ext4: validate readdir offset before accessing dirent

From: Yao Kai <[email protected]>

commit bc4b7b0414c33b2c8898eb04386df0d21a13dad8 upstream.

A corrupted directory can trigger the following KASAN report when ext4_readdir() resumes from an invalid position:

  BUG: KASAN: use-after-free in __ext4_check_dir_entry+0x5ef/0x820
  Read of size 2 at addr ffff88810a646000 by task repro_linear/509

  Call Trace:
   <TASK>
   dump_stack_lvl+0x53/0x70
   print_report+0xd0/0x630
   kasan_report+0xce/0x100
   __ext4_check_dir_entry+0x5ef/0x820
   ext4_readdir+0xcde/0x2b70
   iterate_dir+0x1a1/0x520
   __x64_sys_getdents64+0x12b/0x220
   do_syscall_64+0xf9/0x540
   entry_SYSCALL_64_after_hwframe+0x77/0x7f
   </TASK>

KASAN reports use-after-free because the out-of-bounds access lands in an adjacent freed page. The directory buffer itself is still referenced.

ext4_dir_llseek() invalidates the directory cookie so that ext4_readdir() rescans directory entries from the start of the block. The rescan checks only the lower bound of rec_len before advancing. A corrupted rec_len can therefore place the offset where the block has insufficient space for a complete directory entry. The rescan itself may dereference that truncated entry, or the main loop may pass it to __ext4_check_dir_entry(). The latter reads de->rec_len before validating the range. For example:

block offset 0 4092 4096

                |---- de1.rec_len = 4092 -----|----|
                                               de2.inode
                                                    | de2.rec_len
                                                    ^ OOB, reported as UAF

de2 starts at offset 4092 in this 4 KiB block. Its four-byte inode fits in the block, but its rec_len starts at offset 4096 and crosses the boundary.

The minimum safe length is inode-dependent. Encrypted and casefolded directory entries need eight additional hash bytes, while a valid metadata checksum tail is only 12 bytes.

Cache the metadata checksum feature state and derive the minimum directory entry length from the on-disk format. Use it to bound both the rescan and the offset passed to the main loop. Report an offset in a truncated block tail and skip the remainder of the block, while continuing to accept an offset exactly at the block boundary.

---
 fs/ext4/dir.c | 20 ++++++++++++++++++--
 1 file changed, 18 insertions(+), 2 deletions(-)

index b278b5703c197..265e544cc3482 100644
@@ -137,6 +137,7 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 	struct buffer_head *bh = NULL;
 	struct fscrypt_str fstr = FSTR_INIT(NULL, 0);
 	struct dir_private_info *info = file->private_data;
+	bool has_csum = ext4_has_metadata_csum(sb);
 
 	err = fscrypt_prepare_readdir(inode);
 	if (err)

@@ -148,7 +149,7 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 			return err;
 
 		/* Can we just clear INDEX flag to ignore htree information? */
-		if (!ext4_has_metadata_csum(sb)) {
+		if (!has_csum) {
 			/*
 			 * We don't set the inode dirty flag since it's not
 			 * critical that it gets flushed back to the disk.

@@ -234,7 +235,10 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 		 * dirent right now.  Scan from the start of the block
 		 * to make sure. */
 		if (!inode_eq_iversion(inode, info->cookie)) {
-			for (i = 0; i < sb->s_blocksize && i < offset; ) {
+			for (i = 0;
+			     i <= sb->s_blocksize -
+				  ext4_dir_rec_len(1, has_csum ? NULL : inode) &&
+			     i < offset;) {
 				de = (struct ext4_dir_entry_2 *)
 					(bh->b_data + i);
 				/* It's too expensive to do a full

@@ -256,6 +260,17 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 			info->cookie = inode_query_iversion(inode);
 		}
 
+		if (unlikely(offset < sb->s_blocksize &&
+			     offset > sb->s_blocksize -
+			     ext4_dir_rec_len(1, has_csum ? NULL : inode))) {
+			EXT4_ERROR_FILE(file, bh->b_blocknr,
+					"bad entry in directory: %s - offset=%u, size=%lu",
+					"directory entry too close to block end",
+					offset, sb->s_blocksize);
+			ctx->pos = round_up(ctx->pos, sb->s_blocksize);
+			goto next_block;
+		}
+
 		while (ctx->pos < inode->i_size
 		       && offset < sb->s_blocksize) {
 			de = (struct ext4_dir_entry_2 *) (bh->b_data + offset);

@@ -311,6 +326,7 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 			ctx->pos += ext4_rec_len_from_disk(de->rec_len,
 						sb->s_blocksize);
 		}
+next_block:
 		if ((ctx->pos < inode->i_size) && !dir_relax_shared(inode))
 			goto done;
 		brelse(bh);

base-commit: 85d82e0d7c2dfd3e75d34f35b7ebbf41f45c9c4e

2.56.0

[+Cc recipients that Sashiko dropped]

Should the Sashiko configuration for linux-ext4 be updated to enable reply-all? That would be my preference, and various other lists already use that setting.

Sounds like this fix was incomplete but the backport is still safe.

- Eric