From mboxrd@z Thu Jan 1 00:00:00 1970 Received: with ECARTIS (v1.0.0; list linux-mips); Mon, 18 Feb 2008 21:28:27 +0000 (GMT) Received: from localhost.localdomain ([127.0.0.1]:63720 "EHLO dl5rb.ham-radio-op.net") by ftp.linux-mips.org with ESMTP id S28574393AbYBRV2Z (ORCPT ); Mon, 18 Feb 2008 21:28:25 +0000 Received: from denk.linux-mips.net (denk.linux-mips.net [127.0.0.1]) by dl5rb.ham-radio-op.net (8.14.1/8.13.8) with ESMTP id m1ILSNMc010480; Mon, 18 Feb 2008 21:28:24 GMT Received: (from ralf@localhost) by denk.linux-mips.net (8.14.1/8.14.1/Submit) id m1ILSNsf010479; Mon, 18 Feb 2008 21:28:23 GMT Date: Mon, 18 Feb 2008 21:28:23 +0000 From: Ralf Baechle To: Thiemo Seufer Cc: linux-mips@linux-mips.org Subject: Re: [PATCH] Reimplement clear_page/copy_page Message-ID: <20080218212823.GA10420@linux-mips.org> References: <20080218193249.GD4747@networkno.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20080218193249.GD4747@networkno.de> User-Agent: Mutt/1.5.17 (2007-11-01) Return-Path: X-Envelope-To: <"|/home/ecartis/ecartis -s linux-mips"> (uid 0) X-Orcpt: rfc822;linux-mips@linux-mips.org Original-Recipient: rfc822;linux-mips@linux-mips.org X-archive-position: 18260 X-ecartis-version: Ecartis v1.0.0 Sender: linux-mips-bounce@linux-mips.org Errors-to: linux-mips-bounce@linux-mips.org X-original-sender: ralf@linux-mips.org Precedence: bulk X-list: linux-mips On Mon, Feb 18, 2008 at 07:32:49PM +0000, Thiemo Seufer wrote: > Fold the SB-1 specific implementation of clear_page/copy_page in the > generic version, and rewrite that one in tlbex style. The immediate > benefits: > - It converts the compile-time workaround for SB-1 pass 1 prefetches > to a more efficient run-time check. > - It allows adjustment of loop unfolling, which helps to reduce the > number of redundant cdex cache ops. > - It fixes some esoteric cornercases (the cache line length calculations > can go wrong, and support for 64k pages without prefetch instructions > will overflow the addiu immediate). > - Somewhat better guesses of "good" prefetch values. > > > Signed-off-by: Thiemo Seufer > --- > > Lmbench3 running on a BCM1480 system shows improvements for some > benchmarks (three runs with the original kernel, then three runs > with the patched kernel), most markedly (~5%) for open/close and > exec: The patch is certainly a good thing but I somehow doubt for many of the improvments that they are a result of the patch and not the unavoidable benchmarking noise. Queued for 2.6.26, Ralf