2006-08-23 10:53:05

by Gerrit Renker

[permalink] [raw]
Subject: [RFC][PATCH 1/3] net/ipv4: UDP-Lite extensions

[Net/IPv4]: UDP-Lite standalone support and shared UDP/-Lite socket structure.


Signed-off-by: Gerrit Renker <[email protected]>
---

include/linux/udp.h | 19 ++++
include/net/udplite.h | 35 ++++++++
net/ipv4/udplite.c | 209 ++++++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 262 insertions(+), 1 deletion(-)


diff --git a/include/linux/udp.h b/include/linux/udp.h
index 90223f0..76fb1a1 100644
--- a/include/linux/udp.h
+++ b/include/linux/udp.h
@@ -19,6 +19,12 @@ #define _LINUX_UDP_H

#include <linux/types.h>

+/**
+ * struct udphdr - UDP/-Lite header
+ *
+ * UDP (RFC 768) and UDP-Lite (RFC 3828) share the same header structure,
+ * the only difference is that UDP-Lite interprets `len' as checksum coverage.
+ */
struct udphdr {
__u16 source;
__u16 dest;
@@ -50,13 +56,24 @@ struct udp_sock {
* when the socket is uncorked.
*/
__u16 len; /* total length of pending frames */
+/*
+ * Fields specific to UDP-Lite.
+ */
+ __u16 pcslen;
+ __u16 pcrlen;
+/* indicator bits used by pcflag: */
+#define UDPLITE_BIT 0x1 /* set by udplite proto init function */
+#define UDPLITE_SEND_CC 0x2 /* set via udplite setsockopt */
+#define UDPLITE_RECV_CC 0x4 /* set via udplite setsocktopt */
+ __u8 pcflag; /* marks socket as UDP-Lite if > 0 */
};

static inline struct udp_sock *udp_sk(const struct sock *sk)
{
return (struct udp_sock *)sk;
}
+#define IS_UDPLITE(__sk) (udp_sk(__sk)->pcflag)

