Re: [PATCH v4 1/2] cxl/region: Find free cxl decoder by device_for_each_child()

From: Dan Williams
Date: Tue Sep 10 2024 - 00:21:38 EST


Dan Williams wrote:
[..]
> So I wanted to write a comment here to stop the next person from
> tripping over this dependency on decoder 'add' order, but there is a
> problem. For this simple version to work it needs 3 things:
>
> 1/ decoders are added in hardware id order: done,
> devm_cxl_enumerate_decoders() handles that
>
> 2/ search for decoders in their added order: done, device_find_child()
> guarantees this, although it is not obvious without reading the internals
> of device_add().
>
> 3/ regions are de-allocated from decoders in reverse decoder id order.
> This is not enforced, in fact it is impossible to enforce. Consider that
> any memory device can be removed at any time and may not be removed in
> the order in which the device allocated switch decoders in the topology.
>
> So, that existing comment of needing to enforce ordered allocation is
> still relevant even though the implementation fails to handle the
> out-of-order region deallocation problem.
>
> I alluded to the need for a "tear down the world" implementation back in
> 2022 [1], but never got around to finishing that.
>
> Now, the cxl_port.hdm_end attribute tracks the "last" decoder to be
> allocated for endpoint ports. That same tracking needs to be added for
> switch ports, then this routine could check for ordering constraints by:
>
> /* enforce hardware ordered allocation */
> if (!cxld->region && port->hdm_end + 1 == cxld->id)
> return 1;
> return 0;
>
> As it stands now @hdm_end is never updated for switch ports.
>
> [1]: 176baefb2eb5 cxl/hdm: Commit decoder state to hardware

--- cut the reply here ---

> Yes, that looks simple enough for now, although lets not use a ternary
> condition and lets leave a comment for the next person:
>
> /* decoders are added in hardware id order
> * (devm_cxl_enumerate_decoders), allocated to regions in id order
> * (device_find_child() walks children in 'add' order)
> */

This is garbage I forgot to delete after realizing there was missing
logic to make this simple proposal work in practice.