From mboxrd@z Thu Jan 1 00:00:00 1970 From: "H. Peter Anvin" Subject: Re: [RFC] - Mapping ACPI tables as CACHED Date: Tue, 17 Aug 2010 08:51:41 -0700 Message-ID: <4C6AB00D.8020901@zytor.com> References: <20100722152220.GA18290@sgi.com> <1279849573.13929.28.camel@localhost.localdomain> <20100723072301.GC23461@elte.hu> <1279895165.13929.61.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from terminus.zytor.com ([198.137.202.10]:48868 "EHLO mail.zytor.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751468Ab0HQPwg (ORCPT ); Tue, 17 Aug 2010 11:52:36 -0400 In-Reply-To: <1279895165.13929.61.camel@localhost.localdomain> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: ykzhao Cc: Ingo Molnar , Jack Steiner , "tglx@linutronix.de" , "lenb@kernel.org" , "linux-acpi@vger.kernel.org" , "linux-kernel@vger.kernel.org" , Linus Torvalds On 07/23/2010 07:26 AM, ykzhao wrote: > > Yes. We can't map the corresponding ACPI region as cached under the > following case: > >No E820_ACPI region is reported by BIOS. In such case the ACPI > table resides in the NVS region > Why could we not map the NVS region as cached? That doesn't seem to make sense. In practice, at least, on all BIOSes I've seen the NVS region is just another hunk of RAM. Sample from a real system: BIOS-e820: 000000007d6b0000 - 000000007d6cc000 (ACPI data) BIOS-e820: 000000007d6cc000 - 000000007d700000 (ACPI NVS) Both are clearly RAM. If you're not talking about the e820 NVS region, that might be a different thing, but for the ROM region in the legacy area, the fixed MTRRs are often set up to allow caching, and we should be able to map them cacheable, e.g. on this system: 00000-9FFFF write-back A0000-BFFFF uncachable C0000-CFFFF write-protect D0000-DFFFF uncachable E0000-FFFFF write-protect Clearly cacheable. -hpa