***SPAM*** Re: [ofa-general] Re: [PATCH] opensm/osm_port_info_rcv.c: don't clear sw->need_update if port 0 is active

Hal Rosenstock hal.rosenstock at gmail.com
Thu Jan 15 06:40:40 PST 2009


On Sun, Jan 11, 2009 at 4:48 PM, Yevgeny Kliteynik
<kliteyn at dev.mellanox.co.il> wrote:
> Sasha Khapyorsky wrote:
>>
>> On 10:57 Thu 01 Jan     , Yevgeny Kliteynik wrote:
>>>
>>> When switch is coming up after reset, port 0 always reports
>>> logical state ACTIVE.
>>> OpenSM shouldn't clear sw->need_update flag because of port 0.
>>>
>>> Signed-off-by: Yevgeny Kliteynik <kliteyn at dev.mellanox.co.il>
>>
>> Applied. Thanks.
>>
>> BTW is it legal from IBA point of view for switch to setup logical state
>> of port 0 without SM intervention after reset?
>
> Good question. I don't remember any special port 0 treatment in the spec
> WRT the port state...

For base port 0 (BSP0), port state is not used. For enhanced port 0
(ESP0), it behaves like a TCA. I would expect that on reset, the port
would be INIT and require the SM to walk it up to ACTIVE but once
there it would likely remain there.

I don't know if this impacts the patch incorporated or not or whether
this discussion is an aside to that.

-- Hal

> -- Yevgeny
>
>> Sasha



More information about the general mailing list