Merge samba.org:/scratch/anton/linux-2.5

into samba.org:/scratch/anton/linux-2.5_ppc64_new

arch/alpha/kernel/traps.c

@@ -171,6 +171,11 @@ void show_stack(unsigned long *sp)
 	dik_show_trace(sp);
 }
 
+void dump_stack(void)
+{
+	show_stack(NULL);
+}
+
 void
 die_if_kernel(char * str, struct pt_regs *regs, long err, unsigned long *r9_15)
 {

arch/cris/kernel/traps.c

@@ -230,8 +230,12 @@ watchdog_bite_hook(struct pt_regs *regs)
 #endif	
 }
 
-/* This is normally the 'Oops' routine */
+void dump_stack(void)
+{
+	show_stack(NULL);
+}
 
+/* This is normally the 'Oops' routine */
 void 
 die_if_kernel(const char * str, struct pt_regs * regs, long err)
 {

arch/i386/Config.help

@@ -25,6 +25,15 @@ CONFIG_SMP
 
   If you don't know what to do here, say N.
 
+CONFIG_HUGETLB_PAGE
+  This enables support for huge pages.  User space applications
+  can make use of this support with the sys_alloc_hugepages and
+  sys_free_hugepages system calls.  If your applications are
+  huge page aware and your processor (Pentium or later for x86)
+  supports this, then say Y here.
+
+  Otherwise, say N.
+
 CONFIG_PREEMPT
   This option reduces the latency of the kernel when reacting to
   real-time or interactive events by allowing a low priority process to

arch/i386/config.in

@@ -154,6 +154,8 @@ if [ "$CONFIG_MWINCHIP3D" = "y" ]; then
    define_bool CONFIG_X86_OOSTORE y
 fi
 
+bool 'IA-32 Huge TLB Page Support (if available on processor)' CONFIG_HUGETLB_PAGE
+
 bool 'Symmetric multi-processing support' CONFIG_SMP
 bool 'Preemptible Kernel' CONFIG_PREEMPT
 if [ "$CONFIG_SMP" != "y" ]; then

arch/i386/kernel/entry.S

@@ -759,8 +759,8 @@ ENTRY(sys_call_table)
 	.long sys_io_getevents
 	.long sys_io_submit
 	.long sys_io_cancel
-	.long sys_ni_syscall	/* 250 */	/* sys_alloc_hugepages */
-	.long sys_ni_syscall			/* sys_free_hugepages */
+	.long sys_alloc_hugepages /* 250 */
+	.long sys_free_hugepages
 	.long sys_exit_group
 
 	.rept NR_syscalls-(.-sys_call_table)/4

arch/i386/kernel/sys_i386.c

@@ -246,3 +246,94 @@ asmlinkage int sys_olduname(struct oldold_utsname * name)
 
 	return error;
 }
+
+#ifdef CONFIG_HUGETLB_PAGE
+#define HPAGE_ALIGN(x)  (((unsigned long)x + (HPAGE_SIZE -1)) & HPAGE_MASK)
+extern long     sys_munmap(unsigned long, size_t);
+/* get_addr function gets the currently unused virtaul range in
+ * current process's address space.  It returns the LARGE_PAGE_SIZE
+ * aligned address (in cases of success).  Other kernel generic
+ * routines only could gurantee that allocated address is PAGE_SIZSE aligned.
+ */
+static unsigned long
+get_addr(unsigned long addr, unsigned long len)
+{
+	struct vm_area_struct   *vma;
+	if (addr) {
+		addr = HPAGE_ALIGN(addr);
+		vma = find_vma(current->mm, addr);
+		if (((TASK_SIZE - len) >= addr) &&
+				(!vma || addr + len <= vma->vm_start))
+			goto found_addr;
+	}
+	addr = HPAGE_ALIGN(TASK_UNMAPPED_BASE);
+	for (vma = find_vma(current->mm, addr); ; vma = vma->vm_next) {
+		if (TASK_SIZE - len < addr)
+			return -ENOMEM;
+		if (!vma || ((addr + len) < vma->vm_start))
+			goto found_addr;
+		addr = vma->vm_end;
+	}
+found_addr:
+	addr = HPAGE_ALIGN(addr);
+	return addr;
+}
+
+asmlinkage unsigned long
+sys_alloc_hugepages(int key, unsigned long addr, unsigned long len, int prot, int flag)
+{
+	struct mm_struct *mm = current->mm;
+	unsigned long   raddr;
+	int     retval = 0;
+	extern int alloc_hugetlb_pages(int, unsigned long, unsigned long, int, int);
+	if (!(cpu_has_pse))
+		return -EINVAL;
+	if (key < 0)
+		return -EINVAL;
+	if (len & (HPAGE_SIZE - 1))
+		return -EINVAL;
+	down_write(&mm->mmap_sem);
+	raddr = get_addr(addr, len);
+	if (raddr == -ENOMEM)
+		goto raddr_out;
+	retval = alloc_hugetlb_pages(key, raddr, len, prot, flag);
+
+raddr_out: up_write(&mm->mmap_sem);
+	if (retval < 0)
+		return (unsigned long) retval;
+	return raddr;
+}
+
+asmlinkage int
+sys_free_hugepages(unsigned long addr)
+{
+	struct mm_struct *mm = current->mm;
+	struct vm_area_struct   *vma;
+	int	retval;
+	extern int free_hugepages(struct vm_area_struct *);
+
+	vma = find_vma(current->mm, addr);
+	if ((!vma) || (!is_vm_hugetlb_page(vma)) || (vma->vm_start!=addr))
+		return -EINVAL;
+	down_write(&mm->mmap_sem);
+	spin_lock(&mm->page_table_lock);
+	retval =  free_hugepages(vma);
+	spin_unlock(&mm->page_table_lock);
+	up_write(&mm->mmap_sem);
+	return retval;
+}
+
+#else
+
+asmlinkage unsigned long
+sys_alloc_hugepages(int key, unsigned long addr, size_t len, int prot, int flag)
+{
+	return -ENOSYS;
+}
+asmlinkage int
+sys_free_hugepages(unsigned long addr)
+{
+	return -ENOSYS;
+}
+
+#endif

arch/i386/kernel/traps.c

@@ -189,6 +189,14 @@ void show_stack(unsigned long * esp)
 	show_trace(esp);
 }
 
+/*
+ * The architecture-independent dump_stack generator
+ */
+void dump_stack(void)
+{
+	show_stack(0);
+}
+
 void show_registers(struct pt_regs *regs)
 {
 	int i;

arch/i386/mm/Makefile

@@ -12,5 +12,6 @@ O_TARGET := mm.o
 obj-y	 := init.o pgtable.o fault.o ioremap.o extable.o pageattr.o 
 obj-$(CONFIG_DISCONTIGMEM)	+= discontig.o
 export-objs := pageattr.o
+obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o
 
 include $(TOPDIR)/Rules.make

arch/i386/mm/init.c

@@ -215,19 +215,14 @@ void __init permanent_kmaps_init(pgd_t *pgd_base)
 
 void __init one_highpage_init(struct page *page, int pfn, int bad_ppro)
 {
-	if (!page_is_ram(pfn)) {
-		SetPageReserved(page);
-		return;
-	}
-	if (bad_ppro && page_kills_ppro(pfn)) {
-		SetPageReserved(page);
-		return;
-	}
+	if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn))) {
 		ClearPageReserved(page);
 		set_bit(PG_highmem, &page->flags);
-	atomic_set(&page->count, 1);
+		set_page_count(page, 1);
 		__free_page(page);
 		totalhigh_pages++;
+	} else
+		SetPageReserved(page);
 }
 
 #ifndef CONFIG_DISCONTIGMEM
@@ -431,6 +426,13 @@ static void __init set_max_mapnr_init(void)
 extern void set_max_mapnr_init(void);
 #endif /* !CONFIG_DISCONTIGMEM */
 
+#ifdef CONFIG_HUGETLB_PAGE
+long    htlbpagemem = 0;
+int     htlbpage_max;
+long    htlbzone_pages;
+extern struct list_head htlbpage_freelist;
+#endif
+
 void __init mem_init(void)
 {
 	extern int ppro_with_ram_bug(void);
@@ -493,6 +495,30 @@ void __init mem_init(void)
 #ifndef CONFIG_SMP
 	zap_low_mappings();
 #endif
+#ifdef CONFIG_HUGETLB_PAGE
+	{
+		long	i, j;
+		struct	page	*page, *map;
+		/*For now reserve quarter for hugetlb_pages.*/
+		htlbzone_pages = (max_low_pfn >> ((HPAGE_SHIFT - PAGE_SHIFT) + 2)) ;
+		/*Will make this kernel command line. */
+		INIT_LIST_HEAD(&htlbpage_freelist);
+		for (i=0; i<htlbzone_pages; i++) {
+			page = alloc_pages(GFP_ATOMIC, HUGETLB_PAGE_ORDER);
+			if (page == NULL)
+				break;
+			map = page;
+			for (j=0; j<(HPAGE_SIZE/PAGE_SIZE); j++) {
+				SetPageReserved(map);
+				map++;
+			}
+			list_add(&page->list, &htlbpage_freelist);
+		}
+		printk("Total Huge_TLB_Page memory pages allocated %ld\n", i);
+		htlbzone_pages = htlbpagemem = i;
+		htlbpage_max = i;
+	}
+#endif
 }
 
