Received-SPF: pass (google.com: domain of linux-kernel+bounces-82851-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Tue, 27 Feb 2024 08:10:48 +0000
Message-ID: <874jdu9vjr.wl-maz@kernel.org>
From: Marc Zyngier <maz@kernel.org>
To: Wei-Lin Chang <r09922117@csie.ntu.edu.tw>
Cc: oliver.upton@linux.dev,
	james.morse@arm.com,
	suzuki.poulose@arm.com,
	yuzenghui@huawei.com,
	catalin.marinas@arm.com,
	will@kernel.org,
	linux-arm-kernel@lists.infradead.org,
	kvmarm@lists.linux.dev,
	linux-kernel@vger.kernel.org,
	sauravsc@amazon.com,
	eric.auger@redhat.com
Subject: Re: [PATCH 1/1] KVM: arm64: Affinity level 3 support
In-Reply-To: <20240227022708.795214-1-r09922117@csie.ntu.edu.tw>
References: <86frxg3i6u.wl-maz@kernel.org>
	<20240227022708.795214-1-r09922117@csie.ntu.edu.tw>
User-Agent: Wanderlust/2.15.9 (Almost Unreal) SEMI-EPG/1.14.7 (Harue)
 FLIM-LB/1.14.9 (=?UTF-8?B?R29qxY0=?=) APEL-LB/10.8 EasyPG/1.0.0 Emacs/28.2
 (x86_64-pc-linux-gnu) MULE/6.0 (HANACHIRUSATO)
Precedence: bulk
MIME-Version: 1.0 (generated by SEMI-EPG 1.14.7 - "Harue")
Content-Type: text/plain; charset=US-ASCII

On Tue, 27 Feb 2024 02:27:08 +0000,
Wei-Lin Chang <r09922117@csie.ntu.edu.tw> wrote:
> 
> On Sun, 25 Feb 2024 11:19:05 +0000,
> Marc Zyngier <maz@kernel.org> wrote:

[...]

> > I can see multiple problems with this:
> > 
> > - this is the host state, which shouldn't necessarily represent the
> >   guest state. It should be possible to restore a VM that have a
> >   different A3V value and still have the same guarantees.  There is
> >   however a small nit around ICV_CTLR_EL1.A3V, which would require
> >   trapping to emulate the A3V bit.
> > 
> > - this assumes GICv3, which is definitely not universal (we support
> >   GICv2, for which no such restriction actually exists).
> > 
> > Finally, I don't see VM save/restore being addressed here, and I
> > suspect it hasn't been looked at.
> > 
> > Overall, this patch does too many things, and it should be split in
> > discrete changes. I also want to see an actual justification for Aff3
> > support. And if we introduce it, it must be fully virtualised
> > (independent of the A3V support on the host).
> 
> Hi Marc,
> 
> Really appreciate for the feedback. I think I understand most of your
> comments and agree with them. It appears that I don't fully understand
> the changes that I am doing with this. Thanks for explaining.

I hope this doesn't deter you from working on this feature. I'll
happily answer questions and discuss the above points.

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.