Hi Christoph,
On 20/07/2020 14:47, Christoph Hellwig wrote:
> Rework the remaining setsockopt code to pass a sockptr_t instead of a
> plain user pointer. This removes the last remaining set_fs(KERNEL_DS)
> outside of architecture specific code.
>
> Signed-off-by: Christoph Hellwig <[email protected]>
> ---
...
> diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
> index 27b6f250b87dfd..30a8e697b9db9c 100644
> --- a/net/mptcp/protocol.c
> +++ b/net/mptcp/protocol.c
> @@ -1627,7 +1627,7 @@ static void mptcp_destroy(struct sock *sk)
> }
>
> static int mptcp_setsockopt_sol_socket(struct mptcp_sock *msk, int optname,
> - char __user *optval, unsigned int optlen)
> + sockptr_t optval, unsigned int optlen)
> {
> struct sock *sk = (struct sock *)msk;
> struct socket *ssock;
> @@ -1643,8 +1643,8 @@ static int mptcp_setsockopt_sol_socket(struct mptcp_sock *msk, int optname,
> return -EINVAL;
> }
>
> - ret = sock_setsockopt(ssock, SOL_SOCKET, optname,
> - USER_SOCKPTR(optval), optlen);
> + ret = sock_setsockopt(ssock, SOL_SOCKET, optname, optval,
> + optlen);
A very small detail related to the modifications in MPTCP code, only if
you have to send a v2 and if you don't mind: may you move "optlen" to
the previous line like it was before your patch 7/24. Same below at the
end of the function.
That would reduce the global diff in MPTCP files to function signatures
only.
Cheers,
Matt
--
Tessares | Belgium | Hybrid Access Solutions
http://www.tessares.net