]> git.karo-electronics.de Git - karo-tx-linux.git/commitdiff
page_alloc: introduce zone_movable_limit[] to keep movable limit for nodes
authorTang Chen <tangchen@cn.fujitsu.com>
Wed, 20 Feb 2013 02:14:26 +0000 (13:14 +1100)
committerStephen Rothwell <sfr@canb.auug.org.au>
Wed, 20 Feb 2013 05:52:37 +0000 (16:52 +1100)
Introduce a new array zone_movable_limit[] to store the ZONE_MOVABLE limit
from movablemem_map boot option for all nodes.  The function
sanitize_zone_movable_limit() will find out to which node the ranges in
movable_map.map[] belongs, and calculates the low boundary of ZONE_MOVABLE
for each node.

Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Liu Jiang <jiang.liu@huawei.com>
Reviewed-by: Wen Congyang <wency@cn.fujitsu.com>
Cc: Wu Jianguo <wujianguo@huawei.com>
Reviewed-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Tested-by: Lin Feng <linfeng@cn.fujitsu.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
mm/page_alloc.c

index 01991c40daf55645de077681d0bc838b74a39d65..080fac750c97ef4e2f78436a62e1c9180cbb7de8 100644 (file)
@@ -210,6 +210,7 @@ static unsigned long __meminitdata arch_zone_highest_possible_pfn[MAX_NR_ZONES];
 static unsigned long __initdata required_kernelcore;
 static unsigned long __initdata required_movablecore;
 static unsigned long __meminitdata zone_movable_pfn[MAX_NUMNODES];
+static unsigned long __meminitdata zone_movable_limit[MAX_NUMNODES];
 
 /* movable_zone is the "real" zone pages in ZONE_MOVABLE are taken from */
 int movable_zone;
@@ -4375,6 +4376,77 @@ static unsigned long __meminit zone_absent_pages_in_node(int nid,
        return __absent_pages_in_range(nid, zone_start_pfn, zone_end_pfn);
 }
 
+/**
+ * sanitize_zone_movable_limit - Sanitize the zone_movable_limit array.
+ *
+ * zone_movable_limit is initialized as 0. This function will try to get
+ * the first ZONE_MOVABLE pfn of each node from movablemem_map, and
+ * assigne them to zone_movable_limit.
+ * zone_movable_limit[nid] == 0 means no limit for the node.
+ *
+ * Note: Each range is represented as [start_pfn, end_pfn)
+ */
+static void __meminit sanitize_zone_movable_limit(void)
+{
+       int map_pos = 0, i, nid;
+       unsigned long start_pfn, end_pfn;
+
+       if (!movablemem_map.nr_map)
+               return;
+
+       /* Iterate all ranges from minimum to maximum */
+       for_each_mem_pfn_range(i, MAX_NUMNODES, &start_pfn, &end_pfn, &nid) {
+               /*
+                * If we have found lowest pfn of ZONE_MOVABLE of the node
+                * specified by user, just go on to check next range.
+                */
+               if (zone_movable_limit[nid])
+                       continue;
+
+#ifdef CONFIG_ZONE_DMA
+               /* Skip DMA memory. */
+               if (start_pfn < arch_zone_highest_possible_pfn[ZONE_DMA])
+                       start_pfn = arch_zone_highest_possible_pfn[ZONE_DMA];
+#endif
+
+#ifdef CONFIG_ZONE_DMA32
+               /* Skip DMA32 memory. */
+               if (start_pfn < arch_zone_highest_possible_pfn[ZONE_DMA32])
+                       start_pfn = arch_zone_highest_possible_pfn[ZONE_DMA32];
+#endif
+
+#ifdef CONFIG_HIGHMEM
+               /* Skip lowmem if ZONE_MOVABLE is highmem. */
+               if (zone_movable_is_highmem() &&
+                   start_pfn < arch_zone_lowest_possible_pfn[ZONE_HIGHMEM])
+                       start_pfn = arch_zone_lowest_possible_pfn[ZONE_HIGHMEM];
+#endif
+
+               if (start_pfn >= end_pfn)
+                       continue;
+
+               while (map_pos < movablemem_map.nr_map) {
+                       if (end_pfn <= movablemem_map.map[map_pos].start_pfn)
+                               break;
+
+                       if (start_pfn >= movablemem_map.map[map_pos].end_pfn) {
+                               map_pos++;
+                               continue;
+                       }
+
+                       /*
+                        * The start_pfn of ZONE_MOVABLE is either the minimum
+                        * pfn specified by movablemem_map, or 0, which means
+                        * the node has no ZONE_MOVABLE.
+                        */
+                       zone_movable_limit[nid] = max(start_pfn,
+                                       movablemem_map.map[map_pos].start_pfn);
+
+                       break;
+               }
+       }
+}
+
 #else /* CONFIG_HAVE_MEMBLOCK_NODE_MAP */
 static inline unsigned long __meminit zone_spanned_pages_in_node(int nid,
                                        unsigned long zone_type,
@@ -4393,6 +4465,10 @@ static inline unsigned long __meminit zone_absent_pages_in_node(int nid,
        return zholes_size[zone_type];
 }
 
+static void __meminit sanitize_zone_movable_limit(void)
+{
+}
+
 #endif /* CONFIG_HAVE_MEMBLOCK_NODE_MAP */
 
 static void __meminit calculate_node_totalpages(struct pglist_data *pgdat,
@@ -4839,7 +4915,6 @@ static void __init find_zone_movable_pfns_for_nodes(void)
                goto out;
 
        /* usable_startpfn is the lowest possible pfn ZONE_MOVABLE can be at */
-       find_usable_zone_for_movable();
        usable_startpfn = arch_zone_lowest_possible_pfn[movable_zone];
 
 restart:
@@ -4998,6 +5073,8 @@ void __init free_area_init_nodes(unsigned long *max_zone_pfn)
 
        /* Find the PFNs that ZONE_MOVABLE begins at in each node */
        memset(zone_movable_pfn, 0, sizeof(zone_movable_pfn));
+       find_usable_zone_for_movable();
+       sanitize_zone_movable_limit();
        find_zone_movable_pfns_for_nodes();
 
        /* Print out the zone ranges */