sh: bootmem tidying for discontig/sparsemem preparation.

This reworks some of the node 0 bootmem initialization in
preparation for discontigmem and sparsemem support.

ARCH_POPULATES_NODE_MAP is switched to as a result of this.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

5 files changed, 207 insertions, 181 deletions

diff --git a/arch/sh/kernel/setup.c b/arch/sh/kernel/setup.c
index 4d6d891..60cc216 100644
--- a/arch/sh/kernel/setup.c
+++ b/arch/sh/kernel/setup.c
@@ -4,7 +4,7 @@
  * This file handles the architecture-dependent parts of initialization
  *
  *  Copyright (C) 1999  Niibe Yutaka
- *  Copyright (C) 2002 - 2006 Paul Mundt
+ *  Copyright (C) 2002 - 2007 Paul Mundt
  */
 #include <linux/screen_info.h>
 #include <linux/ioport.h>
@@ -15,15 +15,18 @@
 #include <linux/seq_file.h>
 #include <linux/root_dev.h>
 #include <linux/utsname.h>
+#include <linux/nodemask.h>
 #include <linux/cpu.h>
 #include <linux/pfn.h>
 #include <linux/fs.h>
+#include <linux/mm.h>
 #include <asm/uaccess.h>
 #include <asm/io.h>
 #include <asm/sections.h>
 #include <asm/irq.h>
 #include <asm/setup.h>
 #include <asm/clock.h>
+#include <asm/mmu_context.h>
 
 extern void * __rd_start, * __rd_end;
 
@@ -202,53 +205,33 @@ static int __init sh_mv_setup(char **cmdline_p)
 	return 0;
 }
 
-void __init setup_arch(char **cmdline_p)
+/*
+ * Register fully available low RAM pages with the bootmem allocator.
+ */
+static void __init register_bootmem_low_pages(void)
 {
-	unsigned long bootmap_size;
-	unsigned long start_pfn, max_pfn, max_low_pfn;
-
-#ifdef CONFIG_CMDLINE_BOOL
-        strcpy(COMMAND_LINE, CONFIG_CMDLINE);
-#endif
-
-	ROOT_DEV = old_decode_dev(ORIG_ROOT_DEV);
-
-#ifdef CONFIG_BLK_DEV_RAM
-	rd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK;
-	rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0);
-	rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0);
-#endif
-
-	if (!MOUNT_ROOT_RDONLY)
-		root_mountflags &= ~MS_RDONLY;
-	init_mm.start_code = (unsigned long) _text;
-	init_mm.end_code = (unsigned long) _etext;
-	init_mm.end_data = (unsigned long) _edata;
-	init_mm.brk = (unsigned long) _end;
-
-	code_resource.start = (unsigned long)virt_to_phys(_text);
-	code_resource.end = (unsigned long)virt_to_phys(_etext)-1;
-	data_resource.start = (unsigned long)virt_to_phys(_etext);
-	data_resource.end = (unsigned long)virt_to_phys(_edata)-1;
-
-	sh_mv_setup(cmdline_p);
-
+	unsigned long curr_pfn, last_pfn, pages;
 
 	/*
-	 * Find the highest page frame number we have available
+	 * We are rounding up the start address of usable memory:
 	 */
-	max_pfn = PFN_DOWN(__pa(memory_end));
+	curr_pfn = PFN_UP(__MEMORY_START);
 
 	/*
-	 * Determine low and high memory ranges:
+	 * ... and at the end of the usable range downwards:
 	 */
-	max_low_pfn = max_pfn;
+	last_pfn = PFN_DOWN(__pa(memory_end));
 
-	/*
-	 * Partially used pages are not usable - thus
-	 * we are rounding upwards:
-	 */
-	start_pfn = PFN_UP(__pa(_end));
+	if (last_pfn > max_low_pfn)
+		last_pfn = max_low_pfn;
+
+	pages = last_pfn - curr_pfn;
+	free_bootmem(PFN_PHYS(curr_pfn), PFN_PHYS(pages));
+}
+
+void __init setup_bootmem_allocator(unsigned long start_pfn)
+{
+	unsigned long bootmap_size;
 
