From patchwork Thu Apr 23 01:48:19 2026 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Yi Zhao X-Patchwork-Id: 86677 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from aws-us-west-2-korg-lkml-1.web.codeaurora.org (localhost.localdomain [127.0.0.1]) by smtp.lore.kernel.org (Postfix) with ESMTP id 465AAFAD3EE for ; Thu, 23 Apr 2026 01:48:48 +0000 (UTC) Received: from mx0a-0064b401.pphosted.com (mx0a-0064b401.pphosted.com [205.220.166.238]) by mx.groups.io with SMTP id smtpd.msgproc02-g2.3473.1776908920922290729 for ; Wed, 22 Apr 2026 18:48:40 -0700 Authentication-Results: mx.groups.io; dkim=fail reason="dkim: body hash did not verify" header.i=@windriver.com header.s=PPS06212021 header.b=cSPZBW4C; spf=permerror, err=parse error for token &{10 18 %{ir}.%{v}.%{d}.spf.has.pphosted.com}: invalid domain name (domain: windriver.com, ip: 205.220.166.238, mailfrom: prvs=857337a20d=yi.zhao@windriver.com) Received: from pps.filterd (m0250810.ppops.net [127.0.0.1]) by mx0a-0064b401.pphosted.com (8.18.1.11/8.18.1.11) with ESMTP id 63N101rf1195472 for ; Wed, 22 Apr 2026 18:48:40 -0700 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=windriver.com; h=content-transfer-encoding:content-type:date:from:in-reply-to :message-id:mime-version:references:subject:to; s=PPS06212021; bh=/aoHdd+CJRLuSONsVYI7i1n20X4SqzfzZ+5fgulm2cw=; b=cSPZBW4Cz2Ux OpBfo54SV9LYOlKLYKuc9gxwJZzT0mH4+c6w3UtlPg+2Tjbjm7emVbQtHOEEIUfm sb1I3lsdgPMueAUuZx4vR+u8sQK+ecXHvUw6DEIojuu+XHvrWRBp/2jfVk9kuC0N XlxPgMe97x/P5XV0nRN1BBxYK+gXXVNaVkrxSIss4SDolfTKp/hkMBu3MZ5m1C94 GEGF6tbHh4feX3T61D2IUGhlTrBBLW4ZryxtvTxSNMQdgdlgrPb7DTe4gBM/RlCD MJ3ZXfxjBV9bML1vtmeFLKNszHGXuPOznQB31wRP/LL2uZsAJITbR9cUStb1KN68 Y/Jk9v0ffg== Received: from sa9pr02cu001.outbound.protection.outlook.com (mail-southcentralusazon11013031.outbound.protection.outlook.com [40.93.196.31]) by mx0a-0064b401.pphosted.com (PPS) with ESMTPS id 4dq8rcr21r-1 (version=TLSv1.3 cipher=TLS_AES_256_GCM_SHA384 bits=256 verify=NOT) for ; Wed, 22 Apr 2026 18:48:40 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; s=arcselector10001; d=microsoft.com; cv=none; b=MiuZsHj7FdzT7UFyGZi1obVlscOQZnf34hxiwoXbAy2vKcyAGj4gROrI9TXxKRC/jNXjvMVmjJAbIum51XXFt7Vru4sKTzkmsRTbhw8FSxY0qcgrzJHimng2t3WVScxePLeNN82mzfWf/hUx4g+kytTDglatR7zImrD6FZt9/B1zuAA8CVtDFk4S4U7wrSxqntPpPsC8XJ0Jub58uj/hCpL7hZWaHZuxu/6hIM9Wjz8vBKZ/NQZlHh/OYlFWk2NI+OD9rjai/W/1ME9ZMja58Ydw2uGSTa8jDKgB6/T97jNEJCWCqqPRCh0YKAs/tIaNGxDs0oWSv8CyJiOrgOnWmg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector10001; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=/aoHdd+CJRLuSONsVYI7i1n20X4SqzfzZ+5fgulm2cw=; b=g1M+ACBh85ofsAtsoQDvE/ImzUe2eNI79wo9lZo/JC17CUarP71uPDnSTbpHRYUo201k13+f2BFb7sQ9IgyxfuMZvRA5r2bkHr5meFL7V/XWLncgWP41z6pHBY3AoGG1GGOtkpBTPvl7vP3MSOvMHzt+2+SqYtg2iQvjTKophPRMMIVX3xH36v6ijBic9OqHudC0NIUPxOFM6+hOfdwhRtaIQuICCzwDE2rfqiTYQVh5PxrKD1tdMUbCizX2YWT1GKCZSgqPu/o+g4Mzjs7Jhg9i3V3GHwCxzFb7dVDGjlHeQHxVlztWM8D/OamXABBWMr2x7Pvv25ZRMjR5wMdmdg== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=windriver.com; dmarc=pass action=none header.from=windriver.com; dkim=pass header.d=windriver.com; arc=none Received: from DS0PR11MB6399.namprd11.prod.outlook.com (2603:10b6:8:c8::5) by IA3PR11MB9421.namprd11.prod.outlook.com (2603:10b6:208:578::9) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.9846.16; Thu, 23 Apr 2026 01:48:36 +0000 Received: from DS0PR11MB6399.namprd11.prod.outlook.com ([fe80::3432:2eb3:d0a5:7831]) by DS0PR11MB6399.namprd11.prod.outlook.com ([fe80::3432:2eb3:d0a5:7831%6]) with mapi id 15.20.9846.019; Thu, 23 Apr 2026 01:48:36 +0000 From: Yi Zhao To: openembedded-devel@lists.openembedded.org Subject: [meta-networking][PATCH 2/2] frr: fix mgmtd crash on ARM32 Date: Thu, 23 Apr 2026 09:48:19 +0800 Message-Id: <20260423014819.945909-2-yi.zhao@windriver.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20260423014819.945909-1-yi.zhao@windriver.com> References: <20260423014819.945909-1-yi.zhao@windriver.com> X-ClientProxiedBy: TYCP286CA0333.JPNP286.PROD.OUTLOOK.COM (2603:1096:400:38e::19) To DS0PR11MB6399.namprd11.prod.outlook.com (2603:10b6:8:c8::5) MIME-Version: 1.0 X-MS-PublicTrafficType: Email X-MS-TrafficTypeDiagnostic: DS0PR11MB6399:EE_|IA3PR11MB9421:EE_ X-MS-Office365-Filtering-Correlation-Id: 13103f2d-41b8-49a1-8a28-08dea0da6efc X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam: BCL:0;ARA:13230040|366016|52116014|376014|1800799024|18002099003|56012099003|22082099003|38350700014; X-Microsoft-Antispam-Message-Info: V2UJpVZ3F1jKlaGKIqoF1BXOG8uxd+6ONB+3utpz540/9xxU/FNz7EM1mcUsZrBaLXdCQ+CvhGecmJFW9iYK0p4is2SchpVla8Fe17/Xdr2+DUxoiT8ncRNHevEaC8+FLHjKq49whg29CbCH8FTB8leAMWPqUh/ogI4LccHvikzdhLyMjjQ2kw23ODlIREy6UB5G5dH+/iORvtKgfm691xoOq/bs3lucQ3D7OREMJC+N6yOf45TxQm/+rhX8+eFQLnh5OLvEMhAp79x19DhMDZX4TkTP/UB6pXVZL3FbedV2NjTHO7tau6GjDUaiJUNpXWEyq8KHHYoNbM+og0QpGuifsFHrYeEDM6Vp7TlkpkeYMUovsezF0nGefLiMg/gXP+L4nma5bznSaBriioohQNpX2S6SDR1bwhIL6YYL0AKE3/jds/inbaoi65UK84qnq5zD2fWtLfW9ueyYUocLWLSvEpLI0J6A0pMAEqmwYyF93LhtTD1CXYXu/pPd9P2Fs2DuYIbAuIZM1njIK9/oaf7rJdWt1LTsDC0QZEJ7O2jmqkyJDsL5ZHE0gONj68RfB3cMfVRHkPq9MTRrS811Q7aPkQxJGMvM67YD7LawEKT2VdCu6HY1iM9bvmgYhvTSYxMEGfgsSWebcUiaW4RJclAq4GfK2g8rkctremHiIoHwPiFvpgZiiRSBt9y67LC6Bx8jMLOX7nsN6nIrr9mJhYr2SfAtBY9niUt3oUDW+Ho1GlZUjkeCyCZaRGi8pdgH X-Forefront-Antispam-Report: CIP:255.255.255.255;CTRY:;LANG:en;SCL:1;SRV:;IPV:NLI;SFV:NSPM;H:DS0PR11MB6399.namprd11.prod.outlook.com;PTR:;CAT:NONE;SFS:(13230040)(366016)(52116014)(376014)(1800799024)(18002099003)(56012099003)(22082099003)(38350700014);DIR:OUT;SFP:1101; X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1 X-MS-Exchange-AntiSpam-MessageData-0: =?utf-8?q?6izwq0fFJ76FGE8JFUqwlCdFgcPX?= =?utf-8?q?wT6KBTf0fOiRHnyQ3q6wVJQRK9vVn1Wj9qnMNoGkr+Eoumr6hIJJ1gQMK6hc0NFri?= =?utf-8?q?Rl6j7FXCvNIw7AR1oBADO8OYpxFcCDsTTJAQVMcBdJ02TtaDOC4TaUQlszhD+yiko?= =?utf-8?q?5RvvrfkEGThcfhHuaIlb6dGmTYZ9+caNiLIJ4/Mr9JrlI8fccJ0nM+St85FafuLEe?= =?utf-8?q?uiLwqLQI//3LynvjGUoL5DB0zoSROFxCA+uzIosdfjkwIxw1WWCNTZaMQa/DFE5e4?= =?utf-8?q?FDb3VR7tXPX+g3pqkIqZVPMyxOgKr8Ax47Pl20bb7L9vFoJkyM3et7XWoPlyf5H4A?= =?utf-8?q?JEYkiO+zrtEW73Hm44gNunkTvb/G4Cxf5Qu0RPPWRabs3eOdYLqY/ZA4JOU37W38f?= =?utf-8?q?LMP3qe47tisUScxs27ezVJzc+rBgB/+EYd4Rbizy3+JiOTurdWQLLQLIzTS75c4Sh?= =?utf-8?q?0Nl9OcGVdnMcwFeepg/QmaSofoklqc/1W5OeMnQHS5G46ije27RWzNhAR7g67balr?