Content-Type: multipart/mixed; boundary="===============0052027621677873607==" MIME-Version: 1.0 From: Christoph Paasch To: mptcp at lists.01.org Subject: [MPTCP] [PATCH v4 mptcp-net-next 06/12] mptcp: Remove reinjection of data across subflows Date: Fri, 21 Dec 2018 08:13:32 -0800 Message-ID: <20181221161338.72325-7-cpaasch@apple.com> In-Reply-To: 20181221161338.72325-1-cpaasch@apple.com X-Status: X-Keywords: X-UID: 960 --===============0052027621677873607== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable It's an "optimisation" that we don't need in net-next. Signed-off-by: Christoph Paasch --- include/linux/tcp.h | 3 - include/net/mptcp.h | 9 -- net/mptcp/mptcp_ctrl.c | 10 +- net/mptcp/mptcp_fullmesh.c | 12 -- net/mptcp/mptcp_input.c | 20 +-- net/mptcp/mptcp_output.c | 358 +----------------------------------------= ---- net/mptcp/mptcp_sched.c | 32 ++-- 7 files changed, 21 insertions(+), 423 deletions(-) diff --git a/include/linux/tcp.h b/include/linux/tcp.h index cc896f842c31..6dd9c45c52d6 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -470,9 +470,6 @@ struct tcp_sock { request_mptcp:1, /* Did we send out an MP_CAPABLE? * (this speeds up mptcp_doit() in tcp_recvmsg) */ - pf:1, /* Potentially Failed state: when this flag is set, we - * stop using the subflow - */ mp_killed:1, /* Killed with a tcp_done in mptcp? */ was_meta_sk:1, /* This was a meta sk (in case of reuse) */ is_master_sk:1, diff --git a/include/net/mptcp.h b/include/net/mptcp.h index 9e7dd681fb71..5e3fc3e7adba 100644 --- a/include/net/mptcp.h +++ b/include/net/mptcp.h @@ -265,12 +265,6 @@ struct mptcp_cb { u8 mptcp_sched[MPTCP_SCHED_DATA_SIZE] __aligned(8); struct mptcp_sched_ops *sched_ops; = - struct sk_buff_head reinject_queue; - /* First cache-line boundary is here minus 8 bytes. But from the - * reinject-queue only the next and prev pointers are regularly - * accessed. Thus, the whole data-path is on a single cache-line. - */ - u64 infinite_rcv_seq; = /***** Start of fields, used for connection closure */ @@ -421,7 +415,6 @@ extern bool mptcp_init_failed; #define MPTCPHDR_SEQ64_OFO 0x20 /* Is it not in our circular array? */ /* MPTCP flags: TX only */ #define MPTCPHDR_INF 0x08 -#define MPTCP_REINJECT 0x10 /* Did we reinject this segment? */ = struct mptcp_option { __u8 kind; @@ -719,7 +712,6 @@ int mptcp_add_sock(struct sock *meta_sk, struct sock *s= k, u8 loc_id, u8 rem_id, gfp_t flags); void mptcp_del_sock(struct sock *sk); void mptcp_update_metasocket(const struct sock *meta_sk); -void mptcp_reinject_data(struct sock *orig_sk, int clone_it); void mptcp_update_sndbuf(const struct tcp_sock *tp); void mptcp_send_fin(struct sock *meta_sk); void mptcp_send_active_reset(struct sock *meta_sk, gfp_t priority); @@ -1256,7 +1248,6 @@ static inline int is_master_tp(const struct tcp_sock = *tp) } static inline void mptcp_del_sock(const struct sock *sk) {} static inline void mptcp_update_metasocket(const struct sock *meta_sk) {} -static inline void mptcp_reinject_data(struct sock *orig_sk, int