diff mbox series

[net-next,v3] mptcp: fix length of MP_PRIO suboption

Message ID 07135f8a644672e1847788341e39f6dd25b1c32c.1611852408.git.dcaratti@redhat.com
State Accepted, archived
Commit 91384be37ad9b9408c9321a76eddba6439cac9da
Delegated to: Matthieu Baerts
Headers show
Series [net-next,v3] mptcp: fix length of MP_PRIO suboption | expand

Commit Message

Davide Caratti Jan. 28, 2021, 5:34 p.m. UTC
With version 0 of the protocol it was legal to encode the 'Subflow Id' in
the MP_PRIO suboption, to specify which subflow would change its 'Backup'
flag. This has been removed from v1 specification: thus, according to RFC
8684 §3.3.8, the resulting 'Length' for MP_PRIO changed from 4 to 3 byte.

Current Linux generates / parses MP_PRIO according to the old spec, using
'Length' equal to 4, and hardcoding 1 as 'Subflow Id'; RFC compliance can
improve if we change 'Length' in other to become 3, leaving a 'Nop' after
the MP_PRIO suboption. In this way the kernel will emit and accept *only*
MP_PRIO suboptions that are compliant to version 1 of the MPTCP protocol.

 unpatched 5.11-rc kernel:
 [root@bottarga ~]# tcpdump -tnnr unpatched.pcap | grep prio
 reading from file unpatched.pcap, link-type LINUX_SLL (Linux cooked v1)
 dropped privs to tcpdump
 IP 10.0.3.2.48433 > 10.0.1.1.10006: Flags [.], ack 1, win 502, options [nop,nop,TS val 4032325513 ecr 1876514270,mptcp prio non-backup id 1,mptcp dss ack 14084896651682217737], length 0

 patched 5.11-rc kernel:
 [root@bottarga ~]# tcpdump -tnnr patched.pcap | grep prio
 reading from file patched.pcap, link-type LINUX_SLL (Linux cooked v1)
 dropped privs to tcpdump
 IP 10.0.3.2.49735 > 10.0.1.1.10006: Flags [.], ack 1, win 502, options [nop,nop,TS val 1276737699 ecr 2686399734,mptcp prio non-backup,nop,mptcp dss ack 18433038869082491686], length 0

Changes since v2:
 - when accounting for option space, don't increment 'TCPOLEN_MPTCP_PRIO'
   and use 'TCPOLEN_MPTCP_PRIO_ALIGN' instead, thanks to Matthieu Baerts.
Changes since v1:
 - refactor patch to avoid using 'TCPOLEN_MPTCP_PRIO' with its old value,
   thanks to Geliang Tang.

Fixes: 067065422fcd ("mptcp: add the outgoing MP_PRIO support")
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
---
 net/mptcp/options.c  | 5 +++--
 net/mptcp/protocol.h | 3 ++-
 2 files changed, 5 insertions(+), 3 deletions(-)

Comments

Mat Martineau Jan. 28, 2021, 11:37 p.m. UTC | #1
On Thu, 28 Jan 2021, Davide Caratti wrote:

> With version 0 of the protocol it was legal to encode the 'Subflow Id' in
> the MP_PRIO suboption, to specify which subflow would change its 'Backup'
> flag. This has been removed from v1 specification: thus, according to RFC
> 8684 §3.3.8, the resulting 'Length' for MP_PRIO changed from 4 to 3 byte.
>
> Current Linux generates / parses MP_PRIO according to the old spec, using
> 'Length' equal to 4, and hardcoding 1 as 'Subflow Id'; RFC compliance can
> improve if we change 'Length' in other to become 3, leaving a 'Nop' after
> the MP_PRIO suboption. In this way the kernel will emit and accept *only*
> MP_PRIO suboptions that are compliant to version 1 of the MPTCP protocol.
>
> unpatched 5.11-rc kernel:
> [root@bottarga ~]# tcpdump -tnnr unpatched.pcap | grep prio
> reading from file unpatched.pcap, link-type LINUX_SLL (Linux cooked v1)
> dropped privs to tcpdump
> IP 10.0.3.2.48433 > 10.0.1.1.10006: Flags [.], ack 1, win 502, options [nop,nop,TS val 4032325513 ecr 1876514270,mptcp prio non-backup id 1,mptcp dss ack 14084896651682217737], length 0
>
> patched 5.11-rc kernel:
> [root@bottarga ~]# tcpdump -tnnr patched.pcap | grep prio
> reading from file patched.pcap, link-type LINUX_SLL (Linux cooked v1)
> dropped privs to tcpdump
> IP 10.0.3.2.49735 > 10.0.1.1.10006: Flags [.], ack 1, win 502, options [nop,nop,TS val 1276737699 ecr 2686399734,mptcp prio non-backup,nop,mptcp dss ack 18433038869082491686], length 0
>
> Changes since v2:
> - when accounting for option space, don't increment 'TCPOLEN_MPTCP_PRIO'
>   and use 'TCPOLEN_MPTCP_PRIO_ALIGN' instead, thanks to Matthieu Baerts.
> Changes since v1:
> - refactor patch to avoid using 'TCPOLEN_MPTCP_PRIO' with its old value,
>   thanks to Geliang Tang.
>
> Fixes: 067065422fcd ("mptcp: add the outgoing MP_PRIO support")
> Signed-off-by: Davide Caratti <dcaratti@redhat.com>
> ---
> net/mptcp/options.c  | 5 +++--
> net/mptcp/protocol.h | 3 ++-
> 2 files changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/net/mptcp/options.c b/net/mptcp/options.c
> index 775f0576592e..bb874c5d663a 100644
> --- a/net/mptcp/options.c
> +++ b/net/mptcp/options.c
> @@ -699,10 +699,11 @@ static bool mptcp_established_options_mp_prio(struct sock *sk,
> 	if (!subflow->send_mp_prio)
> 		return false;
>
> -	if (remaining < TCPOLEN_MPTCP_PRIO)
> +	/* account for the trailing 'nop' option */
> +	if (remaining < TCPOLEN_MPTCP_PRIO_ALIGN)
> 		return false;
>
> -	*size = TCPOLEN_MPTCP_PRIO;
> +	*size = TCPOLEN_MPTCP_PRIO_ALIGN;
> 	opts->suboptions |= OPTION_MPTCP_PRIO;
> 	opts->backup = subflow->request_bkup;
>
> diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
> index 447ce4631b43..d31edbae8da8 100644
> --- a/net/mptcp/protocol.h
> +++ b/net/mptcp/protocol.h
> @@ -61,7 +61,8 @@
> #define TCPOLEN_MPTCP_ADD_ADDR6_BASE_PORT	24
> #define TCPOLEN_MPTCP_PORT_LEN		4
> #define TCPOLEN_MPTCP_RM_ADDR_BASE	4
> -#define TCPOLEN_MPTCP_PRIO		4
> +#define TCPOLEN_MPTCP_PRIO		3
> +#define TCPOLEN_MPTCP_PRIO_ALIGN	4
> #define TCPOLEN_MPTCP_FASTCLOSE		12
>
> /* MPTCP MP_JOIN flags */
> -- 
> 2.29.2

