Change to remove MPTCP members from tcp_skb_cb
by Rao Shoaib
Hi,
Here is an attempt to not use tcp_skb_cb with minimal change to MPTCP
code and skb code. I have tested the change by using wget from
multipath-tcp.org. pcap file is attached.
The approach that I have taken is very straight forward. On Tx when an
skb is allocated and the underlying socket is MPTCP , 32 bytes extra are
allocated[1] and reserved as the head room. This space is used for
passing what was being passed in tcp_skb_cb. An used bit in sk_buff is
used to indicate this special case and when skb is copied the extra data
is also copied. clone is not an issue as the data is shared. This work
for retransmission also because tcp_transmit_skb clones the original skb.
On the Rx side, options are parsed in tcp at that point we can reuse the
space that was used for ether and IP headers and no extra space is
required.
The space used is always right below the skb head pointer because the
location of head never changes.
Please take a look and let me know what I have missed.
Regards,
Rao
2 years, 11 months
[PATCH] tcp: Register handlers for extra TCP options
by Mat Martineau
Allow additional TCP options to be handled by registered hook
functions.
Option parsing will call the provided 'parse' function when a TCP option
number is not recognized by the normal option parsing code.
The 'prepare' and 'prepare_req' functions determine the required space
for registered options and store associated data. There are two variants
because options must be generated for connections that are associated
with either tcp_sock or request_sock structures.
A static key and RCU synchronization are used to minimize the
performance impact of these extensible TCP features.
Before this is posted upstream, I plan to split the patch in two parts
(at least), since making tcp_out_options public should be separate. The
extra parameter for tcp_parse_options may also warrant its own
patch. I'd also like to discuss where to store information for the
options that are parsed or generated using these hooks, since the
tcp_options_received and tcp_out_options structs don't have places to
store extra information yet.
I've confirmed that this code builds, but haven't tested yet.
This commit is also available at
https://git.kernel.org/pub/scm/linux/kernel/git/martineau/linux.git
branch: mptcp
Signed-off-by: Mat Martineau <mathew.j.martineau(a)linux.intel.com>
---
drivers/infiniband/hw/cxgb4/cm.c | 2 +-
include/linux/tcp.h | 17 +++++
include/net/tcp.h | 48 +++++++++++++-
net/ipv4/syncookies.c | 2 +-
net/ipv4/tcp.c | 135 +++++++++++++++++++++++++++++++++++++++
net/ipv4/tcp_input.c | 15 +++--
net/ipv4/tcp_minisocks.c | 4 +-
net/ipv4/tcp_output.c | 35 +++++-----
net/ipv6/syncookies.c | 2 +-
9 files changed, 232 insertions(+), 28 deletions(-)
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index 03a1b0e64fc3..1d1dbfdf5a05 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -3748,7 +3748,7 @@ static void build_cpl_pass_accept_req(struct sk_buff *skb, int stid , u8 tos)
*/
memset(&tmp_opt, 0, sizeof(tmp_opt));
tcp_clear_options(&tmp_opt);
- tcp_parse_options(skb, &tmp_opt, 0, NULL);
+ tcp_parse_options(skb, &tmp_opt, 0, NULL, NULL);
req = (struct cpl_pass_accept_req *)__skb_push(skb, sizeof(*req));
memset(req, 0, sizeof(*req));
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index cfc2d9506ce8..9e625a68f5e1 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -112,6 +112,23 @@ static inline void tcp_clear_options(struct tcp_options_received *rx_opt)
rx_opt->wscale_ok = rx_opt->snd_wscale = 0;
}
+#define OPTION_SACK_ADVERTISE (1 << 0)
+#define OPTION_TS (1 << 1)
+#define OPTION_MD5 (1 << 2)
+#define OPTION_WSCALE (1 << 3)
+#define OPTION_FAST_OPEN_COOKIE (1 << 8)
+
+struct tcp_out_options {
+ u16 options; /* bit field of OPTION_* */
+ u16 mss; /* 0 to disable */
+ u8 ws; /* window scale, 0 to disable */
