[38/41] ext3: fix rec_len overflow with 64KB block size

From: Christoph Lameter
Date: Tue Sep 11 2007 - 02:19:08 EST

Next message: Christoph Lameter: "[24/41] compound pages: Use new compound vmstat functions in SLUB"
Previous message: Christoph Lameter: "[40/41] Do not use f_mapping in simple_prepare_write()"
In reply to: Christoph Lameter: "[40/41] Do not use f_mapping in simple_prepare_write()"
Next in thread: Christoph Lameter: "[24/41] compound pages: Use new compound vmstat functions in SLUB"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

[3/4] ext3: fix rec_len overflow
- prevent rec_len from overflow with 64KB blocksize

Signed-off-by: Takashi Sato <sho@xxxxxxxxxxxxxx>
Signed-off-by: Mingming Cao <cmm@xxxxxxxxxx>
Signed-off-by: Christoph Lameter <clameter@xxxxxxx>

---
fs/ext3/dir.c | 13 +++++--
fs/ext3/namei.c | 88 +++++++++++++++++++++++++++++++++++++++--------
include/linux/ext3_fs.h | 9 +++++
3 files changed, 91 insertions(+), 19 deletions(-)

diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c
index a65b5a7..a544da8 100644
--- a/fs/ext3/dir.c
+++ b/fs/ext3/dir.c
@@ -100,10 +100,11 @@ static int ext3_readdir(struct file * filp,
unsigned long offset;
int i, stored;
struct ext3_dir_entry_2 *de;
- struct super_block *sb;
int err;
struct inode *inode = filp->f_path.dentry->d_inode;
int ret = 0;
+ struct super_block *sb = inode->i_sb;
+ unsigned tail = sb->s_blocksize;

sb = inode->i_sb;

@@ -168,8 +169,11 @@ revalidate:
* readdir(2), then we might be pointing to an invalid
* dirent right now. Scan from the start of the block
* to make sure. */
- if (filp->f_version != inode->i_version) {
- for (i = 0; i < sb->s_blocksize && i < offset; ) {
+ if (tail > EXT3_DIR_MAX_REC_LEN) {
+ tail = EXT3_DIR_MAX_REC_LEN;
+ }
+ if (filp->f_version != inode->i_version) {
+ for (i = 0; i < tail && i < offset; ) {
de = (struct ext3_dir_entry_2 *)
(bh->b_data + i);
/* It's too expensive to do a full
@@ -190,7 +194,7 @@ revalidate:
}

while (!error && filp->f_pos < inode->i_size
- && offset < sb->s_blocksize) {
+ && offset < tail) {
de = (struct ext3_dir_entry_2 *) (bh->b_data + offset);
if (!ext3_check_dir_entry ("ext3_readdir", inode, de,
bh, offset)) {
@@ -226,6 +230,7 @@ revalidate:
}
filp->f_pos += le16_to_cpu(de->rec_len);
}
+ filp->f_pos = EXT3_DIR_ADJUST_TAIL_OFFS(filp->f_pos, sb->s_blocksize);
offset = 0;
brelse (bh);
}
diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c
index 1586807..fe7f433 100644
--- a/fs/ext3/namei.c
+++ b/fs/ext3/namei.c
@@ -262,9 +262,13 @@ static struct stats dx_show_leaf(struct dx_hash_info *hinfo, struct ext3_dir_ent
unsigned names = 0, space = 0;
char *base = (char *) de;
struct dx_hash_info h = *hinfo;
+ unsigned tail = size;

printk("names: ");
- while ((char *) de < base + size)
+ if (tail > EXT3_DIR_MAX_REC_LEN) {
+ tail = EXT3_DIR_MAX_REC_LEN;
+ }
+ while ((char *) de < base + tail)
{
if (de->inode)
{
@@ -677,8 +681,12 @@ static int dx_make_map (struct ext3_dir_entry_2 *de, int size,
int count = 0;
char *base = (char *) de;
struct dx_hash_info h = *hinfo;
+ unsigned tail = size;

- while ((char *) de < base + size)
+ if (tail > EXT3_DIR_MAX_REC_LEN) {
+ tail = EXT3_DIR_MAX_REC_LEN;
+ }
+ while ((char *) de < base + tail)
{
if (de->name_len && de->inode) {
ext3fs_dirhash(de->name, de->name_len, &h);
@@ -775,9 +783,13 @@ static inline int search_dirblock(struct buffer_head * bh,
int de_len;
const char *name = dentry->d_name.name;
int namelen = dentry->d_name.len;
+ unsigned tail = dir->i_sb->s_blocksize;

de = (struct ext3_dir_entry_2 *) bh->b_data;
- dlimit = bh->b_data + dir->i_sb->s_blocksize;
+ if (tail > EXT3_DIR_MAX_REC_LEN) {
+ tail = EXT3_DIR_MAX_REC_LEN;
+ }
+ dlimit = bh->b_data + tail;
while ((char *) de < dlimit) {
/* this code is executed quadratically often */
/* do minimal checking `by hand' */
@@ -1115,6 +1127,9 @@ static struct ext3_dir_entry_2* dx_pack_dirents(char *base, int size)
unsigned rec_len = 0;

