From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-qt1-f175.google.com (mail-qt1-f175.google.com [209.85.160.175]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 4EFA1313E38 for ; Tue, 21 Jul 2026 01:58:30 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.160.175 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784599111; cv=none; b=ZTjyvumFEyzR3yM8QDhLFQeH7zMeIIDtWTCCLf0FMresKaXWJaHpJmm1x3XuTJKLgImkkKOiCFt8KE4LNptDUqcp3jIkZrYdi8DUU3LgUQ/h4JDFmuh+9MYbwgMnY5+r8Ehv4obYj3CHd3mu4+Xuw8xlEIWsI2gxZTz/Fa53Hno= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784599111; c=relaxed/simple; bh=kfcA0jtQM2vxhoUMmTZaEx+SFxxAGDbBXmXwON70ieU=; h=From:To:Cc:Subject:Date:Message-ID:MIME-Version; b=MtpQahSGiOgRfxOGJdUjaoCteWbcKR6WhirlOlu6zoj4ep7Lor+/TVR+aHgfSy8OEGOPVX2uYyd086R30ig8+JglqXRe88cGJTqIAlfUSSnhbsJn1eRWTlN1aevdmPQZ3GCrnffBgYOsNsWvsDkSxT20a5c5Rt/qNNYfBb305tE= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=openai.com; spf=pass smtp.mailfrom=openai.com; dkim=pass (1024-bit key) header.d=openai.com header.i=@openai.com header.b=N5G03A+f; arc=none smtp.client-ip=209.85.160.175 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=openai.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=openai.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=openai.com header.i=@openai.com header.b="N5G03A+f" Received: by mail-qt1-f175.google.com with SMTP id d75a77b69052e-51c2a818fc4so91720981cf.2 for ; Mon, 20 Jul 2026 18:58:30 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=openai.com; s=google; t=1784599109; x=1785203909; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:from:to:cc:subject:date:message-id:reply-to:content-type; bh=gvfid7rw/l1nHM0s2D/GVV5QJykAVKQqpYZKzlHL3SE=; b=N5G03A+fG9/GlYMdhKRuHKgKKLXQZaAboT50X/Rzcq9c45wDBMVflYY/k6jTQGB59t 8oX408a2wmfvjcJ21X9q/RODA2skyu77M/2YKBnQv434bvd3bI3bV5zMmE+EPFr5uSuI OS5o4QOlNKypOD7Rdj4ZlZAsrVXwKET8i93UE= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1784599109; x=1785203909; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:x-gm-gg:x-gm-message-state:from:to:cc:subject:date :message-id:reply-to:content-type; bh=gvfid7rw/l1nHM0s2D/GVV5QJykAVKQqpYZKzlHL3SE=; b=aCjz5BvhlKruWNX9XyhLOgX3jHhHIeEHl5P+GoZQmFvdq920TMPSJYFZmg1XSVHmUp eZ0ogsR9oUw4W0t0SS0BaYJF6zIH73nAPFonlKhkz4kCyBweubOQeevkuy4Wsv0L5WMd XqfyAJDsEJC4+aR5p0tRsBLRL990q8N6OzgPI3dhiBhG9roXlTvkUKq3ZE52tFEORY4g gKONpbZuPVUwjyHawTJaAz6v+HkQik33f4H7GbCcBAL4WnRQdDekAsTdhBUjw/o/P1k0 oHSuuAB1J/COCDZ8ao07H8BDuSkX08JVu9/Px0k8jyCI60mreClw0thheZsX4mqUe4xT LrYA== X-Forwarded-Encrypted: i=1; AHgh+RrEwF+3iQBqSAhocUISGeNO2U4kxkWgthn9XMpvdEtM5fgZgdbaXaTqfMdS7NNGJ/fiPHOgNGs=@vger.kernel.org X-Gm-Message-State: AOJu0YySGt1tGd7sHtI9gEHHunYLSql1wXqjQXLEEkP51RHe1JBRoAlc QUMKldlFVnHRchwF5H3F0QtEsUXHotI0lnv2lkFp0e6KORWtQIOB43/0PO4q6TvuBsI= X-Gm-Gg: AfdE7cknn0Z/ZEc+OhTf4dMVh0BqKX5Hu0VjrEC6hXpSDosIn7qBH5KRxLequ8Tgaxb OSaxraRBuD+Z+apyuqaGJQTWHDgGwS/1DpXdwB5+CKl3pVKB+h54bmyDfd47tBqxX2qFRF1V9nN e4dOAenirSGV22BrkdSFkwJJIjrSzKzFB21KtvUp0HaEKNcJ1NAFVoGS49SL97DBXiVoYAJAe/Z CvVoKoE36qoK3lRCNDOPnCFshx7TMqPB9UPheZenL1moT2NcHoTIb3c98jl7Jk8P1ZemPflv3jL a2wl5HAfIATxh1SfqsTCtx++p+cQhFRz9MUAMsxRXUJN8Zp6XUw5og3wBKEgPP9i4kf+w8m+Est 