On Thu, Jul 29, 2021 at 10:23:35PM +0300, Благодаренко Артём wrote:

Hello,

It looks like the fix b5776e7524afbd4569978ff790864755c438bba7 "ext4: fix potential htree index checksum corruption” introduced this regression.
I reverted it and my test from previous message passed the dangerous level of 1570000 names count.
Now test is still in progress. 2520000 names are already created.

I am searching the way to fix this.

Best regards,
Artem Blagodarenko.

Hi Artem, did you have a chance to take a look at some of the possible
fixes which I floated on this thread?

Do you have any objections if I take this and send it to Linus?

Thanks,

					- Ted

From fa8db30806b4e83981c65f18f98de33f804012d9 Mon Sep 17 00:00:00 2001
From: Theodore Ts'o <tytso@mit.edu>
Date: Wed, 4 Aug 2021 14:23:55 -0400
Subject: [PATCH] ext4: fix potential htree correuption when growing large_dir
 directories
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Commit b5776e7524af ("ext4: fix potential htree index checksum
corruption) removed a required restart when multiple levels of index
nodes need to be split.  Fix this to avoid directory htree corruptions
when using the large_dir feature.

Cc: stable@kernel.org # v5.11
Cc: Благодаренко Артём <redacted>
Fixes: b5776e7524af ("ext4: fix potential htree index checksum corruption)
Reported-by: Denis <redacted>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
---
 fs/ext4/namei.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 5fd56f616cf0..f3bbcd4efb56 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c

@@ -2517,7 +2517,7 @@ static int ext4_dx_add_entry(handle_t *handle, struct ext4_filename *fname,
 				goto journal_error;
 			err = ext4_handle_dirty_dx_node(handle, dir,
 							frame->bh);
-			if (err)
+			if (restart || err)
 				goto journal_error;
 		} else {
 			struct dx_root *dxroot;

-- 
2.31.0