Microsoft Publishes exFAT Specification, Encourages Linux Support

[pal666]

You better read the OIN license again. The very first section is plainly obvious....

you better learn that you trying to play lawyer on tv is looking silly, leave it to real lawyers who are advising many of licensees

[duby229]

you better learn that you trying to play lawyer on tv is looking silly, leave it to real lawyers who are advising many of licensees

I'm sorry that I'm not a lawyer I guess? But why should I trust lawyers paid by MS?

[pal666]

I'm sorry that I'm not a lawyer I guess? But why should I trust lawyers paid by MS?

i never offered you ms lawyer. hire lawyer yourself and stop your maddening fantasies

[duby229]

i never offered you ms lawyer. hire lawyer yourself and stop your maddening fantasies

You're doing just fine at that all by yourself, you don't need any help.... Good day....

[mitcoes]

Not being FREE - only open source -I think it must not go to the kernel space.
Or the FSF must fork the kernel and make a FREE version, and let the Linux Foundation with MS sell themselves to MS.
PS: As ext4 is far better filesystem than exFAT... Why not help to extend ext4 as default filesystem to sd cards making good drivers for MS WOS and Mac OS?

[Bubbles_by_day]

I see. I thought reads/writes to the OS disk were forbidden...

VirtualBox and VMware let you optionally map just certain partitions to a raw VMDK. The raw VMDK still appears to expose all the original partitions, but only the ones you specify are actually accessible.

I routinely map whole drives and just certain partitions, in order to run a real linux installation in a VM in Windows. You can go back and forth between real and VM at will. Linux is smart enough to load real hardware drivers or VM guest additions depending on context. And yes, if you map and access your Windows partition from Linux in a VM hosted on that Windows host, even read-only, Windows will crash and might not come back. (I've had mixed outcomes. Best just to make sure you accurately map only the partitions you need in the VMDK.)

This is hands-down the best way to access data on ZFS and Btrfs from Windows. If you use the Virtio network driver for the Linux VM, access can be nearly as fast as native. Especially if you have enough RAM and processing power. I'm sure benchmarking reveals reduced throughput due to the VM overhead, but I haven't noticed.

You can also do the reverse, in theory, but it's very tricky: Alternately boot a native Windows install both real and in a VM. Parallels does this - or at least did. It set up all the fakery needed to trick Windows into not noticing enough differences between real and VM. (But did require one re-authorization of the license in the process, which is easy enough.) In theory you can do this manually with vbox or vmware, but I've never tried. I just use a standard Windows install in a regular virtual disk file. (Requiring two separate Windows licenses and installations.) With Windows as the VM, you're also limited to the bandwidth of whatever virtual network adapter being emulated.

Edit: Just be aware that the raw VMDK mapping won't be valid across reboots, unless done with VMware, which re-inspects hardware mapping each time. With Virtualbox, you have to recreate the raw VMDK after every host reboot, or reinsertion if using a USB drive. You're probably realizing this would be easier if done with a script, and you'd be correct. (But vbox doesn't make this easy since you can't remove old, inaccurate raw VMDKs from a VM, without removing the whole virtual controller. So the script needs to also create and delete a controller for the raw VMDK[s].)

And yes, you can create a full bootable Linux install on SD card or thumbrive this way, trivially easy - by mapping it to raw VMDK, install Linux in a VM to it, then voila - it's bootable either in a VM, or on metal.

