[U-Boot] [PATCH v2 01/11] Blackfin: bf60x: new processor header files

[Sonic Zhang]

Hi Wolfgang,


On Tue, Feb 19, 2013 at 6:15 PM, Wolfgang Denk <wd at denx.de> wrote:
> Dear Sonic Zhang,
>
> In message <1361257266-3722-2-git-send-email-sonic.adi at gmail.com> you wrote:
>>
>> Add header files for blackfin new processor bf60x.
> ...
>>  arch/blackfin/include/asm/mach-bf609/BF609_def.h  | 3758 +++++++++++++++++++++
>
> This is a bit excessive, isn't it?  Do we really all thse many
> thousands lines of defines?
>
>
>> diff --git a/arch/blackfin/include/asm/mach-bf609/BF609_def.h b/arch/blackfin/include/asm/mach-bf609/BF609_def.h
>> new file mode 100644
>> index 0000000..39b740f
>> --- /dev/null
>> +++ b/arch/blackfin/include/asm/mach-bf609/BF609_def.h
>> @@ -0,0 +1,3758 @@
>> +/* DO NOT EDIT THIS FILE
>> + * Automatically generated by generate-def-headers.xsl
>> + * DO NOT EDIT THIS FILE
>> + */
>> +
>> +#ifndef __BFIN_DEF_ADSP_BF609_proc__
>> +#define __BFIN_DEF_ADSP_BF609_proc__
>> +
>> +#include "../mach-common/ADSP-EDN-core_def.h"
>> +
>> +#define CNT_CFG                     0xFFC00400         /* CNT0 Configuration Register */
>> +#define CNT_IMSK                    0xFFC00404         /* CNT0 Interrupt Mask Register */
>> +#define CNT_STAT                    0xFFC00408         /* CNT0 Status Register */
>> +#define CNT_CMD                     0xFFC0040C         /* CNT0 Command Register */
>> +#define CNT_DEBNCE                  0xFFC00410         /* CNT0 Debounce Register */
>> +#define CNT_CNTR                    0xFFC00414         /* CNT0 Counter Register */
>> +#define CNT_MAX                     0xFFC00418         /* CNT0 Maximum Count Register */
>> +#define CNT_MIN                     0xFFC0041C         /* CNT0 Minimum Count Register */
>> +
>> +#define RSI_CLK_CONTROL             0xFFC00604         /* RSI0 Clock Control Register */
>> +#define RSI_ARGUMENT                0xFFC00608         /* RSI0 Argument Register */
>> +#define RSI_COMMAND                 0xFFC0060C         /* RSI0 Command Register */
>> +#define RSI_RESP_CMD                0xFFC00610         /* RSI0 Response Command Register */
>> +#define RSI_RESPONSE0               0xFFC00614         /* RSI0 Response 0 Register */
>> +#define RSI_RESPONSE1               0xFFC00618         /* RSI0 Response 1 Register */
>> +#define RSI_RESPONSE2               0xFFC0061C         /* RSI0 Response 2 Register */
>> +#define RSI_RESPONSE3               0xFFC00620         /* RSI0 Response 3 Register */
>> +#define RSI_DATA_TIMER              0xFFC00624         /* RSI0 Data Timer Register */
>> +#define RSI_DATA_LGTH               0xFFC00628         /* RSI0 Data Length Register */
>> +#define RSI_DATA_CONTROL            0xFFC0062C         /* RSI0 Data Control Register */
>> +#define RSI_DATA_CNT                0xFFC00630         /* RSI0 Data Count Register */
>> +#define RSI_STATUS                  0xFFC00634         /* RSI0 Status Register */
>> +#define RSI_STATUSCL                0xFFC00638         /* RSI0 Status Clear Register */
>> +#define RSI_IMSK0                   0xFFC0063C         /* RSI0 Interrupt 0 Mask Register */
>> +#define RSI_IMSK1                   0xFFC00640         /* RSI0 Interrupt 1 Mask Register */
>> +#define RSI_FIFO_CNT                0xFFC00648         /* RSI0 FIFO Counter Register */
> ...
>
> I think I mentioned before that device registers have to be described
> using C structs in U-Boot, and proper I/O accessors must be used to
> access these.
>
> In any case, please trim to the really needed definitions.

OK. I will trim the head file to those that are really used in drivers.

>
>> --- /dev/null
>> +++ b/arch/blackfin/include/asm/mach-bf609/anomaly.h
>> @@ -0,0 +1,103 @@
>> +/*
>> + * DO NOT EDIT THIS FILE
>> + * This file is under version control at
>> + *   svn://sources.blackfin.uclinux.org/toolchain/trunk/proc-defs/header-frags/
>> + * and can be replaced with that version at any time
>> + * DO NOT EDIT THIS FILE
>
> This makes no sense to me.

OK, I will remove these lines.

>
>> + * Copyright 2004-2012 Analog Devices Inc.
>> + * Licensed under the ADI BSD license.
>> + *   https://docs.blackfin.uclinux.org/doku.php?id=adi_bsd
>> + */
>
> I'm not sure if this has been discussed before - is this really GPL
> compatible?
>

The adi_bsd license comment is the same as those in the head files for
BF5xx, which were already merged in to the U-boot mainline git tree.
As the document says "The ADI BSD License is similar to the Clear BSD
License, which the Free Software Foundation indicates is compatible
with both GPLv2 and GPLv3."

I think this has already been discussed when Mike Frysinger got the
Blackfin architecture code merged upstream.


Regards,

Sonic

>
>
> Best regards,
>
> Wolfgang Denk
>
> --
> DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
> HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
> Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
> Harrison's Postulate:
>         For every action, there is an equal and opposite criticism.