 #if CONFIG_X86_PAE

arch/ia64/kernel/perfmon.c

@@ -2345,22 +2345,19 @@ static int
 check_task_state(struct task_struct *task)
 {
 	int ret = 0;
-#ifdef CONFIG_SMP
+
 	/* We must wait until the state has been completely
 	 * saved. There can be situations where the reader arrives before
 	 * after the task is marked as STOPPED but before pfm_save_regs()
 	 * is completed.
 	 */
-	if (task->state != TASK_ZOMBIE && task->state != TASK_STOPPED) return -EBUSY;
-	DBprintk(("before wait_task_inactive [%d] state %ld\n", task->pid, task->state));
-	wait_task_inactive(task);
-	DBprintk(("after wait_task_inactive [%d] state %ld\n", task->pid, task->state));
-#else
 	if (task->state != TASK_ZOMBIE && task->state != TASK_STOPPED) {
 		DBprintk(("warning [%d] not in stable state %ld\n", task->pid, task->state));
 		ret = -EBUSY;
 	}
-#endif
+	DBprintk(("before wait_task_inactive [%d] state %ld\n", task->pid, task->state));
+	wait_task_inactive(task);
+	DBprintk(("after wait_task_inactive [%d] state %ld\n", task->pid, task->state));
 	return ret;
 }
 

fs/block_dev.c

@@ -332,6 +332,29 @@ struct block_device *bdget(dev_t dev)
 	return bdev;
 }
 
+long nr_blockdev_pages(void)
+{
+	long ret = 0;
+	int i;
+
+	spin_lock(&bdev_lock);
+	for (i = 0; i < ARRAY_SIZE(bdev_hashtable); i++) {
+		struct list_head *head = &bdev_hashtable[i];
+		struct list_head *lh;
+
+		if (head == NULL)
+			continue;
+		list_for_each(lh, head) {
+			struct block_device *bdev;
+
+			bdev = list_entry(lh, struct block_device, bd_hash);
+			ret += bdev->bd_inode->i_mapping->nrpages;
+		}
+	}
+	spin_unlock(&bdev_lock);
+	return ret;
+}
+
 static inline void __bd_forget(struct inode *inode)
 {
 	list_del_init(&inode->i_devices);

fs/buffer.c

@@ -61,10 +61,8 @@ void __buffer_error(char *file, int line)
 		return;
 	enough++;
 	printk("buffer layer error at %s:%d\n", file, line);
-#ifdef CONFIG_X86
 	printk("Pass this trace through ksymoops for reporting\n");
-	show_stack(0);
-#endif
+	dump_stack();
 }
 EXPORT_SYMBOL(__buffer_error);
 

fs/ext3/balloc.c

@@ -46,18 +46,18 @@ struct ext3_group_desc * ext3_get_group_desc(struct super_block * sb,
 	unsigned long desc;
 	struct ext3_group_desc * gdp;
 
-	if (block_group >= sb->u.ext3_sb.s_groups_count) {
+	if (block_group >= EXT3_SB(sb)->s_groups_count) {
 		ext3_error (sb, "ext3_get_group_desc",
 			    "block_group >= groups_count - "
 			    "block_group = %d, groups_count = %lu",
-			    block_group, sb->u.ext3_sb.s_groups_count);
+			    block_group, EXT3_SB(sb)->s_groups_count);
 
 		return NULL;
 	}
 	
 	group_desc = block_group / EXT3_DESC_PER_BLOCK(sb);
 	desc = block_group % EXT3_DESC_PER_BLOCK(sb);
-	if (!sb->u.ext3_sb.s_group_desc[group_desc]) {
+	if (!EXT3_SB(sb)->s_group_desc[group_desc]) {
 		ext3_error (sb, "ext3_get_group_desc",
 			    "Group descriptor not loaded - "
 			    "block_group = %d, group_desc = %lu, desc = %lu",
@@ -66,9 +66,9 @@ struct ext3_group_desc * ext3_get_group_desc(struct super_block * sb,
 	}
 	
 	gdp = (struct ext3_group_desc *) 
-	      sb->u.ext3_sb.s_group_desc[group_desc]->b_data;
+	      EXT3_SB(sb)->s_group_desc[group_desc]->b_data;
 	if (bh)
-		*bh = sb->u.ext3_sb.s_group_desc[group_desc];
+		*bh = EXT3_SB(sb)->s_group_desc[group_desc];
 	return gdp + desc;
 }
 
@@ -119,7 +119,7 @@ void ext3_free_blocks (handle_t *handle, struct inode * inode,
 		return;
 	}
 	lock_super (sb);
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 	if (block < le32_to_cpu(es->s_first_data_block) || 
 	    (block + count) > le32_to_cpu(es->s_blocks_count)) {
 		ext3_error (sb, "ext3_free_blocks",
@@ -155,9 +155,9 @@ void ext3_free_blocks (handle_t *handle, struct inode * inode,
 	if (in_range (le32_to_cpu(gdp->bg_block_bitmap), block, count) ||
 	    in_range (le32_to_cpu(gdp->bg_inode_bitmap), block, count) ||
 	    in_range (block, le32_to_cpu(gdp->bg_inode_table),
-		      sb->u.ext3_sb.s_itb_per_group) ||
+		      EXT3_SB(sb)->s_itb_per_group) ||
 	    in_range (block + count - 1, le32_to_cpu(gdp->bg_inode_table),
-		      sb->u.ext3_sb.s_itb_per_group))
+		      EXT3_SB(sb)->s_itb_per_group))
 		ext3_error (sb, "ext3_free_blocks",
 			    "Freeing blocks in system zones - "
 			    "Block = %lu, count = %lu",
@@ -183,8 +183,8 @@ void ext3_free_blocks (handle_t *handle, struct inode * inode,
 	if (err)
 		goto error_return;
 
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "get_write_access");
-	err = ext3_journal_get_write_access(handle, sb->u.ext3_sb.s_sbh);
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "get_write_access");
+	err = ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh);
 	if (err)
 		goto error_return;
 
@@ -253,8 +253,8 @@ void ext3_free_blocks (handle_t *handle, struct inode * inode,
 	if (!err) err = ret;
 
 	/* And the superblock */
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "dirtied superblock");
-	ret = ext3_journal_dirty_metadata(handle, sb->u.ext3_sb.s_sbh);
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "dirtied superblock");
+	ret = ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh);
 	if (!err) err = ret;
 
 	if (overflow && !err) {
@@ -408,12 +408,12 @@ ext3_new_block(handle_t *handle, struct inode *inode, unsigned long goal,
 	}
 
 	lock_super(sb);
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 	if (le32_to_cpu(es->s_free_blocks_count) <=
 			le32_to_cpu(es->s_r_blocks_count) &&
-	    ((sb->u.ext3_sb.s_resuid != current->fsuid) &&
-	     (sb->u.ext3_sb.s_resgid == 0 ||
-	      !in_group_p(sb->u.ext3_sb.s_resgid)) && 
+	    ((EXT3_SB(sb)->s_resuid != current->fsuid) &&
+	     (EXT3_SB(sb)->s_resgid == 0 ||
+	      !in_group_p(EXT3_SB(sb)->s_resgid)) && 
 	     !capable(CAP_SYS_RESOURCE)))
 		goto out;
 
@@ -464,9 +464,9 @@ ext3_new_block(handle_t *handle, struct inode *inode, unsigned long goal,
 	 * Now search the rest of the groups.  We assume that 
 	 * i and gdp correctly point to the last group visited.
 	 */
-	for (bit = 0; bit < sb->u.ext3_sb.s_groups_count; bit++) {
+	for (bit = 0; bit < EXT3_SB(sb)->s_groups_count; bit++) {
 		group_no++;
-		if (group_no >= sb->u.ext3_sb.s_groups_count)
+		if (group_no >= EXT3_SB(sb)->s_groups_count)
 			group_no = 0;
 		gdp = ext3_get_group_desc(sb, group_no, &gdp_bh);
 		if (!gdp) {
@@ -518,8 +518,8 @@ ext3_new_block(handle_t *handle, struct inode *inode, unsigned long goal,
 	if (fatal)
 		goto out;
 
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "get_write_access");
-	fatal = ext3_journal_get_write_access(handle, sb->u.ext3_sb.s_sbh);
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "get_write_access");
+	fatal = ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh);
 	if (fatal)
 		goto out;
 
