v2.4.10.0.3 -> v2.4.10.0.4

  - more buffers-in-pagecache coherency

fs/affs/inode.c

@@ -300,7 +300,7 @@ affs_new_inode(struct inode *dir)
 	u32			 block;
 	struct buffer_head	*bh;
 
-	if (!(inode = get_empty_inode()))
+	if (!(inode = new_inode(sb)))
 		goto err_inode;
 
 	if (!(block = affs_alloc_block(dir, dir->i_ino)))
@@ -312,8 +312,6 @@ affs_new_inode(struct inode *dir)
 	mark_buffer_dirty_inode(bh, inode);
 	affs_brelse(bh);
 
-	inode->i_sb      = sb;
-	inode->i_dev     = sb->s_dev;
 	inode->i_uid     = current->fsuid;
 	inode->i_gid     = current->fsgid;
 	inode->i_ino     = block;

fs/block_dev.c

@@ -47,7 +47,7 @@ static inline unsigned int block_size(kdev_t dev)
 	return retval;
 }
 
-static unsigned int max_block(kdev_t dev)
+static unsigned long max_block(kdev_t dev)
 {
 	unsigned int retval = ~0U;
 	int major = MAJOR(dev);
@@ -86,302 +86,74 @@ static void kill_bdev(struct block_device *bdev)
 	truncate_inode_pages(bdev->bd_inode->i_mapping, 0);
 }	
 
-static inline void kill_buffers(kdev_t dev)
+int set_blocksize(kdev_t dev, int size)
 {
-	struct block_device *bdev = bdget(dev);
-	if (bdev) {
-		kill_bdev(bdev);
-		bdput(bdev);
-	}
-}
-
-void set_blocksize(kdev_t dev, int size)
-{
-	extern int *blksize_size[];
-
-	if (!blksize_size[MAJOR(dev)])
-		return;
+	int oldsize;
+	struct block_device *bdev;
 
 	/* Size must be a power of two, and between 512 and PAGE_SIZE */
 	if (size > PAGE_SIZE || size < 512 || (size & (size-1)))
-		panic("Invalid blocksize passed to set_blocksize");
+		return -EINVAL;
 
-	if (blksize_size[MAJOR(dev)][MINOR(dev)] == 0 && size == BLOCK_SIZE) {
+	/* Size cannot be smaller than the size supported by the device */
+	if (size < get_hardsect_size(dev))
+		return -EINVAL;
+
+	/* No blocksize array? Implies hardcoded BLOCK_SIZE */
+	if (!blksize_size[MAJOR(dev)]) {
+		if (size == BLOCK_SIZE)
+			return 0;
+		return -EINVAL;
+	}
+
+	oldsize = blksize_size[MAJOR(dev)][MINOR(dev)];
+	if (oldsize == size)
+		return 0;
+
+	if (!oldsize && size == BLOCK_SIZE) {
 		blksize_size[MAJOR(dev)][MINOR(dev)] = size;
-		return;
+		return 0;
 	}
-	if (blksize_size[MAJOR(dev)][MINOR(dev)] == size)
-		return;
 
+	/* Ok, we're actually changing the blocksize.. */
+	bdev = bdget(dev);
 	sync_buffers(dev, 2);
 	blksize_size[MAJOR(dev)][MINOR(dev)] = size;
-	kill_buffers(dev);
+	bdev->bd_inode->i_blkbits = blksize_bits(size);
+	kill_bdev(bdev);
+	bdput(bdev);
+	return 0;
 }
 
-static inline int blkdev_get_block(struct inode * inode, long iblock, struct buffer_head * bh_result)
+static int blkdev_get_block(struct inode * inode, long iblock, struct buffer_head * bh, int create)
 {
-	int err;
-
-	err = -EIO;
 	if (iblock >= max_block(inode->i_rdev))
-		goto out;
-
-	bh_result->b_blocknr = iblock;
-	bh_result->b_state |= 1UL << BH_Mapped;
-	err = 0;
+		return -EIO;
 