= =?utf-8?q?Q0ueOAMEMQwdBHB0aLefqVlep2CX4E98LJL//aWrN8r/YUdcPgLUtkeNEwrrMy3oj?= =?utf-8?q?7EAFGiURsRymzZHha6xAIBeVVW1411Giu/hqJ9vROl0CYNkmoW1xrbX+IzL0f2AnW?= =?utf-8?q?MiT/oCPSEUPeeNuzKp8zG6/n7+t2EMDU+NLzj98FOwcfFl4AhTjeKZq6pV0JeTP7q?= =?utf-8?q?GWyPCBmKmOszq+4unaKy6z2KKx8NWaTsIvDmp3jQYgPzNNUcixN2bZaCNylrtMzKB?= =?utf-8?q?R4wI6YkduedUth6QlcaUrE9AkP94FUwcDvm9sfAuqDpzD4bX36bkToIwf5JzSMe/J?= =?utf-8?q?EE0EqvC++d37o2XP19zMkUqC6N48KefieRMAUN+R01d+lFiTnYK/4VuNPOyImj+B9?= =?utf-8?q?FXG1N+hzK+tEHGJ2Q1agddHJFeZ4S/HG7LX70D80gk5H1ClCb13ya94F7FokIHi3G?= =?utf-8?q?LGIeeX2q9Vj0FlVjMJTxDaFvi5VSXqBjGWpMpWbJRhoaGA5sphfegilMVGPP/vcR8?= =?utf-8?q?VmZ+Hi4UuWRcWuIbyKmX0/FVpqS0Z39WOAK530rHMl/4hPZhgBNvVsSkeHEbq0Ae6?= =?utf-8?q?MyeWBwUGz+vX71WBdQiETrqGzxMmgdt6b2SEKztaizv45DfVGC3uvhIS4kirI2brY?= =?utf-8?q?Wc7mxRFx5qD2P/Q1H/VOZEVgaw+MJM5wHeqcrIWHcZUykQKHEKfBbICLcuyi/eKZ7?= =?utf-8?q?iy0G4rmF0Sc7VNsayFW0db5UAsgvOFscdVZZ/DMWrTyD37bNqaXuzgrtIp/ab4ypg?= =?utf-8?q?Q8gWGVgRJ+emDf2M548JjvPhEXCkgWaOS2LDgOcw64bEjz39RUt6v8OBHImkITdk1?= =?utf-8?q?VShJiL0BGIGxRLZngn+6V9otqW89OvyPGVO8h0GcEt1a3ZmGqjM8smpyjxNNjsP6F?= =?utf-8?q?kg/RkoNYo4gtUVve67tWWfYnYQktEPgdHc8OrOyO0Fb7M7R/CcvFE4zSyyOQiNaYH?= =?utf-8?q?sdGOp2ArLPZOfqPfGlL0atrHpudeFTT7fs551/ILyeTLyRRXprx92sJsasbUyXrsh?= =?utf-8?q?JWy9fQTdUOh6hgaNO7QGmiTsSHLpFk8gkoPi7CjbPVFIhhEF0jzHIfEEnbCvdymIA?= =?utf-8?q?KB790OSuAP3pzUm2Jif2FUwA4GVBljeEbVIj8jBSiG9aLQdDL7BOy6IWzLSyg8cKJ?= =?utf-8?q?bx9dcZTYN/fy+DR4LRVlL6B5i/KMF3dg=3D=3D?= X-Exchange-RoutingPolicyChecked: KLDLOuSPlYLwHodWXuEMms9f+/F8h1aDGjrRyqAyOa4C0BwZvpXgcSOjWI8woQnAnCNa8tIqd1EduF9wUN0AKg0g9i5OlvLrMAFeojcxPeWg//q09dZupDXisWPSnhtBFYmaXshO+QTRnmyhAgUNgR8/d5dwZmB/LV9m/zBGxJMj9pFBNf8fqrMzYoQdQxYWy+W8HtktDK/FX5FXvK8OFs0ds2Y2wIXcdZIWdPgkqKgpdPy9xd0IDwTdPGLJyViop/GxzXncb50I0EVmlXXqLRYXSQ7xf91tj/IhO7+ULu6lsJ0WXGOQA2cGlJfiPrlmbVBR9A6ifNqwx2Yw3BFRDw== X-OriginatorOrg: windriver.com X-MS-Exchange-CrossTenant-Network-Message-Id: 13103f2d-41b8-49a1-8a28-08dea0da6efc X-MS-Exchange-CrossTenant-AuthSource: DS0PR11MB6399.namprd11.prod.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Internal X-MS-Exchange-CrossTenant-OriginalArrivalTime: 23 Apr 2026 01:48:36.2371 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-CrossTenant-Id: 8ddb2873-a1ad-4a18-ae4e-4644631433be X-MS-Exchange-CrossTenant-MailboxType: HOSTED X-MS-Exchange-CrossTenant-UserPrincipalName: PRnawqPym050etR4fHZ05iV99RG/GkS6/ZOJoXjOStsR3O2clzocnehRgeywoIligGK5XeRvPLlQLSclols0dw== X-MS-Exchange-Transport-CrossTenantHeadersStamped: IA3PR11MB9421 X-Proofpoint-ORIG-GUID: Vi2uA3gTS7AE7DJx3mPW8aBWgMOHcT6R X-Authority-Analysis: v=2.4 cv=MYhcfZ/f c=1 sm=1 tr=0 ts=69e97a78 cx=c_pps a=GI5TgwD/LrSAXmlNZGMSjA==:117 a=6eWqkTHjU83fiwn7nKZWdM+Sl24=:19 a=z/mQ4Ysz8XfWz/Q5cLBRGdckG28=:19 a=lCpzRmAYbLLaTzLvsPZ7Mbvzbb8=:19 a=xqWC_Br6kY4A:10 a=IkcTkHD0fZMA:10 a=A5OVakUREuEA:10 a=VkNPw1HP01LnGYTKEx00:22 a=bi6dqmuHe4P4UrxVR6um:22 a=HK-ge7EqtdluswH-FwHe:22 a=NEAV23lmAAAA:8 a=nZ40scU-AAAA:20 a=t7CeM3EgAAAA:8 a=wU2YTnxGAAAA:8 a=u9XiZHwLAAAA:8 a=P-IC7800AAAA:8 a=ZPpYJ6LdIiJ9JR2BXO4A:9 a=3ZKOabzyN94A:10 a=QEXdDO2ut3YA:10 a=FdTzh2GWekK77mhwV6Dw:22 a=Yz9wTY_ffGCQnEDHKrcv:22 a=_dVK9oMV_cxv_JnhfDLr:22 a=d3PnA9EDa4IxuAV0gXij:22 a=bA3UWDv6hWIuX7UZL3qL:22 