[ofa-general] Re: [PATCH 0/6 v2] opensm: Unicast Routing Cache

[Yevgeny Kliteynik]

Hi Sasha,

Sasha Khapyorsky wrote:
> Hi Yevgeny,
> 
> On 23:29 Wed 15 Oct     , Yevgeny Kliteynik wrote:
>> I'm sending v2 of the patches:
> 
> All applied. Thanks.
> 
>> The job that still needs to be done:
>>  - Check how the cache handles port moving during discovery.
>>    Might be a bug there.
>>  - Check how unicast manager handles fast reset of switches.
>>    AFAIK SM will now write the LFT there - need to fix it
>>    (unrelated to cache, general ucast mgr issue)
> 
> Seeing again at this I think that it handles this properly. This is
> the code in osm_ucast_mgr_set_fwd_table():

Right, I noticed it too.

> 	if (!p_sw->need_update &&
> 	    !memcmp(block, p_sw->lft_buf + block_id_ho * 64, 64))
> 		continue;
> 
>>  - Optimize LFT usage - simplify current switch LFT,

I've already done this - two patches were posted to list
couple of days ago.

>>    hold two LFTs (current and cached) only when these LFTs
>>    are not identical.

I already have a draft for this too - it's very short.
Waiting for the LFT simplification patches to be applied
first.

-- Yevgeny


> This would be a nice improvement.
> 
> Sasha
>