From patchwork Fri Nov 10 22:37:05 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Joshua Watt X-Patchwork-Id: 10053905 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id E7AE46032D for ; Fri, 10 Nov 2017 22:37:41 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id DBCB4223C6 for ; Fri, 10 Nov 2017 22:37:41 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id D0A932B4BB; Fri, 10 Nov 2017 22:37:41 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.8 required=2.0 tests=BAYES_00, DKIM_ADSP_CUSTOM_MED, DKIM_SIGNED, FREEMAIL_FROM, RCVD_IN_DNSWL_HI, T_DKIM_INVALID autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 44B59223C6 for ; Fri, 10 Nov 2017 22:37:41 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754733AbdKJWhj (ORCPT ); Fri, 10 Nov 2017 17:37:39 -0500 Received: from mail-it0-f49.google.com ([209.85.214.49]:55069 "EHLO mail-it0-f49.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754707AbdKJWhh (ORCPT ); Fri, 10 Nov 2017 17:37:37 -0500 Received: by mail-it0-f49.google.com with SMTP id 72so3398699itk.3 for ; Fri, 10 Nov 2017 14:37:37 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references; bh=Plfx+az2n6gbStJzUe7yC+UbofETWsRNn7H47JOJGdc=; b=A3NX+CQPGveG6BoVuyfj3q1XomfZFfa6U9IYjpSet1hfP/6LwGhWw3zyEqltEW0hhc ntxebV/hy7E5L+m4kQd7adr2mlv2ClGWa647w2XiADGfEjZd1/+7qmerrVRihH1QNYhz jFXylX7/dO16qzlatsDkp37IcsoIPrDPy+ibs0wL4V2xb8scsVLuxa22oTSfBcOyKeVd B9XRIa9ptaFTD00qljnCnY+151A0PA/OU2C+NNjpxaDtVe5uitVkqC9AbUnDu8WRJnPq oA6fBuU7EHUmSvtUAIKJ+iySHTmJa0JM5Z7KKDalo0IaxTOZ1nLHFgREUp5R8whYGAUk qQlw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references; bh=Plfx+az2n6gbStJzUe7yC+UbofETWsRNn7H47JOJGdc=; b=twfI0axvYjyKi/4+ThA4xG+M80mJnmhM1Uf0YQic3beyQ6pB2uzZei+zEjITnKA+D9 iOWP2qJ812n2CdOaO92oRK07eumKvQeUoC0Jh7xEZZacXHqkXhwhEwQhncHLZaKD01st XqHJ4qqE2P6JQSVEo/PQiWOZCUOSE4LVLkqqkBcS9Mk7HPYC8iVYLrGNcr2YimixYrsb 1YeEnMJlJgxqfJ4aIYe8WpXodgN9BEdeNW1KwtYoJWuGqiOT1LNn8N/Ih7tbUKWKV6Kv pVoPPqKo4koNNHAyerHVxajH3XWWeNUVbhElSULCGuRBML8y3s6nt9nFY/gOH8cr/um6 CuCg== X-Gm-Message-State: AJaThX76gi6OxIbwxW0uXIBKhTQQYqVwYm5+kvg9mYP2TDpqMf907bRy H2Xzav9L//2to7AdXmnW4Ck= X-Google-Smtp-Source: AGs4zMZQv73dZVEZywfhPfOZWyXMa1HZ3YKFjQ5fHXiV/oHVPi78j/ELzJwwViaFeWvto/22lIB6rA== X-Received: by 10.36.217.15 with SMTP id p15mr2364311itg.14.1510353456803; Fri, 10 Nov 2017 14:37:36 -0800 (PST) Received: from ola-842mrw1.ad.garmin.com ([204.77.163.55]) by smtp.googlemail.com with ESMTPSA id f3sm5001090iob.70.2017.11.10.14.37.36 (version=TLS1_2 cipher=ECDHE-RSA-CHACHA20-POLY1305 bits=256/256); Fri, 10 Nov 2017 14:37:36 -0800 (PST) From: Joshua Watt X-Google-Original-From: Joshua Watt To: NeilBrown , Jeff Layton , Trond Myklebust Cc: linux-nfs@vger.kernel.org, Al Viro , "J . Bruce Fields" , David Howells , Joshua Watt Subject: [RFC v2 5/7] NFS: Add serverfailed mount option Date: Fri, 10 Nov 2017 16:37:05 -0600 Message-Id: <20171110223707.17098-6-JPEWhacker@gmail.com> X-Mailer: git-send-email 2.13.6 In-Reply-To: <20171110223707.17098-1-JPEWhacker@gmail.com> References: <20171110223707.17098-1-JPEWhacker@gmail.com> Sender: linux-nfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-nfs@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Specifying the serverfailed mount option causes all subsequent RPC tasks that are queued to fail with -EIO instead of timing out. For example, if a server has disappeared, the sequence: mount -o remount,serverfailed umount -f will ensure that all pending I/O requests are cancelled, and any new requests will also fail. In the event that the server returns, the flag can be removed with a remount: mount -o remount,noserverfailed Although bringing the server back may result in a loss of data Signed-off-by: Joshua Watt --- fs/nfs/inode.c | 6 ++++++ fs/nfs/super.c | 30 +++++++++++++++++++++++++----- include/uapi/linux/nfs_mount.h | 1 + 3 files changed, 32 insertions(+), 5 deletions(-) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 134d9f560240..55b7cd40508d 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -723,6 +723,12 @@ static void nfs_readdirplus_parent_cache_hit(struct dentry *dentry) static bool nfs_need_revalidate_inode(struct inode *inode) { + /* If the server has failed, it is not going to respond, so don't try + * and revalidated (otherwise, the