X-Proofpoint-GUID: Vi2uA3gTS7AE7DJx3mPW8aBWgMOHcT6R X-Proofpoint-Spam-Details-Enc: AW1haW4tMjYwNDIzMDAxNSBTYWx0ZWRfX1ZHZk1dGkn/C FjhogI/d+aWmrSN3xIE6hT+MlniKYcPO6sSru31FeG/1eh1U8She3gPCL+PFDRE5DvdGL8aRPjC RSpr5hJ4WMODARMVUZ9pF639gZvn6Kp1LHsifnD9tBazOqq1smy9LBWodQqlOvXPbKmnn1Pjgcb ZCuwVBBknttQ0SFhmCda33zh1EGcOz/Tfq7Ud3JUL2V/v97kn7NN3mntSQtqzLoCwBDCgXaZuMO 6/IY/lAtXkHdA4E6OL97LHlq54P7Dai4R1dMwSfjI0wQzKs8/I/s8yW/TJe3KRpns6uagSpLpdk Koy2Jd85QMjb8A0vVfMVLifkgHrim/sYkvPoislIirbieB/pIBKlB6t5fuacsoQKw+kbOllGlZi 5FnaF3LsJBaoIL9u6ONOKbEs6XA5UpdtWT0ZpNqGyAkzcmg6sTHSw/VZzanqFchoQwQ+U+JMBJs +LC9rnRwC1hO9/dlg3Q== X-Proofpoint-Virus-Version: vendor=baseguard engine=ICAP:2.0.293,Aquarius:18.0.1143,Hydra:6.1.51,FMLib:17.12.100.49 definitions=2026-04-22_04,2026-04-21_02,2025-10-01_01 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 phishscore=0 bulkscore=0 suspectscore=0 lowpriorityscore=0 priorityscore=1501 malwarescore=0 adultscore=0 spamscore=0 clxscore=1015 impostorscore=0 classifier=typeunknown authscore=0 authtc= authcc= route=outbound adjust=0 reason=mlx scancount=1 engine=8.22.0-2604200000 definitions=main-2604230015 X-MIME-Autoconverted: from 8bit to quoted-printable by mx0a-0064b401.pphosted.com id 63N101rf1195472 List-Id: X-Webhook-Received: from 45-33-107-173.ip.linodeusercontent.com [45.33.107.173] by aws-us-west-2-korg-lkml-1.web.codeaurora.org with HTTPS for ; Thu, 23 Apr 2026 01:48:48 -0000 X-Groupsio-URL: https://lists.openembedded.org/g/openembedded-devel/message/126565 Backport fix[1] for MGMT crash on first start on ARM32 platforms[2]. [1] https://github.com/FRRouting/frr/pull/21651 [2] https://github.com/FRRouting/frr/issues/20087 Signed-off-by: Yi Zhao --- ..._msg-recv-to-deal-with-mis-alignment.patch | 352 ++++++++++++++++++ .../recipes-protocols/frr/frr_10.6.1.bb | 1 + 2 files changed, 353 insertions(+) create mode 100644 meta-networking/recipes-protocols/frr/frr/0001-lib-fix-mgmt_msg-recv-to-deal-with-mis-alignment.patch diff --git a/meta-networking/recipes-protocols/frr/frr/0001-lib-fix-mgmt_msg-recv-to-deal-with-mis-alignment.patch b/meta-networking/recipes-protocols/frr/frr/0001-lib-fix-mgmt_msg-recv-to-deal-with-mis-alignment.patch new file mode 100644 index 0000000000..22cc10cf31 --- /dev/null +++ b/meta-networking/recipes-protocols/frr/frr/0001-lib-fix-mgmt_msg-recv-to-deal-with-mis-alignment.patch @@ -0,0 +1,352 @@ +From 5959a3d0cbc73b0c41134bf0d9944a6bd40ba510 Mon Sep 17 00:00:00 2001 +From: Christian Hopps +Date: Sat, 18 Apr 2026 03:01:46 +0000 +Subject: [PATCH] lib: fix mgmt_msg recv to deal with mis-alignment +MIME-Version: 1.0 +Content-Type: text/plain; charset=UTF-8 +Content-Transfer-Encoding: 8bit + +We need our messages to start on 64 bit boundaries as the message buffer +is accessed directly as structured data. In particular on ARM32 arch +using the data this way lead to unaligned access and SIGBUS. + +The minor optimization of reading multiple messages into a single stream +buffer complicated this. Instead we KISS and switch to one message per +stream buffer. + +Fixes #20087. + +Signed-off-by: Christian Hopps +Co-developed-by: Samir MOUHOUNE +Co-developed-by: