F2FS In Linux 6.7 Supports Larger Page Size, Continued Zone Block Device Work

[phoronix]

Phoronix: F2FS In Linux 6.7 Supports Larger Page Size, Continued Zone Block Device Work

The Flash-Friendly File-System (F2FS) continues to be improved upon in the mainline Linux kernel and with the ongoing v6.7 merge window has received some additional enhancements...

https://www.phoronix.com/news/Linux-6.7-F2FS

[darkbasic]

Does it support 64K?

[zboszor]

Previously formatted and populated F2FS filesystems (which use 4K blocks) cannot be used anymore if the system happens to use a different page size. Nice way to make systems unbootable.

Also, flash drives formatted with F2FS won't be portable across different architectures.

[stormcrow]

Previously formatted and populated F2FS filesystems (which use 4K blocks) cannot be used anymore if the system happens to use a different page size. Nice way to make systems unbootable.

Also, flash drives formatted with F2FS won't be portable across different architectures.

So don't change the page size post filesystem creation. That's not rocket science. The article doesn't say you can't mount and read these devices on a platform with different page sizes. It says that if youdo change page sizes for the filesystem the block size must be equal to the memory page size as currently implemented. The only difference is you'll lose performance if your filesystem doesn't reflect the memory page size of the host CPU, it won't be unreadable.

[zboszor]

So don't change the page size post filesystem creation. That's not rocket science.

What happens on systems that already use >4K page size?

F2F2 filesystems formatted using f2f2-tools up to 1.16.0 only supported 4K block size regardless of the page size on the system.

Now both f2fs-tools git and the kernel wants the F2FS block size match the page size and don't support anything else. Relaxing it comes later. This is exactly what the change says.