prev = to = de;
+ if (size > EXT3_DIR_MAX_REC_LEN) {
+ size = EXT3_DIR_MAX_REC_LEN;
+ }
while ((char*)de < base + size) {
next = (struct ext3_dir_entry_2 *) ((char *) de +
le16_to_cpu(de->rec_len));
@@ -1180,8 +1195,15 @@ static struct ext3_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
/* Fancy dance to stay within two buffers */
de2 = dx_move_dirents(data1, data2, map + split, count - split);
de = dx_pack_dirents(data1,blocksize);
- de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de);
- de2->rec_len = cpu_to_le16(data2 + blocksize - (char *) de2);
+ if (blocksize < EXT3_DIR_MAX_REC_LEN) {
+ de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de);
+ de2->rec_len = cpu_to_le16(data2 + blocksize - (char *) de2);
+ } else {
+ de->rec_len = cpu_to_le16(data1 + EXT3_DIR_MAX_REC_LEN -
+ (char *) de);
+ de2->rec_len = cpu_to_le16(data2 + EXT3_DIR_MAX_REC_LEN -
+ (char *) de2);
+ }
dxtrace(dx_show_leaf (hinfo, (struct ext3_dir_entry_2 *) data1, blocksize, 1));
dxtrace(dx_show_leaf (hinfo, (struct ext3_dir_entry_2 *) data2, blocksize, 1));

@@ -1236,11 +1258,15 @@ static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry,
unsigned short reclen;
int nlen, rlen, err;
char *top;
+ unsigned tail = dir->i_sb->s_blocksize;

+ if (tail > EXT3_DIR_MAX_REC_LEN) {
+ tail = EXT3_DIR_MAX_REC_LEN;
+ }
reclen = EXT3_DIR_REC_LEN(namelen);
if (!de) {
de = (struct ext3_dir_entry_2 *)bh->b_data;
- top = bh->b_data + dir->i_sb->s_blocksize - reclen;
+ top = bh->b_data + tail - reclen;
while ((char *) de <= top) {
if (!ext3_check_dir_entry("ext3_add_entry", dir, de,
bh, offset)) {
@@ -1354,13 +1380,21 @@ static int make_indexed_dir(handle_t *handle, struct dentry *dentry,
/* The 0th block becomes the root, move the dirents out */
fde = &root->dotdot;
de = (struct ext3_dir_entry_2 *)((char *)fde + le16_to_cpu(fde->rec_len));
- len = ((char *) root) + blocksize - (char *) de;
+ if (blocksize < EXT3_DIR_MAX_REC_LEN) {
+ len = ((char *) root) + blocksize - (char *) de;
+ } else {
+ len = ((char *) root) + EXT3_DIR_MAX_REC_LEN - (char *) de;
+ }
memcpy (data1, de, len);
de = (struct ext3_dir_entry_2 *) data1;
top = data1 + len;
while ((char *)(de2=(void*)de+le16_to_cpu(de->rec_len)) < top)
de = de2;
- de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de);
+ if (blocksize < EXT3_DIR_MAX_REC_LEN) {
+ de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de);
+ } else {
+ de->rec_len = cpu_to_le16(data1 + EXT3_DIR_MAX_REC_LEN - (char *) de);
+ }
/* Initialize the root; the dot dirents already exist */
de = (struct ext3_dir_entry_2 *) (&root->dotdot);
de->rec_len = cpu_to_le16(blocksize - EXT3_DIR_REC_LEN(2));
@@ -1450,7 +1484,11 @@ static int ext3_add_entry (handle_t *handle, struct dentry *dentry,
return retval;
de = (struct ext3_dir_entry_2 *) bh->b_data;
de->inode = 0;
- de->rec_len = cpu_to_le16(blocksize);
+ if (blocksize < EXT3_DIR_MAX_REC_LEN) {
+ de->rec_len = cpu_to_le16(blocksize);
+ } else {
+ de->rec_len = cpu_to_le16(EXT3_DIR_MAX_REC_LEN);
+ }
return add_dirent_to_buf(handle, dentry, inode, de, bh);
}

