Re: [PATCH 1/2] mm/memory_hotplug: fix memory block reference leak on remove

[Oscar Salvador]

On Mon, Apr 27, 2026 at 04:02:17PM +0800, Muchun Song wrote:
> 
> 
> > On Apr 27, 2026, at 15:49, Oscar Salvador <osalvador@xxxxxxx> wrote:
> > 2) I kind of dislike having an internal put_device() lingering here in
> >   memory-hotplug code, it feels like it does not really belong here.
> >   Ideally we should have a high-level function in drivers/base/memory.c
> >   that calls put_device itself.
> >   Something like "put_memblock_dev", dunno, names are hard.
> > 
> 
> I share your perspective. The current naming of find_memory_block_by_id
> is ambiguous as it fails to signal the internal 'get' operation. To improve
> clarity and reduce errors, it should be renamed to memory_block_get_by_id.
> Pairing this with a new memory_block_put function to wrap put_device
> would ensure a more robust and intuitive API.

Even better, yes.

Thanks!


-- 
Oscar Salvador
SUSE Labs