[ofa-general] Re: [RFC] [PATCH 0/2] for 2.6.23: ib/sa - add local path recordcaching

[Michael S. Tsirkin]

> Quoting Sean Hefty <sean.hefty at intel.com>:
> Subject: RE: [RFC] [PATCH 0/2] for 2.6.23: ib/sa - add local path recordcaching
> 
> >This code seems to be significantly different from what
> >OFED 1.2 has (and on the flip side, it does look better),
> >but this means that it's all new code in a core component.
> >Might it be prudent to have this sit in -mm for awhile?
> >Another approach might be to have it disabled by default in 2.6.23 -
> >that would make it a low-risk change, and give people time to experiment
> >with it.
> 
> I would rather disable it by default, since I'm not sure how many IB people run
> -mm.

Yes, seems to be true.

> I go back and forth on whether the cache should be enabled by default and
> figured we could decide on the list.


-- 
MST