From nobody Sat Jul 25 07:27:27 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (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 C67592931D7 for ; Wed, 15 Jul 2026 11:04:03 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113444; cv=none; b=TgfgOVDivG1APW/yNKjn/5NeR4fsew3fTj/I1mV73gT9H34cR+39LyJe8RnDtV/Qj42NDHQHm9L5+MH2+QrCSFp1LwE90eoxWRxOEn/E+yq9OQgyf9R8pJjzebgwDLVYC3wznp/kQQ2WgG1bB20f4aZOZsJNHjUNs5SCKa+6vok= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113444; c=relaxed/simple; bh=jME7UwAhFGdpy0DC+LD1XLRwwMwYlVOX++7xseGcuGo=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=SMQ4ojtp3+xpjEz3Dpx6tPOGjuT+gK+dah2tDyexJ7xAl0F1JrZj6CV80lWIupm1sfI3drK2K0xZRFuLZ0H05+Acu5pYg5vlkl1qjxPiSFpROvel6E3ytUO95tv5qEGUq2dNFncVSI267a9iymr76AbSkIs66hHzuQGdQ/q/FMU= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=D2xeuj4f; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="D2xeuj4f" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 55D351F00A3A; Wed, 15 Jul 2026 11:04:02 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1784113443; bh=EYeyAMGYwgrEvRqPliDbY8/sG7cAJUpla97WO7oBIDM=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=D2xeuj4fkwb4w8qJLyDjmZwufiWzwX+SAKmztQuRCnzjFK/7t5aJnhFGJ/DNSkbZJ srQUU3AcNeVMQ7M01Q8MDrCHExxD5ze8CF/LfjBMIDFtzkvgpy7FHBW4Gn3h5Fh4+L Obc/xYexE7nSnI/quw/fYWu+M7Kqdsb5hPRWsKJmreb/xzja5UpAD+7NqEfKR4NbQA fDxD21qYDb00kpR88G1i6h582T78Q4YGrm61P31KrlFon/OXP6f9rOLTb66LrYXbbT IM0rskgh5SvfVJ+ygvoKtVp4/3gAk6u4FfoMJ15dnFp0tc79DuXhClEAN35wIQVWZ6 2t6xP/BKxUL+w== From: Geliang Tang To: mptcp@lists.linux.dev Cc: Geliang Tang , Paolo Abeni Subject: [PATCH mptcp-next v5 1/8] mptcp: use local variable tp in sendmsg_frag Date: Wed, 15 Jul 2026 19:03:46 +0800 Message-ID: X-Mailer: git-send-email 2.53.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Geliang Tang In mptcp_sendmsg_frag(), tcp_sk(ssk) is called multiple times. Introduce a local variable 'tp' to hold the tcp_sock pointer and use it throughout the function. This avoids repeated dereferencing and improves code readability, with no functional change. Suggested-by: Paolo Abeni Signed-off-by: Geliang Tang --- net/mptcp/protocol.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index ffcf5a1788f6..1cab07b1b9b1 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -1375,6 +1375,7 @@ static int mptcp_sendmsg_frag(struct sock *sk, struct= sock *ssk, u64 data_seq =3D dfrag->data_seq + info->sent; int offset =3D dfrag->offset + info->sent; struct mptcp_sock *msk =3D mptcp_sk(sk); + struct tcp_sock *tp =3D tcp_sk(ssk); bool zero_window_probe =3D false; struct mptcp_ext *mpext =3D NULL; bool can_coalesce =3D false; @@ -1410,14 +1411,14 @@ static int mptcp_sendmsg_frag(struct sock *sk, stru= ct sock *ssk, mpext =3D mptcp_get_ext(skb); if (!mptcp_skb_can_collapse_to(data_seq, skb, mpext)) { TCP_SKB_CB(skb)->eor =3D 1; - tcp_mark_push(tcp_sk(ssk), skb); + tcp_mark_push(tp, skb); goto alloc_skb; } =20 i =3D skb_shinfo(skb)->nr_frags; can_coalesce =3D skb_can_coalesce(skb, i, dfrag->page, offset); if (!can_coalesce && i >=3D READ_ONCE(net_hotdata.sysctl_max_skb_frags))= { - tcp_mark_push(tcp_sk(ssk), skb); + tcp_mark_push(tp, skb); goto alloc_skb; } =20 @@ -1471,7 +1472,7 @@ static int mptcp_sendmsg_frag(struct sock *sk, struct= sock *ssk, skb->truesize +=3D copy; sk_wmem_queued_add(ssk, copy); sk_mem_charge(ssk, copy); - WRITE_ONCE(tcp_sk(ssk)->write_seq, tcp_sk(ssk)->write_seq + copy); + WRITE_ONCE(tp->write_seq, tp->write_seq + copy); TCP_SKB_CB(skb)->end_seq +=3D copy; tcp_skb_pcount_set(skb, 0); =20 --=20 2.53.0 From nobody Sat Jul 25 07:27:27 