From nobody Fri Apr 19 13:09:33 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id B55371FB8 for ; Thu, 25 May 2023 07:17:30 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1684999049; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=zMrtPYLc8N0yEkLzvOfgi6NGz/GHoHjbi/eai7DYQdI=; b=clahLJjpbyUAEU3KSL8bQzzmWAVy8auIbLmsfAVYXA2WRf0mBifeDR+x1ey4UU14C6vXfq jybxMykMvWZk64P6olQpfUejPI9Md1GOi4kAD8z64+vAwhWCvG1x2IPJCoSHN2dSFuCLFo zWgixqOVsaxLzN8fxYq5reE01QJ8RQ0= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-515-nGq-ylE-NkOYhr2TdvUmgg-1; Thu, 25 May 2023 03:17:27 -0400 X-MC-Unique: nGq-ylE-NkOYhr2TdvUmgg-1 Received: from smtp.corp.redhat.com (int-mx10.intmail.prod.int.rdu2.redhat.com [10.11.54.10]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 98533101A53A; Thu, 25 May 2023 07:17:27 +0000 (UTC) Received: from gerbillo.redhat.com (unknown [10.39.193.113]) by smtp.corp.redhat.com (Postfix) with ESMTP id 08BC8492B0A; Thu, 25 May 2023 07:17:26 +0000 (UTC) From: Paolo Abeni To: mptcp@lists.linux.dev Cc: Florian Westphal Subject: [PATCH v6 mptcp-next 1/6] mptcp: move snd_una update earlier for fallback socket. Date: Thu, 25 May 2023 09:17:08 +0200 Message-Id: <1759edf2231943dcdff5d74b4aa89ac839d0ef8e.1684997011.git.pabeni@redhat.com> In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.1 on 10.11.54.10 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8"; x-default="true" That will avoid an unneeded conditional in both the fast-path and in the fallback case and will simplify a bit the next patch. Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts --- net/mptcp/options.c | 6 ++++++ net/mptcp/protocol.c | 6 ------ 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/net/mptcp/options.c b/net/mptcp/options.c index 8a8083207be4..4bdcd2b326bd 100644 --- a/net/mptcp/options.c +++ b/net/mptcp/options.c @@ -1119,6 +1119,12 @@ bool mptcp_incoming_options(struct sock *sk, struct = sk_buff *skb) mptcp_data_lock(subflow->conn); if (sk_stream_memory_free(sk)) __mptcp_check_push(subflow->conn, sk); + + /* on fallback we just need to ignore the msk-level snd_una, as + * this is really plain TCP + */ + msk->snd_una =3D READ_ONCE(msk->snd_nxt); + __mptcp_data_acked(subflow->conn); mptcp_data_unlock(subflow->conn); return true; diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 763f709fd5f5..7ecd2f250909 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -1004,12 +1004,6 @@ static void __mptcp_clean_una(struct sock *sk) struct mptcp_data_frag *dtmp, *dfrag; u64 snd_una; =20 - /* on fallback we just need to ignore snd_una, as this is really - * plain TCP - */ - if (__mptcp_check_fallback(msk)) - msk->snd_una =3D READ_ONCE(msk->snd_nxt); - snd_una =3D msk->snd_una; list_for_each_entry_safe(dfrag, dtmp, &msk->rtx_queue, list) { if (after64(dfrag->data_seq + dfrag->data_len, snd_una)) --=20 2.40.1 From nobody Fri Apr 19 13:09:33 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 86FD3BA22 for ; Thu, 25 May 2023 07:17:31 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1684999050; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=GaghEVFzWj+1Y7oosGLFmPUZRQ7aDLAB35YZuGT5SZ4=; b=IfhrzsTKSThjG0IvNgiTT7JxklyUqQgP9Lx/XYIuriUzMrDkWFMgzskIxRZR2T/IVyU0HE pPkPrZWHuHcXZMdw50No42q8lfecvV5CjxcOM0QeyNw1akkTVqtYvNpRkTVtMIhB46Gnf4 rs6khvOel3nVRul5R637z0coK5oj4gA= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-542-dA7OEFlAOGOhm201yuDdmg-1; Thu, 25 May 2023 03:17:28 -0400 X-MC-Unique: dA7OEFlAOGOhm201yuDdmg-1 Received: from smtp.corp.redhat.com (int-mx10.intmail.prod.int.rdu2.redhat.com [10.11.54.10]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 7DA1C85A5BE; Thu, 25 May 2023 07:17:28 +0000 (UTC) Received: from gerbillo.redhat.com (unknown [10.39.193.113]) by smtp.corp.redhat.com (Postfix) with ESMTP id DCA3A492B0B; Thu, 25 May 2023 07:17:27 +0000 (UTC) From: Paolo Abeni To: mptcp@lists.linux.dev Cc: Florian Westphal Subject: [PATCH v6 mptcp-next 2/6] mptcp: track some aggregate data counters. Date: Thu, 25 May 2023 09:17:09 +0200 Message-Id: <9da66102af6d8df4ae08a5a5ff8526f52e2f34b5.1684997011.git.pabeni@redhat.com> In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.1 on 10.11.54.10 