bootmem crash fix
A message that is not in the archive (sent off-list or not yet received): [email protected]
From: Linux Kernel Mailing List <[email protected]>
Date: Fri, 22 Nov 2002 19:31:53 +0000
From Roman Zippel. Don't assume that physical memory starts at physical address zero.
Which would be a nice improvement, however:
- free_area_init_node(0, &contig_page_data, NULL, zones_size, 0, NULL); + free_area_init_node(0, &contig_page_data, NULL, zones_size, + __pa(PAGE_OFFSET) >> PAGE_SHIFT, NULL);
__pa(PAGE_OFFSET) is not necessarily the first physical address in the system either.
I've never seen this even implied to be the case :-) In any event, it isn't well defined and we should make it so.
Quoted 5 lines · David S. Miller
__pa(PAGE_OFFSET)is not necessarily the first physical address in the system either.
I've never seen this even implied to be the case :-)
In any event, it isn't well defined and we should make it so.
Would a first_pfn variable be in order?
Bill
From: William Lee Irwin III <[email protected]>
Date: Fri, 22 Nov 2002 13:24:19 -0800
Would a first_pfn variable be in order?
Just some silly "first_possible_paddr" macro would be fine.
It would even be fine, for now, to put it in mm.h and make it this PAGE_OFFSET expression.
I just hate these expressions that "we know" mean something.