Re: [PATCH v2 2/5] memory-hotplug: update mce_bad_pages when removingthe memory

From: Dave Hansen
Date: Fri Oct 19 2012 - 09:42:50 EST

On 10/18/2012 03:20 PM, Andrew Morton wrote:
> On Wed, 17 Oct 2012 08:09:55 -0700
> Dave Hansen <dave@xxxxxxxxxxxxxxxxxx> wrote:
>>> +static void clear_hwpoisoned_pages(struct page *memmap, int nr_pages)
>>> +{
>>> + int i;
>>> +
>>> + if (!memmap)
>>> + return;
>> I guess free_section_usemap() does the same thing.
> What does this observation mean?

sparse_remove_one_section() has an if(ms->section_mem_map) statement.
Inside that if() block is the only place in the function where 'memmap'
can get set.

Currently, sparse_remove_one_section() calls in to free_section_usemap()
ouside of that if() block. With this patch new call to
clear_hwpoisoned_pages() is done in the same place, both passing 'memmap'.

However, both free_section_usemap() and clear_hwpoisoned_pages() check
'memmap' for NULL and immediately return if so. That's a bit silly
since it could hide garbage coming back from sparse_decode_mem_map().
Seems like we should just call them both inside that if() block, or
reorganize sparse_remove_one_section(), maybe like this:

void sparse_remove_one_section(struct zone *zone, struct mem_section *ms)
struct page *memmap = NULL;
unsigned long *usemap = NULL;

if (!ms->section_mem_map)

usemap = ms->pageblock_flags;
memmap = sparse_decode_mem_map(ms->section_mem_map,
ms->section_mem_map = 0;
ms->pageblock_flags = NULL;

free_section_usemap(memmap, usemap);
clear_hwpoisoned_pages(usemap, ...);

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at