Alexis Lothoré + +See also PR #20985 + +This issue was identified and another solution was provided by Samir +MOUHOUNE with the following commit message comments: + +On ARM32 systems, mgmtd crashes at startup on an alignment fault: + +``` + frrinit.sh[158]: Starting watchfrr with command: ' /usr/sbin/watchfrr -d mgmtd zebra staticd' + watchfrr[168]: [T83RR-8SM5G] watchfrr 10.5.2 starting: vty@0 + watchfrr[168]: [ZCJ3S-SPH5S] mgmtd state -> down : initial connection attempt failed + watchfrr[168]: [ZCJ3S-SPH5S] zebra state -> down : initial connection attempt failed + watchfrr[168]: [ZCJ3S-SPH5S] staticd state -> down : initial connection attempt failed + watchfrr[168]: [YFT0P-5Q5YX] Forked background command [pid 169]: /usr/sbin/watchfrr.sh restart all + frrinit.sh[180]: 2026/02/27 09:14:13 ZEBRA: [KGY44-D47GD][EC 4043309111] Disabling MPLS support (no kernel support) + watchfrr[168]: [QDG3Y-BY5TN] zebra state -> up : connect succeeded + kernel: Alignment trap: not handling instruction edc30b02 at [<004c3c1c>] + kernel: 8<--- cut here --- + kernel: Unhandled fault: alignment exception (0x801) at 0x008879f6 + kernel: [008879f6] *pgd=9baf6831 + watchfrr[168]: [YFT0P-5Q5YX] Forked background command [pid 189]: /usr/sbin/watchfrr.sh restart mgmtd + frrinit.sh[189]: Cannot stop mgmtd: pid 179 not running + watchfrr.sh[196]: Cannot stop mgmtd: pid 179 not running + frrinit.sh[202]: [202|zebra] sending configuration + frrinit.sh[202]: [202|zebra] done + frrinit.sh[216]: [216|watchfrr] sending configuration + frrinit.sh[218]: [218|staticd] sending configuration + watchfrr[168]: [VTVCM-Y2NW3] Configuration Read in Took: 00:00:00 + frrinit.sh[199]: Waiting for children to finish applying config... + frrinit.sh[216]: [216|watchfrr] done +``` + +When checking crashlogs in /var/tmp/frr, mgmt gives the following: + +``` + MGMTD: Received signal 7 at 1772183653 (si_addr 0x8879f6); aborting... + MGMTD: /lib/libfrr.so.0(zlog_backtrace_sigsafe+0x5c) [0xb6e89c90] + MGMTD: /lib/libfrr.so.0(zlog_signal+0xe0) [0xb6e89e80] + MGMTD: /lib/libfrr.so.0(+0xd4374) [0xb6ed3374] + MGMTD: /lib/libc.so.6(__default_rt_sa_restorer+0) [0xb6ab4d90] + MGMTD: /usr/sbin/mgmtd(mgmt_fe_adapter_send_notify+0x6b8) [0x4c3c20] + MGMTD: /lib/libfrr.so.0(mgmt_msg_procbufs+0x124) [0xb6e976b8] + MGMTD: /lib/libfrr.so.0(+0x98798) [0xb6e97798] + MGMTD: /lib/libfrr.so.0(event_call+0xa8) [0xb6ee739c] + MGMTD: /lib/libfrr.so.0(frr_run+0xd4) [0xb6e80fc8] + MGMTD: /usr/sbin/mgmtd(main+0x188) [0x4bd7ec] + MGMTD: /lib/libc.so.6(+0x236b0) [0xb6a9f6b0] + MGMTD: /lib/libc.so.6(__libc_start_main+0x98) [0xb6a9f790] + MGMTD: in thread msg_conn_proc_msgs scheduled from lib/mgmt_msg.c:543 msg_conn_sched_proc_msgs() +``` + +The issue is that messages are queued for sending/receive back-to-back +with no padding. This means that when mgmt creates a pointer back to the +data waiting in queue and tries to access fields inside the dereferenced +message, those accesses are not performed with the alignment constraints +required by some architectures. For example, ARM ABI AAPCS32 ([1]) +states that structures alignment should be the same as the "most +aligned" member; so a