linux-kernel - Re: [PATCH v3 06/15] KVM: arm64: Implement kvm_pgtable_hyp

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] [day] [month] [year] [list]

Message-ID: <YboRe+ZRnqtCymHV@google.com>
Date:   Wed, 15 Dec 2021 16:02:03 +0000
From:   Quentin Perret <qperret@...gle.com>
To:     Andrew Walbran <qwandor@...gle.com>
Cc:     Marc Zyngier <maz@...nel.org>, James Morse <james.morse@....com>,
        Alexandru Elisei <alexandru.elisei@....com>,
        Suzuki K Poulose <suzuki.poulose@....com>,
        Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will@...nel.org>,
        linux-arm-kernel@...ts.infradead.org, kvmarm@...ts.cs.columbia.edu,
        linux-kernel@...r.kernel.org, kernel-team@...roid.com
Subject: Re: [PATCH v3 06/15] KVM: arm64: Implement kvm_pgtable_hyp_unmap()
 at EL2

Hey Andrew,

On Wednesday 08 Dec 2021 at 14:40:54 (+0000), Andrew Walbran wrote:
> Aha, I didn't realise that block mappings, but it makes sense in that
> case. How about adding a note to the function comment here explaining
> that reasoning? Otherwise it sounds like the caller has to handle it
> somehow.

Well in fact the caller does have to handle it if it decided to use
block mappings. We just decided not to use them in pkvm for now for
simplicity, but I guess that might change at some point. The page-table
library is meant to be architecturally compliant (it can create block
mappings or not, ...) but the decision of which mappings to create is
left to the caller and the handling logic belongs there. At least that's
my view of it. So, I actually like the comment the way it is, it
describes clearly what the function does, and what the caller should
expect. I have a bunch of changes for v4 queued locally, so I'll send it
now with that comment left untouched, but please shout if you have an
idea on how to make it better.

Cheers,
Quentin