From mboxrd@z Thu Jan 1 00:00:00 1970 From: Steven Whitehouse Date: Tue, 12 Aug 2008 10:10:04 +0100 Subject: [Cluster-devel] [GFS2 Patch] bz458289: rm on multiple nodes causes panic In-Reply-To: <1218531200.3426.1.camel@localhost.localdomain> References: <1218512508.9521.117.camel@technetium.msp.redhat.com> <1218531200.3426.1.camel@localhost.localdomain> Message-ID: <1218532204.3426.3.camel@localhost.localdomain> List-Id: To: cluster-devel.redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Hi, Actually now I read the code again, I think its your diff thats backwards :-) Can you send me one the other way around? Steve. On Tue, 2008-08-12 at 09:53 +0100, Steven Whitehouse wrote: > Hi, > > This looks like the correct lock order to me. I guess it must be the > "other" process thats violating the correct order? > > Steve. > > On Mon, 2008-08-11 at 22:41 -0500, Bob Peterson wrote: > > This patch fixes a problem whereby simultaneous delete operations > > (e.g. rm -fR *) from multiple nodes on the same GFS2 file system > > can cause kernel panics, hangs, and/or memory corruption. > > > > Regards, > > > > Bob Peterson > > > > Signed-off-by: Bob Peterson > > -- > > fs/gfs2/ops_inode.c | 24 ++++++------------------ > > 1 files changed, 6 insertions(+), 18 deletions(-) > > > > diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c > > index e2c62f7..a072c9a 100644 > > --- a/fs/gfs2/ops_inode.c > > +++ b/fs/gfs2/ops_inode.c > > @@ -288,25 +288,17 @@ static int gfs2_unlink(struct inode *dir, struct dentry *dentry) > > gfs2_holder_init(rgd->rd_gl, LM_ST_EXCLUSIVE, 0, ghs + 2); > > > > > > - error = gfs2_glock_nq(ghs); /* parent */ > > - if (error) > > - goto out_parent; > > - > > - error = gfs2_glock_nq(ghs + 1); /* child */ > > - if (error) > > - goto out_child; > > - > > - error = gfs2_glock_nq(ghs + 2); /* rgrp */ > > + error = gfs2_glock_nq_m(3, ghs); > > if (error) > > - goto out_rgrp; > > + goto out; > > > > error = gfs2_unlink_ok(dip, &dentry->d_name, ip); > > if (error) > > - goto out_rgrp; > > + goto out; > > > > error = gfs2_trans_begin(sdp, 2*RES_DINODE + RES_LEAF + RES_RG_BIT, 0); > > if (error) > > - goto out_rgrp; > > + goto out; > > > > error = gfs2_dir_del(dip, &dentry->d_name); > > if (error) > > @@ -316,14 +308,10 @@ static int gfs2_unlink(struct inode *dir, struct dentry *dentry) > > > > out_end_trans: > > gfs2_trans_end(sdp); > > - gfs2_glock_dq(ghs + 2); > > -out_rgrp: > > +out: > > + gfs2_glock_dq_m(3, ghs); > > gfs2_holder_uninit(ghs + 2); > > - gfs2_glock_dq(ghs + 1); > > -out_child: > > gfs2_holder_uninit(ghs + 1); > > - gfs2_glock_dq(ghs); > > -out_parent: > > gfs2_holder_uninit(ghs); > > gfs2_glock_dq_uninit(&ri_gh); > > return error; > > > > >