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8"; x-default="true" Currently there are no data transfer counters accounting for all the subflows used by a given MPTCP socket. The user-space can compute such figures aggregating the subflow info, but that is inaccurate if any subflow is closed before the MPTCP socket itself. Add the new counters in the MPTCP socket itself and expose them via the existing diag and sockopt. While touching mptcp_diag_fill_info(), acquire the relevant locks before fetching the msk data, to ensure better data consistency Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts --- include/uapi/linux/mptcp.h | 5 +++++ net/mptcp/options.c | 10 ++++++++-- net/mptcp/protocol.c | 12 +++++++++++- net/mptcp/protocol.h | 4 ++++ net/mptcp/sockopt.c | 22 +++++++++++++++++----- 5 files changed, 45 insertions(+), 8 deletions(-) diff --git a/include/uapi/linux/mptcp.h b/include/uapi/linux/mptcp.h index 32af2d278cb4..a124be6ebbba 100644 --- a/include/uapi/linux/mptcp.h +++ b/include/uapi/linux/mptcp.h @@ -123,6 +123,11 @@ struct mptcp_info { __u8 mptcpi_local_addr_used; __u8 mptcpi_local_addr_max; __u8 mptcpi_csum_enabled; + __u32 mptcpi_retransmits; + __u64 mptcpi_bytes_retrans; + __u64 mptcpi_bytes_sent; + __u64 mptcpi_bytes_received; + __u64 mptcpi_bytes_acked; }; =20 /* diff --git a/net/mptcp/options.c b/net/mptcp/options.c index 4bdcd2b326bd..c254accb14de 100644 --- a/net/mptcp/options.c +++ b/net/mptcp/options.c @@ -1026,6 +1026,12 @@ u64 __mptcp_expand_seq(u64 old_seq, u64 cur_seq) return cur_seq; } =20 +static void __mptcp_snd_una_update(struct mptcp_sock *msk, u64 new_snd_una) +{ + msk->bytes_acked +=3D new_snd_una - msk->snd_una; + msk->snd_una =3D new_snd_una; +} + static void ack_update_msk(struct mptcp_sock *msk, struct sock *ssk, struct mptcp_options_received *mp_opt) @@ -1057,7 +1063,7 @@ static void ack_update_msk(struct mptcp_sock *msk, __mptcp_check_push(sk, ssk); =20 if (after64(new_snd_una, old_snd_una)) { - msk->snd_una =3D new_snd_una; + __mptcp_snd_una_update(msk, new_snd_una); __mptcp_data_acked(sk); } mptcp_data_unlock(sk); @@ -1123,7 +1129,7 @@ bool mptcp_incoming_options(struct sock *sk, struct s= k_buff *skb) /* on fallback we just need to ignore the msk-level snd_una, as * this is really plain TCP */ - msk->snd_una =3D READ_ONCE(msk->snd_nxt); + __mptcp_snd_una_update(msk, READ_ONCE(msk->snd_nxt)); =20 __mptcp_data_acked(subflow->conn); mptcp_data_unlock(subflow->conn); diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 7ecd2f250909..79b3059ca3ef 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -377,6 +377,7 @@ static bool __mptcp_move_skb(struct mptcp_sock *msk, st= ruct sock *ssk, =20 if (MPTCP_SKB_CB(skb)->map_seq =3D=3D msk->ack_seq) { /* in sequence */ + msk->bytes_received +=3D copy_len; WRITE_ONCE(msk->ack_seq, msk->ack_seq + copy_len); tail =3D skb_peek_tail(&sk->sk_receive_queue); if (tail && mptcp_try_coalesce(sk, tail, skb)) @@ -760,6 +761,7 @@ static bool __mptcp_ofo_queue(struct mptcp_sock *msk) MPTCP_SKB_CB(skb)->map_seq +=3D delta; __skb_queue_tail(&sk->sk_receive_queue, skb); } + msk->bytes_received +=3D end_seq - msk->ack_seq; msk->ack_seq =3D end_seq; moved =3D true; } @@ -1511,8 +1513,10 @@ static void mptcp_update_post_push(struct mptcp_sock= *msk, * that has been handed to the subflow for transmission * and skip update in case it was old dfrag. */ - if (likely(after64(snd_nxt_new, msk->snd_nxt))) + if (likely(after64(snd_nxt_new, msk->snd_nxt))) { + msk->bytes_sent +=3D snd_nxt_new - msk->snd_nxt; msk->snd_nxt =3D snd_nxt_new; + } } =20 void mptcp_check_and_set_pending(struct sock *sk) @@ -2637,6 +2641,8 @@ static void __mptcp_retrans(struct sock *sk) msk->last_snd =3D ssk; } } + + msk->bytes_retrans +=3D len; dfrag->already_sent =3D max(dfrag->already_sent, len); =20 reset_timer: @@ -3150,6 +3156,10 @@ static int mptcp_disconnect(struct sock *sk, int fla= gs) WRITE_ONCE(msk->csum_enabled, mptcp_is_checksum_enabled(sock_net(sk))); mptcp_pm_data_reset(msk); mptcp_ca_reset(sk); + msk->bytes_acked =3D 0; + msk->bytes_received =3D 0; + msk->bytes_sent =3D 0; + msk->bytes_retrans =3D 0; =20 WRITE_ONCE(sk->sk_shutdown, 