From mboxrd@z Thu Jan  1 00:00:00 1970
Content-Type: multipart/mixed; boundary="===============5147960362129724043=="
MIME-Version: 1.0
From: Valentin Schneider <valentin.schneider@arm.com>
To: kbuild-all@lists.01.org
Subject: Re: [RFC PATCH 1/2] sched/topology: Get rid of NUMA overlapping groups
Date: Wed, 03 Feb 2021 19:26:51 +0000
Message-ID: <jhjczxh2br8.mognet@arm.com>
In-Reply-To: <20210203183840.GY2696@kadam>
List-Id: <oe-kbuild-all.lists.linux.dev>

--===============5147960362129724043==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable


Hi Dan,

On 03/02/21 21:38, Dan Carpenter wrote:
> Hi Valentin,
>
> url:    https://github.com/0day-ci/linux/commits/Valentin-Schneider/sched=
-topology-Get-rid-of-overlapping-groups/20210204-000142
> base:   https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git 075a2=
8439d0c8eb6d3c799e1eed24bb9bc7750cd
> config: s390-randconfig-m031-20210204 (attached as .config)
> compiler: s390-linux-gcc (GCC) 9.3.0
>
> If you fix the issue, kindly add following tag as appropriate
> Reported-by: kernel test robot <lkp@intel.com>
> Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
>
> smatch warnings:
> kernel/sched/topology.c:1012 find_node_domain() error: uninitialized symb=
ol 'parent'.
>
> vim +/parent +1012 kernel/sched/topology.c
>
> 02d81c8f49452e Valentin Schneider   2021-02-03   985  static struct sched=
_domain *find_node_domain(struct sched_domain *sd)
> 02d81c8f49452e Valentin Schneider   2021-02-03   986  {
> 02d81c8f49452e Valentin Schneider   2021-02-03   987  	struct sched_domai=
n *parent;
> 02d81c8f49452e Valentin Schneider   2021-02-03   988  =

> 02d81c8f49452e Valentin Schneider   2021-02-03   989  	BUG_ON(!(sd->flags=
 & SD_NUMA));
> 02d81c8f49452e Valentin Schneider   2021-02-03   990  =

> 02d81c8f49452e Valentin Schneider   2021-02-03   991  	/* Get to the leve=
l above NODE */
> 02d81c8f49452e Valentin Schneider   2021-02-03   992  	while (sd && sd->c=
hild) {
>
> Presumably this is always true when the function is called so this
> warning is a false positive?
>

Yeah, the passed sd has SD_NUMA so it for sure has at least one child (with
or without SD_NUMA). Not the cleanest way to go about it for sure, I'll see
about reformulating this if this doesn't get shot down :-)

Thanks!

--===============5147960362129724043==--