diff mbox

[1/1] Fix getsockname (ipv4/ipv6)

Message ID 20090126220931.A8F0154A11@localhost
State Rejected, archived
Delegated to: David Miller
Headers show

Commit Message

John Reumann Jan. 23, 2009, 9:45 p.m. UTC
This change will return ENOSPC if the return address provided by the caller of getsockname
 is not sufficiently large to hold the return value.  This is consistent with the rest of the
 API. However, there are some implemenations that copy only as many bytes of sockaddr from the
 kernel to user-space as specified in the *addrlen attribute.  (1) This wasn't the behavior now
 (2) we should not adopt this behavior because it is quite messy, e.g., copy only part of a port,
 IP-address? Most partial retrivals do not make sense.  This is why I suggest returning EINVAL,
 when the return struct is too small.
---
 net/ipv4/af_inet.c  |    7 +++++--
 net/ipv6/af_inet6.c |    3 +++
 2 files changed, 8 insertions(+), 2 deletions(-)

Comments

Florian Westphal Jan. 26, 2009, 10:37 p.m. UTC | #1
John Reumann <reumann.linux@gmail.com> wrote:
> This change will return ENOSPC if the return address provided by the caller of getsockname
>  is not sufficiently large to hold the return value.

Those interfaces allow applications to query how large the sockaddr storage needs to
be by checking *addrlen after a successful call.

> This is consistent with the rest of the
>  API. However, there are some implemenations that copy only as many bytes of sockaddr from the
>  kernel to user-space as specified in the *addrlen attribute.  (1) This wasn't the behavior now

It is, and it has been for ages.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller Jan. 26, 2009, 10:42 p.m. UTC | #2
From: Florian Westphal <fw@strlen.de>
Date: Mon, 26 Jan 2009 23:37:33 +0100

> John Reumann <reumann.linux@gmail.com> wrote:
> > This change will return ENOSPC if the return address provided by the caller of getsockname
> >  is not sufficiently large to hold the return value.
> 
> Those interfaces allow applications to query how large the sockaddr storage needs to
> be by checking *addrlen after a successful call.

Right.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Arnaldo Carvalho de Melo Jan. 26, 2009, 10:48 p.m. UTC | #3
Em Fri, Jan 23, 2009 at 04:45:52PM -0500, John Reumann escreveu:
> This change will return ENOSPC if the return address provided by the caller of getsockname
>  is not sufficiently large to hold the return value.  This is consistent with the rest of the
>  API. However, there are some implemenations that copy only as many bytes of sockaddr from the
>  kernel to user-space as specified in the *addrlen attribute.  (1) This wasn't the behavior now
>  (2) we should not adopt this behavior because it is quite messy, e.g., copy only part of a port,
>  IP-address? Most partial retrivals do not make sense.  This is why I suggest returning EINVAL,
>  when the return struct is too small.
> ---
>  net/ipv4/af_inet.c  |    7 +++++--
>  net/ipv6/af_inet6.c |    3 +++
>  2 files changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
> index 743f554..9a8c5b0 100644
> --- a/net/ipv4/af_inet.c
> +++ b/net/ipv4/af_inet.c
> @@ -677,10 +677,13 @@ do_err:
>  int inet_getname(struct socket *sock, struct sockaddr *uaddr,
>  			int *uaddr_len, int peer)
>  {
> -	struct sock *sk		= sock->sk;
> -	struct inet_sock *inet	= inet_sk(sk);
> +	struct sock *sk = sock->sk;
> +	struct inet_sock *inet = inet_sk(sk);

Besides the other problem pointed out by Florian, why the gratuitous
reformatting above?

>  	struct sockaddr_in *sin	= (struct sockaddr_in *)uaddr;
>  
> +        if (*uaddr_len < sizeof(struct sockaddr_in))
> +	  return -ENOSPC;
> +
>  	sin->sin_family = AF_INET;
>  	if (peer) {
>  		if (!inet->dport ||
> diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
> index c802bc1..f2772ee 100644
> --- a/net/ipv6/af_inet6.c
> +++ b/net/ipv6/af_inet6.c
> @@ -401,6 +401,9 @@ int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
>  	struct inet_sock *inet = inet_sk(sk);
>  	struct ipv6_pinfo *np = inet6_sk(sk);
>  
> +	if (*uaddr_len < sizeof(struct sockaddr_in6))
> +		return -ENOSPC;
> +
>  	sin->sin6_family = AF_INET6;
>  	sin->sin6_flowinfo = 0;
>  	sin->sin6_scope_id = 0;
> -- 
> 1.5.4.5
> 
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 743f554..9a8c5b0 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -677,10 +677,13 @@  do_err:
 int inet_getname(struct socket *sock, struct sockaddr *uaddr,
 			int *uaddr_len, int peer)
 {
-	struct sock *sk		= sock->sk;
-	struct inet_sock *inet	= inet_sk(sk);
+	struct sock *sk = sock->sk;
+	struct inet_sock *inet = inet_sk(sk);
 	struct sockaddr_in *sin	= (struct sockaddr_in *)uaddr;
 
+        if (*uaddr_len < sizeof(struct sockaddr_in))
+	  return -ENOSPC;
+
 	sin->sin_family = AF_INET;
 	if (peer) {
 		if (!inet->dport ||
diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
index c802bc1..f2772ee 100644
--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -401,6 +401,9 @@  int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
 	struct inet_sock *inet = inet_sk(sk);
 	struct ipv6_pinfo *np = inet6_sk(sk);
 
+	if (*uaddr_len < sizeof(struct sockaddr_in6))
+		return -ENOSPC;
+
 	sin->sin6_family = AF_INET6;
 	sin->sin6_flowinfo = 0;
 	sin->sin6_scope_id = 0;