Messages in this thread | | | Date | Wed, 02 Mar 2011 11:06:41 -0800 | From | Yinghai Lu <> | Subject | Re: [PATCH x86/mm UPDATED] x86-64, NUMA: Fix distance table handling |
| |
On 03/02/2011 11:02 AM, Tejun Heo wrote: > Hey, Yinghai. > > On Wed, Mar 02, 2011 at 10:52:28AM -0800, Yinghai Lu wrote: >>> Complexity of a solution should match the benefit of the complexity. >>> Code complexity is one of the most important metrics that we need to >>> keep an eye on. If you don't do that, the code base becomes very ugly >>> and difficult to maintain very quickly. So, yes, some amount of >>> execution inefficiency is acceptable depending on circumstances. >>> Efficiency too is something which should be traded off against other >>> benefits. >> >> No. it is not acceptable in your case. >> >> We can accept that something like: during init stage, do some probe >> and call pathes to be happy. like subarch. > > Hmmm? I can't really follow your sentence. This is init stage. > Anyways, why can't it just walk over the enabled nodes? What would be > the difference?
my point is that we really not need to go over it if original is not there.
> >> Also why did you omit my first question? > > Yeah, well, because that wasn't completely consistent with what I said > earlier. I wanted to tell you to take the assignments out of if () on > your earlier patch but I just let it pass and now I had this another > patch touching the same code, so I just had to do it. > > I know it's a petty style thing but it's my pet peeve and I can't help > it when related change goes through me, so there it is. I'm sorry but > I'll probaly do it again. I beg your understanding.
never mind.
Thanks
Yinghai
| |