[PATCH] ext4: Fix loff_t overflow in ext4_max_bitmap_size()

We should use unsigned long long rather than loff_t to avoid
overflow in ext4_max_bitmap_size() for comparison before returning.
w/o this patch sbi->s_bitmap_maxbytes was becoming a negative
value due to overflow of upper_limit (with has_huge_files as true)

Below is a quick test to trigger it on a 64KB pagesize system.

sudo mkfs.ext4 -b 65536 -O ^has_extents,^64bit /dev/loop2
sudo mount /dev/loop2 /mnt
sudo echo "hello" > /mnt/hello 	-> This will error out with
				"echo: write error: File too large"

Signed-off-by: Ritesh Harjani <redacted>
---
 fs/ext4/super.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 7dc94f3e18e6..bedb66386966 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c

@@ -3189,17 +3189,17 @@ static loff_t ext4_max_size(int blkbits, int has_huge_files)
  */
 static loff_t ext4_max_bitmap_size(int bits, int has_huge_files)
 {
-	loff_t res = EXT4_NDIR_BLOCKS;
+	unsigned long long upper_limit, res = EXT4_NDIR_BLOCKS;
 	int meta_blocks;
-	loff_t upper_limit;
-	/* This is calculated to be the largest file size for a dense, block
+
+	/*
+	 * This is calculated to be the largest file size for a dense, block
 	 * mapped file such that the file's total number of 512-byte sectors,
 	 * including data and all indirect blocks, does not exceed (2^48 - 1).
 	 *
 	 * __u32 i_blocks_lo and _u16 i_blocks_high represent the total
 	 * number of 512-byte sectors of the file.
 	 */
-
 	if (!has_huge_files) {
 		/*
 		 * !has_huge_files or implies that the inode i_block field

@@ -3242,7 +3242,7 @@ static loff_t ext4_max_bitmap_size(int bits, int has_huge_files)
 	if (res > MAX_LFS_FILESIZE)
 		res = MAX_LFS_FILESIZE;
 
-	return res;
+	return (loff_t)res;
 }
 
 static ext4_fsblk_t descriptor_loc(struct super_block *sb,

-- 
2.31.1

On Sat 05-06-21 10:39:32, Ritesh Harjani wrote:

We should use unsigned long long rather than loff_t to avoid
overflow in ext4_max_bitmap_size() for comparison before returning.
w/o this patch sbi->s_bitmap_maxbytes was becoming a negative
value due to overflow of upper_limit (with has_huge_files as true)

Below is a quick test to trigger it on a 64KB pagesize system.

sudo mkfs.ext4 -b 65536 -O ^has_extents,^64bit /dev/loop2
sudo mount /dev/loop2 /mnt
sudo echo "hello" > /mnt/hello 	-> This will error out with
				"echo: write error: File too large"

Signed-off-by: Ritesh Harjani <redacted>

OK, this works (although it's really tight ;). Won't it be somewhat safer
if we compared upper_limit and res before shifting both by blocksize_bits
to the left? Basically we need to shift only for comparison with
MAX_LFS_FILESIZE which is in bytes... But either way feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

-- 
Jan Kara [off-list ref]
SUSE Labs, CR

On 21/06/07 12:00PM, Jan Kara wrote:

On Sat 05-06-21 10:39:32, Ritesh Harjani wrote:

quoted

We should use unsigned long long rather than loff_t to avoid
overflow in ext4_max_bitmap_size() for comparison before returning.
w/o this patch sbi->s_bitmap_maxbytes was becoming a negative
value due to overflow of upper_limit (with has_huge_files as true)

Below is a quick test to trigger it on a 64KB pagesize system.

sudo mkfs.ext4 -b 65536 -O ^has_extents,^64bit /dev/loop2
sudo mount /dev/loop2 /mnt
sudo echo "hello" > /mnt/hello 	-> This will error out with
				"echo: write error: File too large"

Signed-off-by: Ritesh Harjani <redacted>

OK, this works (although it's really tight ;). Won't it be somewhat safer
if we compared upper_limit and res before shifting both by blocksize_bits
to the left? Basically we need to shift only for comparison with
MAX_LFS_FILESIZE which is in bytes... But either way feel free to add:

Yes, at 1st I did think that, but since for comparing "res" against
MAX_LFS_FILESIZE we will be anyway doing the bit shifting and since this logic
too was (just) fitting into the limits so I thought of keeping it the same.
But, if absolutely required, I can make those changes.

Reviewed-by: Jan Kara <jack@suse.cz>

Thanks for the review :)

-ritesh

								Honza

quoted

---
 fs/ext4/super.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 7dc94f3e18e6..bedb66386966 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c

@@ -3189,17 +3189,17 @@ static loff_t ext4_max_size(int blkbits, int has_huge_files)
  */
 static loff_t ext4_max_bitmap_size(int bits, int has_huge_files)
 {
-	loff_t res = EXT4_NDIR_BLOCKS;
+	unsigned long long upper_limit, res = EXT4_NDIR_BLOCKS;
 	int meta_blocks;
-	loff_t upper_limit;
-	/* This is calculated to be the largest file size for a dense, block
+
+	/*
+	 * This is calculated to be the largest file size for a dense, block
 	 * mapped file such that the file's total number of 512-byte sectors,
 	 * including data and all indirect blocks, does not exceed (2^48 - 1).
 	 *
 	 * __u32 i_blocks_lo and _u16 i_blocks_high represent the total
 	 * number of 512-byte sectors of the file.
 	 */
-
 	if (!has_huge_files) {
 		/*
 		 * !has_huge_files or implies that the inode i_block field

@@ -3242,7 +3242,7 @@ static loff_t ext4_max_bitmap_size(int bits, int has_huge_files)
 	if (res > MAX_LFS_FILESIZE)
 		res = MAX_LFS_FILESIZE;

-	return res;
+	return (loff_t)res;
 }

 static ext4_fsblk_t descriptor_loc(struct super_block *sb,
--

2.31.1

--
Jan Kara [off-list ref]
SUSE Labs, CR

On Sat, 5 Jun 2021 10:39:32 +0530, Ritesh Harjani wrote:

We should use unsigned long long rather than loff_t to avoid
overflow in ext4_max_bitmap_size() for comparison before returning.
w/o this patch sbi->s_bitmap_maxbytes was becoming a negative
value due to overflow of upper_limit (with has_huge_files as true)

Below is a quick test to trigger it on a 64KB pagesize system.

[...]

Applied, thanks!

[1/1] ext4: Fix loff_t overflow in ext4_max_bitmap_size()
      commit: f9b9e1afe996e8b4a0a2ea8481c41756fff53d08

Best regards,
-- 
Theodore Ts'o [off-list ref]