From: Jeremy Fitzhardinge Subject: Re: [PATCH 30 of 38] xen: implement io_apic_ops Date: Thu, 20 Nov 2008 11:38:06 -0800 Message-ID: <4925BC9E.9010006@goop.org> References: <20081120093506.GB6885@elte.hu> <492597B9.8070506@goop.org> <20081120192218.GF3955@elte.hu> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20081120192218.GF3955@elte.hu> Sender: linux-kernel-owner@vger.kernel.org To: Ingo Molnar Cc: linux-kernel@vger.kernel.org, Xen-devel , the arch/x86 maintainers , Ian Campbell , Thomas Gleixner , "H. Peter Anvin" , Yinghai Lu , "Eric W. Biederman" List-Id: xen-devel@lists.xenproject.org Ingo Molnar wrote: > it certainly looks thin enough to me although i'm really not sure we > want to virtualize at the IO-APIC level. Peter, what's your > opinion/preference? > Given that Xen's requirements here are pretty Xen-specific (I don't imagine that any other virtualization system would work in the same way), I didn't bother trying to come up with a general "virtualization" layer at this level - that's why I was pretty blunt about putting xen_* calls in without any indirection. But the code could certainly be restructured in a way which would make it simpler to hook Xen as a side-effect, so long as it was achieving some other goal as well (general cleanup, big iron architecture support, better msi handling, whatever...). J