Re: [patch iproute2 1/6] iproute2: ipa: show switch id

[ebiederm@xmission.com (Eric W. Biederman)]

Jiri Pirko <jiri@resnulli.us> writes:

>>No.  phys_port_id is not broken in the same way, and phys_port_id does
>>not have the same set of properties.
>>
>>phys_port_id's in practice all have an IEEE prefix that identifies the
>>manufacturer and a manufacture assigned serial number.  Aka a mac
>>address or a EUID-64.  What the mlx4 ethernet driver is doing retunring
>>a 64bit EUID-64 I don't know.  If there are problems in the worst
>>case issues with phys_port_id are fixable by simple driver tweaks,
>>because fundamentally we are working with globally uniuqe identifiers.
>>Well globally unique baring manufacturing bugs in eeproms.
>
> Well the fact that phys_post_id's are now implemented mostly by putting
> mac into it does not mean that other drivers cannot do it differently.
> So once again, phys_port_id and phys_switch_id are the same in this
> matter.

No exclusively implemented that way.

And yes other drivers can implement bugs, that doesn't mean those bugs
will be toleraged and won't break userspace.

>>I agree with you that the switch id concept can be saved.  But I think
>>we should fix switch id before we export it to userspace so we don't
>>have to break userspace later.
>>
>>My intuition says we want something like ifindex, but I am not at all
>>certain how switch id is planned to be used.  Given that it is single
>>box I don't expect you are sending it out over the wire.
>
> No, it is not to be send out.
>
>>
>>*shrug*
>>
>>Why does switch id need to be persistent?  Why can't switch id be
>>property like ifindex?
>
> Well I can imagine that multiple ports of the same switch chip could be
> passed through to the virtual machines (similar to SR-IOV pf/vf).

In that case you do need something that is globally unique.  Because
if I migrate your virtual machine onto a different physical machine 
seeing the same switch id and I might make the mistaken assumption
that I am remain on the same switch if the ids are not global.

>>What are the actual requirements.
>
>
> They are actually very similar to phys_port_id. Therefore I made that
> the same.

Then please for rocket use a non-buggy implementation with a globally
unique id.

Given your descriptions of the requirements I can't see how any other
implementation isn't buggy.

Eric