+ u8 num_sack_blocks; /* number of SACK blocks to include */
+ u8 hash_size; /* bytes in hash_location */
+ __u8 *hash_location; /* temporary pointer, overloaded */
+ __u32 tsval, tsecr; /* need to include OPTION_TS */
+ struct tcp_fastopen_cookie *fastopen_cookie; /* Fast open cookie */
+};
+
/* This is the max number of SACKS that we'll generate and process. It's safe
* to increase this, although since:
* size = TCPOLEN_SACK_BASE_ALIGNED (4) + n * TCPOLEN_SACK_PERBLOCK (8)
diff --git a/include/net/tcp.h b/include/net/tcp.h
index cc6ae0a95201..36785427ae90 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -429,7 +429,8 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
int flags, int *addr_len);
void tcp_parse_options(const struct sk_buff *skb,
struct tcp_options_received *opt_rx,
- int estab, struct tcp_fastopen_cookie *foc);
+ int estab, struct tcp_fastopen_cookie *foc,
+ struct tcp_sock *tp);
const u8 *tcp_parse_md5sig_option(const struct tcphdr *th);
/*
@@ -1937,4 +1938,49 @@ static inline void tcp_listendrop(const struct sock *sk)
__NET_INC_STATS(sock_net(sk), LINUX_MIB_LISTENDROPS);
}
+extern struct static_key_false tcp_extra_options_enabled;
+
+struct tcp_extra_option_ops {
+ struct list_head list;
+ unsigned char option_kind;
+ void (*parse)(int opsize, const unsigned char *opptr,
+ const struct sk_buff *skb,
+ struct tcp_options_received *opt_rx,
+ struct tcp_sock *tp);
+ /* Return the number of bytes consumed */
+ unsigned int (*prepare)(struct sk_buff *skb, u8 flags,
+ unsigned int remaining,
+ struct tcp_out_options *opts,
+ struct tcp_sock *tp);
+ /* Return the number of bytes consumed */
+ unsigned int (*prepare_req)(struct sk_buff *skb, u8 flags,
+ unsigned int remaining,
+ struct tcp_out_options *opts,
+ struct request_sock *req);
+ void (*write)(__be32 *ptr, struct tcp_out_options *opts,
+ struct tcp_sock *tp);
+ struct module *owner;
+};
+
+void tcp_extra_options_parse(int opcode, int opsize, const unsigned char *opptr,
+ const struct sk_buff *skb,
+ struct tcp_options_received *opt_rx,
+ struct tcp_sock *tp);
+
+unsigned int tcp_extra_options_prepare(struct sk_buff *skb, u8 flags,
+ unsigned int remaining,
+ struct tcp_out_options *opts,
+ struct tcp_sock *tp);
+
+unsigned int tcp_extra_options_prepare_req(struct sk_buff *skb, u8 flags,
+ unsigned int remaining,
+ struct tcp_out_options *opts,
+ struct request_sock *req);
+
+void tcp_extra_options_write(__be32 *ptr, struct tcp_out_options *opts,
+ struct tcp_sock *tp);
+
+int tcp_register_extra_option(struct tcp_extra_option_ops *ops);
+void tcp_unregister_extra_option(struct tcp_extra_option_ops *ops);
+
#endif /* _TCP_H */
diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index 496b97e17aaf..5697083a2c77 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -309,7 +309,7 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb)
/* check for timestamp cookie support */
memset(&tcp_opt, 0, sizeof(tcp_opt));
- tcp_parse_options(skb, &tcp_opt, 0, NULL);
+ tcp_parse_options(skb, &tcp_opt, 0, NULL, tp);
if (!cookie_timestamp_decode(&tcp_opt))
goto out;
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 04843ae77b9e..e0d0504f005a 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -269,6 +269,7 @@
#include <linux/err.h>
#include <linux/time.h>
#include <linux/slab.h>
+#include <linux/static_key.h>
#include <net/icmp.h>
#include <net/inet_common.h>
@@ -306,6 +307,13 @@ struct percpu_counter tcp_sockets_allocated;
EXPORT_SYMBOL(tcp_sockets_allocated);
/*
+ * Optional TCP option handlers
+ */
+static DEFINE_SPINLOCK(tcp_option_list_lock);
+static LIST_HEAD(tcp_option_list);
+DEFINE_STATIC_KEY_FALSE(tcp_extra_options_enabled);
+
+/*
* TCP splice context
*/
struct tcp_splice_state {
@@ -3279,6 +3287,132 @@ EXPORT_SYMBOL(tcp_md5_hash_key);
#endif
+/* Linear search, few entries are expected. The RCU read lock must
+ * be held before calling.
