From patchwork Thu Apr 7 17:43:22 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chuck Lever X-Patchwork-Id: 12805588 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 20AA9C433F5 for ; Thu, 7 Apr 2022 17:43:33 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232372AbiDGRpa (ORCPT ); Thu, 7 Apr 2022 13:45:30 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:44146 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231470AbiDGRpa (ORCPT ); Thu, 7 Apr 2022 13:45:30 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C3CC422A292 for ; Thu, 7 Apr 2022 10:43:27 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 9D36761AFE for ; Thu, 7 Apr 2022 17:43:24 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id DCEBEC385AA; Thu, 7 Apr 2022 17:43:23 +0000 (UTC) Subject: [PATCH v3 1/2] SUNRPC: Fix NFSD's request deferral on RDMA transports From: Chuck Lever To: linux-nfs@vger.kernel.org Cc: rostedt@goodmis.org Date: Thu, 07 Apr 2022 13:43:22 -0400 Message-ID: <164935340273.76813.7096678268046264254.stgit@klimt.1015granger.net> In-Reply-To: <164935330144.76813.17862521591948764594.stgit@klimt.1015granger.net> References: <164935330144.76813.17862521591948764594.stgit@klimt.1015granger.net> User-Agent: StGit/1.5.dev1+g8516920 MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-nfs@vger.kernel.org Trond Myklebust reports an NFSD crash in svc_rdma_sendto(). Further investigation shows that the crash occurred while NFSD was handling a deferred request. This patch addresses two inter-related issues that prevent request deferral from working correctly for RPC/RDMA requests: 1. Prevent the crash by ensuring that the original svc_rqst::rq_xprt_ctxt value is available when the request is revisited. Otherwise svc_rdma_sendto() does not have a Receive context available with which to construct its reply. 2. Possibly since before commit 71641d99ce03 ("svcrdma: Properly compute .len and .buflen for received RPC Calls"), svc_rdma_recvfrom() did not include the transport header in the returned xdr_buf. There should have been no need for svc_defer() and friends to save and restore that header, as of that commit. This issue is addressed in a backport-friendly way by simply having svc_rdma_recvfrom() set rq_xprt_hlen to zero unconditionally, just as svc_tcp_recvfrom() does. This enables svc_deferred_recv() to correctly reconstruct an RPC message received via RPC/RDMA. Reported-by: Trond Myklebust Link: https://lore.kernel.org/linux-nfs/82662b7190f26fb304eb0ab1bb04279072439d4e.camel@hammerspace.com/ Signed-off-by: Chuck Lever Cc: --- include/linux/sunrpc/svc.h | 1 + net/sunrpc/svc_xprt.c | 3 +++ net/sunrpc/xprtrdma/svc_rdma_recvfrom.c | 2 +- 3 files changed, 5 insertions(+), 1 deletion(-) diff --git a/include/linux/sunrpc/svc.h b/include/linux/sunrpc/svc.h index a5dda4987e8b..217711fc9cac 100644 --- a/include/linux/sunrpc/svc.h +++ b/include/linux/sunrpc/svc.h @@ -395,6 +395,7 @@ struct svc_deferred_req { size_t addrlen; struct sockaddr_storage daddr; /* where reply must come from */ size_t daddrlen; + void *xprt_ctxt; struct cache_deferred_req handle; size_t xprt_hlen; int argslen; diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index 0c117d3bfda8..b42cfffa7395 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -1231,6 +1231,8 @@ static struct cache_deferred_req *svc_defer(struct cache_req *req) dr->daddr = rqstp->rq_daddr; dr->argslen = rqstp->rq_arg.len >> 2; dr->xprt_hlen = rqstp->rq_xprt_hlen; + dr->xprt_ctxt = rqstp->rq_xprt_ctxt; + rqstp->rq_xprt_ctxt = NULL; /* back up head to the start of the buffer and copy */ skip = rqstp->rq_arg.len - rqstp->rq_arg.head[0].iov_len; @@ -1269,6 +1271,7 @@ static noinline int svc_deferred_recv(struct svc_rqst *rqstp) rqstp->rq_xprt_hlen = dr->xprt_hlen; rqstp->rq_daddr = dr->daddr; rqstp->rq_respages = rqstp->rq_pages; + rqstp->rq_xprt_ctxt = dr->xprt_ctxt; svc_xprt_received(rqstp->rq_xprt); return (dr->argslen<<2) - dr->xprt_hlen; } diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c