[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200325155806.GC65658@google.com>
Date: Wed, 25 Mar 2020 08:58:06 -0700
From: Jaegeuk Kim <jaegeuk@...nel.org>
To: Chao Yu <yuchao0@...wei.com>
Cc: linux-f2fs-devel@...ts.sourceforge.net,
linux-kernel@...r.kernel.org, chao@...nel.org
Subject: Re: [PATCH RFC] f2fs: don't inline compressed inode
On 03/25, Chao Yu wrote:
> f2fs_may_inline_data() only check compressed flag on current inode
> rather than on parent inode, however at this time compressed flag
> has not been inherited yet.
When write() or other allocation happens, it'll be reset.
>
> Signed-off-by: Chao Yu <yuchao0@...wei.com>
> ---
>
> Jaegeuk,
>
> I'm not sure about this, whether inline_data flag can be compatible with
> compress flag, thoughts?
>
> fs/f2fs/namei.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c
> index f54119da2217..3807d1b4c4bc 100644
> --- a/fs/f2fs/namei.c
> +++ b/fs/f2fs/namei.c
> @@ -86,7 +86,8 @@ static struct inode *f2fs_new_inode(struct inode *dir, umode_t mode)
> if (test_opt(sbi, INLINE_XATTR))
> set_inode_flag(inode, FI_INLINE_XATTR);
>
> - if (test_opt(sbi, INLINE_DATA) && f2fs_may_inline_data(inode))
> + if (test_opt(sbi, INLINE_DATA) && f2fs_may_inline_data(inode) &&
> + !f2fs_compressed_file(dir))
> set_inode_flag(inode, FI_INLINE_DATA);
> if (f2fs_may_inline_dentry(inode))
> set_inode_flag(inode, FI_INLINE_DENTRY);
> --
> 2.18.0.rc1
Powered by blists - more mailing lists