Received-SPF: pass (google.com: domain of linux-kernel+bounces-171299-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Subject: Re: [PATCH 4/9] jbd2: move repeat tag around to remove a repeat check
 of b_frozen_data
To: Kemeng Shi <shikemeng@huaweicloud.com>
Cc: linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org, tytso@mit.edu,
 jack@suse.com
References: <20240506141801.1165315-1-shikemeng@huaweicloud.com>
 <20240506141801.1165315-5-shikemeng@huaweicloud.com>
From: Zhang Yi <yi.zhang@huaweicloud.com>
Message-ID: <b2fa97c7-3f1c-321d-038e-7fb46420d24b@huaweicloud.com>
Date: Tue, 7 May 2024 20:41:43 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101
 Thunderbird/78.12.0
Precedence: bulk
MIME-Version: 1.0
In-Reply-To: <20240506141801.1165315-5-shikemeng@huaweicloud.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit

On 2024/5/6 22:17, Kemeng Shi wrote:
> We make sure b_frozen_data is not NULL before jump to "repeat" tag, move
> "repeat" tag around to remove repeat check of b_frozen_data.
> 
> Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
> ---
>  fs/jbd2/journal.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
> index 9a35d0c5b38c..77fcdc76fdfd 100644
> --- a/fs/jbd2/journal.c
> +++ b/fs/jbd2/journal.c
> @@ -353,12 +353,12 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  	atomic_set(&new_bh->b_count, 1);
>  
>  	spin_lock(&jh_in->b_state_lock);
> -repeat:
>  	/*
>  	 * If a new transaction has already done a buffer copy-out, then
>  	 * we use that version of the data for the commit.
>  	 */
>  	if (jh_in->b_frozen_data) {
> +repeat:
>  		done_copy_out = 1;
>  		new_folio = virt_to_folio(jh_in->b_frozen_data);
>  		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
> 

I suppose we could drop the repeat tag entirely, just set the new_folio and
new_offset, and then goto handle do_escape. We don't need to call
jbd2_buffer_frozen_trigger() and check for escaping again, is that right?

Thanks,
Yi.