread_mutex_unlock(&my_mutex);	
		std::cerr<<"\n T1: mutex_lock ,cond_wait
,cond_signal , mutex_unlock:
"<<a<<","<<b<<","<<c<<","<<d; 
	}
};

void* routineT2(void* obj)
{
	while(1)
	{
		int a = pthread_mutex_lock(&my_mutex);	
		int b = 146; //dummy default
		while(!condizione) b =
pthread_cond_wait(&my_condition, &my_mutex);
		std::cerr<< "\n T2: I do my job ( I read the
counter: "<<shared_counter<<" ) and T1 waits ";
		//some dummy and HEAVY work
		for(int q = 0; q< 10000000; q++) 
		{
			q = q + 1;
			q--;
			int f = q-2;
			f++;
		}
		usleep(1000000);
		condizione = false;
		int c = pthread_cond_signal(&my_condition);
		int d = pthread_mutex_unlock(&my_mutex);
		std::cerr<<"\n T2: mutex_lock ,cond_wait
,cond_signal , mutex_unlock:
"<<a<<","<<b<<","<<c<<","<<d; 
		std::cerr<<"\n *";
		std::cerr<<"\n *";
	}
};



int main(int argc, char *argv[])
{
  pthread_t* tt1 =  new pthread_t;
  pthread_t* tt2 = new pthread_t; 
  pthread_mutex_init(&my_mutex,NULL);
  pthread_cond_init(&my_condition,NULL);
  pthread_create(tt1,NULL,routineT1,NULL);
  sleep(1);
  pthread_create(tt2,NULL,routineT2,NULL);
  while(1);
}


	

	
		
___________________________________ 
L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail: 
http://it.docs.yahoo.com/nowyoucan.html


From schabi at logix-tt.com  Fri Mar 16 13:00:18 2007
From: schabi at logix-tt.com (Markus Schaber)
Date: Fri, 16 Mar 2007 14:00:18 +0100
Subject: pthreads and operator new
Message-ID: <20070316140018.5e86a7ea@kingfisher.sec.intern.logix-tt.com>

Hi,

I did not find a definitive source on the net (due to lots of false
positives when googling), but I found this list that seems to discuss
C++ / pthread library issues.

My question is simple: Does the combination of C++98 and Posix pthread
standards give any guarantee that the operator new is "thread safe",
means that several threads can allocate objects via "new" concurrently
without explicit locking by the user code? 

Our current research seems to indicate that pthread standards don't
mention C++ in any way, and C++ does not mention threads, and it is
pure benevolence (or plain sanity?) of the implementations I have seen
up to now to provide a thread safe implementation of new (mostly based
on a global lock).

Some pointers to mandatory sources or better places to discuss this
issue are welcome.

Thanks a lot,
Markus

-- 
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf.     | Software Development GIS

Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org


From hinnant at twcny.rr.com  Fri Mar 16 20:55:52 2007
From: hinnant at twcny.rr.com (Howard Hinnant)
Date: Fri, 16 Mar 2007 16:55:52 -0400
Subject: [c++-pthreads] pthreads and operator new
In-Reply-To: <20070316140018.5e86a7ea@kingfisher.sec.intern.logix-tt.com>
References: <20070316140018.5e86a7ea@kingfisher.sec.intern.logix-tt.com>
Message-ID: <39F6AA06-F932-4577-8C29-5040D37EF4E4@twcny.rr.com>

On Mar 16, 2007, at 9:00 AM, Markus Schaber wrote:

> My question is simple: Does the combination of C++98 and Posix pthread
> standards give any guarantee that the operator new is "thread safe",
> means that several threads can allocate objects via "new" concurrently
> without explicit locking by the user code?
>
> Our current research seems to indicate that pthread standards don't
> mention C++ in any way, and C++ does not mention threads, and it is
> pure benevolence (or plain sanity?) of the implementations I have seen
> up to now to provide a thread safe implementation of new (mostly based
> on a global lock).

Your research is correct.  Though you might also add "market value" to  
your list of reasons.  C++ vendors would otherwise go out of business.

The C++ committee is attempting to address this shortcoming in an  
upcoming standard colloquially termed C++0X (we hope the X will become  
a 9).

-Howard



From mohsen at pahlevanzadeh.org  Sat Mar 24 09:51:53 2007
From: mohsen at pahlevanzadeh.org (Mohsen Pahlevanzadeh)
Date: Sat, 24 Mar 2007 05:51:53 -0400
Subject: an error on Debian 
Message-ID: <4604F4B9.6020406@pahlevanzadeh.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear all,
I wrote following code :
#include <stdio.h>
#include <stdlib.h>
#include <pthread.h>
int g1=0;
int g2=0;
void *task1(int *counter);
//void task2(int *counter);
void cleanup(int counter1,int counter2);
int main(int argc,char *argv[])
{
        pthread_t thr1,thr2;

        if ( pthread_create(&thr1,NULL,task1,(void*)g1))
        {
                perror("pthread_create : task1");
                exit(EXIT_FAILURE);
        }

/*      if ((ret = pthread_create(&thr2,NULL,task2,(void *)&g2)))
        {
                perror("pthread_create : task2");
                exit(EXIT_FAILURE);
        }*/


//      pthread_join(thr2,NULL);
        pthread_join(thr1,NULL);

        cleanup(g1,g2);
        exit(EXIT_SUCCESS);
}//end of main program
void *task1(int *counter)
{
        while(*counter < 5 ){
                printf("task1 count: %d\n",*counter);
                (*counter)++;
        }//end of while
void cleanup(int counter1,int counter2)
{
        printf("Total iterations: %d\n",counter1+counter2);
}//end of cleanup function



But i receive following error:


mohsen at debian:~/test/learning/pthread$ make
gcc -c -Wall test.c -lpthread
In file included from test.c:3:
/usr/include/pthread.h:285: error: conflicting types for ?pthread_t?
/usr/include/bits/pthreadtypes.h:152: error: previous declaration of
?pthread_t? was here
/usr/include/pthread.h:286: error: conflicting types for ?pthread_attr_t?
/usr/include/bits/pthreadtypes.h:54: error: previous declaration of
?pthread_attr_t? was here
/usr/include/pthread.h:287: error: conflicting types for ?pthread_key_t?
/usr/include/bits/pthreadtypes.h:82: error: previous declaration of
?pthread_key_t? was here
/usr/include/pthread.h:289: error: conflicting types for
?pthread_mutexattr_t?
/usr/include/bits/pthreadtypes.h:102: error: previous declaration of
?pthread_mutexattr_t? was here
/usr/include/pthread.h:290: error: conflicting types for ?pthread_mutex_t?
/usr/include/bits/pthreadtypes.h:95: error: previous declaration of
?pthread_mutex_t? was here
/usr/include/pthread.h:291: error: conflicting types for
?pthread_condattr_t?
/usr/include/bits/pthreadtypes.h:79: error: previous declaration of
?pthread_condattr_t? was here
/usr/include/pthread.h:292: error: conflicting types for ?pthread_cond_t?
/usr/include/bits/pthreadtypes.h:72: error: previous declaration of
?pthread_cond_t? was here
test.c: In function ?main?:
test.c:13: warning: passing argument 3 of ?pthread_create? from
incompatible pointer type
test.c:11: warning: unused variable ?thr2?
test.c: In function ?task1?:
test.c:38: warning: control reaches end of non-void function
make: *** [main.o] Error 1
mohsen at debian:~/test/learning/pthread$

My distro is Debian Etch.
kernel is 2.6.18
Please help me....
Yours,Mohsen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGBPS5/ZBAvBh9bHIRAmTtAJ4pcg3fHT5Iq3ootfiAWaGFwb94iQCfdKM4
o+gfENU0CfgL1Xk/NX8Lufw=
=fLVY
-----END PGP SIGNATURE-----


From mohsen at pahlevanzadeh.org  Sat Mar 24 09:49:10 2007
From: mohsen at pahlevanzadeh.org (Mohsen Pahlevanzadeh)
Date: Sat, 24 Mar 2007 05:49:10 -0400
Subject: an error on pthread
Message-ID: <4604F416.1050306@pahlevanzadeh.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear all,
I wrote following code :
#include <stdio.h>
#include <stdlib.h>
#include <pthread.h>
int g1=0;
int g2=0;
void *task1(int *counter);
//void task2(int *counter);
void cleanup(int counter1,int counter2);
int main(int argc,char *argv[])
{
        pthread_t thr1,thr2;

        if ( pthread_create(&thr1,NULL,task1,(void*)g1))
        {
                perror("pthread_create : task1");
                exit(EXIT_FAILURE);
        }

/*      if ((ret = pthread_create(&thr2,NULL,task2,(void *)&g2)))
        {
                perror("pthread_create : task2");
                exit(EXIT_FAILURE);
        }*/


