KEYS: trusted: fix writing past end of buffer in trusted_read()

When calling keyctl_read() on a key of type "trusted", if the
user-supplied buffer was too small, the kernel ignored the buffer length
and just wrote past the end of the buffer, potentially corrupting
userspace memory.  Fix it by instead returning the size required, as per
the documentation for keyctl_read().

We also don't even fill the buffer at all in this case, as this is
slightly easier to implement than doing a short read, and either
behavior appears to be permitted.  It also makes it match the behavior
of the "encrypted" key type.

Fixes: d00a1c72f7 ("keys: add new trusted key-type")
Reported-by: Ben Hutchings <ben@decadent.org.uk>
Cc: <stable@vger.kernel.org> # v2.6.38+
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Mimi Zohar <zohar@linux.vnet.ibm.com>
Reviewed-by: James Morris <james.l.morris@oracle.com>
Signed-off-by: James Morris <james.l.morris@oracle.com>
This commit is contained in:
Eric Biggers 2017-11-02 00:47:12 +00:00 committed by James Morris
parent 3239b6f29b
commit a3c812f7cf

View File

@ -1147,20 +1147,21 @@ static long trusted_read(const struct key *key, char __user *buffer,
p = dereference_key_locked(key);
if (!p)
return -EINVAL;
if (!buffer || buflen <= 0)
return 2 * p->blob_len;
ascii_buf = kmalloc(2 * p->blob_len, GFP_KERNEL);
if (!ascii_buf)
return -ENOMEM;
bufp = ascii_buf;
for (i = 0; i < p->blob_len; i++)
bufp = hex_byte_pack(bufp, p->blob[i]);
if ((copy_to_user(buffer, ascii_buf, 2 * p->blob_len)) != 0) {
if (buffer && buflen >= 2 * p->blob_len) {
ascii_buf = kmalloc(2 * p->blob_len, GFP_KERNEL);
if (!ascii_buf)
return -ENOMEM;
bufp = ascii_buf;
for (i = 0; i < p->blob_len; i++)
bufp = hex_byte_pack(bufp, p->blob[i]);
if (copy_to_user(buffer, ascii_buf, 2 * p->blob_len) != 0) {
kzfree(ascii_buf);
return -EFAULT;
}
kzfree(ascii_buf);
return -EFAULT;
}
kzfree(ascii_buf);
return 2 * p->blob_len;
}