ext4fs: Add ext4 extent cache for read operations

In an ext4 filesystem, the inode corresponding to a file has a 60-byte
area which contains an extent header structure and up to 4 extent
structures (5 x 12 bytes).

For files that need more than 4 extents to be represented (either files
larger than 4 x 128MB = 512MB or smaller files but very fragmented),
ext4 creates extent index structures. Each extent index points to a 4KB
physical block where one extent header and additional 340 extents could
be stored.

The current u-boot ext4 code is very inefficient when it tries to load a
file which has extent indexes. For each logical file block the code will
read over and over again the same blocks of 4096 bytes from the disk.

Since the extent tree in a file is always the same, we can cache the
extent structures in memory before actually starting to read the file.

This patch creates a simple linked list of structures holding information
about all the extents used to represent a file. The list is sorted by
the logical block number (ee_block) so that we can easily find the
proper extent information for any file block.

Without this patch, a 69MB file which had just one extent index pointing
to a block with another 6 extents was read in approximately 3 minutes.
With this patch applied the same file can be read in almost 20 seconds.

Signed-off-by: Ionut Nicu <ioan.nicu.ext@nsn.com>

fs/ext4/ext4_common.c

@@ -26,6 +26,7 @@
 #include <stddef.h>
 #include <linux/stat.h>
 #include <linux/time.h>
+#include <linux/list.h>
 #include <asm/byteorder.h>
 #include "ext4_common.h"
 
@@ -44,6 +45,14 @@ int ext4fs_indir3_blkno = -1;
 struct ext2_inode *g_parent_inode;
 static int symlinknest;
 
+struct ext4_extent_node {
+	uint32_t block;
+	uint16_t len;
+	uint64_t start;
+	struct list_head lh;
+};
+static LIST_HEAD(ext4_extent_lh);
+
 #if defined(CONFIG_EXT4_WRITE)
 uint32_t ext4fs_div_roundup(uint32_t size, uint32_t n)
 {
@@ -1407,43 +1416,100 @@ void ext4fs_allocate_blocks(struct ext2_inode *file_inode,
 
 #endif
 
-static struct ext4_extent_header *ext4fs_get_extent_block
+static void ext4fs_extent_cache_insert(struct ext4_extent_node *new)
-	(struct ext2_data *data, char *buf,
-		struct ext4_extent_header *ext_block,
-		uint32_t fileblock, int log2_blksz)
 {
+	struct ext4_extent_node *node;
+
+	list_for_each_entry(node, &ext4_extent_lh, lh)
+		if (node->block > new->block) {
+			list_add_tail(&new->lh, &node->lh);
+			return;
+		}
+	list_add_tail(&new->lh, &ext4_extent_lh);
+}
+
+static int __ext4fs_build_extent_cache(struct ext2_data *data,
+		struct ext4_extent_header *ext_block)
+{
+	int blksz = EXT2_BLOCK_SIZE(data);
+	int log2_blksz = LOG2_BLOCK_SIZE(data)
+		- get_fs()->dev_desc->log2blksz;
+	struct ext4_extent_node *node;
 	struct ext4_extent_idx *index;
+	struct ext4_extent *extent;
 	unsigned long long block;
-	int blksz = EXT2_BLOCK_SIZE(data);
+	char *buf;
-	int i;
+	int i, err;
-
-	while (1) {
-		index = (struct ext4_extent_idx *)(ext_block + 1);
 
 	if (le16_to_cpu(ext_block->eh_magic) != EXT4_EXT_MAGIC)
-			return 0;
+		return -EINVAL;
-
-		if (ext_block->eh_depth == 0)
-			return ext_block;
-		i = -1;
-		do {
-			i++;
-			if (i >= le16_to_cpu(ext_block->eh_entries))
-				break;
-		} while (fileblock >= le32_to_cpu(index[i].ei_block));
 
