[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <e6a2233f-5f32-4e1a-87b3-4ecb935e3997@kernel.org>
Date: Tue, 17 Jun 2025 17:30:43 +0800
From: Chao Yu <chao@...nel.org>
To: wangzijie <wangzijie1@...or.com>, linux-f2fs-devel@...ts.sourceforge.net
Cc: chao@...nel.org, feng.han@...or.com, jaegeuk@...nel.org,
linux-kernel@...r.kernel.org
Subject: Re: [f2fs-dev] [PATCH 1/2] f2fs: avoid non-section-aligned size
pinned file generation
On 6/17/25 15:36, wangzijie wrote:
>> On 6/17/25 11:57, wangzijie wrote:
>>> To prevent non-section-aligned size pinned file generated from truncation,
>>> add check condition in setattr.
>>>
>>> Signed-off-by: wangzijie <wangzijie1@...or.com>
>>> ---
>>> fs/f2fs/file.c | 6 ++++++
>>> 1 file changed, 6 insertions(+)
>>>
>>> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
>>> index 6bd3de64f..72f7d1b4a 100644
>>> --- a/fs/f2fs/file.c
>>> +++ b/fs/f2fs/file.c
>>> @@ -1026,6 +1026,7 @@ int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
>>> {
>>> struct inode *inode = d_inode(dentry);
>>> struct f2fs_inode_info *fi = F2FS_I(inode);
>>> + struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
>>> int err;
>>>
>>> if (unlikely(f2fs_cp_error(F2FS_I_SB(inode))))
>>> @@ -1047,6 +1048,11 @@ int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
>>> !IS_ALIGNED(attr->ia_size,
>>> F2FS_BLK_TO_BYTES(fi->i_cluster_size)))
>>> return -EINVAL;
>>> + if (f2fs_is_pinned_file(inode) &&
>>> + attr->ia_size < i_size_read(inode) &&
>>
>> Do we need to consider attr->ia_size > i_size case?
>>
>> Thanks,
>
> Hi, Chao
> After commit 3fdd89b452c2("f2fs: prevent writing without fallocate() for pinned
> files"), when we want to write data to pinned file, we need to use pin+fallocate,
> and we did CAP_BLKS_PER_SEC roundup align when fallocate pinned file:
>
> block_t sec_len = roundup(map.m_len, sec_blks);
>
> Even if we truncate the file to a larger size(maybe larger than sec_len, section
> align or not), and write data to offset beyond sec_len, the write will fail(commit
> 3fdd89b452c2 prevent it). The scattered pin block cannot be generated by this way,
> so I did not consider attr->ia_size > i_size case.
> Do you have some suggestions?
Ah, correct, so what about adding comments here to describe why we don't
need consider attr->ia_size > i_size case?
>
>>> + !IS_ALIGNED(attr->ia_size,
>>> + F2FS_BLK_TO_BYTES(CAP_BLKS_PER_SEC(sbi))))
>>> + return -EINVAL;
>>> }
>>>
>>> err = setattr_prepare(idmap, dentry, attr);
>
>
>
Powered by blists - more mailing lists