- out:
-	return err;
+	bh->b_dev = inode->i_rdev;
+	bh->b_blocknr = iblock;
+	bh->b_state |= 1UL << BH_Mapped;
+	return 0;
 }
 
 static int blkdev_writepage(struct page * page)
 {
-	int err, i;
-	unsigned int blocksize;
-	unsigned long block;
-	struct buffer_head *bh, *head;
-	struct inode *inode = page->mapping->host;
-
-	if (!PageLocked(page))
-		BUG();
-	blocksize = block_size(inode->i_rdev);	
-
-	if (!page->buffers)
-		create_empty_buffers(page, inode->i_rdev, blocksize);
-	head = page->buffers;
-
-	block = page->index << (PAGE_CACHE_SHIFT - blksize_bits(blocksize));
-
-	bh = head;
-	i = 0;
-
-	/* Stage 1: make sure we have all the buffers mapped! */
-	do {
-		/*
-		 * If the buffer isn't up-to-date, we can't be sure
-		 * that the buffer has been initialized with the proper
-		 * block number information etc..
-		 *
-		 * Leave it to the low-level FS to make all those
-		 * decisions (block #0 may actually be a valid block)
-		 */
-		if (!buffer_mapped(bh)) {
-			err = blkdev_get_block(inode, block, bh);
-			if (err)
-				goto out;
-		}
-		bh = bh->b_this_page;
-		block++;
-	} while (bh != head);
-
-	/* Stage 2: lock the buffers, mark them clean */
-	do {
-		lock_buffer(bh);
-		set_buffer_async_io(bh);
-		set_bit(BH_Uptodate, &bh->b_state);
-		clear_bit(BH_Dirty, &bh->b_state);
-		bh = bh->b_this_page;
-	} while (bh != head);
-
-	/* Stage 3: submit the IO */
-	do {
-		submit_bh(WRITE, bh);
-		bh = bh->b_this_page;
-	} while (bh != head);
-
-	/* Done - end_buffer_io_async will unlock */
-	SetPageUptodate(page);
-	return 0;
-
-out:
-	ClearPageUptodate(page);
-	UnlockPage(page);
-	return err;
+	return block_write_full_page(page, blkdev_get_block);
 }
 
