Re: [PATCH RFC v7 7/7] iommu/arm-smmu-v3: Wire up shared-SID STE ordering and feature gating

[Peng Fan <peng.fan@linux.dev>]

On Mon, Nov 02, 2026 at 10:49:20AM -0800, Nicolin Chen wrote:
>On Fri, Oct 30, 2026 at 08:09:39PM +0800, Peng Fan (OSS) wrote:
>> From: Peng Fan <peng.fan@nxp.com>
>> 
>> With the per-domain CD table, every non-PASID sharer of a SID builds the
>> same normal-domain STE: same S1ContextPtr, S1DSS and S1CDMAX. Disable
>> stall and ATS for shared SIDs as well, so the remaining STE fields are
>> common too, and let any sharer program the STE. This also handles the
>> case where one device's driver attaches before the stream owner's driver:
>> the first attach now installs a valid STE instead of leaving the boot or
>> firmware STE in place.
>> 
>> A release or blocked attach must not abort a shared STE while other
>> sharers remain, otherwise it would break their ongoing DMA. Skip those
>> writes while the SID is shared; the last master (no longer shared) still
>> tears the STE down. This also means removing one device from a shared
>> group no longer leaves the survivors aborted.
>> 
>> Fault events on shared SIDs cannot be attributed to a specific master,
>> so arm_smmu_find_master() returns NULL when the stream's shared_masters
>> list is non-empty.
>> 
>> Disable SVA, PASID, vSMMU nesting, ATS and IOPF/stall for shared-SID
>> masters since all of them require an unambiguous SID-to-device mapping.
>> Reject sharing at insert_master() time if either master has stall enabled,
>> or if the existing master has PASID or ATS enabled.
>> 
>> Introduce arm_smmu_master_stall_enabled() to describe the policy actually
>> programmed into the hardware: a master stalls only if it asked for it and
>> its StreamID is not shared.
>
>Why do we need this? A stall_enabled master shouldn't share its
>SID in the first place.
>
>> Add arm_smmu_is_shared_sid() as an inline helper in arm-smmu-v3.h so that
>
>Use arm_smmu_master_shares_sid or arm_smmu_master_has_shared_sid.
>
>> arm-smmu-v3-sva.c and arm-smmu-v3-iommufd.c can use it without repeating
>> the list_empty() pattern.
>> 
>> Signed-off-by: Peng Fan <peng.fan@nxp.com>
>
>Please check sashiko findings; all seems valid.
>
>https://sashiko.dev/#/patchset/20261030-smmu-shared-sid-ver7-2026-10-26-v7-0-961726ebd6ab%40Example.Com

Yeah. I checked them every time. Seems the patchset will be bloating again.

Thanks
Peng

>
>Nicolin
>