linux-kernel - Re: [PATCH v4 02/13] iommu/arm-smmu-v3: Add smmu_s1_cfg to smmu

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]

Date:   Thu, 13 Jul 2023 16:34:56 +0800
From:   Michael Shavit <mshavit@...gle.com>
To:     Nicolin Chen <nicolinc@...dia.com>
Cc:     Will Deacon <will@...nel.org>, Robin Murphy <robin.murphy@....com>,
        Joerg Roedel <joro@...tes.org>, jean-philippe@...aro.org,
        jgg@...dia.com, baolu.lu@...ux.intel.com,
        linux-arm-kernel@...ts.infradead.org, iommu@...ts.linux.dev,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v4 02/13] iommu/arm-smmu-v3: Add smmu_s1_cfg to smmu_master

On Thu, Jul 13, 2023 at 9:22 AM Nicolin Chen <nicolinc@...dia.com> wrote:
>
> > Except for Nested domains, arm_smmu_master will own the STEs that are
> > inserted into the arm_smmu_device's STE table.
>
> I think that the master still owns an STE when attached to a
> nested domain. Though an IOMMU_DOMAIN_NESTED iommu_domain is
> an opaque object to the STE in the guest, the host still has
> a real STE for the nested configuration somewhere -- and it's
> likely still to be owned by the master that's attached to the
> opaque NESTED iommu_domain in the host kernel.

> I am a bit confused by this naming. If only master would own
> an s1_cfg, perhaps we can just make it "s1_cfg" and drop the
> s1_cfg pointer in the next patch.

Could be that the naming is causing some confusion. This owned_s1_cfg
is very different from the s1_cfg set-up by Nested domains in your
patch series. It's better to think of it as the default s1_cfg used
for DMA/SVA/UNMANAGED domains. Because stage 1 domains represent a
single page table, it doesn't make sense for them to own an entire CD
table. In contrast, nested domains map an entire CD table and it
therefore makes sense for them to own the s1_cfg representing that
table.
Would renaming this as default_s1_cfg make more sense?