-#endif
+#endif /* __KERNEL__ */

#endif /* _LINUX_UDP_H */
diff --git a/net/ipv4/udplite.c b/net/ipv4/udplite.c
new file mode 100644
index 0000000..f5597e9
--- /dev/null
+++ b/net/ipv4/udplite.c
@@ -0,0 +1,209 @@
+/*
+ * UDPLITE An implementation of the UDP-Lite protocol (RFC 3828).
+ *
+ * Version: $Id: udplite.c,v 1.22 2006/08/22 13:01:52 gerrit Exp gerrit $
+ *
+ * Authors: Gerrit Renker <[email protected]>
+ *
+ * Changes:
+ * Fixes:
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+struct hlist_head udplite_hash[UDP_HTABLE_SIZE];
+int udplite_port_rover;
+DEFINE_SNMP_STAT(struct udp_mib, udplite_statistics) __read_mostly;
+
+/* these functions are called by UDP-Lite with protocol-specific parameters */
+static int __udp_get_port(struct sock *, unsigned short,
+ struct hlist_head *, int * );
+static struct sock *__udp_lookup(u32 , u16, u32, u16, int, struct hlist_head *);
+static int __udp_mcast_deliver(struct sk_buff *, struct udphdr *,
+ u32, u32, struct hlist_head * );
+static int __udp_common_rcv(struct sk_buff *, int is_udplite);
+static void __udp_err(struct sk_buff *, u32, struct hlist_head *);
+#ifdef CONFIG_PROC_FS
+static int udp4_seq_show(struct seq_file *, void *);
+#endif
+
+/*
+ * Designate sk as UDP-Lite socket
+ */
+static inline int udplite_sk_init(struct sock *sk)
+{
+ udp_sk(sk)->pcflag = UDPLITE_BIT;
+ return 0;
+}
+
+static __inline__ int udplite_v4_get_port(struct sock *sk, unsigned short snum)
+{
+ return __udp_get_port(sk, snum, udplite_hash, &udplite_port_rover);
+}
+
+static __inline__ struct sock *udplite_v4_lookup(u32 saddr, u16 sport,
+ u32 daddr, u16 dport, int dif)
+{
+ return __udp_lookup(saddr, sport, daddr, dport, dif, udplite_hash);
+}
+
+static __inline__ int udplite_v4_mcast_deliver(struct sk_buff *skb,
+ struct udphdr *uh, u32 saddr, u32 daddr)
+{
+ return __udp_mcast_deliver(skb, uh, saddr, daddr, udplite_hash);
+}
+
+__inline__ int udplite_rcv(struct sk_buff *skb)
+{
+ return __udp_common_rcv(skb, 1);
+}
+
+__inline__ void udplite_err(struct sk_buff *skb, u32 info)
+{
+ return __udp_err(skb, info, udplite_hash);
+}
+
+static int udplite_checksum_init(struct sk_buff *skb, struct udphdr *uh,
+ unsigned short len, u32 saddr, u32 daddr)
+{
+ u16 cscov;
+
+ /* In UDPv4 a zero checksum means that the transmitter generated no
+ * checksum. UDP-Lite (like IPv6) mandates checksums, hence packets
+ * with a zero checksum field are illegal. */
+ if (uh->check == 0) {
+ LIMIT_NETDEBUG(KERN_DEBUG "UDPLITE: zeroed csum field"
+ "(%d.%d.%d.%d:%d -> %d.%d.%d.%d:%d)\n", NIPQUAD(saddr),
+ ntohs(uh->source), NIPQUAD(daddr), ntohs(uh->dest) );
+ return 0;
+ }
+
+ UDP_SKB_CB(skb)->partial_cov = 0;
+ cscov = ntohs(uh->len);
+
+ if (cscov == 0) /* Indicates that full coverage is required. */
+ cscov = len;
+ else if (cscov < 8 || cscov > len) {
+ /*
+ * Coverage length violates RFC 3828: log and discard silently.
+ */
+ LIMIT_NETDEBUG(KERN_DEBUG "UDPLITE: bad csum coverage %d/%d "
+ "(%d.%d.%d.%d:%d -> %d.%d.%d.%d:%d)\n", cscov, len,
+ NIPQUAD(saddr), ntohs(uh->source),
+ NIPQUAD(daddr), ntohs(uh->dest) );
+ return 0;
+
+ } else if (cscov < len)
+ UDP_SKB_CB(skb)->partial_cov = 1;
+
+ UDP_SKB_CB(skb)->cscov = cscov;
+
+ /*
+ * Initialise pseudo-header for checksum computation.
+ *
+ * There is no known NIC manufacturer supporting UDP-Lite yet,
+ * hence ip_summed is always (re-)set to CHECKSUM_NONE.
+ */
+ skb->csum = csum_tcpudp_nofold(saddr, daddr, len, IPPROTO_UDPLITE, 0);
