The Linux Kernel Mailing List
 help / color / mirror / Atom feed
From: Chao Yu <chao@kernel.org>
To: Zhiguo Niu <niuzhiguo84@gmail.com>
Cc: chao@kernel.org, Zhiguo Niu <zhiguo.niu@unisoc.com>,
	daehojeong@google.com, jaegeuk@kernel.org,
	linux-f2fs-devel@lists.sourceforge.net,
	linux-kernel@vger.kernel.org, ke.wang@unisoc.com,
	Hao_hao.Wang@unisoc.com
Subject: Re: [RFC PATCH] f2fs: remove some redundant flow about FI_ATOMIC_DIRTIED
Date: Wed, 26 Mar 2025 20:10:44 +0800	[thread overview]
Message-ID: <8a72691e-28ef-4b6b-a855-62fd6c2743dc@kernel.org> (raw)
In-Reply-To: <CAHJ8P3JWABsntymD3u5=0YR7=0it5x0PP49S4ftwBEjC1UreOA@mail.gmail.com>

On 3/26/25 18:51, Zhiguo Niu wrote:
> Chao Yu <chao@kernel.org> 于2025年3月26日周三 17:26写道:
>>
>> On 3/26/25 16:46, Zhiguo Niu wrote:
>>> Commit fccaa81de87e ("f2fs: prevent atomic file from being dirtied before commit")
>>> adds the processing of FI_ATOMIC_DIRTIED in the following two positions,
>>> [1]
>>> f2fs_commit_atomic_write
>>>  - __f2fs_commit_atomic_write
>>>   - sbi->committed_atomic_block += fi->atomic_write_cnt;
>>>   - set_inode_flag(inode, FI_ATOMIC_COMMITTED);
>>>   - if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) {
>>>   -    clear_inode_flag(inode, FI_ATOMIC_DIRTIED);
>>>   -    f2fs_mark_inode_dirty_sync(inode, true);
>>>   - }
>>> [2]
>>> f2fs_abort_atomic_write
>>>   - if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) {
>>>   -    clear_inode_flag(inode, FI_ATOMIC_DIRTIED);
>>>   -    f2fs_mark_inode_dirty_sync(inode, true);
>>>   - }
>>>
>>> but [1] seems to be redundant:
>>> The atomic file flag FI_ATOMIC_FILE is still set here, so f2fs_mark_inode_dirty_sync
>>> still does not set the dirty state to vfs. If FI_ATOMIC_DIRTIED was originally set
>>> when atomic file is committing, then FI_ATOMIC_DIRTIED is just cleared here, and
>>> then do the repeating action of setting FI_ATOMIC_DIRTIED?
>>> So is it enough to do this only in [2]?
>>
>> Hi Zhiguo,
>>
>> I checked the code again, finally, I got this, could you please take
>> a look?
>>
>> Ping Daeho as well.
>>
>> Subject: [PATCH] f2fs: fix to set atomic write status more clear
>>
>> 1. After we start atomic write in a database file, before committing
>> all data, we'd better not set inode w/ vfs dirty status to avoid
>> redundant updates, instead, we only set inode w/ atomic dirty status.
>>
>> 2. After we commit all data, before committing metadata, we need to
>> clear atomic dirty status, and set vfs dirty status to allow vfs flush
>> dirty inode.
>>
> Hi Chao,
> these looks more clear.
>> Signed-off-by: Chao Yu <chao@kernel.org>
>> ---
>>  fs/f2fs/inode.c   |  4 +++-
>>  fs/f2fs/segment.c | 10 ++++++----
>>  fs/f2fs/super.c   |  4 +++-
>>  3 files changed, 12 insertions(+), 6 deletions(-)
>>
>> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
>> index 5c8634eaef7b..f5991e8751b9 100644
>> --- a/fs/f2fs/inode.c
>> +++ b/fs/f2fs/inode.c
>> @@ -34,7 +34,9 @@ void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync)
>>         if (f2fs_inode_dirtied(inode, sync))
>>                 return;
>>
>> -       if (f2fs_is_atomic_file(inode))
>> +       /* only atomic file w/ FI_ATOMIC_COMMITTED can be set vfs dirty */
>> +       if (f2fs_is_atomic_file(inode) &&
>> +                       !is_inode_flag_set(inode, FI_ATOMIC_COMMITTED))
>>                 return;
>>
>>         mark_inode_dirty_sync(inode);
>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
>> index dc360b4b0569..28659a71891a 100644
>> --- a/fs/f2fs/segment.c
>> +++ b/fs/f2fs/segment.c
>> @@ -376,10 +376,12 @@ static int __f2fs_commit_atomic_write(struct inode *inode)
>>         } else {
>>                 sbi->committed_atomic_block += fi->atomic_write_cnt;
>>                 set_inode_flag(inode, FI_ATOMIC_COMMITTED);
>> -               if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) {
>> -                       clear_inode_flag(inode, FI_ATOMIC_DIRTIED);
>> -                       f2fs_mark_inode_dirty_sync(inode, true);
>> -               }
>> +
>> +               f2fs_bug_on(sbi, !is_inode_flag_set(inode, FI_ATOMIC_DIRTIED));
> but FI_ATOMIC_DIRTIED may  not be set when atomic file is committing?
> thanks!

