Message-ID: <b50b7c01-c4fc-87bd-f5d9-80e6438cf670@redhat.com>
Date:   Wed, 15 Dec 2021 11:58:45 -0500
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.3.0
Subject: Re: [PATCH-next v3] mm/memcg: Properly handle memcg_stock access for
 PREEMPT_RT
Content-Language: en-US
To:     Johannes Weiner <hannes@cmpxchg.org>,
        Michal Hocko <mhocko@kernel.org>,
        Vladimir Davydov <vdavydov.dev@gmail.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-kernel@vger.kernel.org, cgroups@vger.kernel.org,
        linux-mm@kvack.org,
        Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        Thomas Gleixner <tglx@linutronix.de>
References: <20211214144412.447035-1-longman@redhat.com>
 <20211215043657.ngmxlk6rgc2ysbmz@offworld>
From:   Waiman Long <longman@redhat.com>
In-Reply-To: <20211215043657.ngmxlk6rgc2ysbmz@offworld>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 12/14/21 23:36, Davidlohr Bueso wrote:
> On Tue, 14 Dec 2021, Waiman Long wrote:
>
>> @@ -2189,7 +2194,7 @@ static void drain_local_stock(struct 
>> work_struct *dummy)
>>      * drain_stock races is that we always operate on local CPU stock
>>      * here with IRQ disabled
>>      */
>> -    local_irq_save(flags);
>> +    local_lock_irqsave(&memcg_stock.lock, flags);
>>
>>     stock = this_cpu_ptr(&memcg_stock);
>>     drain_obj_stock(&stock->irq_obj);
>
> So here there is still the problem that you can end up taking sleeping 
> locks
> with irqs disabled via obj_cgroup_put() >> obj_cgroup_release() - ie: the
> percpu_ref_switch_lock and css_set_lock. It had occurred to me to promote
> the former to a raw spinlock, but doubt we can get away with the latter.

Yes, it is tricky to avoid all these. One possibility to to delay 
obj_cgroup_release() through RCU. However, this is outside the scope of 
this patch.

Cheers,
Longman