Re: [PATCH 1/2] scalable rw_mutex

[Oleg Nesterov <oleg@tv-sign.ru>]

On 05/12, Peter Zijlstra wrote:
>
> On Sat, 2007-05-12 at 20:04 +0400, Oleg Nesterov wrote:
> > 
> > this code roughly does (the only reader does unlock)
> > 
> > 	READER			WRITER
> > 
> > 	readers = 0;		state = 1;
> > 	wmb();			wmb();
> > 	CHECK(state != 0)	CHECK(readers == 0)
> > 
> > We need to ensure that we can't miss both CHECKs. Either reader
> > should see RW_MUTEX_READER_SLOW, o writer sees "readers == 0"
> > and does not sleep.
> > 
> > In that case both barriers should be converted to smp_mb(). There
> > was a _long_ discussion about STORE-MB-LOAD behaviour, and experts
> > seem to believe everething is ok.
> 
> Ah, but note that both those CHECK()s have a rmb(), so that ends up
> being:
> 
> 	READER				WRITER
> 
> 	readers = 0;			state = 1;
> 	wmb();				wmb();
> 
> 	rmb();				rmb();		
> 	if (state != 0)			if (readers == 0)
> 
> and a wmb+rmb is a full mb, right?

I used to think the same, but this is wrong: wmb+rmb != mb. wmb+rmb
doesn't provide LOAD,STORE or STORE,LOAD ordering.

for example,

	LOAD;
	rmb(); wmb();
	STORE;

it is still possible that STORE comes before LOAD. At least this
is my understanding.

Oleg.

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