Some ideas from an Algorithms perspective:

For an always sorted list, you could make use of data structures like AVL or Red-Black Trees to maintain sorted order of FileList upon creating, deleting or renaming a file as per your choice.

To improve performance with large directories:

Use a hash table for filename lookups, inspired by the linux kernel's dcache

Map filenames (or file IDs) to FileNode pointers in the Red Black Tree for O(1) lookups.

To handle rename cases, I would:

1. Retrieve the FileNode pointer using Hash table.
2. Free the occupied memory by FileNode->name. Delete the node from Red-black tree and rebalance.
3. Reallocate the necessary memory required.
4. Copy the newName to FileNode->name. Insert the node in Red-black tree and rebalance again.

Further scope of improvement:

• There's a concept called relaxed balancing for RB trees/AVL trees. You can explore the tradeoffs in this article.
• For higher degree of concurrent rebalancing: Chromatic binary search trees and HyperRed-Black trees