Re: sleep_on, wake_up question

Richard B. Johnson (root@chaos.analogic.com)
Thu, 30 Dec 1999 13:01:07 -0500 (EST)

Mark,
I checked a later kernel and it can get stuck. However, you can
use sleep_on_timeout() to make sure you do get awakened. A
check of something the interrupt should have written will let you
know what happened, i.e,...

spin_lock_irqsave(&device_lock, flags);
if(irq_flag)
{
spin_unlock_irqrestore(&device_lock, flags);
//// You already got interrupted, fix this first!

spin_lock_irqsave(&device_lock, flags);
}
irq_flag = 0;
tickle_hardware(); /* Interrupts are off */
spin_unlock_irqrestore(&device_lock, flags);

interruptible_sleep_on_timeout(&wait, time * HZ);
spin_lock_irqsave(&device_lock, flags);
if(irq_flag)
/// Was interrupted
else
/// Was not
spin_unlock_irqrestore(&device_lock, flags);

isr()
{
irq_flag = 1;
wake_up_interruptible(&wait);
}

Just check program flow under a spin-lock and it should be safe
in SMP. Keep the code under the locks short. I have a GPIB driver
which will have to be modifed once a new kernel is released. It
does about 100,000 interrupts per second without losing data, but
it won't even compile with a new kernel because some types have
changed. So I haven't done anything with it recently.

Cheers,
Dick Johnson

Penguin : Linux version 2.3.13 on an i686 machine (400.59 BogoMips).
Warning : The end of the world as we know it requires a new calendar.
Seconds : 125933 (until Y2K)

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/