[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <9b709ef7-bf81-e5d0-bf84-5d61ee0b92e5@cumulusnetworks.com>
Date: Mon, 27 Nov 2017 09:12:47 -0700
From: David Ahern <dsa@...ulusnetworks.com>
To: Arkadi Sharshevsky <arkadis@...lanox.com>,
Jiri Pirko <jiri@...nulli.us>, netdev@...r.kernel.org
Cc: davem@...emloft.net, mlxsw@...lanox.com, andrew@...n.ch,
vivien.didelot@...oirfairelinux.com, f.fainelli@...il.com,
michael.chan@...adcom.com, ganeshgr@...lsio.com,
saeedm@...lanox.com, matanb@...lanox.com, leonro@...lanox.com,
idosch@...lanox.com, jakub.kicinski@...ronome.com, ast@...nel.org,
daniel@...earbox.net, simon.horman@...ronome.com,
pieter.jansenvanvuuren@...ronome.com, john.hurley@...ronome.com,
alexander.h.duyck@...el.com, linville@...driver.com,
gospo@...adcom.com, steven.lin1@...adcom.com, yuvalm@...lanox.com,
ogerlitz@...lanox.com, roopa@...ulusnetworks.com
Subject: Re: [patch net-next RFC v2 08/11] mlxsw: spectrum_dpipe: Connect
dpipe tables to resources
On 11/23/17 6:40 AM, Arkadi Sharshevsky wrote:
>
>
> On 11/19/2017 05:58 PM, David Ahern wrote:
>> On 11/19/17 2:16 AM, Arkadi Sharshevsky wrote:
>>>
>>>
>>> On 11/18/2017 09:19 PM, David Ahern wrote:
>>>> On 11/14/17 9:18 AM, Jiri Pirko wrote:
>>>>> From: Arkadi Sharshevsky <arkadis@...lanox.com>
>>>>>
>>>>> Connect current dpipe tables to resources. The tables are connected
>>>>> in the following fashion:
>>>>> 1. IPv4 host - KVD hash single
>>>>> 2. IPv6 host - KVD hash double
>>>>> 3. Adjacency - KVD linear
>>>>
>>>> Those descriptions would be helpful to the user. A description attribute
>>>> for the resources?
>>>>
>>>
>>> As described in the cover letter this resources are used by the
>>> majority of the ASICs lookup processes. So currently there is one
>>> to one mapping but is should increase as more tables are exposed,
>>> so I don't think its a good idea to maintain such an attribute.
>>>
>>
>> 'IPv4 host' yes, but I mean the term 'KVD hash single'? Is it the same
>> across all h/w vendors? I have only seen that in the context of MLX. If
>> it is a MLX term then a description to the user that KVD hash single ==
>> IPv4 host is warranted.
>>
>
> But this relation is wrong, there is no equality here. The LPM, FDB and
> VID to FID mapping are all can be modeled as lookup tables (via dpipe)
> that use KVD hash single resource.
>
> This description string will grow very long. I dont think this is the
> right place to document such thing, eitherway, the user can dump the
> dpipe tables and see which is mapped to what resource.
Users should not have to find a PRM or user guide for *each version of
their hardware* to program something so fundamental. This is software.
We can make it user friendly. Use of vendor specific terms is fine --
allows correlation to vendor docs. But there should also be text to help
the user correlate vendor terms to generic industry terms.
Powered by blists - more mailing lists