 	/*
 	 * Find a proper area for the bootmem bitmap. After this
@@ -256,31 +239,11 @@ void __init setup_arch(char **cmdline_p)
 	 * is intact) must be done via bootmem_alloc().
 	 */
 	bootmap_size = init_bootmem_node(NODE_DATA(0), start_pfn,
-					 __MEMORY_START>>PAGE_SHIFT,
-					 max_low_pfn);
-	/*
-	 * Register fully available low RAM pages with the bootmem allocator.
-	 */
-	{
-		unsigned long curr_pfn, last_pfn, pages;
-
-		/*
-		 * We are rounding up the start address of usable memory:
-		 */
-		curr_pfn = PFN_UP(__MEMORY_START);
-		/*
-		 * ... and at the end of the usable range downwards:
-		 */
-		last_pfn = PFN_DOWN(__pa(memory_end));
+					 min_low_pfn, max_low_pfn);
 
-		if (last_pfn > max_low_pfn)
-			last_pfn = max_low_pfn;
-
-		pages = last_pfn - curr_pfn;
-		free_bootmem_node(NODE_DATA(0), PFN_PHYS(curr_pfn),
-				  PFN_PHYS(pages));
-	}
+	register_bootmem_low_pages();
 
+	node_set_online(0);
 
 	/*
 	 * Reserve the kernel text and
@@ -289,14 +252,14 @@ void __init setup_arch(char **cmdline_p)
 	 * case of us accidentally initializing the bootmem allocator with
 	 * an invalid RAM area.
 	 */
-	reserve_bootmem_node(NODE_DATA(0), __MEMORY_START+PAGE_SIZE,
+	reserve_bootmem(__MEMORY_START+PAGE_SIZE,
 		(PFN_PHYS(start_pfn)+bootmap_size+PAGE_SIZE-1)-__MEMORY_START);
 
 	/*
 	 * reserve physical page 0 - it's a special BIOS page on many boxes,
 	 * enabling clean reboots, SMP operation, laptop functions.
 	 */
-	reserve_bootmem_node(NODE_DATA(0), __MEMORY_START, PAGE_SIZE);
+	reserve_bootmem(__MEMORY_START, PAGE_SIZE);
 
 #ifdef CONFIG_BLK_DEV_INITRD
 	ROOT_DEV = MKDEV(RAMDISK_MAJOR, 0);
@@ -310,8 +273,8 @@ void __init setup_arch(char **cmdline_p)
 
 	if (LOADER_TYPE && INITRD_START) {
 		if (INITRD_START + INITRD_SIZE <= (max_low_pfn << PAGE_SHIFT)) {
-			reserve_bootmem_node(NODE_DATA(0), INITRD_START +
-						__MEMORY_START, INITRD_SIZE);
+			reserve_bootmem(INITRD_START + __MEMORY_START,
+					INITRD_SIZE);
 			initrd_start = INITRD_START + PAGE_OFFSET +
 					__MEMORY_START;
 			initrd_end = initrd_start + INITRD_SIZE;
@@ -324,6 +287,71 @@ void __init setup_arch(char **cmdline_p)
 		}
 	}
 #endif
+}
+
+#ifndef CONFIG_NEED_MULTIPLE_NODES
+static void __init setup_memory(void)
+{
+	unsigned long start_pfn;
+
+	/*
+	 * Partially used pages are not usable - thus
+	 * we are rounding upwards:
+	 */
+	start_pfn = PFN_UP(__pa(_end));
+	setup_bootmem_allocator(start_pfn);
+}
+#else
+extern void __init setup_memory(void);
+#endif
+
+void __init setup_arch(char **cmdline_p)
+{
+	enable_mmu();
+
+#ifdef CONFIG_CMDLINE_BOOL
+	strcpy(COMMAND_LINE, CONFIG_CMDLINE);
+#endif
+
+	ROOT_DEV = old_decode_dev(ORIG_ROOT_DEV);
+
+#ifdef CONFIG_BLK_DEV_RAM
+	rd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK;
+	rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0);
+	rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0);
+#endif
+
+	if (!MOUNT_ROOT_RDONLY)
+		root_mountflags &= ~MS_RDONLY;
+	init_mm.start_code = (unsigned long) _text;
+	init_mm.end_code = (unsigned long) _etext;
+	init_mm.end_data = (unsigned long) _edata;
+	init_mm.brk = (unsigned long) _end;
+
+	code_resource.start = virt_to_phys(_text);
+	code_resource.end = virt_to_phys(_etext)-1;
+	data_resource.start = virt_to_phys(_etext);
+	data_resource.end = virt_to_phys(_edata)-1;
+
+	parse_early_param();
+
+	sh_mv_setup(cmdline_p);
+
+	/*
+	 * Find the highest page frame number we have available
+	 */
+	max_pfn = PFN_DOWN(__pa(memory_end));
+
+	/*
+	 * Determine low and high memory ranges:
+	 */
+	max_low_pfn = max_pfn;
+	min_low_pfn = __MEMORY_START >> PAGE_SHIFT;
+
+	nodes_clear(node_online_map);
+	setup_memory();
+	paging_init();
+	sparse_init();
 
