Re: [PATCH] Linux: Implement membarrier function

[Alan Stern <stern@rowland.harvard.edu>]

On Thu, 13 Dec 2018, Paul E. McKenney wrote:

> > > A good next step would be to automatically generate random tests along
> > > with an automatically generated prediction, like I did for RCU a few
> > > years back.  I should be able to generalize my time-based cheat for RCU to
> > > also cover SRCU, though sys_membarrier() will require a bit more thought.
> > > (The time-based cheat was to have fixed duration RCU grace periods and
> > > RCU read-side critical sections, with the grace period duration being
> > > slightly longer than that of the critical sections.  The number of
> > > processes is of course limited by the chosen durations, but that limit
> > > can easily be made insanely large.)
> > 
> > Imagine that each sys_membarrier call takes a fixed duration and each 
> > other instruction takes slightly less (the idea being that each 
> > instruction is a critical section).  Instructions can be reordered 
> > (although not across a sys_membarrier call), but no matter how the 
> > reordering is done, the result is disallowed.

This turns out not to be right.  Instead, imagine that each 
sys_membarrier call takes a fixed duration, T.  Other instructions can 
take arbitrary amounts of time and can be reordered abitrarily, with 
two restrictions:

	Instructions cannot be reordered past a sys_membarrier call;

	If instructions A and B are reordered then the time duration
	from B to A must be less than T.

If you prefer, you can replace the second restriction with something a 
little more liberal:

	If A and B are reordered and A ends up executing after a 
	sys_membarrier call (on any CPU) then B cannot execute before 
	that sys_membarrier call.

Of course, this form is a consequence of the more restrictive form.

> It gets a bit trickier with interleavings of different combinations
> of RCU, SRCU, and sys_membarrier().  Yes, your cat code very elegantly
> sorts this out, but my goal is to be able to explain a given example
> to someone.

I don't think you're going to be able to fit different combinations of
RCU, SRCU, and sys_membarrier into this picture.  How would you allow
tests with incorrect interleaving, such as GP - memb - RSCS - nothing,
while forbidding similar tests with correct interleaving?

Alan