@@ -529,7 +529,7 @@ ext3_new_block(handle_t *handle, struct inode *inode, unsigned long goal,
 	if (target_block == le32_to_cpu(gdp->bg_block_bitmap) ||
 	    target_block == le32_to_cpu(gdp->bg_inode_bitmap) ||
 	    in_range(target_block, le32_to_cpu(gdp->bg_inode_table),
-		      sb->u.ext3_sb.s_itb_per_group))
+		      EXT3_SB(sb)->s_itb_per_group))
 		ext3_error(sb, "ext3_new_block",
 			    "Allocating block in system zone - "
 			    "block = %u", target_block);
@@ -594,9 +594,9 @@ ext3_new_block(handle_t *handle, struct inode *inode, unsigned long goal,
 	if (!fatal)
 		fatal = err;
 
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh,
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh,
 			"journal_dirty_metadata for superblock");
-	err = ext3_journal_dirty_metadata(handle, sb->u.ext3_sb.s_sbh);
+	err = ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh);
 	if (!fatal)
 		fatal = err;
 
@@ -637,11 +637,11 @@ unsigned long ext3_count_free_blocks(struct super_block *sb)
 	int i;
 	
 	lock_super(sb);
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 	desc_count = 0;
 	bitmap_count = 0;
 	gdp = NULL;
-	for (i = 0; i < sb->u.ext3_sb.s_groups_count; i++) {
+	for (i = 0; i < EXT3_SB(sb)->s_groups_count; i++) {
 		gdp = ext3_get_group_desc(sb, i, NULL);
 		if (!gdp)
 			continue;
@@ -662,7 +662,7 @@ unsigned long ext3_count_free_blocks(struct super_block *sb)
 	unlock_super(sb);
 	return bitmap_count;
 #else
-	return le32_to_cpu(sb->u.ext3_sb.s_es->s_free_blocks_count);
+	return le32_to_cpu(EXT3_SB(sb)->s_es->s_free_blocks_count);
 #endif
 }
 
@@ -671,7 +671,7 @@ static inline int block_in_use(unsigned long block,
 				unsigned char * map)
 {
 	return ext3_test_bit ((block -
-		le32_to_cpu(sb->u.ext3_sb.s_es->s_first_data_block)) %
+		le32_to_cpu(EXT3_SB(sb)->s_es->s_first_data_block)) %
 			 EXT3_BLOCKS_PER_GROUP(sb), map);
 }
 
@@ -738,11 +738,11 @@ void ext3_check_blocks_bitmap (struct super_block * sb)
 	struct ext3_group_desc *gdp;
 	int i;
 
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 	desc_count = 0;
 	bitmap_count = 0;
 	gdp = NULL;
-	for (i = 0; i < sb->u.ext3_sb.s_groups_count; i++) {
+	for (i = 0; i < EXT3_SB(sb)->s_groups_count; i++) {
 		gdp = ext3_get_group_desc (sb, i, NULL);
 		if (!gdp)
 			continue;
@@ -776,7 +776,7 @@ void ext3_check_blocks_bitmap (struct super_block * sb)
 				    "Inode bitmap for group %d is marked free",
 				    i);
 
-		for (j = 0; j < sb->u.ext3_sb.s_itb_per_group; j++)
+		for (j = 0; j < EXT3_SB(sb)->s_itb_per_group; j++)
 			if (!block_in_use (le32_to_cpu(gdp->bg_inode_table) + j,
 							sb, bitmap_bh->b_data))
 				ext3_error (sb, "ext3_check_blocks_bitmap",

fs/ext3/dir.c

@@ -54,7 +54,7 @@ int ext3_check_dir_entry (const char * function, struct inode * dir,
 	else if (((char *) de - bh->b_data) + rlen > dir->i_sb->s_blocksize)
 		error_msg = "directory entry across blocks";
 	else if (le32_to_cpu(de->inode) >
-			le32_to_cpu(dir->i_sb->u.ext3_sb.s_es->s_inodes_count))
+			le32_to_cpu(EXT3_SB(dir->i_sb)->s_es->s_inodes_count))
 		error_msg = "inode out of bounds";
 
 	if (error_msg != NULL)

fs/ext3/ialloc.c

@@ -127,7 +127,7 @@ void ext3_free_inode (handle_t *handle, struct inode * inode)
 	clear_inode (inode);
 
 	lock_super (sb);
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 	if (ino < EXT3_FIRST_INO(sb) || ino > le32_to_cpu(es->s_inodes_count)) {
 		ext3_error (sb, "ext3_free_inode",
 			    "reserved or nonexistent inode %lu", ino);
@@ -155,8 +155,8 @@ void ext3_free_inode (handle_t *handle, struct inode * inode)
 		fatal = ext3_journal_get_write_access(handle, bh2);
 		if (fatal) goto error_return;
 
-		BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "get write access");
-		fatal = ext3_journal_get_write_access(handle, sb->u.ext3_sb.s_sbh);
+		BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "get write access");
+		fatal = ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh);
 		if (fatal) goto error_return;
 
 		if (gdp) {
@@ -171,9 +171,9 @@ void ext3_free_inode (handle_t *handle, struct inode * inode)
 		if (!fatal) fatal = err;
 		es->s_free_inodes_count =
 			cpu_to_le32(le32_to_cpu(es->s_free_inodes_count) + 1);
-		BUFFER_TRACE(sb->u.ext3_sb.s_sbh,
+		BUFFER_TRACE(EXT3_SB(sb)->s_sbh,
 					"call ext3_journal_dirty_metadata");
-		err = ext3_journal_dirty_metadata(handle, sb->u.ext3_sb.s_sbh);
+		err = ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh);
 		if (!fatal) fatal = err;
 	}
 	BUFFER_TRACE(bitmap_bh, "call ext3_journal_dirty_metadata");
@@ -222,16 +222,16 @@ struct inode *ext3_new_inode(handle_t *handle, struct inode * dir, int mode)
 	ei = EXT3_I(inode);
 
 	lock_super (sb);
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 repeat:
 	gdp = NULL;
 	i = 0;
 
 	if (S_ISDIR(mode)) {
 		avefreei = le32_to_cpu(es->s_free_inodes_count) /
-			sb->u.ext3_sb.s_groups_count;
+			EXT3_SB(sb)->s_groups_count;
 		if (!gdp) {
-			for (j = 0; j < sb->u.ext3_sb.s_groups_count; j++) {
+			for (j = 0; j < EXT3_SB(sb)->s_groups_count; j++) {
 				struct buffer_head *temp_buffer;
 				tmp = ext3_get_group_desc (sb, j, &temp_buffer);
 				if (tmp &&
@@ -261,10 +261,10 @@ struct inode *ext3_new_inode(handle_t *handle, struct inode * dir, int mode)
 			 * Use a quadratic hash to find a group with a
 			 * free inode
 			 */
-			for (j = 1; j < sb->u.ext3_sb.s_groups_count; j <<= 1) {
+			for (j = 1; j < EXT3_SB(sb)->s_groups_count; j <<= 1) {
 				i += j;
-				if (i >= sb->u.ext3_sb.s_groups_count)
-					i -= sb->u.ext3_sb.s_groups_count;
+				if (i >= EXT3_SB(sb)->s_groups_count)
+					i -= EXT3_SB(sb)->s_groups_count;
 				tmp = ext3_get_group_desc (sb, i, &bh2);
 				if (tmp &&
 				    le16_to_cpu(tmp->bg_free_inodes_count)) {
@@ -278,8 +278,8 @@ struct inode *ext3_new_inode(handle_t *handle, struct inode * dir, int mode)
 			 * That failed: try linear search for a free inode
 			 */
 			i = EXT3_I(dir)->i_block_group + 1;
-			for (j = 2; j < sb->u.ext3_sb.s_groups_count; j++) {
-				if (++i >= sb->u.ext3_sb.s_groups_count)
+			for (j = 2; j < EXT3_SB(sb)->s_groups_count; j++) {
+				if (++i >= EXT3_SB(sb)->s_groups_count)
 					i = 0;
 				tmp = ext3_get_group_desc (sb, i, &bh2);
 				if (tmp &&
@@ -357,13 +357,13 @@ struct inode *ext3_new_inode(handle_t *handle, struct inode * dir, int mode)
 	err = ext3_journal_dirty_metadata(handle, bh2);
 	if (err) goto fail;
 	
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "get_write_access");
-	err = ext3_journal_get_write_access(handle, sb->u.ext3_sb.s_sbh);
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "get_write_access");
+	err = ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh);
 	if (err) goto fail;
 	es->s_free_inodes_count =
 		cpu_to_le32(le32_to_cpu(es->s_free_inodes_count) - 1);
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "call ext3_journal_dirty_metadata");
-	err = ext3_journal_dirty_metadata(handle, sb->u.ext3_sb.s_sbh);
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "call ext3_journal_dirty_metadata");
+	err = ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh);
 	sb->s_dirt = 1;
 	if (err) goto fail;
 