struct mgmt_msg_header, which contains some +uint64_t fields (which are 8-bytes alignes), should be 8-bytes aligned +as well. + +On x86, this goes unnoticed because the CPU handles unaligned access +transparently. On ARM 32-bit with NEON/VFP, the compiler generates +64-bit store instructions that trap on unaligned addresses. The kernel +cannot emulate these instructions and kills the process with SIGBUS. + +[1] https://github.com/ARM-software/abi-aa/blob/main/aapcs32/aapcs32.rst#data-types-and-alignment + +Upstream-Status: Backport [https://github.com/FRRouting/frr/commit/5959a3d0cbc73b0c41134bf0d9944a6bd40ba510] + +Signed-off-by: Christian Hopps +(cherry picked from commit ae7d79f8ff25d5750e5796567ff6317030900d40) +Signed-off-by: Yi Zhao +--- + lib/mgmt_be_client.h | 3 +- + lib/mgmt_fe_client.h | 3 +- + lib/mgmt_msg.c | 157 ++++++++++++++++++------------------------- + 3 files changed, 68 insertions(+), 95 deletions(-) + +diff --git a/lib/mgmt_be_client.h b/lib/mgmt_be_client.h +index f5627e3c4e..2f412a6fbd 100644 +--- a/lib/mgmt_be_client.h ++++ b/lib/mgmt_be_client.h +@@ -21,7 +21,8 @@ extern "C" { + + #define MGMTD_BE_MAX_NUM_MSG_PROC 500 + #define MGMTD_BE_MAX_NUM_MSG_WRITE 1000 +-#define MGMTD_BE_MAX_MSG_LEN (64 * 1024) ++/* Messages can be any size, this is just the preallocated buffer size */ ++#define MGMTD_BE_MAX_MSG_LEN (4 * 1024) + + #define MGMTD_BE_CONTAINER_NODE_VAL "<>" + +diff --git a/lib/mgmt_fe_client.h b/lib/mgmt_fe_client.h +index 8ff08b566a..3005c5dd01 100644 +--- a/lib/mgmt_fe_client.h ++++ b/lib/mgmt_fe_client.h +@@ -29,7 +29,8 @@ extern "C" { + + #define MGMTD_FE_MAX_NUM_MSG_PROC 500 + #define MGMTD_FE_MAX_NUM_MSG_WRITE 100 +-#define MGMTD_FE_MAX_MSG_LEN (64 * 1024) ++/* Messages can be any size, this is just the preallocated buffer size */ ++#define MGMTD_FE_MAX_MSG_LEN (4 * 1024) + + /*************************************************************** + * Data-structures +diff --git a/lib/mgmt_msg.c b/lib/mgmt_msg.c +index f299b52873..fb56f58ab5 100644 +--- a/lib/mgmt_msg.c ++++ b/lib/mgmt_msg.c +@@ -13,6 +13,7 @@ + #include "network.h" + #include "sockopt.h" + #include "stream.h" ++#include "zlog.h" + #include "frrevent.h" + #include "mgmt_msg.h" + #include "mgmt_msg_native.h" +@@ -39,8 +40,8 @@ static bool trace; + DEFINE_MTYPE(LIB, MSG_CONN, "msg connection state"); + + /** +- * Read data from a socket into streams containing 1 or more full msgs headed by +- * mgmt_msg_hdr which contain API messages (currently protobuf). ++ * Read data from a socket into a stream containing 1 full msg headed by ++ * mgmt_msg_hdr. + * + * Args: + * ms: mgmt_msg_state for this process. +@@ -57,96 +58,80 @@ enum mgmt_msg_rsched mgmt_msg_read(struct mgmt_msg_state *ms, int fd, + bool debug) + { + const char *dbgtag = debug ? ms->idtag : NULL; +- size_t avail = STREAM_WRITEABLE(ms->ins); + struct mgmt_msg_hdr *mhdr = NULL; +- size_t total = 0; +- size_t mcount = 0; +- ssize_t n, left; ++ struct stream *news; ++ size_t nread; ++ ssize_t n; + + assert(ms && fd != -1); +- MGMT_MSG_TRACE(dbgtag, "enter with %zu bytes available to read on fd %d", avail, fd); ++ MGMT_MSG_TRACE(dbgtag, "enter to read from