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (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 3A4D22931D7 for ; Wed, 15 Jul 2026 11:04:04 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113446; cv=none; b=oEnLpV64Shqjrw/c3B1SZzuZbvijE5cQyPV7ygDjN9c1dgUElasOdwtipCIcbqtCAahP/sg3z+WznhBijlgNa8KN0goGMqJa78h/p5HemwNdOFulsvkckhX9Yh3bLmd00Wlq9PJpSWA20WQOraI0gkkGABb3Wl8bopEvOfYii1s= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113446; c=relaxed/simple; bh=cs9ZDRNpHUGOZhVgZLHEw9C7VMUcJdt8rlQlk1KLNQw=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=ZGIzrhIJCSXyvTdgC1fN4jeliYIM+zPuqBsQeNxQdrNwJEVK2EGDo1g7i0O5d8337KNpzB7r+3XoGQKegDtmWTG3qZ9cQQpDlFtutVpAx5JgiYoucGSfFXSzmaDgstYsbShjLcjROYZjklRTTS65cIhlC19omS8+qOIaujlt+xw= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=VCDUJE06; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="VCDUJE06" Received: by smtp.kernel.org (Postfix) with ESMTPSA id D071C1F000E9; Wed, 15 Jul 2026 11:04:03 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1784113444; bh=ElSZlA4Rb1X0Hy0iYkwooMfxGtyLfN6d3LNYXmaf+tE=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=VCDUJE06+cMzau7YyzBvmqK9d/hnSVrgdrvBZ5QjrQ81gjV/TZ4+EBcg9KK3Eod44 wQaqnK4LBUFAgwFZEg9/OmmY6FdhokhO2EzYudOJCSlWywkAQ3hWGmyGzqJ3pu19Ua JGYtpUkmQsTP7Os/FSTvq7A4tldoIWKUkOe2880BGsNbays6urvKYrnmGwMMaTrWZk ZcY+wNEhi0UHO5tT085rChANVI7TO8Hp3ogZf8XJANYbF5WqZzRUmesCbRLm4B/Ls6 +0WXJ8rF2PA52qCyn5/v1wI7IWVR81tG3YBI4eBzXDg55xld+zeYeUAKKwWG8cBbxr kpwpfE1tvnWSg== From: Geliang Tang To: mptcp@lists.linux.dev Cc: Geliang Tang Subject: [PATCH mptcp-next v5 2/8] mptcp: align struct mptcp_data_frag to 8 bytes on 32-bit Date: Wed, 15 Jul 2026 19:03:47 +0800 Message-ID: <05eacad4eb98fd960fde99acda2e339c6f64649e.1784113088.git.tanggeliang@kylinos.cn> X-Mailer: git-send-email 2.53.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Geliang Tang mptcp_carve_data_frag() places a struct mptcp_data_frag inside a page-frag region, using ALIGN(orig_offset, sizeof(long)) to compute the offset where the struct lives. On 64-bit Linux sizeof(long) is 8 and this works as intended. On 32-bit Linux sizeof(long) is 4, which only guarantees 4-byte alignment for the struct. struct mptcp_data_frag contains a u64 data_seq member, which requires 8-byte alignment. 4-byte-aligned 64-bit loads fault on 32-bit architectures that do not handle unaligned 64-bit accesses in hardware (armv7, mips, sparc, ...). The in-source comment near BUILD_BUG_ON that states "ALIGN(1, sizeof(long)) - 1, so 8-1" is misleading because it implicitly assumes sizeof(long) is always 8. Switch both the ALIGN() and the matching BUILD_BUG_ON to sizeof(u64) so the 8-byte alignment is enforced on every architecture, and the overhead u8 budget still fits the worst-case 7-byte padding plus the struct size. This is a no-op on 64-bit and a correctness fix on 32-bit. Signed-off-by: Geliang Tang --- net/mptcp/protocol.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 1cab07b1b9b1..ca0c02f3bfd3 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -1237,7 +1237,7 @@ static struct mptcp_data_frag * mptcp_carve_data_frag(const struct mptcp_sock *msk, struct page_frag *pfra= g, int orig_offset) { - int offset =3D ALIGN(orig_offset, sizeof(long)); + int offset =3D ALIGN(orig_offset, sizeof(u64)); struct mptcp_data_frag *dfrag; =20 dfrag =3D (struct mptcp_data_frag *)(page_to_virt(pfrag->page) + offset); @@ -4789,9 +4789,9 @@ void __init mptcp_proto_init(void) BUILD_BUG_ON(sizeof(struct mptcp_skb_cb) > sizeof_field(struct sk_buff, c= b)); =20 /* struct mptcp_data_frag: 'overhead' corresponds