Re: [PATCH 2/2] jffs2: fix buffer dump debug output

[Cyrille Pitchen <cyrille.pitchen@atmel.com>]

Le 21/10/2014 11:54, Artem Bityutskiy a écrit :
> On Tue, 2014-10-21 at 11:47 +0200, Cyrille Pitchen wrote:
>> thanks for your comment. Indeed using print_hex_dump() would be a good idea
>> since it would avoid each driver to implement its own version of buffer dumps.
>> Reading the source code of print_hex_dump(), the output format would change a
>> little bit:
>> the output would no longer be aligned to JFFS2_BUFDUMP_BYTES_PER_LINE boundary
>> using leading spaces.
>> If it's ok to change the output format it's good for me as well.
> 
> Changing format is OK, this is just a debugging cruft. But it is OK only
> if you test the changes. If you are unable to test the changes, it is
> better to avoid doing non-trivial changes.
> 
The version provided in this patch was tested. However I don't know how to cope
with the offs parameter of __jffs2_dbg_dump_buffer() if I replace the current
source code by a call to print_hex_dump().
If I use DUMP_PREFIX_ADDRESS as prefix_type value, the printed offsets are
relative to the address of buf in RAM, not to the hardware offset in my
dataflash provided by the offs parameter.
Also, if I use DUMP_PREFIX_OFFSET, the printed offsets start from 0 instead of
offs value.

I think keeping the right offsets in the output dump is helpfull.