Re: Question of resource_size() implementation

[Ben Nizette]

On 30/12/2009, at 11:16 AM, Felipe Balbi wrote:

> On Tue, 2009-12-29 at 16:12 -0800, Joe Perches wrote:
>> On Wed, 2009-12-30 at 01:43 +0200, Felipe Balbi wrote:
>>> I'm wondering whether the +1 in resource_size() is actually necessary.
>>> resource_size() is defined as:
>> []
>>> static inline resource_size_t resource_size(const struct resource *res)
>>> {
>>> 	return res->end - res->start + 1;
>>> }
>>> Are we off-by-one
>>> here ? Or is this all expected ?
>> 
>> Imagine you have 1 byte sized resources.
>> 
>> AREA1 = 0x40000000
>> AREA2 = 0x40000001
>> 
>> area1.start = 0x40000000
>> area1.end   = 0x40000000
>> 
>> area2.start = 0x40000001
>> area2.end   = 0x40000001
> 
> (adding lkml back to the loop)
> 
> in that you wouldn't use any of the SZ_* macros and simply hardcode
> start and end, right ? then you would define:
> 
> area1.start = 0x40000000
> area1.end = 0x40000001

No-o, you'd hardcode

area1.start=0x40000000
area1.end=0x40000000

As Joe wrote.  Then the resource_size would return 1 and ioremap would map the 1 byte starting at 0x400000000, i.e. just 0x400000000.  This is correct.

In your original example you had

		.start	= MEM_AREA1_BASE,
		.end	= MEM_AREA1_BASE + SZ_4K - 1

So resource_size would return SZ_4K and ioremap would map the SZ_4K bytes starting at MEM_AREA1_BASE, i.e. MEM_AREA1_BASE to MEM_AREA1_BASE - 1 /inclusive/.  This is also correct.

	--Ben.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/