Linux's exFAT Driver Looking To Still Be Replaced By A Newer Driver From Samsung

[billbo]

I do wonder though... FAT32 is also patent encumbered and it's been enforced though the case was settled - see Microsoft v. TomTom. There's a strong argument that, thanks to UEFI requiring a small FAT32 formatted partition on all modern PCs, people now have a right to use alternative implementations not sanctioned by Microsoft. This has never been tested in court, and I doubt Microsoft is willing to fight that Pyrrhic battle under Satya Nadella's leadership.

Patents in the US at least have max. terms of 20 years. As FAT32 was first released in Windows 95 OSR2, that ship may have already sailed. The TomTom case was back in 2009. I suppose if one dug deep enough one could find which patents Microsoft cited in that case to verify that they have expired.

[CommunityMember]

Too bad they just went silent. I would have liked to get some insight into the quality of their drivers (and had some high hopes for that).

My conjecture (and it is just a guess, I have no knowledge) is the company decided to invest their limited programmer resources in their remaining products rather than invest in the effort to split out the exfat code (as I understand it, the exfat code was part of their entire file system compatibility suite they sell for many platforms), especially as reducing the sales/licenses when exfat made it to stable was going to impact their revenue going forward.

But I guess Samsung's driver should be quite battle proven to be a good addition. We'll see :-)

The Samsung code has been part of Samsung's shipping products for years, so it likely is reasonably good. Even Samsung has had to split exfat out of their combined vfat/exfat file system driver, but Samsung sees the value (and has the deep resources) to do that and then eliminate their independent work going forward (it is almost certainly a net win for Samsung not having to maintain an out-of-tree file system driver).

[caligula]

Patents in the US at least have max. terms of 20 years. As FAT32 was first released in Windows 95 OSR2, that ship may have already sailed. The TomTom case was back in 2009. I suppose if one dug deep enough one could find which patents Microsoft cited in that case to verify that they have expired.

I think they've made updates to FAT32 after that. XBOX has a special version of FAT32. Some later versions increased the maximum block size. So, some FAT32 may still be active.

[oiaohm]

I think they've made updates to FAT32 after that. XBOX has a special version of FAT32. Some later versions increased the maximum block size. So, some FAT32 may still be active.

Not quite. Linux kernel vfat driver for fat32 included all the block sizes from day one. Windows would read/write fat32 with the largest block size from get go as well just the format tools provided with Windows would refuse to make it and the diagnostic tools would also die.. XBOX first version used ntfs. Later versions uses xfat variation that is fairly close fat32 but not quite.

[Nille_kungen]

Does the Samsung code include any exfatfsck?
A well supported filesystem needs fsck to find and fix errors.

[Nille_kungen]

Patents in the US at least have max. terms of 20 years. As FAT32 was first released in Windows 95 OSR2, that ship may have already sailed. The TomTom case was back in 2009. I suppose if one dug deep enough one could find which patents Microsoft cited in that case to verify that they have expired.

You'll find all the related patents in this article.

https://lwn.net/Articles/320737/

According to the article there was 8 patents used and 5 of those have expired the remaining 3 unexpired patents isn't related to FAT.
The short version is that all the FAT patents used against TomTom has expired.

[rogerx]

Relavent with Patent incumbered code and Microsoft's ownership of GitHub, why are the Linux kernel sources using GitHub rather than GitLab?

[oiaohm]

Relavent with Patent incumbered code and Microsoft's ownership of GitHub, why are the Linux kernel sources using GitHub rather than GitLab?

That is not 100 percent true.

https://gitlab.com/linux-kernel/linux

Linux kernel source are minored to github and gitlab but the masters are on kernel.org. There is talk about kernel.org running its own gitlab instances in time as well. Just doing that change is not a simple or painless process.

[SciK]

On top of that, what difference would it actually make?

[caligula]

On top of that, what difference would it actually make?

The pull requests need to be transferred.