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: <CAFJ_xbqT8h2Exix3S6AGgB7W1N0u-=WKffAyb7Hk9-8K8FBwKA@mail.gmail.com>
 <20210127100454.GK196782@linux.ibm.com> <CAFJ_xboaFNQ9NuZ1rhH8WdejoFRzvez9cp2AQ59rKY6T_xZ-_w@mail.gmail.com>
 <20210127111858.GA273567@linux.ibm.com> <CAFJ_xbo8Zv9VdJibC106sFOqoYsVhifm0eh=VWtMzeoUE4KVWA@mail.gmail.com>
 <CAFJ_xbrwLwgDfCyHA=PmJ8j_3dJXqVNxmv7e+ATQAAa9n3de2w@mail.gmail.com> <20210127182651.GA281042@linux.ibm.com>
In-Reply-To: <20210127182651.GA281042@linux.ibm.com>
From:   =?UTF-8?Q?=C5=81ukasz_Majczak?= <lma@semihalf.com>
Date:   Wed, 27 Jan 2021 20:18:34 +0100
Message-ID: <CAFJ_xbqHq-P8z-xpkVDAQsG1s4j5FMmXFnyXLzd=ja8y_=8LfA@mail.gmail.com>
Subject: Re: PROBLEM: Crash after mm: fix initialization of struct page for
 holes in memory layout
To:     Mike Rapoport <rppt@linux.ibm.com>
Cc:     Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org,
        =?UTF-8?Q?Rados=C5=82aw_Biernacki?= <rad@semihalf.com>,
        Marcin Wojtas <mw@semihalf.com>,
        Alex Levin <levinale@google.com>,
        Guenter Roeck <groeck@google.com>,
        Jesse Barnes <jsbarnes@google.com>,
        Chris Wilson <chris@chris-wilson.co.uk>,
        "Sarvela, Tomi P" <tomi.p.sarvela@intel.com>,
        =?UTF-8?Q?=C5=81ukasz_Bartosik?= <lb@semihalf.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi Mike,

Great ! it seems to work well - I have built a valila kernel v5.11-rc5
with your patch and it boots properly.
Full log available here:
https://gist.github.com/semihalf-majczak-lukasz/ea89bf52f6fad7907a18d1870e7=
ce9bd

