Clocks on Barebox

Clocks on Barebox

[andreas.willig]

Hi all
i just updated from 2014 to latest Barebox and hit the following problem:
of_clk_init from clk.c:451 does not find any entries. matches is set to __clk_of_table start where compatible=0 and data=0,
but from my thoughts the dt is ok on this point:
{
...
clocks {
#address-cells = <0x1>;
#size-cells = <0x0>;
ckil {
compatible = "fsl,imx-ckil", "fixed-clock";
#clock-cells = <0x0>;
clock-frequency = <0x8000>;
};
ckih1 {
compatible = "fsl,imx-ckih1", "fixed-clock";
#clock-cells = <0x0>;
clock-frequency = <0x0>;
};
osc {
compatible = "fsl,imx-osc", "fixed-clock";
#clock-cells = <0x0>;
clock-frequency = <0x16e3600>;
};
clock@0 {
compatible = "fixed-clock";
reg = <0x0>;
#clock-cells = <0x0>;
clock-frequency = <0x16e3600>;
};
}
aips-bus@02000000 {
...
ccm@020c4000 {
compatible = "fsl,imx6q-ccm", "fsl,imx6dl-ccm";
reg = <0x20c4000 0x4000>;
interrupts = <0x0 0x57 0x4 0x0 0x58 0x4>;
#clock-cells = <0x1>;
linux,phandle = <0x3>;
phandle = <0x3>;
};
...
};
...
};
Is there a pretty howto added? Am I missing something?
from arch/arm/mach-imx/clk-imx6.c are removed:
static int imx6_ccm_probe()
{
- unsigned long ckil_rate = 32768;
- unsigned long ckih_rate = 0;
- unsigned long osc_rate = 24000000;
- clks[dummy] = clk_fixed("dummy", 0);
- clks[ckil] = clk_fixed("ckil", ckil_rate);
- clks[ckih] = clk_fixed("ckih", ckih_rate);
- clks[osc] = clk_fixed("osc", osc_rate);
}
and on gpt_probe the clock tree refers to a frequency of 0 while seeking for osc as clk parent which cannot be found and leads to a div by zero and therefore in a panic in further context. This prevents bb from starting (of course).

System is Freescale i.MX6 Solo on our on DIMM module.


Mit freundlichen Grüßen aus Berg/ With kind regards

i.A. Andreas Willig
Dipl.-Ing. (FH)
Entwicklung Elektronik

RAFI GmbH & Co. KG
Postfach 2060
88276 Ravensburg
Tel.: +49 (0) 751 89-6142

mailto:andreas.willig@rafi.de, http://www.rafi.de

Kommanditgesellschaft: Sitz Berg, Amtsgericht Ulm, HRA 550059
Komplementär: RAFI Beteiligungs-GmbH, Sitz Berg, Amtsgericht Ulm HRB 550074
Geschäftsführer: Albert Wasmeier, Dipl.-Kfm. Gerhard Schenk
USt-Ident.-Nr.: DE 146 392 319, Steuer-Nr.: 77079/00291

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[Sascha Hauer]

Hi Andreas,

On Mon, Jun 01, 2015 at 10:12:55AM +0200, andreas.willig@rafi.de wrote:
> Hi all
> i just updated from 2014 to latest Barebox and hit the following problem:
> of_clk_init from clk.c:451 does not find any entries. matches is set to __clk_of_table start where compatible=0 and data=0,
> but from my thoughts the dt is ok on this point:

I have no idea currently. You are doing nothing obviously wrong.

> from arch/arm/mach-imx/clk-imx6.c are removed:
> static int imx6_ccm_probe()
> {
> - unsigned long ckil_rate = 32768;
> - unsigned long ckih_rate = 0;
> - unsigned long osc_rate = 24000000;
> - clks[dummy] = clk_fixed("dummy", 0);
> - clks[ckil] = clk_fixed("ckil", ckil_rate);
> - clks[ckih] = clk_fixed("ckih", ckih_rate);
> - clks[osc] = clk_fixed("osc", osc_rate);
> }

These are removed because they come from the OF clock providers now. For
this you need COMMON_CLK_OF_PROVIDER enabled, but this should be
selected by the imx6 support automatically though. Let me test this on
some hardware here. What barebox version do you use? master, next, or
the latest release?

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[andreas.willig]

Hi Sascha,
I'm currently on the next branch, but i did not see changes at this point since master or stable release. Like issued earlier the clock table enumeration is empty. Is there a specific compatible attribute required? 
 

