From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1LFwVB-000531-47 for qemu-devel@nongnu.org; Thu, 25 Dec 2008 15:08:13 -0500 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1LFwV8-00052n-L4 for qemu-devel@nongnu.org; Thu, 25 Dec 2008 15:08:11 -0500 Received: from [199.232.76.173] (port=47281 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1LFwV8-00052k-GI for qemu-devel@nongnu.org; Thu, 25 Dec 2008 15:08:10 -0500 Received: from rv-out-0708.google.com ([209.85.198.248]:3444) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1LFwV8-0004fP-0Q for qemu-devel@nongnu.org; Thu, 25 Dec 2008 15:08:10 -0500 Received: by rv-out-0708.google.com with SMTP id f25so3180795rvb.22 for ; Thu, 25 Dec 2008 12:08:07 -0800 (PST) Message-ID: Date: Thu, 25 Dec 2008 21:08:07 +0100 From: "andrzej zaborowski" Subject: Re: [Qemu-devel] Re: [PATCH 1/5] re-register whole area upon lfb unmap. In-Reply-To: <494966BA.7000100@codemonkey.ws> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <1229546822-11972-1-git-send-email-glommer@redhat.com> <1229546822-11972-2-git-send-email-glommer@redhat.com> <494966BA.7000100@codemonkey.ws> Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Glauber Costa , Ian.Jackson@eu.citrix.com, avi@redhat.com, kvm@vger.kernel.org, stefano.stabellini@eu.citrix.com 2008/12/17 Anthony Liguori : > Glauber Costa wrote: >> >> set phys_offset correctly for the whole vga area when unmapping linear >> vram >> (for vga optimization). We first register the old pieces as unassigned >> memory, to make things easier for kvm (and possibly other slot based >> implementations in the future). Replacing the region directly would >> make the slot management significantly more complex. >> > > This change worries me because it involves explicitly unassigning slots and > then assigning a new, bigger slot. This is not necessary for TCG. It > suggests to me that there's a bug in the kvm slot code and that we're > changing QEMU to work around it. > > That will means there may be other places in the code that are completely > valid, but exercise this bug. > > Or is this purely an optimization? It also changes the semantics because IO callbacks are now passed offsets from regions starts instead of absolute addresses. I'm not able to tell if the change is for good or for bad though. Cheers