The exFAT filesystem is coming to Linux | Hacker News Reader

The exFAT filesystem is coming to Linux

arstechnica.com

110 pointsnotRobot30 comments

LWN has some good coverage of the licensing and technical aspects: https://lwn.net/Articles/797963/

I tried to put the license and patent info into the Wikipedia page on exFAT a few years ago and it was reverted a few minutes later as “not germane”. Hmm...I decided not to try again.

How does this square with Microsoft removing or deprecating exFAT support from Windows [0]? Isn't the whole point to have a file system with cross-platform support? I mean, is Microsoft being nefarious by allowing exFAT to come to Linux while at the same time beginning to drop support in Windows? Or am I reading too much into this?

[0]: I came across the comment below yesterday and was able to verify that any mention of exFAT has been removed from the official "diskpart" documentation.

> Well it appears that Windows 10 support for exFAT has been dropped in a recent update. The option no longer appears in the drive format dialog, nor does it appear in the diskpart utility. Any mention of exFAT has been expunged from the official diskpart documentation making me believe that Microsoft has officially terminated support for it, even so I cannot find any official announcements or confirmations of that. source - https://superuser.com/a/1420186

A comment on that response explains it; some partitions marked for NTFS filesystems will not allow it to be formatted as exFAT, it's still present on the latest version AFAICT.

By far the worst part of using Linux, Windows, and Mac is trying to find a compatible filesystem for them. Every time I have a file on my Mac that I need to transfer to Windows or from Windows to Linux, trying to write to a filesystem that is supported by all three platforms is just an exercise in frustration.

Sadly (or not sadly depending on your perspective), most people today will use a remote host via Google Drive or DropBox (or even email) and not have this problem.

It's absolutely insane that this is the most painless workaround in 2020

I agree it’s a niche problem. It also sucks that I’ve had this problem for almost 20 years now and it’s still a problem.

Not helpful though if you want to move large files quickly.

Ever tried a spare partition formatted to an older UDF? Like 2.0x? Or if that is too wonky a spare partition formatted to some FAT, and "burn" an ISO into that, then mounting that ISO later on from whichever target you are on. Mounting ISOs without burning them is universally supported. Only remaining annoyance then is the wasted drive capacity, but that depends on your needs.

I used UDF on USB media for years, and only past tense because I don't have much need of passing files between systems on physical media anymore. Iirc, there were some sharp edges, but it was an overall refreshing experience compared to dealing with FAT or dealing with any other experience. I would recommend UDF based on my own experience.

The article quotes Paragon's blog post without linking to it. Here's the original:

https://www.embedded-computing.com/guest-blogs/three-cases-w...

Fourth time this has popped up. Here is the previous discussion https://news.ycombinator.com/item?id=22688058

I think HN rules allow this since it wasn't really discussed before.

It's no coincidence that his is being pushed so aggressively. This change is about to break tons of userland & competition, just not at Microsoft.

Break userland? Sorry.. explanation?

I think he said "break" where it would have been more accurate to say "obsolete". I can see either one inspiring similar emotions from anyone too invested in the current hard ways of dealing with exFAT.

Differences in implementations already cause problems. Adding one will not mitigate this problem, only worsen it. A bit like this https://xkcd.com/927/ Furthermore, anything contemporary should steer away from this heavily outdated technology unless they desire to accumulate even more technical debt.

I hate DOS as much as the next UNIX person but in filesystems it pays to be nuanced. We need to be able to swap data across OS and read/write enabled FS support for >4GB partitions with sensible names not 8.3 and with ACLS and even being able to boot from them would be nice. I wouldn't choose to live in exVFAT but I would welcome integration of support.

All those arguments are valid and where already somewhere around 2k. Today, many more matter. Prominently the FS being aware of it being at flash storage or magnetic etc. We surely should oppose artificially prolonging the *fat shelf life.

Edit: @Guy claiming exFAT was developed for flash drives. Not sure if your clever spindoktering is astroturfing or trolling. >_<

That’s nice. Seeing as exFAT was developed for flash drives, it seems like you should be happy with that :)

> Prominently the FS being aware of it being at flash storage or magnetic etc. We surely should oppose artificially prolonging the fat shelf life.
All of those are good things in isolation. However, the reality is, Linux/BSD/etc. doesn't have the marketshare to push any filesystem cross platform.

Any filesystem with the characteristics you want that goes cross platform will start on MacOS (unlikely) or Windows (much, much more likely).

Until then, people need a way to accomplish tasks and exFat is the best we've got today for cross platform flash storage.

* I suppose it's possible it could come from Android or iOS, although iOS is pretty unlikely since it's so hostile to the idea of filesystems. For this to happen, though, it would have to be a regular task to plug flash drives into phones. That is currently not the case and won't be the case in the near future.

You're not going to be able to boot from exFAT partitions: the UEFI boot partition spec is strictly based on plain old FAT format, not the new exFAT.

Its good to have dreams. It took a few years but PS/2 dropped from the supply chain mobo. We might move on from multi-voltage power supplies in the next decade.

I think people don't use 3.5" floppy drives any more. Or parallel ports.

So yes: the UEFI boot spec is FAT. Once, we didn't have UEFI and maybe once, in the future, boot spec won't be locked in FAT.

What was wrong with PS/2? It worked. Even with less latency than USB. And you could feel the right orientation of the plug, whereas USB was more difficult to get orientated right. And it was usually color coded.

I still think using a PS/2 mouse was an advantage in games like Quake and Counter Strike back in the day.

Perhaps so that when WSL2 is available, they'll be able to mount all Windows filesystems easier?

I believe, Its just part of Microsoft's plan for world domination and to neutralize Linux and other competitors. They are fully in the embrace phase at present. I'd question if it will have ALL the features on Linux that it does on Windows...

The first paragraph of the article states the code is "Microsoft-approved, largely Samsung-authored". The code is open source and missing features, if any, could be added. Microsoft has explicitly endorsed its usage in the kernel [0] and its architecture is fully documented [1]. What more do you want?

[0] https://cloudblogs.microsoft.com/opensource/2019/08/28/exfat...

[1] https://docs.microsoft.com/en-us/windows/win32/fileio/exfat-...

There is a line between caution and paranoia, and you are on the wrong side of it.

What's the reasoning behind this? How does exFat support in Linux give Microsoft leverage? I don't see how this isn;t just rehashing of the same old chestnut.

Reply on news.ycombinator.com