Re: GEN2:Lager: Only 1 core works when turning off the SW8-PIN4

From: Geert Uytterhoeven <geert@linux-m68k.org>
To: Hiep Cao Minh <cm-hiep@jinso.co.jp>
Cc: "Xuan Truong Nguyen" <nx-truong@jinso.co.jp>,
	"Laurent Pinchart" <laurent.pinchart+renesas@ideasonboard.com>,
	duclm <lm-duc@jinso.co.jp>,
	"Ryusuke Sakato" <ryusuke.sakato.bx@renesas.com>,
	"Kuninori Morimoto" <kuninori.morimoto.gx@renesas.com>,
	"Magnus Damm" <magnus.damm@gmail.com>,
	"Geert Uytterhoeven" <geert+renesas@glider.be>,
	稲吉 <h-inayoshi@jinso.co.jp>,
	"Yoshihiro Shimoda" <yoshihiro.shimoda.uh@renesas.com>,
	Dung：人ソ <nv-dung@jinso.co.jp>,
	"Simon Horman" <horms+renesas@verge.net.au>,
	Linux-Renesas <linux-renesas-soc@vger.kernel.org>,
	"Wolfram Sang" <wsa@sang-engineering.com>
Subject: Re: GEN2:Lager: Only 1 core works when turning off the SW8-PIN4
Date: Thu, 29 Sep 2016 15:40:41 +0200	[thread overview]
Message-ID: <CAMuHMdUw_k_SLUO1r5v12Cun40cfJCfA7Em4e-zupEDsxcEdhw@mail.gmail.com> (raw)
In-Reply-To: <57ECF4EA.4010704@jinso.co.jp>

Hi Hiep,

On Thu, Sep 29, 2016 at 1:03 PM, Hiep Cao Minh <cm-hiep@jinso.co.jp> wrote:
> I'd like to report the issue of the CPU operation.
> We tested and found it on Lager board at linux-v4.8-rcx.
>
> The test procedure is the following:
>
> Confirmation command:
>
> # cat /proc/interrupts
> CPU0
> 19: 2509 GIC-0 27 Level arch_timer
> 21: 0 GIC-0 36 Level e6050000.gpio
> 22: 0 GIC-0 37 Level e6051000.gpio
> 23: 0 GIC-0 38 Level e6052000.gpio
> 24: 0 GIC-0 39 Level e6053000.gpio
> 25: 0 GIC-0 40 Level e6054000.gpio
> 26: 0 GIC-0 41 Level e6055000.gpio
> 27: 23 GIC-0 101 Level e61f0000.thermal
> …”
>
> This issue appears when we changed the SW8-PIN4 of Lager board.
>
> SW8-PIN4: ON
>
> + At linux-v4.7: OK (4 cores work together normally).
> + At linux-v4.8-rc2: OK (4 cores work together normally).
>
> SW8-PIN4: OFF
>
> + At linux-v4.7: -> OK(4 cores work together normally).
> + At linux-v4.8-rc2: -> NG(Only 1 core works).

And the kernel prints "Unable to boot CPU%u when MD21 is set", right?

> We tried to find out the issued patch and we realize that it happens from
> the following patch:
> " 043248c Merge tag 'armsoc-dt' of
> git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc"

The issue you're seeing is due to a combination of commits 5f3bca0db8ac01a7
("ARM: shmobile: apmu: Add APMU DT support via Enable method") and
dc378795156d980c ("ARM: dts: r8a7790: Add APMU nodes").

When debug mode is enabled (SW8-4 = OFF), trying to boot secondary CPUs may
lock up the system after a cold boot, depending of boot load version.  Hence
we explicitly prohibit that. BTW, this has been the case on Koelsch since commit
277efd30cfc72ec2 ("ARM: shmobile: Check r8a7791 MD21 at SMP boot").

Now, does series "[PATCH/RFT 0/4] ARM: shmobile: R-Car Gen2: Allow booting
secondary CPU cores in debug mode" (included in all renesas-drivers releases
during September) fix it for you?

Thanks!

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds