[PATCH] convert node/zone_start_paddr to pfns

I've had ia32-discontigmem under test for a month, uneventfully.  Possibly
because I don't have a machine to test it on....

A major part of this work is a general move to convert the low-level
memory management to consistently use pageframe numbers.  It's a bit
schizo at present..

This patch was written by Martin Bligh.  A version of this patch is in
the 2.4 aa tree.

It changes the unsigned longs node_start_paddr and zone_start_paddr to
page frame numbers.  This is necessary because a PAE address is 36 bits
and cannot be represented in an unsigned long.

 - The per-node physical memory start address node_start_paddr becomes
   a pfn, node_start_pfn.

 - The per-zone physical memory start address zone_start_paddr becomes
   a pfn, zone_start_pfn.

 - free_area_init_node() takes a pfn rather than a physical address.

Patricia has tested this patch on the following configurations: UP,
SMP, SMP PAE, multiquad, multiquad PAE, multiquad DISCONTIGMEM,
multiquad DISCONTIGMEM PAE.

arch/alpha/mm/numa.c

@@ -294,7 +294,7 @@ void __init paging_init(void)
 			zones_size[ZONE_DMA] = dma_local_pfn;
 			zones_size[ZONE_NORMAL] = (end_pfn - start_pfn) - dma_local_pfn;
 		}
-		free_area_init_node(nid, NODE_DATA(nid), NULL, zones_size, start_pfn<<PAGE_SHIFT, NULL);
+		free_area_init_node(nid, NODE_DATA(nid), NULL, zones_size, start_pfn, NULL);
 		lmax_mapnr = PLAT_NODE_DATA_STARTNR(nid) + PLAT_NODE_DATA_SIZE(nid);
 		if (lmax_mapnr > max_mapnr) {
 			max_mapnr = lmax_mapnr;
@@ -371,7 +371,7 @@ void __init mem_init(void)
 		totalram_pages += free_all_bootmem_node(NODE_DATA(nid));
 
 		lmem_map = NODE_MEM_MAP(nid);
-		pfn = NODE_DATA(nid)->node_start_paddr >> PAGE_SHIFT;
+		pfn = NODE_DATA(nid)->node_start_pfn;
 		for (i = 0; i < PLAT_NODE_DATA_SIZE(nid); i++, pfn++)
 			if (page_is_ram(pfn) && PageReserved(lmem_map+i))
 				reservedpages++;

arch/arm/mm/init.c

@@ -512,7 +512,7 @@ void __init paging_init(struct meminfo *mi, struct machine_desc *mdesc)
 		arch_adjust_zones(node, zone_size, zhole_size);
 
 		free_area_init_node(node, pgdat, 0, zone_size,
-				bdata->node_boot_start, zhole_size);
+				bdata->node_boot_start >> PAGE_SHIFT, zhole_size);
 	}
 
 	/*

arch/cris/mm/init.c

@@ -345,7 +345,7 @@ paging_init(void)
 	 * mem_map page array.
 	 */
 
-	free_area_init_node(0, 0, 0, zones_size, PAGE_OFFSET, 0);
+	free_area_init_node(0, 0, 0, zones_size, PAGE_OFFSET >> PAGE_SHIFT, 0);
 
 }
 

arch/mips64/sgi-ip27/ip27-memory.c

@@ -253,7 +253,7 @@ void __init paging_init(void)
 
 		zones_size[ZONE_DMA] = end_pfn + 1 - start_pfn;
 		free_area_init_node(node, NODE_DATA(node), 0, zones_size, 
-						start_pfn << PAGE_SHIFT, 0);
+						start_pfn, 0);
 		if ((PLAT_NODE_DATA_STARTNR(node) + 
 					PLAT_NODE_DATA_SIZE(node)) > pagenr)
 			pagenr = PLAT_NODE_DATA_STARTNR(node) +

arch/sh/mm/init.c

@@ -123,11 +123,11 @@ void __init paging_init(void)
 			zones_size[ZONE_DMA] = max_dma - start_pfn;
 			zones_size[ZONE_NORMAL] = low - max_dma;
 		}
-		free_area_init_node(0, NODE_DATA(0), 0, zones_size, __MEMORY_START, 0);
+		free_area_init_node(0, NODE_DATA(0), 0, zones_size, __MEMORY_START >> PAGE_SHIFT, 0);
 #ifdef CONFIG_DISCONTIGMEM
 		zones_size[ZONE_DMA] = __MEMORY_SIZE_2ND >> PAGE_SHIFT;
 		zones_size[ZONE_NORMAL] = 0;
-		free_area_init_node(1, NODE_DATA(1), 0, zones_size, __MEMORY_START_2ND, 0);
+		free_area_init_node(1, NODE_DATA(1), 0, zones_size, __MEMORY_START_2ND >> PAGE_SHIFT, 0);
 #endif
  	}
 }

arch/sparc/mm/srmmu.c

@@ -1311,7 +1311,7 @@ void __init srmmu_paging_init(void)
 		zholes_size[ZONE_HIGHMEM] = npages - calc_highpages();
 
 		free_area_init_node(0, NULL, NULL, zones_size,
-				    phys_base, zholes_size);
+				    phys_base >> PAGE_SHIFT, zholes_size);
 	}
 
 /* P3: easy to fix, todo. Current code is utterly broken, though. */

arch/sparc/mm/sun4c.c

@@ -2074,7 +2074,7 @@ void __init sun4c_paging_init(void)
 		zholes_size[ZONE_HIGHMEM] = npages - calc_highpages();
 
 		free_area_init_node(0, NULL, NULL, zones_size,
-				    phys_base, zholes_size);
+				    phys_base >> PAGE_SHIFT, zholes_size);
 	}
 
 	cnt = 0;

arch/sparc64/mm/init.c

@@ -1559,7 +1559,7 @@ void __init paging_init(void)
 		zholes_size[ZONE_DMA] = npages - pages_avail;
 
 		free_area_init_node(0, NULL, NULL, zones_size,
-				    phys_base, zholes_size);
+				    phys_base >> PAGE_SHIFT, zholes_size);
 	}
 
 	device_scan();

include/asm-alpha/mmzone.h

@@ -52,14 +52,14 @@ extern plat_pg_data_t *plat_node_data[];
 
 #if 1
 #define PLAT_NODE_DATA_LOCALNR(p, n)	\
-	(((p) - PLAT_NODE_DATA(n)->gendata.node_start_paddr) >> PAGE_SHIFT)
+	(((p) >> PAGE_SHIFT) - PLAT_NODE_DATA(n)->gendata.node_start_pfn)
 #else
 static inline unsigned long
 PLAT_NODE_DATA_LOCALNR(unsigned long p, int n)
 {
 	unsigned long temp;
-	temp = p - PLAT_NODE_DATA(n)->gendata.node_start_paddr;
-	return (temp >> PAGE_SHIFT);
+	temp = p >> PAGE_SHIFT;
+	return temp - PLAT_NODE_DATA(n)->gendata.node_start_pfn;
 }
 #endif
 
@@ -96,7 +96,7 @@ PLAT_NODE_DATA_LOCALNR(unsigned long p, int n)
  * and returns the kaddr corresponding to first physical page in the
  * node's mem_map.
  */
-#define LOCAL_BASE_ADDR(kaddr)	((unsigned long)__va(NODE_DATA(KVADDR_TO_NID(kaddr))->node_start_paddr))
+#define LOCAL_BASE_ADDR(kaddr)	((unsigned long)__va(NODE_DATA(KVADDR_TO_NID(kaddr))->node_start_pfn << PAGE_SHIFT))
 
 #define LOCAL_MAP_NR(kvaddr) \
 	(((unsigned long)(kvaddr)-LOCAL_BASE_ADDR(kvaddr)) >> PAGE_SHIFT)

include/asm-alpha/pgtable.h

@@ -195,8 +195,8 @@ extern unsigned long __zero_page(void);
 #define PAGE_TO_PA(page)	((page - mem_map) << PAGE_SHIFT)
 #else
 #define PAGE_TO_PA(page) \
-		((((page)-(page)->zone->zone_mem_map) << PAGE_SHIFT) \
-		+ (page)->zone->zone_start_paddr)
+		((( (page) - (page)->zone->zone_mem_map ) \
+		+ (page)->zone->zone_start_pfn) << PAGE_SHIFT)
 #endif
 
 #ifndef CONFIG_DISCONTIGMEM
@@ -216,7 +216,7 @@ extern unsigned long __zero_page(void);
 	unsigned long pfn;							\
 										\
 	pfn = ((unsigned long)((page)-(page)->zone->zone_mem_map)) << 32;	\
-	pfn += (page)->zone->zone_start_paddr << (32-PAGE_SHIFT);		\
+	pfn += (page)->zone->zone_start_pfn << 32);				\
 	pte_val(pte) = pfn | pgprot_val(pgprot);				\
 										\
 	pte;									\

include/asm-arm/memory.h

@@ -80,8 +80,8 @@ static inline void *phys_to_virt(unsigned long x)
  * around in memory.
  */
 #define page_to_pfn(page)					\
-	(((page) - page_zone(page)->zone_mem_map)		\
-	  + (page_zone(page)->zone_start_paddr >> PAGE_SHIFT))
+	(( (page) - page_zone(page)->zone_mem_map)		\
+	  + page_zone(page)->zone_start_pfn)
 
 #define pfn_to_page(pfn)					\
 	(PFN_TO_MAPBASE(pfn) + LOCAL_MAP_NR((pfn) << PAGE_SHIFT))

include/asm-mips64/mmzone.h

@@ -27,7 +27,7 @@ extern plat_pg_data_t *plat_node_data[];
 #define PLAT_NODE_DATA_STARTNR(n)    (PLAT_NODE_DATA(n)->gendata.node_start_mapnr)
 #define PLAT_NODE_DATA_SIZE(n)	     (PLAT_NODE_DATA(n)->gendata.node_size)
 #define PLAT_NODE_DATA_LOCALNR(p, n) \
-		(((p) - PLAT_NODE_DATA(n)->gendata.node_start_paddr) >> PAGE_SHIFT)
+		(((p) >> PAGE_SHIFT) - PLAT_NODE_DATA(n)->gendata.node_start_pfn)
 
 #define numa_node_id()	cputocnode(current->processor)
 

include/asm-mips64/pgtable.h

@@ -484,8 +484,8 @@ static inline pgprot_t pgprot_noncached(pgprot_t _prot)
 #define PAGE_TO_PA(page)	((page - mem_map) << PAGE_SHIFT)
 #else
 #define PAGE_TO_PA(page) \
-		((((page)-(page)->zone->zone_mem_map) << PAGE_SHIFT) \
-		+ ((page)->zone->zone_start_paddr))
+		(( ((page)-(page)->zone->zone_mem_map) + \
+		   (page)->zone->zone_start_pfn) << PAGE_SHIFT)
 #endif
 #define mk_pte(page, pgprot)						\
 ({									\

include/asm-ppc64/mmzone.h

@@ -31,7 +31,7 @@ extern plat_pg_data_t plat_node_data[];
 	(PLAT_NODE_DATA(n)->gendata.node_start_mapnr)
 #define PLAT_NODE_DATA_SIZE(n)		(PLAT_NODE_DATA(n)->gendata.node_size)
 #define PLAT_NODE_DATA_LOCALNR(p, n)	\
-	(((p) - PLAT_NODE_DATA(n)->gendata.node_start_paddr) >> PAGE_SHIFT)
+	(((p) >> PAGE_SHIFT) - PLAT_NODE_DATA(n)->gendata.node_start_pfn)
 
 #ifdef CONFIG_DISCONTIGMEM
 
@@ -67,7 +67,7 @@ extern plat_pg_data_t plat_node_data[];
  * node's mem_map.
  */
 #define LOCAL_BASE_ADDR(kaddr) \
-	((unsigned long)__va(NODE_DATA(KVADDR_TO_NID(kaddr))->node_start_paddr))
+	((unsigned long)__va(NODE_DATA(KVADDR_TO_NID(kaddr))->node_start_pfn << PAGE_SHIFT))
 
 #define LOCAL_MAP_NR(kvaddr) \
 	(((unsigned long)(kvaddr)-LOCAL_BASE_ADDR(kvaddr)) >> PAGE_SHIFT)

include/asm-ppc64/page.h

@@ -207,7 +207,7 @@ static inline int get_order(unsigned long size)
 #ifdef CONFIG_DISCONTIGMEM
 #define page_to_pfn(page) \
 		((page) - page_zone(page)->zone_mem_map + \
-		(page_zone(page)->zone_start_paddr >> PAGE_SHIFT))
+		(page_zone(page)->zone_start_pfn))
 #define pfn_to_page(pfn)	discontigmem_pfn_to_page(pfn)
 #else
 #define pfn_to_page(pfn)	(mem_map + (pfn))

include/linux/mm.h

@@ -316,8 +316,8 @@ static inline void set_page_zone(struct page *page, unsigned long zone_num)
 #else /* CONFIG_HIGHMEM || WANT_PAGE_VIRTUAL */
 
 #define page_address(page)						\
-	__va( (((page) - page_zone(page)->zone_mem_map) << PAGE_SHIFT)	\
-			+ page_zone(page)->zone_start_paddr)
+	__va( ( ((page) - page_zone(page)->zone_mem_map)		\
+			+ page_zone(page)->zone_start_pfn) << PAGE_SHIFT)
 
 #endif /* CONFIG_HIGHMEM || WANT_PAGE_VIRTUAL */
 
@@ -398,7 +398,7 @@ static inline pmd_t *pmd_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long a
 
 extern void free_area_init(unsigned long * zones_size);
 extern void free_area_init_node(int nid, pg_data_t *pgdat, struct page *pmap,
-	unsigned long * zones_size, unsigned long zone_start_paddr, 
+	unsigned long * zones_size, unsigned long zone_start_pfn, 
 	unsigned long *zholes_size);
 extern void mem_init(void);
 extern void show_mem(void);

include/linux/mmzone.h

@@ -110,7 +110,8 @@ struct zone {
 	 */
 	struct pglist_data	*zone_pgdat;
 	struct page		*zone_mem_map;
-	unsigned long		zone_start_paddr;
+	/* zone_start_pfn == zone_start_paddr >> PAGE_SHIFT */
+	unsigned long		zone_start_pfn;
 	unsigned long		zone_start_mapnr;
 
 	/*
@@ -161,7 +162,7 @@ typedef struct pglist_data {
 	struct page *node_mem_map;
 	unsigned long *valid_addr_bitmap;
 	struct bootmem_data *bdata;
-	unsigned long node_start_paddr;
+	unsigned long node_start_pfn;
 	unsigned long node_start_mapnr;
 	unsigned long node_size;
 	int node_id;

mm/highmem.c

@@ -383,7 +383,7 @@ void blk_queue_bounce(request_queue_t *q, struct bio **bio_orig)
 		/*
 		 * is destination page below bounce pfn?
 		 */
-		if ((page - page_zone(page)->zone_mem_map) + (page_zone(page)->zone_start_paddr >> PAGE_SHIFT) < pfn)
+		if ((page - page_zone(page)->zone_mem_map) + (page_zone(page)->zone_start_pfn) < pfn)
 			continue;
 
 		/*

mm/numa.c

@@ -22,11 +22,11 @@ pg_data_t contig_page_data = { .bdata = &contig_bootmem_data };
  * Should be invoked with paramters (0, 0, unsigned long *[], start_paddr).
  */
 void __init free_area_init_node(int nid, pg_data_t *pgdat, struct page *pmap,
-	unsigned long *zones_size, unsigned long zone_start_paddr, 
+	unsigned long *zones_size, unsigned long zone_start_pfn, 
 	unsigned long *zholes_size)
 {
 	free_area_init_core(0, &contig_page_data, &mem_map, zones_size, 
-				zone_start_paddr, zholes_size, pmap);
+				zone_start_pfn, zholes_size, pmap);
 }
 
 #endif /* !CONFIG_DISCONTIGMEM */
@@ -48,7 +48,7 @@ struct page * alloc_pages_node(int nid, unsigned int gfp_mask, unsigned int orde
  * Nodes can be initialized parallely, in no particular order.
  */
 void __init free_area_init_node(int nid, pg_data_t *pgdat, struct page *pmap,
-	unsigned long *zones_size, unsigned long zone_start_paddr, 
+	unsigned long *zones_size, unsigned long zone_start_pfn, 
 	unsigned long *zholes_size)
 {
 	int i, size = 0;
@@ -57,7 +57,7 @@ void __init free_area_init_node(int nid, pg_data_t *pgdat, struct page *pmap,
 	if (mem_map == NULL)
 		mem_map = (struct page *)PAGE_OFFSET;
 
-	free_area_init_core(nid, pgdat, &discard, zones_size, zone_start_paddr,
+	free_area_init_core(nid, pgdat, &discard, zones_size, zone_start_pfn,
 					zholes_size, pmap);
 	pgdat->node_id = nid;
 

mm/page_alloc.c

@@ -773,7 +773,7 @@ static inline unsigned long wait_table_bits(unsigned long size)
  *   - clear the memory bitmaps
  */
 void __init free_area_init_core(int nid, pg_data_t *pgdat, struct page **gmap,
-	unsigned long *zones_size, unsigned long zone_start_paddr, 
+	unsigned long *zones_size, unsigned long zone_start_pfn, 
 	unsigned long *zholes_size, struct page *lmem_map)
 {
 	unsigned long i, j;
@@ -781,8 +781,6 @@ void __init free_area_init_core(int nid, pg_data_t *pgdat, struct page **gmap,
 	unsigned long totalpages, offset, realtotalpages;
 	const unsigned long zone_required_alignment = 1UL << (MAX_ORDER-1);
 
-	BUG_ON(zone_start_paddr & ~PAGE_MASK);
-
 	totalpages = 0;
 	for (i = 0; i < MAX_NR_ZONES; i++) {
 		unsigned long size = zones_size[i];
@@ -810,7 +808,7 @@ void __init free_area_init_core(int nid, pg_data_t *pgdat, struct page **gmap,
 	}
 	*gmap = pgdat->node_mem_map = lmem_map;
 	pgdat->node_size = totalpages;
-	pgdat->node_start_paddr = zone_start_paddr;
+	pgdat->node_start_pfn = zone_start_pfn;
 	pgdat->node_start_mapnr = (lmem_map - mem_map);
 	pgdat->nr_zones = 0;
 
@@ -868,9 +866,9 @@ void __init free_area_init_core(int nid, pg_data_t *pgdat, struct page **gmap,
 
 		zone->zone_mem_map = mem_map + offset;
 		zone->zone_start_mapnr = offset;
-		zone->zone_start_paddr = zone_start_paddr;
+		zone->zone_start_pfn = zone_start_pfn;
 
-		if ((zone_start_paddr >> PAGE_SHIFT) & (zone_required_alignment-1))
+		if ((zone_start_pfn) & (zone_required_alignment-1))
 			printk("BUG: wrong zone alignment, it will crash\n");
 
 		/*
@@ -885,8 +883,12 @@ void __init free_area_init_core(int nid, pg_data_t *pgdat, struct page **gmap,
 			SetPageReserved(page);
 			INIT_LIST_HEAD(&page->list);
 			if (j != ZONE_HIGHMEM)
-				set_page_address(page, __va(zone_start_paddr));
-			zone_start_paddr += PAGE_SIZE;
+				/*
+				 * The shift left won't overflow because the
+				 * ZONE_NORMAL is below 4G.
+				 */
+				set_page_address(page, __va(zone_start_pfn << PAGE_SHIFT));
+			zone_start_pfn++;
 		}
 
 		offset += size;