 #ifdef CONFIG_DUMMY_CONSOLE
 	conswitchp = &dummy_con;
@@ -332,8 +360,6 @@ void __init setup_arch(char **cmdline_p)
 	/* Perform the machine specific initialisation */
 	if (likely(sh_mv.mv_setup))
 		sh_mv.mv_setup(cmdline_p);
-
-	paging_init();
 }
 
 struct sh_machine_vector* __init get_mv_byname(const char* name)
diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig
index 3cac22f..5359f3d 100644
--- a/arch/sh/mm/Kconfig
+++ b/arch/sh/mm/Kconfig
@@ -291,6 +291,17 @@ config VSYSCALL
 	  For systems with an MMU that can afford to give up a page,
 	  (the default value) say Y.
 
+config NODES_SHIFT
+	int
+	default "1"
+	depends on NEED_MULTIPLE_NODES
+
+config ARCH_FLATMEM_ENABLE
+	def_bool y
+
+config ARCH_POPULATES_NODE_MAP
+	def_bool y
+
 choice
 	prompt "Kernel page size"
 	default PAGE_SIZE_4KB
diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c
index ae957a9..4d03098 100644
--- a/arch/sh/mm/init.c
+++ b/arch/sh/mm/init.c
@@ -1,37 +1,20 @@
-/* $Id: init.c,v 1.19 2004/02/21 04:42:16 kkojima Exp $
- *
- *  linux/arch/sh/mm/init.c
+/*
+ * linux/arch/sh/mm/init.c
  *
  *  Copyright (C) 1999  Niibe Yutaka
- *  Copyright (C) 2002, 2004  Paul Mundt
+ *  Copyright (C) 2002 - 2007  Paul Mundt
  *
  *  Based on linux/arch/i386/mm/init.c:
  *   Copyright (C) 1995  Linus Torvalds
  */
-
-#include <linux/signal.h>
-#include <linux/sched.h>
-#include <linux/kernel.h>
-#include <linux/errno.h>
-#include <linux/string.h>
-#include <linux/types.h>
-#include <linux/ptrace.h>
-#include <linux/mman.h>
 #include <linux/mm.h>
 #include <linux/swap.h>
-#include <linux/smp.h>
 #include <linux/init.h>
-#include <linux/highmem.h>
 #include <linux/bootmem.h>
-#include <linux/pagemap.h>
 #include <linux/proc_fs.h>
-#include <asm/processor.h>
-#include <asm/system.h>
-#include <asm/uaccess.h>
-#include <asm/pgtable.h>
-#include <asm/pgalloc.h>
+#include <linux/percpu.h>
+#include <linux/io.h>
 #include <asm/mmu_context.h>
-#include <asm/io.h>
 #include <asm/tlb.h>
 #include <asm/cacheflush.h>
 #include <asm/cache.h>
@@ -39,37 +22,51 @@
 DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
 pgd_t swapper_pg_dir[PTRS_PER_PGD];
 
-#ifdef CONFIG_MMU
-/* It'd be good if these lines were in the standard header file. */
-#define START_PFN	(NODE_DATA(0)->bdata->node_boot_start >> PAGE_SHIFT)
-#define MAX_LOW_PFN	(NODE_DATA(0)->bdata->node_low_pfn)
-#endif
-
 void (*copy_page)(void *from, void *to);
 void (*clear_page)(void *to);
 
 void show_mem(void)
 {
-	int i, total = 0, reserved = 0;
-	int shared = 0, cached = 0;
+	int total = 0, reserved = 0, free = 0;
+	int shared = 0, cached = 0, slab = 0;
+	pg_data_t *pgdat;
 
 	printk("Mem-info:\n");
 	show_free_areas();
-	printk("Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
-	i = max_mapnr;
-	while (i-- > 0) {
-		total++;
-		if (PageReserved(mem_map+i))
-			reserved++;
-		else if (PageSwapCache(mem_map+i))
-			cached++;
-		else if (page_count(mem_map+i))
-			shared += page_count(mem_map+i) - 1;
+
+	for_each_online_pgdat(pgdat) {
+		struct page *page, *end;
+		unsigned long flags;
+
+		pgdat_resize_lock(pgdat, &flags);
+		page = pgdat->node_mem_map;
+		end = page + pgdat->node_spanned_pages;
+
+		do {
+			total++;
+			if (PageReserved(page))
+				reserved++;
+			else if (PageSwapCache(page))
+				cached++;
+			else if (PageSlab(page))
+				slab++;
+			else if (!page_count(page))
+				free++;
+			else
+				shared += page_count(page) - 1;
+			page++;
+		} while (page < end);
+
+		pgdat_resize_unlock(pgdat, &flags);
 	}
-	printk("%d pages of RAM\n",total);
-	printk("%d reserved pages\n",reserved);
-	printk("%d pages shared\n",shared);
-	printk("%d pages swap cached\n",cached);
+
+	printk("Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
+	printk("%d pages of RAM\n", total);
+	printk("%d free pages\n", free);
+	printk("%d reserved pages\n", reserved);
+	printk("%d slab pages\n", slab);
+	printk("%d pages shared\n", shared);
+	printk("%d pages swap cached\n", cached);
 }
 
 #ifdef CONFIG_MMU
@@ -147,52 +144,38 @@ extern char __init_begin, __init_end;
  */
 void __init paging_init(void)
 {
-	unsigned long zones_size[MAX_NR_ZONES] = { 0, };
+	int nid;
 
-	/*
-	 * Setup some defaults for the zone sizes.. these should be safe
-	 * regardless of distcontiguous memory or MMU settings.
-	 */
-	zones_size[ZONE_NORMAL] = __MEMORY_SIZE >> PAGE_SHIFT;
-#ifdef CONFIG_HIGHMEM
-	zones_size[ZONE_HIGHMEM] = 0 >> PAGE_SHIFT;
-#endif
-
-#ifdef CONFIG_MMU
-	/*
-	 * If we have an MMU, and want to be using it .. we need to adjust
-	 * the zone sizes accordingly, in addition to turning it on.
-	 */
-	{
-		/* We don't need to map the kernel through the TLB, as
-		 * it is permanatly mapped using P1. So clear the
-		 * entire pgd. */
-		memset(swapper_pg_dir, 0, sizeof(swapper_pg_dir));
-
-		/* Turn on the MMU */
-		enable_mmu();
-		zones_size[ZONE_NORMAL] = MAX_LOW_PFN - START_PFN;
-	}
+	/* We don't need to map the kernel through the TLB, as
+	 * it is permanatly mapped using P1. So clear the
+	 * entire pgd. */
+	memset(swapper_pg_dir, 0, sizeof(swapper_pg_dir));
 
 	/* Set an initial value for the MMU.TTB so we don't have to
 	 * check for a null value. */
 	set_TTB(swapper_pg_dir);
 
-#elif defined(CONFIG_CPU_SH3) || defined(CONFIG_CPU_SH4)
-	/*
-	 * If we don't have CONFIG_MMU set and the processor in question
-	 * still has an MMU, care needs to be taken to make sure it doesn't
-	 * stay on.. Since the boot loader could have potentially already
-	 * turned it on, and we clearly don't want it, we simply turn it off.
-	 *
-	 * We don't need to do anything special for the zone sizes, since the
-	 * default values that were already configured up above should be
-	 * satisfactory.
-	 */
-	disable_mmu();
-#endif
-	NODE_DATA(0)->node_mem_map = NULL;
-	free_area_init_node(0, NODE_DATA(0), zones_size, __MEMORY_START >> PAGE_SHIFT, 0);
+	for_each_online_node(nid) {
+		pg_data_t *pgdat = NODE_DATA(nid);
+		unsigned long max_zone_pfns[MAX_NR_ZONES];
+		unsigned long low, start_pfn;
+
+		memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
+
+		start_pfn = pgdat->bdata->node_boot_start >> PAGE_SHIFT;
+		low = pgdat->bdata->node_low_pfn;
+
+		max_zone_pfns[ZONE_NORMAL] = low;
+		add_active_range(nid, start_pfn, low);
+
+		printk("Node %u: start_pfn = 0x%lx, low = 0x%lx\n",
+		       nid, start_pfn, low);
+
+		free_area_init_nodes(max_zone_pfns);
+
+		printk("Node %u: mem_map starts at %p\n",
+		       pgdat->node_id, pgdat->node_mem_map);
+	}
 }
 
 static struct kcore_list kcore_mem, kcore_vmalloc;
@@ -200,18 +183,33 @@ static struct kcore_list kcore_mem, kcore_vmalloc;
 void __init mem_init(void)
 {
 	int codesize, reservedpages, datasize, initsize;
-	int tmp;
-	extern unsigned long memory_start;
+	int nid;
 
-#ifdef CONFIG_MMU
-	high_memory = (void *)__va(MAX_LOW_PFN * PAGE_SIZE);
-#else
-	extern unsigned long memory_end;
+	reservedpages = 0;
 
-	high_memory = (void *)(memory_end & PAGE_MASK);
-#endif
+	for_each_online_node(nid) {
+		pg_data_t *pgdat = NODE_DATA(nid);
+		unsigned long node_pages = 0;
+		void *node_high_memory;
+		int i;
+
+		num_physpages += pgdat->node_present_pages;
+
+		if (pgdat->node_spanned_pages)
+			node_pages = free_all_bootmem_node(pgdat);
+
+		totalram_pages += node_pages;
 
-	max_mapnr = num_physpages = MAP_NR(high_memory) - MAP_NR(memory_start);
+		for (i = 0; i < node_pages; i++)
+			if (PageReserved(pgdat->node_mem_map + i))
+				reservedpages++;
+
+		node_high_memory = (void *)((pgdat->node_start_pfn +
+					     pgdat->node_spanned_pages) <<
+						PAGE_SHIFT);
+		if (node_high_memory > high_memory)
+			high_memory = node_high_memory;
+	}
 
 	/* clear the zero-page */
 	memset(empty_zero_page, 0, PAGE_SIZE);
@@ -229,16 +227,6 @@ void __init mem_init(void)
 	clear_page = clear_page_nommu;
 #endif
 
