Re: [PATCH next] fs: fix oob in do_handle_open

From: Gustavo A. R. Silva
Date: Wed Apr 03 2024 - 08:59:53 EST




On 03/04/24 02:48, Jeff Layton wrote:
On Wed, 2024-04-03 at 14:54 +0800, Edward Adam Davis wrote:
[Syzbot reported]
BUG: KASAN: slab-out-of-bounds in instrument_copy_from_user_before include/linux/instrumented.h:129 [inline]
BUG: KASAN: slab-out-of-bounds in _copy_from_user+0x7b/0xe0 lib/usercopy.c:22
Write of size 48 at addr ffff88802b8cbc88 by task syz-executor333/5090

CPU: 0 PID: 5090 Comm: syz-executor333 Not tainted 6.9.0-rc2-next-20240402-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 03/27/2024
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0x241/0x360 lib/dump_stack.c:114
print_address_description mm/kasan/report.c:377 [inline]
print_report+0x169/0x550 mm/kasan/report.c:488
kasan_report+0x143/0x180 mm/kasan/report.c:601
kasan_check_range+0x282/0x290 mm/kasan/generic.c:189
instrument_copy_from_user_before include/linux/instrumented.h:129 [inline]
_copy_from_user+0x7b/0xe0 lib/usercopy.c:22
copy_from_user include/linux/uaccess.h:183 [inline]
handle_to_path fs/fhandle.c:203 [inline]
do_handle_open+0x204/0x660 fs/fhandle.c:226
do_syscall_64+0xfb/0x240
entry_SYSCALL_64_after_hwframe+0x72/0x7a
[Fix]
When copying data to f_handle, the length of the copied data should not include
the length of "struct file_handle".

Reported-by: syzbot+4139435cb1b34cf759c2@xxxxxxxxxxxxxxxxxxxxxxxxx
Signed-off-by: Edward Adam Davis <eadavis@xxxxxx>
---
fs/fhandle.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/fhandle.c b/fs/fhandle.c
index 53ed54711cd2..8a7f86c2139a 100644
--- a/fs/fhandle.c
+++ b/fs/fhandle.c
@@ -202,7 +202,7 @@ static int handle_to_path(int mountdirfd, struct file_handle __user *ufh,
*handle = f_handle;
if (copy_from_user(&handle->f_handle,
&ufh->f_handle,
- struct_size(ufh, f_handle, f_handle.handle_bytes))) {
+ f_handle.handle_bytes)) {
retval = -EFAULT;
goto out_handle;
}

cc'ing Gustavo, since it looks like his patch in -next is what broke
this.


Oh, sorry about that folks. That looks pretty much like a copy/paste error.

The fix is correct.

Thanks, Edward!
--
Gustavo