[oi-dev] ntfs-3g and fuse

[ken mays]

Jean-Pierre,

What is your current analysis (if any) with ntfs-3g_ntfsprogs-2012.1.15 ?? Just fuse kernel module issue?


~ Ken Mays


________________________________
 From: Jean-Pierre André <>
To: Milan Jurik <milan.jurik at xylab.cz> 
Cc: OpenIndiana Developer mailing list <oi-dev at openindiana.org> 
Sent: Monday, February 13, 2012 9:06 AM
Subject: Re: [oi-dev] ntfs-3g and fuse
 
Hi Milan,

Milan Jurik wrote:
> Hi Jean-Pierre,
> 
> On 13.02.2012 11:16, Jean-Pierre André wrote:
>> Hi,
>> 
>> Milan Jurik wrote:
>>> Hi Jean-Pierre,
>>> 
>>> Jean-Pierre ANDRE píše v út 24. 01. 2012 v 15:16 +0100:
>>> 
>>>> Hi,
>>>> 
>>>> As a maintainer of ntfs-3g, I have received bug reports on OpenIndiana. Digging into them, I found there were almost all caused by the buggy fuse kernel module, which nobody seems to care about. So I had to do it myself....
>>>> 
>> 
>> I think I have now finished the first level debugging of
>> the fuse kernel module....
>> 
>> With the fixes, ntfs-3g passes the Posix file system test
>> http://www.tuxera.com/community/posix-test-suite/
>> except for a single test :
>> 
>> Failed Test                               Stat Wstat Total Fail  List
>> of Failed
>> 
>> ------------------------------------------------------------------------------- 
>> /home/linux/fstest/pjd-fstest/tests/open/                3    1  2
>> Failed 1/191 test scripts. 1/1964 subtests failed.
>> Files=191, Tests=1964, 351 wallclock secs (19.50 cusr + 99.55 csys =
>> 119.05 CPU)
>> Failed 1/191 test programs. 1/1964 subtests failed.
>> 
>> The test which fails is the test for opening a fifo
>> named by ntfs-3g. Obviously this opening should be
>> bounced elsewhere, but I have no idea where....
>> 
>> There are a few other obscure things, which I cannot
>> clarify without the help of a Solaris VFS expert.
>> Reviewing of the code by a Solaris developer would
>> also be useful : there were a lot of memory leaks,
>> missing lockings, and forgotten unlockings. There
>> were obvious deficiencies in the quality management
>> of the initial version.
>> 
> 
> no surprise, the original developers were not filesystem experts.
> 
>> In the meantime, I consider the result as a release
>> candidate.
>> 
>> Please see http://b.andre.pagesperso-orange.fr/openindiana-ntfs-3g.html
>> 
> 
> Wow.
> 
>> By the way, where is the official OpenIndiana source
>> repository for this ?
> 
> Nowhere. Currently it is build from spec-files-extra (http://pkgbuild.sourceforge.net/spec-files-extra/) which is using the last OpenSolaris.org FUSE tarball (it contains head of the FUSE repo) + your patches, then it is pulled by SFE OI. What I did it was that I took the tarballs (opensolaris.org and your) and created diff and pushed the diff to spec-files-extra.
> 
>> I saw Milan posting the patches,
>> but is not clear to me against what the patches were.
>> I can provide a list of patches per fix for a source
>> repository management.
>> 
> 
> I think the best way would be to clone opensolaris.org repo to some public repository site and commit your patches there.

Sure, but I am not the maintainer of fuse (and I do
not want to), nor the packager for OpenIndiana.
My job is only to answer problem reports on ntfs-3g.

> It makes no sense to continue with opensolaris.org repository as the latest updates to that site brought more closed source management features than community things.

I agree.

> I can create one later this week (I will be mostly off till weekend) if nobody will stand up.

That is a nice proposal. My patches are at your disposal.

Regards

Jean-Pierre



_______________________________________________
oi-dev mailing list
oi-dev at openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://openindiana.org/pipermail/oi-dev/attachments/20120213/c7104342/attachment-0001.html>