@@ -417,7 +417,7 @@ struct inode *ext3_new_inode(handle_t *handle, struct inode * dir, int mode)
 	if (IS_DIRSYNC(inode))
 		handle->h_sync = 1;
 	insert_inode_hash(inode);
-	inode->i_generation = sb->u.ext3_sb.s_next_generation++;
+	inode->i_generation = EXT3_SB(sb)->s_next_generation++;
 
 	ei->i_state = EXT3_STATE_NEW;
 	err = ext3_mark_inode_dirty(handle, inode);
@@ -512,11 +512,11 @@ unsigned long ext3_count_free_inodes (struct super_block * sb)
 	int i;
 
 	lock_super (sb);
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 	desc_count = 0;
 	bitmap_count = 0;
 	gdp = NULL;
-	for (i = 0; i < sb->u.ext3_sb.s_groups_count; i++) {
+	for (i = 0; i < EXT3_SB(sb)->s_groups_count; i++) {
 		gdp = ext3_get_group_desc (sb, i, NULL);
 		if (!gdp)
 			continue;
@@ -537,7 +537,7 @@ unsigned long ext3_count_free_inodes (struct super_block * sb)
 	unlock_super(sb);
 	return desc_count;
 #else
-	return le32_to_cpu(sb->u.ext3_sb.s_es->s_free_inodes_count);
+	return le32_to_cpu(EXT3_SB(sb)->s_es->s_free_inodes_count);
 #endif
 }
 
@@ -551,11 +551,11 @@ void ext3_check_inodes_bitmap (struct super_block * sb)
 	struct ext3_group_desc * gdp;
 	int i;
 
-	es = sb->u.ext3_sb.s_es;
+	es = EXT3_SB(sb)->s_es;
 	desc_count = 0;
 	bitmap_count = 0;
 	gdp = NULL;
-	for (i = 0; i < sb->u.ext3_sb.s_groups_count; i++) {
+	for (i = 0; i < EXT3_SB(sb)->s_groups_count; i++) {
 		gdp = ext3_get_group_desc (sb, i, NULL);
 		if (!gdp)
 			continue;

fs/ext3/inode.c

@@ -471,7 +471,7 @@ static inline unsigned long ext3_find_near(struct inode *inode, Indirect *ind)
 	 * the same cylinder group then.
 	 */
 	return (ei->i_block_group * EXT3_BLOCKS_PER_GROUP(inode->i_sb)) +
-	       le32_to_cpu(inode->i_sb->u.ext3_sb.s_es->s_first_data_block);
+	       le32_to_cpu(EXT3_SB(inode->i_sb)->s_es->s_first_data_block);
 }
 
 /**
@@ -2141,20 +2141,20 @@ int ext3_get_inode_loc (struct inode *inode, struct ext3_iloc *iloc)
 		inode->i_ino != EXT3_JOURNAL_INO &&
 		inode->i_ino < EXT3_FIRST_INO(inode->i_sb)) ||
 		inode->i_ino > le32_to_cpu(
-			inode->i_sb->u.ext3_sb.s_es->s_inodes_count)) {
+			EXT3_SB(inode->i_sb)->s_es->s_inodes_count)) {
 		ext3_error (inode->i_sb, "ext3_get_inode_loc",
 			    "bad inode number: %lu", inode->i_ino);
 		goto bad_inode;
 	}
 	block_group = (inode->i_ino - 1) / EXT3_INODES_PER_GROUP(inode->i_sb);
-	if (block_group >= inode->i_sb->u.ext3_sb.s_groups_count) {
+	if (block_group >= EXT3_SB(inode->i_sb)->s_groups_count) {
 		ext3_error (inode->i_sb, "ext3_get_inode_loc",
 			    "group >= groups count");
 		goto bad_inode;
 	}
 	group_desc = block_group >> EXT3_DESC_PER_BLOCK_BITS(inode->i_sb);
 	desc = block_group & (EXT3_DESC_PER_BLOCK(inode->i_sb) - 1);
-	bh = inode->i_sb->u.ext3_sb.s_group_desc[group_desc];
+	bh = EXT3_SB(inode->i_sb)->s_group_desc[group_desc];
 	if (!bh) {
 		ext3_error (inode->i_sb, "ext3_get_inode_loc",
 			    "Descriptor not loaded");
@@ -2224,7 +2224,7 @@ void ext3_read_inode(struct inode * inode)
 	 */
 	if (inode->i_nlink == 0) {
 		if (inode->i_mode == 0 ||
-		    !(inode->i_sb->u.ext3_sb.s_mount_state & EXT3_ORPHAN_FS)) {
+		    !(EXT3_SB(inode->i_sb)->s_mount_state & EXT3_ORPHAN_FS)) {
 			/* this inode is deleted */
 			brelse (bh);
 			goto bad_inode;
@@ -2394,7 +2394,7 @@ static int ext3_do_update_inode(handle_t *handle,
 				* created, add a flag to the superblock.
 				*/
 				err = ext3_journal_get_write_access(handle,
-						sb->u.ext3_sb.s_sbh);
+						EXT3_SB(sb)->s_sbh);
 				if (err)
 					goto out_brelse;
 				ext3_update_dynamic_rev(sb);
@@ -2403,7 +2403,7 @@ static int ext3_do_update_inode(handle_t *handle,
 				sb->s_dirt = 1;
 				handle->h_sync = 1;
 				err = ext3_journal_dirty_metadata(handle,
-						sb->u.ext3_sb.s_sbh);
+						EXT3_SB(sb)->s_sbh);
 			}
 		}
 	}

fs/ext3/ioctl.c

@@ -159,12 +159,12 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd,
 			int ret = 0;
 
 			set_current_state(TASK_INTERRUPTIBLE);
-			add_wait_queue(&sb->u.ext3_sb.ro_wait_queue, &wait);
-			if (timer_pending(&sb->u.ext3_sb.turn_ro_timer)) {
+			add_wait_queue(&EXT3_SB(sb)->ro_wait_queue, &wait);
+			if (timer_pending(&EXT3_SB(sb)->turn_ro_timer)) {
 				schedule();
 				ret = 1;
 			}
-			remove_wait_queue(&sb->u.ext3_sb.ro_wait_queue, &wait);
+			remove_wait_queue(&EXT3_SB(sb)->ro_wait_queue, &wait);
 			return ret;
 		}
 #endif

fs/ext3/namei.c

@@ -729,8 +729,8 @@ int ext3_orphan_add(handle_t *handle, struct inode *inode)
 	J_ASSERT ((S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
 		S_ISLNK(inode->i_mode)) || inode->i_nlink == 0);
 
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "get_write_access");
-	err = ext3_journal_get_write_access(handle, sb->u.ext3_sb.s_sbh);
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "get_write_access");
+	err = ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh);
 	if (err)
 		goto out_unlock;
 	
@@ -741,7 +741,7 @@ int ext3_orphan_add(handle_t *handle, struct inode *inode)
 	/* Insert this inode at the head of the on-disk orphan list... */
 	NEXT_ORPHAN(inode) = le32_to_cpu(EXT3_SB(sb)->s_es->s_last_orphan);
 	EXT3_SB(sb)->s_es->s_last_orphan = cpu_to_le32(inode->i_ino);
-	err = ext3_journal_dirty_metadata(handle, sb->u.ext3_sb.s_sbh);
+	err = ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh);
 	rc = ext3_mark_iloc_dirty(handle, inode, &iloc);
 	if (!err)
 		err = rc;

fs/ext3/super.c

@@ -120,7 +120,7 @@ static int ext3_error_behaviour(struct super_block *sb)
 	/* If no overrides were specified on the mount, then fall back
 	 * to the default behaviour set in the filesystem's superblock
 	 * on disk. */
-	switch (le16_to_cpu(sb->u.ext3_sb.s_es->s_errors)) {
+	switch (le16_to_cpu(EXT3_SB(sb)->s_es->s_errors)) {
 	case EXT3_ERRORS_PANIC:
 		return EXT3_ERRORS_PANIC;
 	case EXT3_ERRORS_RO:
@@ -268,9 +268,9 @@ void ext3_abort (struct super_block * sb, const char * function,
 		return;
 	
 	printk (KERN_CRIT "Remounting filesystem read-only\n");
-	sb->u.ext3_sb.s_mount_state |= EXT3_ERROR_FS;
+	EXT3_SB(sb)->s_mount_state |= EXT3_ERROR_FS;
 	sb->s_flags |= MS_RDONLY;
-	sb->u.ext3_sb.s_mount_opt |= EXT3_MOUNT_ABORT;
+	EXT3_SB(sb)->s_mount_opt |= EXT3_MOUNT_ABORT;
 	journal_abort(EXT3_SB(sb)->s_journal, -EIO);
 }
 
@@ -439,7 +439,8 @@ void ext3_put_super (struct super_block * sb)
 		ext3_blkdev_remove(sbi);
 	}
 	clear_ro_after(sb);
-
+	sb->u.generic_sbp = NULL;
+	kfree(sbi);
 	return;
 }
 
@@ -877,7 +878,7 @@ static void ext3_orphan_cleanup (struct super_block * sb,
 		sb->s_flags &= ~MS_RDONLY;
 	}
 