Mit freundlichen Grüßen aus Berg/ With kind regards

i.A. Andreas Willig
Dipl.-Ing. (FH)
Entwicklung Elektronik

RAFI GmbH & Co. KG
Postfach 2060
88276 Ravensburg
Tel.: +49 (0) 751 89-6142

mailto:andreas.willig@rafi.de, http://www.rafi.de

Kommanditgesellschaft: Sitz Berg, Amtsgericht Ulm, HRA 550059
Komplementär: RAFI Beteiligungs-GmbH, Sitz Berg, Amtsgericht Ulm HRB 550074
Geschäftsführer: Albert Wasmeier, Dipl.-Kfm. Gerhard Schenk
USt-Ident.-Nr.: DE 146 392 319, Steuer-Nr.: 77079/00291 

 
-----Sascha Hauer <s.hauer@pengutronix.de> schrieb: -----

 =======================
 An: andreas.willig@rafi.de
 Von: Sascha Hauer <s.hauer@pengutronix.de>
 Datum: 01.06.2015 13:41 
 Kopie: barebox@lists.infradead.org
 Betreff: Re: Clocks on Barebox
 =======================
   Hi Andreas,

On Mon, Jun 01, 2015 at 10:12:55AM +0200, andreas.willig@rafi.de wrote:
> Hi all
> i just updated from 2014 to latest Barebox and hit the following problem:
> of_clk_init from clk.c:451 does not find any entries. matches is set to __clk_of_table start where compatible=0 and data=0,
> but from my thoughts the dt is ok on this point:

I have no idea currently. You are doing nothing obviously wrong.

> from arch/arm/mach-imx/clk-imx6.c are removed:
> static int imx6_ccm_probe()
> {
> - unsigned long ckil_rate = 32768;
> - unsigned long ckih_rate = 0;
> - unsigned long osc_rate = 24000000;
> - clks[dummy] = clk_fixed("dummy", 0);
> - clks[ckil] = clk_fixed("ckil", ckil_rate);
> - clks[ckih] = clk_fixed("ckih", ckih_rate);
> - clks[osc] = clk_fixed("osc", osc_rate);
> }

These are removed because they come from the OF clock providers now. For
this you need COMMON_CLK_OF_PROVIDER enabled, but this should be
selected by the imx6 support automatically though. Let me test this on
some hardware here. What barebox version do you use? master, next, or
the latest release?

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
    

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[Sascha Hauer]

On Tue, Jun 02, 2015 at 05:42:49AM +0200, andreas.willig@rafi.de wrote:
> Hi Sascha,
> I'm currently on the next branch, but i did not see changes at this
> point since master or stable release. Like issued earlier the clock
> table enumeration is empty. Is there a specific compatible attribute
> required?

I just tested barebox -next on the SabreLite board with the
imx_v7_defconfig. It works as expected. Do you have any patches applied
or use a different config file? I have no idea what goes wrong here.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[andreas.willig]

Hi Sascha,

any news on the clocking Problem, did you receive my patch file? I did a workaround and readded the hardcoded osc settings, but i think this is not the way on long term.

Congrats for the new barebox.org website, pretty nice :)



Mit freundlichen Grüßen aus Berg/ With kind regards

i.A. Andreas Willig
Dipl.-Ing. (FH)
Entwicklung Elektronik

RAFI GmbH & Co. KG
Postfach 2060
88276 Ravensburg
Tel.: +49 (0) 751 89-6142

mailto:andreas.willig@rafi.de, http://www.rafi.de

Kommanditgesellschaft: Sitz Berg, Amtsgericht Ulm, HRA 550059
Komplementär: RAFI Beteiligungs-GmbH, Sitz Berg, Amtsgericht Ulm HRB 550074
Geschäftsführer: Albert Wasmeier, Dipl.-Kfm. Gerhard Schenk
USt-Ident.-Nr.: DE 146 392 319, Steuer-Nr.: 77079/00291


-----Sascha Hauer <s.hauer@pengutronix.de> schrieb: ----- 
An: andreas.willig@rafi.de
Von: Sascha Hauer <s.hauer@pengutronix.de>
Datum: 03.06.2015 10:19
Kopie: barebox@lists.infradead.org
Betreff: Re: Clocks on Barebox


On Tue, Jun 02, 2015 at 05:42:49AM +0200, andreas.willig@rafi.de wrote:
> Hi Sascha,
> I'm currently on the next branch, but i did not see changes at this
> point since master or stable release. Like issued earlier the clock
> table enumeration is empty. Is there a specific compatible attribute
> required?

