linux-kernel - Re: [PATCH v3 18/18] KVM: arm64: Plumb the pKVM MMU in KVM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Z2GLR8KZzJn1YpUu@google.com>
Date: Tue, 17 Dec 2024 14:31:35 +0000
From: Quentin Perret <qperret@...gle.com>
To: Marc Zyngier <maz@...nel.org>
Cc: Oliver Upton <oliver.upton@...ux.dev>, Joey Gouly <joey.gouly@....com>,
	Suzuki K Poulose <suzuki.poulose@....com>,
	Zenghui Yu <yuzenghui@...wei.com>,
	Catalin Marinas <catalin.marinas@....com>,
	Will Deacon <will@...nel.org>, Fuad Tabba <tabba@...gle.com>,
	Vincent Donnefort <vdonnefort@...gle.com>,
	Sebastian Ene <sebastianene@...gle.com>,
	linux-arm-kernel@...ts.infradead.org, kvmarm@...ts.linux.dev,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 18/18] KVM: arm64: Plumb the pKVM MMU in KVM

On Tuesday 17 Dec 2024 at 14:03:37 (+0000), Marc Zyngier wrote:
> My gripe with this is that it makes it much harder to follow what is
> happening by using tags (ctags, etags, whatever). I ended up with the
> hack below, which is super ugly, but preserves the tagging
> functionality for non-pKVM.

Ack.

> I'll scratch my head to find something more elegant...

I find your proposal pretty reasonable -- I had a few different ideas
but they were all really over-engineered, so I figured relying on a
naming convention was the simplest. And any divergence will be flagged
at compile time, so that shouldn't be too hard to maintain looking
forward.

The __S2 name isn't massively descriptive though. Maybe KVM_PGT_CALL()
or something? Thinking about it, this abstraction doesn't need to be
restricted to stage-2 stuff. We could most likely hide the
__pkvm_host_{un}share_hyp() logic behind a pkvm_pgtable_hyp_{un}map()
implementation in pkvm.c as well...