linux-nilfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ryusuke Konishi <konishi.ryusuke@gmail.com>
To: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: linux-fsdevel@vger.kernel.org, linux-nilfs@vger.kernel.org
Subject: Re: [PATCH 4/4] nilfs2: Convert metadata aops from writepage to writepages
Date: Thu, 3 Oct 2024 20:49:33 +0900	[thread overview]
Message-ID: <CAKFNMon0yfSPNcPG0h90DG=c9-anbri1pgZR6Qtc2a8kPG7+LQ@mail.gmail.com> (raw)
In-Reply-To: <20241002150036.1339475-5-willy@infradead.org>

On Thu, Oct 3, 2024 at 12:00 AM Matthew Wilcox (Oracle) wrote:
>
> By implementing ->writepages instead of ->writepage, we remove a
> layer of indirect function calls from the writeback path and the
> last use of struct page in nilfs2.
>
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> ---
>  fs/nilfs2/mdt.c | 19 +++++++++++++++----
>  1 file changed, 15 insertions(+), 4 deletions(-)
>
> diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
> index ceb7dc0b5bad..4f4a935fcdc5 100644
> --- a/fs/nilfs2/mdt.c
> +++ b/fs/nilfs2/mdt.c
> @@ -396,10 +396,9 @@ int nilfs_mdt_fetch_dirty(struct inode *inode)
>         return test_bit(NILFS_I_DIRTY, &ii->i_state);
>  }
>
> -static int
> -nilfs_mdt_write_page(struct page *page, struct writeback_control *wbc)
> +static int nilfs_mdt_write_folio(struct folio *folio,
> +               struct writeback_control *wbc)
>  {
> -       struct folio *folio = page_folio(page);
>         struct inode *inode = folio->mapping->host;
>         struct super_block *sb;
>         int err = 0;
> @@ -432,11 +431,23 @@ nilfs_mdt_write_page(struct page *page, struct writeback_control *wbc)
>         return err;
>  }
>
> +static int nilfs_mdt_writeback(struct address_space *mapping,
> +               struct writeback_control *wbc)
> +{
> +       struct folio *folio = NULL;
> +       int error;
> +

> +       while ((folio = writeback_iter(mapping, wbc, folio, &error)))
> +               nilfs_mdt_write_folio(folio, wbc);

In order to catch and return the error returned by
nilfs_mdt_write_folio, I think it's necessary to assign the return
value to the variable "error" in the loop as follows:

+       while ((folio = writeback_iter(mapping, wbc, folio, &error)))
+               error = nilfs_mdt_write_folio(folio, wbc);

> +
> +       return error;
> +}
>
>  static const struct address_space_operations def_mdt_aops = {
>         .dirty_folio            = block_dirty_folio,
>         .invalidate_folio       = block_invalidate_folio,
> -       .writepage              = nilfs_mdt_write_page,
> +       .writepages             = nilfs_mdt_writeback,

> +       .migrate_folio          = buffer_migrate_folio,
>  };

And, this also caused kernel panics with the fsstress command.  As
with patch 1/4, it was necessary to use buffer_migrate_folio_norefs
for migrate_folio:

+ .migrate_folio = buffer_migrate_folio_norefs,


Thanks,
Ryusuke Konishi


>
>  static const struct inode_operations def_mdt_iops;
> --
> 2.43.0
>

  reply	other threads:[~2024-10-03 11:49 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-10-02 15:00 [PATCH 0/4] nilfs2: Finish folio conversion Matthew Wilcox (Oracle)
2024-10-02 15:00 ` [PATCH 1/4] nilfs2: Remove nilfs_writepage Matthew Wilcox (Oracle)
2024-10-03 11:47   ` Ryusuke Konishi
2024-10-02 15:00 ` [PATCH 2/4] nilfs2: Convert nilfs_page_count_clean_buffers() to take a folio Matthew Wilcox (Oracle)
2024-10-03 11:51   ` Ryusuke Konishi
2024-10-02 15:00 ` [PATCH 3/4] nilfs2: Convert nilfs_recovery_copy_block() " Matthew Wilcox (Oracle)
2024-10-03 11:54   ` Ryusuke Konishi
2024-10-02 15:00 ` [PATCH 4/4] nilfs2: Convert metadata aops from writepage to writepages Matthew Wilcox (Oracle)
2024-10-03 11:49   ` Ryusuke Konishi [this message]
2024-10-02 15:40 ` [PATCH 0/4] nilfs2: Finish folio conversion Ryusuke Konishi
2024-10-03 12:20   ` Ryusuke Konishi

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='CAKFNMon0yfSPNcPG0h90DG=c9-anbri1pgZR6Qtc2a8kPG7+LQ@mail.gmail.com' \
    --to=konishi.ryusuke@gmail.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-nilfs@vger.kernel.org \
    --cc=willy@infradead.org \
    /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;
as well as URLs for NNTP newsgroup(s).