[U-Boot] [PATCH 04/10] arm: socfpga: arria10: Added clock manager and pin mux compat macro

Marek Vasut marex at denx.de
Mon Dec 19 11:04:52 CET 2016


On 12/19/2016 09:54 AM, Chee, Tien Fong wrote:
> On Isn, 2016-12-19 at 09:43 +0100, Marek Vasut wrote:
>> On 12/19/2016 09:40 AM, Chee, Tien Fong wrote:
>>>
>>> On Isn, 2016-12-19 at 08:55 +0100, Marek Vasut wrote:
>>>>
>>>> On 12/19/2016 05:10 AM, Chee, Tien Fong wrote:
>>>>>
>>>>>
>>>>> On Rab, 2016-12-07 at 14:54 +0100, Marek Vasut wrote:
>>>>>>
>>>>>>
>>>>>> On 12/07/2016 11:48 AM, Chee, Tien Fong wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On Sel, 2016-12-06 at 13:49 +0100, Marek Vasut wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On 12/06/2016 08:52 AM, Chee Tien Fong wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> From: Tien Fong Chee <tien.fong.chee at intel.com>
>>>>>>>>>
>>>>>>>>> These compat macros would be used by clock manager and
>>>>>>>>> pin
>>>>>>>>> mux
>>>>>>>>> drivers
>>>>>>>>> to look the required HW info from DTS for hardware
>>>>>>>>> initialization.
>>>>>>>>>
>>>>>>>>> Signed-off-by: Tien Fong Chee <tien.fong.chee at intel.com
>>>>>>>>>>
>>>>>>>>> Cc: Marek Vasut <marex at denx.de>
>>>>>>>>> Cc: Dinh Nguyen <dinguyen at kernel.org>
>>>>>>>>> Cc: Chin Liang See <chin.liang.see at intel.com>
>>>>>>>>> Cc: Tien Fong <skywindctf at gmail.com>
>>>>>>>>> ---
>>>>>>>>>  include/fdtdec.h |    8 ++++++++
>>>>>>>>>  lib/fdtdec.c     |    2 ++
>>>>>>>>>  2 files changed, 10 insertions(+), 0 deletions(-)
>>>>>>>>>
>>>>>>>>> diff --git a/include/fdtdec.h b/include/fdtdec.h
>>>>>>>>> index 27887c8..68cb199 100644
>>>>>>>>> --- a/include/fdtdec.h
>>>>>>>>> +++ b/include/fdtdec.h
>>>>>>>>> @@ -155,6 +155,14 @@ enum fdt_compat_id {
>>>>>>>>>  	COMPAT_INTEL_BAYTRAIL_FSP_MDP,	/* Intel
>>>>>>>>> FSP
>>>>>>>>> memory-
>>>>>>>>> down params */
>>>>>>>>>  	COMPAT_INTEL_IVYBRIDGE_FSP,	/* Intel
>>>>>>>>> Ivy
>>>>>>>>> Bridge
>>>>>>>>> FSP
>>>>>>>>> */
>>>>>>>>>  	COMPAT_SUNXI_NAND,		/* SUNXI
>>>>>>>>> NAND
>>>>>>>>> controller
>>>>>>>>> */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_CLK,	/* SoCFPGA
>>>>>>>>> Clock
>>>>>>>>> initialization */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_PINCTRL_SINGLE,	/
>>>>>>>>> *
>>>>>>>>> pinctrl-
>>>>>>>>> single */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_H2F_BRG,		
>>>>>>>>> /*
>>>>>>>>> Arria10
>>>>>>>>> hps2fpga bridge */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_LWH2F_BRG,	/*
>>>>>>>>> Arria10
>>>>>>>>> lwhps2fpga bridge */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_F2H_BRG,		
>>>>>>>>> /*
>>>>>>>>> Arria10
>>>>>>>>> fpga2hps bridge */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_F2SDR0,		/
>>>>>>>>> *
>>>>>>>>> Arria10
>>>>>>>>> fpga2SDRAM0 bridge */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_F2SDR1,		/
>>>>>>>>> *
>>>>>>>>> Arria10
>>>>>>>>> fpga2SDRAM1 bridge */
>>>>>>>>> +	COMPAT_ALTERA_SOCFPGA_F2SDR2,		/
>>>>>>>>> *
>>>>>>>>> Arria10
>>>>>>>>> fpga2SDRAM2 bridge */
>>>>>>>> Is all of this needed ? You're only adding two entries in
>>>>>>>> the
>>>>>>>> FDTDEC
>>>>>>>> below.
>>>>>>>>
>>>>>>> This is to avoid compilation error, we have some functions
>>>>>>> ported
>>>>>>> from
>>>>>>> our internal branch, which using above COMPAT macro. Soon,
>>>>>>> in
>>>>>>> upcoming
>>>>>>> patches, we will need those functions and more entries will
>>>>>>> be
>>>>>>> added
>>>>>>> into FDTDEC below.
>>>>>> You can add the compat strings when you really need them.
>>>>>> Still,
>>>>>> with
>>>>>> DM, you shouldn't even need them AFAIK.
>>>>>>
>>>>> We have some drivers in these series of patches contain some
>>>>> COMPAT
>>>>> strings, without these compact strings, the compilation would
>>>>> fail
>>>>> due
>>>>> to error compact string is not defined.
>>>> All of them ? Mind you, with DM you should not need to add those
>>>> at
>>>> all.
>>>>
>>>>>
>>>>>
>>>>> I think having compact string
>>>>> would giving us flexbility to put our nodes where we want
>>>>> without
>>>>> worrying to break our existing codes?
>>>> Which existing codes ?
>>>>
>>> let say we have version A, we found the node based on COMPAT
>>> STRING.
>>> One day, we move the node to somewhere else in version B, we still
>>> can
>>> find them based on COMPAT STRING without checking the node path.
>>> What
>>> do you think?
>> I do not understand what you're trying to tell me here. The DM core
>> will
>> walk the DT and bind drivers according to compat strings, so if
>> you ever move a node, it will still be bound.
>>
>> Also, you didn't answer my question -- which existing codes do you
>> refer
>> to. Or is this some hypothetical concern ?
>>
> existing codes is hypothetical example, when someone move the node
> somewhere, the DM still can find the node based on COMPAT STRING.

Correct. If you only have hypothetical concerns, we'll deal with those
when they become real.

> Without COMPAT STRING, we need to change the driver code for the node
> path right?

Uh no, the driver must never encode a fixed DT path. It should match on
the compat string and that's it. And the compat string is part of the
driver, not the FDT decoder. AFAIK, the strings in FDT decoder are just
temporary helpers during the DT conversion of drivers.

-- 
Best regards,
Marek Vasut


More information about the U-Boot mailing list