Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <e2561303-8853-7e16-7eba-001415d34e09@suse.cz>
Date:   Mon, 17 Apr 2023 13:05:40 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.10.0
Subject: Re: [PATCH v2 2/2] mm/slab: break up RCU readers on
 SLAB_TYPESAFE_BY_RCU example code
To:     SeongJae Park <sj@kernel.org>, akpm@linux-foundation.org
Cc:     willy@infradead.org, paulmck@kernel.org, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
References: <20230415033159.4249-1-sj@kernel.org>
 <20230415033159.4249-3-sj@kernel.org>
Content-Language: en-US
From:   Vlastimil Babka <vbabka@suse.cz>
In-Reply-To: <20230415033159.4249-3-sj@kernel.org>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 4/15/23 05:31, SeongJae Park wrote:
> The SLAB_TYPESAFE_BY_RCU example code snippet is having not tiny RCU

Since "tiny RCU" means something quite specific in the RCU world, it can be
confusing to read it in this sense. We could say e.g. "... snippet uses a
single RCU read-side critical section for retries"?

> read-side critical section.  'Documentation/RCU/rculist_nulls.rst' has
> similar example code snippet, and commit da82af04352b ("doc: Update and
> wordsmith rculist_nulls.rst") has broken it.

"has broken it" has quite different meaning than "has broken it up" :) I
guess we could just add the "up", unless someone has an even better wording.

> Apply the change to
> SLAB_TYPESAFE_BY_RCU example code snippet, too.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>
> ---
>  include/linux/slab.h | 8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/include/linux/slab.h b/include/linux/slab.h
> index b18e56c6f06c..6acf1b7c6551 100644
> --- a/include/linux/slab.h
> +++ b/include/linux/slab.h
> @@ -53,16 +53,18 @@
>   * stays valid, the trick to using this is relying on an independent
>   * object validation pass. Something like:
>   *
> + * begin:
>   *  rcu_read_lock();
> - * again:
>   *  obj = lockless_lookup(key);
>   *  if (obj) {
>   *    if (!try_get_ref(obj)) // might fail for free objects
> - *      goto again;
> + *      rcu_read_unlock();
> + *      goto begin;
>   *
>   *    if (obj->key != key) { // not the object we expected
>   *      put_ref(obj);
> - *      goto again;
> + *      rcu_read_unlock();
> + *      goto begin;
>   *    }
>   *  }
>   *  rcu_read_unlock();