linux-ext4 - Re: [PATCH 0/6 v2] Introduce FALLOC_FL_ZERO

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:	Sun, 16 Mar 2014 15:08:20 -0400
From:	tytso@....edu
To:	Lukas Czerner <lczerner@...hat.com>
Cc:	linux-ext4@...r.kernel.org
Subject: Re: [PATCH 0/6 v2] Introduce FALLOC_FL_ZERO_RANGE flag for fallocate

On Tue, Feb 25, 2014 at 08:14:33PM +0100, Lukas Czerner wrote:
> Introduce new FALLOC_FL_ZERO_RANGE flag for fallocate. This has the same
> functionality as xfs ioctl XFS_IOC_ZERO_RANGE.

Hi Lukas,

I've been trying to merge these patches into the ext4 tree, and I'm
running into a large number of test failures.  Could you take a quick
eyeball and see if there's anything obvious wrong?

The "dev2" branch in the ext4 git tree has all the patches in this
series (1, 2, and 5) --- 3 was included earlier applied on top of the
"dev" branch.  The "test" branch is the dev2 branch with the
xfs-collapse-range branch pulled in, which actually enables the
ZERO_RANGE flags (as well as the collapse range patches).

I have tested the "dev" branch both standalone and with the
xfs-collapse-range branch pulled in, and things seem to be pretty
stable.  I'm doing more comprehensive testing now to confirm things.
(I'm using xfstests commit id 3948694eb12 which has the latest tests
to exercise the zero_range codepath.)

When I tried testing with the "test" branch, things failed pretty
quickly.  I've attached two of these in this patch set.  I'm guessing
it's some kind of memory corruption problem.  These failures are
pretty repeatable, and it fails fast.

If I try running the "dev2" branch, without the xfstests collapse
range branch pulled in, things are much better (so there's clearly a
bug in the ZERO_RANGE code path), but there was still a few more
errors than the baseline.  I'm rerunning those tests so I can be sure
that the results are repeatable.

I suspect the problem is that something in the dev branch isn't
playing well with your patches, or I screwed up while fixing up some
merge conflicts -- but the merge conflicts were pretty minimal, so
that seems a bit unlikely.

Anyway, if you could take a look, I'd really appreciate it.  Thanks!!

	       	     	    	      	     - Ted

Download attachment "test-fail.tar.gz" of type "application/octet-stream" (34334 bytes)