[PATCH] nandwrite: Factor out buffer checking code

[Marek Vasut]

On 12/08/2016 07:35 AM, Boris Brezillon wrote:
> Hi Marek,
> 
> On Thu,  8 Dec 2016 05:13:42 +0100
> Marek Vasut <marex at denx.de> wrote:
> 
>> Pull the buffer content checking code into separate function and
>> simplify the code invoking it slightly.
> 
> Thanks for taking the time to prepare and send this patch.
> 
>>
>> Signed-off-by: Marek Vasut <marex at denx.de>
>> Cc: Boris Brezillon <boris.brezillon at free-electrons.com>
>> Cc: David Oberhollenzer <david.oberhollenzer at sigma-star.at>
>> Cc: Kees Trommel <ctrommel at aimvalley.nl>
>> ---
>> NOTE: This is only compile-tested.
>> ---
>>  nand-utils/nandwrite.c | 42 ++++++++++++++++++++++++++++--------------
>>  1 file changed, 28 insertions(+), 14 deletions(-)
>>
>> diff --git a/nand-utils/nandwrite.c b/nand-utils/nandwrite.c
>> index b376869..c5bd1c2 100644
>> --- a/nand-utils/nandwrite.c
>> +++ b/nand-utils/nandwrite.c
>> @@ -231,6 +231,31 @@ static void erase_buffer(void *buffer, size_t size)
>>  		memset(buffer, kEraseByte, size);
>>  }
>>  
>> +/* Check whether buffer is filled with character 'pattern' */
>> +static int buffer_check_pattern(unsigned char *buffer, size_t size,
>> +				unsigned char pattern)
>> +{
>> +	/* Invalid input */
>> +	if (!buffer || (size == 0))
>> +		return 0;
>> +
>> +	/* No match on first byte */
>> +	if (*buffer != pattern)
>> +		return 0;
>> +
>> +	/* First byte matched and buffer is 1 byte long, OK. */
>> +	if (size == 1)
>> +		return 1;
>> +
>> +	/*
>> +	 * Check buffer longer than 1 byte. We already know that buffer[0]
>> +	 * matches the pattern, so the test below only checks whether the
>> +	 * buffer[0...size-2] == buffer[1...size-1] , which is a test for
>> +	 * whether the buffer is filled with constant value.
>> +	 */
>> +	return !memcmp(buffer, buffer + 1, size - 1);
>> +}
>> +
> 
> Can we put that in include/common.h as an inline function so that other
> tools/libs can re-use it?
> I just grepped for the 0xff/0xFF pattern and found several places that
> could use this function instead of open-coding it.
> 
> Looks good otherwise.
> 
> Reviewed-by: Boris Brezillon <boris.brezillon at free-electrons.com>
> 

OK, done and V2 out. But it'd be great if someone could actually test it
:-) Kees ?


-- 
Best regards,
Marek Vasut