Linus Torvalds [off-list ref] writes:

On Mon, 2 Sept 2024 at 13:49, Andrew Morton [off-list ref] wrote:

quoted

uprobe_clear_state() is a pretty simple low-level thing.  Side-effects
seem unlikely?

I think uprobe_clear_state() should be removed from fork.c entirely,
made 'static', and then we'd have

        area->xol_mapping.close = uprobe_clear_state;

in __create_xol_area() instead (ok, the arguments change, instead of
looking up "mm->uprobes_state.xol_area", it would get it as the vma
argument)

That's how it should always have been, except we didn't have a close() function.

Hmm?

Indeed, that's much better. I'll prepare a patch.

Thanks!