0); sk_error_report(sk); diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h index bd3771c7d79d..ec733955f18e 100644 --- a/net/mptcp/protocol.h +++ b/net/mptcp/protocol.h @@ -261,10 +261,13 @@ struct mptcp_sock { u64 local_key; u64 remote_key; u64 write_seq; + u64 bytes_sent; u64 snd_nxt; + u64 bytes_received; u64 ack_seq; atomic64_t rcv_wnd_sent; u64 rcv_data_fin_seq; + u64 bytes_retrans; int rmem_fwd_alloc; struct sock *last_snd; int snd_burst; @@ -273,6 +276,7 @@ struct mptcp_sock { * recovery related fields are under data_lock * protection */ + u64 bytes_acked; u64 snd_una; u64 wnd_end; unsigned long timer_ival; diff --git a/net/mptcp/sockopt.c b/net/mptcp/sockopt.c index d4258869ac48..770279e0a598 100644 --- a/net/mptcp/sockopt.c +++ b/net/mptcp/sockopt.c @@ -888,7 +888,9 @@ static int mptcp_getsockopt_first_sf_only(struct mptcp_= sock *msk, int level, int =20 void mptcp_diag_fill_info(struct mptcp_sock *msk, struct mptcp_info *info) { + struct sock *sk =3D (struct sock *)msk; u32 flags =3D 0; + bool slow; =20 memset(info, 0, sizeof(*info)); =20 @@ -914,11 +916,21 @@ void mptcp_diag_fill_info(struct mptcp_sock *msk, str= uct mptcp_info *info) if (READ_ONCE(msk->can_ack)) flags |=3D MPTCP_INFO_FLAG_REMOTE_KEY_RECEIVED; info->mptcpi_flags =3D flags; - info->mptcpi_token =3D READ_ONCE(msk->token); - info->mptcpi_write_seq =3D READ_ONCE(msk->write_seq); - info->mptcpi_snd_una =3D READ_ONCE(msk->snd_una); - info->mptcpi_rcv_nxt =3D READ_ONCE(msk->ack_seq); - info->mptcpi_csum_enabled =3D READ_ONCE(msk->csum_enabled); + mptcp_data_lock(sk); + info->mptcpi_snd_una =3D msk->snd_una; + info->mptcpi_rcv_nxt =3D msk->ack_seq; + info->mptcpi_bytes_acked =3D msk->bytes_acked; + mptcp_data_unlock(sk); + + slow =3D lock_sock_fast(sk); + info->mptcpi_csum_enabled =3D msk->csum_enabled; + info->mptcpi_token =3D msk->token; + info->mptcpi_write_seq =3D msk->write_seq; + info->mptcpi_retransmits =3D inet_csk(sk)->icsk_retransmits; + info->mptcpi_bytes_sent =3D msk->bytes_sent; + info->mptcpi_bytes_received =3D msk->bytes_received; + info->mptcpi_bytes_retrans =3D msk->bytes_retrans; + unlock_sock_fast(sk, slow); } EXPORT_SYMBOL_GPL(mptcp_diag_fill_info); =20 --=20 2.40.1 From nobody Fri Apr 19 13:09:33 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 3CCC8BA35 for ; Thu, 25 May 2023 07:17:32 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1684999051; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=LWH2EfHSYdsRTUncJhIKmnCKyeCK8zpc1t1M/miAwfM=; b=jUq9et1ALs38XIEV+84sfj4zb2LdH5HtwjP/GUuIRWqfzssCy4iHVNNKekrHgcy4BDGPvd RBGtCtlbyHqmMRIlCcNTZr/rsA7nfx+OwKkEv7SQ0yt0WwBPUE1R5l2omYuPn3dbfaUXQE Moo3LKZezi5Dc/I8GUb0NBZgbsLcvcI= Received: from mimecast-mx02.redhat.com (mx3-rdu2.redhat.com [66.187.233.73]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-265-e2ZQLKH7M_uKZ39O1QRwTQ-1; Thu, 25 May 2023 03:17:29 -0400 X-MC-Unique: e2ZQLKH7M_uKZ39O1QRwTQ-1 Received: from smtp.corp.redhat.com (int-mx10.intmail.prod.int.rdu2.redhat.com [10.11.54.10]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 62EDF29AA38F; Thu, 25 May 2023 07:17:29 +0000 (UTC) Received: from gerbillo.redhat.com (unknown [10.39.193.113]) by smtp.corp.redhat.com (Postfix) with ESMTP id C1D44492B0A; Thu, 25 May 2023 07:17:28 +0000 (UTC) From: Paolo Abeni To: mptcp@lists.linux.dev Cc: Florian Westphal Subject: [PATCH v6 mptcp-next 3/6] selftests: mptcp: explicitly tests aggregate counters Date: Thu, 25 May 2023 09:17:10 +0200 Message-Id: <955947697cb63678edf9b674f95bece320056746.1684997011.git.pabeni@redhat.com> In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.1 on 10.11.54.10 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8"; x-default="true" Update the existing sockopt test-case to do some basic checks on the newly added counters. Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts --- v2 -> v3: - be more kind with older kernel (Matttbe) --- .../selftests/net/mptcp/mptcp_sockopt.c | 27 ++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_sockopt.c b/tools/test= ing/selftests/net/mptcp/mptcp_sockopt.c index ae61f39556ca..ff8fcdfccf76 100644 --- a/tools/testing/selftests/net/mptcp/mptcp_sockopt.c +++ b/tools/testing/selftests/net/mptcp/mptcp_sockopt.c @@ -51,6 +51,11 @@ struct mptcp_info { __u8 mptcpi_local_addr_used; __u8 mptcpi_local_addr_max; __u8 mptcpi_csum_enabled; + __u32 mptcpi_retransmits; + __u64 mptcpi_bytes_retrans; + __u64 mptcpi_bytes_sent; + __u64 mptcpi_bytes_received; + __u64 mptcpi_bytes_acked; }; =20 struct mptcp_subflow_data { @@ -83,8 +88,10 @@ struct mptcp_subflow_addrs { =20 struct so_state { struct mptcp_info mi; + struct mptcp_info last_sample; uint64_t mptcpi_rcv_delta; uint64_t tcpi_rcv_delta; + bool pkt_stats_avail; }; =20 static void die_perror(const char *msg) @@ -318,8 +325,9 @@ static void do_getsockopt_mptcp_info(struct so_state *s= , int fd, size_t w) if (ret < 0) die_perror("getsockopt MPTCP_INFO"); =20 - assert(olen =3D=3D sizeof(i)); + s->pkt_stats_avail =3D olen >=3D sizeof(i); =20 + s->last_sample =3D i; if (s->mi.mptcpi_write_seq =3D=3D 0) s->mi =3D i; =20 @@ -556,6 +564,23 @@ static void process_one_client(int fd, int pipefd) do_getsockopts(&s, fd, ret, ret2); if (s.mptcpi_rcv_delta !=3D (uint64_t)ret + 1) xerror("mptcpi_rcv_delta %" PRIu64 ", expect %" PRIu64, s.mptcpi_rcv_del= ta, ret + 1, s.mptcpi_rcv_delta - ret); + + /* be nice when running on top of older kernel */ + if (s.pkt_stats_avail) { + if (s.last_sample.mptcpi_bytes_sent !=3D ret2) + xerror("mptcpi_bytes_sent %" PRIu64 ", expect %" PRIu64, + s.last_sample.mptcpi_bytes_sent, ret2, + s.last_sample.mptcpi_bytes_sent - ret2); + if (s.last_sample.mptcpi_bytes_received !=3D ret) + xerror("mptcpi_bytes_received %" PRIu64 ", expect %" PRIu64, + s.last_sample.mptcpi_bytes_received, ret, + s.last_sample.mptcpi_bytes_received - ret); + if (s.last_sample.mptcpi_bytes_acked !=3D ret) + xerror("mptcpi_bytes_acked %" PRIu64 ", expect %" PRIu64, + s.last_sample.mptcpi_bytes_acked, ret2, + s.last_sample.mptcpi_bytes_acked - ret2); + } + close(fd); } =20 --=20 2.40.1 From nobody Fri Apr 19 13:09:33 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C88B5BA2F for ; Thu, 25 May 2023 07:17:34 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1684999053; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=Gsb4908GkOnJZxmQ0qWl6DCfidAj/to7N51dZR2wTOw=; b=V+CLlAwspv2m+m8CXTc3Vv+hmSwxqECK4sinYbv9M2H/dbYQy/JgzsftUM93BrAuakg0bO iBwzSsoiNyzo48aBNgGq1wlmn985L7KzSUGZTbfVGQ9EaSPo07DTiFIDDWcZ5xdMVODefp Uypq8OExiPgwxPaSq2+tXyGJKWGrooo= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-152-ga3MLYFPPgmqVUFK7XnrLA-1; Thu, 25 May 2023 03:17:30 -0400 X-MC-Unique: ga3MLYFPPgmqVUFK7XnrLA-1 Received: from smtp.corp.redhat.com (int-mx10.intmail.prod.int.rdu2.redhat.com [10.11.54.10]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 42FB5811E7F; Thu, 25 May 2023 07:17:30 +0000 (UTC) Received: from gerbillo.redhat.com (unknown [10.39.193.113]) by smtp.corp.redhat.com (Postfix) with ESMTP id A7392492B0A; Thu, 25 May 2023 07:17:29 +0000 (UTC) From: Paolo Abeni To: mptcp@lists.linux.dev Cc: Florian Westphal Subject: [PATCH v6 mptcp-next 4/6] mptcp: add subflow unique id Date: Thu, 25 May 2023 09:17:11 +0200 Message-Id: In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.1 on 10.11.54.10 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8"; x-default="true" The user-space need to preperly account the data received/sent by individual subflows. When additional subflows are created and/or closed during the MPTCP socket lifetime, the information currently exposed via MPTCP_TCPINFO are not enough: subflows are identifed only by the sequential position inside the info dumps, and that will change with the above mentioned events. To solve the above problem, this patch introduces a new subflow identifier that is unique inside the given mptcp socket scope. The initial subflow get the id 1 and the other subflows get incremental values at join time. Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts --- v2 -> v3: - fix msk subflow_id init (Matttbe) v1 -> v2: - properly set subflow_id for the first passive subflow and active subflow= s, too - drop the tcpi_fackets overload --- net/mptcp/protocol.c | 6 ++++++ net/mptcp/protocol.h | 5 ++++- net/mptcp/subflow.c | 2 ++ 3 files changed, 12 insertions(+), 1 deletion(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 79b3059ca3ef..895eda3d79e8 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -96,6 +96,7 @@ static int __mptcp_socket_create(struct mptcp_sock *msk) list_add(&subflow->node, &msk->conn_list); sock_hold(ssock->sk); subflow->request_mptcp =3D 1; + subflow->subflow_id =3D msk->subflow_id++; =20 /* This is the first subflow, always with id 0 */ subflow->local_id_valid =3D 1; @@ -847,6 +848,7 @@ static bool __mptcp_finish_join(struct mptcp_sock *msk,= struct sock *ssk) if (sk->sk_socket && !ssk->sk_socket) mptcp_sock_graft(ssk, sk->sk_socket); =20 + mptcp_subflow_ctx(ssk)->subflow_id =3D msk->subflow_id++; mptcp_sockopt_sync_locked(msk, ssk); mptcp_subflow_joined(msk, ssk); return true; @@ -2775,6 +2777,7 @@ static int __mptcp_init_sock(struct sock *sk) WRITE_ONCE(msk->csum_enabled, mptcp_is_checksum_enabled(sock_net(sk))); WRITE_ONCE(msk->allow_infinite_fallback, true); msk->recovery =3D false; + msk->subflow_id =3D 1; =20 mptcp_pm_data_init(msk); =20 @@ -3210,6 +3213,9 @@ struct sock *mptcp_sk_clone_init(const struct sock *s= k, msk->setsockopt_seq =3D mptcp_sk(sk)->setsockopt_seq; mptcp_init_sched(msk, mptcp_sk(sk)->sched); =20 + /* passive msk is created after the first/MPC subflow */ + msk->subflow_id =3D 2; + sock_reset_flag(nsk, SOCK_RCU_FREE); security_inet_csk_clone(nsk, req); =20 diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h index ec733955f18e..b4b5c5004f34 100644 --- a/net/mptcp/protocol.h +++ b/net/mptcp/protocol.h @@ -323,7 +323,8 @@ struct mptcp_sock { u64 rtt_us; /* last maximum rtt of subflows */ } rcvq_space; =20 - u32 setsockopt_seq; + u32 subflow_id; + u32 setsockopt_seq; char ca_name[TCP_CA_NAME_MAX]; struct mptcp_sock *dl_next; }; @@ -505,6 +506,8 @@ struct mptcp_subflow_context { u8 reset_reason:4; u8 stale_count; =20 + u32 subflow_id; + long delegated_status; unsigned long fail_tout; =20 diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c index 63ac4dc621d4..c7001a23550a 100644 --- a/net/mptcp/subflow.c +++ b/net/mptcp/subflow.c @@ -819,6 +819,7 @@ static struct sock *subflow_syn_recv_sock(const struct = sock *sk, if (!ctx->conn) goto fallback; =20 + ctx->subflow_id =3D 1; owner =3D mptcp_sk(ctx->conn); mptcp_pm_new_connection(owner, child, 1); =20 @@ -1574,6 +1575,7 @@ int __mptcp_subflow_connect(struct sock *sk, const st= ruct mptcp_addr_info *loc, subflow->remote_id =3D remote_id; subflow->request_join =3D 1; subflow->request_bkup =3D !!(flags & MPTCP_PM_ADDR_FLAG_BACKUP); + subflow->subflow_id =3D msk->subflow_id++; mptcp_info2sockaddr(remote, &addr, ssk->sk_family); =20 sock_hold(ssk); --=20 2.40.1 From nobody Fri Apr 19 13:09:33 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 458121FB8 for ; Thu, 25 May 2023 07:17:39 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1684999058; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=TB87SoBDvcKQNNXi0xbWf4SFsnaqUO9Fj+iF7wlsBSM=; b=RkODbgeci50zIgABah6T4ZhyYP16Y/W7sbEVuk99QKBMsfDcjOkP3nc+NOEtYYJtyfkMDG tc4EtbsfFMfskN4fksUZ1QjkRH1ezQ9XhUiBtJIJqgvABCfoPp04d/6ZsSbMFAmmvQ+c2P BH5g7fZz5bkirePrFYg7SqM24TnTnm8= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-569-lZ6l5gUlNYKU647FXJXKNg-1; Thu, 25 May 2023 03:17:31 -0400 X-MC-Unique: lZ6l5gUlNYKU647FXJXKNg-1 Received: from smtp.corp.redhat.com (int-mx10.intmail.prod.int.rdu2.redhat.com [10.11.54.10]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 2359985A5BE; Thu, 25 May 2023 07:17:31 +0000 (UTC) Received: from gerbillo.redhat.com (unknown [10.39.193.113]) by smtp.corp.redhat.com (Postfix) with ESMTP id 87A01492B0A; Thu, 25 May 2023 07:17:30 +0000 (UTC) From: Paolo Abeni To: mptcp@lists.linux.dev Cc: Florian Westphal Subject: [PATCH v6 mptcp-next 5/6] mptcp: introduce MPTCP_FULL_INFO getsockopt Date: Thu, 25 May 2023 09:17:12 +0200 Message-Id: <6445ff289e2784aca7f93847fea85abf304c9d50.1684997011.git.pabeni@redhat.com> In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.1 on 10.11.54.10 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8"; x-default="true" Some user-space applications want to monitor the subflows utilization. Dumping the per subflow tcp_info is not enough, as the PM could close and re-create the subflows under-the-hood, fooling the accounting. Even checking the src/dst addresses used by each subflow could not be enough, because new subflows could re-use the same address/port of the just closed one. This patch introduces a new socket option, allow dumping all the relevant information all-at-once (everything, everywhere...), in a consistent manner. Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts --- v5 -> v6: - fix compiler warning - unused variable v4 -> v5: - full_info struct re-design (Florian) - fix build issue on 32 bit hosts v3 -> v4: - full_info struct re-design (Florian) v2 -> v3: - added missing changelog (oops) --- include/uapi/linux/mptcp.h | 24 +++++++ net/mptcp/sockopt.c | 127 ++++++++++++++++++++++++++++++++++++- 2 files changed, 149 insertions(+), 2 deletions(-) diff --git a/include/uapi/linux/mptcp.h b/include/uapi/linux/mptcp.h index a124be6ebbba..ee9c49f949a2 100644 --- a/include/uapi/linux/mptcp.h +++ b/include/uapi/linux/mptcp.h @@ -249,9 +249,33 @@ struct mptcp_subflow_addrs { }; }; =20 +struct mptcp_subflow_info { + __u32 id; + struct mptcp_subflow_addrs addrs; +}; + +struct mptcp_full_info { + __u32 size_tcpinfo_kernel; /* must be 0, set by kernel */ + __u32 size_tcpinfo_user; + __u32 size_sfinfo_kernel; /* must be 0, set by kernel */ + __u32 size_sfinfo_user; + __u32 num_subflows; /* must be 0, set by kernel (real subflow count) */ + __u32 size_arrays_user; /* max subflows that userspace is interested in; + * the buffers at subflow_info/tcp_info + * are respectively at least: + * size_arrays * size_sfinfo_user + * size_arrays * size_tcpinfo_user + * bytes wide + */ + __aligned_u64 subflow_info; + __aligned_u64 tcp_info; + struct mptcp_info mptcp_info; +}; + /* MPTCP socket options */ #define MPTCP_INFO 1 #define MPTCP_TCPINFO 2 #define MPTCP_SUBFLOW_ADDRS 3 +#define MPTCP_FULL_INFO 4 =20 #endif /* _UAPI_MPTCP_H */ diff --git a/net/mptcp/sockopt.c b/net/mptcp/sockopt.c index 770279e0a598..bfbe0378e997 100644 --- a/net/mptcp/sockopt.c +++ b/net/mptcp/sockopt.c @@ -14,7 +14,8 @@ #include #include "protocol.h" =20 -#define MIN_INFO_OPTLEN_SIZE 16 +#define MIN_INFO_OPTLEN_SIZE 16 +#define MIN_FULL_INFO_OPTLEN_SIZE 40 =20 static struct sock *__mptcp_tcp_fallback(struct mptcp_sock *msk) { @@ -977,7 +978,8 @@ static int mptcp_put_subflow_data(struct mptcp_subflow_= data *sfd, } =20 static int mptcp_get_subflow_data(struct mptcp_subflow_data *sfd, - char __user *optval, int __user *optlen) + char __user *optval, + int __user *optlen) { int len, copylen; =20 @@ -1158,6 +1160,125 @@ static int mptcp_getsockopt_subflow_addrs(struct mp= tcp_sock *msk, char __user *o return 0; } =20 +static int mptcp_get_full_info(struct mptcp_full_info *mfi, + char __user *optval, + int __user *optlen) +{ + int len; + + BUILD_BUG_ON(offsetof(struct mptcp_full_info, mptcp_info) !=3D + MIN_FULL_INFO_OPTLEN_SIZE); + + if (get_user(len, optlen)) + return -EFAULT; + + if (len < MIN_FULL_INFO_OPTLEN_SIZE) + return -EINVAL; + + memset(mfi, 0, sizeof(*mfi)); + if (copy_from_user(mfi, optval, MIN_FULL_INFO_OPTLEN_SIZE)) + return -EFAULT; + + if (mfi->size_tcpinfo_kernel || + mfi->size_sfinfo_kernel || + mfi->num_subflows) + return -EINVAL; + + if (mfi->size_sfinfo_user > INT_MAX || + mfi->size_tcpinfo_user > INT_MAX) + return -EINVAL; + + return len - MIN_FULL_INFO_OPTLEN_SIZE; +} + +static int mptcp_put_full_info(struct mptcp_full_info *mfi, + char __user *optval, + u32 copylen, + int __user *optlen) +{ + copylen +=3D MIN_FULL_INFO_OPTLEN_SIZE; + if (put_user(copylen, optlen)) + return -EFAULT; + + if (copy_to_user(optval, mfi, copylen)) + return -EFAULT; + return 0; +} + +static int mptcp_getsockopt_full_info(struct mptcp_sock *msk, char __user = *optval, + int __user *optlen) +{ + unsigned int sfcount =3D 0, copylen =3D 0; + struct mptcp_subflow_context *subflow; + struct sock *sk =3D (struct sock *)msk; + void __user *tcpinfoptr, *sfinfoptr; + struct mptcp_full_info mfi; + int len; + + len =3D mptcp_get_full_info(&mfi, optval, optlen); + if (len < 0) + return len; + + /* don't bother filling the mptcp info if there is not enough + * user-space-provided storage + */ + if (len > 0) { + mptcp_diag_fill_info(msk, &mfi.mptcp_info); + copylen +=3D min_t(unsigned int, len, sizeof(struct