From nobody Tue Dec 2 02:42:39 2025 Received: from mail-pj1-f52.google.com (mail-pj1-f52.google.com [209.85.216.52]) (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 4F6C93559FE for ; Tue, 18 Nov 2025 11:32:19 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.216.52 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1763465541; cv=none; b=EboJQHU0MnukxjDcYpR2wZa0nvb/MDp5VCOD2SnnTGV3lsQJZQsGTTFDv8JvoHRHJ7vdxsE/CT81zIu6FEA0zm+plhvTEv9CvHa7EwXp7tbsjJ/xLNXooI3qmYepCe803UfPqYlz2iOS+J7C0TBshXoH6CDobnJKrn/NUJeEwU0= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1763465541; c=relaxed/simple; bh=K8Wyd2To1BmAKFVUmUbQOinghXJ6DrKZBtPEPb4ilFQ=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=JBjfs1iIGtghlZ6Q85H808ZmJt2BooXEUAGHgE0zqCWXJ+6xTMsNzZvuj0qzzwH4gzztSQSHkQTwn1rWZrjgkfc+amkteRupgSHb4O0T99pTm5kDIS357uNizj9QEzrt5N8tN7Yqzs9XfRtPiLcugA2TjTVCjVKkplqCr0gd/gY= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=DyUt7bx2; arc=none smtp.client-ip=209.85.216.52 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="DyUt7bx2" Received: by mail-pj1-f52.google.com with SMTP id 98e67ed59e1d1-340ba29d518so3786734a91.3 for ; Tue, 18 Nov 2025 03:32:19 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1763465539; x=1764070339; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=U36EJ4GyvUj1c+KWVQBGEQGDIxcP9pTwbj0Nc88XDlE=; b=DyUt7bx2GXMi8+9f09ff1Db3Ptbgbilvu41PK7cNugTb8O4Slv7rpO18YVhSWSs/3b w1Vc8G8u/PofGGeQE1W+zJpZz4ddZU07S2XvTHhGf6ZNW2l1F0AWvFJZmtYZCLnu+42Y 1lgAmniLcS41LWTB400k3bSYfqredbE/aPdXyiURjtXadGgxay5k4IC6ybuXfZa/qTV4 y/cv2MN+uODKjX7uZnmpJZR+diqThGndYdnISGWCQ3q5MdA7r+gSU3lx7n5RB2/1Hra2 fIf6Db3txOL1jRjGBL9cnGCNNyUlLZJ+y1e2oOz51JZO8lZOfm6spwib89uDMP5sVWj7 9BzA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1763465539; x=1764070339; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-gg:x-gm-message-state:from :to:cc:subject:date:message-id:reply-to; bh=U36EJ4GyvUj1c+KWVQBGEQGDIxcP9pTwbj0Nc88XDlE=; b=OE/QldGQFoOw0D2Dv/9Mfam8+vOjwnWNadlJDoUC58QM2dpN6taPD2Exn5rOUk5FT8 6O9VaY7xn12zN+E7sOmFYIBR9/CO6NzrO0J8rT8+40KDnorV1+PF/N/2hdAwiw3/2oa5 b6nglpnjzfLRX3QOK/6mJ0nrsvLMDmFSSWBdcOx+jQdEYIwC2o9FHeYiXZL9OvkEm2YU j/2oWGisK4tK1AJ6bRR8BSr6pzzY4+Oj0zP2owklfndRunRxKjoPcGDtpl4dIIH4s28j k+BjxWlBwpmE/5mWgzHyTfQKBfj1S/4KIAknaLMc/YRWdMlRvMHUyx2RG+x7n9ob+m9m nPwQ== X-Forwarded-Encrypted: i=1; AJvYcCVn2jE0A49Gmriu+8FELrOzmP3Qlokq92/gFwX+qtycRkkwxUBibk0ksepBSRFO/MnQFp/iFdr1xwAZJXc=@vger.kernel.org X-Gm-Message-State: AOJu0YyK0RuahmWJ+oqntM07mjy4rPBRmDM1HKEBVBAyLHo3+12uGEMZ rE1joGavB392QgzssNBKe8FXLNgCnRKEBJr1FZImTbxDLtrH9p6Vh/7I X-Gm-Gg: ASbGncv67uJgyl2+xLk4+iuBJFvYgn65QzImkrOew/QxHE8IFSK/841EiwAi6u34eph ClNhwCudHZgmwmgf2wch1zyrqKC7qnKsxbcv617VgAt3Pck+jSMFo70Vdn/s8tRiEXGHL4y19t3 Eog2feYJNHdfbyKAik1H62P5AuFsh0c55QCmT4W3XAiOTAseW2P3jtmfurDb9CKUl749dIC7ExP ttfJFc8aVu7/skPFeWstD7rFQyoc88/3Nu2NLahkjd+8c65Q7bAnsRp15s7QrbtFsaaamXiok5p OoJEpaZxJd3RjuE2oA0Emm3CxDi7oGq3VszAJ4XX9INwgGrR0G0t30E8EGhVHcIULGBrzfYZMn1 hcrVeX+oeZ1OHDBJ5ggrwqZKeLIKuBn92vbfyob4KekGAz8a/Vy8Dk5WNcsI02j+DrFI7X93WVE CW5wicDyRzTL+F5Acj0ElWMziOm2L7J6ZStqLSqTPiDnsXz5m0FQ== X-Google-Smtp-Source: AGHT+IEXgXhQS/PPrYJbpAo06A+LbwNc11z1gcGQgtP5bDqE6cvj/3xMoRUdO7/TieS5wkj2K+854Q== X-Received: by 2002:a17:90b:2552:b0:340:ec8f:82d8 with SMTP id 98e67ed59e1d1-343f9ea40c1mr19277977a91.12.1763465538566; Tue, 18 Nov 2025 03:32:18 -0800 (PST) Received: from localhost.localdomain ([103.246.102.164]) by smtp.gmail.com with ESMTPSA id 98e67ed59e1d1-3456516d4a9sm12583736a91.11.2025.11.18.03.32.12 (version=TLS1_3 cipher=TLS_CHACHA20_POLY1305_SHA256 bits=256/256); Tue, 18 Nov 2025 03:32:18 -0800 (PST) From: Alessandro Decina To: netdev@vger.kernel.org Cc: Maciej Fijalkowski , "David S. Miller" , Alexei Starovoitov , Andrew Lunn , Daniel Borkmann , Eric Dumazet , Jakub Kicinski , Jesper Dangaard Brouer , John Fastabend , Paolo Abeni , Przemek Kitszel , Stanislav Fomichev , Tirthendu Sarkar , Tony Nguyen , bpf@vger.kernel.org, intel-wired-lan@lists.osuosl.org, linux-kernel@vger.kernel.org, Alessandro Decina Subject: [PATCH net v4 1/1] i40e: xsk: advance next_to_clean on status descriptors Date: Tue, 18 Nov 2025 22:31:17 +1100 Message-Id: <20251118113117.11567-2-alessandro.d@gmail.com> X-Mailer: git-send-email 2.39.3 (Apple Git-146) In-Reply-To: <20251118113117.11567-1-alessandro.d@gmail.com> References: <20251118113117.11567-1-alessandro.d@gmail.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Whenever a status descriptor is received, i40e processes and skips over it, correctly updating next_to_process but forgetting to update next_to_clean. In the next iteration this accidentally causes the creation of an invalid multi-buffer xdp_buff where the first fragment is the status descriptor. If then a skb is constructed from such an invalid buffer - because the eBPF program returns XDP_PASS - a panic occurs: [ 5866.367317] BUG: unable to handle page fault for address: ffd31c37eab1c9= 80 [ 5866.375050] #PF: supervisor read access in kernel mode [ 5866.380825] #PF: error_code(0x0000) - not-present page [ 5866.386602] PGD 0 [ 5866.388867] Oops: Oops: 0000 [#1] SMP NOPTI [ 5866.393575] CPU: 34 UID: 0 PID: 0 Comm: swapper/34 Not tainted 6.17.0-cu= stom #1 PREEMPT(voluntary) [ 5866.403740] Hardware name: Supermicro AS -2115GT-HNTR/H13SST-G, BIOS 3.2= 03/20/2025 [ 5866.412339] RIP: 0010:memcpy+0x8/0x10 [ 5866.416454] Code: cc cc 90 cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc = 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 66 90 48 89 f8 48 89 d1 a4 e9 fc 26 c0 fe 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 [ 5866.437538] RSP: 0018:ff428d9ec0bb0ca8 EFLAGS: 00010286 [ 5866.443415] RAX: ff2dd26dbd8f0000 RBX: ff2dd265ad161400 RCX: 00000000000= 004e1 [ 5866.451435] RDX: 00000000000004e1 RSI: ffd31c37eab1c980 RDI: ff2dd26dbd8= f0000 [ 5866.459454] RBP: ff428d9ec0bb0d40 R08: 0000000000000000 R09: 00000000000= 00000 [ 5866.467470] R10: 0000000000000000 R11: 0000000000000000 R12: ff428d9eec7= 26ef8 [ 5866.475490] R13: ff2dd26dbd8f0000 R14: ff2dd265ca2f9fc0 R15: ff2dd265485= 48b80 [ 5866.483509] FS: 0000000000000000(0000) GS:ff2dd2c363592000(0000) knlGS:= 0000000000000000 [ 5866.492600] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 5866.499060] CR2: ffd31c37eab1c980 CR3: 0000000178d7b040 CR4: 0000000000f= 71ef0 [ 5866.507079] PKRU: 55555554 [ 5866.510125] Call Trace: [ 5866.512867] [ 5866.515132] ? i40e_clean_rx_irq_zc+0xc50/0xe60 [i40e] [ 5866.520921] i40e_napi_poll+0x2d8/0x1890 [i40e] [ 5866.526022] ? srso_alias_return_thunk+0x5/0xfbef5 [ 5866.531408] ? raise_softirq+0x24/0x70 [ 5866.535623] ? srso_alias_return_thunk+0x5/0xfbef5 [ 5866.541011] ? srso_alias_return_thunk+0x5/0xfbef5 [ 