Content-Type: multipart/mixed; boundary="===============6304001648104068115==" MIME-Version: 1.0 From: Christoph Paasch To: mptcp at lists.01.org Subject: Re: [MPTCP] [PATCH 03/12] mptcp: Use correct tcp_queue when calling mptcp_fragment Date: Sun, 08 Apr 2018 21:04:52 -0700 Message-ID: <20180409040452.GQ16289@MacBook-Pro-6.local> In-Reply-To: bdfbbb60-356a-1e56-8452-bc75ef3c1903@tessares.net X-Status: X-Keywords: X-UID: 518 --===============6304001648104068115== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hello, On 04/04/18 - 17:55:27, Matthieu Baerts wrote: > Hi Christoph, > = > On 26/03/2018 09:32, Christoph Paasch wrote: > > The skb may come from the write-queue, the rtx-queue or the reinject > > queue. > = > Good catch! :) > = > Should we not clearly document that in mptcp_fragment() code? I mean the = skb > may come from the write-queue, the rtx-queue or the reinject-queue but a > reinjection is not a retransmission on the same sk, it should then go in = the > write-queue. This is quite tricky and specific to MPTCP, we need to have > that in mind when looking at this code. sounds good! I will add a comment at the top of the function. Christoph > = > I guess it is not easy to debug when this kind of fragments are in the wr= ong > queue. :) > = > Best regards, > Matt > = > > When it comes from the rtx-queue, we need to give mptcp_fragment the > > right tcp_queue. > > = > > To handle the case when it comes from the reinject-queue, I will post a > > follow-up patch as this seems to be a bug in the stable release as well. > > = > > Fixes: 8a1051d977cf ("Merge tag 'v4.15' into mptcp_trunk") > > Signed-off-by: Christoph Paasch > > --- > > net/mptcp/mptcp_output.c | 12 +++++++++--- > > 1 file changed, 9 insertions(+), 3 deletions(-) > > = > > diff --git a/net/mptcp/mptcp_output.c b/net/mptcp/mptcp_output.c > > index 43a21f66ac34..fc950a98ce31 100644 > > --- a/net/mptcp/mptcp_output.c > > +++ b/net/mptcp/mptcp_output.c > > @@ -607,7 +607,10 @@ static int mptcp_fragment(struct sock *meta_sk, en= um tcp_queue tcp_queue, > > if (ret) > > return ret; > > - buff =3D skb->next; > > + if (tcp_queue =3D=3D TCP_FRAG_IN_WRITE_QUEUE) > > + buff =3D skb->next; > > + else > > + buff =3D skb_rb_next(skb); > > flags =3D TCP_SKB_CB(skb)->mptcp_flags; > > TCP_SKB_CB(skb)->mptcp_flags =3D flags & ~(MPTCPHDR_FIN); > > @@ -740,6 +743,7 @@ bool mptcp_write_xmit(struct sock *meta_sk, unsigne= d int mss_now, int nonagle, > > while ((skb =3D mpcb->sched_ops->next_segment(meta_sk, &reinject, &s= ubsk, > > &sublimit))) { > > + enum tcp_queue tcp_queue =3D TCP_FRAG_IN_WRITE_QUEUE; > > unsigned int limit; > > WARN(TCP_SKB_CB(skb)->sacked, "sacked: %u reinject: %u", > > @@ -755,6 +759,8 @@ bool mptcp_write_xmit(struct sock *meta_sk, unsigne= d int mss_now, int nonagle, > > __kfree_skb(skb); > > continue; > > } > > + } else if (reinject =3D=3D -1) { > > + tcp_queue =3D TCP_FRAG_IN_RTX_QUEUE; > > } > > /* If the segment was cloned (e.g. a meta retransmission), > > @@ -817,7 +823,7 @@ bool mptcp_write_xmit(struct sock *meta_sk, unsigne= d int mss_now, int nonagle, > > limit =3D min(limit, sublimit); > > if (skb->len > limit && > > - unlikely(mptcp_fragment(meta_sk, TCP_FRAG_IN_WRITE_QUEUE, > > + unlikely(mptcp_fragment(meta_sk, tcp_queue, > > skb, limit, gfp, reinject))) > > break; > > @@ -1545,7 +1551,7 @@ int mptcp_retransmit_skb(struct sock *meta_sk, st= ruct sk_buff *skb) > > TCP_NAGLE_OFF); > > if (skb->len > limit && > > - unlikely(mptcp_fragment(meta_sk, TCP_FRAG_IN_WRITE_QUEUE, skb, > > + unlikely(mptcp_fragment(meta_sk, TCP_FRAG_IN_RTX_QUEUE, skb, > > limit, GFP_ATOMIC, 0))) > > goto failed; > > = > = > -- = > Matthieu Baerts | R&D Engineer > matthieu.baerts(a)tessares.net > Tessares SA | Hybrid Access Solutions > www.tessares.net > 1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium > = > -- = > = > = > DISCLAIMER. > This email and any files transmitted with it are confidential and intended > solely for the use of the individual or entity to whom they are addressed. > If you have received this email in error please notify the system manager. > This message contains confidential information and is intended only for t= he > individual named. If you are not the named addressee you should not > disseminate, distribute or copy this e-mail. Please notify the sender > immediately by e-mail if you have received this e-mail by mistake and del= ete > this e-mail from your system. If you are not the intended recipient you a= re > notified that disclosing, copying, distributing or taking any action in > reliance on the contents of this information is strictly prohibited. --===============6304001648104068115==--