linux-kernel - Re: [PATCH 1/2] f2fs: reorganize __f2fs_add

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150724155200.GA30823@jaegeuk-mac02.hsd1.ca.comcast.net>
Date:	Fri, 24 Jul 2015 08:52:00 -0700
From:	Jaegeuk Kim <jaegeuk@...nel.org>
To:	Chao Yu <chao2.yu@...sung.com>
Cc:	linux-f2fs-devel@...ts.sourceforge.net,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] f2fs: reorganize __f2fs_add_link

Hi Chao,

On Fri, Jul 24, 2015 at 06:31:16PM +0800, Chao Yu wrote:
> Hi Jaegeuk,
> 
> > -----Original Message-----
> > From: Jaegeuk Kim [mailto:jaegeuk@...nel.org]
> > Sent: Friday, July 24, 2015 2:15 AM
> > To: Chao Yu
> > Cc: linux-f2fs-devel@...ts.sourceforge.net; linux-kernel@...r.kernel.org
> > Subject: Re: [PATCH 1/2] f2fs: reorganize __f2fs_add_link
> > 
> > Hi Chao,
> > 
> > On Wed, Jul 22, 2015 at 06:21:52PM +0800, Chao Yu wrote:
> > > Firstly, spliting init_inode_metadata into two parts as below since they
> > > are relatively independent.
> > > 1) init_inode_metadata: init datas belong to newly created inode;
> > > 2) update_inode_metadata: update inode info for linking inode to new
> > > dentry.
> > >
> > > Secondly, move init_inode_metadata to the front of __f2fs_add_link,
> > > So the flow of __f2fs_add_link will be reorganized as:
> > > 1) init inode meta data for new creatation;
> > > 2) lookup dentry page and insert new directory entry in the page;
> > > 3) update meta data of inode and parent.
> > >
> > > Signed-off-by: Chao Yu <chao2.yu@...sung.com>
> > > ---
> > >  fs/f2fs/dir.c    | 129 ++++++++++++++++++++++++++++++++-----------------------
> > >  fs/f2fs/f2fs.h   |   4 +-
> > >  fs/f2fs/inline.c |   2 +-
> > >  3 files changed, 78 insertions(+), 57 deletions(-)
> > >
> > > diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c
> > > index a34ebd8..0ad9a1c 100644
> > > --- a/fs/f2fs/dir.c
> > > +++ b/fs/f2fs/dir.c
> > > @@ -381,43 +381,68 @@ static int make_empty_dir(struct inode *inode,
> > >  	return 0;
> > >  }
> > >
> > > -struct page *init_inode_metadata(struct inode *inode, struct inode *dir,
> > > +int init_inode_metadata(struct inode *inode, struct inode *dir,
> > >  			const struct qstr *name, struct page *dpage)
> > >  {
> > >  	struct page *page;
> > >  	int err;
> > >
> > > -	if (is_inode_flag_set(F2FS_I(inode), FI_NEW_INODE)) {
> > > -		page = new_inode_page(inode);
> > > -		if (IS_ERR(page))
> > > -			return page;
> > > +	if (!is_inode_flag_set(F2FS_I(inode), FI_NEW_INODE))
> > > +		return 0;
> > >
> > > -		if (S_ISDIR(inode->i_mode)) {
> > > -			err = make_empty_dir(inode, dir, page);
> > > -			if (err)
> > > -				goto error;
> > > -		}
> > > +	page = new_inode_page(inode);
> > > +	if (IS_ERR(page))
> > > +		return PTR_ERR(page);
> > >
> > > -		err = f2fs_init_acl(inode, dir, page, dpage);
> > > +	if (S_ISDIR(inode->i_mode)) {
> > > +		err = make_empty_dir(inode, dir, page);
> > >  		if (err)
> > > -			goto put_error;
> > > +			goto error;
> > > +	}
> > >
> > > -		err = f2fs_init_security(inode, dir, name, page);
> > > +	err = f2fs_init_acl(inode, dir, page, dpage);
> > > +	if (err)
> > > +		goto put_error;
> > > +
> > > +	err = f2fs_init_security(inode, dir, name, page);
> > > +	if (err)
> > > +		goto put_error;
> > > +
> > > +	if (f2fs_encrypted_inode(dir) && f2fs_may_encrypt(inode)) {
> > > +		err = f2fs_inherit_context(dir, inode, page);
> > >  		if (err)
> > >  			goto put_error;
> > > +	}
> > >
> > > -		if (f2fs_encrypted_inode(dir) && f2fs_may_encrypt(inode)) {
> > > -			err = f2fs_inherit_context(dir, inode, page);
> > > -			if (err)
> > > -				goto put_error;
> > > -		}
> > > -	} else {
> > > -		page = get_node_page(F2FS_I_SB(dir), inode->i_ino);
> > > -		if (IS_ERR(page))
> > > -			return page;
> > > +	if (f2fs_encrypted_inode(dir))
> > > +		file_set_enc_name(inode);
> > > +
> > > +	update_inode(inode, page);
> > > +	f2fs_put_page(page, 1);
> > 
> > We should not put the inode page here.
> > Since checkpoint can be called and flush all the node pages with valid
> > NAT entres. Once power-off happens after then, we can see unreachable NAT
> > entries.
> 
> IMO, all callers of init_inode_metadata such as __f2fs_add_link/f2fs_do_tmpfile,
> they are all protected well by cp_rwsem, so checkpoint & SPO will not easily
> bother us.
> 
> 1) Lately in __f2fs_add_link, ino of ipage will be inserted into directory entry
> of parent's inode, this makes our nat entry of ipage reachable.
> 2) And also in __f2fs_tmpfile, ino will be added in orphan list, so all data
> resource of the inode will be evicted after CP->SPO->RFR procedure.
> 
> Is there anything I missed?

Let me think about this for a while.
I think two patches are clean-ups with a little bit big changes.
Currently, we've touched many parts including extent_cache, so I need to focus
on stabilizing them first.
After then, I'd like to dig two clean-up patches. Ok?

Thanks,

> 
> Thanks,
> 
> > 
> > Thanks,
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/