Re: [PATCH v3] NFS: put net on idr allocation failure

From: Myklebust, Trond
Date: Tue May 22 2012 - 10:45:08 EST


On Tue, 2012-05-22 at 18:41 +0400, Stanislav Kinsbursky wrote:
> On 22.05.2012 18:37, Myklebust, Trond wrote:
> > On Tue, 2012-05-22 at 12:03 +0400, Stanislav Kinsbursky wrote:
> >> Signed-off-by: Stanislav Kinsbursky<skinsbursky@xxxxxxxxxxxxx>
> >> ---
> >> fs/nfs/client.c | 4 +++-
> >> 1 files changed, 3 insertions(+), 1 deletions(-)
> >>
> >> diff --git a/fs/nfs/client.c b/fs/nfs/client.c
> >> index 44cd70f..ae29d4f 100644
> >> --- a/fs/nfs/client.c
> >> +++ b/fs/nfs/client.c
> >> @@ -176,8 +176,10 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_
> >>
> >> #ifdef CONFIG_NFS_V4
> >> err = nfs_get_cb_ident_idr(clp, cl_init->minorversion);
> >> - if (err)
> >> + if (err) {
> >> + put_net(clp->net);
> >> goto error_cleanup;
> >> + }
> >>
> >> spin_lock_init(&clp->cl_lock);
> >> INIT_DELAYED_WORK(&clp->cl_renewd, nfs4_renew_state);
> >>
> >
> > Let's just move the actual get_net() call to the end of the function.
> >
>
> Ok. But it will look a bit strange, like this:
>
> clp->net = net;
> ...
> nfs_get_cb_ident_idr
> ...
> get_net(clp->net)
>
> Or I can pass net to nfs_get_cb_ident_idr() as a parameter.
> Which solution is more preferable&?

Just put the get_net() at the end. It doesn't matter if it looks a bit
odd: the intent is still obvious.

--
Trond Myklebust
Linux NFS client maintainer

NetApp
Trond.Myklebust@xxxxxxxxxx
www.netapp.com

¢éì®&Þ~º&¶¬–+-±éÝ¥Šw®žË±Êâmébžìdz¹Þ)í…æèw*jg¬±¨¶‰šŽŠÝj/êäz¹ÞŠà2ŠÞ¨è­Ú&¢)ß«a¶Úþø®G«éh®æj:+v‰¨Šwè†Ù>Wš±êÞiÛaxPjØm¶Ÿÿà -»+ƒùdš_