+ */
+static struct tcp_extra_option_ops *tcp_extra_options_find_kind(unsigned char kind)
+{
+ struct tcp_extra_option_ops *entry;
+
+ list_for_each_entry_rcu(entry, &tcp_option_list, list) {
+ if (entry->option_kind == kind)
+ return entry;
+ }
+
+ return NULL;
+}
+
+void tcp_extra_options_parse(int opcode, int opsize, const unsigned char *opptr,
+ const struct sk_buff *skb,
+ struct tcp_options_received *opt_rx,
+ struct tcp_sock *tp)
+{
+ struct tcp_extra_option_ops *entry;
+
+ rcu_read_lock();
+ entry = tcp_extra_options_find_kind(opcode);
+ if (entry && entry->parse)
+ entry->parse(opsize, opptr, skb, opt_rx, tp);
+ rcu_read_unlock();
+}
+
+unsigned int tcp_extra_options_prepare(struct sk_buff *skb, u8 flags,
+ unsigned int remaining,
+ struct tcp_out_options *opts,
+ struct tcp_sock *tp)
+{
+ struct tcp_extra_option_ops *entry;
+ unsigned int used = 0;
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(entry, &tcp_option_list, list) {
+ if (unlikely(!entry->prepare))
+ continue;
+
+ used += entry->prepare(skb, flags, remaining - used, opts, tp);
+ }
+ rcu_read_unlock();
+
+ return used;
+}
+
+unsigned int tcp_extra_options_prepare_req(struct sk_buff *skb, u8 flags,
+ unsigned int remaining,
+ struct tcp_out_options *opts,
+ struct request_sock *req)
+{
+ struct tcp_extra_option_ops *entry;
+ unsigned int used = 0;
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(entry, &tcp_option_list, list) {
+ if (unlikely(!entry->prepare_req))
+ continue;
+
+ used += entry->prepare_req(skb, flags, remaining - used, opts,
+ req);
+ }
+ rcu_read_unlock();
+
+ return used;
+}
+
+void tcp_extra_options_write(__be32 *ptr, struct tcp_out_options *opts,
+ struct tcp_sock *tp)
+{
+ struct tcp_extra_option_ops *entry;
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(entry, &tcp_option_list, list) {
+ if (unlikely(!entry->write))
+ continue;
+
+ entry->write(ptr, opts, tp);
+ }
+ rcu_read_unlock();
+}
+
+int tcp_register_extra_option(struct tcp_extra_option_ops *ops)
+{
+ int ret = 0;
+
+ if (!ops->option_kind)
+ return -EINVAL;
+
+ if (!try_module_get(ops->owner))
+ return -ENOENT;
+
+ spin_lock(&tcp_option_list_lock);
+ if (tcp_extra_options_find_kind(ops->option_kind)) {
+ pr_notice("Option kind %u already registered\n",
+ ops->option_kind);
+ spin_unlock(&tcp_option_list_lock);
+ module_put(ops->owner);
+ return -EEXIST;
+ }
+
+ list_add_tail_rcu(&ops->list, &tcp_option_list);
+ pr_debug("Option kind %u registered\n", ops->option_kind);
+ spin_unlock(&tcp_option_list_lock);
+
+ static_branch_inc(&tcp_extra_options_enabled);
+
+ return ret;
+}
+
+void tcp_unregister_extra_option(struct tcp_extra_option_ops *ops)
+{
+ spin_lock(&tcp_option_list_lock);
+ list_del_rcu(&ops->list);
+ spin_unlock(&tcp_option_list_lock);
+
+ synchronize_net();
+
+ static_branch_dec(&tcp_extra_options_enabled);
+
+ module_put(ops->owner);
+}
+
void tcp_done(struct sock *sk)
{
struct request_sock *req = tcp_sk(sk)->fastopen_rsk;
@@ -3425,6 +3559,7 @@ void __init tcp_init(void)
INIT_HLIST_HEAD(&tcp_hashinfo.bhash[i].chain);
}
+ INIT_LIST_HEAD(&tcp_option_list);
cnt = tcp_hashinfo.ehash_mask + 1;
