Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Waiman Long <llong@redhat.com>
Subject: Re: [PATCH-next] mm/memcontrol.c: Fix potential uninitialized
 variable warning
To:     Andrew Morton <akpm@linux-foundation.org>
Cc:     Johannes Weiner <hannes@cmpxchg.org>,
        Michal Hocko <mhocko@kernel.org>,
        Vladimir Davydov <vdavydov.dev@gmail.com>,
        linux-kernel@vger.kernel.org, cgroups@vger.kernel.org,
        linux-mm@kvack.org, Muchun Song <songmuchun@bytedance.com>
References: <20210526193602.8742-1-longman@redhat.com>
 <20210526134321.42bbd4a9dcbcf53e855c5b1b@linux-foundation.org>
Message-ID: <9fb60ba4-258d-2bb5-57af-c174df9b1a31@redhat.com>
Date:   Wed, 26 May 2021 18:29:02 -0400
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.9.0
MIME-Version: 1.0
In-Reply-To: <20210526134321.42bbd4a9dcbcf53e855c5b1b@linux-foundation.org>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-US
Precedence: bulk

On 5/26/21 4:43 PM, Andrew Morton wrote:
> On Wed, 26 May 2021 15:36:02 -0400 Waiman Long <longman@redhat.com> wrote:
>
>> If the -Wno-maybe-uninitialized gcc option is not specified, compilation
>> of memcontrol.c may generate the following warnings:
>>
>> mm/memcontrol.c: In function ‘refill_obj_stock’:
>> ./arch/x86/include/asm/irqflags.h:127:17: warning: ‘flags’ may be used uninitialized in this function [-Wmaybe-uninitialized]
>>    return !(flags & X86_EFLAGS_IF);
>>            ~~~~~~~^~~~~~~~~~~~~~~~
>> mm/memcontrol.c:3216:16: note: ‘flags’ was declared here
>>    unsigned long flags;
>>                  ^~~~~
>> In file included from mm/memcontrol.c:29:
>> mm/memcontrol.c: In function ‘uncharge_page’:
>> ./include/linux/memcontrol.h:797:2: warning: ‘objcg’ may be used uninitialized in this function [-Wmaybe-uninitialized]
>>    percpu_ref_put(&objcg->refcnt);
>>    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>
>> Fix that by properly initializing *pflags in get_obj_stock() and
>> introducing a use_objcg bool variable in uncharge_page() to avoid
>> potentially accessing the struct page data twice.
>>
> Thanks.  I'll queue this as a fix against your "mm/memcg: optimize user
> context object stock access".
>
Thanks for that.

Cheers,
Longman