From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff Layton Subject: Re: [PATCH 5/5] ceph: avoid null pointer derefernece in case of utsname() return NULL Date: Tue, 12 Sep 2017 09:15:20 -0400 Message-ID: <1505222120.28831.10.camel@redhat.com> References: <20170912025351.42147-1-zyan@redhat.com> <20170912025351.42147-6-zyan@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-yw0-f178.google.com ([209.85.161.178]:34638 "EHLO mail-yw0-f178.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751379AbdILNPX (ORCPT ); Tue, 12 Sep 2017 09:15:23 -0400 Received: by mail-yw0-f178.google.com with SMTP id r85so28103514ywg.1 for ; Tue, 12 Sep 2017 06:15:23 -0700 (PDT) In-Reply-To: <20170912025351.42147-6-zyan@redhat.com> Sender: ceph-devel-owner@vger.kernel.org List-ID: To: "Yan, Zheng" , ceph-devel@vger.kernel.org, idryomov@gmail.com On Tue, 2017-09-12 at 10:53 +0800, Yan, Zheng wrote: > utsname() can return NULL while process is exiting. kernel releases > file locks during process exits. We send request to mds when releasing > file lock. So it's possible that we open mds session while process is > exiting. utsname() is called in create_session_open_msg() > > Signed-off-by: "Yan, Zheng" > --- > fs/ceph/mds_client.c | 6 ++++-- > fs/ceph/mds_client.h | 3 +++ > 2 files changed, 7 insertions(+), 2 deletions(-) > > diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c > index 26893cc1fbee..ab6b998d3f44 100644 > --- a/fs/ceph/mds_client.c > +++ b/fs/ceph/mds_client.c > @@ -884,8 +884,8 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6 > void *p; > > const char* metadata[][2] = { > - {"hostname", utsname()->nodename}, > - {"kernel_version", utsname()->release}, > + {"hostname", mdsc->nodename}, > + {"kernel_version", init_utsname()->release}, > {"entity_id", opt->name ? : ""}, > {"root", fsopt->server_path ? : "/"}, > {NULL, NULL} > @@ -3558,6 +3558,8 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc) > init_rwsem(&mdsc->pool_perm_rwsem); > mdsc->pool_perm_tree = RB_ROOT; > > + strncpy(mdsc->nodename, utsname()->nodename, > + sizeof(mdsc->nodename) - 1); > return 0; > } > > diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h > index db57ae98ed34..636d6b2ec49c 100644 > --- a/fs/ceph/mds_client.h > +++ b/fs/ceph/mds_client.h > @@ -8,6 +8,7 @@ > #include > #include > #include > +#include > > #include > #include > @@ -368,6 +369,8 @@ struct ceph_mds_client { > > struct rw_semaphore pool_perm_rwsem; > struct rb_root pool_perm_tree; > + > + char nodename[__NEW_UTS_LEN + 1]; > }; > > extern const char *ceph_mds_op_name(int op); (nit: might want to add a link to the tracker bug in the description) Reviewed-by: Jeff Layton