mptcp_info)); + } + + mfi.size_tcpinfo_kernel =3D sizeof(struct tcp_info); + mfi.size_tcpinfo_user =3D min_t(unsigned int, mfi.size_tcpinfo_user, + sizeof(struct tcp_info)); + sfinfoptr =3D u64_to_user_ptr(mfi.subflow_info); + mfi.size_sfinfo_kernel =3D sizeof(struct mptcp_subflow_info); + mfi.size_sfinfo_user =3D min_t(unsigned int, mfi.size_sfinfo_user, + sizeof(struct mptcp_subflow_info)); + tcpinfoptr =3D u64_to_user_ptr(mfi.tcp_info); + + lock_sock(sk); + mptcp_for_each_subflow(msk, subflow) { + struct sock *ssk =3D mptcp_subflow_tcp_sock(subflow); + struct mptcp_subflow_info sfinfo; + struct tcp_info tcp_info; + + if (sfcount++ >=3D mfi.size_arrays_user) + continue; + + /* fetch addr/tcp_info only if the user space buffers + * are wide enough + */ + memset(&sfinfo, 0, sizeof(sfinfo)); + sfinfo.id =3D subflow->subflow_id; + if (mfi.size_sfinfo_user > + offsetof(struct mptcp_subflow_info, addrs)) + mptcp_get_sub_addrs(ssk, &sfinfo.addrs); + if (copy_to_user(sfinfoptr, &sfinfo, mfi.size_sfinfo_user)) + goto fail_release; + + if (mfi.size_tcpinfo_user) { + tcp_get_info(ssk, &tcp_info); + if (copy_to_user(tcpinfoptr, &tcp_info, + mfi.size_tcpinfo_user)) + goto fail_release; + } + + tcpinfoptr +=3D mfi.size_tcpinfo_user; + sfinfoptr +=3D mfi.size_sfinfo_user; + } + release_sock(sk); + + mfi.num_subflows =3D sfcount; + if (mptcp_put_full_info(&mfi, optval, copylen, optlen)) + return -EFAULT; + + return 0; + +fail_release: + release_sock(sk); + return -EFAULT; +} + static int mptcp_put_int_option(struct mptcp_sock *msk, char __user *optva= l, int __user *optlen, int val) { @@ -1231,6 +1352,8 @@ static int mptcp_getsockopt_sol_mptcp(struct mptcp_so= ck *msk, int optname, switch (optname) { case MPTCP_INFO: return mptcp_getsockopt_info(msk, optval, optlen); + case MPTCP_FULL_INFO: + return mptcp_getsockopt_full_info(msk, optval, optlen); case MPTCP_TCPINFO: return mptcp_getsockopt_tcpinfo(msk, optval, optlen); case MPTCP_SUBFLOW_ADDRS: --=20 2.40.1 From nobody Fri Apr 19 13:09:33 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 9BE27BA22 for ; Thu, 25 May 2023 07:17:34 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1684999053; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=L74qtqh8bMuRnovWRiTK9aMYmptT5utFOEWeDjxrmkc=; b=IX2M4Sy+Kfc3qVrNWT+6ChhU3swvPAV+2fgJqIb6ZuMae0F/rOyUPt8WQif6BqzWkVi0/K 9DP0Lz7+mKo0+BnxzcOcpuDHXtPZjjjjxc9R0BQk2oY5m0nHoFUasa/0xDG5Zu0CG92c/V d2gf29BKcHwZOqbbTl6Skti1l0nzxZE= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-581-x-hM45xXNz6e3nFRj0hxww-1; Thu, 25 May 2023 03:17:32 -0400 X-MC-Unique: x-hM45xXNz6e3nFRj0hxww-1 Received: from smtp.corp.redhat.com (int-mx10.intmail.prod.int.rdu2.redhat.com [10.11.54.10]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 03050101A53B; Thu, 25 May 2023 07:17:32 +0000 (UTC) Received: from gerbillo.redhat.com (unknown [10.39.193.113]) by smtp.corp.redhat.com (Postfix) with ESMTP id 673BA492B0A; Thu, 25 May 2023 07:17:31 +0000 (UTC) From: Paolo Abeni To: mptcp@lists.linux.dev Cc: Florian Westphal Subject: [PATCH v6 mptcp-next 6/6] selftests: mptcp: add MPTCP_FULL_INFO testcase Date: Thu, 25 May 2023 09:17:13 +0200 Message-Id: <576f5cf229e5b91d27e8411fab1ba7223cec0c17.1684997011.git.pabeni@redhat.com> In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.1 on 10.11.54.10 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8"; x-default="true" Add a testcase explicitly triggering the newly introduce MPTCP_FULL_INFO getsockopt. Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts --- .../selftests/net/mptcp/mptcp_sockopt.c | 91 ++++++++++++++++++- 1 file changed, 89 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_sockopt.c b/tools/test= ing/selftests/net/mptcp/mptcp_sockopt.c index ff8fcdfccf76..568e95599355 100644 --- a/tools/testing/selftests/net/mptcp/mptcp_sockopt.c +++ b/tools/testing/selftests/net/mptcp/mptcp_sockopt.c @@ -86,9 +86,38 @@ struct mptcp_subflow_addrs { #define MPTCP_SUBFLOW_ADDRS 3 #endif =20 +#ifndef MPTCP_FULL_INFO +struct mptcp_subflow_info { + __u32 id; + struct mptcp_subflow_addrs addrs; +}; + +struct mptcp_full_info { + __u32 size_tcpinfo_kernel; /* must be 0, set by kernel */ + __u32 