[Bubbles_by_day]

I just remind you of "Those who do not learn history are doomed to repeat it."
Microsoft had stolen DOS and also Windows NT (VMS derivate - much older than Unix - as update by DEC too bad to be usable) - and called kids pirates.
Microsoft urged to only pre-install Windows - those wanting to pre-install Linux won't get Windows licenses or only by a huge price tag.
The only company being more evil than Microsoft is Apple.
Currently I don't see that anything has changed - Microsoft just had to much paying customers who wanted Linux.
Google has aims of their own - it makes money by advertising so selling personal data in some way or let other companies benefit from that knowledge.
And yes - they make money with Linux - as IBM - and both pay little back in comparison of their gains. It's called business.
But at least both are no threat to Linux.
The entire SecureBoot mess is not caused by Google, nor the HW problems: caused by the Windows monopoly.
Microsoft is good at paying people to use Windows - even schools. It's the same with Apple. And politicians are responsible for this.
So I really can not think of any ethics (based on logic by definition - just to spoil the next discussion in advance) being able to say Google is more evil than Microsoft.
And running Linux on Windows is a really bad idea - a buggy proprietary OS full of backdoors running the Linux kernel - just think of it.
Even letting a Windows guest running under Linux should make anyone feel uncomfortable - similar to running a virus under Wine.
But concerning this thread: the kernel hackers are well prepared to just get all info (reverse engineered and directly from Microsoft) and make a driver which has no backdoors and not that many bugs as the original code ...
They would never accept a binary blob (so no danaer's gift - maybe chalk for the wolf ) ... and will improve that code further over time.
So yes, Microsoft may deserve a "thank you" for this small but appreciated move.

You are applying black-or-white/fandboy/dogmatic thinking to a large world of nuance and complexity. Microsoft isn't "Evil". They are an organization of 140,000 people full of drama, power struggles, internal conflict, disagreements, competing goals, terrible decisions, good decisions, failures, successes. They are no more or less "Evil" than any other large collection of human beings working towards common objectives, including surviving and/or thriving.

Many have argued that if a company (any company) was a person, they'd be diagnosed as sociopathic. I find no fault in that idea. Like "The Cube", it's end results nudged by thousands of well-intentioned hands that can wind up causing broader social harm. Sometimes charismatic leaders can shift that internal tension towards more good or more harm, but the more you internalize the notion that almost no one is the villain in their own narrative, the less paranoid you'll tend to be and the happier and more positively productive you'll find yourself. ("You" speaking generally. I can't comment on you personally other than what you just wrote.) I was never a fan of Ballmer. Bill Gates, while as flawed as any human, has literally saved millions of lives single-handedly (at least in terms of decisions, influence, and bringing capital to bear). Is he "Evil"?

Is Linus Torvalds "Evil"? Many people say so. I'd say he's said and done some crappy things. But he's just a poop-flinging monkey like the rest of us, that some choose to worship as a brilliant celebrity, sometimes even god-like, for some weird reason. (Inasmuch as celebrity-worship at all is dumb.) He's a bright guy who did the right thing at the right time. Opportunity met preparation, so to speak. Some would say he fosters a bullying culture and hostile working environment. Maybe so. Maybe not. I haven't worked with/for/near him. Personally, I find some of the news reports involving him repugnant, but take with a heaping grain of salt, and it doesn't directly affect me.

Regardless, never attribute to evil, what can be explained by incompetence or other factors. At least, if you don't want to be a functional adult capable of experiencing occasional serenity.

I worked for microsoft until 2009. No one I worked with, up and down the chain of command, believed they were doing "Evil". Quite the opposite, most people found joy, meaning, and some level of altruism in their work. Few people hated their job there, on balance. Microsoft is exactly like every other large, reasonably successful organization comprised of arguably "high-functioning" primates. (But with better food.)

People need to feel like they belong to a tribe and feel good about their work, and that happens both organically and by design at every big organization, whether a for-profit company, or looser organization of unpaid volunteers that is long-lived and successful in their objectives. And like most companies, Microsoft routinely shoots themselves in the foot with internal politics and misaligned interdepartmental objectives, even highly internally destructive power struggles that net result in compromised market effectivness. The point is, there's no grand internal conspiracy to be evil.