to the alignment - * (ALIGN(1, sizeof(long)) - 1, so 8-1) + the struct's size + * (ALIGN(1, sizeof(u64)) - 1, so 8-1) + the struct's size */ - BUILD_BUG_ON(ALIGN(1, sizeof(long)) - 1 + sizeof(struct mptcp_data_frag) + BUILD_BUG_ON(ALIGN(1, sizeof(u64)) - 1 + sizeof(struct mptcp_data_frag) > U8_MAX); } =20 --=20 2.53.0 From nobody Sat Jul 25 07:27:27 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (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 35A35334692 for ; Wed, 15 Jul 2026 11:04:06 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113447; cv=none; b=SpZerCpfflNKJG7YysBpEUJ1E870GzGZfdYEcHre9O10iWGgL4Zimmy0uFiDuv757349V52Hj8smwzXE62pHTaE4XiiiSfyT/htjtxD5RNseDa350PChE5hgxlSsiuziL//u0n6yVuYs27ysMUCodqwuT3Xeg5nbi6GzspY5d94= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113447; c=relaxed/simple; bh=VO2lWM9yY6PIOaTJ4sd3MLKK0esuT4PcPjMrYkyKJfU=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=arttRasrkMQGdlVZPi2Fkm1hoqaNTjsnGpM50ZqCjM2z5xKSkVmGucEsusyUYswmmLPWh5LRBfwLZj2AxmM7a2qW5DvtzL3ElsSz5YsVPsqgfj8GG62Yz9haZ6kEKJW6Ue/iveBCI9IpgAbgTK9qDs7xEhPhh9nhfBzVCyJBG9c= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=VGmkgLlm; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="VGmkgLlm" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 50F751F00A3A; Wed, 15 Jul 2026 11:04:05 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1784113446; bh=hBTp4gJcdKx312WpXFgr4fskzyl+43n/D4lcsgFrgyg=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=VGmkgLlm6pJ7awVOXC7c9tG8Ol1RI5J6HQml7EI+2eDYIbZEltW6Td+qhP9H7DNC5 VaafwqWEYuxRiOaDysEaNietvHQBhiSTOK04RPOBTTrKUsLDKoq+oQYmi624DM4O70 fG3jHPPrO4+8zQJTz7H7vwQtMmEhEkiPu0egfbS68W3DHah3BQWa01F8kzZ/NPTR56 H1DXX+DypyUUSlgv4IjKxeNqTPfMofj5p6ogKbuEZZIiIKOfu3lusgUznYOZOqKhKY 59+Z1MICHQ+kynI+qvxhsCCbsdr6iKkJyoENt4zw9lm0fOZ8rCVhpVxQ6Gf7IxFrjp 7l0mjusRx1WVg== From: Geliang Tang To: mptcp@lists.linux.dev Cc: Geliang Tang Subject: [PATCH mptcp-next v5 3/8] mptcp: remove redundant orig_offset in carve_data_frag Date: Wed, 15 Jul 2026 19:03:48 +0800 Message-ID: <0241a104ea5183010ac867dffb7fe8366fab6da7.1784113088.git.tanggeliang@kylinos.cn> X-Mailer: git-send-email 2.53.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Geliang Tang mptcp_carve_data_frag() already receives struct page_frag via its 'pfrag' parameter, making the separate 'orig_offset' parameter redundant since it merely duplicates pfrag->offset. Simplify the function signature by removing this unnecessary parameter and directly accessing the offset from struct pfrag internally. Signed-off-by: Geliang Tang --- net/mptcp/protocol.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index ca0c02f3bfd3..6a9daa187dad 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -1234,12 +1234,13 @@ static bool mptcp_page_frag_refill(struct sock *sk,= struct page_frag *pfrag) } =20 static struct mptcp_data_frag * -mptcp_carve_data_frag(const struct mptcp_sock *msk, struct page_frag *pfra= g, - int orig_offset) +mptcp_carve_data_frag(const struct mptcp_sock *msk, struct page_frag *pfra= g) { - int offset =3D ALIGN(orig_offset, sizeof(u64)); + int orig_offset =3D pfrag->offset; struct mptcp_data_frag *dfrag; + int offset; =20 + offset =3D ALIGN(orig_offset, sizeof(u64)); dfrag =3D (struct mptcp_data_frag *)(page_to_virt(pfrag->page) + offset); dfrag->data_len =3D 0; dfrag->data_seq =3D msk->write_seq; @@ -2031,7 +2032,7 @@ static int mptcp_sendmsg(struct sock *sk, struct msgh= dr *msg, size_t len) if (!mptcp_page_frag_refill(sk, pfrag)) goto wait_for_memory; =20 - dfrag =3D mptcp_carve_data_frag(msk, pfrag, pfrag->offset); + dfrag =3D mptcp_carve_data_frag(msk, pfrag); frag_truesize =3D dfrag->overhead; } =20 --=20 2.53.0 From nobody Sat Jul 25 07:27:27 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (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 BF54E334692 for ; Wed, 15 Jul 2026 11:04:07 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113449; cv=none; b=rs6wwSzQvD7Wnww+7SDEVjDavfs506sNUDNGYOu2G+uaPO+k8pAXCPNUYuA2WaWhluU5sASNyCoTu8uKBaWn7EQbEP5YKZIOESq8yUNzVtH5r9/IBs3Lwur/RLh6picX3wjbE1FrmZKTmLX1EW7Br1mrIn7g0qpXYV4xjTR1dOw= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113449; c=relaxed/simple; bh=M4Q7y0BFTZO7cMUUi332vVpjkeFLm+EJpc7xnDakPQw=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=jvgGYqomel+LgRbAy5e/K8BZSR5A7v1SuhEn9w7H3gJ8JhAT8PQ96QJAAYrWGUtP/l3ehKEssOwy0ZW5DWxlhWdEPGoubrf1Y3Fuj6nn5VIZdWiobaC0No/4X2/KWhvLFzPj6+/tnTCOlYW0QstPUBdRLb/MuWeUnUHA7824/zU= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=UOY1h75z; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="UOY1h75z" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 980BA1F000E9; Wed, 15 Jul 2026 11:04:06 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1784113447; bh=RKtj0w3YErCE4EZCPSjpmGTgi2Qyo5lWeqaiKMTVwM0=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=UOY1h75zmYSmmADM8QiCc3aHHxbBWUbtRHAvuyFGue91r1GRknRAABsPFjId6tdtn dPGbOfRymvxrZMbfSDNLk9gSQyfOZQ61fQslsxjxJeJBzdvq7dUM9htE3n3u9s61ML L5apWVo42cOKC7T+AlWOy5EMy7UiHOs8KQDDeNinqS7WlSv46/T3cgSEMkFwTMzVpg wd+UkwvH9doDAcJPpkcMGp+A9VQm8KqfNCf7taYG5IZPaEjQhCTwXUftVNCnxn4tlm ImFAdjDnQD0x3TKSTIBFnxI3kNoMIx1SqG2cwhSYc0UZSCYboZeZ0ST5m5EGs8bOb3 mzo093+Kh87Cw== From: Geliang Tang To: mptcp@lists.linux.dev Cc: Geliang Tang Subject: [PATCH mptcp-next v5 4/8] mptcp: add MSG_ZEROCOPY support Date: Wed, 15 Jul 2026 19:03:49 +0800 Message-ID: <844992bd333b4605819e347d259e0c94c71709b3.1784113088.git.tanggeliang@kylinos.cn> X-Mailer: git-send-email 2.53.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Geliang Tang This patch enables MSG_ZEROCOPY support for MPTCP sockets, allowing user-space to transmit data without intermediate kernel copies. The implementation follows the same lifecycle model as TCP: A single ubuf_info_msgzc is allocated per sendmsg via msg_zerocopy_realloc(). Its reference count consists of three parts: one 'caller' reference held until sendmsg returns, one reference per MPTCP dfrag in the retransmission queue (released when the dfrag is acked and cleared), and one reference per subflow skb that consumes dfrag bytes (released by TCP's skb_zcopy_clear() when the skb is acked). Completion is reported only after all bytes are acknowledged at both MPTCP and subflow levels. User pages are pinned via iov_iter_get_pages2() and stored directly in a newly allocated mptcp_data_frag. The dfrag takes one reference on the ubuf_info and one on the page. Subflow skbs take additional page references via get_page() and are marked with skb_zcopy_set() and SKBFL_PURE_ZEROCOPY, mirroring TCP's behavior and bypassing kernel memory accounting for those pages. The feature is silently downgraded to regular copy when MSG_FASTOPEN is set or the application has not opted in via setsockopt(SO_ZEROCOPY). If msg_zerocopy_realloc() fails (e.g. RLIMIT_MEMLOCK pressure), sendmsg(2) aborts with -ENOBUFS. The implementation prevents coalescing of dfrags with different ubuf instances into the same subflow skb to avoid premature completion notifications. The data_len field is widened from u16 to u32 to prevent truncation on 64KB-page kernels (ARM64, PPC64). The dfrag metadata overhead is accounted for in