linux-ext4 - Re: [PATCH 0/4] ext4: DAX fixes

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 May 2016 16:26:16 +0200
From:	Jan Kara <jack@...e.cz>
To:	Theodore Ts'o <tytso@....edu>
Cc:	Jan Kara <jack@...e.cz>, linux-ext4@...r.kernel.org,
	linux-nvdimm@...1.01.org, Vishal Verma <vishal.l.verma@...el.com>,
	Ross Zwisler <ross.zwisler@...ux.intel.com>,
	Dan Williams <dan.j.williams@...el.com>,
	linux-fsdevel@...r.kernel.org
Subject: Re: [PATCH 0/4] ext4: DAX fixes

On Mon 16-05-16 11:35:25, Jan Kara wrote:
> On Fri 13-05-16 09:56:00, Ted Tso wrote:
> > On Fri, May 13, 2016 at 01:24:19AM -0400, Theodore Ts'o wrote:
> > > Queued and I'm running tests before I push them out.  The patches
> > > passed the smoke tests[1], and the full tests should be running for the
> > > next seven hours or so at:   http://104.197.222.91
> > 
> > So the good news is that the full tests ("gce-xfstests full") look
> > really good[1]) Unfortunately the DAX tests ("gce-xfstests -c dax -g
> > auto") show a regression.  Is this expected?  (i.e., were you
> > expecting test regressions that will be fixed by the DAX patch
> > series?)
> 
> No, I was not expecting to get new xfstests failures from just ext4
> patches. That being said I didn't test ext4 patches on their own, just in
> combination with other DAX changes so there may be some interaction I
> forgot about. I'll see whether I can reproduce the failures and understand
> what's going on. Thanks for having a look!

So I have nailed this down. The patch "ext4: Handle transient ENOSPC
properly for DAX" changes how ext4_dax_mmap_get_block() (later renamed to
ext4_dax_get_block()) behaves for calls with create == 0 and that trips
over an issue in __dax_fault() (a bug that __dax_fault() doesn't call
get_blocks() with create == 1 for write faults into unwritten extents)
which gets fixed only in "dax: Remove dead zeroing code from fault handlers".

I see two options here:

1) Just push patches as is and have ext4 dax broken between ext4 merge and
nvdimm merge.

2) Split out the one-line change from "dax: Remove dead zeroing code from
fault handlers" in __dax_fault() which fixes the behavior for ext4 and
merge it through ext4 tree. Merge the rest through nvdimm tree.

Dan? Ted?

								Honza

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