-	if (sb->u.ext3_sb.s_mount_state & EXT3_ERROR_FS) {
+	if (EXT3_SB(sb)->s_mount_state & EXT3_ERROR_FS) {
 		if (es->s_last_orphan)
 			jbd_debug(1, "Errors on filesystem, "
 				  "clearing orphan list.\n");
@@ -949,7 +950,7 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 {
 	struct buffer_head * bh;
 	struct ext3_super_block *es = 0;
-	struct ext3_sb_info *sbi = EXT3_SB(sb);
+	struct ext3_sb_info *sbi;
 	unsigned long sb_block = 1;
 	unsigned long logic_sb_block = 1;
 	unsigned long offset = 0;
@@ -970,7 +971,11 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	 * This is important for devices that have a hardware
 	 * sectorsize that is larger than the default.
 	 */
-
+	sbi = kmalloc(sizeof(*sbi), GFP_KERNEL);
+	if (!sbi)
+		return -ENOMEM;
+	sb->u.generic_sbp = sbi;
+	memset(sbi, 0, sizeof(*sbi));
 	sbi->s_mount_opt = 0;
 	sbi->s_resuid = EXT3_DEF_RESUID;
 	sbi->s_resgid = EXT3_DEF_RESGID;
@@ -1266,6 +1271,8 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	ext3_blkdev_remove(sbi);
 	brelse(bh);
 out_fail:
+	sb->u.generic_sbp = NULL;
+	kfree(sbi);
 	return -EINVAL;
 }
 
@@ -1520,11 +1527,11 @@ static void ext3_commit_super (struct super_block * sb,
 			       int sync)
 {
 	es->s_wtime = cpu_to_le32(CURRENT_TIME);
-	BUFFER_TRACE(sb->u.ext3_sb.s_sbh, "marking dirty");
-	mark_buffer_dirty(sb->u.ext3_sb.s_sbh);
+	BUFFER_TRACE(EXT3_SB(sb)->s_sbh, "marking dirty");
+	mark_buffer_dirty(EXT3_SB(sb)->s_sbh);
 	if (sync) {
-		ll_rw_block(WRITE, 1, &sb->u.ext3_sb.s_sbh);
-		wait_on_buffer(sb->u.ext3_sb.s_sbh);
+		ll_rw_block(WRITE, 1, &EXT3_SB(sb)->s_sbh);
+		wait_on_buffer(EXT3_SB(sb)->s_sbh);
 	}
 }
 
@@ -1575,7 +1582,7 @@ static void ext3_clear_journal_err(struct super_block * sb,
 		ext3_warning(sb, __FUNCTION__, "Marking fs in need of "
 			     "filesystem check.");
 		
-		sb->u.ext3_sb.s_mount_state |= EXT3_ERROR_FS;
+		EXT3_SB(sb)->s_mount_state |= EXT3_ERROR_FS;
 		es->s_state |= cpu_to_le16(EXT3_ERROR_FS);
 		ext3_commit_super (sb, es, 1);
 

fs/proc/array.c

@@ -487,7 +487,18 @@ int proc_pid_statm(struct task_struct *task, char * buffer)
 		while (vma) {
 			pgd_t *pgd = pgd_offset(mm, vma->vm_start);
 			int pages = 0, shared = 0, dirty = 0, total = 0;
+			if (is_vm_hugetlb_page(vma)) {
+				int num_pages = ((vma->vm_end - vma->vm_start)/PAGE_SIZE);
 
+				resident += num_pages;
+				if (!(vma->vm_flags & VM_DONTCOPY))
+					share += num_pages;
+				if (vma->vm_flags & VM_WRITE)
+					dt += num_pages;
+				drs += num_pages;
+				vma = vma->vm_next;
+				continue;
+			}
 			statm_pgd_range(pgd, vma->vm_start, vma->vm_end, &pages, &shared, &dirty, &total);
 			resident += pages;
 			share += shared;

fs/proc/proc_misc.c

@@ -136,16 +136,8 @@ static int meminfo_read_proc(char *page, char **start, off_t off,
 	struct sysinfo i;
 	int len, committed;
 	struct page_state ps;
-	int cpu;
 	unsigned long inactive;
 	unsigned long active;
-	unsigned long flushes = 0;
-	unsigned long non_flushes = 0;
-
-	for (cpu = 0; cpu < NR_CPUS; cpu++) {
-		flushes += mmu_gathers[cpu].flushes;
-		non_flushes += mmu_gathers[cpu].avoided_flushes;
-	}
 
 	get_page_state(&ps);
 	get_zone_counts(&active, &inactive);
@@ -165,6 +157,7 @@ static int meminfo_read_proc(char *page, char **start, off_t off,
 		"MemTotal:     %8lu kB\n"
 		"MemFree:      %8lu kB\n"
 		"MemShared:    %8lu kB\n"
+		"Buffers:      %8lu kB\n"
 		"Cached:       %8lu kB\n"
 		"SwapCached:   %8lu kB\n"
 		"Active:       %8lu kB\n"
@@ -177,15 +170,15 @@ static int meminfo_read_proc(char *page, char **start, off_t off,
 		"SwapFree:     %8lu kB\n"
 		"Dirty:        %8lu kB\n"
 		"Writeback:    %8lu kB\n"
+		"Mapped:       %8lu kB\n"
 		"Committed_AS: %8u kB\n"
 		"PageTables:   %8lu kB\n"
-		"ReverseMaps:  %8lu\n"
-		"TLB flushes:  %8lu\n"
-		"non flushes:  %8lu\n",
+		"ReverseMaps:  %8lu\n",
 		K(i.totalram),
 		K(i.freeram),
 		K(i.sharedram),
-		K(ps.nr_pagecache-swapper_space.nrpages),
+		K(i.bufferram),
+		K(ps.nr_pagecache-swapper_space.nrpages-i.bufferram),
 		K(swapper_space.nrpages),
 		K(active),
 		K(inactive),
@@ -197,13 +190,25 @@ static int meminfo_read_proc(char *page, char **start, off_t off,
 		K(i.freeswap),
 		K(ps.nr_dirty),
 		K(ps.nr_writeback),
+		K(ps.nr_mapped),
 		K(committed),
 		K(ps.nr_page_table_pages),
-		ps.nr_reverse_maps,
-		flushes,
-		non_flushes
+		ps.nr_reverse_maps
 		);
 
+#ifdef CONFIG_HUGETLB_PAGE
+	{
+		extern unsigned long htlbpagemem, htlbzone_pages;
+		len += sprintf(page + len,
+				"HugePages:    %8lu\n"
+				"Available:    %8lu\n"
+				"Size:         %8lu kB\n",
+				htlbzone_pages,
+				htlbpagemem,
+				HPAGE_SIZE/1024);
+	}
+
+#endif
 	return proc_calc_metrics(page, start, off, count, eof, len);
 #undef K
 }

include/asm-generic/tlb.h

@@ -21,7 +21,7 @@
  * and page free order so much..
  */
 #ifdef CONFIG_SMP
-  #define FREE_PTE_NR	507
+  #define FREE_PTE_NR	506
   #define tlb_fast_mode(tlb) ((tlb)->nr == ~0U)
 #else
   #define FREE_PTE_NR	1
@@ -40,8 +40,6 @@ typedef struct free_pte_ctx {
 	unsigned int		fullmm; /* non-zero means full mm flush */
 	unsigned long		freed;
 	struct page *		pages[FREE_PTE_NR];
-	unsigned long		flushes;/* stats: count avoided flushes */
-	unsigned long		avoided_flushes;
 } mmu_gather_t;
 
 /* Users of the generic TLB shootdown code must declare this storage space. */
@@ -67,17 +65,10 @@ static inline mmu_gather_t *tlb_gather_mmu(struct mm_struct *mm, unsigned int fu
 
 static inline void tlb_flush_mmu(mmu_gather_t *tlb, unsigned long start, unsigned long end)
 {
-	unsigned long nr;
-
-	if (!tlb->need_flush) {
-		tlb->avoided_flushes++;
+	if (!tlb->need_flush)
 		return;
-	}
 	tlb->need_flush = 0;
-	tlb->flushes++;
-
 	tlb_flush(tlb);
-	nr = tlb->nr;
 	if (!tlb_fast_mode(tlb)) {
 		free_pages_and_swap_cache(tlb->pages, tlb->nr);
 		tlb->nr = 0;

include/asm-i386/page.h

@@ -44,14 +44,22 @@ typedef struct { unsigned long pte_low, pte_high; } pte_t;
 typedef struct { unsigned long long pmd; } pmd_t;
 typedef struct { unsigned long long pgd; } pgd_t;
 #define pte_val(x)	((x).pte_low | ((unsigned long long)(x).pte_high << 32))
+#define HPAGE_SHIFT	21
 #else
 typedef struct { unsigned long pte_low; } pte_t;
 typedef struct { unsigned long pmd; } pmd_t;
 typedef struct { unsigned long pgd; } pgd_t;
 #define pte_val(x)	((x).pte_low)
+#define HPAGE_SHIFT	22
 #endif
 #define PTE_MASK	PAGE_MASK
 