I just tested barebox -next on the SabreLite board with the
imx_v7_defconfig. It works as expected. Do you have any patches applied
or use a different config file? I have no idea what goes wrong here.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[Sascha Hauer]

On Mon, Jun 01, 2015 at 10:12:55AM +0200, andreas.willig@rafi.de wrote:
> Hi all
> i just updated from 2014 to latest Barebox and hit the following problem:
> of_clk_init from clk.c:451 does not find any entries. matches is set to __clk_of_table start where compatible=0 and data=0,

Do you mean the clk_of_table is empty? Could you post an excerpt from
you barebox.map file containing the __clk_of_table? Here it looks like:

                0x0000000000092178                __clk_of_table_start =
.
 *(.__clk_of_table_*)
 .__clk_of_table_fixed_clk
                0x0000000000092178        0x8 drivers/built-in.o
                0x0000000000092178
__clk_of_table_fixed_clk
 .__clk_of_table_fixed_factor_clk
                0x0000000000092180        0x8 drivers/built-in.o
                0x0000000000092180
__clk_of_table_fixed_factor_clk
 .__clk_of_table_end
                0x0000000000092188        0x8 drivers/built-in.o
                0x0000000000092188
__clk_of_table_sentinel
                0x0000000000092190                __clk_of_table_end = .

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[andreas.willig]

Hi Sascha,

here my oftables section from barebox.map:

.oftables       0x278f7bc0       0x18
                0x278f7bc0                . = ALIGN (0x8)
                0x278f7bc0                __clk_of_table_start = .
 *(.__clk_of_table_*)
 .__clk_of_table_end
                0x278f7bc0        0x8 drivers/built-in.o
                0x278f7bc0                __clk_of_table_sentinel
 .__clk_of_table_fixed_clk
                0x278f7bc8        0x8 drivers/built-in.o
                0x278f7bc8                __clk_of_table_fixed_clk
 .__clk_of_table_fixed_factor_clk
                0x278f7bd0        0x8 drivers/built-in.o
                0x278f7bd0                __clk_of_table_fixed_factor_clk
                0x278f7bd8                __clk_of_table_end = .

from imx6qdl.dtsi:
...
    clocks {
        #address-cells = <1>;
        #size-cells = <0>;
        ckil {
            compatible = "fsl,imx-ckil", "fixed-clock";
            #clock-cells = <0>;
            clock-frequency = <32768>;
        };
        ckih1 {
            compatible = "fsl,imx-ckih1", "fixed-clock";
            #clock-cells = <0>;
            clock-frequency = <0>;
        };
        osc {
            compatible = "fsl,imx-osc", "fixed-clock";
            #clock-cells = <0>;
            clock-frequency = <24000000>;
        };
    };
...

I' m not quite sure how to read the barebox.map at this Point. Do I have clocks in Image or not? The parsercode just breaks out without finding any enumerated items.... 


Mit freundlichen Grüßen aus Berg/ With kind regards

i.A. Andreas Willig
Dipl.-Ing. (FH)
Entwicklung Elektronik

RAFI GmbH & Co. KG
Postfach 2060
88276 Ravensburg
Tel.: +49 (0) 751 89-6142

mailto:andreas.willig@rafi.de, http://www.rafi.de

Kommanditgesellschaft: Sitz Berg, Amtsgericht Ulm, HRA 550059
Komplementär: RAFI Beteiligungs-GmbH, Sitz Berg, Amtsgericht Ulm HRB 550074
Geschäftsführer: Albert Wasmeier, Dipl.-Kfm. Gerhard Schenk
USt-Ident.-Nr.: DE 146 392 319, Steuer-Nr.: 77079/00291


-----Sascha Hauer <s.hauer@pengutronix.de> schrieb: ----- 
An: andreas.willig@rafi.de
Von: Sascha Hauer <s.hauer@pengutronix.de>
Datum: 09.06.2015 12:18
Kopie: barebox@lists.infradead.org
Betreff: Re: Clocks on Barebox


On Mon, Jun 01, 2015 at 10:12:55AM +0200, andreas.willig@rafi.de wrote:
> Hi all
> i just updated from 2014 to latest Barebox and hit the following problem:
> of_clk_init from clk.c:451 does not find any entries. matches is set to __clk_of_table start where compatible=0 and data=0,

Do you mean the clk_of_table is empty? Could you post an excerpt from
you barebox.map file containing the __clk_of_table? Here it looks like:

                0x0000000000092178                __clk_of_table_start =
.
 *(.__clk_of_table_*)
 .__clk_of_table_fixed_clk
                0x0000000000092178        0x8 drivers/built-in.o
                0x0000000000092178
__clk_of_table_fixed_clk
 .__clk_of_table_fixed_factor_clk
                0x0000000000092180        0x8 drivers/built-in.o
                0x0000000000092180
__clk_of_table_fixed_factor_clk
 .__clk_of_table_end
                0x0000000000092188        0x8 drivers/built-in.o
                0x0000000000092188
__clk_of_table_sentinel
                0x0000000000092190                __clk_of_table_end = .

