Re: [patch 0/3] j_state_lock, j_list_lock, remove-bitlocks

[Ingo Molnar]

* Andrew Morton <akpm@xxxxxxxx> wrote:

> Ingo Molnar <mingo@xxxxxxx> wrote:
> >
> >  > There's a little lock ranking diagram in jbd.h which tells us that
> >  > these locks nest inside j_list_lock and j_state_lock.  So I guess
> >  > you'll need to turn those into semaphores.
> > 
> >  indeed. I did this (see the three followup patches, against BK-curr),
> >  and it builds/boots/works just fine on an ext3 box. Do we want to try
> >  this in -mm?
> 
> ooh, I'd rather not.  I spent an intense three days removing all the
> sleeping locks from ext3 (and three months debugging the result). 
> Ended up gaining 1000% on 16-way.
> 
> Putting them back in will really hurt the SMP performance.

ah. Yeah. Sniff.

if we gain 1000% on a 16-way then there's something really wrong about
semaphores (or scheduling) though. A semaphore is almost a spinlock, in
the uncontended case - and even under contention we really (should) just
spend the cycles that we'd spend spinning. There will be some
intermediate contention level where semaphores hurt, but 1000% sounds
truly excessive.

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