-		if (--i < 0)
+	if (ext_block->eh_depth == 0) {
+		extent = (struct ext4_extent *)(ext_block + 1);
+		for (i = 0; i < le16_to_cpu(ext_block->eh_entries); i++) {
+			node = malloc(sizeof(*node));
+			if (!node)
+				return -ENOMEM;
+			node->block = le32_to_cpu(extent[i].ee_block);
+			node->len = le16_to_cpu(extent[i].ee_len);
+			node->start = le16_to_cpu(extent[i].ee_start_hi);
+			node->start = (node->start << 32) +
+				le32_to_cpu(extent[i].ee_start_lo);
+			ext4fs_extent_cache_insert(node);
+		}
 		return 0;
+	}
+
+	index = (struct ext4_extent_idx *)(ext_block + 1);
+	for (i = 0; i < le16_to_cpu(ext_block->eh_entries); i++) {
+		buf = malloc(blksz);
+		if (!buf)
+			return -ENOMEM;
 
 		block = le16_to_cpu(index[i].ei_leaf_hi);
 		block = (block << 32) + le32_to_cpu(index[i].ei_leaf_lo);
 
-		if (ext4fs_devread((lbaint_t)block << log2_blksz, 0, blksz,
+		if (!ext4fs_devread(block << log2_blksz, 0, blksz, buf)) {
-				   buf))
+			free(buf);
-			ext_block = (struct ext4_extent_header *)buf;
+			return -EIO;
-		else
+		}
+
+		err = __ext4fs_build_extent_cache(data,
+				(struct ext4_extent_header *) buf);
+		free(buf);
+		if (err < 0)
+			return err;
+	}
+
 	return 0;
 }
+
+int ext4fs_build_extent_cache(struct ext2_inode *inode)
+{
+	return __ext4fs_build_extent_cache(ext4fs_root,
+			(struct ext4_extent_header *)
+			inode->b.blocks.dir_blocks);
+}
+
+void ext4fs_free_extent_cache(void)
+{
+	struct ext4_extent_node *node, *tmp;
+
+	list_for_each_entry_safe(node, tmp, &ext4_extent_lh, lh) {
+		list_del(&node->lh);
+		free(node);
+	}
+}
+
+static struct ext4_extent_node *ext4fs_extent_cache_get(uint32_t block)
+{
+	struct ext4_extent_node *node;
+
+	list_for_each_entry(node, &ext4_extent_lh, lh)
+		if (block >= node->block && block < node->block + node->len)
+			return node;
+
+	return NULL;
 }
 
 static int ext4fs_blockgroup
@@ -1508,56 +1574,24 @@ long int read_allocated_block(struct ext2_inode *inode, int fileblock)
 	long int rblock;
 	long int perblock_parent;
 	long int perblock_child;
-	unsigned long long start;
+
 	/* get the blocksize of the filesystem */
 	blksz = EXT2_BLOCK_SIZE(ext4fs_root);
 	log2_blksz = LOG2_BLOCK_SIZE(ext4fs_root)
 		- get_fs()->dev_desc->log2blksz;
 
 	if (le32_to_cpu(inode->flags) & EXT4_EXTENTS_FL) {
-		char *buf = zalloc(blksz);
+		struct ext4_extent_node *node;
-		if (!buf)
-			return -ENOMEM;
-		struct ext4_extent_header *ext_block;
-		struct ext4_extent *extent;
-		int i = -1;
-		ext_block =
-			ext4fs_get_extent_block(ext4fs_root, buf,
-						(struct ext4_extent_header *)
-						inode->b.blocks.dir_blocks,
-						fileblock, log2_blksz);
-		if (!ext_block) {
-			printf("invalid extent block\n");
-			free(buf);
-			return -EINVAL;
-		}
-
-		extent = (struct ext4_extent *)(ext_block + 1);
-
-		do {
-			i++;
-			if (i >= le16_to_cpu(ext_block->eh_entries))
-				break;
-		} while (fileblock >= le32_to_cpu(extent[i].ee_block));
-		if (--i >= 0) {
-			fileblock -= le32_to_cpu(extent[i].ee_block);
-			if (fileblock >= le16_to_cpu(extent[i].ee_len)) {
-				free(buf);
-				return 0;
-			}
-
-			start = le16_to_cpu(extent[i].ee_start_hi);
-			start = (start << 32) +
-					le32_to_cpu(extent[i].ee_start_lo);
-			free(buf);
-			return fileblock + start;
-		}
 