clone_it)= {} static inline void mptcp_update_sndbuf(const struct tcp_sock *tp) {} static inline void mptcp_clean_rtx_infinite(const struct sk_buff *skb, const struct sock *sk) {} diff --git a/net/mptcp/mptcp_ctrl.c b/net/mptcp/mptcp_ctrl.c index 16d388fa9454..92c2bf6e6431 100644 --- a/net/mptcp/mptcp_ctrl.c +++ b/net/mptcp/mptcp_ctrl.c @@ -561,7 +561,7 @@ struct sock *mptcp_select_ack_sock(const struct sock *m= eta_sk) struct tcp_sock *tp =3D tcp_sk(sk); u32 elapsed; = - if (!mptcp_sk_can_send_ack(sk) || tp->pf) + if (!mptcp_sk_can_send_ack(sk) || inet_csk(sk)->icsk_ca_state =3D=3D TCP= _CA_Loss) continue; = elapsed =3D keepalive_time_elapsed(tp); @@ -693,8 +693,6 @@ void mptcp_destroy_sock(struct sock *sk) struct mptcp_tcp_sock *mptcp; struct hlist_node *tmp; = - __skb_queue_purge(&tcp_sk(sk)->mpcb->reinject_queue); - /* We have to close all remaining subflows. Normally, they * should all be about to get closed. But, if the kernel is * forcing a closure (e.g., tcp_write_err), the subflows might @@ -1230,7 +1228,6 @@ static int mptcp_alloc_mpcb(struct sock *meta_sk, __u= 64 remote_key, meta_tp->was_meta_sk =3D 0; = /* Initialize the queues */ - skb_queue_head_init(&mpcb->reinject_queue); master_tp->out_of_order_queue =3D RB_ROOT; master_sk->tcp_rtx_queue =3D RB_ROOT; INIT_LIST_HEAD(&master_tp->tsq_node); @@ -1461,9 +1458,6 @@ void mptcp_del_sock(struct sock *sk) tp->mptcp->attached =3D 0; mpcb->path_index_bits &=3D ~(1 << tp->mptcp->path_index); = - if (!tcp_write_queue_empty(sk) || !tcp_rtx_queue_empty(sk)) - mptcp_reinject_data(sk, 0); - if (is_master_tp(tp)) { struct sock *meta_sk =3D mptcp_meta_sk(sk); struct tcp_sock *meta_tp =3D tcp_sk(meta_sk); @@ -1921,8 +1915,6 @@ void mptcp_disconnect(struct sock *sk) struct tcp_sock *tp =3D tcp_sk(sk); struct hlist_node *tmp; = - __skb_queue_purge(&tp->mpcb->reinject_queue); - if (tp->inside_tk_table) mptcp_hash_remove_bh(tp); = diff --git a/net/mptcp/mptcp_fullmesh.c b/net/mptcp/mptcp_fullmesh.c index 4dad457246b1..2446f75b0b2c 100644 --- a/net/mptcp/mptcp_fullmesh.c +++ b/net/mptcp/mptcp_fullmesh.c @@ -923,12 +923,6 @@ static void mptcp_address_worker(struct work_struct *w= ork) !ipv6_addr_equal(&inet6_sk(sk)->saddr, &event->addr.in6)) continue; = - /* Reinject, so that pf =3D 1 and so we - * won't select this one as the - * ack-sock. - */ - mptcp_reinject_data(sk, 0); - /* We announce the removal of this id */ announce_remove_addr(tcp_sk(sk)->mptcp->loc_id, meta_sk); = @@ -1440,12 +1434,6 @@ static void full_mesh_release_sock(struct sock *meta= _sk) } = if (shall_remove) { - /* Reinject, so that pf =3D 1 and so we - * won't select this one as the - * ack-sock. - */ - mptcp_reinject_data(sk, 0); - announce_remove_addr(tcp_sk(sk)->mptcp->loc_id, meta_sk); = diff --git a/net/mptcp/mptcp_input.c b/net/mptcp/mptcp_input.c index 652da0439644..f247e9ef2853 100644 --- a/net/mptcp/mptcp_input.c +++ b/net/mptcp/mptcp_input.c @@ -92,7 +92,7 @@ static inline int mptcp_tso_acked_reinject(const struct s= ock *meta_sk, /* Cleans