Re: [RFC PATCH 1/2] maple_tree: add __mtree_insert_range function

[Andrew Ballance]

On Sat, Apr 05, 2025 at 04:22:34PM +0100, Matthew Wilcox wrote:
> On Sat, Apr 05, 2025 at 01:01:53AM -0500, Andrew Ballance wrote:
> > adds the __mtree_insert_range which is identical to mtree_insert_range
> > but does not aquire ma_lock.
> > This function is needed for the rust bindings for maple trees because
> > the locking is handled on the rust side.
> 
> No.
> 
> The support for external locking is a TEMPORARY HACK.  I've talked
> before about why this is and don't feel like explaining it again.

this does use the maple_tree's internal ma_lock. the locking is
done on the rust side using bindings to mtree_lock(). it is done this
way so that rust can track the lifetimes of any values from mtree_load.

the easiest way to do this was to add code to the c side. 
for the v2 I can make changes so that this does not touch the c side.