5866.546397] ? rcu_sched_clock_irq+0x225/0x1800 [ 5866.551493] __napi_poll+0x30/0x230 [ 5866.555423] net_rx_action+0x20b/0x3f0 [ 5866.559643] handle_softirqs+0xe4/0x340 [ 5866.563962] __irq_exit_rcu+0x10e/0x130 [ 5866.568283] irq_exit_rcu+0xe/0x20 [ 5866.572110] common_interrupt+0xb6/0xe0 [ 5866.576425] [ 5866.578791] Advance next_to_clean to ensure invalid xdp_buff(s) aren't created. Move the common logic to i40e_clean_programming_status and update both i40e_clean_rx_irq and i40e_clean_rx_irq_zc accordingly. Fixes: 1c9ba9c14658 ("i40e: xsk: add RX multi-buffer support") Signed-off-by: Alessandro Decina --- drivers/net/ethernet/intel/i40e/i40e_txrx.c | 59 +++++++++++-------- .../ethernet/intel/i40e/i40e_txrx_common.h | 5 +- drivers/net/ethernet/intel/i40e/i40e_xsk.c | 14 ++--- 3 files changed, 42 insertions(+), 36 deletions(-) diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethe= rnet/intel/i40e/i40e_txrx.c index cc0b9efc2637..fe2190f4b9bc 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c @@ -1386,22 +1386,36 @@ static void i40e_reuse_rx_page(struct i40e_ring *rx= _ring, * @rx_ring: the rx ring that has this descriptor * @qword0_raw: qword0 * @qword1: qword1 representing status_error_len in CPU ordering + * @next_to_process: pointer to next_to_process index + * @next_to_clean: pointer to next_to_clean index * * Flow director should handle FD_FILTER_STATUS to check its filter progra= mming * status being successful or not and take actions accordingly. FCoE should * handle its context/filter programming/invalidation status and take acti= ons. * - * Returns an i40e_rx_buffer to reuse if the cleanup occurred, otherwise N= ULL. + * Returns false if what is passed is not a status descriptor. **/ -void i40e_clean_programming_status(struct i40e_ring *rx_ring, u64 qword0_r= aw, - u64 qword1) +bool i40e_clean_programming_status(struct i40e_ring *rx_ring, u64 qword0_r= aw, + u64 qword1, u16 *next_to_process, + u16 *next_to_clean) { + u16 ntp =3D *next_to_process; u8 id; =20 + if (!i40e_rx_is_programming_status(qword1)) + return false; + id =3D FIELD_GET(I40E_RX_PROG_STATUS_DESC_QW1_PROGID_MASK, qword1); =20 if (id =3D=3D I40E_RX_PROG_STATUS_DESC_FD_FILTER_STATUS) i40e_fd_handle_status(rx_ring, qword0_raw, qword1, id); + + if (++*next_to_process =3D=3D rx_ring->count) + *next_to_process =3D 0; + if (ntp =3D=3D *next_to_clean) + *next_to_clean =3D *next_to_process; + + return true; } =20 /** @@ -1971,19 +1985,18 @@ static void i40e_rx_buffer_flip(struct i40e_rx_buff= er *rx_buffer, } =20 /** - * i40e_get_rx_buffer - Fetch Rx buffer and synchronize data for use + * i40e_prepare_rx_buffer - Synchronize the buffer for use by the CPU * @rx_ring: rx descriptor ring to transact packets on + * @rx_buffer: the rx buffer * @size: size of buffer to add to skb * - * This function will pull an Rx buffer from the ring and synchronize it - * for use by the CPU. + * This function will synchronize the given buffer for use by the CPU. */ -static struct i40e_rx_buffer *i40e_get_rx_buffer(struct i40e_ring *rx_ring, - const unsigned int size) +static struct i40e_rx_buffer * +i40e_prepare_rx_buffer(struct i40e_ring *rx_ring, + struct i40e_rx_buffer *rx_buffer, + const unsigned int size) { - struct i40e_rx_buffer *rx_buffer; - - rx_buffer =3D i40e_rx_bi(rx_ring, rx_ring->next_to_process); rx_buffer->page_count =3D #if (PAGE_SIZE < 8192) page_count(rx_buffer->page); @@ -2450,6 +2463,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_rin= g, int budget, =20 while (likely(total_rx_packets < (unsigned int)budget)) { u16 ntp =3D rx_ring->next_to_process; + u16 ntc =3D rx_ring->next_to_clean; struct i40e_rx_buffer *rx_buffer; union i40e_rx_desc *rx_desc; struct sk_buff *skb; @@ -2480,21 +2494,15 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_r= ing, int budget, */ dma_rmb(); =20 - if (i40e_rx_is_programming_status(qword)) { - i40e_clean_programming_status(rx_ring, - rx_desc->raw.qword[0], - qword); - rx_buffer =3D i40e_rx_bi(rx_ring, ntp); - i40e_inc_ntp(rx_ring); + rx_buffer =3D i40e_rx_bi(rx_ring, ntp); + + if (i40e_clean_programming_status(rx_ring, + rx_desc->raw.qword[0], qword, + &rx_ring->next_to_process, + &rx_ring->next_to_clean)) { i40e_reuse_rx_page(rx_ring, rx_buffer); - /* Update ntc and bump cleaned count if not in the - * middle of mb packet. - */ - if (rx_ring->next_to_clean =3D=3D ntp) { - rx_ring->next_to_clean =3D - rx_ring->next_to_process; + if (ntc !=3D rx_ring->next_to_clean) cleaned_count++; - } continue; } =20 @@ -2503,8 +2511,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_rin= g, int budget, break; =20 i40e_trace(clean_rx_irq, rx_ring, rx_desc, xdp); - /* retrieve a buffer from the ring */ - rx_buffer =3D i40e_get_rx_buffer(rx_ring, size); + i40e_prepare_rx_buffer(rx_ring, rx_buffer, size); =20 neop =3D i40e_is_non_eop(rx_ring, rx_desc); i40e_inc_ntp(rx_ring); diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx_common.h b/drivers/n= et/ethernet/intel/i40e/i40e_txrx_common.h index e26807fd2123..21d9ed878bf0 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx_common.h +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx_common.h @@ -7,8 +7,9 @@ #include "i40e.h" =20 int i40e_xmit_xdp_tx_ring(struct xdp_buff *xdp, struct i40e_ring *xdp_ring= ); -void i40e_clean_programming_status(struct i40e_ring *rx_ring, u64 qword0_r= aw, - u64 qword1); +bool i40e_clean_programming_status(struct i40e_ring *rx_ring, u64 qword0_r= aw, + u64 qword1, u16 *next_to_clean, + u16 *next_to_process); void i40e_process_skb_fields(struct i40e_ring *rx_ring, union i40e_rx_desc *rx_desc, struct sk_buff *skb); void i40e_xdp_ring_update_tail(struct i40e_ring *xdp_ring); diff --git a/drivers/net/ethernet/intel/i40e/i40e_xsk.c b/drivers/net/ether= net/intel/i40e/i40e_xsk.c index 9f47388eaba5..f8accc266c2c 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_xsk.c +++ b/drivers/net/ethernet/intel/i40e/i40e_xsk.c @@ -440,14 +440,13 @@ int i40e_clean_rx_irq_zc(struct i40e_ring *rx_ring, i= nt budget) */ dma_rmb(); =20 - if (i40e_rx_is_programming_status(qword)) { - i40e_clean_programming_status(rx_ring, - rx_desc->raw.qword[0], - qword); - bi =3D *i40e_rx_bi(rx_ring, next_to_process); + bi =3D *i40e_rx_bi(rx_ring, next_to_process); + + if (i40e_clean_programming_status(rx_ring, + rx_desc->raw.qword[0], + qword, &next_to_process, + &next_to_clean)) { xsk_buff_free(bi); - if (++next_to_process =3D=3D count) - next_to_process =3D 0; continue; } =20 @@ -455,7 +454,6 @@ int i40e_clean_rx_irq_zc(struct i40e_ring *rx_ring, int= budget) if (!size) break; =20 - bi =3D *i40e_rx_bi(rx_ring, next_to_process); xsk_buff_set_size(bi, size); xsk_buff_dma_sync_for_cpu(bi); =20 --=20 2.43.0