sk_wmem_queued_add() to prevent bypassing socket memory limits. When iov_iter_get_pages2() returns 0, the function returns -EFAULT to properly signal iterator exhaustion. Signed-off-by: Geliang Tang --- net/mptcp/protocol.c | 114 ++++++++++++++++++++++++++++++++++++++++--- net/mptcp/protocol.h | 3 +- 2 files changed, 109 insertions(+), 8 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 6a9daa187dad..628c5902afde 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -1118,9 +1118,11 @@ static bool mptcp_frag_can_collapse_to(const struct = mptcp_sock *msk, df->data_seq + df->data_len =3D=3D msk->write_seq; } =20 -static void dfrag_uncharge(struct sock *sk, int len) +static void dfrag_uncharge(struct sock *sk, int len, + struct mptcp_data_frag *dfrag) { - sk_mem_uncharge(sk, len); + if (!dfrag->ubuf) + sk_mem_uncharge(sk, len); sk_wmem_queued_add(sk, -len); } =20 @@ -1129,7 +1131,13 @@ static void dfrag_clear(struct sock *sk, struct mptc= p_data_frag *dfrag) int len =3D dfrag->data_len + dfrag->overhead; =20 list_del(&dfrag->list); - dfrag_uncharge(sk, len); + dfrag_uncharge(sk, len, dfrag); + if (dfrag->ubuf) { + net_zcopy_put(dfrag->ubuf); + put_page(dfrag->page); + kfree(dfrag); + return; + } put_page(dfrag->page); } =20 @@ -1174,7 +1182,7 @@ static void __mptcp_clean_una(struct sock *sk) dfrag->data_len -=3D delta; dfrag->already_sent -=3D delta; =20 - dfrag_uncharge(sk, delta); + dfrag_uncharge(sk, delta, dfrag); } =20 /* all retransmitted data acked, recovery completed */ @@ -1249,6 +1257,7 @@ mptcp_carve_data_frag(const struct mptcp_sock *msk, s= truct page_frag *pfrag) dfrag->already_sent =3D 0; dfrag->page =3D pfrag->page; dfrag->eor =3D 0; + dfrag->ubuf =3D NULL; =20 return dfrag; } @@ -1418,6 +1427,17 @@ static int mptcp_sendmsg_frag(struct sock *sk, struc= t sock *ssk, =20 i =3D skb_shinfo(skb)->nr_frags; can_coalesce =3D skb_can_coalesce(skb, i, dfrag->page, offset); + if (skb_zcopy_pure(skb) !=3D !!dfrag->ubuf) { + can_coalesce =3D false; + tcp_mark_push(tp, skb); + goto alloc_skb; + } + if (can_coalesce && dfrag->ubuf && skb_zcopy(skb) && + skb_zcopy(skb) !=3D dfrag->ubuf) { + can_coalesce =3D false; + tcp_mark_push(tp, skb); + goto alloc_skb; + } if (!can_coalesce && i >=3D READ_ONCE(net_hotdata.sysctl_max_skb_frags))= { tcp_mark_push(tp, skb); goto alloc_skb; @@ -1456,7 +1476,7 @@ static int mptcp_sendmsg_frag(struct sock *sk, struct= sock *ssk, } =20 copy =3D min_t(size_t, copy, info->limit - info->sent); - if (!sk_wmem_schedule(ssk, copy)) { + if (!dfrag->ubuf && !sk_wmem_schedule(ssk, copy)) { tcp_remove_empty_skb(ssk); return -ENOMEM; } @@ -1466,13 +1486,19 @@ static int mptcp_sendmsg_frag(struct sock *sk, stru= ct sock *ssk, } else { get_page(dfrag->page); skb_fill_page_desc(skb, i, dfrag->page, offset, copy); + + if (dfrag->ubuf && !skb_zcopy(skb)) { + skb_zcopy_set(skb, dfrag->ubuf, NULL); + skb_shinfo(skb)->flags |=3D SKBFL_PURE_ZEROCOPY; + } } =20 skb->len +=3D copy; skb->data_len +=3D copy; skb->truesize +=3D copy; sk_wmem_queued_add(ssk, copy); - sk_mem_charge(ssk, copy); + if (!skb_zcopy_pure(skb)) + sk_mem_charge(ssk, copy); WRITE_ONCE(tp->write_seq, tp->write_seq + copy); TCP_SKB_CB(skb)->end_seq +=3D copy; tcp_skb_pcount_set(skb, 0); @@ -1969,22 +1995,81 @@ static void mptcp_rps_record_subflows(const struct = mptcp_sock *msk) } } =20 +static int mptcp_sendmsg_zerocopy_iter(struct sock *sk, struct msghdr *msg, + struct ubuf_info *ubuf, u32 copy_limit, + ssize_t *copied, long *timeo) +{ + struct mptcp_sock *msk =3D mptcp_sk(sk); + struct mptcp_data_frag *dfrag; + struct page *pages[1] =3D {}; + size_t