linux-ext4 - Re: [PATCH 2/5] ext4: move inode extension/truncate code out from ext4_iomap

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: <20190829081831.GB19156@quack2.suse.cz>
Date:   Thu, 29 Aug 2019 10:18:31 +0200
From:   Jan Kara <jack@...e.cz>
To:     Matthew Bobrowski <mbobrowski@...browski.org>
Cc:     Jan Kara <jack@...e.cz>, linux-ext4@...r.kernel.org,
        linux-fsdevel@...r.kernel.org, tytso@....edu, riteshh@...ux.ibm.com
Subject: Re: [PATCH 2/5] ext4: move inode extension/truncate code out from
 ext4_iomap_end()

On Thu 29-08-19 07:54:21, Matthew Bobrowski wrote:
> On Wed, Aug 28, 2019 at 09:59:14PM +0200, Jan Kara wrote:
> > > @@ -257,6 +308,13 @@ ext4_dax_write_iter(struct kiocb *iocb, struct iov_iter *from)
> > >  		goto out;
> > >  
> > >  	ret = dax_iomap_rw(iocb, from, &ext4_iomap_ops);
> > > +
> > > +	if (ret > 0 && iocb->ki_pos > i_size_read(inode)) {
> > > +		err = ext4_handle_inode_extension(inode, iocb->ki_pos,
> > > +						  iov_iter_count(from));
> > > +		if (err)
> > > +			ret = err;
> > > +	}
> 
> I noticed that within ext4_dax_write_iter() we're not accommodating
> for error cases. Subsequently, there's no clean up code that goes with
> that. So, for example, in the case where we've added the inode onto
> the orphan list as a result of an extension and we bump into any error
> i.e. -ENOSPC, we'll be left with inconsistencies. Perhaps it might be
> worthwhile to introduce a helper here
> i.e. ext4_dax_handle_failed_write(), which would be the path taken to
> perform any necessary clean up routines in the case of a failed write?
> I think it'd be better to have this rather than polluting
> ext4_dax_write_iter(). What do you think?

Esentially you'll need the same error-handling code as you have in
ext4_dio_write_end_io(). So probably just factor that out into a common
helper like ext4_handle_failed_inode_extension() and call it from
ext4_dio_write_end_io() and ext4_dax_write_iter().

								Honza
-- 
Jan Kara <jack@...e.com>
SUSE Labs, CR