+#ifdef CONFIG_HUGETLB_PAGE
+#define HPAGE_SIZE	((1UL) << HPAGE_SHIFT)
+#define HPAGE_MASK	(~(HPAGE_SIZE - 1))
+#define HUGETLB_PAGE_ORDER	(HPAGE_SHIFT - PAGE_SHIFT)
+#endif
+
 typedef struct { unsigned long pgprot; } pgprot_t;
 
 #define pmd_val(x)	((x).pmd)

include/linux/blkdev.h

@@ -327,7 +327,7 @@ extern struct backing_dev_info *blk_get_backing_dev_info(struct block_device *bd
 extern int blk_rq_map_sg(request_queue_t *, struct request *, struct scatterlist *);
 extern void blk_dump_rq_flags(struct request *, char *);
 extern void generic_unplug_device(void *);
-
+extern long nr_blockdev_pages(void);
 
 /*
  * tag stuff

include/linux/ext3_fs.h

@@ -97,9 +97,9 @@
 # define EXT3_BLOCK_SIZE_BITS(s)	((s)->s_log_block_size + 10)
 #endif
 #ifdef __KERNEL__
-#define	EXT3_ADDR_PER_BLOCK_BITS(s)	((s)->u.ext3_sb.s_addr_per_block_bits)
-#define EXT3_INODE_SIZE(s)		((s)->u.ext3_sb.s_inode_size)
-#define EXT3_FIRST_INO(s)		((s)->u.ext3_sb.s_first_ino)
+#define	EXT3_ADDR_PER_BLOCK_BITS(s)	(EXT3_SB(s)->s_addr_per_block_bits)
+#define EXT3_INODE_SIZE(s)		(EXT3_SB(s)->s_inode_size)
+#define EXT3_FIRST_INO(s)		(EXT3_SB(s)->s_first_ino)
 #else
 #define EXT3_INODE_SIZE(s)	(((s)->s_rev_level == EXT3_GOOD_OLD_REV) ? \
 				 EXT3_GOOD_OLD_INODE_SIZE : \
@@ -116,8 +116,8 @@
 #define	EXT3_MAX_FRAG_SIZE		4096
 #define EXT3_MIN_FRAG_LOG_SIZE		  10
 #ifdef __KERNEL__
-# define EXT3_FRAG_SIZE(s)		((s)->u.ext3_sb.s_frag_size)
-# define EXT3_FRAGS_PER_BLOCK(s)	((s)->u.ext3_sb.s_frags_per_block)
+# define EXT3_FRAG_SIZE(s)		(EXT3_SB(s)->s_frag_size)
+# define EXT3_FRAGS_PER_BLOCK(s)	(EXT3_SB(s)->s_frags_per_block)
 #else
 # define EXT3_FRAG_SIZE(s)		(EXT3_MIN_FRAG_SIZE << (s)->s_log_frag_size)
 # define EXT3_FRAGS_PER_BLOCK(s)	(EXT3_BLOCK_SIZE(s) / EXT3_FRAG_SIZE(s))
@@ -164,10 +164,10 @@ struct ext3_group_desc
  * Macro-instructions used to manage group descriptors
  */
 #ifdef __KERNEL__
-# define EXT3_BLOCKS_PER_GROUP(s)	((s)->u.ext3_sb.s_blocks_per_group)
-# define EXT3_DESC_PER_BLOCK(s)		((s)->u.ext3_sb.s_desc_per_block)
-# define EXT3_INODES_PER_GROUP(s)	((s)->u.ext3_sb.s_inodes_per_group)
-# define EXT3_DESC_PER_BLOCK_BITS(s)	((s)->u.ext3_sb.s_desc_per_block_bits)
+# define EXT3_BLOCKS_PER_GROUP(s)	(EXT3_SB(s)->s_blocks_per_group)
+# define EXT3_DESC_PER_BLOCK(s)		(EXT3_SB(s)->s_desc_per_block)
+# define EXT3_INODES_PER_GROUP(s)	(EXT3_SB(s)->s_inodes_per_group)
+# define EXT3_DESC_PER_BLOCK_BITS(s)	(EXT3_SB(s)->s_desc_per_block_bits)
 #else
 # define EXT3_BLOCKS_PER_GROUP(s)	((s)->s_blocks_per_group)
 # define EXT3_DESC_PER_BLOCK(s)		(EXT3_BLOCK_SIZE(s) / sizeof (struct ext3_group_desc))
@@ -346,7 +346,7 @@ struct ext3_inode {
 #ifndef _LINUX_EXT2_FS_H
 #define clear_opt(o, opt)		o &= ~EXT3_MOUNT_##opt
 #define set_opt(o, opt)			o |= EXT3_MOUNT_##opt
-#define test_opt(sb, opt)		((sb)->u.ext3_sb.s_mount_opt & \
+#define test_opt(sb, opt)		(EXT3_SB(sb)->s_mount_opt & \
 					 EXT3_MOUNT_##opt)
 #else
 #define EXT2_MOUNT_NOLOAD		EXT3_MOUNT_NOLOAD
@@ -444,7 +444,10 @@ struct ext3_super_block {
 };
 
 #ifdef __KERNEL__
-#define EXT3_SB(sb)	(&((sb)->u.ext3_sb))
+static inline struct ext3_sb_info * EXT3_SB(struct super_block *sb)
+{
+	return sb->u.generic_sbp;
+}
 static inline struct ext3_inode_info *EXT3_I(struct inode *inode)
 {
 	return container_of(inode, struct ext3_inode_info, vfs_inode);

include/linux/kernel.h

@@ -96,6 +96,8 @@ extern const char *print_tainted(void);
 #define TAINT_FORCED_MODULE		(1<<1)
 #define TAINT_UNSAFE_SMP		(1<<2)
 
+extern void dump_stack(void);
+
 #if DEBUG
 #define pr_debug(fmt,arg...) \
 	printk(KERN_DEBUG fmt,##arg)

include/linux/mm.h

@@ -19,9 +19,6 @@ extern unsigned long max_mapnr;
 extern unsigned long num_physpages;
 extern void * high_memory;
 extern int page_cluster;
-/* The inactive_clean lists are per zone. */
-extern struct list_head active_list;
-extern struct list_head inactive_list;
 
 #include <asm/page.h>
 #include <asm/pgtable.h>
@@ -104,6 +101,7 @@ struct vm_area_struct {
 #define VM_DONTEXPAND	0x00040000	/* Cannot expand with mremap() */
 #define VM_RESERVED	0x00080000	/* Don't unmap it from swap_out */
 #define VM_ACCOUNT	0x00100000	/* Is a VM accounted object */
+#define VM_HUGETLB	0x00400000	/* Huge TLB Page VM */
 
 #define VM_STACK_FLAGS	(0x00000100 | VM_DATA_DEFAULT_FLAGS | VM_ACCOUNT)
 
@@ -377,6 +375,20 @@ int get_user_pages(struct task_struct *tsk, struct mm_struct *mm, unsigned long
 int __set_page_dirty_buffers(struct page *page);
 int __set_page_dirty_nobuffers(struct page *page);
 
+#ifdef CONFIG_HUGETLB_PAGE
+#define is_vm_hugetlb_page(vma) (vma->vm_flags & VM_HUGETLB)
+extern int copy_hugetlb_page_range(struct mm_struct *, struct mm_struct *, struct vm_area_struct *);
+extern int follow_hugetlb_page(struct mm_struct *, struct vm_area_struct *, struct page **, struct vm_area_struct **, unsigned long *, int *, int);
+extern	int free_hugepages(struct vm_area_struct *);
+
+#else
+#define is_vm_hugetlb_page(vma) (0)
+#define follow_hugetlb_page(mm, vma, pages, vmas, start, len, i) (0)
+#define copy_hugetlb_page_range(dst, src, vma) (0)
+#define free_hugepages(mpnt)  do { } while(0)
+#endif
+
+
 /*
  * If the mapping doesn't provide a set_page_dirty a_op, then
  * just fall through and assume that it wants buffer_heads.

include/linux/mmzone.h

@@ -16,7 +16,7 @@
  */
 
 #ifndef CONFIG_FORCE_MAX_ZONEORDER
-#define MAX_ORDER 10
+#define MAX_ORDER 11
 #else
 #define MAX_ORDER CONFIG_FORCE_MAX_ZONEORDER
 #endif
@@ -151,8 +151,8 @@ struct zonelist {
  * On NUMA machines, each NUMA node would have a pg_data_t to describe
  * it's memory layout.
  *
- * XXX: we need to move the global memory statistics (active_list, ...)
- *      into the pg_data_t to properly support NUMA.
+ * Memory statistics and page replacement data structures are maintained on a
+ * per-zone basis.
  */
 struct bootmem_data;
 typedef struct pglist_data {

include/linux/page-flags.h

@@ -78,6 +78,7 @@ extern struct page_state {
 	unsigned long nr_pagecache;
 	unsigned long nr_page_table_pages;
 	unsigned long nr_reverse_maps;
+	unsigned long nr_mapped;
 } ____cacheline_aligned_in_smp page_states[NR_CPUS];
 
 extern void get_page_state(struct page_state *ret);

include/linux/sched.h

@@ -690,7 +690,11 @@ extern void FASTCALL(add_wait_queue(wait_queue_head_t *q, wait_queue_t * wait));
 extern void FASTCALL(add_wait_queue_exclusive(wait_queue_head_t *q, wait_queue_t * wait));
 extern void FASTCALL(remove_wait_queue(wait_queue_head_t *q, wait_queue_t * wait));
 