inc_valid_block_count() will set FI_ATOMIC_DIRTIED for inode at least?

- __f2fs_commit_atomic_write
 - __replace_atomic_write_block
  - inc_valid_block_count
   - f2fs_i_blocks_write
    - f2fs_mark_inode_dirty_sync

Thanks,

>> +
>> +               /* clear atomic dirty status and set vfs dirty status */
>> +               clear_inode_flag(inode, FI_ATOMIC_DIRTIED);
>> +               f2fs_mark_inode_dirty_sync(inode, true);
>>         }
>>
>>         __complete_revoke_list(inode, &revoke_list, ret ? true : false);
>> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
>> index 9a42a1323f42..a5cc9f6ee16a 100644
>> --- a/fs/f2fs/super.c
>> +++ b/fs/f2fs/super.c
>> @@ -1532,7 +1532,9 @@ int f2fs_inode_dirtied(struct inode *inode, bool sync)
>>         }
>>         spin_unlock(&sbi->inode_lock[DIRTY_META]);
>>
>> -       if (!ret && f2fs_is_atomic_file(inode))
>> +       /* if atomic write is not committed, set inode w/ atomic dirty */
>> +       if (!ret && f2fs_is_atomic_file(inode) &&
>> +                       !is_inode_flag_set(inode, FI_ATOMIC_COMMITTED))
>>                 set_inode_flag(inode, FI_ATOMIC_DIRTIED);
>>
>>         return ret;
>> --
>> 2.48.1
>>
>>
>>>
>>> Cc: Daeho Jeong <daehojeong@google.com>
>>> Fixes: fccaa81de87e ("f2fs: prevent atomic file from being dirtied before commit")
>>> Signed-off-by: Zhiguo Niu <zhiguo.niu@unisoc.com>
>>> ---
>>>  fs/f2fs/segment.c | 4 ----
>>>  1 file changed, 4 deletions(-)
>>>
>>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
>>> index 396ef71..d4ea3af 100644
>>> --- a/fs/f2fs/segment.c
>>> +++ b/fs/f2fs/segment.c
>>> @@ -376,10 +376,6 @@ static int __f2fs_commit_atomic_write(struct inode *inode)
>>>       } else {
>>>               sbi->committed_atomic_block += fi->atomic_write_cnt;
>>>               set_inode_flag(inode, FI_ATOMIC_COMMITTED);
>>> -             if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) {
>>> -                     clear_inode_flag(inode, FI_ATOMIC_DIRTIED);
>>> -                     f2fs_mark_inode_dirty_sync(inode, true);
>>> -             }
>>>       }
>>>
>>>       __complete_revoke_list(inode, &revoke_list, ret ? true : false);
>>


  reply	other threads:[~2025-03-26 12:10 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-03-26  8:46 [RFC PATCH] f2fs: remove some redundant flow about FI_ATOMIC_DIRTIED Zhiguo Niu
2025-03-26  9:26 ` Chao Yu
2025-03-26 10:51   ` Zhiguo Niu
2025-03-26 12:10     ` Chao Yu [this message]
2025-03-26 18:51       ` [f2fs-dev] " Daeho Jeong
2025-03-27  3:21         ` Chao Yu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=8a72691e-28ef-4b6b-a855-62fd6c2743dc@kernel.org \
    --to=chao@kernel.org \
    --cc=Hao_hao.Wang@unisoc.com \
    --cc=daehojeong@google.com \
    --cc=jaegeuk@kernel.org \
    --cc=ke.wang@unisoc.com \
    --cc=linux-f2fs-devel@lists.sourceforge.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=niuzhiguo84@gmail.com \
    --cc=zhiguo.niu@unisoc.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox