gfs2: Fix error handling in init_statfs

commit 5d49d3508b3c67201bd3e1bf7f4ef049111b7051 upstream.

On an error path, init_statfs calls iput(pn) after pn has already been put.
Fix that by setting pn to NULL after the initial iput.

Fixes: 97fd734ba17e ("gfs2: lookup local statfs inodes prior to journal recovery")
Cc: stable@vger.kernel.org # v5.10+
Reported-by: Jing Xiangfeng <jingxiangfeng@huawei.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index f2c6bbe5cdb81568bb8b4b7195536112c5192f57..ae9c5c1bdc508dd716138b4e61df23a75309c334 100644 (file)
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -670,6 +670,7 @@ static int init_statfs(struct gfs2_sbd *sdp)
        }
 
        iput(pn);
+       pn = NULL;
        ip = GFS2_I(sdp->sd_sc_inode);
        error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0,
                                   &sdp->sd_sc_gh);