-	/* this will put all low memory onto the freelists */
-	totalram_pages += free_all_bootmem_node(NODE_DATA(0));
-	reservedpages = 0;
-	for (tmp = 0; tmp < num_physpages; tmp++)
-		/*
-		 * Only count reserved RAM pages
-		 */
-		if (PageReserved(mem_map+tmp))
-			reservedpages++;
-
 	codesize =  (unsigned long) &_etext - (unsigned long) &_text;
 	datasize =  (unsigned long) &_edata - (unsigned long) &_etext;
 	initsize =  (unsigned long) &__init_end - (unsigned long) &__init_begin;
@@ -250,7 +238,7 @@ void __init mem_init(void)
 	printk(KERN_INFO "Memory: %luk/%luk available (%dk kernel code, "
 	       "%dk reserved, %dk data, %dk init)\n",
 		(unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
-		max_mapnr << (PAGE_SHIFT-10),
+		totalram_pages << (PAGE_SHIFT-10),
 		codesize >> 10,
 		reservedpages << (PAGE_SHIFT-10),
 		datasize >> 10,
@@ -289,4 +277,3 @@ void free_initrd_mem(unsigned long start, unsigned long end)
 	printk ("Freeing initrd memory: %ldk freed\n", (end - start) >> 10);
 }
 #endif
-
diff --git a/include/asm-sh/mmu_context.h b/include/asm-sh/mmu_context.h
index 01acaaa..199662b 100644
--- a/include/asm-sh/mmu_context.h
+++ b/include/asm-sh/mmu_context.h
@@ -169,6 +169,8 @@ enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
 #define destroy_context(mm)		do { } while (0)
 #define set_asid(asid)			do { } while (0)
 #define get_asid()			(0)
+#define set_TTB(pgd)			do { } while (0)
+#define get_TTB()			(0)
 #define activate_context(mm,cpu)	do { } while (0)
 #define switch_mm(prev,next,tsk)	do { } while (0)
 #define deactivate_mm(tsk,mm)		do { } while (0)
@@ -211,8 +213,8 @@ static inline void disable_mmu(void)
  * MMU control handlers for processors lacking memory
  * management hardware.
  */
-#define enable_mmu()	do { BUG(); } while (0)
-#define disable_mmu()	do { BUG(); } while (0)
+#define enable_mmu()	do { } while (0)
+#define disable_mmu()	do { } while (0)
 #endif
 
 #endif /* __KERNEL__ */
diff --git a/include/asm-sh/page.h b/include/asm-sh/page.h
index ac4b467..7464de4 100644
--- a/include/asm-sh/page.h
+++ b/include/asm-sh/page.h
@@ -59,6 +59,7 @@ extern void (*clear_page)(void *to);
 extern void (*copy_page)(void *to, void *from);
 
 extern unsigned long shm_align_mask;
+extern unsigned long max_low_pfn, min_low_pfn;
 
 #ifdef CONFIG_MMU
 extern void clear_page_slow(void *to);
@@ -124,17 +125,16 @@ typedef struct { unsigned long pgd; } pgd_t;
 #define PAGE_OFFSET		CONFIG_PAGE_OFFSET
 #define __pa(x)			((unsigned long)(x)-PAGE_OFFSET)
 #define __va(x)			((void *)((unsigned long)(x)+PAGE_OFFSET))
+#define pfn_to_kaddr(pfn)	__va((pfn) << PAGE_SHIFT)
 
-#define MAP_NR(addr)		(((unsigned long)(addr)-PAGE_OFFSET) >> PAGE_SHIFT)
-
-#define phys_to_page(phys)	(mem_map + (((phys)-__MEMORY_START) >> PAGE_SHIFT))
-#define page_to_phys(page)	(((page - mem_map) << PAGE_SHIFT) + __MEMORY_START)
+#define phys_to_page(phys)	(pfn_to_page(phys >> PAGE_SHIFT))
+#define page_to_phys(page)	(page_to_pfn(page) << PAGE_SHIFT)
 
 /* PFN start number, because of __MEMORY_START */
 #define PFN_START		(__MEMORY_START >> PAGE_SHIFT)
 #define ARCH_PFN_OFFSET		(PFN_START)
 #define virt_to_page(kaddr)	pfn_to_page(__pa(kaddr) >> PAGE_SHIFT)
-#define pfn_valid(pfn)		(((pfn) - PFN_START) < max_mapnr)
+#define pfn_valid(pfn)		((pfn) >= min_low_pfn && (pfn) < max_low_pfn)
 #define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
 
 #define VM_DATA_DEFAULT_FLAGS	(VM_READ | VM_WRITE | VM_EXEC | \