Why not implement the inode interface provided by kernel fuse to replace the current path interface？

Question

Why not implement the inode interface provided by kernel fuse to replace the current path interface？

T-TRz879 opened this issue 2 years ago · comments

I've seen the specification of kernel fuse before. For example, rename provides methods such as rename, renameat, and renameat2. I think that using inode of the parent directory for Rename is very effective and useful

T-TRz879 commented 2 years ago

@SerCeMan

T-TRz879 commented 2 years ago

@SerCeMan

T-TRz879 · Answer 1 · Fri Jul 15 2022 15:36:04 GMT+0800 (China Standard Time)

@jenschurchill @melix @overheadhunter @czeidler

Sergey Tselovalnikov · Answer 2 · Sat Jul 16 2022 15:54:56 GMT+0800 (China Standard Time)

Hey, @T-TRz879! Could you clarify the question? I'm not sure I understand it.

T-TRz879 · Answer 3 · Wed Jul 20 2022 16:12:14 GMT+0800 (China Standard Time)

When I use go-fuse to implement my own file system. For example, in the rename method, the parameters of go-fuse contain parentPathNodeId. When I use our fuse rename, it's just srcPath and dstPath. If the directory is very deep, it's time-consuming. So can we implement the interface of the higher version of Lib fuse?
I don't know if my understanding is correct？Thank's you reply to me :)