sysctl_tcp_max_orphans = cnt / 2;
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index a5838858c362..13ab3dff7c3f 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -3726,7 +3726,7 @@ static void tcp_parse_fastopen_option(int len, const unsigned char *cookie,
*/
void tcp_parse_options(const struct sk_buff *skb,
struct tcp_options_received *opt_rx, int estab,
- struct tcp_fastopen_cookie *foc)
+ struct tcp_fastopen_cookie *foc, struct tcp_sock *tp)
{
const unsigned char *ptr;
const struct tcphdr *th = tcp_hdr(skb);
@@ -3828,6 +3828,11 @@ void tcp_parse_options(const struct sk_buff *skb,
ptr + 2, th->syn, foc, true);
break;
+ default:
+ tcp_extra_options_parse(opcode, opsize, ptr,
+ skb, opt_rx, tp);
+ break;
+
}
ptr += opsize-2;
length -= opsize;
@@ -3873,7 +3878,7 @@ static bool tcp_fast_parse_options(const struct sk_buff *skb,
return true;
}
- tcp_parse_options(skb, &tp->rx_opt, 1, NULL);
+ tcp_parse_options(skb, &tp->rx_opt, 1, NULL, tp);
if (tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr)
tp->rx_opt.rcv_tsecr -= tp->tsoffset;
@@ -5600,7 +5605,7 @@ static bool tcp_rcv_fastopen_synack(struct sock *sk, struct sk_buff *synack,
/* Get original SYNACK MSS value if user MSS sets mss_clamp */
tcp_clear_options(&opt);
opt.user_mss = opt.mss_clamp = 0;
- tcp_parse_options(synack, &opt, 0, NULL);
+ tcp_parse_options(synack, &opt, 0, NULL, tp);
mss = opt.mss_clamp;
}
@@ -5653,7 +5658,7 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
struct tcp_fastopen_cookie foc = { .len = -1 };
int saved_clamp = tp->rx_opt.mss_clamp;
- tcp_parse_options(skb, &tp->rx_opt, 0, &foc);
+ tcp_parse_options(skb, &tp->rx_opt, 0, &foc, tp);
if (tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr)
tp->rx_opt.rcv_tsecr -= tp->tsoffset;
@@ -6316,7 +6321,7 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
tcp_clear_options(&tmp_opt);
tmp_opt.mss_clamp = af_ops->mss_clamp;
tmp_opt.user_mss = tp->rx_opt.user_mss;
- tcp_parse_options(skb, &tmp_opt, 0, want_cookie ? NULL : &foc);
+ tcp_parse_options(skb, &tmp_opt, 0, want_cookie ? NULL : &foc, tp);
if (want_cookie && !tmp_opt.saw_tstamp)
tcp_clear_options(&tmp_opt);
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 8f6373b0cd77..c51bb1b30f08 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -98,7 +98,7 @@ tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
tmp_opt.saw_tstamp = 0;
if (th->doff > (sizeof(*th) >> 2) && tcptw->tw_ts_recent_stamp) {
- tcp_parse_options(skb, &tmp_opt, 0, NULL);
+ tcp_parse_options(skb, &tmp_opt, 0, NULL, NULL);
if (tmp_opt.saw_tstamp) {
if (tmp_opt.rcv_tsecr)
@@ -558,7 +558,7 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
tmp_opt.saw_tstamp = 0;
if (th->doff > (sizeof(struct tcphdr)>>2)) {
- tcp_parse_options(skb, &tmp_opt, 0, NULL);
+ tcp_parse_options(skb, &tmp_opt, 0, NULL, NULL);
if (tmp_opt.saw_tstamp) {
tmp_opt.ts_recent = req->ts_recent;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index ffc9274b2706..840883531996 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -41,6 +41,7 @@
#include <linux/compiler.h>
#include <linux/gfp.h>
#include <linux/module.h>
