Re: [PATCH 2/2] mm/sparsemem: Fix a bug in free_map_bootmem when CONFIG_SPARSEMEM_VMEMMAP

[Wanpeng Li <liwanp@linux.vnet.ibm.com>]

Hi Yanfei,
On Thu, Oct 03, 2013 at 11:32:02AM +0800, Zhang Yanfei wrote:
>From: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
>
>We pass the number of pages which hold page structs of a memory
>section to function free_map_bootmem. This is right when
>!CONFIG_SPARSEMEM_VMEMMAP but wrong when CONFIG_SPARSEMEM_VMEMMAP.
>When CONFIG_SPARSEMEM_VMEMMAP, we should pass the number of pages
>of a memory section to free_map_bootmem.
>
>So the fix is removing the nr_pages parameter. When
>CONFIG_SPARSEMEM_VMEMMAP, we directly use the prefined marco
>PAGES_PER_SECTION in free_map_bootmem. When !CONFIG_SPARSEMEM_VMEMMAP,
>we calculate page numbers needed to hold the page structs for a
>memory section and use the value in free_map_bootmem.
>
>Signed-off-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
>---
> mm/sparse.c |   17 +++++++----------
> 1 files changed, 7 insertions(+), 10 deletions(-)
>
>diff --git a/mm/sparse.c b/mm/sparse.c
>index fbb9dbc..908c134 100644
>--- a/mm/sparse.c
>+++ b/mm/sparse.c
>@@ -603,10 +603,10 @@ static void __kfree_section_memmap(struct page *memmap)
> 	vmemmap_free(start, end);
> }
> #ifdef CONFIG_MEMORY_HOTREMOVE
>-static void free_map_bootmem(struct page *memmap, unsigned long nr_pages)
>+static void free_map_bootmem(struct page *memmap)
> {
> 	unsigned long start = (unsigned long)memmap;
>-	unsigned long end = (unsigned long)(memmap + nr_pages);
>+	unsigned long end = (unsigned long)(memmap + PAGES_PER_SECTION);
>
> 	vmemmap_free(start, end);
> }
>@@ -648,11 +648,13 @@ static void __kfree_section_memmap(struct page *memmap)
> }
>
> #ifdef CONFIG_MEMORY_HOTREMOVE
>-static void free_map_bootmem(struct page *memmap, unsigned long nr_pages)
>+static void free_map_bootmem(struct page *memmap)
> {
> 	unsigned long maps_section_nr, removing_section_nr, i;
> 	unsigned long magic;
> 	struct page *page = virt_to_page(memmap);
>+	unsigned long nr_pages = get_order(sizeof(struct page) *
>+					   PAGES_PER_SECTION);

Why replace PAGE_ALIGN(XXX) >> PAGE_SHIFT by get_order(XXX)? This will result 
in memory leak.

Regards,
Wanpeng Li 

>
> 	for (i = 0; i < nr_pages; i++, page++) {
> 		magic = (unsigned long) page->lru.next;
>@@ -756,7 +758,6 @@ static inline void clear_hwpoisoned_pages(struct page *memmap, int nr_pages)
> static void free_section_usemap(struct page *memmap, unsigned long *usemap)
> {
> 	struct page *usemap_page;
>-	unsigned long nr_pages;
>
> 	if (!usemap)
> 		return;
>@@ -777,12 +778,8 @@ static void free_section_usemap(struct page *memmap, unsigned long *usemap)
> 	 * on the section which has pgdat at boot time. Just keep it as is now.
> 	 */
>
>-	if (memmap) {
>-		nr_pages = PAGE_ALIGN(PAGES_PER_SECTION * sizeof(struct page))
>-			>> PAGE_SHIFT;
>-
>-		free_map_bootmem(memmap, nr_pages);
>-	}
>+	if (memmap)
>+		free_map_bootmem(memmap);
> }
>
> void sparse_remove_one_section(struct zone *zone, struct mem_section *ms)
>-- 
>1.7.1

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>