Re: [PATCH 4/8] arm64: dts: qcom: shikra: Add CAMSS node

[Loic Poulain]

On Mon, Jun 1, 2026 at 4:15 PM Bryan O'Donoghue
<bryan.odonoghue@xxxxxxxxxx> wrote:
>
> On 01/06/2026 07:42, Nihal Kumar Gupta wrote:
> > Here is the full IOMMU SID list for both platforms:
> > Agatti (QCM2290):
> > - 0x0400 0x00 — VFE non-protected
> > - 0x0800 0x00 — CDM non-protected
> > - 0x0820 0x00 — OPE read non-protected
> > - 0x0840 0x00 — OPE write non-protected
> >
> > Shikra:
> > - 0x0400 0x00 — VFE non-protected
> > - 0x0600 0x00 — CDM non-protected
> > - 0x0620 0x00 — OPE read non-protected
> > - 0x0640 0x00 — OPE write non-protected
>
> What this tells me is that the Agatti binding is incorrect for addition
> of OPE as a separate device.
>
> > Only VFE SID (0x0400) is shared between the two platforms.
>
> Both platforms have the same IOMMU list, its just that you are
> correcting a bug in Agatti's defintion.

Yes, Agatti is declaring too much if we want to go the camss multi-node way.

>
> Again only IFE SIDs should be included in CAMSS...
>
> > The current Shikra submission enumerates only the VFE SID.
>
> Great, this is what we want.
>
> I think for future submissions we need to be commenting, naming,
> something each IOMMU entry as we have a problem now in Agatti and a
> legacy problem - already - in VIDC as a result of misallocated IOMMU
> entries.
> > Should CDM and OPE each get a separate YAML binding like [1], or
> > should both be part of the CAMSS YAML binding? If part of CAMSS,
> > please confirm and we will add their SIDs in the DTS in the next revision.
> > [1]https://lore.kernel.org/all/20260508-camss-isp-ope-v3-9-
> > bb1055274603@xxxxxxxxxxxxxxxx/
>
> Yes I think CDM should be a standalone block and BTW use standard DMA
> APIs with its own IOMMU entry.

Interesting, so CDM should be a dedicated subnode as well, and not
part of the OPE?
Then, If we want to use it from OPE at some point we would need some
'cdm' phandle in OPE node.

Regards,
Loic