Re: i.MX watchdog druver regression

[Sascha Hauer <sha@pengutronix.de>]

On Fri, Oct 21, 2022 at 01:47:35PM +0200, Ahmad Fatoum wrote:
> On 21.10.22 13:43, Sascha Hauer wrote:
> > On Fri, Oct 21, 2022 at 12:53:59PM +0300, Alexander Shiyan wrote:
> >> Hello!
> >>
> >> An error occurred while registering the i.MX watchdog driver:
> >>
> >> barebox 2022.06.0-00530-g31d9ff81c63c-dirty #44 Fri Oct 21 12:13:36 MSK 2022
> >> ...
> >> ERROR: imx-watchdog 73f98000.watchdog@73f98000.of: probe failed: No
> >> such file or directory
> >> ...
> >>
> >> Thought it was the result of a patch "watchdog: imxwd: get and enable clock",
> >> and added the line:
> >> clks[IMX5_CLK_DUMMY] = clk_fixed("dummy", 0);
> >> into the beginning of the mx5_clocks_common_init() function, but this
> >> does not help...
> > 
> > If the missing clock is the problem then you also have to tell the clock
> > core that the dummy clock shall be provided to the watchdog, something
> > like:
> > 
> > 	clkdev_add_physbase(clks[IMX5_CLK_DUMMY], MX53_WDOG1_BASE_ADDR, NULL);
> > 	clkdev_add_physbase(clks[IMX5_CLK_DUMMY], MX53_WDOG2_BASE_ADDR, NULL);
> > 
> > similar is missing for i.MX51 as well
> 
> But the imx50.dtsi device tree already has
> 
> 	clocks = <&clks IMX5_CLK_DUMMY>;

I thought the device tree clocks are unused, because otherwise we could
remove the clkdev_add_physbase() calls. Ok, these are there for the
remaining non device tree boards.

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |