Help with device tree binding for SMC serial

Help with device tree binding for SMC serial

[Rune Torgersen]

Hi

We're trying to get a SMC serial port on a8280 to work.
I cannot find any ecxamples on the binding, so we've tried to make one.

is this anywhere close to workable?

serial@11a80 {
	device_type =3D "serial";
	compatible =3D "fsl,mpc8280-smc-uart",
				 "fsl,cpm2-smc-uart";
	reg =3D <11a80 10 87fc 2>;
	interrupts =3D <4 8>;
	interrupt-parent =3D <&PIC>;
	fsl,cpm-brg =3D <7>;
	fsl,cpm-command =3D <01000000>;
};

Re: Help with device tree binding for SMC serial

[Scott Wood]

Rune Torgersen wrote:
> Hi
> 
> We're trying to get a SMC serial port on a8280 to work.
> I cannot find any ecxamples on the binding, so we've tried to make one.
> 
> is this anywhere close to workable?
> 
> serial@11a80 {
> 	device_type = "serial";
> 	compatible = "fsl,mpc8280-smc-uart",
> 				 "fsl,cpm2-smc-uart";
> 	reg = <11a80 10 87fc 2>;

The current binding has the actual parameter ram block as the second reg 
property, not the two-byte pointer.  If your firmware doesn't initialize 
the pointer, you'll have to do it in platform code.

> 	interrupts = <4 8>;
> 	interrupt-parent = <&PIC>;
> 	fsl,cpm-brg = <7>;
> 	fsl,cpm-command = <01000000>;
> };

cpm-command should be <1d000000> for SMC1.

Otherwise, it looks OK.

-Scott

RE: Help with device tree binding for SMC serial

[Rune Torgersen]

> From: Scott Wood=20
> Sent: Wednesday, January 09, 2008 1:46 PM
> > We're trying to get a SMC serial port on a8280 to work.
> > I cannot find any ecxamples on the binding, so we've tried=20
> to make one.
> >=20
> > is this anywhere close to workable?
> >=20
> > serial@11a80 {
> > 	device_type =3D "serial";
> > 	compatible =3D "fsl,mpc8280-smc-uart",
> > 				 "fsl,cpm2-smc-uart";
> > 	reg =3D <11a80 10 87fc 2>;
>=20
> The current binding has the actual parameter ram block as the=20
> second reg=20
> property, not the two-byte pointer.  If your firmware doesn't=20
> initialize=20
> the pointer, you'll have to do it in platform code.
>=20
> > 	interrupts =3D <4 8>;
> > 	interrupt-parent =3D <&PIC>;
> > 	fsl,cpm-brg =3D <7>;
> > 	fsl,cpm-command =3D <01000000>;
> > };
>=20
> cpm-command should be <1d000000> for SMC1.
>=20
> Otherwise, it looks OK.

Ok we're now using
serial@11a80 {
	device_type =3D "serial";
	compatible =3D "fsl,mpc8280-smc-uart",
				 "fsl,cpm2-smc-uart";
	reg =3D <11a80 10 0 40>;            // <base_address length
parameter_ram_address length>
	interrupts =3D <4 8>;                 // Interrupt from table 4.3
of mpc8280rm, interrupt is level or edge
	interrupt-parent =3D <&PIC>;
	fsl,cpm-brg =3D <7>;
	fsl,cpm-command =3D <1d000000>;       // Page and Sub-block code
of the CPCR
};

Right now we're trying to just get a kernel to give us some serial
output, so we can continue the porting job.
We're unsig a cuImage (using the pq2fads code right now).

Now our problem is that the serial port is spitting out 0x0a's as fast
as it can.
Seems to be repeadtin the linefeed at the end of "Uncompressing Kernel
Image ... OK"

Re: Help with device tree binding for SMC serial

[Scott Wood]

Rune Torgersen wrote:
> Ok we're now using
> serial@11a80 {
> 	device_type = "serial";
> 	compatible = "fsl,mpc8280-smc-uart",
> 				 "fsl,cpm2-smc-uart";
> 	reg = <11a80 10 0 40>;            // <base_address length
> parameter_ram_address length>
> 	interrupts = <4 8>;                 // Interrupt from table 4.3
> of mpc8280rm, interrupt is level or edge
> 	interrupt-parent = <&PIC>;
> 	fsl,cpm-brg = <7>;
> 	fsl,cpm-command = <1d000000>;       // Page and Sub-block code
> of the CPCR
> };

Looks good.