fd %d", fd); ++ ++ assert(stream_get_getp(ms->ins) == 0); ++ nread = stream_get_endp(ms->ins); + + /* +- * Read as much as we can into the stream. ++ * Get header, validate, and resize the stream, if needed, to fit incoming message. + */ +- while (avail > sizeof(struct mgmt_msg_hdr)) { +- n = stream_read_try(ms->ins, fd, avail); +- +- /* -2 is normal nothing read, and to retry */ +- if (n == -2) { +- MGMT_MSG_TRACE(dbgtag, "nothing more to read on fd %d", fd); +- break; +- } +- if (n <= 0) { +- if (n == 0) +- MGMT_MSG_ERR(ms, "got EOF/disconnect on fd %d", fd); +- else +- MGMT_MSG_ERR(ms, "got error while reading on fd %d: '%s'", fd, +- safe_strerror(errno)); +- return MSR_DISCONNECT; ++ if (nread < sizeof(struct mgmt_msg_hdr)) { ++ while (nread < sizeof(struct mgmt_msg_hdr)) { ++ n = stream_read_try(ms->ins, fd, sizeof(struct mgmt_msg_hdr) - nread); ++ if (n <= 0) ++ goto not_done; ++ nread += n; ++ ms->nrxb += n; + } +- MGMT_MSG_TRACE(dbgtag, "read %zd bytes on fd %d", n, fd); +- ms->nrxb += n; +- avail -= n; +- } + +- /* +- * Check if we have read a complete messages or not. +- */ +- assert(stream_get_getp(ms->ins) == 0); +- left = stream_get_endp(ms->ins); +- while (left > (ssize_t)sizeof(struct mgmt_msg_hdr)) { +- mhdr = (struct mgmt_msg_hdr *)(STREAM_DATA(ms->ins) + total); ++ /* Validate the header is sane */ ++ mhdr = (struct mgmt_msg_hdr *)STREAM_DATA(ms->ins); + if (!MGMT_MSG_IS_MARKER(mhdr->marker)) { + MGMT_MSG_DBG(dbgtag, "recv corrupt buffer on fd %d, disconnect", fd); + return MSR_DISCONNECT; ++ } else if (mhdr->len <= sizeof(struct mgmt_msg_hdr)) { ++ MGMT_MSG_DBG(dbgtag, "recv invalid message length %u on fd %d, disconnect", ++ mhdr->len, fd); ++ return MSR_DISCONNECT; + } +- if ((ssize_t)mhdr->len > left) +- break; +- +- MGMT_MSG_TRACE(dbgtag, "read full message on fd %d len %u", fd, mhdr->len); +- total += mhdr->len; +- left -= mhdr->len; +- mcount++; +- } + +- if (!mcount) { +- /* Didn't manage to read a full message */ +- if (mhdr && avail == 0) { +- struct stream *news; +- /* +- * Message was longer than what was left and we have no +- * available space to read more in. B/c mcount == 0 the +- * message starts at the beginning of the stream so +- * therefor the stream is too small to fit the message.. +- * Resize the stream to fit. +- */ ++ /* See if message will fit in the stream, realloc if not */ ++ if (mhdr->len > ms->ins->size) { ++ MGMT_MSG_DBG(dbgtag, ++ "message length %u is greater than available %zu on fd %d", ++ mhdr->len, ms->ins->size, fd); + news = stream_new(mhdr->len); +- stream_put(news, mhdr, left); +- stream_set_endp(news, left); ++ stream_put(news, mhdr, sizeof(struct mgmt_msg_hdr)); + stream_free(ms->ins); + ms->ins = news; + } +- return MSR_SCHED_STREAM; + } + +- /* +- * We have read at least one message into the stream, queue it up. +- */ +- mhdr = (struct mgmt_msg_hdr *)(STREAM_DATA(ms->ins) + total); +- stream_set_endp(ms->ins, total); +- stream_fifo_push(&ms->inq, ms->ins); +- if (left < (ssize_t)sizeof(struct mgmt_msg_hdr)) +- ms->ins = stream_new(ms->max_msg_sz); +- else +- /* handle case where message is greater than max */ +- ms->ins = stream_new(MAX(ms->max_msg_sz, mhdr->len)); +- if (left) { +- stream_put(ms->ins, mhdr, left); +- stream_set_endp(ms->ins, left); ++ /* Read the rest of the message. */ ++ mhdr = (struct mgmt_msg_hdr *)STREAM_DATA(ms->ins); ++ while (nread < mhdr->len) { ++ n = stream_read_try(ms->ins, fd, mhdr->len - nread); ++ if (n <= 0) ++ goto not_done; ++ nread += n; ++ ms->nrxb += n; ++ MGMT_MSG_TRACE(dbgtag, "read %zd from fd %d (%zu of %u)", n, fd, nread, mhdr->len); + } + ++ /* We've got a full message, push it onto the FIFO and setup for the next message. */ ++ MGMT_MSG_TRACE(dbgtag, "read full msg %zu/%u from fd %d", nread, mhdr->len, fd); ++ stream_fifo_push(&ms->inq, ms->ins); ++ ms->ins = stream_new(ms->max_msg_sz); + return MSR_SCHED_BOTH; ++ ++not_done: ++ if (n == -2) { ++ MGMT_MSG_TRACE(dbgtag, "nothing more to read on fd %d", fd); ++ return MSR_SCHED_STREAM; ++ } ++ if (n == 0) ++ MGMT_MSG_ERR(ms, "got EOF/disconnect on fd %d", fd); ++ else ++ MGMT_MSG_ERR(ms, "got error while reading on fd %d: '%s'", fd, ++ safe_strerror(errno)); ++ return MSR_DISCONNECT; + } + + /** +@@ -171,7 +156,6 @@ bool mgmt_msg_procbufs(struct mgmt_msg_state *ms, + const char *dbgtag = debug ? ms->idtag : NULL; + struct mgmt_msg_hdr *mhdr; + struct stream *work; +- uint8_t *data; + size_t left, nproc; + + MGMT_MSG_TRACE(dbgtag, "Have %zu streams to process", ms->inq.count); +@@ -182,30 +166,17 @@ bool mgmt_msg_procbufs(struct mgmt_msg_state *ms, + if (!work) + break; + +- data = STREAM_DATA(work); + left = stream_get_endp(work); + MGMT_MSG_TRACE(dbgtag, "Processing stream of len %zu", left); +- +- for (; left > sizeof(struct mgmt_msg_hdr); +- left -= mhdr->len, data += mhdr->len) { +- mhdr = (struct mgmt_msg_hdr *)data; +- +- assert(MGMT_MSG_IS_MARKER(mhdr->marker)); +- assert(left >= mhdr->len); +- +- /* +- * Q: if the handler disconnects should stop/flush? +- */ +- handle_msg(MGMT_MSG_MARKER_VERSION(mhdr->marker), (uint8_t *)(mhdr + 1), +- mhdr->len - sizeof(struct mgmt_msg_hdr), user); +- ms->nrxm++; +- nproc++; +- } +- +- if (work != ms->ins) +- stream_free(work); /* Free it up */ +- else +- stream_reset(work); /* Reset stream for next read */ ++ /* ++ * Q: if the handler disconnects should we stop/flush? ++ */ ++ mhdr = (struct mgmt_msg_hdr *)STREAM_DATA(work); ++ handle_msg(MGMT_MSG_MARKER_VERSION(mhdr->marker), (uint8_t *)(mhdr + 1), ++ mhdr->len - sizeof(struct mgmt_msg_hdr), user); ++ ms->nrxm++; ++ nproc++; ++ stream_free(work); /* Free it up */ + } + + /* return true if should reschedule b/c more to process. */ +-- +2.43.0 + diff --git a/meta-networking/recipes-protocols/frr/frr_10.6.1.bb b/meta-networking/recipes-protocols/frr/frr_10.6.1.bb index e86e0f3153..1cd102f0da 100644 --- a/meta-networking/recipes-protocols/frr/frr_10.6.1.bb +++ b/meta-networking/recipes-protocols/frr/frr_10.6.1.bb @@ -12,6 +12,7 @@ LIC_FILES_CHKSUM = "file://doc/licenses/GPL-2.0;md5=b234ee4d69f5fce4486a80fdaf4a SRC_URI = "git://github.com/FRRouting/frr.git;protocol=https;branch=stable/10.6;tag=frr-${PV} \ file://frr.pam \ + file://0001-lib-fix-mgmt_msg-recv-to-deal-with-mis-alignment.patch \ " SRCREV = "71da51baee6fb2a02b24262defc46591c86e8a81"