Message ID | 88918657d985bc0e55e64ca232dcd5f2b76b7cb4.1440410910.git.lucien.xin@gmail.com |
---|---|
State | Changes Requested, archived |
Delegated to: | David Miller |
Headers | show |
On Mon, Aug 24, 2015 at 06:08:30PM +0800, Xin Long wrote: > as RFC 4960, 6.10 said, *if the receiver detects a partial chunk, it MUST drop > the chunk*, we should not send the abort. but if we put this discard to inside > state machine, it will send abort. > > so we just drop the partial chunk there, never let this chunk go into the state > machine. > > Signed-off-by: Xin Long <lucien.xin@gmail.com> > --- This is basically reverting a chunk of Daniel's and Vlad's 26b87c788100 ("net: sctp: fix remote memory pressure from excessive queueing") . Isn't it going to re-introduce the initial issue then? > net/sctp/inqueue.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > diff --git a/net/sctp/inqueue.c b/net/sctp/inqueue.c > index 7e8a16c..a22ca57 100644 > --- a/net/sctp/inqueue.c > +++ b/net/sctp/inqueue.c > @@ -183,9 +183,9 @@ struct sctp_chunk *sctp_inq_pop(struct sctp_inq *queue) > /* This is not a singleton */ > chunk->singleton = 0; > } else if (chunk->chunk_end > skb_tail_pointer(chunk->skb)) { > - /* Discard inside state machine. */ > - chunk->pdiscard = 1; > - chunk->chunk_end = skb_tail_pointer(chunk->skb); > + sctp_chunk_free(chunk); > + chunk = queue->in_progress = NULL; > + return NULL; > } else { > /* We are at the end of the packet, so mark the chunk > * in case we need to send a SACK. > -- > 2.1.0 > -- 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
On 08/24/2015 06:08 AM, Xin Long wrote: > as RFC 4960, 6.10 said, *if the receiver detects a partial chunk, it MUST drop > the chunk*, we should not send the abort. but if we put this discard to inside > state machine, it will send abort. > Actually, silently dropping this is _very_ bad. There reason is that you've already processed the leading chunks and may have potentially queued a response... Now, you reach the end of the packet and find that the last chunk is partial. You end up dropping the packet, but still handing the responses. This actually lead to some very interesting issues we were seeing. It is better to terminate the association in this case. -vlad > so we just drop the partial chunk there, never let this chunk go into the state > machine. > > Signed-off-by: Xin Long <lucien.xin@gmail.com> > --- > net/sctp/inqueue.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > diff --git a/net/sctp/inqueue.c b/net/sctp/inqueue.c > index 7e8a16c..a22ca57 100644 > --- a/net/sctp/inqueue.c > +++ b/net/sctp/inqueue.c > @@ -183,9 +183,9 @@ struct sctp_chunk *sctp_inq_pop(struct sctp_inq *queue) > /* This is not a singleton */ > chunk->singleton = 0; > } else if (chunk->chunk_end > skb_tail_pointer(chunk->skb)) { > - /* Discard inside state machine. */ > - chunk->pdiscard = 1; > - chunk->chunk_end = skb_tail_pointer(chunk->skb); > + sctp_chunk_free(chunk); > + chunk = queue->in_progress = NULL; > + return NULL; > } else { > /* We are at the end of the packet, so mark the chunk > * in case we need to send a SACK. > -- 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
On 08/24/2015 02:47 PM, Marcelo Ricardo Leitner wrote: > On Mon, Aug 24, 2015 at 06:08:30PM +0800, Xin Long wrote: >> as RFC 4960, 6.10 said, *if the receiver detects a partial chunk, it MUST drop >> the chunk*, we should not send the abort. but if we put this discard to inside >> state machine, it will send abort. >> >> so we just drop the partial chunk there, never let this chunk go into the state >> machine. >> >> Signed-off-by: Xin Long <lucien.xin@gmail.com> >> --- > > This is basically reverting a chunk of Daniel's and Vlad's 26b87c788100 > ("net: sctp: fix remote memory pressure from excessive queueing") . > Isn't it going to re-introduce the initial issue then? Yes, seems so. -- 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
> > Actually, silently dropping this is _very_ bad. There reason is that you've already > processed the leading chunks and may have potentially queued a response... Now, you > reach the end of the packet and find that the last chunk is partial. You end up > dropping the packet, but still handing the responses. This actually lead to some very > interesting issues we were seeing. > > It is better to terminate the association in this case. > > -vlad > make sense, I just cannot ensure we are doing this as RFC, after all, it doesnot say we should send abort in this case clearly. -- 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 --git a/net/sctp/inqueue.c b/net/sctp/inqueue.c index 7e8a16c..a22ca57 100644 --- a/net/sctp/inqueue.c +++ b/net/sctp/inqueue.c @@ -183,9 +183,9 @@ struct sctp_chunk *sctp_inq_pop(struct sctp_inq *queue) /* This is not a singleton */ chunk->singleton = 0; } else if (chunk->chunk_end > skb_tail_pointer(chunk->skb)) { - /* Discard inside state machine. */ - chunk->pdiscard = 1; - chunk->chunk_end = skb_tail_pointer(chunk->skb); + sctp_chunk_free(chunk); + chunk = queue->in_progress = NULL; + return NULL; } else { /* We are at the end of the packet, so mark the chunk * in case we need to send a SACK.
as RFC 4960, 6.10 said, *if the receiver detects a partial chunk, it MUST drop the chunk*, we should not send the abort. but if we put this discard to inside state machine, it will send abort. so we just drop the partial chunk there, never let this chunk go into the state machine. Signed-off-by: Xin Long <lucien.xin@gmail.com> --- net/sctp/inqueue.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)