[PATCH v2 11/12] [POWERPC] Promess Motion-PRO DTS

Marian Balakowicz m8 at semihalf.com
Wed Nov 7 07:02:10 EST 2007


David Gibson wrote:
> On Sun, Nov 04, 2007 at 12:53:17AM +0100, Marian Balakowicz wrote:
>> Add device tree source file for Motion-PRO board.
> [snip]
>> +		motionpro-statusled at 660 {	// Motion-PRO status LED
>> +			compatible = "promess,motionpro-statusled";
>> +			cell-index = <6>;
>> +			reg = <660 10>;
>> +			interrupts = <1 f 0>;
>> +			interrupt-parent = <&mpc5200_pic>;
>> +			blink-delay = <64>; // 100 msec
>> +		};
>> +
>> +		motionpro-readyled at 670 {	// Motion-PRO ready LED
>> +			compatible = "promess,motionpro-readyled";
>> +			cell-index = <7>;
> 
> These cell-index values for the LEDs look very strange.  How are they
> used?
[snip]
>> +
>> +		mscan at 980 {
>> +			compatible = "mpc5200b-mscan","mpc5200-mscan";
>> +			cell-index = <1>;
> 
> As for serial and gpt, is cell-index really suitable here?

Ok, removed those unused "cell-index" properties.

> [snip]
>> +		spi at f00 {
>> +			device_type = "spi";
>> +			compatible = "mpc5200b-spi","mpc5200-spi";
>> +			reg = <f00 20>;
>> +			interrupts = <2 d 0 2 e 0>;
>> +			interrupt-parent = <&mpc5200_pic>;
>> +		};
> [snip]
>> +		// PSC2 in spi master mode 
>> +		spi at 2200 {		// PSC2
>> +			device_type = "spi";
>> +			compatible = "mpc5200b-psc-spi","mpc5200-psc-spi";
>> +			cell-index = <1>;
> 
> cell-index present for one spi, but not the other makes be even more
> suspicious about it's applicability here...

These are two different SPIs, both are part of the SoC but first one
is dedicated SPI interface while the second one is PSC port in a SPI
mode. Driver for the second one (mpc52xx_psc_spi.c) is actually using
cell-index to get the PSC port number it's controlling.

> [snip]
>> +	lpb {
>> +		model = "fsl,lpb";
>> +		compatible = "fsl,lpb";
> 
> Is lpb another one of these chipselect/offset configurable bus bridge
> things?  If so, you should use a 2-cell addressing convention for the
> subnodes like fsl localbus and 4xx EBC.

Yes, it is, so I have switched to 2-cell addressing. Please see the v3
respin of the patches.


>> +		// 8-bit custom Anybus Module on LocalPlus Bus CS3
>> +		anybus50020000 {
> 
> Missing '@'.

Fixed.

Thanks,
m.



More information about the Linuxppc-dev mailing list