pg_off =3D 0; + ssize_t pg_len; + + pg_len =3D iov_iter_get_pages2(&msg->msg_iter, pages, + copy_limit, 1, &pg_off); + if (pg_len < 0) + return pg_len; + /* pg_len =3D=3D 0 means the iterator is exhausted */ + if (pg_len =3D=3D 0) + return -EFAULT; + + dfrag =3D kzalloc_obj(*dfrag, GFP_KERNEL_ACCOUNT); + if (!dfrag) { + iov_iter_revert(&msg->msg_iter, pg_len); + put_page(pages[0]); + return -ENOMEM; + } + dfrag->data_len =3D pg_len; + dfrag->data_seq =3D msk->write_seq; + dfrag->offset =3D pg_off; + dfrag->page =3D pages[0]; + dfrag->ubuf =3D ubuf; + dfrag->overhead =3D sizeof(struct mptcp_data_frag); + + /* one ref for the dfrag; released in dfrag_clear */ + net_zcopy_get(ubuf); + + *copied +=3D pg_len; + WRITE_ONCE(msk->write_seq, msk->write_seq + pg_len); + sk_wmem_queued_add(sk, pg_len + dfrag->overhead); + + list_add_tail(&dfrag->list, &msk->rtx_queue); + if (!msk->first_pending) + msk->first_pending =3D dfrag; + + return 0; +} + static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len) { struct mptcp_sock *msk =3D mptcp_sk(sk); + struct ubuf_info *ubuf =3D NULL; struct page_frag *pfrag; size_t copied =3D 0; + bool zc =3D false; int ret =3D 0; long timeo; =20 /* silently ignore everything else */ msg->msg_flags &=3D MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL | - MSG_FASTOPEN | MSG_EOR; + MSG_FASTOPEN | MSG_EOR | MSG_ZEROCOPY; =20 lock_sock(sk); =20 mptcp_rps_record_subflows(msk); =20 + if (msg->msg_flags & MSG_FASTOPEN || !sock_flag(sk, SOCK_ZEROCOPY)) + msg->msg_flags &=3D ~MSG_ZEROCOPY; + + if ((msg->msg_flags & MSG_ZEROCOPY) && len) { + ubuf =3D msg_zerocopy_realloc(sk, len, NULL, false); + if (!ubuf) { + ret =3D -ENOBUFS; + goto do_error; + } + zc =3D true; + } + if (unlikely(inet_test_bit(DEFER_CONNECT, sk) || msg->msg_flags & MSG_FASTOPEN)) { int copied_syn =3D 0; @@ -2023,6 +2108,15 @@ static int mptcp_sendmsg(struct sock *sk, struct msg= hdr *msg, size_t len) if (!copy_limit) goto wait_for_memory; =20 + if (zc) { + ret =3D mptcp_sendmsg_zerocopy_iter(sk, msg, + ubuf, copy_limit, + &copied, &timeo); + if (ret) + goto do_error; + continue; + } + /* reuse tail pfrag, if possible, or carve a new one from the * page allocator */ @@ -2098,6 +2192,8 @@ static int mptcp_sendmsg(struct sock *sk, struct msgh= dr *msg, size_t len) } =20 out: + if (zc) + net_zcopy_put(ubuf); release_sock(sk); return copied; =20 @@ -2106,6 +2202,10 @@ static int mptcp_sendmsg(struct sock *sk, struct msg= hdr *msg, size_t len) goto out; =20 copied =3D sk_stream_error(sk, msg->msg_flags, ret); + if (zc) { + net_zcopy_put_abort(ubuf, true); + zc =3D false; + } goto out; } =20 diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h index da40c6f3705f..145c0f1a0007 100644 --- a/net/mptcp/protocol.h +++ b/net/mptcp/protocol.h @@ -262,12 +262,13 @@ struct mptcp_pm_addr_entry { struct mptcp_data_frag { struct list_head list; u64 data_seq; - u16 data_len; + u32 data_len; /* u16 truncates on 64KB-page kernels */ u16 offset; u8 overhead; u8 eor; /* currently using 1 bit */ u16 already_sent; struct page *page; + struct ubuf_info *ubuf; }; =20 /* Arbitrary compromise between as low as possible to react timely to subf= low --=20 2.53.0 From nobody Sat Jul 25 07:27:27 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (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 C4B662931D7 for ; Wed, 15 Jul 2026 11:04:08 