the meta-socket retransmission queue and the reinject-queue. */ static void mptcp_clean_rtx_queue(struct sock *meta_sk, u32 prior_snd_una) { - struct sk_buff *skb, *tmp, *next; + struct sk_buff *skb, *next; struct tcp_sock *meta_tp =3D tcp_sk(meta_sk); struct mptcp_cb *mpcb =3D meta_tp->mpcb; bool acked =3D false; @@ -149,20 +149,6 @@ static void mptcp_clean_rtx_queue(struct sock *meta_sk= , u32 prior_snd_una) } sk_wmem_free_skb(meta_sk, skb); } - /* Remove acknowledged data from the reinject queue */ - skb_queue_walk_safe(&mpcb->reinject_queue, skb, tmp) { - if (before(meta_tp->snd_una, TCP_SKB_CB(skb)->end_seq)) { - if (tcp_skb_pcount(skb) =3D=3D 1 || - !after(meta_tp->snd_una, TCP_SKB_CB(skb)->seq)) - break; - - mptcp_tso_acked_reinject(meta_sk, skb); - break; - } - - __skb_unlink(skb, &mpcb->reinject_queue); - __kfree_skb(skb); - } = if (likely(between(meta_tp->snd_up, prior_snd_una, meta_tp->snd_una))) meta_tp->snd_up =3D meta_tp->snd_una; @@ -1242,9 +1228,6 @@ static void mptcp_data_ack(struct sock *sk, const str= uct sk_buff *skb) u32 nwin, data_ack, data_seq; u16 data_len =3D 0; = - /* A valid packet came in - subflow is operational again */ - tp->pf =3D 0; - /* Even if there is no data-ack, we stop retransmitting. * Except if this is a SYN/ACK. Then it is just a retransmission */ @@ -1400,7 +1383,6 @@ static void mptcp_send_reset_rem_id(const struct mptc= p_cb *mpcb, u8 rem_id) struct sock *sk_it =3D mptcp_to_sock(mptcp); = if (tcp_sk(sk_it)->mptcp->rem_id =3D=3D rem_id) { - mptcp_reinject_data(sk_it, 0); mptcp_send_reset(sk_it); } } diff --git a/net/mptcp/mptcp_output.c b/net/mptcp/mptcp_output.c index af695fafaab2..26cf41f75922 100644 --- a/net/mptcp/mptcp_output.c +++ b/net/mptcp/mptcp_output.c @@ -54,309 +54,6 @@ int mptcp_sub_len_remove_addr_align(u16 bitfield) } EXPORT_SYMBOL(mptcp_sub_len_remove_addr_align); = -/* get the data-seq and end-data-seq and store them again in the - * tcp_skb_cb - */ -static bool mptcp_reconstruct_mapping(struct sk_buff *skb) -{ - const struct mp_dss *mpdss =3D (struct mp_dss *)TCP_SKB_CB(skb)->dss; - u32 *p32; - u16 *p16; - - if (!mptcp_is_data_seq(skb)) - return false; - - if (!mpdss->M) - return false; - - /* Move the pointer to the data-seq */ - p32 =3D (u32 *)mpdss; - p32++; - if (mpdss->A) { - p32++; - if (mpdss->a) - p32++; - } - - TCP_SKB_CB(skb)->seq =3D ntohl(*p32); - - /* Get the data_len to calculate the end_data_seq */ - p32++; - p32++; - p16 =3D (u16 *)p32; - TCP_SKB_CB(skb)->end_seq =3D ntohs(*p16) + TCP_SKB_CB(skb)->seq; - - return true; -} - -static bool mptcp_is_reinjected(const struct sk_buff *skb) -{ - return TCP_SKB_CB(skb)->mptcp_flags & MPTCP_REINJECT; -} - -static void mptcp_find_and_set_pathmask(struct sock *meta_sk, struct sk_bu= ff *skb) -{ - struct rb_node **p =3D &meta_sk->tcp_rtx_queue.rb_node; - struct rb_node *parent; - struct sk_buff *skb_it; - - while (*p) { - parent =3D *p; - skb_it =3D rb_to_skb(parent); - if (before(TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb_it)->seq)) { - p =3D &parent->rb_left; - continue; - } - if (after(TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb_it)->seq)) { - p =3D &parent->rb_right; - continue; - } - - TCP_SKB_CB(skb)->path_mask =3D TCP_SKB_CB(skb_it)->path_mask; - break; - } -} - -/* Reinject data from one TCP subflow to the meta_sk. If sk =3D=3D NULL, w= e are - * coming from the meta-retransmit-timer - */ -static void __mptcp_reinject_data(struct sk_buff *orig_skb, struct sock *m= eta_sk, - struct sock *sk, int clone_it, - enum tcp_queue tcp_queue) -{ - struct sk_buff *skb, *skb1; - const struct tcp_sock *meta_tp =3D tcp_sk(meta_sk); - struct mptcp_cb *mpcb =3D meta_tp->mpcb; - u32 seq, end_seq; - - if (clone_it) { - /* pskb_copy is necessary here, because the TCP/IP-headers - * will be changed when it's going to be reinjected on another - * subflow. - */ - tcp_skb_tsorted_save(orig_skb) { - skb =3D pskb_copy_for_clone(orig_skb, GFP_ATOMIC); - } tcp_skb_tsorted_restore(orig_skb); - } else { - if (tcp_queue =3D=3D TCP_FRAG_IN_WRITE_QUEUE) - __skb_unlink(orig_skb, &sk->sk_write_queue); - else - rb_erase(&orig_skb->rbnode, &sk->tcp_rtx_queue); - sock_set_flag(sk, SOCK_QUEUE_SHRUNK); - sk->sk_wmem_queued -=3D orig_skb->truesize; - sk_mem_uncharge(sk, orig_skb->truesize); - skb =3D orig_skb; - } - if (unlikely(!skb)) - return; - - if (sk && !mptcp_reconstruct_mapping(skb)) { - __kfree_skb(skb); - return; - } - - skb->sk =3D meta_sk; - - /* Make sure that this list is clean */ - tcp_skb_tsorted_anchor_cleanup(skb); - - /* Reset subflow-specific TCP control-data */ - TCP_SKB_CB(skb)->sacked =3D 0; - TCP_SKB_CB(skb)->tcp_flags &=3D (TCPHDR_ACK | TCPHDR_PSH); - - /* If it reached already the destination, we don't have to reinject it */ - if (!after(TCP_SKB_CB(skb)->end_seq, meta_tp->snd_una)) { - __kfree_skb(skb); - return; - } - - /* Only reinject segments that are fully covered by the mapping */ - if (skb->len + (mptcp_is_data_fin(skb) ? 1 : 0) !=3D - TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq) { - struct rb_node *parent, **p =3D &meta_sk->tcp_rtx_queue.rb_node; - u32 end_seq =3D TCP_SKB_CB(skb)->end_seq; - u32 seq =3D TCP_SKB_CB(skb)->seq; - - __kfree_skb(skb); - - /* Ok, now we have to look for the full mapping in the meta - * send-queue :S - */ - - /* First, find the first skb that covers us */ - while (*p) { - parent =3D *p; - skb =3D rb_to_skb(parent); - - /* Not yet at the mapping? */ - if (!after(end_seq, TCP_SKB_CB(skb)->seq)) { - p =3D &parent->rb_left; - continue; - } - - if (!before(seq, TCP_SKB_CB(skb)->end_seq)) { - p =3D &parent->rb_right; - continue; - } - - break; - } - - if (*p) { - /* We found it, now let's reinject everything */ - skb =3D rb_to_skb(*p); - - skb_rbtree_walk_from(skb) { - if (after(TCP_SKB_CB(skb)->end_seq, end_seq)) - return; - __mptcp_reinject_data(skb, meta_sk, NULL, 1, - TCP_FRAG_IN_RTX_QUEUE); - } - } - return; - } - - /* Segment goes back to the MPTCP-layer. So, we need to zero the - * path_mask/dss. - */ - memset(TCP_SKB_CB(skb)->dss, 0 , mptcp_dss_len); - - /* We