size_tcpinfo_user; + __u32 size_sfinfo_kernel; /* must be 0, set by kernel */ + __u32 size_sfinfo_user; + __u32 num_subflows; /* must be 0, set by kernel (real subflow count) */ + __u32 size_arrays_user; /* max subflows that userspace is interested in; + * the buffers at subflow_info/tcp_info + * are respectively at least: + * size_arrays * size_sfinfo_user + * size_arrays * size_tcpinfo_user + * bytes wide + */ + __aligned_u64 subflow_info; + __aligned_u64 tcp_info; + struct mptcp_info mptcp_info; +}; + +#define MPTCP_FULL_INFO 4 +#endif + struct so_state { struct mptcp_info mi; struct mptcp_info last_sample; + struct tcp_info tcp_info; + struct mptcp_subflow_addrs addrs; uint64_t mptcpi_rcv_delta; uint64_t tcpi_rcv_delta; bool pkt_stats_avail; @@ -365,6 +394,8 @@ static void do_getsockopt_tcp_info(struct so_state *s, = int fd, size_t r, size_t olen -=3D sizeof(struct mptcp_subflow_data); assert(olen =3D=3D sizeof(struct tcp_info)); =20 + s->tcp_info =3D ti.ti[0]; + if (ti.ti[0].tcpi_bytes_sent =3D=3D w && ti.ti[0].tcpi_bytes_received =3D=3D r) goto done; @@ -386,7 +417,7 @@ static void do_getsockopt_tcp_info(struct so_state *s, = int fd, size_t r, size_t do_getsockopt_bogus_sf_data(fd, MPTCP_TCPINFO); } =20 -static void do_getsockopt_subflow_addrs(int fd) +static void do_getsockopt_subflow_addrs(struct so_state *s, int fd) { struct sockaddr_storage remote, local; socklen_t olen, rlen, llen; @@ -433,6 +464,7 @@ static void do_getsockopt_subflow_addrs(int fd) =20 assert(memcmp(&local, &addrs.addr[0].ss_local, sizeof(local)) =3D=3D 0); assert(memcmp(&remote, &addrs.addr[0].ss_remote, sizeof(remote)) =3D=3D 0= ); + s->addrs =3D addrs.addr[0]; =20 memset(&addrs, 0, sizeof(addrs)); =20 @@ -453,13 +485,68 @@ static void do_getsockopt_subflow_addrs(int fd) do_getsockopt_bogus_sf_data(fd, MPTCP_SUBFLOW_ADDRS); } =20 +static void do_getsockopt_mptcp_full_info(struct so_state *s, int fd) +{ + size_t data_size =3D sizeof(struct mptcp_full_info); + struct mptcp_subflow_info sfinfo[2]; + struct tcp_info tcp_info[2]; + struct mptcp_full_info mfi; + socklen_t olen; + int ret; + + memset(&mfi, 0, data_size); + memset(tcp_info, 0, sizeof(tcp_info)); + memset(sfinfo, 0, sizeof(sfinfo)); + + mfi.size_tcpinfo_user =3D sizeof(struct tcp_info); + mfi.size_sfinfo_user =3D sizeof(struct mptcp_subflow_info); + mfi.size_arrays_user =3D 2; + mfi.subflow_info =3D (unsigned long) &sfinfo[0]; + mfi.tcp_info =3D (unsigned long) &tcp_info[0]; + olen =3D data_size; + + ret =3D getsockopt(fd, SOL_MPTCP, MPTCP_FULL_INFO, &mfi, &olen); + if (ret < 0) { + if (errno =3D=3D EOPNOTSUPP) { + perror("MPTCP_FULL_INFO test skipped"); + return; + } + xerror("getsockopt MPTCP_FULL_INFO"); + } + + assert(olen <=3D data_size); + assert(mfi.size_tcpinfo_user =3D=3D mfi.size_tcpinfo_kernel); + assert(mfi.size_tcpinfo_user =3D=3D sizeof(struct tcp_info)); + assert(mfi.size_sfinfo_user =3D=3D mfi.size_sfinfo_kernel); + assert(mfi.size_sfinfo_user =3D=3D sizeof(struct mptcp_subflow_info)); + assert(mfi.num_subflows =3D=3D 1); + + /* Tolerate future extension to mptcp_info struct and running newer + * test on top of older kernel. + * Anyway any kernel supporting MPTCP_FULL_INFO must at least include + * the following in mptcp_info. + */ + assert(olen > (socklen_t)__builtin_offsetof(struct mptcp_full_info, tcp_i= nfo)); + assert(mfi.mptcp_info.mptcpi_subflows =3D=3D 0); + assert(mfi.mptcp_info.mptcpi_bytes_sent =3D=3D s->last_sample.mptcpi_byte= s_sent); + assert(mfi.mptcp_info.mptcpi_bytes_received =3D=3D s->last_sample.mptcpi_= bytes_received); + + assert(sfinfo[0].id =3D=3D 1); + assert(tcp_info[0].tcpi_bytes_sent =3D=3D s->tcp_info.tcpi_bytes_sent); + assert(tcp_info[0].tcpi_bytes_received =3D=3D s->tcp_info.tcpi_bytes_rece= ived); + assert(!memcmp(&sfinfo->addrs, &s->addrs, sizeof(struct mptcp_subflow_add= rs))); +} + static void do_getsockopts(struct so_state *s, int fd, size_t r, size_t w) { do_getsockopt_mptcp_info(s, fd, w); =20 do_getsockopt_tcp_info(s, fd, r, w); =20 - do_getsockopt_subflow_addrs(fd); + do_getsockopt_subflow_addrs(s, fd); + + if (r) + do_getsockopt_mptcp_full_info(s, fd); } =20 static void connect_one_server(int fd, int pipefd) --=20 2.40.1