RE: [PATCH V1] serial: imx: revert setup DCEDTE early and ensure DCD and RI irqs to be off

From: Steve Twiss
Date: Wed May 24 2017 - 12:09:19 EST


Hi Fabio,

On 24 May 2017 14:53 Fabio Estevam wrote,
> Subject: Re: [PATCH V1] serial: imx: revert setup DCEDTE early and ensure DCD and RI irqs to be off
> On Wed, May 24, 2017 at 9:49 AM, Steve Twiss wrote:
>
> > I used this tag: jb4.3_1.1.1-ga
> > http://git.freescale.com/git/cgit.cgi/imx/uboot-imx.git/commit/?id=jb4.3_1.1.1-ga
> >
> > This repo is slightly modified for a project I am working on, but it only queries
> > the I2C for devices before printing out some messages, so minor changes I
> > think. Here's the output from the console:
>
> I am not able to reproduce it using this same U-Boot version on my
> mx6sabresd board:
>
> https://pastebin.com/ppnrnidr

That's very interesting.
Once I apply the test change from Uwe, I can get a console trace.

> Uwe Kleine-KÃnig wrote:
> > Can you try to just remove the line
> > writel(0, sport->port.membase + UFCR);

The differences in the kernel console logs are ...

My board : Your board differences
Board: i.MX6Q-SABRESD: unknown-board Board: 0x63012 [POR ] : Board: i.MX6Q-SABRESD: unknown-board Board: 0x63011 [POR ]
Found PFUZE100! deviceid=10,revid=11 : Found PFUZE100! deviceid=10,revid=10

... and others, but.. I do not have "console=ttymxc0,115200" in my kernel
command line.

Yes. I can verify, using the raw linux-next/v4.12-rc2 I get the console error.
When I add "console=ttymxc0,115200" to my U-boot environment and the
kernel command-line, the kernel console starts working again.

That would indicate the test on the i.MX6DL might be misleading, since my
colleague ran that using their bootloader set-up. So, my assertion this was
only visible on the i.MX6Q could be wrong.

There's no way to to check this until tomorrow now.
I'll check check on the i.MX6DL using my bootloader environment when I
get access to an i.MX6DL board.

Regards,
Stephen