From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752827AbaFEEVD (ORCPT ); Thu, 5 Jun 2014 00:21:03 -0400 Received: from mail.linuxfoundation.org ([140.211.169.12]:43989 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752658AbaFEEUt (ORCPT ); Thu, 5 Jun 2014 00:20:49 -0400 From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Trond Myklebust , Benny Halevy , Ben Hutchings , Rui Xiang Subject: [PATCH 3.4 092/214] NFSv4.1: Fix a race in pNFS layoutcommit Date: Wed, 4 Jun 2014 21:17:35 -0700 Message-Id: <20140605041652.013739566@linuxfoundation.org> X-Mailer: git-send-email 2.0.0 In-Reply-To: <20140605041639.638675216@linuxfoundation.org> References: <20140605041639.638675216@linuxfoundation.org> User-Agent: quilt/0.60-1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org 3.4-stable review patch. If anyone has any objections, please let me know. ------------------ From: Trond Myklebust commit a073dbff359f4741013ae4b8395f5364c5e00b48 upstream. We need to clear the NFS_LSEG_LAYOUTCOMMIT bits atomically with the NFS_INO_LAYOUTCOMMIT bit, otherwise we may end up with situations where the two are out of sync. The first half of the problem is to ensure that pnfs_layoutcommit_inode clears the NFS_LSEG_LAYOUTCOMMIT bit through pnfs_list_write_lseg. We still need to keep the reference to those segments until the RPC call is finished, so in order to make it clear _where_ those references come from, we add a helper pnfs_list_write_lseg_done() that cleans up after pnfs_list_write_lseg. Signed-off-by: Trond Myklebust Acked-by: Benny Halevy [bwh: Backported to 3.2: s/pnfs_put_lseg/put_lseg/] Signed-off-by: Ben Hutchings Cc: Rui Xiang Signed-off-by: Greg Kroah-Hartman --- fs/nfs/nfs4proc.c | 14 -------------- fs/nfs/pnfs.c | 19 ++++++++++++++++++- 2 files changed, 18 insertions(+), 15 deletions(-) --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -6297,22 +6297,8 @@ nfs4_layoutcommit_done(struct rpc_task * static void nfs4_layoutcommit_release(void *calldata) { struct nfs4_layoutcommit_data *data = calldata; - struct pnfs_layout_segment *lseg, *tmp; - unsigned long *bitlock = &NFS_I(data->args.inode)->flags; pnfs_cleanup_layoutcommit(data); - /* Matched by references in pnfs_set_layoutcommit */ - list_for_each_entry_safe(lseg, tmp, &data->lseg_list, pls_lc_list) { - list_del_init(&lseg->pls_lc_list); - if (test_and_clear_bit(NFS_LSEG_LAYOUTCOMMIT, - &lseg->pls_flags)) - put_lseg(lseg); - } - - clear_bit_unlock(NFS_INO_LAYOUTCOMMITTING, bitlock); - smp_mb__after_clear_bit(); - wake_up_bit(bitlock, NFS_INO_LAYOUTCOMMITTING); - put_rpccred(data->cred); kfree(data); } --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1381,11 +1381,27 @@ static void pnfs_list_write_lseg(struct list_for_each_entry(lseg, &NFS_I(inode)->layout->plh_segs, pls_list) { if (lseg->pls_range.iomode == IOMODE_RW && - test_bit(NFS_LSEG_LAYOUTCOMMIT, &lseg->pls_flags)) + test_and_clear_bit(NFS_LSEG_LAYOUTCOMMIT, &lseg->pls_flags)) list_add(&lseg->pls_lc_list, listp); } } +static void pnfs_list_write_lseg_done(struct inode *inode, struct list_head *listp) +{ + struct pnfs_layout_segment *lseg, *tmp; + unsigned long *bitlock = &NFS_I(inode)->flags; + + /* Matched by references in pnfs_set_layoutcommit */ + list_for_each_entry_safe(lseg, tmp, listp, pls_lc_list) { + list_del_init(&lseg->pls_lc_list); + put_lseg(lseg); + } + + clear_bit_unlock(NFS_INO_LAYOUTCOMMITTING, bitlock); + smp_mb__after_clear_bit(); + wake_up_bit(bitlock, NFS_INO_LAYOUTCOMMITTING); +} + void pnfs_set_lo_fail(struct pnfs_layout_segment *lseg) { if (lseg->pls_range.iomode == IOMODE_RW) { @@ -1434,6 +1450,7 @@ void pnfs_cleanup_layoutcommit(struct nf if (nfss->pnfs_curr_ld->cleanup_layoutcommit) nfss->pnfs_curr_ld->cleanup_layoutcommit(data); + pnfs_list_write_lseg_done(data->args.inode, &data->lseg_list); } /*