[PATCH] mm: clear N_HIGH_MEMORY map before se set it again -v4

[Yinghai Lu]

Nathan Lynch wrote:
> Yinghai Lu <yinghai at kernel.org> writes:
>>> I bisected to the commit containing this change.  Reverting fixes the
>>> problem.
>>>
>> can you use following patch to see what happens to that nodemask?
> 
> # dmesg | grep node_states
> before clear: node_states [3]: 1
> after online check: node_states [3]: 0
> 

/*
 * Array of node states.
 */
nodemask_t node_states[NR_NODE_STATES] __read_mostly = {
        [N_POSSIBLE] = NODE_MASK_ALL,
        [N_ONLINE] = { { [0] = 1UL } },
#ifndef CONFIG_NUMA
        [N_NORMAL_MEMORY] = { { [0] = 1UL } },
#ifdef CONFIG_HIGHMEM
        [N_HIGH_MEMORY] = { { [0] = 1UL } },
#endif
        [N_CPU] = { { [0] = 1UL } },
#endif  /* NUMA */
};
EXPORT_SYMBOL(node_states);

preset that


and

Zone PFN ranges:
  DMA      0x00000000 -> 0x00001000
  Normal   0x00001000 -> 0x0001fff0
  HighMem  0x0001fff0 -> 0x0001fff0
Movable zone start PFN for each node
early_node_map[2] active PFN ranges
    0: 0x00000000 -> 0x0000009f
    0: 0x00000100 -> 0x0001fff0
before clear: node_states [3]: 1
On node 0 totalpages: 130959
free_area_init_node: node 0, pgdat c0976960, node_mem_map c1151000
  DMA zone: 60 pages used for memmap
  DMA zone: 0 pages reserved
  DMA zone: 3939 pages, LIFO batch:0
  Normal zone: 1860 pages used for memmap
  Normal zone: 125100 pages, LIFO batch:31
after online check: node_states [3]: 0


you don't have high mem, so we should it.

really hate those preset value assume node 0 has RAM...


YH