Message-ID: <937f1320-6b7e-9aa2-2a21-7fd2f94eeb32@colorfullife.com>
Date:   Thu, 23 Dec 2021 12:52:17 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.4.0
Subject: Re: [PATCH] mm/util.c: Make kvfree() safe for calling while holding
 spinlocks
Content-Language: en-US
To:     Vasily Averin <vvs@virtuozzo.com>,
        LKML <linux-kernel@vger.kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>
Cc:     cgel.zte@gmail.com, shakeelb@google.com, rdunlap@infradead.org,
        dbueso@suse.de, unixbhaskar@gmail.com, chi.minghao@zte.com.cn,
        arnd@arndb.de, Zeal Robot <zealci@zte.com.cn>, linux-mm@kvack.org,
        1vier1@web.de, stable@vger.kernel.org
References: <20211222194828.15320-1-manfred@colorfullife.com>
 <3ca4dec8-f0bd-740f-73c8-34fc6fc1cf66@virtuozzo.com>
From:   Manfred Spraul <manfred@colorfullife.com>
In-Reply-To: <3ca4dec8-f0bd-740f-73c8-34fc6fc1cf66@virtuozzo.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hello Vasily,

On 12/23/21 08:21, Vasily Averin wrote:
>
> I would prefer to release memory ASAP if it's possible.
> What do you think about this change?
> --- a/mm/util.c
> +++ b/mm/util.c
> @@ -614,9 +614,12 @@ EXPORT_SYMBOL(kvmalloc_node);
>    */
>   void kvfree(const void *addr)
>   {
> -       if (is_vmalloc_addr(addr))
> -               vfree(addr);
> -       else
> +       if (is_vmalloc_addr(addr)) {
> +               if (in_atomic())
> +                       vfree_atomic();
> +               else
> +                       vfree(addr);
> +       } else
>                  kfree(addr);
>   }
>   EXPORT_SYMBOL(kvfree);
>
Unfortunately this cannot work:

> /*
> * Are we running in atomic context?  WARNING: this macro cannot
> * always detect atomic context; in particular, it cannot know about
> * held spinlocks in non-preemptible kernels.  Thus it should not be
> * used in the general case to determine whether sleeping is possible.
> * Do not use in_atomic() in driver code.
> */
> #define in_atomic()     (preempt_count() != 0)
>

--

     Manfred