serverfailed flag can't be cleared by + * a remount) + */ + if (NFS_SERVER(inode)->flags & NFS_MOUNT_SERVERFAILED) + return false; if (NFS_I(inode)->cache_validity & (NFS_INO_INVALID_ATTR|NFS_INO_INVALID_LABEL)) return true; diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 777a0cc22704..bca38e1cdd85 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -90,6 +90,7 @@ enum { Opt_resvport, Opt_noresvport, Opt_fscache, Opt_nofscache, Opt_migration, Opt_nomigration, + Opt_serverfailed, Opt_noserverfailed, /* Mount options that take integer arguments */ Opt_port, @@ -151,6 +152,8 @@ static const match_table_t nfs_mount_option_tokens = { { Opt_nofscache, "nofsc" }, { Opt_migration, "migration" }, { Opt_nomigration, "nomigration" }, + { Opt_serverfailed, "serverfailed" }, + { Opt_noserverfailed, "noserverfailed" }, { Opt_port, "port=%s" }, { Opt_rsize, "rsize=%s" }, @@ -637,6 +640,7 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, { NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" }, { NFS_MOUNT_UNSHARED, ",nosharecache", "" }, { NFS_MOUNT_NORESVPORT, ",noresvport", "" }, + { NFS_MOUNT_SERVERFAILED, ",serverfailed", "" }, { 0, NULL, NULL } }; const struct proc_nfs_info *nfs_infop; @@ -1330,6 +1334,11 @@ static int nfs_parse_mount_options(char *raw, case Opt_nomigration: mnt->options &= ~NFS_OPTION_MIGRATION; break; + case Opt_serverfailed: + case Opt_noserverfailed: + set_flag(mnt, NFS_MOUNT_SERVERFAILED, + token == Opt_serverfailed); + break; /* * options that take numeric values @@ -2192,6 +2201,8 @@ static int nfs_validate_text_mount_data(void *options, return -EINVAL; } +#define NFS_REMOUNT_CHANGE_FLAGS (NFS_MOUNT_SERVERFAILED) + #define NFS_REMOUNT_CMP_FLAGMASK ~(NFS_MOUNT_INTR \ | NFS_MOUNT_SECURE \ | NFS_MOUNT_TCP \ @@ -2200,7 +2211,8 @@ static int nfs_validate_text_mount_data(void *options, | NFS_MOUNT_NONLM \ | NFS_MOUNT_BROKEN_SUID \ | NFS_MOUNT_STRICTLOCK \ - | NFS_MOUNT_LEGACY_INTERFACE) + | NFS_MOUNT_LEGACY_INTERFACE \ + | NFS_REMOUNT_CHANGE_FLAGS) #define NFS_MOUNT_CMP_FLAGMASK (NFS_REMOUNT_CMP_FLAGMASK & \ ~(NFS_MOUNT_UNSHARED | NFS_MOUNT_NORESVPORT)) @@ -2209,12 +2221,15 @@ static int nfs_compare_and_set_remount_data(struct nfs_server *nfss, struct nfs_parsed_mount_data *data) { + int changed_flags_mask = data->flags_mask & NFS_REMOUNT_CHANGE_FLAGS; + struct rpc_clnt *cl = nfss->client; + if ((data->flags ^ nfss->flags) & NFS_REMOUNT_CMP_FLAGMASK || data->rsize != nfss->rsize || data->wsize != nfss->wsize || data->version != nfss->nfs_client->rpc_ops->version || data->minorversion != nfss->nfs_client->cl_minorversion || - !nfs_auth_info_match(&data->auth_info, nfss->client->cl_auth->au_flavor) || + !nfs_auth_info_match(&data->auth_info, cl->cl_auth->au_flavor) || data->acregmin != nfss->acregmin / HZ || data->acregmax != nfss->acregmax / HZ || data->acdirmin != nfss->acdirmin / HZ || @@ -2225,13 +2240,16 @@ nfs_compare_and_set_remount_data(struct nfs_server *nfss, (struct sockaddr *)&nfss->nfs_client->cl_addr)) return -EINVAL; - if (data->retrans != nfss->client->cl_timeout->to_retries || - data->timeo != (10U * nfss->client->cl_timeout->to_initval / HZ)) { + /* Update flags */ + nfss->flags = (nfss->flags & ~changed_flags_mask) | + (data->flags & changed_flags_mask); + + if (data->retrans != cl->cl_timeout->to_retries || + data->timeo != (10U * cl->cl_timeout->to_initval / HZ)) { /* Note that this will affect all mounts from the same server, * that use the same protocol. The timeouts are always forced * to be the same. */ - struct rpc_clnt *cl = nfss->client; if (cl->cl_timeout != &cl->cl_timeout_default) memcpy(&cl->cl_timeout_default, cl->cl_timeout, sizeof(struct rpc_timeout)); @@ -2239,6 +2257,8 @@ nfs_compare_and_set_remount_data(struct nfs_server *nfss, cl->cl_timeout_default.to_initval = data->timeo * HZ / 10U; } + cl->cl_kill_new_tasks = !!(nfss->flags & NFS_MOUNT_SERVERFAILED); + return 0; } diff --git a/include/uapi/linux/nfs_mount.h b/include/uapi/linux/nfs_mount.h index e44e00616ab5..8ad931cdca20 100644 --- a/include/uapi/linux/nfs_mount.h +++ b/include/uapi/linux/nfs_mount.h @@ -74,5 +74,6 @@ struct nfs_mount_data { #define NFS_MOUNT_LOCAL_FLOCK 0x100000 #define NFS_MOUNT_LOCAL_FCNTL 0x200000 +#define NFS_MOUNT_SERVERFAILED 0x400000 #endif