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[Sascha Hauer]

On Tue, Jun 09, 2015 at 01:28:17PM +0200, andreas.willig@rafi.de wrote:
> Hi Sascha,
> 
> here my oftables section from barebox.map:
> 
> .oftables       0x278f7bc0       0x18
>                 0x278f7bc0                . = ALIGN (0x8)
>                 0x278f7bc0                __clk_of_table_start = .
>  *(.__clk_of_table_*)
>  .__clk_of_table_end
>                 0x278f7bc0        0x8 drivers/built-in.o
>                 0x278f7bc0                __clk_of_table_sentinel
>  .__clk_of_table_fixed_clk
>                 0x278f7bc8        0x8 drivers/built-in.o
>                 0x278f7bc8                __clk_of_table_fixed_clk
>  .__clk_of_table_fixed_factor_clk
>                 0x278f7bd0        0x8 drivers/built-in.o
>                 0x278f7bd0                __clk_of_table_fixed_factor_clk
>                 0x278f7bd8                __clk_of_table_end = .

And here's the problem. The sentinel entry containing only NULL marks
the end of the array. It should be the last entry, but here it's the
first entry which effectively makes the table empty. It turned out it
was pure luck that the sentinel entry ended up as last entry normally.

I've just sent out a patch fixing this, please test.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[andreas.willig]

Hi Sascha,

you make my day :) this is it, everything is fine, i removed my Workaround with hardinit osc to tree, and everything works smooth and beautiful :)

Problem solved, looking Forward, seeing this patch on branch next :)

Mit freundlichen Grüßen aus Berg/ With kind regards

i.A. Andreas Willig
Dipl.-Ing. (FH)
Entwicklung Elektronik

RAFI GmbH & Co. KG
Postfach 2060
88276 Ravensburg
Tel.: +49 (0) 751 89-6142

mailto:andreas.willig@rafi.de, http://www.rafi.de

Kommanditgesellschaft: Sitz Berg, Amtsgericht Ulm, HRA 550059
Komplementär: RAFI Beteiligungs-GmbH, Sitz Berg, Amtsgericht Ulm HRB 550074
Geschäftsführer: Albert Wasmeier, Dipl.-Kfm. Gerhard Schenk
USt-Ident.-Nr.: DE 146 392 319, Steuer-Nr.: 77079/00291


-----Sascha Hauer <s.hauer@pengutronix.de> schrieb: ----- 
An: andreas.willig@rafi.de
Von: Sascha Hauer <s.hauer@pengutronix.de>
Datum: 10.06.2015 06:59
Kopie: barebox@lists.infradead.org
Betreff: Re: Clocks on Barebox


On Tue, Jun 09, 2015 at 01:28:17PM +0200, andreas.willig@rafi.de wrote:
> Hi Sascha,
> 
> here my oftables section from barebox.map:
> 
> .oftables       0x278f7bc0       0x18
>                 0x278f7bc0                . = ALIGN (0x8)
>                 0x278f7bc0                __clk_of_table_start = .
>  *(.__clk_of_table_*)
>  .__clk_of_table_end
>                 0x278f7bc0        0x8 drivers/built-in.o
>                 0x278f7bc0                __clk_of_table_sentinel
>  .__clk_of_table_fixed_clk
>                 0x278f7bc8        0x8 drivers/built-in.o
>                 0x278f7bc8                __clk_of_table_fixed_clk
>  .__clk_of_table_fixed_factor_clk
>                 0x278f7bd0        0x8 drivers/built-in.o
>                 0x278f7bd0                __clk_of_table_fixed_factor_clk
>                 0x278f7bd8                __clk_of_table_end = .

And here's the problem. The sentinel entry containing only NULL marks
the end of the array. It should be the last entry, but here it's the
first entry which effectively makes the table empty. It turned out it
was pure luck that the sentinel entry ended up as last entry normally.

I've just sent out a patch fixing this, please test.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: Clocks on Barebox

[Sascha Hauer]

On Wed, Jun 10, 2015 at 07:10:25AM +0200, andreas.willig@rafi.de wrote:
> Hi Sascha,
> 
> you make my day :) this is it, everything is fine, i removed my Workaround with hardinit osc to tree, and everything works smooth and beautiful :)

Ok, I take this as a

Tested-by: Andreas Willig <andreas.willig@rafi.de>

:)

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox