Dear Sir
I thing, that I find in new GDB 6.8 new bug.
example :
#include <pthread.h>
#include <stdio.h>
#include <unistd.h>
static pthread_mutex_t mutex;
void* print_pokus(void *unused)
{
while(1)
{
pthread_mutex_lock(&mutex);
printf(" %s\n", (char *)unused);
pthread_mutex_unlock(&mutex);
}
return(NULL);
}
int main()
{
pthread_t thread_id1;
pthread_t thread_id2;
pthread_t thread_id3;
pthread_mutex_init(&mutex, NULL);
pthread_create(&thread_id1, NULL, print_pokus, "1");
pthread_create(&thread_id2, NULL, print_pokus, "2");
pthread_create(&thread_id3, NULL, print_pokus, "3");
while(1)
{
sleep(1);
}
return(0);
}
I compile source code :
gcc -Wall -pedantic -g -O0 thread.c -lpthread
ulimit -c unlimited
gdb ./a.out
(gdb) break print_pokus
Breakpoint 1 at 0x80484fa: file thread.c, line 12.
(gdb) run
Starting program: /home/dusan/tuxanci-ng-svn/a.out
[Thread debugging using libthread_db enabled]
[New Thread 0xb7dfaad0 (LWP 19310)]
[New Thread 0xb7df9b90 (LWP 19313)]
(gdb) continue
[ press key ENTER 10 second ]
GDB remember last command. I press key ENTER more or less 10 second,
then a use commad step same 10 second and debufer`s program down.
(gdb) step
[ press key ENTER 10 second ]
I analyze coredump :
(gdb) disassemble
0x080484f4 <print_pokus+0>: push %ebp
0x080484f5 <print_pokus+1>: mov %esp,%ebp
0x080484f7 <print_pokus+3>: sub $0x8,%esp
0x080484fa <print_pokus+6>: int3
0x080484fb <print_pokus+7>: add $0x24,%al
0x080484fd <print_pokus+9>: int3
0x080484fe <print_pokus+10>: xchg %eax,%edi
0x080484ff <print_pokus+11>: add $0x8,%al
0x08048501 <print_pokus+13>: call 0x8048440 <[EMAIL PROTECTED]>
0x08048506 <print_pokus+18>: int3
0x08048507 <print_pokus+19>: inc %ebp
0x08048508 <print_pokus+20>: or %cl,-0x38fbdbbc(%ecx)
0x0804850e <print_pokus+26>: add $0x24,%al
0x08048510 <print_pokus+28>: cwtl
0x08048511 <print_pokus+29>: xchg %al,(%eax,%ecx,1)
0x08048514 <print_pokus+32>: call 0x8048420 <[EMAIL PROTECTED]>
0x08048519 <print_pokus+37>: movl $0x80497cc,(%esp)
0x08048520 <print_pokus+44>: call 0x8048410 <[EMAIL PROTECTED]>
0x08048525 <print_pokus+49>: jmp 0x80484fa <print_pokus+6>
I see, that GDB-6.8 abandon in debufer`s program int3 instruction and hereat
program down.
GDB abandon in debufer`s program int3 instruction in function, which share more
threads.
best regarts
PS:
When send this email, My down email client, hereat send email even now.
Dušan Ďurech - Slowakia Republic
_______________________________________________
bug-gdb mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-gdb
