Re: [PATCH] wrong accounting in direct_remap_pfn_range

[Steven Rostedt <srostedt@redhat.com>]

Keir Fraser wrote:
> On 31/8/06 1:37 am, "Steven Rostedt" <srostedt@redhat.com> wrote:
> 
>> grr, I take it back, I am the one that's confused :P
>>
>> OK, this all happens because this whole blob of code is crazy because it
>> is missing a "if (size == 0)" check!
> 
> It's not really missing. We could have a size==0 check *or* we can have the
> v!=u check. We don't need both and I think the latter is more obviously
> correct, as the test is closer to the code that it 'protects'. Also it's a
> fairly idiomatic way of generating and flushing batches of work.
> 

Well it wasn't obvious to me :P

If a size == 0 is passed in (for whatever reason!), couldn't we skip the 
  flush_cache_all, flush_tlb_all and the allocation and freeing of a 
page and just return?

If you want this in mainline Linux, you'll probably have others mention 
that too.

-- Steve