From mboxrd@z Thu Jan  1 00:00:00 1970
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Subject: Re: [kvmarm] [RFC PATCH 0/3] KVM: ARM: Get rid of hardcoded VGIC
 addresses
Date: Mon, 29 Oct 2012 09:19:20 +1100
Message-ID: <1351462760.12271.115.camel@pasglop>
References: <1350173065-35350-1-git-send-email-c.dall@virtualopensystems.com>
	 <507F172D.2030802@suse.de>
	 <CANM98q+gfcp86zVV0B1t2aOKTftFegs6dk-+DNc-78LijueOiw@mail.gmail.com>
	 <20121017221022.GA4333@bloggs.ozlabs.ibm.com>
	 <1350518331.4678.114.camel@pasglop>
	 <CANM98q+4diN7_xBtyo6AesDP=AVzj0xQ99vtgBNa=sxLifOzow@mail.gmail.com>
	 <508675FC.7090901@siemens.com>
	 <20121024005017.GA17834@bloggs.ozlabs.ibm.com>
	 <5089656A.9010704@redhat.com> <5089699D.3050004@siemens.com>
	 <50898479.8060402@redhat.com> <1351194021.2728.170.camel@pasglop>
	 <508A5EE3.6070709@redhat.com>
	 <CAFEAcA8qi7-v5Noyyt_ur-u7GoKgdG4F+xa9AZD-55gqZEYvAQ@mail.gmail.com>
	 <508A62BB.4010209@redhat.com> <1351248264.12271.17.camel@pasglop>
	 <CAFEAcA_+brjafLeXMmWbk3_vCz1UZtqCjdr4byR8Ve8iKhzF=Q@mail.gmail.com>
	 <1351251595.12271.46.camel@pasglop> <508A846F.40700@siemens.com>
	 <1351284302.12271.83.camel@pasglop> <1351289011.12271.106.camel@pasglop>
	 <CAFEAcA_2Tr3_BXKH12OniQ2XoErGDpZTucMt6aRyZ70KLfD7zg@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Cc: Jan Kiszka <jan.kiszka@siemens.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Paul Mackerras <paulus@samba.org>,
	Christoffer Dall <c.dall@virtualopensystems.com>,
	Alexander Graf <agraf@suse.de>,
	"kvmarm@lists.cs.columbia.edu" <kvmarm@lists.cs.columbia.edu>,
	"kvm@vger.kernel.org" <kvm@vger.kernel.org>,
	kvm-ppc <kvm-ppc@vger.kernel.org>
To: Peter Maydell <peter.maydell@linaro.org>
Return-path: <kvm-ppc-owner@vger.kernel.org>
In-Reply-To: <CAFEAcA_2Tr3_BXKH12OniQ2XoErGDpZTucMt6aRyZ70KLfD7zg@mail.gmail.com>
Sender: kvm-ppc-owner@vger.kernel.org
List-Id: kvm.vger.kernel.org

On Sat, 2012-10-27 at 11:01 +0100, Peter Maydell wrote:

> This is more or less how ARM has done it (though our specific encoding
> of interrupt numbers is different, obviously).
> 
> If I were designing an interface for this kind of thing from scratch
> I'd probably want it to look like "create a specific irqchip and give
> me some kind of handle to it" and then have an interface for "assert
> interrupt line X on that irqchip". Lacking that, a plausible encoding
> scheme on the global interrupt numbers works OK given that you know
> there aren't going to be that many irqchips in practice...

Agreed. It feels like the whole business with routing is not useful for
either of us and just adds confusion...

Note: I'll be in KVM forum next week (hopefully) so if necessary (and if
you guys are there too) we can have a "live" discussion.

Cheers,
Ben.