Looks good to me, thanks Davide.

Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>

--
Mat Martineau
Intel
Matthieu Baerts Jan. 29, 2021, 5:08 p.m. UTC | #2
Hi Davide, Mat,

On 28/01/2021 18:34, Davide Caratti wrote:
> With version 0 of the protocol it was legal to encode the 'Subflow Id' in
> the MP_PRIO suboption, to specify which subflow would change its 'Backup'
> flag. This has been removed from v1 specification: thus, according to RFC
> 8684 §3.3.8, the resulting 'Length' for MP_PRIO changed from 4 to 3 byte.
> 
> Current Linux generates / parses MP_PRIO according to the old spec, using
> 'Length' equal to 4, and hardcoding 1 as 'Subflow Id'; RFC compliance can
> improve if we change 'Length' in other to become 3, leaving a 'Nop' after
> the MP_PRIO suboption. In this way the kernel will emit and accept *only*
> MP_PRIO suboptions that are compliant to version 1 of the MPTCP protocol.

Thank you for this new version!

Just added in our tree, at the bottom, with Mat and my RvB tag.

- 91384be37ad9 mptcp: fix length of MP_PRIO suboption

Tests + export have been almost a success! Simult Flow selftest with a 
debug kernel was unstable but it is not new, see:

   https://github.com/multipath-tcp/mptcp_net-next/issues/137

So feel free to send this patch to netdev when you have the opportunity  ;-)
It doesn't conflict with other pending patches from Geliang that have 
been recently sent by Mat.

Cheers,
Matt
Davide Caratti Feb. 1, 2021, 10:23 a.m. UTC | #3
On Fri, 2021-01-29 at 18:08 +0100, Matthieu Baerts wrote:
> Hi Davide, Mat,
> 
> On 28/01/2021 18:34, Davide Caratti wrote:
[...]
> Thank you for this new version!
> 
> Just added in our tree, at the bottom, with Mat and my RvB tag.
> 
> - 91384be37ad9 mptcp: fix length of MP_PRIO suboption
> 
> Tests + export have been almost a success! Simult Flow selftest with a 
> debug kernel was unstable but it is not new, see:
> 
>    https://github.com/multipath-tcp/mptcp_net-next/issues/137
> 
> So feel free to send this patch to netdev when you have the opportunity  ;-)
> It doesn't conflict with other pending patches from Geliang that have 
> been recently sent by Mat.

Ok, I'll post it today.

thanks!
diff mbox series

Patch

diff --git a/net/mptcp/options.c b/net/mptcp/options.c
index 775f0576592e..bb874c5d663a 100644
--- a/net/mptcp/options.c
+++ b/net/mptcp/options.c
@@ -699,10 +699,11 @@  static bool mptcp_established_options_mp_prio(struct sock *sk,
 	if (!subflow->send_mp_prio)
 		return false;
 
-	if (remaining < TCPOLEN_MPTCP_PRIO)
+	/* account for the trailing 'nop' option */
+	if (remaining < TCPOLEN_MPTCP_PRIO_ALIGN)
 		return false;
 
-	*size = TCPOLEN_MPTCP_PRIO;
+	*size = TCPOLEN_MPTCP_PRIO_ALIGN;
 	opts->suboptions |= OPTION_MPTCP_PRIO;
 	opts->backup = subflow->request_bkup;
 
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 447ce4631b43..d31edbae8da8 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -61,7 +61,8 @@ 
 #define TCPOLEN_MPTCP_ADD_ADDR6_BASE_PORT	24
 #define TCPOLEN_MPTCP_PORT_LEN		4
 #define TCPOLEN_MPTCP_RM_ADDR_BASE	4
-#define TCPOLEN_MPTCP_PRIO		4
+#define TCPOLEN_MPTCP_PRIO		3
+#define TCPOLEN_MPTCP_PRIO_ALIGN	4
 #define TCPOLEN_MPTCP_FASTCLOSE		12
 
 /* MPTCP MP_JOIN flags */