On Apr 4, 2013, at 2:03 AM, Leekha Shaveta-B20052 wrote:
>
>
> -----Original Message-----
> From: Wood Scott-B07421
> Sent: Wednesday, April 03, 2013 10:10 PM
> To: Leekha Shaveta-B20052
> Cc: Wood Scott-B07421; linuxppc-dev@lists.ozlabs.org; Zhao Chenhui-B35336;
> Lian Minghuan-B31939; Garg Vakul-B16394; Tang Yuantian-B29983; Fleming
> Andy-AFLEMING; Mehresh Ramneek-B31383; Sethi Varun-B16395
> Subject: Re: [PATCH 2/5] powerpc/fsl-booke: Add initial silicon device tree
> files for B4860 and B4420
>
> On 04/03/2013 01:42:14 AM, Leekha Shaveta-B20052 wrote:
>>
>>
>> -----Original Message-----
>> From: Wood Scott-B07421
>> Sent: Wednesday, April 03, 2013 12:49 AM
>> To: Leekha Shaveta-B20052
>> Cc: linuxppc-dev@lists.ozlabs.org; Zhao Chenhui-B35336; Lian
>> Minghuan-B31939; Leekha Shaveta-B20052; Garg Vakul-B16394; Tang
>> Yuantian-B29983; Fleming Andy-AFLEMING; Mehresh Ramneek-B31383; Sethi
>> Varun-B16395
>> Subject: Re: [PATCH 2/5] powerpc/fsl-booke: Add initial silicon device
>> tree files for B4860 and B4420
>>
>> On 04/02/2013 02:16:05 AM, Shaveta Leekha wrote:
>>> +/ {
>>> + compatible = "fsl,B4860";
>>> +
>>> + cpus {
>>> + cpu1: PowerPC,e6500@1 {
>>> + device_type = "cpu";
>>> + reg = <2 3>;
>>> + next-level-cache = <&L2>;
>>> + };
>>> + cpu2: PowerPC,e6500@2 {
>>> + device_type = "cpu";
>>> + reg = <4 5>;
>>> + next-level-cache = <&L2>;
>>> + };
>>> + cpu3: PowerPC,e6500@3 {
>>> + device_type = "cpu";
>>> + reg = <6 7>;
>>> + next-level-cache = <&L2>;
>>> + };
>>
>> The unit addresses need to match "reg".
>> [SL] You mean "@1" should match to "reg = <2 3>" ?
>
> Yes, it should be "@2" for that node.
>
>> As each e6500 core in B4860 is dual- threaded, reg property here
>> represents the thread's identifier in that PA core.
>>
>> So convention used in T4 and B4 is: core 0 having threads 0 and 1,
>> Core 1 having <2 3> and
>> so on....
>
> The convention used in device trees is that the unit address matches the reg.
>
> -Scott
> [SL] Ok, I can change that. Will make unit address as @2 for <2 3>, @4 for <4
> 5> and so on....
>
> Kumar, please respond here, as I have followed the convention used in T4
> device tree files for
> Dual-threaded cores.
Scott is correct, we probably need to fix the T4 dts.
- k
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev
