[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <tbm6emntje2lcwmg5xa6whlhsghwflbyb6p7m4y72dffenttqz@g2hftzczixxx>
Date: Wed, 6 Aug 2025 11:20:19 +0100
From: Pedro Falcato <pfalcato@...e.de>
To: "Liam R. Howlett" <Liam.Howlett@...cle.com>,
Andrew Morton <akpm@...ux-foundation.org>, Matthew Wilcox <willy@...radead.org>,
maple-tree@...ts.infradead.org, linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] maple_tree: Use kfree_rcu in ma_free_rcu
On Tue, Aug 05, 2025 at 09:25:13PM -0400, Liam R. Howlett wrote:
> * Pedro Falcato <pfalcato@...e.de> [250718 13:21]:
> > kfree_rcu is an optimized version of call_rcu + kfree. It used to not be
> > possible to call it on non-kmalloc objects, but this restriction was
> > lifted ever since SLOB was dropped from the kernel, and since commit
> > 6c6c47b063b5 ("mm, slab: call kvfree_rcu_barrier() from kmem_cache_destroy()").
> >
> > Thus, replace call_rcu + mt_free_rcu with kfree_rcu.
> [snip]
> > static void mt_set_height(struct maple_tree *mt, unsigned char height)
> > @@ -5281,7 +5274,7 @@ static void mt_free_walk(struct rcu_head *head)
> > mt_free_bulk(node->slot_len, slots);
> >
> > free_leaf:
> > - mt_free_rcu(&node->rcu);
> > + mt_free_one(node);
>
> Why are we still using mt_free_one()? Couldn't this also be dropped in
> favour of kfree() or does kfree() not work for kmem_cache?
kfree() also works (since SLOB was dropped). I thought you wanted mt_free_one
for style points, but I can replace all calls with a direct kfree() if you prefer.
--
Pedro
Powered by blists - more mailing lists