iMX7d cpu1 issue

[ilya@compulab.co.il (Ilya Ledvich)]

Hi!

On 02/18/2016 05:24 PM, Frank Li wrote:
>
>
>> -----Original Message-----
>> From: Shawn Guo [mailto:shawnguo at kernel.org]
>> Sent: Thursday, February 18, 2016 8:09 AM
>> To: Ilya Ledvich <ilya@compulab.co.il>
>> Cc: linux-arm-kernel at lists.infradead.org; Frank Li <frank.li@nxp.com>;
>> Yongcai Huang <anson.huang@nxp.com>
>> Subject: Re: iMX7d cpu1 issue
>>
>> On Tue, Feb 16, 2016 at 09:56:47AM +0200, Ilya Ledvich wrote:
>>> Hi Shawn,
>>> I'm working on Compulab CL-SOM-IMX7 device
>>> (http://www.compulab.co.il/products/computer-on-modules/cl-som-
>> imx7-fr
>>> eescale-i-mx-7-system-on-module)
>>> based on Freescale iMX7d CPU.
>>> I use the Freescale latest BSP imx_3.14.52_1.1.0_ga  and the Linux
>>> mainline v4.4-rc6 as a reference.
>>> I observe an issue with bringing up the second CPU core. Briefly, the
>>> kernel can successfully boot only when CPU1 fails to come online.
>>> In this case the boot log shows the following messages:
>>> ...
>>> CPU1: failed to come online
>>> Brought up 1 CPUs
>>> SMP: Total of 1 processors activated (6.00 BogoMIPS).
>>> ...
>>>
>>> Otherwise, the device tries to start kernel several times until the
>>> boot process eventually crashes. Please see attached logs for more
>>> details.
>>> This never happens on mainline Linus kernel, because it never tries to
>>> bring up CPU1.
>>> Unfortunately, we cannot verify this behavior on reference board cause
>>> we have no one.
>>> Have you ever seen such behavior? Could you please advice something?
>
> Did you set arm,cpu-registers-not-fw-configured; in arch timer?

Yes, it's set in the imx7d.dtsi file included by my device .dts.

> How about your uboot version?
>
> Latest upstream uboot already boot A7 as no-security mode.

I actually use v2016.01 from the upstream and u-boot sources are based 
on on mx7dsabresd.c.

>
> Best regards
> Frank Li
>

Kind regards,
Ilya.