need to find out the path-mask from the meta-write-queue - * to properly select a subflow. - */ - mptcp_find_and_set_pathmask(meta_sk, skb); - - /* If it's empty, just add */ - if (skb_queue_empty(&mpcb->reinject_queue)) { - skb_queue_head(&mpcb->reinject_queue, skb); - return; - } - - /* Find place to insert skb - or even we can 'drop' it, as the - * data is already covered by other skb's in the reinject-queue. - * - * This is inspired by code from tcp_data_queue. - */ - - skb1 =3D skb_peek_tail(&mpcb->reinject_queue); - seq =3D TCP_SKB_CB(skb)->seq; - while (1) { - if (!after(TCP_SKB_CB(skb1)->seq, seq)) - break; - if (skb_queue_is_first(&mpcb->reinject_queue, skb1)) { - skb1 =3D NULL; - break; - } - skb1 =3D skb_queue_prev(&mpcb->reinject_queue, skb1); - } - - /* Do skb overlap to previous one? */ - end_seq =3D TCP_SKB_CB(skb)->end_seq; - if (skb1 && before(seq, TCP_SKB_CB(skb1)->end_seq)) { - if (!after(end_seq, TCP_SKB_CB(skb1)->end_seq)) { - /* All the bits are present. Don't reinject */ - __kfree_skb(skb); - return; - } - if (seq =3D=3D TCP_SKB_CB(skb1)->seq) { - if (skb_queue_is_first(&mpcb->reinject_queue, skb1)) - skb1 =3D NULL; - else - skb1 =3D skb_queue_prev(&mpcb->reinject_queue, skb1); - } - } - if (!skb1) - __skb_queue_head(&mpcb->reinject_queue, skb); - else - __skb_queue_after(&mpcb->reinject_queue, skb1, skb); - - /* And clean segments covered by new one as whole. */ - while (!skb_queue_is_last(&mpcb->reinject_queue, skb)) { - skb1 =3D skb_queue_next(&mpcb->reinject_queue, skb); - - if (!after(end_seq, TCP_SKB_CB(skb1)->seq)) - break; - - if (before(end_seq, TCP_SKB_CB(skb1)->end_seq)) - break; - - __skb_unlink(skb1, &mpcb->reinject_queue); - __kfree_skb(skb1); - } - return; -} - -/* Inserts data into the reinject queue */ -void mptcp_reinject_data(struct sock *sk, int clone_it) -{ - struct sock *meta_sk =3D mptcp_meta_sk(sk); - struct sk_buff *skb_it, *tmp; - enum tcp_queue tcp_queue; - - /* It has already been closed - there is really no point in reinjecting */ - if (meta_sk->sk_state =3D=3D TCP_CLOSE) - return; - - skb_queue_walk_safe(&sk->sk_write_queue, skb_it, tmp) { - struct tcp_skb_cb *tcb =3D TCP_SKB_CB(skb_it); - /* Subflow syn's and fin's are not reinjected. - * - * As well as empty subflow-fins with a data-fin. - * They are reinjected below (without the subflow-fin-flag) - */ - if (tcb->tcp_flags & TCPHDR_SYN || - (tcb->tcp_flags & TCPHDR_FIN && !mptcp_is_data_fin(skb_it)) || - (tcb->tcp_flags & TCPHDR_FIN && mptcp_is_data_fin(skb_it) && !skb_it= ->len)) - continue; - - if (mptcp_is_reinjected(skb_it)) - continue; - - tcb->mptcp_flags |=3D MPTCP_REINJECT; - __mptcp_reinject_data(skb_it, meta_sk, sk, clone_it, - TCP_FRAG_IN_WRITE_QUEUE); - } - - skb_rbtree_walk_safe(skb_it, &sk->tcp_rtx_queue, tmp) { - struct tcp_skb_cb *tcb =3D TCP_SKB_CB(skb_it); - /* Subflow syn's and fin's are not reinjected. - * - * As well as empty subflow-fins with a data-fin. - * They are reinjected below (without the subflow-fin-flag) - */ - if (tcb->tcp_flags & TCPHDR_SYN || - (tcb->tcp_flags & TCPHDR_FIN && !mptcp_is_data_fin(skb_it)) || - (tcb->tcp_flags & TCPHDR_FIN && mptcp_is_data_fin(skb_it) && !skb_it= ->len)) - continue; - - if (mptcp_is_reinjected(skb_it)) - continue; - - tcb->mptcp_flags |=3D MPTCP_REINJECT; - __mptcp_reinject_data(skb_it, meta_sk, sk, clone_it, - TCP_FRAG_IN_RTX_QUEUE); - } - - skb_it =3D tcp_write_queue_tail(meta_sk); - tcp_queue =3D TCP_FRAG_IN_WRITE_QUEUE; - - if (!skb_it) { - skb_it =3D skb_rb_last(&meta_sk->tcp_rtx_queue); - tcp_queue =3D TCP_FRAG_IN_RTX_QUEUE; - } - - /* If sk has sent the empty data-fin, we have to reinject it too. */ - if (skb_it && mptcp_is_data_fin(skb_it) && skb_it->len =3D=3D 0 && - TCP_SKB_CB(skb_it)->path_mask & mptcp_pi_to_flag(tcp_sk(sk)->mptcp->p= ath_index)) { - __mptcp_reinject_data(skb_it, meta_sk, NULL, 1, tcp_queue); - } - - tcp_sk(sk)->pf =3D 1; - - mptcp_push_pending_frames(meta_sk); -} -EXPORT_SYMBOL(mptcp_reinject_data); - static void mptcp_combine_dfin(const struct sk_buff *skb, const struct sock *meta_sk, struct sock *subsk) @@ -560,8 +257,7 @@ static bool mptcp_skb_entail(struct sock *sk, struct sk= _buff *skb, int reinject) * - The reinject-queue (reinject =3D=3D -1) */ static int mptcp_fragment(struct sock *meta_sk, enum tcp_queue tcp_queue, - struct sk_buff *skb, u32 len, - gfp_t gfp, int reinject) + struct sk_buff *skb, u32 len, gfp_t gfp) { int ret, diff, old_factor; struct sk_buff *buff; @@ -592,34 +288,6 @@ static int mptcp_fragment(struct sock *meta_sk, enum t= cp_queue tcp_queue, TCP_SKB_CB(buff)->mptcp_flags =3D flags; TCP_SKB_CB(buff)->path_mask =3D TCP_SKB_CB(skb)->path_mask; = - /* If reinject =3D=3D 1, the buff will be added to the reinject - * queue, which is currently not part of memory accounting. So - * undo the changes done by tcp_fragment and update the - * reinject queue. Also, undo changes to the packet counters. - */ - if (reinject =3D=3D 1) { - int undo =3D buff->truesize - diff; - meta_sk->sk_wmem_queued -=3D undo; - sk_mem_uncharge(meta_sk, undo); - - tcp_sk(meta_sk)->mpcb->reinject_queue.qlen++; - if (tcp_queue =3D=3D TCP_FRAG_IN_WRITE_QUEUE) - meta_sk->sk_write_queue.qlen--; - - if (!before(tcp_sk(meta_sk)->snd_nxt, TCP_SKB_CB(buff)->end_seq)) { - undo =3D old_factor - tcp_skb_pcount(skb) - - tcp_skb_pcount(buff); - if (undo) - tcp_adjust_pcount(meta_sk, skb, -undo); - } - - /* tcp_fragment's call to sk_stream_alloc_skb initializes the - * tcp_tsorted_anchor. We need to revert this as it clashes - * with the refdst pointer. - */ - tcp_skb_tsorted_anchor_cleanup(buff); - } - return 0; } = @@ -663,7 +331,7 @@ int mptcp_write_wakeup(struct sock *meta_sk, int mib) seg_size =3D min(seg_size, mss); TCP_SKB_CB(skb)->tcp_flags |=3D TCPHDR_PSH; if (mptcp_fragment(meta_sk, TCP_FRAG_IN_WRITE_QUEUE, - skb, seg_size, GFP_ATOMIC, 0)) + skb, seg_size, GFP_ATOMIC)) return -1; } else if (!tcp_skb_pcount(skb)) { /* see mptcp_write_xmit on why we use UINT_MAX */ @@ -738,14 +406,7 @@ bool mptcp_write_xmit(struct sock *meta_sk, unsigned