Best regards,
Lukasz

=C5=9Br., 27 sty 2021 o 19:27 Mike Rapoport <rppt@linux.ibm.com> napisa=C5=
=82(a):
>
> Hi Lukasz,
>
> On Wed, Jan 27, 2021 at 02:15:53PM +0100, =C5=81ukasz Majczak wrote:
> > Hi Mike,
> >
> > I have started bisecting your patch and I have figured out that there
> > might be something wrong with clamping - with comments out these lines
> > it started to work.
> > The full log (with logs from below patch) can be found here:
> > https://gist.github.com/semihalf-majczak-lukasz/3cecbab0ddc59a6c3ce11dd=
c29645725
> > it's fresh - I haven't analyze it yet, just sharing with hope it will h=
elp.
>
> Thanks, that helps!
>
> The first page is never considered by the kernel as memory and so
> arch_zone_lowest_possible_pfn[ZONE_DMA] is set to 0x1000. As the result,
> init_unavailable_mem() skips pfn 0 and then __SetPageReserved(page) in
> reserve_bootmem_region() panics because the struct page for pfn 0 remains
> poisoned.
>
> Can you please try the below patch on top of v5.11-rc5?
>
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 783913e41f65..3ce9ef238dfc 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -7083,10 +7083,11 @@ void __init free_area_init_memoryless_node(int ni=
d)
>  static u64 __init init_unavailable_range(unsigned long spfn, unsigned lo=
ng epfn,
>                                          int zone, int nid)
>  {
> -       unsigned long pfn, zone_spfn, zone_epfn;
> +       unsigned long pfn, zone_spfn =3D 0, zone_epfn;
>         u64 pgcnt =3D 0;
>
> -       zone_spfn =3D arch_zone_lowest_possible_pfn[zone];
> +       if (zone > 0)
> +               zone_spfn =3D arch_zone_highest_possible_pfn[zone - 1];
>         zone_epfn =3D arch_zone_highest_possible_pfn[zone];
>
>         spfn =3D clamp(spfn, zone_spfn, zone_epfn);
>
>
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index eed54ce26ad1..9f4468c413a1 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -7093,9 +7093,11 @@ static u64 __init
> > init_unavailable_range(unsigned long spfn, unsigned long epfn,
> >         zone_spfn =3D arch_zone_lowest_possible_pfn[zone];
> >         zone_epfn =3D arch_zone_highest_possible_pfn[zone];
> >
> > -       spfn =3D clamp(spfn, zone_spfn, zone_epfn);
> > -       epfn =3D clamp(epfn, zone_spfn, zone_epfn);
> > -
> > +       //spfn =3D clamp(spfn, zone_spfn, zone_epfn);
> > +       //epfn =3D clamp(epfn, zone_spfn, zone_epfn);
> > +       pr_info("LMA DBG: zone_spfn: %llx, zone_epfn %llx\n",
> > zone_spfn, zone_epfn);
> > +       pr_info("LMA DBG: spfn: %llx, epfn %llx\n", spfn, epfn);
> > +       pr_info("LMA DBG: clamp_spfn: %llx, clamp_epfn %llx\n",
> > clamp(spfn, zone_spfn, zone_epfn), clamp(epfn, zone_spfn, zone_epfn));
> >         for (pfn =3D spfn; pfn < epfn; pfn++) {
> >                 if (!pfn_valid(ALIGN_DOWN(pfn, pageblock_nr_pages))) {
> >                         pfn =3D ALIGN_DOWN(pfn, pageblock_nr_pages)
> >
> > Best regards,
> > Lukasz
> >
> >
> > =C5=9Br., 27 sty 2021 o 13:15 =C5=81ukasz Majczak <lma@semihalf.com> na=
pisa=C5=82(a):
> > >
> > > Unfortunately nothing :( my current kernel command line contains:
> > > console=3DttyS0,115200n8 debug earlyprintk=3Dserial loglevel=3D7
> > >
> > > I was thinking about using earlycon, but it seems to be blocked.
> > > (I think the lack of earlycon might be related to Chromebook HW
> > > security design. There is an EC controller which is a part of AP ->
> > > serial chain as kernel messages are considered sensitive from a
> > > security standpoint.)
> > >
> > > Best regards,
> > > Lukasz
> > >
> > > =C5=9Br., 27 sty 2021 o 12:19 Mike Rapoport <rppt@linux.ibm.com> napi=
sa=C5=82(a):
> > > >
> > > > On Wed, Jan 27, 2021 at 11:08:17AM +0100, =C5=81ukasz Majczak wrote=
:
> > > > > Hi Mike,
> > > > >
> > > > > Actually I have a serial console attached (via servo device), but
> > > > > there is no output :( and also the reboot/crash is very fast/imme=
diate
> > > > > after power on.
> > > >
> > > > If you boot with earlyprintk=3Dserial are there any messages?
> > > >
> > > > > Best regards
> > > > > Lukasz
> > > > >
> > > > > =C5=9Br., 27 sty 2021 o 11:05 Mike Rapoport <rppt@linux.ibm.com> =
napisa=C5=82(a):
> > > > > >
> > > > > > Hi Lukasz,
> > > > > >
> > > > > > On Wed, Jan 27, 2021 at 10:22:29AM +0100, =C5=81ukasz Majczak w=
rote:
> > > > > > > Crash after mm: fix initialization of struct page for holes i=
n memory layout
> > > > > > >
> > > > > > > Hi,
> > > > > > > I was trying to run v5.11-rc5 on my Samsung Chromebook Pro (C=
aroline),
> > > > > > > but I've noticed it has crashed - unfortunately it seems to h=
appen at
> > > > > > > a very early stage - No output to the console nor to the scre=
en, so I
> > > > > > > have started a bisect (between 5.11-rc4 - which works just fi=
nd - and
> > > > > > > 5.11-rc5),
> > > > > > > bisect results points to:
> > > > > > >
> > > > > > > d3921cb8be29 mm: fix initialization of struct page for holes =
in memory layout
> > > > > > >
> > > > > > > Reproduction is just to build and load the kernel.
> > > > > > >
> > > > > > > If it will help any how I am attaching:
> > > > > > > - /proc/cpuinfo (from healthy system):
> > > > > > > https://gist.github.com/semihalf-majczak-lukasz/3517867bf39f0=
7377c1a785b64a97066
> > > > > > > - my .config file (for a broken system):
> > > > > > > https://gist.github.com/semihalf-majczak-lukasz/584b329f1bf3e=
43b53efe8e18b5da33c
> > > > > > >
> > > > > > > If there is anything I could add/do/test to help fix this ple=
ase let me know.
> > > > > >
> > > > > > Chris Wilson also reported boot failures on several Chromebooks=
:
> > > > > >
> > > > > > https://lore.kernel.org/lkml/161160687463.28991.354987542182281=
928@build.alporthouse.com
> > > > > >
> > > > > > I presume serial console is not an option, so if you could boot=
 with
> > > > > > earlyprintk=3Dvga and see if there is anything useful printed o=
n the screen
> > > > > > it would be really helpful.
> > > > > >
> > > > > > > Best regards
> > > > > > > Lukasz
> > > > > >
> > > > > > --
> > > > > > Sincerely yours,
> > > > > > Mike.
> > > >
> > > > --
> > > > Sincerely yours,
> > > > Mike.
>
> --
> Sincerely yours,
> Mike.