Re: [tip:core/locking] x86/smp: Move waiting on contended ticket lockout of line

From: Linus Torvalds
Date: Wed Feb 27 2013 - 22:19:44 EST


On Wed, Feb 27, 2013 at 6:58 PM, Rik van Riel <riel@xxxxxxxxxx> wrote:
>
> On the other hand, both MCS and the fast queue locks
> implemented by Michel showed low variability and high
> performance.

On microbenchmarks, and when implemented for only one single subsystem, yes.

> The numbers for Michel's MCS and fast queue lock
> implementations appear to be both fast and stable.

I do think that doing specialized spinlocks for special areas may be a
rasonable approach, and it's quite possible that the SySV IPC thing is
one such area.

But no, I don't think the numbers I've seen for Michel's MCS are AT
ALL comparable to the generic spinlocks, and the interface makes them
incompatible as a replacement to even test in general.

Don't get me wrong: I think the targeted approach is *better*. I just
also happen to think that you spout big words about things that aren't
all that big, and try to make this a bigger deal than it is. The
benchmark numbers you point to are micro-benchmarks, and not all
comparable.

Linus
--
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/