> Right now we're trying to just get a kernel to give us some serial
> output, so we can continue the porting job.
> We're unsig a cuImage (using the pq2fads code right now).
> 
> Now our problem is that the serial port is spitting out 0x0a's as fast
> as it can.

Don't forget to exclude the SMC parameter RAM from the muram data area 
in /soc/cpm/muram/data/reg.  If you have an older device tree binding 
that has no /soc/cpm/muram node, but instead has two resources in 
/soc/cpm/reg, you need to move to head-of-tree to get this to work.

-Scott

RE: Help with device tree binding for SMC serial

[Rune Torgersen]

> From: Scott Wood
> Don't forget to exclude the SMC parameter RAM from the muram=20
> data area=20
> in /soc/cpm/muram/data/reg.  If you have an older device tree binding=20
> that has no /soc/cpm/muram node, but instead has two resources in=20
> /soc/cpm/reg, you need to move to head-of-tree to get this to work.

Did that, now we get e checkstop resert. Time to hook up the
BDI-2000....
Is there a way to have the bootwrapper use the u-boot serial for a while
(like a early srial port) for debugging?

Re: Help with device tree binding for SMC serial

[Scott Wood]

Rune Torgersen wrote:
>> From: Scott Wood
>> Don't forget to exclude the SMC parameter RAM from the muram 
>> data area 
>> in /soc/cpm/muram/data/reg.  If you have an older device tree binding 
>> that has no /soc/cpm/muram node, but instead has two resources in 
>> /soc/cpm/reg, you need to move to head-of-tree to get this to work.
> 
> Did that, now we get e checkstop resert. Time to hook up the
> BDI-2000....
> Is there a way to have the bootwrapper use the u-boot serial for a while
> (like a early srial port) for debugging?

If you mean calling into u-boot for a console, no, there's no code for that.

What do your cpm and muram nodes look like?

-Scott

RE: Help with device tree binding for SMC serial

[Rune Torgersen]

> From: Scott Wood [mailto:scottwood@freescale.com]=20
> Rune Torgersen wrote:
> >> From: Scott Wood
> >> Don't forget to exclude the SMC parameter RAM from the muram=20
> >> data area=20
> >> in /soc/cpm/muram/data/reg.  If you have an older device=20
> tree binding=20
> >> that has no /soc/cpm/muram node, but instead has two resources in=20
> >> /soc/cpm/reg, you need to move to head-of-tree to get this to work.
> >=20
> > Did that, now we get e checkstop resert. Time to hook up the
> > BDI-2000....
> > Is there a way to have the bootwrapper use the u-boot=20
> serial for a while
> > (like a early srial port) for debugging?
>=20
> If you mean calling into u-boot for a console, no, there's no=20
> code for that.
>=20
> What do your cpm and muram nodes look like?

cpm@119c0 {
	#address-cells =3D <1>;
	#size-cells =3D <1>;
	#interrupt-cells =3D <2>;
	compatible =3D "fsl,mpc8280-cpm", "fsl,cpm2";
	reg =3D <119c0 30>;
	ranges;

	muram@0 {
		#address-cells =3D <1>;
		#size-cells =3D <1>;
		ranges =3D <0 0 10000>;

		data@0 {
			compatible =3D "fsl,cpm-muram-data";
			reg =3D <100 1f00 9800 800>;
		};
	};

	brg@119f0 {
		compatible =3D "fsl,mpc8280-brg",
					 "fsl,cpm2-brg",
					 "fsl,cpm-brg";
		reg =3D <119f0 10 115f0 10>;
	};

	serial@11a80 {
		device_type =3D "serial";
		compatible =3D "fsl,mpc8280-smc-uart",
					 "fsl,cpm2-smc-uart";
		reg =3D <11a80 10 0 40>;            // <base_address
length parameter_ram_address length>
		interrupts =3D <4 8>;                 // Interrupt from
table 4.3 of mpc8280rm, interrupt is level or edge
		interrupt-parent =3D <&PIC>;
		fsl,cpm-brg =3D <7>;
		fsl,cpm-command =3D <1d000000>;       // Page and
Sub-block code of the CPCR
	};
};

and chosen node (not sure if needed)
	chosen {
		linux,stdout-path =3D "/soc/cpm/serial@11a80";
		bootargs =3D "console=3DttyS0,115200";
	};

Re: Help with device tree binding for SMC serial

[Scott Wood]

On Wed, Jan 09, 2008 at 05:07:23PM -0600, Rune Torgersen wrote:
> > From: Scott Wood [mailto:scottwood@freescale.com] 
> > Rune Torgersen wrote:
> > > Did that, now we get e checkstop resert. Time to hook up the
> > > BDI-2000.... Is there a way to have the bootwrapper use the u-boot
> > > serial for a while (like a early srial port) for debugging?
> > 
> > If you mean calling into u-boot for a console, no, there's no 
> > code for that.
> > 
> > What do your cpm and muram nodes look like?
[snip tree]

The tree looks OK.  The checkstop may be from erratum SIU18;
I had this issue on the ep8248e board.

