* a bug report for function move_freepages_block
@ 2013-10-06 11:26 martin zhang
0 siblings, 0 replies; only message in thread
From: martin zhang @ 2013-10-06 11:26 UTC (permalink / raw)
To: linux-mm
[-- Attachment #1: Type: text/plain, Size: 1064 bytes --]
HI all,
I think there is a bug in function move_freepages_block.
981 int move_freepages_block(struct zone *zone, struct page *page,
...
987 start_pfn = page_to_pfn(page);
988 start_pfn = start_pfn & ~(pageblock_nr_pages-1);
989 start_page = pfn_to_page(start_pfn);
990 end_page = start_page + pageblock_nr_pages - 1;
991 end_pfn = start_pfn + pageblock_nr_pages - 1;
992
993 /* Do not cross zone boundaries */
994 if (!zone_spans_pfn(zone, start_pfn))
995 start_page = page;
The line 988 will align start_pfn with pageblock_nr_pages,
thus after line988, start_pfn maybe less than zone->pageblock_nr_pages,
in the worst case, start_pfn maybe outof the range of zone->node pfn.
and becomes a invalid pfn.
in this case, line 989 will be wrong.
so I think the check for start_pfn should be done before line 989, just
like:
start_pfn = start_pfn & ~(pageblock_nr_pages-1); <== line 988
if (!zone_spans_pfn(zone, start_pfn))
start_pfn = page_to_pfn(page);
Regards,
Martin
[-- Attachment #2: Type: text/html, Size: 1632 bytes --]
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2013-10-06 11:26 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-10-06 11:26 a bug report for function move_freepages_block martin zhang
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).