+#include <linux/static_key.h>
/* People can turn this off for buggy TCP's found in printers etc. */
int sysctl_tcp_retrans_collapse __read_mostly = 1;
@@ -411,23 +412,6 @@ static inline bool tcp_urg_mode(const struct tcp_sock *tp)
return tp->snd_una != tp->snd_up;
}
-#define OPTION_SACK_ADVERTISE (1 << 0)
-#define OPTION_TS (1 << 1)
-#define OPTION_MD5 (1 << 2)
-#define OPTION_WSCALE (1 << 3)
-#define OPTION_FAST_OPEN_COOKIE (1 << 8)
-
-struct tcp_out_options {
- u16 options; /* bit field of OPTION_* */
- u16 mss; /* 0 to disable */
- u8 ws; /* window scale, 0 to disable */
- u8 num_sack_blocks; /* number of SACK blocks to include */
- u8 hash_size; /* bytes in hash_location */
- __u8 *hash_location; /* temporary pointer, overloaded */
- __u32 tsval, tsecr; /* need to include OPTION_TS */
- struct tcp_fastopen_cookie *fastopen_cookie; /* Fast open cookie */
-};
-
/* Write previously computed TCP options to the packet.
*
* Beware: Something in the Internet is very sensitive to the ordering of
@@ -534,6 +518,9 @@ static void tcp_options_write(__be32 *ptr, struct tcp_sock *tp,
}
ptr += (len + 3) >> 2;
}
+
+ if (static_branch_unlikely(&tcp_extra_options_enabled))
+ tcp_extra_options_write(ptr, opts, tp);
}
/* Compute TCP options for SYN packets. This is not the final
@@ -601,6 +588,10 @@ static unsigned int tcp_syn_options(struct sock *sk, struct sk_buff *skb,
}
}
+ if (static_branch_unlikely(&tcp_extra_options_enabled))
+ remaining -= tcp_extra_options_prepare(skb, TCPHDR_SYN,
+ remaining, opts, tp);
+
return MAX_TCP_OPTION_SPACE - remaining;
}
@@ -661,6 +652,13 @@ static unsigned int tcp_synack_options(struct request_sock *req,
}
}
+ if (static_branch_unlikely(&tcp_extra_options_enabled))
+ remaining -= tcp_extra_options_prepare_req(skb,
+ TCPHDR_SYN |
+ TCPHDR_ACK,
+ remaining, opts,
+ req);
+
return MAX_TCP_OPTION_SPACE - remaining;
}
@@ -694,6 +692,9 @@ static unsigned int tcp_established_options(struct sock *sk, struct sk_buff *skb
size += TCPOLEN_TSTAMP_ALIGNED;
}
+ size += tcp_extra_options_prepare(skb, 0, MAX_TCP_OPTION_SPACE - size,
+ opts, tp);
+
eff_sacks = tp->rx_opt.num_sacks + tp->rx_opt.dsack;
if (unlikely(eff_sacks)) {
const unsigned int remaining = MAX_TCP_OPTION_SPACE - size;
diff --git a/net/ipv6/syncookies.c b/net/ipv6/syncookies.c
index 895ff650db43..1c246a8f41c3 100644
--- a/net/ipv6/syncookies.c
+++ b/net/ipv6/syncookies.c
@@ -160,7 +160,7 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
/* check for timestamp cookie support */
memset(&tcp_opt, 0, sizeof(tcp_opt));
- tcp_parse_options(skb, &tcp_opt, 0, NULL);
+ tcp_parse_options(skb, &tcp_opt, 0, NULL, tp);
if (!cookie_timestamp_decode(&tcp_opt))
goto out;
--
2.12.2
2 years, 11 months
Re: [MPTCP] MPTCP upstreaming introduction
by Mat Martineau
Hi Christoph,
Rao and I talked with David Miller and Eric Dumazet. They aren't
enthusiastic about MPTCP from a protocol perpective, and pointed to the
example of kcm as something that builds on top of multiple TCP connections
rather than modifying TCP itself. I think they're a little skeptical that
there's a committment to put in the necessary effort to upstream.
