Re: [PATCH net] sctp: use the pmtu from the icmp packet to update transport pathmtu
From: Xin Long Date: Mon, 15 Oct 2018 19:58:29 +0800 > Other than asoc pmtu sync from all transports, sctp_assoc_sync_pmtu > is also processing transport pmtu_pending by icmp packets. But it's > meaningless to use sctp_dst_mtu(t->dst) as new pmtu for a transport. > > The right pmtu value should come from the icmp packet, and it would > be saved into transport->mtu_info in this patch and used later when > the pmtu sync happens in sctp_sendmsg_to_asoc or sctp_packet_config. > > Besides, without this patch, as pmtu can only be updated correctly > when receiving a icmp packet and no place is holding sock lock, it > will take long time if the sock is busy with sending packets. > > Note that it doesn't process transport->mtu_info in .release_cb(), > as there is no enough information for pmtu update, like for which > asoc or transport. It is not worth traversing all asocs to check > pmtu_pending. So unlike tcp, sctp does this in tx path, for which > mtu_info needs to be atomic_t. > > Signed-off-by: Xin Long Applied.
Re: [PATCH net] sctp: use the pmtu from the icmp packet to update transport pathmtu
On Mon, Oct 15, 2018 at 07:58:29PM +0800, Xin Long wrote:
> Other than asoc pmtu sync from all transports, sctp_assoc_sync_pmtu
> is also processing transport pmtu_pending by icmp packets. But it's
> meaningless to use sctp_dst_mtu(t->dst) as new pmtu for a transport.
>
> The right pmtu value should come from the icmp packet, and it would
> be saved into transport->mtu_info in this patch and used later when
> the pmtu sync happens in sctp_sendmsg_to_asoc or sctp_packet_config.
>
> Besides, without this patch, as pmtu can only be updated correctly
> when receiving a icmp packet and no place is holding sock lock, it
> will take long time if the sock is busy with sending packets.
>
> Note that it doesn't process transport->mtu_info in .release_cb(),
> as there is no enough information for pmtu update, like for which
> asoc or transport. It is not worth traversing all asocs to check
> pmtu_pending. So unlike tcp, sctp does this in tx path, for which
> mtu_info needs to be atomic_t.
>
> Signed-off-by: Xin Long
Acked-by: Marcelo Ricardo Leitner
> ---
> include/net/sctp/structs.h | 2 ++
> net/sctp/associola.c | 3 ++-
> net/sctp/input.c | 1 +
> net/sctp/output.c | 6 ++
> 4 files changed, 11 insertions(+), 1 deletion(-)
>
> diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
> index 28a7c8e..a11f937 100644
> --- a/include/net/sctp/structs.h
> +++ b/include/net/sctp/structs.h
> @@ -876,6 +876,8 @@ struct sctp_transport {
> unsigned long sackdelay;
> __u32 sackfreq;
>
> + atomic_t mtu_info;
> +
> /* When was the last time that we heard from this transport? We use
>* this to pick new active and retran paths.
>*/
> diff --git a/net/sctp/associola.c b/net/sctp/associola.c
> index 297d9cf..a827a1f 100644
> --- a/net/sctp/associola.c
> +++ b/net/sctp/associola.c
> @@ -1450,7 +1450,8 @@ void sctp_assoc_sync_pmtu(struct sctp_association *asoc)
> /* Get the lowest pmtu of all the transports. */
> list_for_each_entry(t, &asoc->peer.transport_addr_list, transports) {
> if (t->pmtu_pending && t->dst) {
> - sctp_transport_update_pmtu(t, sctp_dst_mtu(t->dst));
> + sctp_transport_update_pmtu(t,
> +atomic_read(&t->mtu_info));
> t->pmtu_pending = 0;
> }
> if (!pmtu || (t->pathmtu < pmtu))
> diff --git a/net/sctp/input.c b/net/sctp/input.c
> index 9bbc5f9..5c36a99 100644
> --- a/net/sctp/input.c
> +++ b/net/sctp/input.c
> @@ -395,6 +395,7 @@ void sctp_icmp_frag_needed(struct sock *sk, struct
> sctp_association *asoc,
> return;
>
> if (sock_owned_by_user(sk)) {
> + atomic_set(&t->mtu_info, pmtu);
> asoc->pmtu_pending = 1;
> t->pmtu_pending = 1;
> return;
> diff --git a/net/sctp/output.c b/net/sctp/output.c
> index 7f849b0..67939ad 100644
> --- a/net/sctp/output.c
> +++ b/net/sctp/output.c
> @@ -120,6 +120,12 @@ void sctp_packet_config(struct sctp_packet *packet,
> __u32 vtag,
> sctp_assoc_sync_pmtu(asoc);
> }
>
> + if (asoc->pmtu_pending) {
> + if (asoc->param_flags & SPP_PMTUD_ENABLE)
> + sctp_assoc_sync_pmtu(asoc);
> + asoc->pmtu_pending = 0;
> + }
> +
> /* If there a is a prepend chunk stick it on the list before
>* any other chunks get appended.
>*/
> --
> 2.1.0
>
[PATCH net] sctp: use the pmtu from the icmp packet to update transport pathmtu
Other than asoc pmtu sync from all transports, sctp_assoc_sync_pmtu
is also processing transport pmtu_pending by icmp packets. But it's
meaningless to use sctp_dst_mtu(t->dst) as new pmtu for a transport.
The right pmtu value should come from the icmp packet, and it would
be saved into transport->mtu_info in this patch and used later when
the pmtu sync happens in sctp_sendmsg_to_asoc or sctp_packet_config.
Besides, without this patch, as pmtu can only be updated correctly
when receiving a icmp packet and no place is holding sock lock, it
will take long time if the sock is busy with sending packets.
Note that it doesn't process transport->mtu_info in .release_cb(),
as there is no enough information for pmtu update, like for which
asoc or transport. It is not worth traversing all asocs to check
pmtu_pending. So unlike tcp, sctp does this in tx path, for which
mtu_info needs to be atomic_t.
Signed-off-by: Xin Long
---
include/net/sctp/structs.h | 2 ++
net/sctp/associola.c | 3 ++-
net/sctp/input.c | 1 +
net/sctp/output.c | 6 ++
4 files changed, 11 insertions(+), 1 deletion(-)
diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index 28a7c8e..a11f937 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -876,6 +876,8 @@ struct sctp_transport {
unsigned long sackdelay;
__u32 sackfreq;
+ atomic_t mtu_info;
+
/* When was the last time that we heard from this transport? We use
* this to pick new active and retran paths.
*/
diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 297d9cf..a827a1f 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -1450,7 +1450,8 @@ void sctp_assoc_sync_pmtu(struct sctp_association *asoc)
/* Get the lowest pmtu of all the transports. */
list_for_each_entry(t, &asoc->peer.transport_addr_list, transports) {
if (t->pmtu_pending && t->dst) {
- sctp_transport_update_pmtu(t, sctp_dst_mtu(t->dst));
+ sctp_transport_update_pmtu(t,
+ atomic_read(&t->mtu_info));
t->pmtu_pending = 0;
}
if (!pmtu || (t->pathmtu < pmtu))
diff --git a/net/sctp/input.c b/net/sctp/input.c
index 9bbc5f9..5c36a99 100644
--- a/net/sctp/input.c
+++ b/net/sctp/input.c
@@ -395,6 +395,7 @@ void sctp_icmp_frag_needed(struct sock *sk, struct
sctp_association *asoc,
return;
if (sock_owned_by_user(sk)) {
+ atomic_set(&t->mtu_info, pmtu);
asoc->pmtu_pending = 1;
t->pmtu_pending = 1;
return;
diff --git a/net/sctp/output.c b/net/sctp/output.c
index 7f849b0..67939ad 100644
--- a/net/sctp/output.c
+++ b/net/sctp/output.c
@@ -120,6 +120,12 @@ void sctp_packet_config(struct sctp_packet *packet, __u32
vtag,
sctp_assoc_sync_pmtu(asoc);
}
+ if (asoc->pmtu_pending) {
+ if (asoc->param_flags & SPP_PMTUD_ENABLE)
+ sctp_assoc_sync_pmtu(asoc);
+ asoc->pmtu_pending = 0;
+ }
+
/* If there a is a prepend chunk stick it on the list before
* any other chunks get appended.
*/
--
2.1.0
