Re: [PATCH 12/16] UML - Memory hotplug

[Daniel Phillips <phillips@google.com>]

Jeff Dike wrote:
> This is the big question with this patch.  How incestuous do I want to
> get with the VM system in order to get it to free up pages?  For now,
> I decided to be fairly hands-off, allocate as many pages as I can get,
> and return the total number to the host.  The host, if it wasn't happy
> with the results, can wait a bit while the UML notices that it is
> really low on memory and frees some up, and then hit up the UML for
> the remainder.

And also wrote:
> page = alloc_page(GFP_ATOMIC);

A slightly different objection than Andrew's: this will rapidly eat up
all the pages available for, e.g., receiving network packets, probably
not what you want.  How about flags=0?  This will dip a little way into
reserves but not as far as interrupts or realtime tasks, and will not
attempt any reclaim.  (Maybe we should have a GFP define for that.)

> INIT_LIST_HEAD(&unplugged->list);
> list_add(&unplugged->list, &unplugged_pages);

You don't need to initialize the list element you are adding.

Regards,

Daniel