Re: PATCH 2.3.26: kmalloc GFP_ZERO
yodaiken@chelm.cs.nmt.edu
Tue, 9 Nov 1999 13:08:10 -0700
On Tue, Nov 09, 1999 at 12:48:49PM -0800, David S. Miller wrote:
> From: yodaiken@chelm.cs.nmt.edu
> Date: Tue, 9 Nov 1999 12:55:58 -0700
>
> > the idle task will be less fast in rescheduling itself.
>
> Why?
>
> He's trying to say that if you're zero'ing pages in the idle loop, the
> latency between checks of current->need_resched will be longer. I
> think he's overestimating the cost of zero'ing a single page.
There is no reason to not check after writing each cacheline, if that's a worry.
And I think this is a case of worrying about the wrong end of the optimization.
The startup costs of memseting a big chunk of memory are significant and do trash
the cache.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/