Re: [PATCH v4] introduce macro spin_event_timeout()

[Timur Tabi <timur@freescale.com>]

Alan Cox wrote:
>> Are you talking about the udelay() inside the loop?  If so, I agree
>> that this is bad and have removed it in the PowerPC-specific version:
> 
> The behaviour you want there is system specific - 10uS is a minimum
> politeness value for x86 PCI bus for example.

So we need to allow for delays between successive rights?  We can
provide that with a third parameter to the macro.

>> rdtsc instruction.  In this case, we're not adding arbitrary delays
>> into the loop, and we're not using jiffies, but we are
>> architecture-dependent.
> 
> and not useful

Is there an architecture-independent method for reading a timebase
register that's not jiffies?

> A macro of this form really needs to be able to look like
> 
> 	spin_until_timeout(readb(foo) & 0x80, 30 * HZ) {
> 		udelay(10);
> 		/* Maybe do other stuff */
> 	}
> 
> to be more generally useful

You mean something like this:

#define spin_until_timeout(condition, timeout) 		\
	for (unsigned long __timeout = jiffies + (timeout);	\
		(!(condition) && time_after(jiffies, __timeout)); )

How do I return a value indicating whether a timeout occurred or
condition came true?

-- 
Timur Tabi
Linux kernel developer at Freescale