Revision da353fac65fede6b8b4cfe207f0d9408e3121105 authored by Daniel Jordan on 27 October 2021, 21:59:20 UTC, committed by David S. Miller on 28 October 2021, 13:41:20 UTC
sk->sk_err appears to expect a positive value, a convention that ktls
doesn't always follow and that leads to memory corruption in other code.
For instance,

    [kworker]
    tls_encrypt_done(..., err=<negative error from crypto request>)
      tls_err_abort(.., err)
        sk->sk_err = err;

    [task]
    splice_from_pipe_feed
      ...
        tls_sw_do_sendpage
          if (sk->sk_err) {
            ret = -sk->sk_err;  // ret is positive

    splice_from_pipe_feed (continued)
      ret = actor(...)  // ret is still positive and interpreted as bytes
                        // written, resulting in underflow of buf->len and
                        // sd->len, leading to huge buf->offset and bogus
                        // addresses computed in later calls to actor()

Fix all tls_err_abort() callers to pass a negative error code
consistently and centralize the error-prone sign flip there, throwing in
a warning to catch future misuse and uninlining the function so it
really does only warn once.

Cc: stable@vger.kernel.org
Fixes: c46234ebb4d1e ("tls: RX path for ktls")
Reported-by: syzbot+b187b77c8474f9648fae@syzkaller.appspotmail.com
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
1 parent a32f07d
History
File Mode Size
ChangeLog -rw-r--r-- 14.0 KB
Kconfig -rw-r--r-- 957 bytes
Makefile -rw-r--r-- 246 bytes
befs.h -rw-r--r-- 3.0 KB
befs_fs_types.h -rw-r--r-- 5.0 KB
btree.c -rw-r--r-- 21.9 KB
btree.h -rw-r--r-- 343 bytes
datastream.c -rw-r--r-- 15.9 KB
datastream.h -rw-r--r-- 575 bytes
debug.c -rw-r--r-- 7.1 KB
endian.h -rw-r--r-- 3.2 KB
inode.c -rw-r--r-- 1.2 KB
inode.h -rw-r--r-- 120 bytes
io.c -rw-r--r-- 1.4 KB
io.h -rw-r--r-- 110 bytes
linuxvfs.c -rw-r--r-- 25.5 KB
super.c -rw-r--r-- 3.3 KB
super.h -rw-r--r-- 131 bytes

back to top