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;
Content-Type: text/plain;
        charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.80.23.2.2\))
Subject: Re: compaction: VM_BUG_ON_PAGE(!zone_spans_pfn(page_zone(page), pfn))
From:   Qian Cai <cai@lca.pw>
In-Reply-To: <20200505124314.GA5029@MiWiFi-R3L-srv>
Date:   Tue, 5 May 2020 09:20:51 -0400
Cc:     Mel Gorman <mgorman@suse.de>, Vlastimil Babka <vbabka@suse.cz>,
        Michal Hocko <mhocko@kernel.org>,
        David Hildenbrand <david@redhat.com>,
        Linux-MM <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Mike Rapoport <rppt@linux.ibm.com>
Content-Transfer-Encoding: quoted-printable
Message-Id: <46D27820-4B67-45F0-898C-B0547D654626@lca.pw>
References: <8C537EB7-85EE-4DCF-943E-3CC0ED0DF56D@lca.pw>
 <20200424034353.GD4247@MiWiFi-R3L-srv>
 <A49F2991-4319-4207-9B6E-CC8563102C27@lca.pw>
 <20200505124314.GA5029@MiWiFi-R3L-srv>
To:     Baoquan He <bhe@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On May 5, 2020, at 8:43 AM, Baoquan He <bhe@redhat.com> wrote:
>=20
> Hi,
>=20
> On 04/24/20 at 09:45am, Qian Cai wrote:
>>=20
>>=20
>>> On Apr 23, 2020, at 11:43 PM, Baoquan He <bhe@redhat.com> wrote:
>>>=20
>>> On 04/23/20 at 05:25pm, Qian Cai wrote:
>>>> Compaction starts to crash below on linux-next today. The faulty =
page belongs to Node 0 DMA32 zone.
>>>> I=E2=80=99ll continue to narrow it down, but just want to give a =
headup in case someone could beat me to it.
>>>>=20
>>>> Debug output from free_area_init_core()
>>>> [    0.000000] KK start page =3D ffffea0000000040, end page =3D =
ffffea0000040000, nid =3D 0 DMA
>>>> [    0.000000] KK start page =3D ffffea0000040000, end page =3D =
ffffea0004000000, nid =3D 0 DMA32
>>>> [    0.000000] KK start page =3D ffffea0004000000, end page =3D =
ffffea0012000000, nid =3D 0 NORMAL
>>>> [    0.000000] KK start page =3D ffffea0012000000, end page =3D =
ffffea0021fc0000, nid =3D 4 NORMAL
>>>=20
>>> Where are these printed? They are the direct mapping address of =
page?
>>=20
>> =46rom this debug patch. Yes, direct mapping.
>=20
> Can you try below patch? I may get why this is caused, not sure if the
> place is right.=20
>=20
> diff --git a/mm/compaction.c b/mm/compaction.c
> index 177c11a8f3b9..e26972f26414 100644
> --- a/mm/compaction.c
> +++ b/mm/compaction.c
> @@ -1409,7 +1409,9 @@ fast_isolate_freepages(struct compact_control =
*cc)
> 				cc->free_pfn =3D highest;
> 			} else {
> 				if (cc->direct_compaction && =
pfn_valid(min_pfn)) {
> -					page =3D pfn_to_page(min_pfn);
> +					page =3D =
pageblock_pfn_to_page(min_pfn,
> +						=
pageblock_end_pfn(min_pfn),
> +						cc->zone);
> 					cc->free_pfn =3D min_pfn;
> 				}
> 			}

I have not had luck to reproduce this again yet, but feel free to move =
forward with the patch anyway if you are comfortable to do so, so at =
least people could review it properly.