Re: [PATCH] lock assertion macros for 2.5.28

[Robert Love <rml@tech9.net>]

On Fri, 2002-07-26 at 10:42, Jesse Barnes wrote:

> Well, I had that in one version of the patch, but people didn't think
> it would be useful.  Maybe you'd like to check out Oliver's comments
> at http://marc.theaimsgroup.com/?l=linux-kernel&m=102644431806734&w=2
> and respond?  If there's demand for MUST_NOT_HOLD, I'd be happy to add
> it since it should be easy.  But if you're using it to enforce lock
> ordering as Oliver suggests, then there are probably more robust
> solutions.

Two other suggestions you could implement are CAN_SLEEP and
CANNOT_SLEEP.  You can implement them via the preempt_count.

Even if CONFIG_PREEMPT is not set, you will get preempt_count values
representing whether or not you are in an interrupt or softirq (and thus
atomic and cannot sleep).  If CONFIG_PREEMPT is set, you get a counter
that represents exactly the atomicity of the code including locks held.

E.g.,

	#define CAN_SLEEP	do { \
		assert(unlikely(!preempt_count())); \
	} while (0)

	#define CANNOT_SLEEP	do { \
		assert(unlikely(preempt_count())); \
	} while (0)

This works great because after the IRQ changes in 2.5.28, preempt_count
is a universal "are we atomic" count.

	Robert Love