The culture and processes any large organization follows are practically identical to each other at any given point in time. People migrate to and from companies and organizations. Linux kernel developer today, Microsoft Research tomorrow. Then Red Hat the next day. I've worked at several leading companies (and many startups). Some of the hands-down smartest people I've ever met in my many years on this rock, were at Microsoft Research. And they contribute to most Microsoft Products - and all the big ones - it's not just basic research.

Yes, plenty of companies have been exposed to have made top-down decisions seen in hindsight and soundbytes, to be evil. Like "Embrace Extend Extinguish" (which, conspiracy theories aside, isn't a thing any more and never was in such simplistic terms - but not a finest moment by any stretch). Google drives me nuts by canceling every product I've ever come to rely on from them. Apple makes arguably the most secure and privacy-focused phone OS (arguably among any OS), but flaws and terrible privacy decisions are routinely exposed.

Misuse of Facebook and social media in general has directly lead to the slaughter of countless lives in Myanmar, thrown elections, and caused unimaginable havoc and suffering.

Is any of that intentionally evil? Or are we just a bunch of frightened monkeys somehow trying to "work together", all of us banging on keyboards trying to make a living writing software, ignorant of the broader consequences of our efforts to simultaneously 1) support our families, 2) "make a dent in the universe", and 3) do good for the world, starting with our family and emanating outwards in concentric circles of compassion in hierarchical importance to each of us?

I have friends that do and/or have worked for Google, Apple, Oracle, Peoplesoft, Facebook, BlackRock, and other big tech and/or finance companies in the valley and Seattle area. (And a middle-manager in a major investment bank implicated in the '08 crash.) None of them are evil, to my knowledge. And yet every single company has done, or allowed to happen due to incompetence or neglect (and very, very occasionally purposefully organized malice with awareness of knock-on effects), Very Bad Things.

As for running Linux in a VM on Windows or vice-versa. You say it's a bad idea. For whom? How do you know it's a bad idea for anyone but you? If you don't like it, fine, don't do it. But to many, it's a professional necessary, and/or joyful solution. And there's a good reason why so much development effort has been put into consumer-grade desktop virtualization on every major platform.

Which is alone enough evidence for any skeptical inquirer to regard your opinion as "uninformed" at best, to "idiotic" at worst. (I'd be inclined to go with the former for benefit of doubt.)

The only indisputable fact here is that any software, especially an incredibly complex operating system with millions of lines of code, will have exploitable flaws. The method of licensing, distribution, and developer compensation is irrelevant to that point.

Anything else is open for debate.

Even machines of war and space exploration are routinely crippled with software flaws.

Windows is insecure and buggy? By what criterion?

In absolute terms, absolutely there are and always will be exploitable flaws. In Windows, Linux, MacOS, Android, iOS, ChromeOS, et. al.

Or do you mean relative to Linux? ...Not relative to source code line count. (Which, in my mind, is the only fair method of evaluating quality of code. You may have different criterea. C'est la vie.)

And according to many objective studies at various points along OS evolution, Windows is more secure and less buggy, than Linux or MacOS even in absolute terms. It depends on when the study is done, their measurable, objective criteria which will always imperfectly model fuzzier specific real-world use-cases, methodology of data analyses, and weighting. Sometimes Windows wins. Sometimes it loses. Sometimes it loses by a small amount, sometimes big. Maybe metastudies would show it losing more than winning. It's all nuance and shades of gray.

But when user behavior is orders of magnitude the biggest single factor in software security, does it really matter if OS xyz loses some fraction more technical security comparisons than others? (Maybe for some. Not for me.)

Black or white thinking is a security blanket for people too emotionally undeveloped to face the harsh reality that there is no heaven for the pure or hell for the wicked.

The only certain black-or-white argument you could make (and be taken seriously), is that Windows the most risky, security-wise, due to size of installed base and therefore target of opportunity.