2gTyzaMwPPZ7RW+MzG3poiEFLX2vvNxFb+N9ORZyZ7kq4N+Kvwq3UV9zotN0XG0g/ntik/pKRrU nzAs6hhv52nulpnuwU2sT2vgN71c7Ba8i6nupJeKoZwXi1zLd58sdi3/KUFchIaw5bqxbgNAWhA gxAqpDI9gGAsj5JxEA= X-Received: by 2002:ac8:7d0c:0:b0:51c:94f:8ce with SMTP id d75a77b69052e-5213c201a70mr153244191cf.34.1784599109046; Mon, 20 Jul 2026 18:58:29 -0700 (PDT) Received: from com-75606.corp.openai.org ([209.249.37.148]) by smtp.gmail.com with ESMTPSA id 6a1803df08f44-907785839f9sm103728686d6.18.2026.07.20.18.58.27 (version=TLS1_3 cipher=TLS_CHACHA20_POLY1305_SHA256 bits=256/256); Mon, 20 Jul 2026 18:58:28 -0700 (PDT) From: Kyle Zeng To: netdev@vger.kernel.org Cc: Jakub Kicinski , Eric Dumazet , "David S . Miller" , Willem de Bruijn , Kyle Zeng , stable@vger.kernel.org Subject: [PATCH net] net/packet: defer vmalloc TX_RING free until skbs finish Date: Mon, 20 Jul 2026 18:58:24 -0700 Message-ID: <20260721015824.45829-1-kylebot@openai.com> X-Mailer: git-send-email 2.54.0 Precedence: bulk X-Mailing-List: stable@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit AF_PACKET TX_RING skbs keep a raw pointer to their ring frame. The skb page references preserve page-backed ring blocks after pg_vec is freed, but they do not preserve a vmalloc mapping. tpacket_destruct_skb() currently drops the pending reference before writing the timestamp and TP_STATUS_AVAILABLE to the frame. Move the decrement after those stores. The smp_wmb() in __packet_set_status() orders the frame stores before the decrement. On socket close, scan every pg_vec entry because allocation can produce a mixture of page-backed and vmalloc-backed blocks. If any block is vmalloc-backed and TX skbs remain pending, defer the whole vector to system_long_wq. After pg_vec is detached, a late destructor can skip the pending decrement. Use socket write-memory accounting as the deferred lifetime gate instead: an skb remains charged through its final sock_wfree(), after all ring-frame accesses. The delayed work retains a socket reference and reschedules itself until no TX skbs remain. Fall back to a synchronous wait if the work allocation fails. Move pending_refcnt release to packet_sock_destruct() so late skb destructors and deferred cleanup can safely use it after packet_release(). Page-backed teardown remains synchronous, and no lock is added to the TX completion hot path. Fixes: b013840810c2 ("packet: use percpu mmap tx frame pending refcount") Cc: stable@vger.kernel.org Suggested-by: Willem de Bruijn Assisted-by: Codex:gpt-5.6 Signed-off-by: Kyle Zeng --- net/packet/af_packet.c | 90 ++++++++++++++++++++++++++++++++++++++---- 1 file changed, 82 insertions(+), 8 deletions(-) diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 8e6f3a734ba0..1a4c548c59ac 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -88,6 +88,7 @@ #include #include #include +#include #ifdef CONFIG_INET #include #endif @@ -1329,6 +1330,8 @@ static void packet_sock_destruct(struct sock *sk) WARN_ON(atomic_read(&sk->sk_rmem_alloc)); WARN_ON(refcount_read(&sk->sk_wmem_alloc)); + packet_free_pending(pkt_sk(sk)); + if (!sock_flag(sk, SOCK_DEAD)) { pr_err("Attempt to release alive packet socket: %p\n", sk); return; @@ -2516,14 +2519,12 @@ static void tpacket_destruct_skb(struct