2011-05-03 14:56:27

by Minchan Kim

[permalink] [raw]
Subject: [PATCH] Check end_pfn in pageblock_is_reserved

[mm-check-if-any-page-in-a-pageblock-is-reserved-before-marking-it-migrate_reserve]
added new function "pageblock_is_reserved". But it doesn't check end_pfn of system
so it meets following BUG in my system. This patch fixes it.

[ 1.859574] type=2000 audit(1304330017.656:1): initialized
[ 1.884130] BUG: unable to handle kernel paging request at f77fe000
[ 1.890557] IP: [<c01dfe3f>] setup_per_zone_wmarks+0x17f/0x280
[ 1.896493] *pde = 013fb067 *pte = 00000000
[ 1.900921] Oops: 0000 [#1] SMP
[ 1.904309] last sysfs file:
[ 1.907324] Modules linked in:
[ 1.910482]
[ 1.912024] Pid: 1, comm: swapper Not tainted 2.6.39-rc5-mm1+ #49 SAMSUNG ELECTRONICS CO.,LTD Samsung DeskT
[ 1.924756] EIP: 0060:[<c01dfe3f>] EFLAGS: 00010006 CPU: 0
[ 1.930282] EIP is at setup_per_zone_wmarks+0x17f/0x280
[ 1.935554] EAX: 00040080 EBX: 000403fe ECX: 00000000 EDX: f77fe000
[ 1.941859] ESI: 00110000 EDI: f77fcfc0 EBP: f5077f88 ESP: f5077f50
[ 1.948170] DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
[ 1.953613] Process swapper (pid: 1, ti=f5076000 task=f5078000 task.ti=f5076000)
[ 1.961056] Stack:
[ 1.963116] 00000091 00000000 000003b5 00000286 c0803c80 c0803ca4 0003709b 00000000
[ 1.971377] 006effe0 006effc0 00040000 00000000 f50783d8 00000001 f5077f94 c0846e40
[ 1.979633] 00000000 f5077fc4 c0101125 00000000 000002c8 f50783d8 f5077fb0 c01a8b44
[ 1.987890] Call Trace:
[ 1.990386] [<c0846e40>] init_per_zone_wmark_min+0x39/0x7c
[ 1.996001] [<c0101125>] do_one_initcall+0x35/0x170
[ 2.001014] [<c01a8b44>] ? irq_to_desc+0x14/0x20
[ 2.005758] [<c0846e07>] ? memmap_init_zone+0xca/0xca
[ 2.010941] [<c080c2bc>] kernel_init+0xbd/0x153
[ 2.015606] [<c080c1ff>] ? parse_early_options+0x25/0x25
[ 2.021047] [<c05e7542>] kernel_thread_helper+0x6/0x10
[ 2.026318] Code: 76 33 8b 4d ec 39 c3 8b 15 80 62 e0 c0 8d 3c 0e 8d 3c 3a 76 67 8b 0f 80 e5 04 75 19 8b 4d
[ 2.049095] EIP: [<c01dfe3f>] setup_per_zone_wmarks+0x17f/0x280 SS:ESP 0068:f5077f50
[ 2.056996] CR2: 00000000f77fe000
[ 2.060362] ---[ end trace 82bc55f036371117 ]---

Cc: John Stultz <[email protected]>
Cc: Arve Hjønnevåg <[email protected]>
Cc: KOSAKI Motohiro <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Dave Hansen <[email protected]>
Signed-off-by: Minchan Kim <[email protected]>
---
mm/page_alloc.c | 11 ++++++-----
1 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 8e4b80c..c5d06c0 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3344,14 +3344,14 @@ static inline unsigned long wait_table_bits(unsigned long size)
/*
* Check if a pageblock contains reserved pages
*/
-static int pageblock_is_reserved(unsigned long start_pfn)
+static int pageblock_is_reserved(unsigned long start_pfn, unsigned long end_pfn)
{
- unsigned long end_pfn = start_pfn + pageblock_nr_pages;
unsigned long pfn;

- for (pfn = start_pfn; pfn < end_pfn; pfn++)
+ for (pfn = start_pfn; pfn < end_pfn; pfn++) {
if (!pfn_valid_within(pfn) || PageReserved(pfn_to_page(pfn)))
return 1;
+ }
return 0;
}

