Re: [PATCH 45/50] x86, memblock: Use memblock_debug to control debug message print out

[Yinghai Lu <yinghai@kernel.org>]

On 07/13/2010 01:37 PM, Bjorn Helgaas wrote:
> On Tuesday, July 13, 2010 01:10:39 am Yinghai Lu wrote:
>> -		pr_cont(" ==> [%010llx-%010llx]\n", final_start, final_end - 1);
>> +		memblock_dbg(" ==> [%#010llx-%#010llx]\n", final_start, final_end - 1);
>>  		reserve_bootmem_generic(final_start, final_end - final_start, BOOTMEM_DEFAULT);
>>  	}
>>  
>> @@ -286,9 +286,11 @@ void __init memblock_x86_reserve_range(u64 start, u64 end, char *name)
>>  	if (start == end)
>>  		return;
>>  
>> -	if (WARN_ONCE(start > end, "memblock_x86_reserve_range: wrong range [%#llx, %#llx]\n", start, end))
>> +	if (WARN_ONCE(start > end, "memblock_x86_reserve_range: wrong range [%#llx, %#llx)\n", start, end))
> 
> Can you print these ranges the same way as the others?  I think
> "invalid range" might be closer to what you mean than "wrong range." 

like to use end instead of end - 1.

> 
> I'm a little dubious about these "(start == end)" and "(start > end)"
> checks anyway.  Who are the callers of these functions?  If "start"
> and "end" are coming from an external source, e.g., some firmware
> interface like an e820 table, the message doesn't give enough of a
> clue about where the problem is.
> 
> If "start" and "end" are internal things, I'd argue that the checks
> are just covering up Linux bugs, and it'd be better to fix those
> bugs and remove the checks.

so we add WARN_ here, could warn the possible wrong usage.

YH