Re: NVFS XFS metadata (was: [PATCH] pmem: export the symbols __copy_user_flushcache and __copy_from_user_flushcache)

[Mikulas Patocka]

On Tue, 22 Sep 2020, Matthew Wilcox wrote:

> On Mon, Sep 21, 2020 at 12:20:42PM -0400, Mikulas Patocka wrote:
> > The same for directories - NVFS hashes the file name and uses radix-tree 
> > to locate a directory page where the directory entry is located. XFS 
> > b+trees would result in much more accesses than the radix-tree.
> 
> What?  Radix trees behave _horribly_ badly when indexed by a hash.
> If you have a 64-bit hash and use 8 bits per level of the tree, you have
> to traverse 8 pointers to get to your destination.  You might as well
> use a linked list!

In NVFS, radix trees are cut off - they have only as much internal levels, 
as is needed to disambiguate the directory entries.

Read this document: http://people.redhat.com/~mpatocka/nvfs/INTERNALS
the section "DIRECTORIES".

Perhaps, I should call it differently than "radix-trees", but I don't 
really know what is the official name for this data structure.

Mikulas