Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230606120854.4170244-1-mshavit@google.com> <20230606120854.4170244-15-mshavit@google.com>
 <ZH9oQMmI8CezMCnx@nvidia.com> <CAKHBV25eDNXEu+6GWDJqpw1h5N-Jq-GPzf81kNH1dyw3Z=KsZA@mail.gmail.com>
 <ZIBxPd1/JCAle6yP@nvidia.com> <CAKHBV25K4BCewMdp3HcRtaX1iNhVpxL_6dMwp1_fmcQ5RWpKBQ@mail.gmail.com>
 <ZImuIPrB2YeulNn3@nvidia.com>
In-Reply-To: <ZImuIPrB2YeulNn3@nvidia.com>
From:   Michael Shavit <mshavit@google.com>
Date:   Wed, 14 Jun 2023 21:30:34 +0800
Message-ID: <CAKHBV25g2M-35ZQ9X-mx-La7jVDgBm6rw-umOkNdgu3_jAiA-A@mail.gmail.com>
Subject: Re: [PATCH v2 14/18] iommu/arm-smmu-v3: Support domains with shared CDs
To:     Jason Gunthorpe <jgg@nvidia.com>
Cc:     Will Deacon <will@kernel.org>, Robin Murphy <robin.murphy@arm.com>,
        Joerg Roedel <joro@8bytes.org>, jean-philippe@linaro.org,
        nicolinc@nvidia.com, baolu.lu@linux.intel.com,
        linux-arm-kernel@lists.infradead.org, iommu@lists.linux.dev,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Jun 14, 2023 at 5:57=E2=80=AFPM Robin Murphy <robin.murphy@arm.com>=
 wrote:
>
> In practice it would be entirely reasonable to only support
> cross-instance attach between instances with matching capabilities such
> that they *can* share the pagetable directly.

On Wed, Jun 14, 2023 at 8:10=E2=80=AFPM Jason Gunthorpe <jgg@nvidia.com> wr=
ote:
> When we talk about multi instance support, it means the iommu_domain -
> in whatever fixed IO page table format it uses - can be attached to
> any SMMU instance that supports it as a compatible page table format.
>
> ARM doesn't quite reach this model, but once it passes the finalize
> step it does. The goal is to move finalize to allocate. So for your
> purposes you can ignore the difference.

Got you. Failing the atach when the page table format is incompatible
with the smmu device is a lot easier to handle. I didn't notice that
SVA was already checking this elsewhere.
I can give the multi-instance support a try (with the rest of these
patches on top) and send it out as a follow-up series to this one.

> I think it is good to make progress, it looked to me like the first
> part stood alone fairly well and was an improvement on its own.

Sorry for the noobie question; it's not 100% obvious to me what the
next step is here. Is there anything I should do to progress that
first part forward?