Hello Teodore,
Your one-line fix looks good.
I have tested it. 1560000 names created successfully.
But the patch with refactoring doesn’t work. I got this messages
1480000 names created
1520000 names created
ln: failed to access 'n0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001520519': Bad message
ln: failed to access 'n0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001520520': Bad message
[ 7699.212018] EXT4-fs error (device loop0): dx_probe:887: inode #2: block 144843: comm ln: Directory index failed checksum
[ 7699.216001] EXT4-fs error (device loop0): dx_probe:887: inode #2: block 144843: comm ln: Directory index failed checksum
I have no objections to send your one-line fix, but we need to double check refactoring.
Best regards,
Artem Blagodarenko
> On 4 Aug 2021, at 22:25, Theodore Ts'o <tytso@xxxxxxx> wrote:
>
> 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@xxxxxxx>
> 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@xxxxxxxxxx # v5.11
> Cc: Благодаренко Артём <artem.blagodarenko@xxxxxxxxx>
> Fixes: b5776e7524af ("ext4: fix potential htree index checksum corruption)
> Reported-by: Denis <denis@xxxxxxxxxxxxx>
> Signed-off-by: Theodore Ts'o <tytso@xxxxxxx>
> ---
> 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
>
