Linux Kernel May Gain VMUFAT File-System Support

[phoronix]

Phoronix: Linux Kernel May Gain VMUFAT File-System Support

A patch has been proposed for the Linux kernel to introduce the VMUFAT file-system...

http://www.phoronix.com/vr.php?view=MTA3NDY

[devius]

I have no idea what it can be used for or if it'll ever get used, but using the word "linux" and a 15 years old console in the same sentence can only be...

A-W-E-S-O-M-E

[RealNC]

I played Virtua Tennis day and night on that thing.

[DaemonFC]

Driver code is a different case entirely than code for other areas of the kernel. It literally affects nothing for people who aren't using that hardware.

Generally if a driver is very small, relatively easy to maintain, isn't legally hazardous (and I'm not sure why something based on FAT16, which is ancient, would be) and is up to the kernel coding style and isn't just plain broken, it should be allowed in even if it's only useful to a small number of people.

It's the reason that Linux works with many thousands of pieces of hardware (even obscure ones) and has support for hundreds of different network protocols.

[chefkoch]

Wouldn't this kind of thing make more sense as a fuse filesystem?

[RealNC]

Wouldn't this kind of thing make more sense as a fuse filesystem?

No. As Torvalds said, FUSE is only for toys

[DaemonFC]

No. As Torvalds said, FUSE is only for toys

Anything that wants to give good performance and minimize CPU overhead should be in the kernel. I don't think anyone is going to argue that 128K FAT16-based storage devices from 1999 need the kind of performance that only a kernel module can provide.

On the other hand, FUSE can be pretty nasty to deal with.

[RealNC]

I was being sarcastic. Linus said FUSE is for toys, and now we have a toy's filesystem in the kernel instead of FUSE.

Oh, the irony ;-P