RE: [PATCH] f2fs: fix to truncate inline data past EOF

[Chao Yu]

Hi Jaegeuk,

> -----Original Message-----
> From: Jaegeuk Kim [mailto:jaegeuk@xxxxxxxxxx]
> Sent: Friday, March 06, 2015 5:34 AM
> To: Chao Yu
> Cc: Changman Lee; linux-f2fs-devel@xxxxxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH] f2fs: fix to truncate inline data past EOF
> 
> Hi Chao,
> 
> On Thu, Mar 05, 2015 at 05:41:32PM +0800, Chao Yu wrote:
> > Previously if inode is with inline data, we will try to invalid partial inline
> > data in page #0 when we truncate size of inode in truncate_partial_data_page().
> > And then we set page #0 to dirty, after this we can synchronize inode page with
> > page #0 at ->writepage().
> >
> > But sometimes we will fail to operate page #0 in truncate_partial_data_page()
> > due to below reason:
> > a) if offset is zero, we will skip setting page #0 to dirty.
> > b) if page #0 is not uptodate, we will fail to update it as it has no mapping
> > data.
> >
> > So with following operations, we will meet recent data which should be
> > truncated.
> >
> > 1.write inline data to file
> > 2.sync first data page to inode page
> > 3.truncate file size to 0
> > 4.truncate file size to max_inline_size
> > 5.echo 1 > /proc/sys/vm/drop_caches
> > 6.read file --> meet original inline data which is remained in inode page.
> >
> > This patch renames truncate_inline_data() to truncate_inline_inode() for code
> > readability, then intruduces new truncate_inline_data() and use it to truncate
> > inline data in page #0 and inode page in truncate_partial_data_page() for
> > fixing.
> >
> > Signed-off-by: Chao Yu <chao2.yu@xxxxxxxxxxx>
> > ---
> >  fs/f2fs/f2fs.h   |  1 +
> >  fs/f2fs/file.c   | 13 ++++++++-----
> >  fs/f2fs/inline.c | 57 +++++++++++++++++++++++++++++++++++++++++++++++++++-----
> >  3 files changed, 61 insertions(+), 10 deletions(-)
> >
> > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > index 511d6cd..55dbbe7 100644
> > --- a/fs/f2fs/f2fs.h
> > +++ b/fs/f2fs/f2fs.h
> > @@ -1754,6 +1754,7 @@ extern struct kmem_cache *inode_entry_slab;
> >   */
> >  bool f2fs_may_inline(struct inode *);
> >  void read_inline_data(struct page *, struct page *);
> > +int truncate_inline_data(struct inode *, u64 from);
> >  int f2fs_read_inline_data(struct inode *, struct page *);
> >  int f2fs_convert_inline_page(struct dnode_of_data *, struct page *);
> >  int f2fs_convert_inline_inode(struct inode *);
> > diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> > index f1341c7..6c4443e 100644
> > --- a/fs/f2fs/file.c
> > +++ b/fs/f2fs/file.c
> > @@ -461,6 +461,9 @@ static int truncate_partial_data_page(struct inode *inode, u64 from)
> >  	unsigned offset = from & (PAGE_CACHE_SIZE - 1);
> >  	struct page *page;
> >
> > +	if (f2fs_has_inline_data(inode))
> > +		return 0;
> > +
> >  	if (!offset)
> >  		return 0;
> >
> > @@ -497,14 +500,14 @@ int truncate_blocks(struct inode *inode, u64 from, bool lock)
> >  	if (lock)
> >  		f2fs_lock_op(sbi);
> >
> > -	ipage = get_node_page(sbi, inode->i_ino);
> > -	if (IS_ERR(ipage)) {
> > -		err = PTR_ERR(ipage);
> > +	if (f2fs_has_inline_data(inode)) {
> > +		err = truncate_inline_data(inode, from);
> >  		goto out;
> >  	}
> >
> > -	if (f2fs_has_inline_data(inode)) {
> > -		f2fs_put_page(ipage, 1);
> > +	ipage = get_node_page(sbi, inode->i_ino);
> > +	if (IS_ERR(ipage)) {
> > +		err = PTR_ERR(ipage);
> >  		goto out;
> >  	}
> >
> > diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c
> > index 4ba9732..a8189c92 100644
> > --- a/fs/f2fs/inline.c
> > +++ b/fs/f2fs/inline.c
> > @@ -50,10 +50,57 @@ void read_inline_data(struct page *page, struct page *ipage)
> >  	SetPageUptodate(page);
> >  }
> >
> > -static void truncate_inline_data(struct page *ipage)
> > +static void truncate_inline_inode(struct page *ipage, u64 from)
> >  {
> > +	void *addr = inline_data_addr(ipage);
> > +
> >  	f2fs_wait_on_page_writeback(ipage, NODE);
> > -	memset(inline_data_addr(ipage), 0, MAX_INLINE_DATA);
> > +	memset(addr + from, 0, MAX_INLINE_DATA - from);
> > +}
> > +
> > +int truncate_inline_data(struct inode *inode, u64 from)
> > +{
> > +	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> > +	struct page *page, *ipage;
> > +	void *addr;
> > +	int err = 0;
> > +
> > +	/*
> > +	 * we should never truncate inline data past max inline data size,
> > +	 * because we always convert inline inode to normal one before
> > +	 * truncating real data if new size is past max inline data size.
> > +	 */
> > +	f2fs_bug_on(sbi, from > MAX_INLINE_DATA);
> > +
> > +	if (from == MAX_INLINE_DATA)
> > +		return 0;
> > +
> > +	page = grab_cache_page(inode->i_mapping, 0);
> > +	if (!page)
> > +		return -ENOMEM;
> 
> Do we need to grab and read #0 page?

No need to read #0 page, sorry for the mistake.

> How about checking its cached page after on-disk truncation?
> 
> 1. page = get_node_page()
> 2. check_inline
> 3. truncate_inline_inode
> 4. f2fs_put_page(ipage)
> 
> 5. truncate_partial_data_page(inode, from, force)
> 	if (!offset && !force)
>  		return 0;
> 
> > +
> > +	ipage = get_node_page(sbi, inode->i_ino);
> > +	if (IS_ERR(ipage)) {
> > +		err = PTR_ERR(ipage);
> > +		goto out;
> > +	}
> > +
> > +	if (!f2fs_has_inline_data(inode)) {
> > +		f2fs_put_page(ipage, 1);
> > +		goto out;
> > +	}
> > +
> > +	truncate_inline_inode(ipage, from);
> > +	set_page_dirty(ipage);
> > +
> > +	read_inline_data(page, ipage);
> > +	f2fs_put_page(ipage, 1);
> > +
> > +	addr = inline_data_addr(page);
> > +	memset(addr + from, 0, MAX_INLINE_DATA - from);
> 
> It needs to do likewise truncate_partial_data_page.

Do you mean like this?