+ skb->ip_summed = CHECKSUM_NONE;
+
+ return 1;
+}
+
+static void udplite_csum_outgoing(struct sock *sk, struct sk_buff *skb,
+ int totlen, int cscov, u32 src, u32 dst)
+{
+ unsigned int csum = 0, len;
+ struct udphdr *uh = skb->h.uh;
+
+ uh->check = 0;
+
+ skb->ip_summed = CHECKSUM_NONE; /* no HW support for checksumming */
+
+ if (skb_queue_len(&sk->sk_write_queue) == 1) {
+ /*
+ * Only one fragment on the socket.
+ */
+ csum = skb_checksum(skb, skb->h.raw - skb->data, cscov, 0);
+
+ } else {
+ skb_queue_walk(&sk->sk_write_queue, skb) {
+ len = skb->tail - skb->h.raw;
+
+ skb->csum = skb_checksum(skb, skb->h.raw - skb->data,
+ (cscov > len)? len : cscov, 0);
+ csum = csum_add(csum, skb->csum);
+
+ if (cscov < len) /* Enough seen. */
+ break;
+ cscov -= len;
+ }
+ }
+
+ uh->check = csum_tcpudp_magic(src, dst, totlen, IPPROTO_UDPLITE, csum);
+ if (uh->check == 0)
+ uh->check = -1;
+}
+
+
+static struct net_protocol udplite_protocol = {
+ .handler = udplite_rcv,
+ .err_handler = udplite_err,
+ .no_policy = 1,
+};
+
+static struct inet_protosw udplite4_protosw = {
+ .type = SOCK_DGRAM,
+ .protocol = IPPROTO_UDPLITE,
+ .prot = &udplite_prot,
+ .ops = &inet_dgram_ops,
+ .capability = -1,
+ .no_check = 0, /* must checksum (RFC 3828) */
+ .flags = INET_PROTOSW_PERMANENT,
+};
+
+void __init udplite4_register(void)
+{
+ if (proto_register(&udplite_prot, 1))
+ goto out_register_err;
+
+ if (inet_add_protocol(&udplite_protocol, IPPROTO_UDPLITE) < 0)
+ goto out_unregister_proto;
+
+ inet_register_protosw(&udplite4_protosw);
+
+ return;
+
+out_unregister_proto:
+ proto_unregister(&udplite_prot);
+out_register_err:
+ printk(KERN_ERR "udplite4_register: Cannot add UDP-Lite protocol\n");
+}
+
+#ifdef CONFIG_PROC_FS
+static struct file_operations udplite4_seq_fops;
+static struct udp_seq_afinfo udplite4_seq_afinfo = {
+ .owner = THIS_MODULE,
+ .name = "udplite",
+ .family = AF_INET,
+ .hashtable = udplite_hash,
+ .seq_show = udp4_seq_show,
+ .seq_fops = &udplite4_seq_fops,
+};
+
+__inline__ int __init udplite4_proc_init(void)
+{
+ return udp_proc_register(&udplite4_seq_afinfo);
+}
+
+__inline__ void udplite4_proc_exit(void)
+{
+ udp_proc_unregister(&udplite4_seq_afinfo);
+}
+#endif /* CONFIG_PROC_FS */
+
+EXPORT_SYMBOL(udplite_hash);
+EXPORT_SYMBOL(udplite_prot);
diff --git a/include/net/udplite.h b/include/net/udplite.h
new file mode 100644
index 0000000..1a32c42
--- /dev/null
+++ b/include/net/udplite.h
@@ -0,0 +1,35 @@
+/*
+ * Definitions for the UDP-Lite (RFC 3828) code.
+ */
+#ifndef _UDPLITE_H
+#define _UDPLITE_H
+
+/* UDP-Lite socket options */
+#define UDPLITE_SEND_CSCOV 10 /* sender partial coverage (as sent) */
+#define UDPLITE_RECV_CSCOV 11 /* receiver partial coverage (threshold ) */
+
+extern struct proto udplite_prot;
+extern struct hlist_head udplite_hash[UDP_HTABLE_SIZE];
+
+/* UDP-Lite does not have a standardized MIB yet, so we inherit from UDP */
+DECLARE_SNMP_STAT(struct udp_mib, udplite_statistics);
+
+/*
+ * Checksum computation is all in software, hence simpler getfrag.
+ */
+static __inline__ int udplite_getfrag(void *from, char *to, int offset,
+ int len, int odd, struct sk_buff *skb)
+{
+ return memcpy_fromiovecend(to, (struct iovec *) from, offset, len);
+}
+
+/*
+ * net/ipv4/udplite.c
+ */
+extern void udplite4_register(void);
+#ifdef CONFIG_PROC_FS
+extern int udplite4_proc_init(void);
+extern void udplite4_proc_exit(void);
+#endif
+
+#endif /* _UDPLITE_H */


