RE: [PATCH] net: mctp: Add MCTP PCIe VDM transport driver

[YH Chung]

Hi Jeremy,

>Just put them in the top-level at drivers/net/mctp/ for now. There's not much in
>there at present, so will be simple to keep organised. If you end up with two
>drivers and a common set of utils between the two, that's a total of four files,
>which I think we can manage.

Understood, thanks for the guidance!
We plan to start with the AST2700 driver in the next submission, along with the shared utility code.

>YH: so we would just have the three-byte format for your proposed
>driver:
>
>   [0]: routing type (bits 0:2, others reserved)
>   [1]: bus
>   [2]: device / function
>
>- assuming you're only handling non-flit mode for now.
>
Sounds good-We'll apply this format in the next patch.


Hi Adam,

> Would the mailbox abstraction work for this?  It is what I am using in the MCTP over PCC code.  Perhaps a PCIe VDM mailbox implementation will have a wider use  than just MCTP.
>
Thanks a lot for the suggestion! For now, we're planning to remove the current abstraction and use a set of common utility functions shared between the AST2600 and AST2700 drivers.
As Jeremy mentioned, we'd prefer to hold off on introducing a PCIe VDM abstraction layer until there are more concrete use cases that would benefit from it.

>My suggestiong  is: Keep it simple.  There are only a handful of mctp device
>drivers thus far, and there seems to be little justification for a deeper hierarchy.
>
>Just focus on the cleanest implementation.  Having two drivers plus a single
>common source code for each in drivers/net/mctp seems to be easier to
>manage for now.

Appreciate the guidance!
We'll proceed as you and Jeremy suggested and keep everything under drivers/net/mctp for now.