Re: [PATCH 11/17] mm: page_alloc: Use unsigned int for order in more places

[Dave Hansen <dave.hansen@intel.com>]

On 05/01/2014 08:11 AM, Mel Gorman wrote:
> On Thu, May 01, 2014 at 07:35:47AM -0700, Dave Hansen wrote:
>> On 05/01/2014 01:44 AM, Mel Gorman wrote:
>>> X86 prefers the use of unsigned types for iterators and there is a
>>> tendency to mix whether a signed or unsigned type if used for page
>>> order. This converts a number of sites in mm/page_alloc.c to use
>>> unsigned int for order where possible.
>>
>> Does this actually generate any different code?  I'd actually expect
>> something like 'order' to be one of the easiest things for the compiler
>> to figure out an absolute range on.
> 
> Yeah, it generates different code. Considering that this patch affects an
> API that can be called external to the code block how would the compiler
> know what the range of order would be in all cases?

The compiler comprehends that if you do a check against a constant like
MAX_ORDER early in the function that the the variable now has a limited
range, like the check we do first-thing in __alloc_pages_slowpath().

The more I think about it, at least in page_alloc.c, I don't see any
checks for order<0, which means the compiler isn't free to do this
anyway.  Your move over to an unsigned type gives that check for free
essentially.

So this makes a lot of sense in any case.  I was just curious if it
affected the code.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>