From: Fred Isaman Subject: Re: [PATCH 07/10] pnfs-submit: avoid race handling return on close Date: Tue, 15 Jun 2010 13:52:48 -0400 Message-ID: References: <1276566375-24566-1-git-send-email-iisaman@netapp.com> <1276566375-24566-2-git-send-email-iisaman@netapp.com> <1276566375-24566-3-git-send-email-iisaman@netapp.com> <1276566375-24566-4-git-send-email-iisaman@netapp.com> <1276566375-24566-5-git-send-email-iisaman@netapp.com> <1276566375-24566-6-git-send-email-iisaman@netapp.com> <1276566375-24566-7-git-send-email-iisaman@netapp.com> <1276566375-24566-8-git-send-email-iisaman@netapp.com> <4C17B304.4070308@panasas.com> <1276623230.8767.48.camel@heimdal.trondhjem.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Cc: Benny Halevy , Fred Isaman , linux-nfs@vger.kernel.org To: Trond Myklebust Return-path: Received: from mail-bw0-f46.google.com ([209.85.214.46]:35904 "EHLO mail-bw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752042Ab0FORwu convert rfc822-to-8bit (ORCPT ); Tue, 15 Jun 2010 13:52:50 -0400 Received: by bwz7 with SMTP id 7so2828763bwz.19 for ; Tue, 15 Jun 2010 10:52:48 -0700 (PDT) In-Reply-To: <1276623230.8767.48.camel-rJ7iovZKK19ZJLDQqaL3InhyD016LWXt@public.gmane.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Tue, Jun 15, 2010 at 1:33 PM, Trond Myklebust wrote: > On Tue, 2010-06-15 at 13:32 -0400, Fred Isaman wrote: >> On Tue, Jun 15, 2010 at 1:06 PM, Benny Halevy = wrote: >> > On Jun. 14, 2010, 21:46 -0400, Fred Isaman wr= ote: >> >> This prepares for the next patch. >> >> >> >> NOTE this doesn't really fix any current race, since >> >> layout going to NULL is OK. =A0But layout changing from NULL to n= onNULL >> >> is a real race that is not fixed >> >> >> >> Signed-off-by: Fred Isaman >> >> --- >> >> =A0fs/nfs/nfs4state.c =A0 =A0 =A0 =A0| =A0 =A05 +++-- >> >> =A0fs/nfs/pnfs.c =A0 =A0 =A0 =A0 =A0 =A0 | =A0 11 +++++++++++ >> >> =A0include/linux/nfs4_pnfs.h | =A0 =A02 ++ >> >> =A03 files changed, 16 insertions(+), 2 deletions(-) >> >> >> >> diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c >> >> index d5144bd..8a7a64c 100644 >> >> --- a/fs/nfs/nfs4state.c >> >> +++ b/fs/nfs/nfs4state.c >> >> @@ -594,11 +594,12 @@ static void __nfs4_close(struct path *path,= struct nfs4_state *state, >> >> =A0 =A0 =A0 } else { >> >> =A0#ifdef CONFIG_NFS_V4_1 >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 struct nfs_inode *nfsi =3D NFS_I(stat= e->inode); >> >> + =A0 =A0 =A0 =A0 =A0 =A0 int roc =3D nfs4_roc_iomode(nfsi); >> >> >> >> - =A0 =A0 =A0 =A0 =A0 =A0 if (has_layout(nfsi) && nfsi->layout.ro= c_iomode) { >> >> + =A0 =A0 =A0 =A0 =A0 =A0 if (roc) { >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 struct nfs4_pnfs_layo= ut_segment range; >> >> >> >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 range.iomode =3D nfsi->= layout.roc_iomode; >> >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 range.iomode =3D roc; >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 range.offset =3D 0; >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 range.length =3D NFS4= _MAX_UINT64; >> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 pnfs_return_layout(st= ate->inode, &range, NULL, >> >> diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c >> >> index 6def09c..bd11ec7 100644 >> >> --- a/fs/nfs/pnfs.c >> >> +++ b/fs/nfs/pnfs.c >> >> @@ -321,6 +321,17 @@ pnfs_unregister_layoutdriver(struct pnfs_lay= outdriver_type *ld_type) >> >> =A0#define BUG_ON_UNLOCKED_LO(lo) do {} while (0) >> >> =A0#endif /* CONFIG_SMP */ >> >> >> >> +int nfs4_roc_iomode(struct nfs_inode *nfsi) >> >> +{ >> >> + =A0 =A0 int rv =3D 0; >> >> + >> >> + =A0 =A0 spin_lock(&pnfs_spinlock); >> > >> > Why take the global lock rather than nfsi->lo_lock? >> > >> > Benny >> >> You are right. That would be a copy-paste error. > > What's an nfsi->lo_lock, and why do we need one? > > Trond > > It protects nfsi->layout and its contents. =46red