Paragon 'optimistic' that its NTFS driver will be accepted into the Linux Kernel
- Reference: 1599592221
- News link: https://www.theregister.co.uk/2020/09/08/paragon_ntfs_linux/
- Source link:
The decision to offer the NTFS code for Linux [1]emerged last month when the company submitted a 27,000 line patch to the kernel mailing list, raising complaints from code maintainers that it was too large to review.
The patch is now at [2]version 4 and has been split into chunks as well as being cleaned up and bugs fixed.
“Though the decision about its acceptance is yet to be made, Paragon feels optimistic … it’s important to us that the Linux community can make use of NTFS included in the Linux kernel with confidence,” the company stated.
Linux currently has two NTFS drivers, a FUSE (Filesystem in Userspace) driver which is read/write, and a kernel driver which is read-only. It is this latter driver which Paragon intends to replace.
“The need for a new native implementation included with the kernel comes as the current NTFS driver remains practically unmaintained, lacks decent write support and has none of the other advanced features,” Paragon said. Its driver is not only read/write but supports additional features including journal replay, compressed and sparse files, and more.
Why has the company chosen to offer its code under the GPL (GNU General Public License) now, many years after first offering it commercially? It was because of Microsoft, Paragon told us. “Microsoft made the decision to allow the use of exFAT in the Linux kernel, and we felt encouraged to give Linux what it has been waiting for over 20 years – quality NTFS supports at the kernel level.”
The driver is one thing, but what about the tools and utilities included in its commercial version, like chkntfs for checking and fixing errors on NTFS partitions?
“We plan to publish the next releases [3]on our GitHub , Paragon said.
We asked whether the company will still offer a commercial NTFS driver for Linux. “We will continue developing it due to its complexity, and offering it also as a proprietary technology because OEMs require us to do so,” the company said. “Note that OEMs (military, government, businesses etc) will never use GPL because rules require making public any modifications - OEMs will always look for proprietary solutions and we remain here to support them.”
Customers may want to amend the code to fine-tune the performance on a specific device, said Paragon. “[A] cell phone or action camera running Linux and accessing NTFS media has different performance, power, footprint and memory requirements compared to a server or desktop running Linux. This kind of fine tuning can make a huge difference.” Customers unwilling to publish the amendments they make have “an option to use Paragon’s proprietary code that comes under commercial license,” we were told.
Paragon also said that it can more easily assure the quality of its proprietary code. “Paragon’s Linux Kernel implementation of NTFS complies with Paragon’s internal test requirements, but we’re unable to provide instant testing of the code once it’s accepted to a Linux Kernel and changed further by the community.”
Since the company is promising to continue maintaining its Linux contribution, and any part of Linux may be "changed further by the community", we're not sure that this last remark makes complete sense; but what is clear is that customers who prefer the commercial licensing model of the existing offer can still have it, even if the company is succesful in getting its driver into the kernel.®
Get our [4]Tech Resources
[1] https://www.theregister.com/2020/08/18/paragon_tries_to_contribute_ntfs/
[2] https://lore.kernel.org/lkml/20200904133231.1769292-1-almaz.alexandrovich@paragon-software.com/t/#u
[3] https://github.com/Paragon-Software-Group
[4] https://whitepapers.theregister.com/
Re: Whatever for?
I have no idea if this new NTFS driver is better than the old ones, but it doesn't seem like "nothing" that someone is offering it gratis for people to make up their own minds. NTFS is a very advanced file system that has gone through many iterations and it is easy to see that a new driver might do things that an existing one doesn't.
Re: Whatever for?
Timeo Danaos et dona ferentes: "I fear Geeks when they come bearing Gits", perhaps? :) (In Sean Connery's voice, ofc)
Re: Whatever for?
Intoned in Jackie Coogan's Uncle Fester voice, Shirley.
Re: Whatever for?
No one is forcing you to use this, and the FUSE driver wouldn't go away. It doesn't cost anything to add a driver built as a module as they pretty much all are in all major distributions, it doesn't make the kernel bigger or use additional memory.
