[PATCH 1/2] linux: sparsemem: Initialize all memmap entries within sections

From: Akira Takeuchi
Date: Thu Mar 29 2012 - 06:00:50 EST

This commit fixes the problem for the kernel

VM subsystem insists that memmap entries within the align to MAX_ORDER_NR_PAGES
must exist and be initialized.

However, in the kernel with CONFIG_SPARSEMEM=y and CONFIG_HAVE_ARCH_PFN_VALID=y,
the kernel only initializes the entries corresponding to the memory regions
specified by "mem=" options. This causes "kernel BUG at mm/page_alloc.c:777!"
This BUG message comes from the following BUG_ON() line in move_freepages().

BUG_ON(page_zone(start_page) != page_zone(end_page));

Signed-off-by: Akira Takeuchi <takeuchi.akr@xxxxxxxxxxxxxxxx>
Signed-off-by: Kiyoshi Owada <owada.kiyoshi@xxxxxxxxxxxxxxxx>
include/linux/mmzone.h | 8 +++++---
1 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index dff7115..1b7538c 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -1088,13 +1088,15 @@ static inline struct mem_section *__pfn_to_section(unsigned long pfn)
return __nr_to_section(pfn_to_section_nr(pfn));

-static inline int pfn_valid(unsigned long pfn)
+static inline int sparsemem_pfn_valid(unsigned long pfn)
if (pfn_to_section_nr(pfn) >= NR_MEM_SECTIONS)
return 0;
return valid_section(__nr_to_section(pfn_to_section_nr(pfn)));
+#define pfn_valid(pfn) sparsemem_pfn_valid(pfn)

static inline int pfn_present(unsigned long pfn)
@@ -1119,7 +1121,7 @@ static inline int pfn_present(unsigned long pfn)
#define pfn_to_nid(pfn) (0)

-#define early_pfn_valid(pfn) pfn_valid(pfn)
+#define early_pfn_valid(pfn) sparsemem_pfn_valid(pfn)
void sparse_init(void);
#define sparse_init() do {} while (0)

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/