index cf76a6ad127b..864131a9fc6e 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c +++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c @@ -831,7 +831,7 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp) goto out_err; if (ret == 0) goto out_drop; - rqstp->rq_xprt_hlen = ret; + rqstp->rq_xprt_hlen = 0; if (svc_rdma_is_reverse_direction_reply(xprt, ctxt)) goto out_backchannel; From patchwork Thu Apr 7 17:43:29 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chuck Lever X-Patchwork-Id: 12805590 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 1D52DC433EF for ; Thu, 7 Apr 2022 17:43:41 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1346528AbiDGRpi (ORCPT ); Thu, 7 Apr 2022 13:45:38 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:44636 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1346519AbiDGRph (ORCPT ); Thu, 7 Apr 2022 13:45:37 -0400 Received: from ams.source.kernel.org (ams.source.kernel.org [IPv6:2604:1380:4601:e00::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9D4E122C1E9 for ; Thu, 7 Apr 2022 10:43:33 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ams.source.kernel.org (Postfix) with ESMTPS id E7D77B82795 for ; Thu, 7 Apr 2022 17:43:31 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 38B4EC385A4; Thu, 7 Apr 2022 17:43:30 +0000 (UTC) Subject: [PATCH v3 2/2] SUNRPC: Fix the svc_deferred_event trace class From: Chuck Lever To: linux-nfs@vger.kernel.org Cc: rostedt@goodmis.org Date: Thu, 07 Apr 2022 13:43:29 -0400 Message-ID: <164935340926.76813.4345516465286175319.stgit@klimt.1015granger.net> In-Reply-To: <164935330144.76813.17862521591948764594.stgit@klimt.1015granger.net> References: <164935330144.76813.17862521591948764594.stgit@klimt.1015granger.net> User-Agent: StGit/1.5.dev1+g8516920 MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-nfs@vger.kernel.org Fix a NULL deref crash that occurs when an svc_rqst is deferred while the sunrpc tracing subsystem is enabled. svc_revisit() sets dr->xprt to NULL, so it can't be relied upon in the tracepoint to provide the remote's address. Unfortunately we can't revert the "svc_deferred_class" hunk in commit ece200ddd54b ("sunrpc: Save remote presentation address in svc_xprt for trace events") because there is now a specific check of event format specifiers for unsafe dereferences. The warning that check emits is: event svc_defer_recv has unsafe dereference of argument 1 A "%pISpc" format specifier with a "struct sockaddr *" is indeed flagged by this check. Instead, take the brute-force approach used by the svcrdma_qp_error tracepoint. Convert the dr::addr field into a presentation address in the TP_fast_assign() arm of the trace event, and store that as a string. This fix can be backported to -stable kernels. In the meantime, commit c6ced22997ad ("tracing: Update print fmt check to handle new __get_sockaddr() macro") is now in v5.18, so this wonky fix can be replaced with __sockaddr() and friends properly during the v5.19 merge window. Fixes: ece200ddd54b ("sunrpc: Save remote presentation address in svc_xprt for trace events") Signed-off-by: Chuck Lever --- include/trace/events/sunrpc.h | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h index ab8ae1f6ba84..4eb706fa5825 100644 --- a/include/trace/events/sunrpc.h +++ b/include/trace/events/sunrpc.h @@ -2017,17 +2017,18 @@ DECLARE_EVENT_CLASS(svc_deferred_event, TP_STRUCT__entry( __field(const void *, dr) __field(u32, xid) - __string(addr, dr->xprt->xpt_remotebuf) + __array(__u8, addr, INET6_ADDRSTRLEN + 10) ), TP_fast_assign( __entry->dr = dr; __entry->xid = be32_to_cpu(*(__be32 *)(dr->args + (dr->xprt_hlen>>2))); - __assign_str(addr, dr->xprt->xpt_remotebuf); + snprintf(__entry->addr, sizeof(__entry->addr) - 1, + "%pISpc", (struct sockaddr *)&dr->addr); ), - TP_printk("addr=%s dr=%p xid=0x%08x", __get_str(addr), __entry->dr, + TP_printk("addr=%s dr=%p xid=0x%08x", __entry->addr, __entry->dr, __entry->xid) );