linux-ext4 - Re: [PATCH 04/12 v2] xfs: pass LLONG_MAX to truncate_inode_pages

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]

Date:	Mon, 16 Jul 2012 09:13:44 +0200 (CEST)
From:	Lukáš Czerner <lczerner@...hat.com>
To:	Dave Chinner <dchinner@...hat.com>
cc:	Lukas Czerner <lczerner@...hat.com>, linux-ext4@...r.kernel.org,
	linux-fsdevel@...r.kernel.org, tytso@....edu,
	achender@...ux.vnet.ibm.com
Subject: Re: [PATCH 04/12 v2] xfs: pass LLONG_MAX to
 truncate_inode_pages_range

On Mon, 16 Jul 2012, Dave Chinner wrote:

> Date: Mon, 16 Jul 2012 09:11:17 +1000
> From: Dave Chinner <dchinner@...hat.com>
> To: Lukas Czerner <lczerner@...hat.com>
> Cc: linux-ext4@...r.kernel.org, linux-fsdevel@...r.kernel.org, tytso@....edu,
>     achender@...ux.vnet.ibm.com
> Subject: Re: [PATCH 04/12 v2] xfs: pass LLONG_MAX to
>     truncate_inode_pages_range
> 
> On Fri, Jul 13, 2012 at 03:19:07PM +0200, Lukas Czerner wrote:
> > Currently we're passing -1 to truncate_inode_pages_range() which is
> > actually really confusing since the argument is signed so we do not get
> > "huge" number as one would expect, but rather just -1. To make things
> > clearer and easier for truncate_inode_pages_range() just pass LLONG_MAX
> > since it is actually what was intended anyway.
> > 
> > It also makes thing easier for allowing truncate_inode_pages_range() to
> > handle non page aligned regions. Moreover letting the lend argument to
> > be negative might actually hide some bugs.
> > 
> > Signed-off-by: Lukas Czerner <lczerner@...hat.com>
> > Cc: Dave Chinner <dchinner@...hat.com>
> > ---
> >  fs/xfs/xfs_fs_subr.c |    6 ++++--
> >  1 files changed, 4 insertions(+), 2 deletions(-)
> > 
> > diff --git a/fs/xfs/xfs_fs_subr.c b/fs/xfs/xfs_fs_subr.c
> > index 652b875..6e9b052 100644
> > --- a/fs/xfs/xfs_fs_subr.c
> > +++ b/fs/xfs/xfs_fs_subr.c
> > @@ -34,7 +34,8 @@ xfs_tosspages(
> >  {
> >  	/* can't toss partial tail pages, so mask them out */
> >  	last &= ~(PAGE_SIZE - 1);
> > -	truncate_inode_pages_range(VFS_I(ip)->i_mapping, first, last - 1);
> > +	truncate_inode_pages_range(VFS_I(ip)->i_mapping, first,
> > +				   last == -1 ? LLONG_MAX : last);
> 
> The last paramter changed from (last -1) to last. so if we pass in
> last = 16384, we now truncate to 16384 (first byte of page index 5)
> instead of 16383 (last byte of page index 4). That's a change of
> behaviour and a potential off-by one error, right?

Right, this could potentially cause off-by-one errors, but as it is
now I do not think this could happen. The only place where it is
used with a proper range is XFS_IOC_ZERO_RANGE and you're going to
convert the whole range to unwritten anyway. But it was unintended
and I\ll fix it.


> 
> > @@ -53,7 +54,8 @@ xfs_flushinval_pages(
> >  	ret = filemap_write_and_wait_range(mapping, first,
> >  				last == -1 ? LLONG_MAX : last);
> >  	if (!ret)
> > -		truncate_inode_pages_range(mapping, first, last);
> > +		truncate_inode_pages_range(mapping, first,
> > +					   last == -1 ? LLONG_MAX : last);
> 
> Given this is also done immediately above in the function, perhaps
> this should be done before anything:
> 
> 	if (last == -1)
> 		last = LLONG_MAX;
> 
> and the parameter simply passed to the two functions without the
> conditional logic?

Yes, it makes sense to do this, I'll change it in the next
iteration.

Thanks for the review Dave.
-Lukas

> 
> Cheers,
> 
> Dave.
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html