mnt: Optimize mntinfo_add_list to O(1) using a tail pointer

[ankushT369]

This patch optimizes the mntinfo_add_list function from O(n) to O(1) by introducing a tail pointer (tailbuffer) to track the end of the linked list. Previously, adding a new mount_info node required traversing the entire list to find the tail, which was inefficient for large lists.

Changes:

1. Added a struct list_head tail to the mount_info struct to maintain the tail pointer.
2. Introduced a global tailbuffer to store the current tail of the list.
3. Modified mntinfo_add_list to use tailbuffer for O(1) insertion at the tail.
4. Updated mntinfo_add_list_before to initialize the tail pointer and maintain consistency when adding nodes at the head.

The optimization significantly improves performance for operations involving large lists of mount information. This change also ensures that the list remains consistent and correctly updated during insertions.

[avagin]

I think it is better to user the list_head structure here.

[ankushT369]

You mean instead of mount_info *tail should I make it list_head tail

[avagin]

I mean we can use the standard list here instead of reinventing a bicycle.

[ankushT369]

@avagin What changes should I make to the code overall?

[avagin]

Here is just one logical change, so it should be just one patch. The subject line starts with the subsystem prefix. In this case, it is mnt:. The commit body should describe the change, reasons and its effects in details.

[ankushT369]

ok got it I am gonna fix this!!