Re: [PATCH v2] of/address: replace printk() with pr_debug() / pr_err()

[Rasmus Villemoes <linux@rasmusvillemoes.dk>]

On Wed, Dec 09 2015, Andy Shevchenko <andy.shevchenko@gmail.com> wrote:

> On Wed, Dec 9, 2015 at 9:28 PM, Joe Perches <joe@perches.com> wrote:
>> ---
>>  include/linux/printk.h |  7 +++++++
>>  lib/hexdump.c          | 39 +++++++++++++++++++++++++++++++--------
>>  2 files changed, 38 insertions(+), 8 deletions(-)
>>
>> diff --git a/include/linux/printk.h b/include/linux/printk.h
>> index 9729565..4be190c 100644
>> --- a/include/linux/printk.h
>> +++ b/include/linux/printk.h
>> @@ -424,6 +424,13 @@ enum {
>>         DUMP_PREFIX_ADDRESS,
>>         DUMP_PREFIX_OFFSET
>>  };
>> +
>> +enum {
>> +       DUMP_TYPE_CPU = 0,
>
> And still open this, do we need it? I think you may just mention in
> the documentation that default behaviour is CPU like.

I think it's best to have a name for the default. In this case it's
unlikely to ever be relevant, but in general one could imagine stuff
like

#ifdef THIS_OR_THAT
#define MY_DUMP_TYPE DUMP_TYPE_LE
#else
#define MY_DUMP_TYPE DUMP_TYPE_CPU
#endif

which is a lot more readable than if the latter was a naked 0.

The feature seems fine to me. It's always been somewhat annoying that
grouping changes the order of the byte values on little-endian machines
(so grouping is actually a wrong name for it; it's more like "treat the
bytes an array of u16/u32/u64"), and one can now get around that by
specifying DUMP_TYPE_BE.

Rasmus