A few specific points of our discussions:
* Intrusive changes to TCP are not acceptable, the code is already complex
and they don't want to introduce bugs or create additional maintenance
burden. Given the need to tie in to TCP for option parsing (and other
changed behaviors), they're willing to look at adding hooks - but of
course reserved judgment until there are actual patches to look at.
* Related to that, skb size cannot increase.
* Regular TCP connections should remain the default and should work like
they do now.
* Expect the review process to take a while.
These points are what drives the goals I listed below. Identifying the
right (limited) set of hooks in to the TCP code is going to be an
important early step.
Best regards,
Mat
On Mon, 10 Apr 2017, Christoph Paasch wrote:
> Hi Mat,
>
> what was the feedback/info you got from netdev 2.1 ?
>
>
> Christoph
>
> On 09/04/17 - 13:12:26, Mat Martineau wrote:
>> Welcome to the MPTCP upstreaming mailing list!
>>
>> This list (mptcp(a)lists.01.org) is focused on integrating MPTCP (RFC 6824 /
>> RFC 6824bis) as part of the standard Linux kernel. In speaking with the
>> Linux networking / TCP maintainers, it is important that MPTCP does not
>> interfere with the performance or reliability of the TCP stack, or
>> complicate the existing TCP code.
>>
>> Given the Linux netdev maintainer's needs, we have the following high-level
>> design goals in mind:
>>
>> * Minimize changes to core TCP.
>>
>> * Favor atomic operations and RCU synchronization over more expensive
>> options like mutexes or spinlocks.
>>
>> * Implement features in userspace where latency isn't important and data
>> flow is not affected
>>
>> * TCP operates as "normal TCP" by default. Individual sockets may opt-in, or
>> systemwide configuration can be used to broadly enable MPTCP.
>>
>> The role of this list is to coordinate upstreaming efforts and share code
>> prior to upstream submission on netdev(a)lists.01.org . This list is separate
>> from the mptcp-dev list associated with the multipath-tcp.org kernel in
>> order to avoid confusion between netdev-focused efforts and work related to
>> multipath-tcp.org . Anyone interested in MPTCP for the upstream kernel is
>> welcome on mptcp(a)lists.01.org, we appreciate your expertise and
>> contributions!
>>
>> You can manage your subscription or view list archives at
>> https://lists.01.org/mailman/listinfo/mptcp
>>
>> Best regards,
>>
>> --
>> Mat Martineau
>> Intel OTC
>
--
Mat Martineau
Intel OTC
2 years, 11 months
MPTCP Linux upstreaming introduction
by Mat Martineau
Welcome to the MPTCP upstreaming mailing list!
This list (mptcp(a)lists.01.org) is focused on integrating MPTCP (RFC 6824 /
RFC 6824bis) as part of the standard
Linux kernel. In speaking with the Linux networking / TCP maintainers, it
is important that MPTCP does not
interfere with the performance or reliability of the TCP stack, or
complicate the existing TCP code.
Given the Linux netdev maintainer's needs, we have the following
high-level design goals in mind:
* Minimize changes to core TCP.
* Favor atomic operations and RCU synchronization over more expensive
options like mutexes or spinlocks.
* Implement features in userspace where latency isn't important and data
flow is not affected
* TCP operates as "normal TCP" by default. Individual sockets may opt-in,
or systemwide configuration can be
used to broadly enable MPTCP.
The role of this list is to coordinate upstreaming efforts and share code
prior to upstream submission on
netdev(a)lists.01.org . This list is separate from the mptcp-dev list
associated with the multipath-tcp.org
kernel in order to avoid confusion between netdev-focused efforts and work
related to multipath-tcp.org .
Anyone interested in MPTCP for the upstream kernel is welcome on
mptcp(a)lists.01.org, we appreciate your
expertise and contributions!
You can manage your subscription or view list archives at
https://lists.01.org/mailman/listinfo/mptcp
Best regards,
--
Mat Martineau
Intel OTC
3 years