PERFORCE change 174261 for review

[Rafal Jaworowski]

On 2010-02-03, at 21:24, Nathan Whitehorn wrote:

> Rafal Jaworowski wrote:
>> http://p4web.freebsd.org/chv.cgi?CH=174261
>> 
>> Change 174261 by raj at raj_fdt on 2010/02/03 20:20:54
>> 
>> 	Bring more endian-safety for FDT uart(4) attachment.
>> 
>> Affected files ...
>> 
>> .. //depot/projects/fdt/sys/dev/uart/uart_bus_fdt.c#5 edit
>> 
>> Differences ...
>> 
>> ==== //depot/projects/fdt/sys/dev/uart/uart_bus_fdt.c#5 (text+ko) ====
>> 
>> @@ -1,5 +1,5 @@
>> /*-
>> - * Copyright (c) 2009 The FreeBSD Foundation
>> + * Copyright (c) 2009-2010 The FreeBSD Foundation
>>  * All rights reserved.
>>  *
>>  * This software was developed by Semihalf under sponsorship from
>> @@ -35,11 +35,15 @@
>> #include <sys/kernel.h>
>> #include <sys/module.h>
>> +#include <machine/bus.h>
>> +
>> #include <dev/ofw/ofw_bus.h>
>> #include <dev/ofw/ofw_bus_subr.h>
>> #include <dev/uart/uart.h>
>> #include <dev/uart/uart_bus.h>
>> +#include "../../contrib/dtc/libfdt/libfdt_env.h"
>> +
>> static int uart_fdt_probe(device_t);
>>  static device_method_t uart_fdt_methods[] = {
>> @@ -81,8 +85,11 @@
>> 		    sizeof(clock))) <= 0)
>> 			clock = 0;
>> +	clock = fdt32_to_cpu(clock);
>> +
> Can't you just use ntoh*? As far as I know, FDT (and Open Firmware) are defined to always use network byte order.

Techinically yes, although the fdt* accessors are native for libfdt, so I'd rather stay with uniform behaviour. Also, they are explict in name while noth* could be a bit confusing, I think, for people not intimate with FDT internals.

Rafal