On 9/21/05, Fawad Lateef <fawadlateef@xxxxxxxxx> wrote:
> On 9/21/05, raja <vnagaraju@xxxxxxxxxxxx> wrote:
> > Hi,
> > I am trying to trace the race condition by creating 2 threads in
> > kernel space. One thread incriments a global value and other
> > decrements that value.
> > But I am unable to trace the race condition.
> > I am listing my code.Will you please help me.
> >
> >
> > static int __init init_module_atomic(void);
> > static void __exit exit_module_atomic(void);
> >
>
> No need for the above declaration ..........
>
> > int value=0;
>
> Can change this to unsigned long value .........
>
> >
> > int thread_function_1(void * data)
> > {
> > printk("Entered into Thread 1\n");
> > int i;
> > for(i=0;i<1000;i++)
> > {
>
> Can change this loop to for( ; ; ) to get the result more properly
>
> > udelay(1000);
>
> Change this to schedule();
>
> > value++;
> > printk("Value Is : %d\t",value);
> > }
> > printk("Exited From Thread 1\n");
> > return 0;
> > }
> >
> > int thread_function_2(void * data)
> > {
> > printk("Entered into Thread 2\n");
> > int i;
> > for(i=0;i<1000;i++)
> > {
>
> Change to for ( ; ; )
>
> > udelay(100);
>
> change to schedule()
>
> > value--;
> > printk("Value Is : %d\t",value);
> > }
> > printk("Exited From Thread 2\n");
> > return 0;
> > }
> >
>
> In your code you won't be able to see just as your thread_1 terminates
> quickly before starting thread_2 and udelay is stalling the whole CPU
> and won't allow CPU to do other stuffs, so changing it to schedule()
> will give you better result ...........
>
And also do kill/send signals to your thread for killing ........ so
that it wont result in seg-fault if you remove your module ........
for the case mentioned above by me .....
--
Fawad Lateef
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