+#ifdef CONFIG_SMP
 extern void wait_task_inactive(task_t * p);
+#else
+#define wait_task_inactive(p)	do { } while (0)
+#endif
 extern void kick_if_running(task_t * p);
 
 #define __wait_event(wq, condition) 					\
@@ -956,6 +960,34 @@ static inline void cond_resched(void)
 		__cond_resched();
 }
 
+#ifdef CONFIG_PREEMPT
+
+/*
+ * cond_resched_lock() - if a reschedule is pending, drop the given lock,
+ * call schedule, and on return reacquire the lock.
+ *
+ * Note: this does not assume the given lock is the _only_ lock held.
+ * The kernel preemption counter gives us "free" checking that we are
+ * atomic -- let's use it.
+ */
+static inline void cond_resched_lock(spinlock_t * lock)
+{
+	if (need_resched() && preempt_count() == 1) {
+		_raw_spin_unlock(lock);
+		preempt_enable_no_resched();
+		__cond_resched();
+		spin_lock(lock);
+	}
+}
+
+#else
+
+static inline void cond_resched_lock(spinlock_t * lock)
+{
+}
+
+#endif
+
 /* Reevaluate whether the task has signals pending delivery.
    This is required every time the blocked sigset_t changes.
    Athread cathreaders should have t->sigmask_lock.  */

include/linux/sysctl.h

@@ -128,6 +128,7 @@ enum
 	KERN_TAINTED=53,	/* int: various kernel tainted flags */
 	KERN_CADPID=54,		/* int: PID of the process to notify on CAD */
 	KERN_PIDMAX=55,		/* int: PID # limit */
+	KERN_HUGETLB_PAGE_NUM=56, /* int: Number of available Huge Pages */
 };
 
 

kernel/exit.c

@@ -55,10 +55,8 @@ void release_task(struct task_struct * p)
 
 	if (p->state != TASK_ZOMBIE)
 		BUG();
-#ifdef CONFIG_SMP
 	if (p != current)
 		wait_task_inactive(p);
-#endif
 	atomic_dec(&p->user->processes);
 	security_ops->task_free_security(p);
 	free_uid(p->user);

kernel/ksyms.c

@@ -605,3 +605,6 @@ EXPORT_SYMBOL(pidhash);
 #if defined(CONFIG_SMP) && defined(__GENERIC_PER_CPU)
 EXPORT_SYMBOL(__per_cpu_offset);
 #endif
+
+/* debug */
+EXPORT_SYMBOL(dump_stack);

kernel/ptrace.c

@@ -69,9 +69,7 @@ int ptrace_check_attach(struct task_struct *child, int kill)
 	if (!kill) {
 		if (child->state != TASK_STOPPED)
 			return -ESRCH;
-#ifdef CONFIG_SMP
 		wait_task_inactive(child);
-#endif		
 	}
 
 	/* All systems go.. */

kernel/sysctl.c

@@ -98,6 +98,11 @@ int proc_dol2crvec(ctl_table *table, int write, struct file *filp,
 extern int acct_parm[];
 #endif
 
+#ifdef CONFIG_HUGETLB_PAGE
+extern	int htlbpage_max;
+extern	int	set_hugetlb_mem_size(int);
+#endif
+
 static int parse_table(int *, int, void *, size_t *, void *, size_t,
 		       ctl_table *, void **);
 static int proc_doutsstring(ctl_table *table, int write, struct file *filp,
@@ -258,6 +263,10 @@ static ctl_table kern_table[] = {
 #endif
 	{KERN_PIDMAX, "pid_max", &pid_max, sizeof (int),
 	 0600, NULL, &proc_dointvec},
+#ifdef CONFIG_HUGETLB_PAGE
+	 {KERN_HUGETLB_PAGE_NUM, "numhugepages", &htlbpage_max, sizeof(int), 0644, NULL, 
+	  &proc_dointvec},
+#endif
 	{0}
 };
 
@@ -897,6 +906,10 @@ static int do_proc_dointvec(ctl_table *table, int write, struct file *filp,
 				val = -val;
 			buffer += len;
 			left -= len;
+#ifdef CONFIG_HUGETLB_PAGE
+			if (i == &htlbpage_max)
+				val = set_hugetlb_mem_size(val);
+#endif
 			switch(op) {
 			case OP_SET:	*i = val; break;
 			case OP_AND:	*i &= val; break;

lib/Makefile

@@ -12,7 +12,7 @@ export-objs := cmdline.o dec_and_lock.o rwsem-spinlock.o rwsem.o \
 	       crc32.o rbtree.o radix-tree.o
 
 obj-y := errno.o ctype.o string.o vsprintf.o brlock.o cmdline.o \
-	 bust_spinlocks.o rbtree.o radix-tree.o
+	 bust_spinlocks.o rbtree.o radix-tree.o dump_stack.o
 
 obj-$(CONFIG_RWSEM_GENERIC_SPINLOCK) += rwsem-spinlock.o
 obj-$(CONFIG_RWSEM_XCHGADD_ALGORITHM) += rwsem.o

lib/dump_stack.c

+/*
+ * Provide a default dump_stack() function for architectures
+ * which don't implement their own.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+
+void dump_stack(void)
+{
+	printk(KERN_NOTICE
+		"This architecture does not implement dump_stack()\n");
+}

mm/memory.c

@@ -208,6 +208,9 @@ int copy_page_range(struct mm_struct *dst, struct mm_struct *src,
 	unsigned long end = vma->vm_end;
 	unsigned long cow = (vma->vm_flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE;
 
+	if (is_vm_hugetlb_page(vma))
+		return copy_hugetlb_page_range(dst, src, vma);
+
 	src_pgd = pgd_offset(src, address)-1;
 	dst_pgd = pgd_offset(dst, address)-1;
 
@@ -389,8 +392,8 @@ void unmap_page_range(mmu_gather_t *tlb, struct vm_area_struct *vma, unsigned lo
 {
 	pgd_t * dir;
 
-	if (address >= end)
-		BUG();
+	BUG_ON(address >= end);
+
 	dir = pgd_offset(vma->vm_mm, address);
 	tlb_start_vma(tlb, vma);
 	do {
@@ -401,30 +404,56 @@ void unmap_page_range(mmu_gather_t *tlb, struct vm_area_struct *vma, unsigned lo
 	tlb_end_vma(tlb, vma);
 }
 
-/*
- * remove user pages in a given range.
+/* Dispose of an entire mmu_gather_t per rescheduling point */
+#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT)
+#define ZAP_BLOCK_SIZE	(FREE_PTE_NR * PAGE_SIZE)
+#endif
+
+/* For UP, 256 pages at a time gives nice low latency */
+#if !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT)
+#define ZAP_BLOCK_SIZE	(256 * PAGE_SIZE)
+#endif
+
+/* No preempt: go for the best straight-line efficiency */
+#if !defined(CONFIG_PREEMPT)
+#define ZAP_BLOCK_SIZE	(~(0UL))
+#endif
+
+/**
+ * zap_page_range - remove user pages in a given range
+ * @vma: vm_area_struct holding the applicable pages
+ * @address: starting address of pages to zap
+ * @size: number of bytes to zap
  */
 void zap_page_range(struct vm_area_struct *vma, unsigned long address, unsigned long size)
 {
 	struct mm_struct *mm = vma->vm_mm;
 	mmu_gather_t *tlb;
-	unsigned long start = address, end = address + size;
+	unsigned long end, block;
+
+	spin_lock(&mm->page_table_lock);
 
   	/*
-	 * This is a long-lived spinlock. That's fine.
-	 * There's no contention, because the page table
-	 * lock only protects against kswapd anyway, and
-	 * even if kswapd happened to be looking at this
-	 * process we _want_ it to get stuck.
+ 	 * This was once a long-held spinlock.  Now we break the
+ 	 * work up into ZAP_BLOCK_SIZE units and relinquish the
+ 	 * lock after each interation.  This drastically lowers
+ 	 * lock contention and allows for a preemption point.
   	 */
-	if (address >= end)
-		BUG();
-	spin_lock(&mm->page_table_lock);
-	flush_cache_range(vma, address, end);
+	while (size) {
+		block = (size > ZAP_BLOCK_SIZE) ? ZAP_BLOCK_SIZE : size;
+ 		end = address + block;
  
+ 		flush_cache_range(vma, address, end);
  		tlb = tlb_gather_mmu(mm, 0);
  		unmap_page_range(tlb, vma, address, end);
-	tlb_finish_mmu(tlb, start, end);
+ 		tlb_finish_mmu(tlb, address, end);
+ 
+ 		cond_resched_lock(&mm->page_table_lock);
+ 
+ 		address += block;
+ 		size -= block;
+ 	}
+
 	spin_unlock(&mm->page_table_lock);
 }
 
