The patch titled
mm: fix memmap init to initialize valid memmap for memory hole
has been removed from the -mm tree. Its filename was
mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole.patch
This patch was dropped because an updated version will be merged
The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
------------------------------------------------------
Subject: mm: fix memmap init to initialize valid memmap for memory hole
From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
If PFN is not in early_node_map[] then struct page for it is not
initialized. If there are holes within a MAX_ORDER_NE_PAGES range of
pages, then PG_reserved will not be set. Code that walks PFNs within
MAX_ORDER_NR_PAGES will the use uninitialized struct pages.
To avoid any problems, this patch initializes holes within a
MAX_ORDER_NR_PAGES that valid memmap exists but is otherwise unused.
Sayeth davem:
What's happening is that the assertion in mm/page_alloc.c:move_freepages()
is triggering:
BUG_ON(page_zone(start_page) != page_zone(end_page));
Once I knew this is what was happening, I added some annotations:
if (unlikely(page_zone(start_page) != page_zone(end_page))) {
printk(KERN_ERR "move_freepages: Bogus zones: "
"start_page[%p] end_page[%p] zone[%p]\n",
start_page, end_page, zone);
printk(KERN_ERR "move_freepages: "
"start_zone[%p] end_zone[%p]\n",
page_zone(start_page), page_zone(end_page));
printk(KERN_ERR "move_freepages: "
"start_pfn[0x%lx] end_pfn[0x%lx]\n",
page_to_pfn(start_page), page_to_pfn(end_page));
printk(KERN_ERR "move_freepages: "
"start_nid[%d] end_nid[%d]\n",
page_to_nid(start_page), page_to_nid(end_page));
...
And here's what I got:
move_freepages: Bogus zones: start_page[2207d0000] end_page[2207dffc0] zone[fffff8103effcb00]
move_freepages: start_zone[fffff8103effcb00] end_zone[fffff8003fffeb00]
move_freepages: start_pfn[0x81f600] end_pfn[0x81f7ff]
move_freepages: start_nid[1] end_nid[0]
My memory layout on this box is:
[ 0.000000] Zone PFN ranges:
[ 0.000000] Normal 0x00000000 -> 0x0081ff5d
[ 0.000000] Movable zone start PFN for each node
[ 0.000000] early_node_map[8] active PFN ranges
[ 0.000000] 0: 0x00000000 -> 0x00020000
[ 0.000000] 1: 0x00800000 -> 0x0081f7ff
[ 0.000000] 1: 0x0081f800 -> 0x0081fe50
[ 0.000000] 1: 0x0081fed1 -> 0x0081fed8
[ 0.000000] 1: 0x0081feda -> 0x0081fedb
[ 0.000000] 1: 0x0081fedd -> 0x0081fee5
[ 0.000000] 1: 0x0081fee7 -> 0x0081ff51
[ 0.000000] 1: 0x0081ff59 -> 0x0081ff5d
So it's a block move in that 0x81f600-->0x81f7ff region which triggers
the problem.
So I did a lot (and I do mean _A LOT_) of digging. And it seems that
unless you set HOLES_IN_ZONE you have to make sure that all of the
memmap regions of free space in a zone begin and end on an HPAGE_SIZE
boundary (the requirement used to be that it had to be MAX_ORDER
sized).
Well, this assumption enterred the tree back in 2005 (!!!) from
the following commit in the history-2.6 tree:
commit 69fba2dd0335abec0b0de9ac53d5bbb67c31fc60
Author: Kamezawa Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Date: Fri Jan 7 22:01:35 2005 -0800
[PATCH] no buddy bitmap patch revisit: for mm/page_alloc.c
Reported-by: David Miller <davem@xxxxxxxxxxxxxx>
Acked-by: Mel Gorman <mel@xxxxxxxxx>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Cc: Heiko Carstens <heiko.carstens@xxxxxxxxxx>
Cc: <stable@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
arch/ia64/mm/numa.c | 12 ++++++++++--
arch/x86/mm/numa_64.c | 6 +++++-
include/linux/mm.h | 1 +
include/linux/mmzone.h | 6 ------
mm/page_alloc.c | 34 +++++++++++++++++++++++++++++++---
5 files changed, 47 insertions(+), 12 deletions(-)
diff -puN include/linux/mmzone.h~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole include/linux/mmzone.h
--- a/include/linux/mmzone.h~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole
+++ a/include/linux/mmzone.h
@@ -1070,12 +1070,6 @@ void sparse_init(void);
#define sparse_index_init(_sec, _nid) do {} while (0)
#endif /* CONFIG_SPARSEMEM */
-#ifdef CONFIG_NODES_SPAN_OTHER_NODES
-#define early_pfn_in_nid(pfn, nid) (early_pfn_to_nid(pfn) == (nid))
-#else
-#define early_pfn_in_nid(pfn, nid) (1)
-#endif
-
#ifndef early_pfn_valid
#define early_pfn_valid(pfn) (1)
#endif
diff -puN mm/page_alloc.c~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole mm/page_alloc.c
--- a/mm/page_alloc.c~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole
+++ a/mm/page_alloc.c
@@ -2557,6 +2557,21 @@ static inline unsigned long wait_table_b
* higher will lead to a bigger reserve which will get freed as contiguous
* blocks as reclaim kicks in
*/
+#ifdef CONFIG_NODE_SPAN_OTHER_NODE
+static inline bool init_pfn_under_nid(unsigned long pfn, int nid)
+{
+ int nid_in_map = early_pfn_to_nid_solid(pfn);
+
+ if (nid_in_map == -1)
+ return true;
+ return (nid_in_map == nid);
+}
+#else
+static inline bool init_pfn_under_nid(unsigned long pfn, int nid)
+{
+ return true;
+}
+#endif
static void setup_zone_migrate_reserve(struct zone *zone)
{
unsigned long start_pfn, pfn, end_pfn;
@@ -2635,7 +2650,11 @@ void __meminit memmap_init_zone(unsigned
if (context == MEMMAP_EARLY) {
if (!early_pfn_valid(pfn))
continue;
- if (!early_pfn_in_nid(pfn, nid))
+ /*
+ * This returns false if the page exists and it's
+ * not under this node.
