From patchwork Mon Sep 24 19:56:30 2012 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Trond Myklebust X-Patchwork-Id: 1499631 Return-Path: X-Original-To: patchwork-linux-nfs@patchwork.kernel.org Delivered-To: patchwork-process-083081@patchwork1.kernel.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by patchwork1.kernel.org (Postfix) with ESMTP id 3C5EA3FE80 for ; Mon, 24 Sep 2012 19:58:26 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932131Ab2IXT6Z (ORCPT ); Mon, 24 Sep 2012 15:58:25 -0400 Received: from mx2.netapp.com ([216.240.18.37]:56248 "EHLO mx2.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932119Ab2IXT6Z (ORCPT ); Mon, 24 Sep 2012 15:58:25 -0400 X-IronPort-AV: E=Sophos;i="4.80,477,1344236400"; d="scan'208";a="693537119" Received: from smtp1.corp.netapp.com ([10.57.156.124]) by mx2-out.netapp.com with ESMTP; 24 Sep 2012 12:58:24 -0700 Received: from lade.trondhjem.org.com ([10.63.230.114]) by smtp1.corp.netapp.com (8.13.1/8.13.1/NTAP-1.6) with ESMTP id q8OJv9TU023988; Mon, 24 Sep 2012 12:58:20 -0700 (PDT) From: Trond Myklebust To: linux-nfs@vger.kernel.org Subject: [PATCH v3 12/28] NFSv4.1: Fix a race in the pNFS return-on-close code Date: Mon, 24 Sep 2012 15:56:30 -0400 Message-Id: <1348516606-22721-12-git-send-email-Trond.Myklebust@netapp.com> X-Mailer: git-send-email 1.7.11.4 In-Reply-To: <1348516606-22721-11-git-send-email-Trond.Myklebust@netapp.com> References: <1348516606-22721-1-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-2-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-3-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-4-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-5-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-6-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-7-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-8-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-9-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-10-git-send-email-Trond.Myklebust@netapp.com> <1348516606-22721-11-git-send-email-Trond.Myklebust@netapp.com> Sender: linux-nfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-nfs@vger.kernel.org If we sleep after dropping the inode->i_lock, then we are no longer atomic with respect to the rpc_wake_up() call in pnfs_layout_remove_lseg(). Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 8 +++----- fs/nfs/pnfs.c | 22 ++++++++++++---------- fs/nfs/pnfs.h | 4 ++-- 3 files changed, 17 insertions(+), 17 deletions(-) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 87702a0..49d5c7d 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2137,6 +2137,7 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data) { struct nfs4_closedata *calldata = data; struct nfs4_state *state = calldata->state; + struct inode *inode = calldata->inode; int call_close = 0; dprintk("%s: begin!\n", __func__); @@ -2170,16 +2171,13 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data) if (calldata->arg.fmode == 0) { task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE]; if (calldata->roc && - pnfs_roc_drain(calldata->inode, &calldata->roc_barrier)) { - rpc_sleep_on(&NFS_SERVER(calldata->inode)->roc_rpcwaitq, - task, NULL); + pnfs_roc_drain(inode, &calldata->roc_barrier, task)) goto out; - } } nfs_fattr_init(calldata->res.fattr); calldata->timestamp = jiffies; - if (nfs4_setup_sequence(NFS_SERVER(calldata->inode), + if (nfs4_setup_sequence(NFS_SERVER(inode), &calldata->arg.seq_args, &calldata->res.seq_res, task)) diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index d0f865f..f48bfef 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -807,27 +807,29 @@ void pnfs_roc_set_barrier(struct inode *ino, u32 barrier) spin_unlock(&ino->i_lock); } -bool pnfs_roc_drain(struct inode *ino, u32 *barrier) +bool pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task) { struct nfs_inode *nfsi = NFS_I(ino); + struct pnfs_layout_hdr *lo; struct pnfs_layout_segment *lseg; + u32 current_seqid; bool found = false; spin_lock(&ino->i_lock); list_for_each_entry(lseg, &nfsi->layout->plh_segs, pls_list) if (test_bit(NFS_LSEG_ROC, &lseg->pls_flags)) { + rpc_sleep_on(&NFS_SERVER(ino)->roc_rpcwaitq, task, NULL); found = true; - break; + goto out; } - if (!found) { - struct pnfs_layout_hdr *lo = nfsi->layout; - u32 current_seqid = be32_to_cpu(lo->plh_stateid.seqid); + lo = nfsi->layout; + current_seqid = be32_to_cpu(lo->plh_stateid.seqid); - /* Since close does not return a layout stateid for use as - * a barrier, we choose the worst-case barrier. - */ - *barrier = current_seqid + atomic_read(&lo->plh_outstanding); - } + /* Since close does not return a layout stateid for use as + * a barrier, we choose the worst-case barrier. + */ + *barrier = current_seqid + atomic_read(&lo->plh_outstanding); +out: spin_unlock(&ino->i_lock); return found; } diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 9735031..aa9fa1b 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -210,7 +210,7 @@ int pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, bool pnfs_roc(struct inode *ino); void pnfs_roc_release(struct inode *ino); void pnfs_roc_set_barrier(struct inode *ino, u32 barrier); -bool pnfs_roc_drain(struct inode *ino, u32 *barrier); +bool pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task); void pnfs_set_layoutcommit(struct nfs_write_data *wdata); void pnfs_cleanup_layoutcommit(struct nfs4_layoutcommit_data *data); int pnfs_layoutcommit_inode(struct inode *inode, bool sync); @@ -442,7 +442,7 @@ pnfs_roc_set_barrier(struct inode *ino, u32 barrier) } static inline bool -pnfs_roc_drain(struct inode *ino, u32 *barrier) +pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task) { return false; }