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;
MIME-Version: 1.0
References: <1598331582-19923-1-git-send-email-iamjoonsoo.kim@lge.com>
 <e83100ae-d687-3b4e-8256-aee242191ada@suse.cz> <CAAmzW4MRizKRbdt_ZBqs_+OQGsP4f2F6iYDk88u78kYqOZA9NQ@mail.gmail.com>
 <20200827133523.GC3090@techsingularity.net> <CAAmzW4PTgbVzf1dkvnzpHpYYTbLyeNM=OzywUBvkFw9rUhtR4w@mail.gmail.com>
 <CAAmzW4NabiAc+Nh=jtg9fUZTodQHHjf=mFTn=w32AJ43L9o2Aw@mail.gmail.com> <d587d057-75f1-288e-49d2-0c6ea387d414@suse.cz>
In-Reply-To: <d587d057-75f1-288e-49d2-0c6ea387d414@suse.cz>
From:   Joonsoo Kim <js1304@gmail.com>
Date:   Fri, 25 Sep 2020 17:58:25 +0900
Message-ID: <CAAmzW4PSwCZuZh=Nm_QJdOchCg5GBixU4sn7YYOzcbnLuG+J+w@mail.gmail.com>
Subject: Re: [PATCH for v5.9] mm/page_alloc: handle a missing case for
 memalloc_nocma_{save/restore} APIs
To:     Vlastimil Babka <vbabka@suse.cz>
Cc:     Mel Gorman <mgorman@techsingularity.net>,
        Andrew Morton <akpm@linux-foundation.org>,
        Linux Memory Management List <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Michal Hocko <mhocko@kernel.org>,
        "Aneesh Kumar K . V" <aneesh.kumar@linux.ibm.com>,
        kernel-team@lge.com, Joonsoo Kim <iamjoonsoo.kim@lge.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

2020=EB=85=84 9=EC=9B=94 25=EC=9D=BC (=EA=B8=88) =EC=98=A4=ED=9B=84 5:55, V=
lastimil Babka <vbabka@suse.cz>=EB=8B=98=EC=9D=B4 =EC=9E=91=EC=84=B1:
>
> On 9/25/20 6:59 AM, Joonsoo Kim wrote:
> > 2020=EB=85=84 8=EC=9B=94 28=EC=9D=BC (=EA=B8=88) =EC=98=A4=EC=A0=84 8:5=
4, Joonsoo Kim <js1304@gmail.com>=EB=8B=98=EC=9D=B4 =EC=9E=91=EC=84=B1:
> >
> > Hello, Andrew and Vlastimil.
> >
> > It's better to fix this possible bug introduced in v5.9-rc1 before
> > v5.9 is released.
> > Which approach do you prefer?
> > If it is determined, I will immediately send a patch as you suggested.
>
> Hmm both Mel and I preferred the bypass approach and nobody else weighted=
 in, so
> if you don't mind, you can use my suggestion. Hmm maybe alloc_flags & ALL=
OC_CMA
> check should precede migratetype check in the if () to optimize for users=
pace
> allocations?

Okay!
I will implement a bypass approach and send it early next week.

Thanks.