Try clearing BCR[PLDP].

-Scott

RE: Help with device tree binding for SMC serial

[Rune Torgersen]

> From: Scott Wood
>=20
> The tree looks OK.  The checkstop may be from erratum SIU18;
> I had this issue on the ep8248e board.
>=20
> Try clearing BCR[PLDP].

Not sure what was wrong. We took a step back, and redid some stuff, and
now we have serial output from the boot-wrapper.
THe checkstop came from the wrapper grying to access the bcsr and doing
the chip select fixup. We don';t have a bcsr on our board, and the cs
layout is different.

Now our problem is that the kernel doesn't want to output anything to
hte serial port.

Re: Help with device tree binding for SMC serial

[Scott Wood]

Rune Torgersen wrote:
> Not sure what was wrong. We took a step back, and redid some stuff, and
> now we have serial output from the boot-wrapper.
> THe checkstop came from the wrapper grying to access the bcsr and doing
> the chip select fixup. We don';t have a bcsr on our board, and the cs
> layout is different.

There's no bcsr access in the upstream bootwrapper (the mention of bcsr 
in cuboot-pq2.c is just one reason why some boards require us to do the 
CS programming), and the CS programming is chip-level, not board-level 
(it just requires that the device tree have a correct localbus node for 
the board).

-Scott

RE: Help with device tree binding for SMC serial

[Rune Torgersen]

> From: Scott Wood [
> Rune Torgersen wrote:
> > Not sure what was wrong. We took a step back, and redid=20
> some stuff, and
> > now we have serial output from the boot-wrapper.
> > THe checkstop came from the wrapper grying to access the=20
> bcsr and doing
> > the chip select fixup. We don';t have a bcsr on our board,=20
> and the cs
> > layout is different.
>=20
> There's no bcsr access in the upstream bootwrapper (the=20
> mention of bcsr=20
> in cuboot-pq2.c is just one reason why some boards require us=20
> to do the=20
> CS programming), and the CS programming is chip-level, not=20
> board-level=20
> (it just requires that the device tree have a correct=20
> localbus node for=20
> the board).

Finally got it (sort-of) working.
Turned out that for some reason the console init is setting the baudrate
to 9600
the options string passed in to the console init fuunction is NULL.

Any idea oon how this should be passed in from u-boot?

RE: Help with device tree binding for SMC serial

[Rune Torgersen]

> From: Rune Torgersen
> Finally got it (sort-of) working.
> Turned out that for some reason the console init is setting=20
> the baudrate
> to 9600
> the options string passed in to the console init fuunction is NULL.
>=20
> Any idea oon how this should be passed in from u-boot?

Ok, needed a valid console=3D line on the command line. WHen we tried
that, we had a typo, so it was not recognized.
Our old 2.6.18 arch/ppc kernel didn't need a console parameter, it got
the baudrate from u-boot somehow.
Anyway of doing that here too?

Re: Help with device tree binding for SMC serial

[Scott Wood]

Rune Torgersen wrote:
> Ok, needed a valid console= line on the command line. WHen we tried
> that, we had a typo, so it was not recognized.
> Our old 2.6.18 arch/ppc kernel didn't need a console parameter, it got
> the baudrate from u-boot somehow.
> Anyway of doing that here too?

You could add something to the cuboot code to fill in current-speed 
based on the value in the bd_t.

-Scott

RE: Help with device tree binding for SMC serial

[Rune Torgersen]

> From: Scott Wood=20
>=20
> You could add something to the cuboot code to fill in current-speed=20
> based on the value in the bd_t.
>=20

Ahh.. That was what I'm missing.
Where in the devicetree is that supposed to be at?

Re: Help with device tree binding for SMC serial

[Scott Wood]

Rune Torgersen wrote:
>> From: Scott Wood 
>>
>> You could add something to the cuboot code to fill in current-speed 
>> based on the value in the bd_t.
>>
> 
> Ahh.. That was what I'm missing.
> Where in the devicetree is that supposed to be at?

In the serial node.

-Scott