From: Paul Mackerras <paulus@samba.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <16056.34210.319959.255815@argo.ozlabs.ibm.com>
Date: Wed, 7 May 2003 14:03:46 +1000
To: William Lee Irwin III <wli@holomorphy.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>, Andrew Morton <akpm@digeo.com>,
       dipankar@in.ibm.com, linux-kernel@vger.kernel.org,
       Bartlomiej Zolnierkiewicz <B.Zolnierkiewicz@elka.pw.edu.pl>
Subject: Re: [PATCH] kmalloc_percpu
In-Reply-To: <20030507024135.GW8978@holomorphy.com>
References: <20030506014745.02508f0d.akpm@digeo.com>
	<20030507023126.12F702C019@lists.samba.org>
	<20030507024135.GW8978@holomorphy.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 926
Lines: 20

William Lee Irwin III writes:
> On Wed, May 07, 2003 at 11:57:13AM +1000, Rusty Russell wrote:
> > Paul Mackerras points out that we could get the numa-aware allocation
> > plus "one big alloc" properties by playing with page mappings: reserve
> > 1MB of virtual address, and map more pages as required.  I didn't
> > think that we'd need that yet, though.
> 
> This is somewhat painful to do (though possible) on i386. The cost of
> task migration would increase at the very least.

Explain?  We're not talking about having the same address mapped
differently on different CPUs, we're just talking about something
which is the equivalent of a sparsely-instantiated vmalloc.

Paul.
-
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/