From mboxrd@z Thu Jan 1 00:00:00 1970 From: CORNELIU ZUZU Subject: Re: [PATCH] ARM: Support for guest-request vm-events Date: Thu, 28 Jan 2016 14:36:14 +0200 Message-ID: <56AA0B3E.9030106@bitdefender.com> References: <1453979874-20037-1-git-send-email-czuzu@bitdefender.com> <1453980216.26591.69.camel@citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1453980216.26591.69.camel@citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Ian Campbell , xen-devel@lists.xen.org Cc: Stefano Stabellini List-Id: xen-devel@lists.xenproject.org On 1/28/2016 1:23 PM, Ian Campbell wrote: > On Thu, 2016-01-28 at 13:17 +0200, Corneliu ZUZU wrote: >> This patch implements ARM support for guest-request vm-events. >> The code has been ported from x86 side w/ minor adjustments. > I've not looked at the patch yet, but if it only involves minor adjustments > from the x86 side can some amount of it not be refactored into common code? > > Ian. > At a first glance it seems to me that parts of monitor vm-events code could be moved to common. But it also seems that it would require a bit of effort and I'm not sure yet if the end result won't actually complicate implementation of monitor vm-events for other architectures in the future. Some of the monitor vm-events implemented are strictly architecture specific, e.g. VM_EVENT_REASON_MOV_TO_MSR will always be an x86-only vm-event, unless it is somehow generalized (maybe somehow merged w/ VM_EVENT_REASON_WRITE_CTRLREG?). But *most* of them indeed don't directly have this kind of specificity, so it would make sense to make most of the code common, if possible. To me personally this seems like a good idea and I'd be willing to give it a try, but as I said, it might require some other changes of the code, including x86 changes. I was about to release another patch after this one to implement control-register writes vm-events for ARM, but I anticipate that doing this move first will actually benefit my effort in that direction as well (I think the patch code will get to be cleaner). So, shall I try it? Thanks, Corneliu.