Pierre PEIFFER wrote:
> Destroying a rt-ipc fifo on which a task is waiting causes a
> segmentation fault;
>
I experienced a similar problem with a real-time task waiting for a
semaphore to be posted before it incremented a counter in an endless loop.
>
> In the init_module:
> I create an rt-ipc fifo;
> I initialize one task;
> I start the task;
>
> What the task does:
> only waits on some data from the created fifo (rtf_receive with timeout
> parameter equal to RT_WAIT_FOREVER). When available, it prints the data
> on the screen and still waits for other data. (in a loop)
>
> In the cleanup_module:
> I delete the task -> done successfully.
> I destroy the fifo -> this causes a segmentation fault, i.e. the call
> to rtf_ipc_destroy crashes, with some stack trace (?) on the screen.
>
I had to include in my real-time task a boolean variable to exit the loop.
After the loop, I execute a rt_task_suspend() to put the task to sleep.
volatile *counter;
InterruptHandler()
{
while(working) {
rt_sem_wait(.......);
if (working) {
(*counter)++;
}
}
rt_task_suspend(MYSELF);
}
cleanup_module()
{
working = 0;
rt_sem_post(.....); // That unblocks the real-time task
rt_sem_destroy(....);
rt_task_ipc_delete(...); // Needed because I use semaphores
}
>
> Bug ?
> Something I've forgot to do ?
> A way to avoid this ?
>
I don't know either if it is a bug....
>
> Thanks !
>
> Pierre
Hope this will help you until a better solution is found !
Stephane Ouellette
Tektrend International inc.
>
> --- [rtl] ---
> To unsubscribe:
> echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
> echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
> ----
> For more information on Real-Time Linux see:
> http://www.rtlinux.org/~rtlinux/
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