//      pthread_join(thr2,NULL);
        pthread_join(thr1,NULL);

        cleanup(g1,g2);
        exit(EXIT_SUCCESS);
}//end of main program
void *task1(int *counter)
{
        while(*counter < 5 ){
                printf("task1 count: %d\n",*counter);
                (*counter)++;
        }//end of while
void cleanup(int counter1,int counter2)
{
        printf("Total iterations: %d\n",counter1+counter2);
}//end of cleanup function



But i receive following error:


mohsen at debian:~/test/learning/pthread$ make
gcc -c -Wall test.c -lpthread
In file included from test.c:3:
/usr/include/pthread.h:285: error: conflicting types for ?pthread_t?
/usr/include/bits/pthreadtypes.h:152: error: previous declaration of
?pthread_t? was here
/usr/include/pthread.h:286: error: conflicting types for ?pthread_attr_t?
/usr/include/bits/pthreadtypes.h:54: error: previous declaration of
?pthread_attr_t? was here
/usr/include/pthread.h:287: error: conflicting types for ?pthread_key_t?
/usr/include/bits/pthreadtypes.h:82: error: previous declaration of
?pthread_key_t? was here
/usr/include/pthread.h:289: error: conflicting types for
?pthread_mutexattr_t?
/usr/include/bits/pthreadtypes.h:102: error: previous declaration of
?pthread_mutexattr_t? was here
/usr/include/pthread.h:290: error: conflicting types for ?pthread_mutex_t?
/usr/include/bits/pthreadtypes.h:95: error: previous declaration of
?pthread_mutex_t? was here
/usr/include/pthread.h:291: error: conflicting types for
?pthread_condattr_t?
/usr/include/bits/pthreadtypes.h:79: error: previous declaration of
?pthread_condattr_t? was here
/usr/include/pthread.h:292: error: conflicting types for ?pthread_cond_t?
/usr/include/bits/pthreadtypes.h:72: error: previous declaration of
?pthread_cond_t? was here
test.c: In function ?main?:
test.c:13: warning: passing argument 3 of ?pthread_create? from
incompatible pointer type
test.c:11: warning: unused variable ?thr2?
test.c: In function ?task1?:
test.c:38: warning: control reaches end of non-void function
make: *** [main.o] Error 1
mohsen at debian:~/test/learning/pthread$

My distro is Debian Etch.
kernel is 2.6.18
Please help me....
Yours,Mohsen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGBPQW/ZBAvBh9bHIRArCkAJ9f3DNW9BtejrB5DRJZijkoeCiGLQCfTQ9n
SrK6gggZbAHQ/YOtV2TeBcA=
=6EVs
-----END PGP SIGNATURE-----


From ben at decadent.org.uk  Sat Mar 24 18:38:57 2007
From: ben at decadent.org.uk (Ben Hutchings)
Date: Sat, 24 Mar 2007 18:38:57 +0000
Subject: [c++-pthreads] an error on Debian
In-Reply-To: <4604F4B9.6020406@pahlevanzadeh.org>
References: <4604F4B9.6020406@pahlevanzadeh.org>
Message-ID: <1174761537.11281.10.camel@deadeye.i.decadent.org.uk>

Answered off-list.

Ben.

-- 
Ben Hutchings
Horngren's Observation:
                   Among economists, the real world is often a special case.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://sourcerytools.com/pipermail/c++-pthreads/attachments/20070324/27877e39/attachment.pgp>

From mohsen at pahlevanzadeh.org  Sun Mar 25 07:50:31 2007
From: mohsen at pahlevanzadeh.org (Mohsen Pahlevanzadeh)
Date: Sun, 25 Mar 2007 03:50:31 -0400
Subject: [c++-pthreads] an error on Debian
In-Reply-To: <1174761537.11281.10.camel@deadeye.i.decadent.org.uk>
References: <4604F4B9.6020406@pahlevanzadeh.org> <1174761537.11281.10.camel@deadeye.i.decadent.org.uk>
Message-ID: <460629C7.5080708@pahlevanzadeh.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ben Hutchings wrote:
> Answered off-list.
> 
> Ben.
> 
I can't understand your purpose.
Mohsen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGBinH/ZBAvBh9bHIRAicJAJ4lY38gaEt1N/en7Ao5t257esJpiACeKUDN
YsfmPLRgcq1RL2oi7ZS9prU=
=Tq0S
-----END PGP SIGNATURE-----