> - if the node is already online.- If the node is offline,
> (It means new node is comming!) then the memory will belongs
> to node 0 yet.
What is the long term plan to address this? Can you make sure
that the new node is always brought online before you get to
this code? Or will you have to bring the node online in the
middle of the memory hot-add code?
Presumably there is a similar issue with hot add cpu.
-Tony
>> - if the node is already online.- If the node is offline,
>> (It means new node is comming!) then the memory will belongs
>> to node 0 yet.
>
> What is the long term plan to address this? Can you make sure
> that the new node is always brought online before you get to
> this code? Or will you have to bring the node online in the
> middle of the memory hot-add code?
>
> Presumably there is a similar issue with hot add cpu.
>
Yes, cpu hot add has the same issue. It's really a performance issue,
not a functional one so you haven't seen people beating down the doors
to fix it.
Technically there isn't anything but manpower standing in the way of
doing it. But it is a good amount of work for the little it gains.
-Joel
> > - if the node is already online.- If the node is offline,
> > (It means new node is comming!) then the memory will belongs
> > to node 0 yet.
>
> What is the long term plan to address this? Can you make sure
> that the new node is always brought online before you get to
> this code? Or will you have to bring the node online in the
> middle of the memory hot-add code?
I still have patches for new pgdat addtion,
But I was afraid that everyone think they are messy due to too much code.
So, I tried to post a part of them step by step.
Ok, I would like to try to rebase and repost them.
It might be good time for it now. :-)
Bye.
--
Yasunori Goto