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113449; cv=none; b=CASaoj7Sak3Slyu2yP5XkcIxU0P2ZL8AqEtGjGRO4JLeiETh0jWyFllBmuKf1ta75NY574a4hgBGKkYIDcYsaBKh7yvBgzC+73NVQf80KsdAaITGdopfrcCupZR38mhhup3FT4k0N4qwV1+P9s9lAt4T/ZQ5AEHquhbRj3wWCEI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113449; c=relaxed/simple; bh=h0KHbTdRLXfAQDlm+Mzn4sOSwJyxsp1fPSLu/6K0DEU=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=CeJbBeVNKpAJ3a7WYHPmMl1eFtyXZ2GCpYpYQezKzWSwHH2Ge7xf/uQLA77puyzXBjZUk788N2eV2rkh6JHZevuMhRd7BDh8LdX8qm+GfbSPmgzEfx9cZ7vs2Vsgj4IrYcAqMa9uJ+8XdttTFVnKpeYn7h9dNLTUpnVAlLuCyMo= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=dtNSLf5k; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="dtNSLf5k" Received: by smtp.kernel.org (Postfix) with ESMTPSA id DFC071F00A3A; Wed, 15 Jul 2026 11:04:07 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1784113448; bh=1aL2W8PM3bCKMiZfqkj5j62ROsq0aloJ4VBeqLorDFw=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=dtNSLf5kUYQgvfP7OSHENUtx54WtgU46sjlw4MbIkU3/tJdK95T5HM7Yk+Z2C41k5 ME3G7Mg0ydmc6WN05hn7uUsOSBD8gaQmJqiVlyrZQhPj4zeADLolY2AIlkDDT/31Kj CMliAcpRlyC5hLUG+70WkdM7DjYDn5MYxd6Ud5cLwZYWHYHj6azWd0F+QGXkvphFPk C3ClNz45lO7o1Kumuw/g+g5dEqHkuSC4OVT6EQ1QSjXfLIQs5rYo67bIgW1APap1T3 1HnwrDFWHk+r2r6T/UBseLAT3kDlP5iEoMra+mFGyetgbDdOCyqAW8sDvrilMch1ZM c5Mw+ndzg3A0g== From: Geliang Tang To: mptcp@lists.linux.dev Cc: Geliang Tang Subject: [PATCH mptcp-next v5 5/8] mptcp: handle SO_ZEROCOPY in setsockopt Date: Wed, 15 Jul 2026 19:03:50 +0800 Message-ID: X-Mailer: git-send-email 2.53.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Geliang Tang SO_ZEROCOPY is currently rejected by MPTCP setsockopt with -EOPNOTSUPP, even though the matching MSG_ZEROCOPY path in mptcp_sendmsg() is wired up. As a result MSG_ZEROCOPY silently falls back to a copy on MPTCP sockets because the opt-in flag is never set. Expose SO_ZEROCOPY through the SOL_SOCKET setsockopt dispatcher. The new mptcp_setsockopt_sol_socket_zerocopy() helper validates that val is 0 or 1 (matching standard TCP sk_setsockopt) and flips SOCK_ZEROCOPY on the msk only. mptcp_sendmsg() consults the same flag directly, so opt-in is honoured on every data path without any cross-socket propagation. Signed-off-by: Geliang Tang --- net/mptcp/sockopt.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/net/mptcp/sockopt.c b/net/mptcp/sockopt.c index fcf6feb2a9eb..abcc7d4452b4 100644 --- a/net/mptcp/sockopt.c +++ b/net/mptcp/sockopt.c @@ -175,6 +175,19 @@ static int mptcp_setsockopt_sol_socket_tstamp(struct m= ptcp_sock *msk, int optnam return 0; } =20 +static int mptcp_setsockopt_sol_socket_zerocopy(struct mptcp_sock *msk, in= t val) +{ + struct sock *sk =3D (struct sock *)msk; + + if (val !=3D 0 && val !=3D 1) + return -EINVAL; + + lock_sock(sk); + sock_valbool_flag(sk, SOCK_ZEROCOPY, val); + release_sock(sk); + return 0; +} + static int mptcp_setsockopt_sol_socket_int(struct mptcp_sock *msk, int opt= name, sockptr_t optval, unsigned int optlen) @@ -203,6 +216,8 @@ static int mptcp_setsockopt_sol_socket_int(struct mptcp= _sock *msk, int optname, case SO_TIMESTAMPNS_OLD: case SO_TIMESTAMPNS_NEW: return mptcp_setsockopt_sol_socket_tstamp(msk, optname, val); + case SO_ZEROCOPY: + return mptcp_setsockopt_sol_socket_zerocopy(msk, val); } =20 return -ENOPROTOOPT; @@ -342,6 +357,7 @@ static int mptcp_setsockopt_sol_socket(struct mptcp_soc= k *msk, int optname, case SO_TIMESTAMP_NEW: case SO_TIMESTAMPNS_OLD: case SO_TIMESTAMPNS_NEW: + case SO_ZEROCOPY: return mptcp_setsockopt_sol_socket_int(msk, optname, optval, optlen); case SO_TIMESTAMPING_OLD: @@ -387,7 +403,6 @@ static int mptcp_setsockopt_sol_socket(struct mptcp_soc= k *msk, int optname, * SO_CNX_ADVICE is currently unsupported, could possibly be relevant, * but likely needs careful design * - * SO_ZEROCOPY is currently unsupported, TODO in sndmsg * SO_TXTIME is currently unsupported */ =20 --=20 2.53.0 From nobody Sat Jul 25 07:27:27 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (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 DB9632931D7 for ; Wed, 15 Jul 2026 11:04:10 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113452; cv=none; b=Ig1kysh6dBtr0GmydgnExuvF9a8gR2PIJfjCA70FJrTi5F6PCL5gEBKhsA0kNIg6VXiA/Z/5OkO84FB2xak9ttzSbwGfqlf4N+H3PKaCiTDnZmTTAQEIxUMNsoL+b3gIZgcFeuRWriwZsiGFGpkMvC7htZApbJsfozhlIBj0330= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784113452; c=relaxed/simple; bh=h025n3Wk0LwTydM3OCfHMUkLdzBmNNCOsfttMCMbrsM=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=YPxY4bFGnBiQzPdFrTo7ZRJ3V0ml4yUyUE1lY/v5hl4vkwxzN8DuIQcCPswXvaF1urBvJ1CIyukyxZ/xeoXcEPZ8R7Yqx7QuOtzwCFxD+94NtWIFBdZTr/Tml6EB7NXc795QSO2vlFcQ0e4c0TwD/xwDmoyTdI0q9djAzkTgwlM= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=oHF13Uoo; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="oHF13Uoo" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 409C91F000E9; Wed, 15 Jul 2026 11:04:09 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1784113450; bh=tiDZpWkvrhgwsDO/miQ045VyimJeEcIeLwgrT2qidqg=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=oHF13UooKLb/kYYOgNAvPP2elx6ZLAdYI0VF2Sf/1CqZWI9GbVgBDu1WK9q/6OQDw HCQhht50PtBeSXNK8e+5BXAGRu3PKJsK+KEt1P6e4l3xSx6NwxMydL85sk+8CLx7uu uPXWP9quWzCOx9vWpGhrsu6rvQymNVOpYPRNZmvyFxROFm5hyibwgJe8XRT6NSvA+K tMaY/whbtqh1Z5AmtwxNQBQ+WFHA7NEo0nxN8ZO9RX9TH2FzU5D+0t9lK4mVghDxCE bq/sJus4el1/AjjToDNyoRKo4fofAlpuu/ud+nxQulCxgHTOv/xe9jLajqUDmtke7i CBY5XtM/6mUbw== From: Geliang Tang To: mptcp@lists.linux.dev Cc: Geliang Tang Subject: [PATCH mptcp-next v5 6/8] selftests: mptcp: connect: add zerocopy io mode Date: Wed, 15 Jul 2026 19:03:51 +0800 Message-ID: X-Mailer: git-send-email 2.53.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Geliang Tang Add a new I/O mode 'zerocopy' to the mptcp_connect selftest, which uses sendmsg() with the MSG_ZEROCOPY flag to transmit data. This enables testing and performance validation of the zero-copy send path for both TCP and MPTCP connections. This mode is useful for benchmarking and for verifying correctness of MSG_ZEROCOPY handling in the MPTCP stack. The completion notification draining uses batched polling to handle large transfers that may generate multiple completion events across different TCP segments. Signed-off-by: Geliang Tang --- .../selftests/net/mptcp/mptcp_connect.c | 122 +++++++++++++++++- 1 file changed, 121 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c b/tools/test= ing/selftests/net/mptcp/mptcp_connect.c index ea4cb6c1bd5e..4670c194672d 100644 --- a/tools/testing/selftests/net/mptcp/mptcp_connect.c +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c @@ -33,6 +33,7 @@ #include #include #include +#include #include =20 extern int optind; @@ -53,6 +54,7 @@ enum cfg_mode { CFG_MODE_MMAP, CFG_MODE_SENDFILE, CFG_MODE_SPLICE, + CFG_MODE_ZEROCOPY, }; =20 enum cfg_peek { @@ -125,7 +127,7 @@ static void die_usage(void) fprintf(stderr, "\t-j -- add additional sleep at connection start and= tear down " "-- for MPJ tests\n"); fprintf(stderr, "\t-l -- listens mode, accepts incoming connection\n"= ); - fprintf(stderr, "\t-m [poll|mmap|sendfile|splice] -- use poll(default)/mm= ap+write/sendfile/splice\n"); + fprintf(stderr, "\t-m [poll|mmap|sendfile|splice|zerocopy] -- use poll(de= fault)/mmap+write/sendfile/splice/zerocopy\n"); fprintf(stderr, "\t-M mark -- set socket packet mark\n"); fprintf(stderr, "\t-o option -- test sockopt