From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([140.186.70.92]:43904) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Ra3Oa-00037g-IU for qemu-devel@nongnu.org; Mon, 12 Dec 2011 05:46:09 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Ra3OU-0005nO-6o for qemu-devel@nongnu.org; Mon, 12 Dec 2011 05:46:08 -0500 Received: from mx1.redhat.com ([209.132.183.28]:20686) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Ra3OT-0005my-VM for qemu-devel@nongnu.org; Mon, 12 Dec 2011 05:46:02 -0500 From: Juan Quintela In-Reply-To: (Blue Swirl's message of "Sat, 10 Dec 2011 16:44:37 +0000") References: Date: Mon, 12 Dec 2011 11:45:57 +0100 Message-ID: <87wra26pay.fsf@trasno.mitica> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Subject: Re: [Qemu-devel] [PATCH 2/6] memory: change dirty setting APIs to take a size Reply-To: quintela@redhat.com List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Blue Swirl Cc: Avi Kivity , qemu-devel Blue Swirl wrote: > Instead of each target knowing or guessing the guest page size, > just pass the desired size of dirtied memory area. This should also > improve performance due to memset() optimizations. My understanding last time I looked at this, is that it is as easy basically to change the interface to work on pages, not on addresses. I looked at the interface from the point of view of migration, and I always needed a "page". My understanding at the time was that every caller could be change to use a page instead, but that is another set of "unfinished" patches on my ToDo list. Actually, what I found what I needed was that I wanted a pointer to the ram_block on that functions. Migration and vga were simple to fix, but TCG got me completely lost :-(, and at that point I moved to someting else. Will try to think what were my problems later today. Later, Juan.