On Wed, 2008-11-05 at 20:10 +1100, Nigel Cunningham wrote: > In your example above, zone_start_pfn will be 1, won't it? If that's the > case, I shouldn't need to subtract ARCH_PFN_OFFSET to get the right > index within the zone and avoid the same wastage that ARCH_PFN_OFFSET > avoids with mem_map. Yeah, I don't think the first zone will ever start before ARCH_PFN_OFFSET. If the code just deals with starting at any random zone_start_pfn and going to any other random zone_end_pfn without any waste, then it should be fine in the presence of ARCH_PFN_OFFSET. The only trouble is if it assumes memory to start at 0x0. -- Dave _______________________________________________ linux-pm mailing list linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/linux-pm