@@ -3364,7 +3364,7 @@ static int pageblock_is_reserved(unsigned long start_pfn)
*/
static void setup_zone_migrate_reserve(struct zone *zone)
{
- unsigned long start_pfn, pfn, end_pfn;
+ unsigned long start_pfn, pfn, end_pfn, block_end_pfn;
struct page *page;
unsigned long block_migratetype;
int reserve;
@@ -3394,7 +3394,8 @@ static void setup_zone_migrate_reserve(struct zone *zone)
continue;

/* Blocks with reserved pages will never free, skip them. */
- if (pageblock_is_reserved(pfn))
+ block_end_pfn = min(pfn + pageblock_nr_pages, end_pfn);
+ if (pageblock_is_reserved(pfn, block_end_pfn))
continue;

block_migratetype = get_pageblock_migratetype(page);
--
1.7.1


2011-05-09 05:49:17

by KOSAKI Motohiro

[permalink] [raw]
Subject: Re: [PATCH] Check end_pfn in pageblock_is_reserved

> [mm-check-if-any-page-in-a-pageblock-is-reserved-before-marking-it-migrate_reserve]
> added new function "pageblock_is_reserved". But it doesn't check end_pfn of system
> so it meets following BUG in my system. This patch fixes it.
>
> [ 1.859574] type=2000 audit(1304330017.656:1): initialized
> [ 1.884130] BUG: unable to handle kernel paging request at f77fe000
> [ 1.890557] IP: [<c01dfe3f>] setup_per_zone_wmarks+0x17f/0x280
> [ 1.896493] *pde = 013fb067 *pte = 00000000
> [ 1.900921] Oops: 0000 [#1] SMP
> [ 1.904309] last sysfs file:
> [ 1.907324] Modules linked in:
> [ 1.910482]
> [ 1.912024] Pid: 1, comm: swapper Not tainted 2.6.39-rc5-mm1+ #49 SAMSUNG ELECTRONICS CO.,LTD Samsung DeskT
> [ 1.924756] EIP: 0060:[<c01dfe3f>] EFLAGS: 00010006 CPU: 0
> [ 1.930282] EIP is at setup_per_zone_wmarks+0x17f/0x280
> [ 1.935554] EAX: 00040080 EBX: 000403fe ECX: 00000000 EDX: f77fe000
> [ 1.941859] ESI: 00110000 EDI: f77fcfc0 EBP: f5077f88 ESP: f5077f50
> [ 1.948170] DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
> [ 1.953613] Process swapper (pid: 1, ti=f5076000 task=f5078000 task.ti=f5076000)
> [ 1.961056] Stack:
> [ 1.963116] 00000091 00000000 000003b5 00000286 c0803c80 c0803ca4 0003709b 00000000
> [ 1.971377] 006effe0 006effc0 00040000 00000000 f50783d8 00000001 f5077f94 c0846e40
> [ 1.979633] 00000000 f5077fc4 c0101125 00000000 000002c8 f50783d8 f5077fb0 c01a8b44
> [ 1.987890] Call Trace:
> [ 1.990386] [<c0846e40>] init_per_zone_wmark_min+0x39/0x7c
> [ 1.996001] [<c0101125>] do_one_initcall+0x35/0x170
> [ 2.001014] [<c01a8b44>] ? irq_to_desc+0x14/0x20
> [ 2.005758] [<c0846e07>] ? memmap_init_zone+0xca/0xca
> [ 2.010941] [<c080c2bc>] kernel_init+0xbd/0x153
> [ 2.015606] [<c080c1ff>] ? parse_early_options+0x25/0x25
> [ 2.021047] [<c05e7542>] kernel_thread_helper+0x6/0x10
> [ 2.026318] Code: 76 33 8b 4d ec 39 c3 8b 15 80 62 e0 c0 8d 3c 0e 8d 3c 3a 76 67 8b 0f 80 e5 04 75 19 8b 4d
> [ 2.049095] EIP: [<c01dfe3f>] setup_per_zone_wmarks+0x17f/0x280 SS:ESP 0068:f5077f50
> [ 2.056996] CR2: 00000000f77fe000
> [ 2.060362] ---[ end trace 82bc55f036371117 ]---
>
> Cc: John Stultz <[email protected]>
> Cc: Arve Hjønnevåg <[email protected]>
> Cc: KOSAKI Motohiro <[email protected]>
> Cc: Mel Gorman <[email protected]>
> Cc: Dave Hansen <[email protected]>
> Signed-off-by: Minchan Kim <[email protected]>

Good catch.
Reviewed-by: KOSAKI Motohiro <[email protected]>