This would allow eventually removing the read only kernel driver currently included, and probably be better supported than the FUSE driver since there's a company with a commercial interest in NTFS standing behind it.
Re: Whatever for?
I really don't think the Linux kernel needs to add another 27,000 lines of code to get ... nothing?
You are free to:
a) not load the ntfs kernel module.
b) compile your kernel leaving out any particular chosen filesystem support. Compile what you need from the choice available to you.
So, with Linux, you are free to not use the Paragon-originated kernel driver. However, some people could well have good reasons for wanting a better performing driver than a FUSE driver, or one with more options (like encryption and sparse file support), so it is good that they could get a GPL-licenced option.
I hope the quality is good enough for inclusion, and that long-term maintenance can be assured.
Re: Whatever for?
"I hope the quality is good enough for inclusion"
I've read it. It isn't. You are free to check my work.
"and that long-term maintenance can be assured."
Are you volunteering? Because I sure as hell am not.
Re: Whatever for?
The FUSE ntfs-3g driver has worked for years certainly, but only for basic tasks. It doesn't include any file-system check/repair utilities, and it can't be used for boot-time use. It would be nice if we could put a Fedora ISO on an NTFS formatted USB thumb drive and boot-up and install from that, directly. There's also the question of performance of userland versus kernel-space.
Good luck.
That's just the same mega-patch chopped into ten, crudely at obvious file boundaries.
There's still no attempt to actually make it friendly.
When one whole patch is "This adds NTFS journal", without barely a word of explanation in the code, it's just never going to get in.
They're trying to code-dump and it's taken them four versions (after the initial junk they just lobbed at people) to get to this? I can't see it ever getting in in that instance.
As I've said before, this is a maintenance burden that you're trying to pass off to 10-20 years worth of deep-level Linux kernel developers because once it's started to be used, it's their responsibility to keep it working, secure and - most importantly - reliable, within the context of the entire kernel.
You're going to have to do a damn sight better than this.
Yes, but some people just don't get it.
As I've said before ...
Quite so, right [1]here .
Well written and thoroughly explained, @Lee D.
All those who think that taking all this Paragon BS on board Linux is such a great and convenient idea should read it instead of pontificating on how generous an offer it is and how useful NTFS would be for the Linux ecosystem.
Unbelievable ...
O.
[1] https://forums.theregister.com/forum/all/2020/08/18/paragon_tries_to_contribute_ntfs/#c_4092145
Are they really this unclear on the concept?
“Note that OEMs (military, government, businesses etc) will never use GPL because rules require making public any modifications - OEMs will always look for proprietary solutions and we remain here to support them.”
Such a seemingly clueless statement clearly designed to misinform and/or misdirect makes me wonder what the real intent behind this "gift" are.
Re: Are they really this unclear on the concept?
Paragon told us. “Microsoft made the decision to allow the use of exFAT in the Linux kernel, and we felt encouraged to give Linux what it has been waiting for over 20 years – quality NTFS supports at the kernel level.”
Pretty sure that line sums it up, exFAT is stealing our customers and we need better integration to compete.
This is interesting.
I wonder if the code will stand up to kernel maintainers scrutiny, both in terms of code quality and legality (let's not have another SCO incident please).
Are Paragon the people who write the Mac hfs+ driver for windows? That has worked well and allowed me to salvage data from failing OS X drives before.
Whatever for?
Linux currently has two NTFS drivers, a FUSE (Filesystem in Userspace) driver which is read/write, and a kernel driver which is read-only.
In my opinion (as a Linux user, not coder, programmer or maintainer) I really don't think the Linux kernel needs to add another 27,000 lines of code to get ... nothing?
I've managed quite well with the FUSE driver when I have needed to look at/write to the odd HDD with a NTFS filesystem and really cannot remember the last time I did.
As far as I'm concerned, it is too little, too late, with a bad attitude and worthless.
Nothing more than a recipe for trouble further on.
--------------------------------------> Timeo Danaos et dona ferentes <--------------------------------------
O.