 static int blkdev_readpage(struct file * file, struct page * page)
 {
-	struct inode *inode = page->mapping->host;
-	kdev_t dev = inode->i_rdev;
-	unsigned long iblock, lblock;
-	struct buffer_head *bh, *head, *arr[MAX_BUF_PER_PAGE];
-	unsigned int blocks, blocksize, blocksize_bits;
-	int nr, i;
-
-	if (!PageLocked(page))
-		PAGE_BUG(page);
-	blocksize = block_size(dev);
-	blocksize_bits = blksize_bits(blocksize);
-	if (!page->buffers)
-		create_empty_buffers(page, dev, blocksize);
-	head = page->buffers;
-
-	blocks = PAGE_CACHE_SIZE >> blocksize_bits;
-	iblock = page->index << (PAGE_CACHE_SHIFT - blocksize_bits);
-	lblock = max_block(dev);
-	bh = head;
-	nr = 0;
-	i = 0;
-
-	do {
-		if (buffer_uptodate(bh))
-			continue;
-
-		if (!buffer_mapped(bh)) {
-			if (iblock <= lblock) {
-				if (blkdev_get_block(inode, iblock, bh))
-					continue;
-			}
-			if (!buffer_mapped(bh)) {
-				memset(kmap(page) + i * blocksize, 0, blocksize);
-				flush_dcache_page(page);
-				kunmap(page);
-				set_bit(BH_Uptodate, &bh->b_state);
-				continue;
-			}
-			/* get_block() might have updated the buffer synchronously */
-			if (buffer_uptodate(bh))
-				continue;
-		}
-
-		arr[nr] = bh;
-		nr++;
-	} while (i++, iblock++, (bh = bh->b_this_page) != head);
-
-	if (!nr) {
-		/*
-		 * all buffers are uptodate - we can set the page
-		 * uptodate as well.
-		 */
-		SetPageUptodate(page);
-		UnlockPage(page);
-		return 0;
-	}
-
-	/* Stage two: lock the buffers */
-	for (i = 0; i < nr; i++) {
-		struct buffer_head * bh = arr[i];
-		lock_buffer(bh);
-		set_buffer_async_io(bh);
-	}
-
-	/* Stage 3: start the IO */
-	for (i = 0; i < nr; i++)
-		submit_bh(READ, arr[i]);
-
-	return 0;
-}
-
-static int __blkdev_prepare_write(struct inode *inode, struct page *page,
-				  unsigned from, unsigned to)
-{
-	kdev_t dev = inode->i_rdev;
-	unsigned block_start, block_end;
-	unsigned long block;
-	int err = 0;
-	struct buffer_head *bh, *head, *wait[2], **wait_bh=wait;
-	unsigned int blocksize, blocksize_bits;
-
-	blocksize = block_size(dev);
-	blocksize_bits = blksize_bits(blocksize);
-	if (!page->buffers)
-		create_empty_buffers(page, dev, blocksize);
-	head = page->buffers;
-
-	block = page->index << (PAGE_CACHE_SHIFT - blocksize_bits);
-
-	for(bh = head, block_start = 0; bh != head || !block_start;
-	    block++, block_start=block_end, bh = bh->b_this_page) {
-		if (!bh)
-			BUG();
-		block_end = block_start + blocksize;
-		if (block_end <= from)
-			continue;
-		if (block_start >= to)
-			break;
-		if (!buffer_mapped(bh)) {
-			err = blkdev_get_block(inode, block, bh);
-			if (err)
-				goto out;
-		}
-		if (Page_Uptodate(page)) {
-			set_bit(BH_Uptodate, &bh->b_state);
-			continue; 
-		}
-		if (!buffer_uptodate(bh) &&
-		     (block_start < from || block_end > to)) {
-			ll_rw_block(READ, 1, &bh);
-			*wait_bh++=bh;
-		}
-	}
-	/*
-	 * If we issued read requests - let them complete.
-	 */
-	while(wait_bh > wait) {
-		wait_on_buffer(*--wait_bh);
-		err = -EIO;
-		if (!buffer_uptodate(*wait_bh))
-			goto out;
-	}
-	return 0;
-out:
-	return err;
+	return block_read_full_page(page, blkdev_get_block);
 }
 
 static int blkdev_prepare_write(struct file *file, struct page *page, unsigned from, unsigned to)
 {
-	struct inode *inode = page->mapping->host;
-	int err = __blkdev_prepare_write(inode, page, from, to);
-	if (err) {
-		ClearPageUptodate(page);
-	}
-	return err;
-}
-
-static int __blkdev_commit_write(struct inode *inode, struct page *page,
-				 unsigned from, unsigned to)
-{
-	unsigned block_start, block_end;
-	int partial = 0, need_balance_dirty = 0;
-	struct buffer_head *bh, *head;
-	unsigned int blocksize;
-
-	blocksize = block_size(inode->i_rdev);
-	for(bh = head = page->buffers, block_start = 0;
-	    bh != head || !block_start;
-	    block_start=block_end, bh = bh->b_this_page) {
-		block_end = block_start + blocksize;
-		if (block_end <= from || block_start >= to) {
-			if (!buffer_uptodate(bh))
-				partial = 1;
-		} else {
-			set_bit(BH_Uptodate, &bh->b_state);
-			if (!atomic_set_buffer_dirty(bh)) {
-				__mark_dirty(bh);
-				need_balance_dirty = 1;
-			}
-		}
-	}
-
-	if (need_balance_dirty)
-		balance_dirty();
-	/*
-	 * is this a partial write that happened to make all buffers
-	 * uptodate then we can optimize away a bogus readpage() for
-	 * the next read(). Here we 'discover' wether the page went
-	 * uptodate as a result of this (potentially partial) write.
-	 */
-	if (!partial)
-		SetPageUptodate(page);
-	return 0;
+	return block_prepare_write(page, from, to, blkdev_get_block);
 }
 