@@ -504,6 +533,11 @@ int get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
 				|| !(flags & vma->vm_flags))
 			return i ? : -EFAULT;
 
+		if (is_vm_hugetlb_page(vma)) {
+			i = follow_hugetlb_page(mm, vma, pages, vmas,
+						&start, &len, i);
+			continue;
+		}
 		spin_lock(&mm->page_table_lock);
 		do {
 			struct page *map;

mm/mempool.c

@@ -196,10 +196,11 @@ void * mempool_alloc(mempool_t *pool, int gfp_mask)
 		return element;
 
 	/*
-	 * If the pool is less than 50% full then try harder
-	 * to allocate an element:
+	 * If the pool is less than 50% full and we can perform effective
+	 * page reclaim then try harder to allocate an element.
 	 */
-	if ((gfp_mask != gfp_nowait) && (pool->curr_nr <= pool->min_nr/2)) {
+	if ((gfp_mask & __GFP_FS) && (gfp_mask != gfp_nowait) &&
+				(pool->curr_nr <= pool->min_nr/2)) {
 		element = pool->alloc(gfp_mask, pool->pool_data);
 		if (likely(element != NULL))
 			return element;

mm/mmap.c

@@ -1031,10 +1031,14 @@ static struct vm_area_struct *touched_by_munmap(struct mm_struct *mm,
 	touched = NULL;
 	do {
 		struct vm_area_struct *next = mpnt->vm_next;
+		if (!(is_vm_hugetlb_page(mpnt))) {
 			mpnt->vm_next = touched;
 			touched = mpnt;
-		mm->map_count--;
 			rb_erase(&mpnt->vm_rb, &mm->mm_rb);
+			mm->map_count--;
+		}
+		else
+			free_hugepages(mpnt);
 		mpnt = next;
 	} while (mpnt && mpnt->vm_start < end);
 	*npp = mpnt;
@@ -1273,7 +1277,10 @@ void exit_mmap(struct mm_struct * mm)
 			vm_unacct_memory((end - start) >> PAGE_SHIFT);
 
 		mm->map_count--;
+		if (!(is_vm_hugetlb_page(mpnt)))
 			unmap_page_range(tlb, mpnt, start, end);
+		else
+			mpnt->vm_ops->close(mpnt);
 		mpnt = mpnt->vm_next;
 	}
 

mm/mprotect.c

@@ -321,6 +321,11 @@ asmlinkage long sys_mprotect(unsigned long start, size_t len, unsigned long prot
 
 		/* Here we know that  vma->vm_start <= nstart < vma->vm_end. */
 
+		if (is_vm_hugetlb_page(vma)) {
+			error = -EACCES;
+			goto out;
+		}
+
 		newflags = prot | (vma->vm_flags & ~(PROT_READ | PROT_WRITE | PROT_EXEC));
 		if ((newflags & ~(newflags >> 4)) & 0xf) {
 			error = -EACCES;

mm/mremap.c

@@ -311,6 +311,10 @@ unsigned long do_mremap(unsigned long addr,
 	vma = find_vma(current->mm, addr);
 	if (!vma || vma->vm_start > addr)
 		goto out;
+	if (is_vm_hugetlb_page(vma)) {
+		ret = -EINVAL;
+		goto out;
+	}
 	/* We can't remap across vm area boundaries */
 	if (old_len > vma->vm_end - addr)
 		goto out;

mm/page_alloc.c

@@ -23,6 +23,7 @@
 #include <linux/module.h>
 #include <linux/suspend.h>
 #include <linux/pagevec.h>
+#include <linux/blkdev.h>
 
 unsigned long totalram_pages;
 unsigned long totalhigh_pages;
@@ -561,6 +562,7 @@ void get_page_state(struct page_state *ret)
 		ret->nr_pagecache += ps->nr_pagecache;
 		ret->nr_page_table_pages += ps->nr_page_table_pages;
 		ret->nr_reverse_maps += ps->nr_reverse_maps;
+		ret->nr_mapped += ps->nr_mapped;
 	}
 }
 
@@ -589,7 +591,7 @@ void si_meminfo(struct sysinfo *val)
 	val->totalram = totalram_pages;
 	val->sharedram = 0;
 	val->freeram = nr_free_pages();
-	val->bufferram = get_page_cache_size();
+	val->bufferram = nr_blockdev_pages();
 #ifdef CONFIG_HIGHMEM
 	val->totalhigh = totalhigh_pages;
 	val->freehigh = nr_free_highpages();

mm/readahead.c

@@ -48,9 +48,9 @@ read_pages(struct file *file, struct address_space *mapping,
 		struct page *page = list_entry(pages->prev, struct page, list);
 		list_del(&page->list);
 		if (!add_to_page_cache(page, mapping, page->index)) {
+			mapping->a_ops->readpage(file, page);
 			if (!pagevec_add(&lru_pvec, page))
 				__pagevec_lru_add(&lru_pvec);
-			mapping->a_ops->readpage(file, page);
 		} else {
 			page_cache_release(page);
 		}

mm/rmap.c

@@ -214,6 +214,7 @@ void page_add_rmap(struct page * page, pte_t * ptep)
 	if (page->pte.direct == 0) {
 		page->pte.direct = pte_paddr;
 		SetPageDirect(page);
+		inc_page_state(nr_mapped);
 		goto out;
 	}
 
@@ -336,6 +337,8 @@ void page_remove_rmap(struct page * page, pte_t * ptep)
 
 out:
 	pte_chain_unlock(page);
+	if (!page_mapped(page))
+		dec_page_state(nr_mapped);
 	return;
 }
 
@@ -447,6 +450,7 @@ int try_to_unmap(struct page * page)
 		ret = try_to_unmap_one(page, page->pte.direct);
 		if (ret == SWAP_SUCCESS) {
 			page->pte.direct = 0;
+			dec_page_state(nr_reverse_maps);
 			ClearPageDirect(page);
 		}
 		goto out;
@@ -500,6 +504,8 @@ int try_to_unmap(struct page * page)
 		}
 	}
 out:
+	if (!page_mapped(page))
+		dec_page_state(nr_mapped);
 	return ret;
 }
 

mm/slab.c

@@ -487,7 +487,7 @@ void __init kmem_cache_sizes_init(void)
 		/* Inc off-slab bufctl limit until the ceiling is hit. */
 		if (!(OFF_SLAB(sizes->cs_cachep))) {
 			offslab_limit = sizes->cs_size-sizeof(slab_t);
-			offslab_limit /= 2;
+			offslab_limit /= sizeof(kmem_bufctl_t);
 		}
 		sizes->cs_dmacachep = kmem_cache_create(
 		    cache_names[sizes-cache_sizes].name_dma, 

mm/vmalloc.c

@@ -11,6 +11,8 @@
 #include <linux/highmem.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
+#include <linux/interrupt.h>
+
 #include <linux/vmalloc.h>
 
 #include <asm/uaccess.h>
@@ -309,6 +311,7 @@ void __vunmap(void *addr, int deallocate_pages)
  */
 void vfree(void *addr)
 {
+	BUG_ON(in_interrupt());
 	__vunmap(addr, 1);
 }
 
@@ -324,6 +327,7 @@ void vfree(void *addr)
  */
 void vunmap(void *addr)
 {
+	BUG_ON(in_interrupt());
 	__vunmap(addr, 0);
 }
 

mm/vmscan.c

@@ -536,6 +536,20 @@ shrink_caches(struct zone *classzone, int priority,
 
 /*
  * This is the main entry point to page reclaim.
+ *
+ * If a full scan of the inactive list fails to free enough memory then we
+ * are "out of memory" and something needs to be killed.
+ *
+ * If the caller is !__GFP_FS then the probability of a failure is reasonably
+ * high - the zone may be full of dirty or under-writeback pages, which this
+ * caller can't do much about.  So for !__GFP_FS callers, we just perform a
+ * small LRU walk and if that didn't work out, fail the allocation back to the
+ * caller.  GFP_NOFS allocators need to know how to deal with it.  Kicking
+ * bdflush, waiting and retrying will work.
+ *
+ * This is a fairly lame algorithm - it can result in excessive CPU burning and
+ * excessive rotation of the inactive list, which is _supposed_ to be an LRU,
+ * yes?
  */
 int
 try_to_free_pages(struct zone *classzone,
@@ -546,12 +560,15 @@ try_to_free_pages(struct zone *classzone,
 
 	KERNEL_STAT_INC(pageoutrun);
 
-	do {
+	for (priority = DEF_PRIORITY; priority; priority--) {
 		nr_pages = shrink_caches(classzone, priority,
 					gfp_mask, nr_pages);
 		if (nr_pages <= 0)
 			return 1;
-	} while (--priority);
+		if (!(gfp_mask & __GFP_FS))
+			break;
+	}
+	if (gfp_mask & __GFP_FS)
 		out_of_memory();
 	return 0;
 }