From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dave Hansen Subject: Re: [linux-pm] [PATCH] hibernation should work ok with memory hotplug Date: Mon, 03 Nov 2008 23:36:14 -0800 Message-ID: <1225784174.12673.547.camel@nimitz> References: <20081029105956.GA16347@atrey.karlin.mff.cuni.cz> <1225751665.12673.511.camel@nimitz> <1225771353.6755.16.camel@nigel-laptop> <200811040808.36464.rjw@sisk.pl> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <200811040808.36464.rjw@sisk.pl> Sender: linux-kernel-owner@vger.kernel.org To: "Rafael J. Wysocki" Cc: Nigel Cunningham , Matt Tolentino , linux-pm@lists.osdl.org, Dave Hansen , linux-kernel@vger.kernel.org, linux-mm@kvack.org, pavel@suse.cz, Mel Gorman , Andy Whitcroft , Andrew Morton List-Id: linux-pm@vger.kernel.org On Tue, 2008-11-04 at 08:08 +0100, Rafael J. Wysocki wrote: > A pfn always refers to specific page frame and/or struct page, so yes. > However, in one of the nodes these pfns are sort of "invalid" (they point > to struct pages belonging to other zones). AFAICS. Part of this problem is getting out of the old zone mindset. It used to be that there were one, two, or three zones, set up at boot, with static ranges. These never had holes, never changed, and were always stacked up nice and tightly on top of one another. It ain't that way no more. Now, the zones are much more truly "allocation pools". They're bunches of memory with similar attributes and hypervisors or firmware can hand them to the OS in very interesting ways. This means that the attributes that help us pool the memory together have less and less to do with physical addresses. A given physical address a decreasing chance of being related to its neighbor. -- Dave