During mount if rootfh changes fsid then fs-core layer dereferences and calls NULL pointer. nfs_fhget() sets rootinode->i_op to nfs_mountpoint_inode_operations. Then d_alloc_and_lookup() calls i_op->lookup() that is NULL. The problem is: rpc_ops->getroot() and rpc_ops->getattr() return different fsid due to server replies. So just refresh fsid, as RFC3530 doesn't specify behavior in case of rootfh changes fsid. Oops: BUG: unable to handle kernel NULL pointer dereference at (null) IP: [< (null)>] (null) stack trace: d_alloc_and_lookup+0x4c/0x74 do_lookup+0x1e3/0x280 link_path_walk+0x12e/0xab0 nfs4_remote_get_sb+0x56/0x2c0 [nfs] path_walk+0x67/0xe0 vfs_path_lookup+0x8e/0x100 nfs_follow_remote_path+0x16f/0x3e0 [nfs] nfs4_try_mount+0x6f/0xd0 [nfs] nfs_get_sb+0x269/0x400 [nfs] vfs_kern_mount+0x8a/0x1f0 do_kern_mount+0x52/0x130 do_mount+0x20a/0x260 sys_mount+0x90/0xe0 system_call_fastpath+0x16/0x1b Signed-off-by: Vitaliy Gusev <gusev.vitaliy@xxxxxxxxxxx> --- fs/nfs/getroot.c | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/fs/nfs/getroot.c b/fs/nfs/getroot.c index b5ffe8f..7979652 100644 --- a/fs/nfs/getroot.c +++ b/fs/nfs/getroot.c @@ -199,6 +199,10 @@ struct dentry *nfs4_get_root(struct super_block *sb, struct nfs_fh *mntfh) goto out; } + if (fattr->valid & NFS_ATTR_FATTR_FSID && + !nfs_fsid_equal(&server->fsid, &fattr->fsid)) + memcpy(&server->fsid, &fattr->fsid, sizeof(server->fsid)); + inode = nfs_fhget(sb, mntfh, fattr); if (IS_ERR(inode)) { dprintk("nfs_get_root: get root inode failed\n"); -- 1.7.1 -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html