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: <1590561903-13186-1-git-send-email-iamjoonsoo.kim@lge.com>
 <1590561903-13186-8-git-send-email-iamjoonsoo.kim@lge.com> <20200609135431.GI22623@dhcp22.suse.cz>
In-Reply-To: <20200609135431.GI22623@dhcp22.suse.cz>
From:   Joonsoo Kim <js1304@gmail.com>
Date:   Wed, 10 Jun 2020 14:12:56 +0900
Message-ID: <CAAmzW4Ou-65wW2iHFDZdOYrEkb2afrSox7sVwd1zEo-VXSE30Q@mail.gmail.com>
Subject: Re: [PATCH v2 07/12] mm/hugetlb: do not modify user provided gfp_mask
To:     Michal Hocko <mhocko@kernel.org>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Linux Memory Management List <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>, kernel-team@lge.com,
        Vlastimil Babka <vbabka@suse.cz>,
        Christoph Hellwig <hch@infradead.org>,
        Roman Gushchin <guro@fb.com>,
        Mike Kravetz <mike.kravetz@oracle.com>,
        Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

2020=EB=85=84 6=EC=9B=94 9=EC=9D=BC (=ED=99=94) =EC=98=A4=ED=9B=84 10:54, M=
ichal Hocko <mhocko@kernel.org>=EB=8B=98=EC=9D=B4 =EC=9E=91=EC=84=B1:
>
> On Wed 27-05-20 15:44:58, Joonsoo Kim wrote:
> > From: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> >
> > It's not good practice to modify user input. Instead of using it to
> > build correct gfp_mask for APIs, this patch introduces another gfp_mask
> > field, __gfp_mask, for internal usage.
>
> Ugh, this is really ugly. It is just hugetlb to add __GFP_THISNODE as a
> special case. This is an ugly hack but I do not think we want to work
> around it by yet another hack. Moreover it seems that the __GFP_THISNODE
> might be not needed anymore as pointed out in a reply to earlier patch.

If you mean __GFP_THISNODE handling is ugly, as you pointed out,
__GFP_THISNODE handling would be removed in the next version.

If you mean introducing __gfp_mask is ugly, I will try to use a local varia=
ble
to keep modified gfp_mask rather than introducing a field in alloc_control.

Thanks.