Re: [PATCH RFC v2 1/4] iommu/arm-smmu-v3: Convert streams from RB tree to XArray

[Nicolin Chen <nicolinc@nvidia.com>]

On Tue, Sep 22, 2026 at 10:04:39PM +0800, Peng Fan wrote:
> >... and slightly mentions "simplifies". Maybe it should write about
> >why this is necessary and how xarray would simplify vs RB tree?
> 
> Jean had a comment that "maybe simplify the driver first by moving to a xarray",
> see [1] [2].
> 
> [1] https://lore.kernel.org/linux-iommu/20230518130459.GA2587493@myrica/
> [2] https://lore.kernel.org/linux-iommu/ecb3725c-27c4-944b-b42c-f4e293521f94@arm.com/#t

You should have put a suggested-by tag and likely one of the links
as well.

With that being said, I don't see the reasoning in those emails very
convincing either. I don't dislike xarray. Yet in this series, there
seems no compelling reason on switching RB tree to xarray.

> A subsequent patch needs to look up an existing stream by SID and
> then decide to share it rather than reject it - two separate
> operations. The RB tree's rb_find_add() fuses lookup and insertion
> atomically: it either inserts or returns the collision, making
> "find, then conditionally share" impossible without restructuring.
> XArray's xa_load() and xa_store() are independent operations that
> naturally support this pattern.

If you avoid duplicated stream structures as I suggested in PATCH-2,
there is no need of that, since you would only insert once to the RB
tree.

Nicolin