-static int blkdev_commit_write(struct file *file, struct page *page,
-			       unsigned from, unsigned to)
+static int blkdev_commit_write(struct file *file, struct page *page, unsigned from, unsigned to)
 {
-	struct inode *inode = page->mapping->host;
-	__blkdev_commit_write(inode,page,from,to);
-	return 0;
+	return block_commit_write(page, from, to);
 }
 
 /*
@@ -565,12 +337,15 @@ struct block_device *bdget(dev_t dev)
 	if (new_bdev) {
 		struct inode *inode = new_inode(bd_mnt->mnt_sb);
 		if (inode) {
+			kdev_t kdev = to_kdev_t(dev);
 			atomic_set(&new_bdev->bd_count,1);
 			new_bdev->bd_dev = dev;
 			new_bdev->bd_op = NULL;
 			new_bdev->bd_inode = inode;
-			inode->i_rdev = to_kdev_t(dev);
+			inode->i_rdev = kdev;
+			inode->i_dev = kdev;
 			inode->i_bdev = new_bdev;
+			inode->i_blkbits = blksize_bits(block_size(kdev));
 			inode->i_data.a_ops = &def_blk_aops;
 			inode->i_data.gfp_mask = GFP_USER;
 			spin_lock(&bdev_lock);

fs/buffer.c

@@ -1482,10 +1482,10 @@ static int __block_write_full_page(struct inode *inode, struct page *page, get_b
 		BUG();
 
 	if (!page->buffers)
-		create_empty_buffers(page, inode->i_dev, inode->i_sb->s_blocksize);
+		create_empty_buffers(page, inode->i_dev, 1 << inode->i_blkbits);
 	head = page->buffers;
 
-	block = page->index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);
+	block = page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
 
 	bh = head;
 	i = 0;
@@ -1547,12 +1547,12 @@ static int __block_prepare_write(struct inode *inode, struct page *page,
 	struct buffer_head *bh, *head, *wait[2], **wait_bh=wait;
 	char *kaddr = kmap(page);
 
-	blocksize = inode->i_sb->s_blocksize;
+	blocksize = 1 << inode->i_blkbits;
 	if (!page->buffers)
 		create_empty_buffers(page, inode->i_dev, blocksize);
 	head = page->buffers;
 
-	bbits = inode->i_sb->s_blocksize_bits;
+	bbits = inode->i_blkbits;
 	block = page->index << (PAGE_CACHE_SHIFT - bbits);
 
 	for(bh = head, block_start = 0; bh != head || !block_start;
@@ -1615,7 +1615,7 @@ static int __block_commit_write(struct inode *inode, struct page *page,
 	unsigned blocksize;
 	struct buffer_head *bh, *head;
 
-	blocksize = inode->i_sb->s_blocksize;
+	blocksize = 1 << inode->i_blkbits;
 
 	for(bh = head = page->buffers, block_start = 0;
 	    bh != head || !block_start;
@@ -1664,14 +1664,14 @@ int block_read_full_page(struct page *page, get_block_t *get_block)
 
 	if (!PageLocked(page))
 		PAGE_BUG(page);
-	blocksize = inode->i_sb->s_blocksize;
+	blocksize = 1 << inode->i_blkbits;
 	if (!page->buffers)
 		create_empty_buffers(page, inode->i_dev, blocksize);
 	head = page->buffers;
 
