From: Ingo Molnar Subject: Re: [PATCH 0 of 9] swiotlb: use phys_addr_t for pages Date: Sun, 28 Dec 2008 10:50:02 +0100 Message-ID: <20081228095002.GG9022@elte.hu> References: <20081227104823.GI14639@elte.hu> <20081228000558I.fujita.tomonori@lab.ntt.co.jp> <4957083B.6090801@goop.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <4957083B.6090801@goop.org> Sender: linux-kernel-owner@vger.kernel.org To: Jeremy Fitzhardinge Cc: FUJITA Tomonori , tony.luck@intel.com, linux-kernel@vger.kernel.org, xen-devel@lists.xensource.com, x86@kernel.org, ian.campbell@citrix.com, beckyb@kernel.crashing.org List-Id: xen-devel@lists.xenproject.org * Jeremy Fitzhardinge wrote: > FUJITA Tomonori wrote: >> Well, I asked Jeremy to use Becky's approach so I don't have though >> this patchset isn't enough for Becky (possibly for Xen too) since >> Jeremy didn't put Becky's map_single_page changes. These changes are >> necessary to support map_map_page with highmem. I think that we can do >> it simpler than his patch. I'll send patches to that in a different >> way soon. >> >> It's not principal objection, but as I said, the patches to revert >> Jeremy's highmem patches in this patchset like the following are >> pointless. >> > > I don't mind either way. I was hoping to go with something that I've > actually tested, but I don't see any reason why Becky's patches won't > work in principle for Xen. I'm on vacation at the moment, but I can > test things out when I get back in a couple of weeks and apply whatever > fixes are needed at whatever stage we're at then. yes, obviously we prefer append-only - i kept your patches and added the Becky lineup and Fujita's cleanups on top of that. If they break your Xen swiotlb use it should all be easily bisectable. Ingo