From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1NRrny-0005lF-AF for qemu-devel@nongnu.org; Mon, 04 Jan 2010 13:37:26 -0500 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1NRrnf-0005YP-Gu for qemu-devel@nongnu.org; Mon, 04 Jan 2010 13:37:18 -0500 Received: from [199.232.76.173] (port=55662 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1NRrnd-0005X9-Of for qemu-devel@nongnu.org; Mon, 04 Jan 2010 13:37:06 -0500 Received: from mx1.redhat.com ([209.132.183.28]:22237) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1NRrnc-0007iA-ML for qemu-devel@nongnu.org; Mon, 04 Jan 2010 13:37:05 -0500 Date: Mon, 4 Jan 2010 20:33:56 +0200 From: "Michael S. Tsirkin" Message-ID: <20100104183356.GA18873@redhat.com> References: <20091227113732.GA3833@redhat.com> <20091227203453.GA30873@volta.aurel32.net> <20091227215023.GB3278@redhat.com> <20091227222326.GA10602@hall.aurel32.net> <20091227224021.GA3443@redhat.com> <4B37E752.1020000@codemonkey.ws> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4B37E752.1020000@codemonkey.ws> Subject: [Qemu-devel] Re: PCI: Fix bus address conversion (was Re: commit rules for common git tree) List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Anthony Liguori Cc: blauwirbel@gmail.com, qemu-devel@nongnu.org, Aurelien Jarno > On Sun, Dec 27, 2009 at 05:01:38PM -0600, Anthony Liguori wrote: > > Likewise, if you see a patch go in that you think would have benefited > > from being on the list, point it out. People are reasonable and if you > > have a good suggestion, they'll value your input and be likely to seek > > it out in the future. Here is another patch that would have benefitted from review before commit: > commit cf616802171905a9b6d087a69caa3b978b9cd741 > Author: Blue Swirl > Date: Sun Dec 27 20:52:36 2009 +0000 > > PCI: Fix bus address conversion > > Pass physical addresses to map functions instead of PCI bus addresses. > > Signed-off-by: Blue Swirl and previous related patches. The issues here that I see are: - IMO mem_base should really be pci_bus_t, as pci address might be 64 bit mapped into 32 bit target - I think pci to pci bridges need mem_base copied from parent to child, this does not seem to be done? - map functions need to get pci_bus_t (for io), and now they get a cpu address there. The real fix IMO is moving the mapping to within pci.c. I think Avi had a patch to do this - any objections to refreshing it? Blue Swirl, could you comment please? -- MST