Skip to contentSkip to conversation

ext4: pass strlen() of the symlink instead of i_size to inode_set_cached_link()

PATCH In review Mateusz Guzik 4 participants 7 messages Feb 5, 2025 – Feb 6, 2025

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ext4: pass strlen() of the symlink instead of i_size to inode_set_cached_link()

The call to nd_terminate_link() clamps the size to min(i_size, sizeof(ei->i_data) - 1), while the subsequent call to inode_set_cached_link() fails the possible update.

The kernel used to always strlen(), so do it now as well.

Reported-by [email protected]
Fixes bae80473f7b0 ("ext4: use inode_set_cached_link()")

---

Per my comments in:
https://lore.kernel.org/all/CAGudoHEv+Diti3r0x9VmF5ixgRVKk4trYnX_skVJNkQoTMaDHg@mail.gmail.com/#t

There is definitely a pre-existing bug in ext4 which the above happens to run into. I suspect the nd_terminate_link thing will disappear once that gets sorted out.

In the meantime the appropriate fix for 6.14 is to restore the original behavior of issuing strlen.

syzbot verified the issue is fixed:
https://lore.kernel.org/linux-hardening/[email protected]/T/#m340e6b52b9547ac85471a1da5980fe0a67c790ac

 fs/ext4/inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ext4/inode.c +1 -1
index 7c54ae5fcbd4..30cff983e601 100644
@@ -5010,7 +5010,7 @@ struct inode *__ext4_iget(struct super_block *sb, unsigned long ino,
 			nd_terminate_link(ei->i_data, inode->i_size,
 				sizeof(ei->i_data) - 1);
 			inode_set_cached_link(inode, (char *)ei->i_data,
-					      inode->i_size);
+					      strlen((char *)ei->i_data));
 		} else {
 			inode->i_op = &ext4_symlink_inode_operations;
 		}
Signature · 1 line
2.43.0
#1 lore raw
Quoted 5 lines · Mateusz Guzik

The call to nd_terminate_link() clamps the size to min(i_size, sizeof(ei->i_data) - 1), while the subsequent call to inode_set_cached_link() fails the possible update.

oof.. that should be:

Quoted 1 line

inode_set_cached_link() fails to take it into account

Signature · 1 line
Mateusz Guzik <mjguzik gmail.com>
#2 lore raw
In reply to Mateusz Guzik
Quoted 18 lines · Mateusz Guzik

The call to nd_terminate_link() clamps the size to min(i_size, sizeof(ei->i_data) - 1), while the subsequent call to inode_set_cached_link() fails the possible update.

The kernel used to always strlen(), so do it now as well.

Reported-by [email protected]
Fixes bae80473f7b0 ("ext4: use inode_set_cached_link()")

---

Per my comments in:
https://lore.kernel.org/all/CAGudoHEv+Diti3r0x9VmF5ixgRVKk4trYnX_skVJNkQoTMaDHg@mail.gmail.com/#t

There is definitely a pre-existing bug in ext4 which the above happens to run into. I suspect the nd_terminate_link thing will disappear once that gets sorted out.

Quoted 5 lines

In the meantime the appropriate fix for 6.14 is to restore the original behavior of issuing strlen.

syzbot verified the issue is fixed:
https://lore.kernel.org/linux-hardening/[email protected]/T/#m340e6b52b9547ac85471a1da5980fe0a67c790ac

Again, this is evidence of inconsistent inode metadata, which should be dealt with by returning EFSCORRUPTED, not arbitrarily truncating the contents of a bad inode.

https://lore.kernel.org/linux-fsdevel/CAGudoHHeHKo6+R86pZTFSzAFRf2v=bc5LOGvbHmC0mCfkjRvgw@mail.gmail.com/T/#mf05b770926225812f8c78c58c6f3b707c7d151d8

To spell this out -- this is ext4_inode_is_fast_symlink:

int ext4_inode_is_fast_symlink(struct inode *inode)

{
	if (!(EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL)) {
		int ea_blocks = EXT4_I(inode)->i_file_acl ?
				EXT4_CLUSTER_SIZE(inode->i_sb) >> 9 : 0;

		if (ext4_has_inline_data(inode))
			return 0;

		return (S_ISLNK(inode->i_mode) && inode->i_blocks - ea_blocks == 0);
	}
	return S_ISLNK(inode->i_mode) && inode->i_size &&
	       (inode->i_size < EXT4_N_BLOCKS * 4);
}

Note in the !EA_INODE and !inlinedata case, the decision is made based on the block count of the file, without checking i_size. The callsite should look more like this:

		} else if (ext4_inode_is_fast_symlink(inode)) {
			if (inode->i_size == 0 ||
			    inode->i_size > sizeof(ei->i_data) - 1) {
				ret = -EFSCORRUPTED;
				ext4_error_inode(..."fast symlink doesn't fit in body??");
				goto bad_inode;
			}
			inode->i_op = &ext4_fast_symlink_inode_operations;
			nd_terminate_link(ei->i_data, inode->i_size,
				sizeof(ei->i_data) - 1);
			inode_set_cached_link(inode, (char *)ei->i_data,
					      inode->i_size);
		} else {

And, seriously, cc the ext4 list on ext4 patches please.