-	blocks = PAGE_CACHE_SIZE >> inode->i_sb->s_blocksize_bits;
-	iblock = page->index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);
-	lblock = (inode->i_size+blocksize-1) >> inode->i_sb->s_blocksize_bits;
+	blocks = PAGE_CACHE_SIZE >> inode->i_blkbits;
+	iblock = page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
+	lblock = (inode->i_size+blocksize-1) >> inode->i_blkbits;
 	bh = head;
 	nr = 0;
 	i = 0;
@@ -1738,7 +1738,7 @@ int cont_prepare_write(struct page *page, unsigned offset, unsigned to, get_bloc
 	unsigned long pgpos;
 	long status;
 	unsigned zerofrom;
-	unsigned blocksize = inode->i_sb->s_blocksize;
+	unsigned blocksize = 1 << inode->i_blkbits;
 	char *kaddr;
 
 	while(page->index > (pgpos = *bytes>>PAGE_CACHE_SHIFT)) {
@@ -1823,6 +1823,14 @@ int block_prepare_write(struct page *page, unsigned from, unsigned to,
 	return err;
 }
 
+int block_commit_write(struct page *page, unsigned from, unsigned to)
+{
+	struct inode *inode = page->mapping->host;
+	__block_commit_write(inode,page,from,to);
+	kunmap(page);
+	return 0;
+}
+
 int generic_commit_write(struct file *file, struct page *page,
 		unsigned from, unsigned to)
 {
@@ -1847,7 +1855,7 @@ int block_truncate_page(struct address_space *mapping, loff_t from, get_block_t
 	struct buffer_head *bh;
 	int err;
 
-	blocksize = inode->i_sb->s_blocksize;
+	blocksize = 1 << inode->i_blkbits;
 	length = offset & (blocksize - 1);
 
 	/* Block boundary? Nothing to do */
@@ -1855,7 +1863,7 @@ int block_truncate_page(struct address_space *mapping, loff_t from, get_block_t
 		return 0;
 
 	length = blocksize - length;
-	iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);
+	iblock = index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
 	
 	page = grab_cache_page(mapping, index);
 	err = -ENOMEM;
@@ -2391,6 +2399,10 @@ int try_to_free_buffers(struct page * page, unsigned int gfp_mask)
 
 	spin_lock(&unused_list_lock);
 	tmp = bh;
+
+	/* if this buffer was hashed, this page counts as buffermem */
+	if (bh->b_pprev)
+		atomic_dec(&buffermem_pages);
 	do {
 		struct buffer_head * p = tmp;
 		tmp = tmp->b_this_page;

fs/ext2/super.c

@@ -408,7 +408,6 @@ struct super_block * ext2_read_super (struct super_block * sb, void * data,
 	unsigned long offset = 0;
 	kdev_t dev = sb->s_dev;
 	int blocksize = BLOCK_SIZE;
-	int hblock;
 	int db_count;
 	int i, j;
 
@@ -429,7 +428,10 @@ struct super_block * ext2_read_super (struct super_block * sb, void * data,
 		return NULL;
 	}
 
-	set_blocksize (dev, blocksize);
+	if (set_blocksize(dev, blocksize) < 0) {
+		printk ("EXT2-fs: unable to set blocksize %d\n", blocksize);
+		return NULL;
+	}
 
 	/*
 	 * If the superblock doesn't start on a sector boundary,
@@ -488,24 +490,19 @@ struct super_block * ext2_read_super (struct super_block * sb, void * data,
 
 	sb->s_maxbytes = ext2_max_size(sb->s_blocksize_bits);
 