2006-08-23 14:22:28

by James Morris

[permalink] [raw]
Subject: Re: [RFC][PATCH 1/3] net/ipv4: UDP-Lite extensions

On Wed, 23 Aug 2006, [email protected] wrote:

> +void __init udplite4_register(void)
> +{
> + if (proto_register(&udplite_prot, 1))
> + goto out_register_err;
> +
> + if (inet_add_protocol(&udplite_protocol, IPPROTO_UDPLITE) < 0)
> + goto out_unregister_proto;
> +
> + inet_register_protosw(&udplite4_protosw);
> +
> + return;
> +
> +out_unregister_proto:
> + proto_unregister(&udplite_prot);
> +out_register_err:
> + printk(KERN_ERR "udplite4_register: Cannot add UDP-Lite protocol\n");
> +}

Other protocols & network components call panic() if they fail during boot
initialization. Not sure if this is a great thing, but it raises the
issue of whether udp-lite should remain consistent here.



- James
--
James Morris
<[email protected]>

2006-08-23 15:04:15

by Gerrit Renker

[permalink] [raw]
Subject: Re: [RFC][PATCH 1/3] net/ipv4: UDP-Lite extensions

Hi James,

Quoting James Morris:
| On Wed, 23 Aug 2006, [email protected] wrote:
|
| > +void __init udplite4_register(void)
| > +{
| > + if (proto_register(&udplite_prot, 1))
| > + goto out_register_err;
| > +
| > + if (inet_add_protocol(&udplite_protocol, IPPROTO_UDPLITE) < 0)
| > + goto out_unregister_proto;
| > +
| > + inet_register_protosw(&udplite4_protosw);
| > +
| > + return;
| > +
| > +out_unregister_proto:
| > + proto_unregister(&udplite_prot);
| > +out_register_err:
| > + printk(KERN_ERR "udplite4_register: Cannot add UDP-Lite protocol\n");
| > +}
|
| Other protocols & network components call panic() if they fail during boot
| initialization. Not sure if this is a great thing, but it raises the
| issue of whether udp-lite should remain consistent here.

The behaviour is consistent (modulo loglevel) with inet_init()
of net/ipv4/af_inet.c:

// ...

rc = proto_register(&udp_prot, 1);
if (rc)
goto out_unregister_tcp_proto;

// ...
if (inet_add_protocol(&udp_protocol, IPPROTO_UDP) < 0)
printk(KERN_CRIT "inet_init: Cannot add UDP protocol\n");
if (inet_add_protocol(&tcp_protocol, IPPROTO_TCP) < 0)
printk(KERN_CRIT "inet_init: Cannot add TCP protocol\n");

// ...
for (q = inetsw_array; q < &inetsw_array[INETSW_ARRAY_LEN]; ++q)
inet_register_protosw(q);

// ...

/* UDP Lite is called here, but the code of udplite4_register()
* could as well be put in place within inet_init() */

udplite4_register();


But your question raises another:
* if TCP cannot init, there is no layer-4 protocol
* if UDP cannot register, TCP is unregistered also
* if RAW cannot register, only UDP is unregistered, but not TCP

>From that I could not deduct a rule what would happen if UDP-Lite failed
to register. If control had reached that above point, it means that all
other protocols have already successfully registered -- if then UDP-Lite
could not register and called a panic(), it would abort the remainder of the
stack.

So how should the code be integrated:
(a) following the same scheme as in af_inet.c (like f.e. raw_prot ) or
(b) keep separate initialisation (as it is used e.g. in net/dccp/ipv4, dccp_v4_init())
(c) ... ?


Gerrit

2006-08-23 15:22:10

by James Morris

[permalink] [raw]
Subject: Re: [RFC][PATCH 1/3] net/ipv4: UDP-Lite extensions

On Wed, 23 Aug 2006, [email protected] wrote:

> | Other protocols & network components call panic() if they fail during boot
> | initialization. Not sure if this is a great thing, but it raises the
> | issue of whether udp-lite should remain consistent here.
>
> The behaviour is consistent (modulo loglevel) with inet_init()
> of net/ipv4/af_inet.c:

Some things will panic there, just deeper in the call chain.

> >From that I could not deduct a rule what would happen if UDP-Lite failed
> to register. If control had reached that above point, it means that all
> other protocols have already successfully registered -- if then UDP-Lite
> could not register and called a panic(), it would abort the remainder of the
> stack.

Other functions can also panic on failure after this, e.g. tcp_init().

I think ideally it'd be best if components did not panic during
initialization unless it _really_ meant that the kernel should not
continue executing. Although, it's not entirely clear how to determine
this, e.g. perhaps the system should panic if netfilter initialization
failed, as it might mean that the systems comes up without a firewall. But
how do we know precisely which components are being used for security
critical purposes?

It seems like a signifcant overhaul of existing code, so probably best
just to leave yours as-is (which I suspect is the correct behavior
anyway).



- James
--
James Morris
<[email protected]>

2006-08-23 19:45:21

by James Morris

[permalink] [raw]
Subject: Re: [RFC][PATCH 1/3] net/ipv4: UDP-Lite extensions

On Wed, 23 Aug 2006, James Morris wrote:

> continue executing. Although, it's not entirely clear how to determine
> this, e.g. perhaps the system should panic if netfilter initialization
> failed, as it might mean that the systems comes up without a firewall.

Actually, I guess it's up to userspace to figure this out.



- James
--
James Morris
<[email protected]>