i= nt mss_now, int nonagle, subtp =3D tcp_sk(subsk); mss_now =3D tcp_current_mss(subsk); = - if (reinject =3D=3D 1) { - if (!after(TCP_SKB_CB(skb)->end_seq, meta_tp->snd_una)) { - /* Segment already reached the peer, take the next one */ - __skb_unlink(skb, &mpcb->reinject_queue); - __kfree_skb(skb); - continue; - } - } else if (reinject =3D=3D -1) { + if (reinject =3D=3D -1) { tcp_queue =3D TCP_FRAG_IN_RTX_QUEUE; } = @@ -810,7 +471,7 @@ bool mptcp_write_xmit(struct sock *meta_sk, unsigned in= t mss_now, int nonagle, = if (skb->len > limit && unlikely(mptcp_fragment(meta_sk, tcp_queue, - skb, limit, gfp, reinject))) + skb, limit, gfp))) break; = if (!mptcp_skb_entail(subsk, skb, reinject)) @@ -833,11 +494,6 @@ bool mptcp_write_xmit(struct sock *meta_sk, unsigned i= nt mss_now, int nonagle, = tcp_minshall_update(meta_tp, mss_now, skb); = - if (reinject > 0) { - __skb_unlink(skb, &mpcb->reinject_queue); - kfree_skb(skb); - } - if (push_one) break; } @@ -1488,7 +1144,7 @@ int mptcp_retransmit_skb(struct sock *meta_sk, struct= sk_buff *skb) = if (skb->len > limit && unlikely(mptcp_fragment(meta_sk, TCP_FRAG_IN_RTX_QUEUE, skb, - limit, GFP_ATOMIC, 0))) + limit, GFP_ATOMIC))) goto failed; = if (!mptcp_skb_entail(subsk, skb, -1)) @@ -1640,10 +1296,8 @@ void mptcp_sub_retransmit_timer(struct sock *sk) = tcp_retransmit_timer(sk); = - if (!tp->fastopen_rsk) { - mptcp_reinject_data(sk, 1); + if (!tp->fastopen_rsk) mptcp_set_rto(sk); - } } = /* Modify values to an mptcp-level for the initial window of new subflows = */ diff --git a/net/mptcp/mptcp_sched.c b/net/mptcp/mptcp_sched.c index 7d4804f8ba91..7f19b4b4bea0 100644 --- a/net/mptcp/mptcp_sched.c +++ b/net/mptcp/mptcp_sched.c @@ -29,7 +29,7 @@ static bool mptcp_is_def_unavailable(struct sock *sk) if (tp->mptcp->pre_established) return true; = - if (tp->pf) + if (inet_csk(sk)->icsk_ca_state =3D=3D TCP_CA_Loss) return true; = return false; @@ -333,25 +333,19 @@ static struct sk_buff *__mptcp_next_segment(struct so= ck *meta_sk, int *reinject) if (mpcb->infinite_mapping_snd || mpcb->send_infinite_mapping) return tcp_send_head(meta_sk); = - skb =3D skb_peek(&mpcb->reinject_queue); + skb =3D tcp_send_head(meta_sk); = - if (skb) { - *reinject =3D 1; - } else { - skb =3D tcp_send_head(meta_sk); - - if (!skb && meta_sk->sk_socket && - test_bit(SOCK_NOSPACE, &meta_sk->sk_socket->flags) && - sk_stream_wspace(meta_sk) < sk_stream_min_wspace(meta_sk)) { - struct sock *subsk =3D get_available_subflow(meta_sk, NULL, - false); - if (!subsk) - return NULL; - - skb =3D mptcp_rcv_buf_optimization(subsk, 0); - if (skb) - *reinject =3D -1; - } + if (!skb && meta_sk->sk_socket && + test_bit(SOCK_NOSPACE, &meta_sk->sk_socket->flags) && + sk_stream_wspace(meta_sk) < sk_stream_min_wspace(meta_sk)) { + struct sock *subsk =3D get_available_subflow(meta_sk, NULL, + false); + if (!subsk) + return NULL; + + skb =3D mptcp_rcv_buf_optimization(subsk, 0); + if (skb) + *reinject =3D -1; } return skb; } -- = 2.16.2 --===============0052027621677873607==--