ocfs2: fix mounting crash if journal is not alloced

Patch series "rewrite error handling during mounting stage".


This patch (of 5):

After commit da5e7c87827e8 ("ocfs2: cleanup journal init and shutdown"),
journal init later than before, it makes NULL pointer access in free
routine.

Crash flow:

ocfs2_fill_super
 + ocfs2_mount_volume
 |  + ocfs2_dlm_init //fail & return, osb->journal is NULL.
 |  + ...
 |  + ocfs2_check_volume //no chance to init osb->journal
 |
 + ...
 + ocfs2_dismount_volume
    ocfs2_release_system_inodes
      ...
       evict
        ...
         ocfs2_clear_inode
          ocfs2_checkpoint_inode
           ocfs2_ci_fully_checkpointed
            time_after(journal->j_trans_id, ci->ci_last_trans)
             + journal is empty, crash!

For fixing, there are three solutions:

1> Partly revert commit da5e7c87827e8

   For avoiding kernel crash, this make sense for us.  We only
   concerned whether there has any non-system inode access before dlm
   init.  The answer is NO.  And all journal replay/recovery handling
   happen after dlm & journal init done.  So this method is not graceful
   but workable.

2> Add osb->journal check in free inode routine (eg ocfs2_clear_inode)

   The fix code is special for mounting phase, but it will continue
   working after mounting stage.  In another word, this method adds
   useless code in normal inode free flow.

3> Do directly free inode in mounting phase

   This method is brutal/complex and may introduce unsafe code,
   currently maintainer didn't like.

At last, we chose method <1> and did partly reverted job.  We reverted
journal init codes, and kept cleanup codes flow.

Link: https://lkml.kernel.org/r/20220424130952.2436-1-heming.zhao@suse.com
Link: https://lkml.kernel.org/r/20220424130952.2436-2-heming.zhao@suse.com
Fixes: da5e7c87827e8 ("ocfs2: cleanup journal init and shutdown")
Signed-off-by: Heming Zhao <heming.zhao@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Jun Piao <piaojun@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

1 files changed, 15 insertions, 0 deletions

diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 477cdf94122e..311433c69a3f 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -2195,6 +2195,15 @@ static int ocfs2_initialize_super(struct super_block *sb,
 
 	get_random_bytes(&osb->s_next_generation, sizeof(u32));
 
+	/*
+	 * FIXME
+	 * This should be done in ocfs2_journal_init(), but any inode
+	 * writes back operation will cause the filesystem to crash.
+	 */
+	status = ocfs2_journal_alloc(osb);
+	if (status < 0)
+		goto bail;
+
 	INIT_WORK(&osb->dquot_drop_work, ocfs2_drop_dquot_refs);
 	init_llist_head(&osb->dquot_drop_list);
 
@@ -2483,6 +2492,12 @@ static void ocfs2_delete_osb(struct ocfs2_super *osb)
 
 	kfree(osb->osb_orphan_wipes);
 	kfree(osb->slot_recovery_generations);
+	/* FIXME
+	 * This belongs in journal shutdown, but because we have to
+	 * allocate osb->journal at the middle of ocfs2_initialize_super(),
+	 * we free it here.
+	 */
+	kfree(osb->journal);
 	kfree(osb->local_alloc_copy);
 	kfree(osb->uuid_str);
 	kfree(osb->vol_label);