+		node = ext4fs_extent_cache_get(fileblock);
+		if (!node) {
 			printf("Extent Error\n");
-		free(buf);
 			return -1;
 		}
 
+		return fileblock - node->block + node->start;
+	}
+
 	/* Direct blocks. */
 	if (fileblock < INDIRECT_BLOCKS)
 		blknr = __le32_to_cpu(inode->b.blocks.dir_blocks[fileblock]);

fs/ext4/ext4_common.h

@@ -57,6 +57,9 @@ int ext4fs_find_file(const char *path, struct ext2fs_node *rootnode,
 int ext4fs_iterate_dir(struct ext2fs_node *dir, char *name,
 			struct ext2fs_node **fnode, int *ftype);
 
+int ext4fs_build_extent_cache(struct ext2_inode *inode);
+void ext4fs_free_extent_cache(void);
+
 #if defined(CONFIG_EXT4_WRITE)
 uint32_t ext4fs_div_roundup(uint32_t size, uint32_t n);
 int ext4fs_checksum_update(unsigned int i);

fs/ext4/ext4fs.c

@@ -63,6 +63,14 @@ int ext4fs_read_file(struct ext2fs_node *node, int pos,
 	char *delayed_buf = NULL;
 	short status;
 
+	if (le32_to_cpu(node->inode.flags) & EXT4_EXTENTS_FL) {
+		if (ext4fs_build_extent_cache(&node->inode)) {
+			printf("Error building extent cache!\n");
+			len = -1;
+			goto out_exit;
+		}
+	}
+
 	/* Adjust len so it we can't read past the end of the file. */
 	if (len > filesize)
 		len = filesize;
@@ -75,8 +83,10 @@ int ext4fs_read_file(struct ext2fs_node *node, int pos,
 		int blockend = blocksize;
 		int skipfirst = 0;
 		blknr = read_allocated_block(&(node->inode), i);
-		if (blknr < 0)
+		if (blknr < 0) {
-			return -1;
+			len = -1;
+			goto out_exit;
+		}
 
 		blknr = blknr << log2_fs_blocksize;
 
@@ -106,8 +116,10 @@ int ext4fs_read_file(struct ext2fs_node *node, int pos,
 							delayed_skipfirst,
 							delayed_extent,
 							delayed_buf);
-					if (status == 0)
+					if (status == 0) {
-						return -1;
+						len = -1;
+						goto out_exit;
+					}
 					previous_block_number = blknr;
 					delayed_start = blknr;
 					delayed_extent = blockend;
@@ -132,8 +144,10 @@ int ext4fs_read_file(struct ext2fs_node *node, int pos,
 							delayed_skipfirst,
 							delayed_extent,
 							delayed_buf);
-				if (status == 0)
+				if (status == 0) {
-					return -1;
+					len = -1;
+					goto out_exit;
+				}
 				previous_block_number = -1;
 			}
 			memset(buf, 0, blocksize - skipfirst);
@@ -145,11 +159,17 @@ int ext4fs_read_file(struct ext2fs_node *node, int pos,
 		status = ext4fs_devread(delayed_start,
 					delayed_skipfirst, delayed_extent,
 					delayed_buf);
-		if (status == 0)
+		if (status == 0) {
-			return -1;
+			len = -1;
+			goto out_exit;
+		}
 		previous_block_number = -1;
 	}
 
+
+out_exit:
+	ext4fs_free_extent_cache();
+
 	return len;
 }