-	if (sb->s_blocksize != BLOCK_SIZE &&
-	    (sb->s_blocksize == 1024 || sb->s_blocksize == 2048 ||
-	     sb->s_blocksize == 4096)) {
-		/*
-		 * Make sure the blocksize for the filesystem is larger
-		 * than the hardware sectorsize for the machine.
-		 */
-		hblock = get_hardsect_size(dev);
-		if (sb->s_blocksize < hblock) {
+	/* If the blocksize doesn't match, re-read the thing.. */
+	if (sb->s_blocksize != blocksize) {
+		blocksize = sb->s_blocksize;
+		brelse(bh);
+
+		if (set_blocksize(dev, blocksize) < 0) {
 			printk(KERN_ERR "EXT2-fs: blocksize too small for device.\n");
-			goto failed_mount;
+			return NULL;
 		}
 
-		brelse (bh);
-		set_blocksize (dev, sb->s_blocksize);
-		logic_sb_block = (sb_block*BLOCK_SIZE) / sb->s_blocksize;
-		offset = (sb_block*BLOCK_SIZE) % sb->s_blocksize;
-		bh = bread (dev, logic_sb_block, sb->s_blocksize);
+		logic_sb_block = (sb_block*BLOCK_SIZE) / blocksize;
+		offset = (sb_block*BLOCK_SIZE) % blocksize;
+		bh = bread (dev, logic_sb_block, blocksize);
 		if(!bh) {
 			printk("EXT2-fs: Couldn't read superblock on "
 			       "2nd try.\n");
@@ -518,6 +515,7 @@ struct super_block * ext2_read_super (struct super_block * sb, void * data,
 			goto failed_mount;
 		}
 	}
+
 	if (le32_to_cpu(es->s_rev_level) == EXT2_GOOD_OLD_REV) {
 		sb->u.ext2_sb.s_inode_size = EXT2_GOOD_OLD_INODE_SIZE;
 		sb->u.ext2_sb.s_first_ino = EXT2_GOOD_OLD_FIRST_INO;

fs/inode.c

@@ -816,6 +816,7 @@ struct inode * get_empty_inode(void)
 		list_add(&inode->i_list, &inode_in_use);
 		inode->i_sb = NULL;
 		inode->i_dev = 0;
+		inode->i_blkbits = 0;
 		inode->i_ino = ++last_ino;
 		inode->i_flags = 0;
 		atomic_set(&inode->i_count, 1);
@@ -849,6 +850,7 @@ static struct inode * get_new_inode(struct super_block *sb, unsigned long ino, s
 			list_add(&inode->i_hash, head);
 			inode->i_sb = sb;
 			inode->i_dev = sb->s_dev;
+			inode->i_blkbits = sb->s_blocksize_bits;
 			inode->i_ino = ino;
 			inode->i_flags = 0;
 			atomic_set(&inode->i_count, 1);

fs/pipe.c

@@ -476,7 +476,7 @@ static struct dentry_operations pipefs_dentry_operations = {
 
 static struct inode * get_pipe_inode(void)
 {
-	struct inode *inode = get_empty_inode();
+	struct inode *inode = new_inode(pipe_mnt->mnt_sb);
 
 	if (!inode)
 		goto fail_inode;
@@ -485,7 +485,6 @@ static struct inode * get_pipe_inode(void)
 		goto fail_iput;
 	PIPE_READERS(*inode) = PIPE_WRITERS(*inode) = 1;
 	inode->i_fop = &rdwr_pipe_fops;
-	inode->i_sb = pipe_mnt->mnt_sb;
 
 	/*
 	 * Mark the inode dirty from the very beginning,

fs/reiserfs/inode.c

@@ -1428,7 +1428,6 @@ struct inode * reiserfs_new_inode (struct reiserfs_transaction_handle *th,
     }
 
     sb = dir->i_sb;
-    inode->i_sb = sb;
     inode->i_flags = 0;//inode->i_sb->s_flags;
 
     /* item head of new item */

fs/reiserfs/namei.c

@@ -533,7 +533,7 @@ int reiserfs_create (struct inode * dir, struct dentry *dentry, int mode)
     struct reiserfs_transaction_handle th ;
 