+ */
+ if (!init_pfn_under_nid(pfn, nid))
continue;
}
page = pfn_to_page(pfn);
@@ -2990,7 +3009,7 @@ static int __meminit next_active_region_
* was used and there are no special requirements, this is a convenient
* alternative
*/
-int __meminit early_pfn_to_nid(unsigned long pfn)
+int __meminit early_pfn_to_nid_solid(unsigned long pfn)
{
int i;
@@ -3001,8 +3020,17 @@ int __meminit early_pfn_to_nid(unsigned
if (start_pfn <= pfn && pfn < end_pfn)
return early_node_map[i].nid;
}
+ return -1;
+}
+/* Allow fallback to 0 */
+int __meminit early_pfn_to_nid(unsigned long pfn)
+{
+ int nid;
- return 0;
+ nid = early_pfn_to_nid_solid(pfn);
+ if (nid < 0)
+ return 0;
+ return nid;
}
#endif /* CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID */
diff -puN arch/ia64/mm/numa.c~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole arch/ia64/mm/numa.c
--- a/arch/ia64/mm/numa.c~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole
+++ a/arch/ia64/mm/numa.c
@@ -58,7 +58,7 @@ paddr_to_nid(unsigned long paddr)
* SPARSEMEM to allocate the SPARSEMEM sectionmap on the NUMA node where
* the section resides.
*/
-int early_pfn_to_nid(unsigned long pfn)
+int early_pfn_to_nid_solid(unsigned long pfn)
{
int i, section = pfn >> PFN_SECTION_SHIFT, ssec, esec;
@@ -70,9 +70,17 @@ int early_pfn_to_nid(unsigned long pfn)
return node_memblk[i].nid;
}
- return 0;
+ return -1;
}
+int early_pfn_to_nid(unsigned long pfn)
+{
+ int nid = early_pfn_to_nid_solid(pfn);
+
+ if (nid < 0) /* see page_alloc.c */
+ return 0;
+ return nid;
+}
#ifdef CONFIG_MEMORY_HOTPLUG
/*
* SRAT information is stored in node_memblk[], then we can use SRAT
diff -puN arch/x86/mm/numa_64.c~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole arch/x86/mm/numa_64.c
--- a/arch/x86/mm/numa_64.c~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole
+++ a/arch/x86/mm/numa_64.c
@@ -166,10 +166,14 @@ int __init compute_hash_shift(struct boo
return shift;
}
-int early_pfn_to_nid(unsigned long pfn)
+int early_pfn_to_nid_solid(unsigned long pfn)
{
return phys_to_nid(pfn << PAGE_SHIFT);
}
+int early_pfn_to_nid(unsigned long pfn)
+{
+ return early_pfn_to_nid_solid(pfn);
+}
static void * __init early_node_mem(int nodeid, unsigned long start,
unsigned long end, unsigned long size,
diff -puN include/linux/mm.h~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole include/linux/mm.h
--- a/include/linux/mm.h~mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole
+++ a/include/linux/mm.h
@@ -1047,6 +1047,7 @@ extern void work_with_active_regions(int
extern void sparse_memory_present_with_active_regions(int nid);
#ifndef CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID
extern int early_pfn_to_nid(unsigned long pfn);
+extern int early_pfn_to_solid(unsigned long pfn);
#endif /* CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID */
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */
extern void set_dma_reserve(unsigned long new_dma_reserve);
_
Patches currently in -mm which might be from kamezawa.hiroyu@xxxxxxxxxxxxxx are
memcg-use-__gfp_nowarn-in-page-cgroup-allocation.patch
linux-next.patch
mm-fix-memmap-init-to-initialize-valid-memmap-for-memory-hole.patch
proc-pid-maps-dont-show-pgoff-of-pure-anon-vmas.patch
proc-pid-maps-dont-show-pgoff-of-pure-anon-vmas-checkpatch-fixes.patch
mm-introduce-for_each_populated_zone-macro.patch
mm-introduce-for_each_populated_zone-macro-cleanup.patch
cgroup-css-id-support.patch
cgroup-fix-frequent-ebusy-at-rmdir.patch
memcg-use-css-id.patch
memcg-hierarchical-stat.patch
memcg-fix-shrinking-memory-to-return-ebusy-by-fixing-retry-algorithm.patch
memcg-fix-oom-killer-under-memcg.patch
memcg-fix-oom-killer-under-memcg-fix2.patch
memcg-fix-oom-killer-under-memcg-fix.patch
memcg-show-memcg-information-during-oom.patch
memcg-show-memcg-information-during-oom-fix2.patch
memcg-show-memcg-information-during-oom-fix.patch
memcg-show-memcg-information-during-oom-fix-fix.patch
memcg-show-memcg-information-during-oom-fix-fix-checkpatch-fixes.patch
memcg-remove-mem_cgroup_calc_mapped_ratio-take2.patch
memcg-remove-mem_cgroup_reclaim_imbalance-remnants.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