But to just call any OS "buggy and virus-prone" because you don't like it and think the company is monolithically/monochromatically "Evil", betrays a fundamental ignorance about large-scale software development, and a deeply flawed, rigid, dogmatic, and emotionally immature thought process. (Which would be fine if you were ten years old. Maybe you are, in which case - congratulations on being on-track for hominid brain development. Unfortunately it's only going to get incrementally better, mostly in the area of fleeting self-awareness of your deeply ingrained and myriad cognitive biases and logical fallacies...if you're lucky.)

[JMB9]

You are applying black-or-white/fandboy/dogmatic ...
But to just call any OS "buggy and virus-prone" because you don't like it and think the company is monolithically/monochromatically "Evil", betrays a fundamental ignorance about large-scale software development, and a deeply flawed, rigid, dogmatic, and emotionally immature thought process. (Which would be fine if you were ten years old.)

Whow - a nice try to make an appearance of political correctness. Lovely!
10 years old - that would be nice - black and white - nice.
Well, I had been a basic researcher and senior Unix consultant working for well known and really big companies.
So I may know what I say about IT industry - especially having seen more internals and signed more NDAs than you may think of, I assume.
For the record - there is no way to make it impossible to get in any computer (other than disconnect it and switch it off). That's right.
I hardened proprietary Unix servers for the banking sector where the level is really high.
And yes, it does make a big difference if you get in minutes - hours - days or months into that system while people are looking at the traffic.
I had never seen any problem with my Linux systems - never lost data - in over 23 years (only 2 dying HDDs - which could be copied).
Ask anyone using Windows saying the same - and everyone in IT industry will smile.
Well, Richard M. Stallman said that stealing may be less evil than creating proprietary SW - and there is a point in that argument.
Just think about it - and then ask yourself - can you think of big companies you would not work for - regardless of the money you may earn?
And now, everyone has his experiences and his point of view.
I am not saying all people working for MS are evil - but MS is a very special company as Apple is ...
But your statements are nice, it was a good laugh.
But just getting more on track - have you followed the kernel mailing list where it was said:
"> The full specification of the filesystem can be found at:
> https://docs.microsoft.com/en-us/win...-specification
This is not truth. This specification is not "full". There are missing
important details, like how is TexFAT implemented. In that specification
is just reference to other unpublished documents. So it is classic MS
way, they release something incomplete and incompatible with their own
NT implementation, like with FAT32 vs fastfat.sys.
I would be very very careful about existence of such documentation until
somebody implement it and do some testing against MS own implementation."
Maybe for you this is a fanboy, too, somebody who sees all things in black and white?
But maybe he made experiences ... and than some ... where it would be stupid to follow the cheer leader squad for MS?
Maybe we all will learn more in the months ahead ... and the kernel hackers know what they are doing.
Wouldn't you say?
"Timeo Danaos et dona ferentes"

[oiaohm]

"> The full specification of the filesystem can be found at:
> https://docs.microsoft.com/en-us/win...-specification
This is not truth. This specification is not "full". There are missing
important details, like how is TexFAT implemented. In that specification
is just reference to other unpublished documents. So it is classic MS
way, they release something incomplete and incompatible with their own
NT implementation, like with FAT32 vs fastfat.sys.
I would be very very careful about existence of such documentation until
somebody implement it and do some testing against MS own implementation."

The horrible reality is what the Microsoft person said could be 100 percent true. As in the exfat specification document that Microsoft just published in the most complete/full document they have other than source.

TexFat has only been implemented on Windows CE 6 and Windows Embedded CE 7.0 not on NT based operating systems. Yes this could be source code compatibility between CE 6 and 7.

"TexFAT specification document" referenced in the exfat specification could be nothing more than the Windows CE source code why I suspect that is all Microsoft specification documents are versioned. Like this exfat specification just published is version 7 or called "Seventh release". "TexFAT specification document" is referenced without a version number followed by word release. Unversioned is normally Microsoft code for it still source and we have not written the documentation yet and what you see is the place holder name for the specification document that has not been written.

The third release of the exfat specification would be a interesting read if you could get mits on it due to extra information about Windows CE permissions having to be removed.