sk_buff *skb) __u32 ts; ph = skb_zcopy_get_nouarg(skb); - packet_dec_pending(&po->tx_ring); ts = __packet_set_timestamp(po, ph, skb); __packet_set_status(po, ph, TP_STATUS_AVAILABLE | ts); - + packet_dec_pending(&po->tx_ring); complete(&po->skb_completion); } - sock_wfree(skb); } @@ -3182,7 +3183,6 @@ static int packet_release(struct socket *sock) /* Purge queues */ skb_queue_purge(&sk->sk_receive_queue); - packet_free_pending(po); sock_put(sk); return 0; @@ -4345,8 +4345,16 @@ static const struct vm_operations_struct packet_mmap_ops = { .close = packet_mm_close, }; -static void free_pg_vec(struct pgv *pg_vec, unsigned int order, - unsigned int len) +struct packet_pg_vec_free { + struct delayed_work work; + struct sock *sk; + struct pgv *pg_vec; + unsigned int order; + unsigned int len; +}; + +static void __free_pg_vec(struct pgv *pg_vec, unsigned int order, + unsigned int len) { int i; @@ -4363,6 +4371,69 @@ static void free_pg_vec(struct pgv *pg_vec, unsigned int order, kfree(pg_vec); } +static void packet_wait_for_tx_skbs(struct packet_sock *po) +{ + while (sk_wmem_alloc_get(&po->sk)) + wait_for_completion_timeout(&po->skb_completion, 1); +} + +static void packet_free_pg_vec_work(struct work_struct *work) +{ + struct packet_pg_vec_free *deferred; + struct sock *sk; + + deferred = container_of(to_delayed_work(work), + struct packet_pg_vec_free, work); + sk = deferred->sk; + if (sk_wmem_alloc_get(sk)) { + queue_delayed_work(system_long_wq, &deferred->work, 1); + return; + } + + __free_pg_vec(deferred->pg_vec, deferred->order, deferred->len); + kfree(deferred); + sock_put(sk); +} + +static bool pg_vec_has_vmalloc(struct pgv *pg_vec, unsigned int len) +{ + int i; + + for (i = 0; i < len; i++) + if (pg_vec[i].buffer && is_vmalloc_addr(pg_vec[i].buffer)) + return true; + + return false; +} + +static void free_pg_vec(struct packet_sock *po, struct pgv *pg_vec, + unsigned int order, unsigned int len) +{ + struct packet_pg_vec_free *deferred; + + if (!pg_vec_has_vmalloc(pg_vec, len) || + !packet_read_pending(&po->tx_ring)) { + __free_pg_vec(pg_vec, order, len); + return; + } + + deferred = kmalloc(sizeof(*deferred), GFP_KERNEL); + if (!deferred) { + packet_wait_for_tx_skbs(po); + __free_pg_vec(pg_vec, order, len); + return; + } + + INIT_DELAYED_WORK(&deferred->work, packet_free_pg_vec_work); + deferred->sk = &po->sk; + deferred->pg_vec = pg_vec; + deferred->order = order; + deferred->len = len; + + sock_hold(deferred->sk); + queue_delayed_work(system_long_wq, &deferred->work, 0); +} + static char *alloc_one_pg_vec_page(unsigned long order) { char *buffer; @@ -4408,7 +4479,7 @@ static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order) return pg_vec; out_free_pgvec: - free_pg_vec(pg_vec, order, block_nr); + __free_pg_vec(pg_vec, order, block_nr); pg_vec = NULL; goto out; } @@ -4574,7 +4645,10 @@ static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u, out_free_pg_vec: if (pg_vec) { bitmap_free(rx_owner_map); - free_pg_vec(pg_vec, order, req->tp_block_nr); + if (tx_ring && closing) + free_pg_vec(po, pg_vec, order, req->tp_block_nr); + else + __free_pg_vec(pg_vec, order, req->tp_block_nr); } out: return err; -- 2.54.0