Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Tarun Sahu <tsahu@linux.ibm.com>
To:     "Verma, Vishal L" <vishal.l.verma@intel.com>,
        "Schofield, Alison" <alison.schofield@intel.com>
Cc:     "Williams, Dan J" <dan.j.williams@intel.com>,
        "Jiang, Dave" <dave.jiang@intel.com>,
        "linux-cxl@vger.kernel.org" <linux-cxl@vger.kernel.org>,
        "nvdimm@lists.linux.dev" <nvdimm@lists.linux.dev>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "aneesh.kumar@linux.ibm.com" <aneesh.kumar@linux.ibm.com>,
        "jaypatel@linux.ibm.com" <jaypatel@linux.ibm.com>
Subject: Re: [PATCH] dax/kmem: Pass valid argument to
 memory_group_register_static
In-Reply-To: <b7e620efa0de6b9f7a8ae9ce51d8dd562f384cdc.camel@intel.com>
References: <20230620140332.30578-1-tsahu@linux.ibm.com>
 <ZJIz07hQiXr/MghO@aschofie-mobl2> <87cz1p1h8c.fsf@linux.ibm.com>
 <b7e620efa0de6b9f7a8ae9ce51d8dd562f384cdc.camel@intel.com>
Date:   Thu, 22 Jun 2023 12:45:40 +0530
Message-ID: <878rcc0xyb.fsf@linux.ibm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk


Hi Vishal,

"Verma, Vishal L" <vishal.l.verma@intel.com> writes:

> On Wed, 2023-06-21 at 11:36 +0530, Tarun Sahu wrote:
>> Hi Alison,
>>=20
>> Alison Schofield <alison.schofield@intel.com> writes:
>>=20
>> > On Tue, Jun 20, 2023 at 07:33:32PM +0530, Tarun Sahu wrote:
>> > > memory_group_register_static takes maximum number of pages as the ar=
gument
>> > > while dev_dax_kmem_probe passes total_len (in bytes) as the argument.
>> >=20
>> > This sounds like a fix. An explanation of the impact and a fixes tag
>> > may be needed. Also, wondering how you found it.
>> >=20
>> Yes, it is a fix, I found it during dry code walk-through.
>> There is not any impact as such. As,
>> memory_group_register_static just set the max_pages limit which
>> is used in auto_movable_zone_for_pfn to determine the zone.
>>=20
>> which might cause these condition to behave differently,
>>=20
>> This will be true always so jump will happen to kernel_zone
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0if (!auto_movable_can_on=
line_movable(NUMA_NO_NODE, group, nr_pages))
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0=C2=A0goto kernel_zone;
>> ---
>> kernel_zone:
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0return default_kernel_zo=
ne_for_pfn(nid, pfn, nr_pages);
>>=20
>> ---
>>=20
>> Here, In below, zone_intersects compare range will be larger as nr_pages
>> will be higher (derived from total_len passed in dev_dax_kmem_probe).
>>=20
>> static struct zone *default_kernel_zone_for_pfn(int nid, unsigned long s=
tart_pfn,
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0=C2=A0unsigned long nr_pages)
>> {
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0struct pglist_data *pgda=
t =3D NODE_DATA(nid);
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0int zid;
>>=20
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0for (zid =3D 0; zid < ZO=
NE_NORMAL; zid++) {
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0=C2=A0struct zone *zone =3D &pgdat->node_zones[zid];
>>=20
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0=C2=A0if (zone_intersects(zone, start_pfn, nr_pages))
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0ret=
urn zone;
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0}
>>=20
>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0return &pgdat->node_zone=
s[ZONE_NORMAL];
>> }
>>=20
>> In Mostly cases, ZONE_NORMAL will be returned. But there is no
>> crash/panic issues involved here, only decision making on selecting zone
>> is affected.
>>=20
>
> Hi Tarun,
>
> Good find! With a Fixes tag, and perhaps inclusion of a bit more of
> this detail described in the commit message, feel free to add:
>
Thanks for reviewing, sent the updated version.

> Reviewed-by: Vishal Verma <vishal.l.verma@intel.com>