--D

Quoted 20 lines
 fs/ext4/inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ext4/inode.c +1 -1
index 7c54ae5fcbd4..30cff983e601 100644
@@ -5010,7 +5010,7 @@ struct inode *__ext4_iget(struct super_block *sb, unsigned long ino,
 			nd_terminate_link(ei->i_data, inode->i_size,
 				sizeof(ei->i_data) - 1);
 			inode_set_cached_link(inode, (char *)ei->i_data,
-					      inode->i_size);
+					      strlen((char *)ei->i_data));
 		} else {
 			inode->i_op = &ext4_symlink_inode_operations;
 		}
Signature · 1 line
2.43.0
#3 lore raw
Quoted 23 lines · Darrick J. Wong
Quoted 21 lines · Mateusz Guzik

The call to nd_terminate_link() clamps the size to min(i_size, sizeof(ei->i_data) - 1), while the subsequent call to inode_set_cached_link() fails the possible update.

The kernel used to always strlen(), so do it now as well.

Reported-by [email protected]
Fixes bae80473f7b0 ("ext4: use inode_set_cached_link()")

---

Per my comments in:
https://lore.kernel.org/all/CAGudoHEv+Diti3r0x9VmF5ixgRVKk4trYnX_skVJNkQoTMaDHg@mail.gmail.com/#t

There is definitely a pre-existing bug in ext4 which the above happens to run into. I suspect the nd_terminate_link thing will disappear once that gets sorted out.

In the meantime the appropriate fix for 6.14 is to restore the original behavior of issuing strlen.

Quoted 7 lines
Quoted 2 lines

syzbot verified the issue is fixed:
https://lore.kernel.org/linux-hardening/[email protected]/T/#m340e6b52b9547ac85471a1da5980fe0a67c790ac

Again, this is evidence of inconsistent inode metadata, which should be dealt with by returning EFSCORRUPTED, not arbitrarily truncating the contents of a bad inode.

I agree, rejecting the inode was something I was advocating for from the get go.

I don't know if a real patch(tm) will materialize for 6.14, so in the meantime I can at least damage-control this back to the original state.

If the ext4 folk do the right fix, I will be delighted to have this patch dropped. :)

Quoted 1 line

And, seriously, cc the ext4 list on ext4 patches please.

Ye that's my bad.

Signature · 1 line
Mateusz Guzik <mjguzik gmail.com>
reviewer
#4 lore raw
Quoted 32 lines · Mateusz Guzik
Quoted 23 lines · Darrick J. Wong
Quoted 21 lines · Mateusz Guzik

The call to nd_terminate_link() clamps the size to min(i_size, sizeof(ei->i_data) - 1), while the subsequent call to inode_set_cached_link() fails the possible update.

The kernel used to always strlen(), so do it now as well.

Reported-by [email protected]
Fixes bae80473f7b0 ("ext4: use inode_set_cached_link()")

---

Per my comments in:
https://lore.kernel.org/all/CAGudoHEv+Diti3r0x9VmF5ixgRVKk4trYnX_skVJNkQoTMaDHg@mail.gmail.com/#t

There is definitely a pre-existing bug in ext4 which the above happens to run into. I suspect the nd_terminate_link thing will disappear once that gets sorted out.

In the meantime the appropriate fix for 6.14 is to restore the original behavior of issuing strlen.

Quoted 7 lines
Quoted 2 lines

syzbot verified the issue is fixed:
https://lore.kernel.org/linux-hardening/[email protected]/T/#m340e6b52b9547ac85471a1da5980fe0a67c790ac

Again, this is evidence of inconsistent inode metadata, which should be dealt with by returning EFSCORRUPTED, not arbitrarily truncating the contents of a bad inode.

Quoted 8 lines

I agree, rejecting the inode was something I was advocating for from the get go.

I don't know if a real patch(tm) will materialize for 6.14, so in the meantime I can at least damage-control this back to the original state.

If the ext4 folk do the right fix, I will be delighted to have this patch dropped. :)

Yeah, let me cook up proper ext4 fix for this (currently under testing).

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#5 lore raw
Quoted 7 lines · Jan Kara
Quoted 2 lines · Mateusz Guzik

If the ext4 folk do the right fix, I will be delighted to have this patch dropped. :)

Yeah, let me cook up proper ext4 fix for this (currently under testing).

I see it got posted and tested:
https://lore.kernel.org/linux-hardening/[email protected]/T/#mb782935cc6926dd5642984189d922135f023ec43

So I consider my patch self-NAKed.

Thanks for sorting it out.

Signature · 1 line
Mateusz Guzik <mjguzik gmail.com>
#6 lore raw
Quoted 6 lines · Mateusz Guzik
Quoted 5 lines · Jan Kara
Quoted 2 lines · Mateusz Guzik

If the ext4 folk do the right fix, I will be delighted to have this patch dropped. :)

Quoted 7 lines
Quoted 2 lines

Yeah, let me cook up proper ext4 fix for this (currently under testing).

I see it got posted and tested:
https://lore.kernel.org/linux-hardening/[email protected]/T/#mb782935cc6926dd5642984189d922135f023ec43

So I consider my patch self-NAKed.

Dropped.