 
-    inode = get_empty_inode() ;
+    inode = new_inode(dir->i_sb) ;
     if (!inode) {
 	return -ENOMEM ;
     }
@@ -586,7 +586,7 @@ int reiserfs_mknod (struct inode * dir, struct dentry *dentry, int mode, int rde
     struct reiserfs_transaction_handle th ;
     int jbegin_count = JOURNAL_PER_BALANCE_CNT * 3; 
 
-    inode = get_empty_inode() ;
+    inode = new_inode(dir->i_sb) ;
     if (!inode) {
 	return -ENOMEM ;
     }
@@ -638,7 +638,7 @@ int reiserfs_mkdir (struct inode * dir, struct dentry *dentry, int mode)
     struct reiserfs_transaction_handle th ;
     int jbegin_count = JOURNAL_PER_BALANCE_CNT * 3; 
 
-    inode = get_empty_inode() ;
+    inode = new_inode(dir->i_sb) ;
     if (!inode) {
 	return -ENOMEM ;
     }
@@ -859,7 +859,7 @@ int reiserfs_symlink (struct inode * dir, struct dentry * dentry, const char * s
     int jbegin_count = JOURNAL_PER_BALANCE_CNT * 3; 
 
 
-    inode = get_empty_inode() ;
+    inode = new_inode(dir->i_sb) ;
     if (!inode) {
 	return -ENOMEM ;
     }

include/linux/fs.h

@@ -278,7 +278,7 @@ void init_buffer(struct buffer_head *, bh_end_io_t *, void *);
 
 extern void set_bh_page(struct buffer_head *bh, struct page *page, unsigned long offset);
 
-#define touch_buffer(bh)	SetPageReferenced(bh->b_page)
+#define touch_buffer(bh)	mark_page_accessed(bh->b_page)
 
 
 #include <linux/pipe_fs_i.h>
@@ -433,6 +433,7 @@ struct inode {
 	time_t			i_atime;
 	time_t			i_mtime;
 	time_t			i_ctime;
+	unsigned int		i_blkbits;
 	unsigned long		i_blksize;
 	unsigned long		i_blocks;
 	unsigned long		i_version;
@@ -1299,12 +1300,14 @@ static inline struct inode *iget(struct super_block *sb, unsigned long ino)
 
 extern void clear_inode(struct inode *);
 extern struct inode * get_empty_inode(void);
+
 static inline struct inode * new_inode(struct super_block *sb)
 {
 	struct inode *inode = get_empty_inode();
 	if (inode) {
 		inode->i_sb = sb;
 		inode->i_dev = sb->s_dev;
+		inode->i_blkbits = sb->s_blocksize_bits;
 	}
 	return inode;
 }
@@ -1331,7 +1334,7 @@ static inline void bforget(struct buffer_head *buf)
 	if (buf)
 		__bforget(buf);
 }
-extern void set_blocksize(kdev_t, int);
+extern int set_blocksize(kdev_t, int);
 extern struct buffer_head * bread(kdev_t, int, int);
 extern void wakeup_bdflush(void);
 
@@ -1349,6 +1352,7 @@ extern int block_read_full_page(struct page*, get_block_t*);
 extern int block_prepare_write(struct page*, unsigned, unsigned, get_block_t*);
 extern int cont_prepare_write(struct page*, unsigned, unsigned, get_block_t*,
 				unsigned long *);
+extern int block_commit_write(struct page *page, unsigned from, unsigned to);
 extern int block_sync_page(struct page *);
 
 int generic_block_bmap(struct address_space *, long, get_block_t *);

mm/filemap.c

@@ -1244,6 +1244,13 @@ void do_generic_file_read(struct file * filp, loff_t *ppos, read_descriptor_t *
 		if (mapping->i_mmap_shared != NULL)
 			flush_dcache_page(page);
 