@@ -1514,7 +1552,12 @@ static int ext3_dx_add_entry(handle_t *handle, struct dentry *dentry,
goto cleanup;
node2 = (struct dx_node *)(bh2->b_data);
entries2 = node2->entries;
- node2->fake.rec_len = cpu_to_le16(sb->s_blocksize);
+ if (sb->s_blocksize < EXT3_DIR_MAX_REC_LEN) {
+ node2->fake.rec_len = cpu_to_le16(sb->s_blocksize);
+ } else {
+ node2->fake.rec_len =
+ cpu_to_le16(EXT3_DIR_MAX_REC_LEN);
+ }
node2->fake.inode = 0;
BUFFER_TRACE(frame->bh, "get_write_access");
err = ext3_journal_get_write_access(handle, frame->bh);
@@ -1602,11 +1645,15 @@ static int ext3_delete_entry (handle_t *handle,
{
struct ext3_dir_entry_2 * de, * pde;
int i;
+ unsigned tail = bh->b_size;

i = 0;
pde = NULL;
de = (struct ext3_dir_entry_2 *) bh->b_data;
- while (i < bh->b_size) {
+ if (tail > EXT3_DIR_MAX_REC_LEN) {
+ tail = EXT3_DIR_MAX_REC_LEN;
+ }
+ while (i < tail) {
if (!ext3_check_dir_entry("ext3_delete_entry", dir, de, bh, i))
return -EIO;
if (de == de_del) {
@@ -1766,7 +1813,11 @@ retry:
de = (struct ext3_dir_entry_2 *)
((char *) de + le16_to_cpu(de->rec_len));
de->inode = cpu_to_le32(dir->i_ino);
- de->rec_len = cpu_to_le16(inode->i_sb->s_blocksize-EXT3_DIR_REC_LEN(1));
+ if (inode->i_sb->s_blocksize < EXT3_DIR_MAX_REC_LEN) {
+ de->rec_len = cpu_to_le16(inode->i_sb->s_blocksize-EXT3_DIR_REC_LEN(1));
+ } else {
+ de->rec_len = cpu_to_le16(EXT3_DIR_MAX_REC_LEN-EXT3_DIR_REC_LEN(1));
+ }
de->name_len = 2;
strcpy (de->name, "..");
ext3_set_de_type(dir->i_sb, de, S_IFDIR);
@@ -1801,10 +1852,10 @@ static int empty_dir (struct inode * inode)
unsigned long offset;
struct buffer_head * bh;
struct ext3_dir_entry_2 * de, * de1;
- struct super_block * sb;
+ struct super_block * sb = inode->i_sb;
int err = 0;
+ unsigned tail = sb->s_blocksize;

- sb = inode->i_sb;
if (inode->i_size < EXT3_DIR_REC_LEN(1) + EXT3_DIR_REC_LEN(2) ||
!(bh = ext3_bread (NULL, inode, 0, 0, &err))) {
if (err)
@@ -1831,11 +1882,17 @@ static int empty_dir (struct inode * inode)
return 1;
}
offset = le16_to_cpu(de->rec_len) + le16_to_cpu(de1->rec_len);
+ if (offset == EXT3_DIR_MAX_REC_LEN) {
+ offset += sb->s_blocksize - EXT3_DIR_MAX_REC_LEN;
+ }
de = (struct ext3_dir_entry_2 *)
((char *) de1 + le16_to_cpu(de1->rec_len));
+ if (tail > EXT3_DIR_MAX_REC_LEN) {
+ tail = EXT3_DIR_MAX_REC_LEN;
+ }
while (offset < inode->i_size ) {
if (!bh ||
- (void *) de >= (void *) (bh->b_data+sb->s_blocksize)) {
+ (void *) de >= (void *) (bh->b_data + tail)) {
err = 0;
brelse (bh);
bh = ext3_bread (NULL, inode,
@@ -1862,6 +1919,7 @@ static int empty_dir (struct inode * inode)
return 0;
}
offset += le16_to_cpu(de->rec_len);
+ offset = EXT3_DIR_ADJUST_TAIL_OFFS(offset, sb->s_blocksize);
de = (struct ext3_dir_entry_2 *)
((char *) de + le16_to_cpu(de->rec_len));
}
diff --git a/include/linux/ext3_fs.h b/include/linux/ext3_fs.h
index 7aa5556..15aacd6 100644
--- a/include/linux/ext3_fs.h
+++ b/include/linux/ext3_fs.h
@@ -660,6 +660,15 @@ struct ext3_dir_entry_2 {
#define EXT3_DIR_ROUND (EXT3_DIR_PAD - 1)
#define EXT3_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT3_DIR_ROUND) & \
~EXT3_DIR_ROUND)
+#define EXT3_DIR_MAX_REC_LEN 65532
+
+/*
+ * Align a tail offset to the end of a directory block
+ */
+#define EXT3_DIR_ADJUST_TAIL_OFFS(offs, bsize) \
+ ((((offs) & ((bsize) -1)) == EXT3_DIR_MAX_REC_LEN) ? \
+ ((offs) + (bsize) - EXT3_DIR_MAX_REC_LEN):(offs))
+
/*
* Hash Tree Directory indexing
* (c) Daniel Phillips, 2001
--
1.5.2.5

--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Christoph Lameter: "[24/41] compound pages: Use new compound vmstat functions in SLUB"
Previous message: Christoph Lameter: "[40/41] Do not use f_mapping in simple_prepare_write()"
In reply to: Christoph Lameter: "[40/41] Do not use f_mapping in simple_prepare_write()"
Next in thread: Christoph Lameter: "[24/41] compound pages: Use new compound vmstat functions in SLUB"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]