Yes Microsoft developers are not much better than Linux kernel developers on skipping out on writing documentation. In some ways the Microsoft developers are worse by in fact titling and referencing a document they have not in fact written. If the reference includes a version the document exists if the reference does not contain a version they document may or may not exist this applies reading any Microsoft Specifications/MSDN. If it does not exist it was a place holder of the old school use the source documentation method.

Titled but not existing documents has been seen by the wine project many times in the MSDN with a few requests for information being with Microsoft staff responding "hey that documentation should be released I will take it to legal and then coming back I cannot release that document it don't exist it will be released in the next MSDN update after we write it." Microsoft own internal documentation is not 100 percent complete itself.

So we do have to be thankful for some mercy the exfat documentation to this level of detail did in fact exist.

Really the big thing we need to know with TexFat is not that Microsoft has given us complete documentation it if Microsoft is willing to let the patent over TexFat go. There are many papers by people who have reversed TexFat for data recovery so the missing TexFat documentation is out there reversed. The problem is implementing Texfat correctly and having Microsoft use the patent they hold. Part of the reason why I would not care if Linux kernel implemented TexFat incorrectly so not in breach of patent but giving the data protection.

[Bubbles_by_day]

Whow - a nice try to make an appearance of political correctness. Lovely!
10 years old - that would be nice - black and white - nice.
Well, I had been a basic researcher and senior Unix consultant working for well known and really big companies.
So I may know what I say about IT industry - especially having seen more internals and signed more NDAs than you may think of, I assume.
For the record - there is no way to make it impossible to get in any computer (other than disconnect it and switch it off). That's right.
I hardened proprietary Unix servers for the banking sector where the level is really high.
And yes, it does make a big difference if you get in minutes - hours - days or months into that system while people are looking at the traffic.
I had never seen any problem with my Linux systems - never lost data - in over 23 years (only 2 dying HDDs - which could be copied).
Ask anyone using Windows saying the same - and everyone in IT industry will smile.
Well, Richard M. Stallman said that stealing may be less evil than creating proprietary SW - and there is a point in that argument.
Just think about it - and then ask yourself - can you think of big companies you would not work for - regardless of the money you may earn?
And now, everyone has his experiences and his point of view.
I am not saying all people working for MS are evil - but MS is a very special company as Apple is ...
But your statements are nice, it was a good laugh.
But just getting more on track - have you followed the kernel mailing list where it was said:
"> The full specification of the filesystem can be found at:
> https://docs.microsoft.com/en-us/win...-specification
This is not truth. This specification is not "full". There are missing
important details, like how is TexFAT implemented. In that specification
is just reference to other unpublished documents. So it is classic MS
way, they release something incomplete and incompatible with their own
NT implementation, like with FAT32 vs fastfat.sys.
I would be very very careful about existence of such documentation until
somebody implement it and do some testing against MS own implementation."
Maybe for you this is a fanboy, too, somebody who sees all things in black and white?
But maybe he made experiences ... and than some ... where it would be stupid to follow the cheer leader squad for MS?
Maybe we all will learn more in the months ahead ... and the kernel hackers know what they are doing.
Wouldn't you say?
"Timeo Danaos et dona ferentes"

That's an unsatisfactory rebuttal. You just reasserted your same black-or-white thinking, even doubling down, and made more statements of absolute. (Which, pro-tip, renders an argument invalid with just a single reference example to the contrary. If anyone cared. And the thing about political correctness...huh?

"What you just said is one of the most insanely idiotic things I have ever heard. At no point in your rambling, incoherent response, were you even close to anything that could be considered a rational thought. Everyone in this room is now dumber for having listened to it...may God have mercy on your soul."

It's been so long since I've argued with someone on the internet, I admittedly forgot how horribly pathetic it is, for everyone involved. I gotta admit, I feel pretty stupid. My bad for triggering you.