+		/*
+		 * Mark the page accessed if we read the
+		 * beginning or we just did an lseek.
+		 */
+		if (!offset || !filp->f_reada)
+			mark_page_accessed(page);
+
 		/*
 		 * Ok, we have the page, and it's up-to-date, so
 		 * now we can copy it to user space...
@@ -1259,7 +1266,6 @@ void do_generic_file_read(struct file * filp, loff_t *ppos, read_descriptor_t *
 		index += offset >> PAGE_CACHE_SHIFT;
 		offset &= ~PAGE_CACHE_MASK;
 
-		mark_page_accessed(page);
 		page_cache_release(page);
 		if (ret == nr && desc->count)
 			continue;

mm/vmscan.c

@@ -406,12 +406,6 @@ static int shrink_cache(int nr_pages, int max_scan, zone_t * classzone, unsigned
 
 			if (try_to_free_buffers(page, gfp_mask)) {
 				if (!page->mapping) {
-					/*
-					 * Account we successfully freed a page
-					 * of buffer cache.
-					 */
-					atomic_dec(&buffermem_pages);
-
 					/*
 					 * We must not allow an anon page
 					 * with no buffers to be visible on
@@ -536,16 +530,20 @@ static void refill_inactive(int nr_pages)
 static int FASTCALL(shrink_caches(int priority, zone_t * classzone, unsigned int gfp_mask, int nr_pages));
 static int shrink_caches(int priority, zone_t * classzone, unsigned int gfp_mask, int nr_pages)
 {
-	int max_scan = nr_inactive_pages / priority;
+	int max_scan;
+	int chunk_size = nr_pages;
+	unsigned long ratio;
 
 	nr_pages -= kmem_cache_reap(gfp_mask);
 	if (nr_pages <= 0)
 		return 0;
 
-	/* Do we want to age the active list? */
-	if (nr_inactive_pages < nr_active_pages*2)
-		refill_inactive(nr_pages);
+	nr_pages = chunk_size;
+	/* try to keep the active list 2/3 of the size of the cache */
+	ratio = (unsigned long) nr_pages * nr_active_pages / ((nr_inactive_pages + 1) * 2);
+	refill_inactive(ratio);
   
+	max_scan = nr_inactive_pages / priority;
 	nr_pages = shrink_cache(nr_pages, max_scan, classzone, gfp_mask);
 	if (nr_pages <= 0)
 		return 0;
@@ -558,11 +556,13 @@ static int shrink_caches(int priority, zone_t * classzone, unsigned int gfp_mask
 
 int try_to_free_pages(zone_t * classzone, unsigned int gfp_mask, unsigned int order)
 {
-	int priority = DEF_PRIORITY;
 	int ret = 0;
 
-	do {
+	for (;;) {
+		int priority = DEF_PRIORITY;
 		int nr_pages = SWAP_CLUSTER_MAX;
+
+		do {
 			nr_pages = shrink_caches(priority, classzone, gfp_mask, nr_pages);
 			if (nr_pages <= 0)
 				return 1;
@@ -570,6 +570,15 @@ int try_to_free_pages(zone_t * classzone, unsigned int gfp_mask, unsigned int or
 			ret |= swap_out(priority, classzone, gfp_mask, SWAP_CLUSTER_MAX << 2);
 		} while (--priority);
 
+		if (likely(ret))
+			break;
+		if (likely(current->pid != 1))
+			break;
+		current->policy |= SCHED_YIELD;
+		__set_current_state(TASK_RUNNING);
+		schedule();
+	}
+
 	return ret;
 }
 

net/socket.c

@@ -440,11 +440,10 @@ struct socket *sock_alloc(void)
 	struct inode * inode;
 	struct socket * sock;
 
-	inode = get_empty_inode();
+	inode = new_inode(sock_mnt->mnt_sb);
 	if (!inode)
 		return NULL;
 
-	inode->i_sb = sock_mnt->mnt_sb;
 	sock = socki_lookup(inode);
 
 	inode->i_mode = S_IFSOCK|S_IRWXUGO;