Re: [PATCH 6/13] maps#2: Move the page walker code to lib/

[Nick Piggin <nickpiggin@yahoo.com.au>]

Andrew Morton wrote:
> On Wed, 11 Apr 2007 16:35:44 +1000 Nick Piggin <nickpiggin@yahoo.com.au> wrote:
> 
> 
>>Matt Mackall wrote:
>>
>>>Move the page walker code to lib/
>>>
>>>This lets it get shared outside of proc/ and linked in only when
>>>needed.
>>
>>Still should go into mm/
> 
> 
> spose so.
> 
> 
>>If it had, you might have also noticed your pagetable walking code
>>is completely different from how everyone else does it, and fixed
>>that too.
> 
> 
> Different in what way?


The form of the loops. It may not seem like a big deal, but before
Hugh did a big cleanup to make them all the same, we had about a
dozen versions which were all very slightly different ;) It was
painful.

>>BTW. Is it the case that unused and unexported symbols don't get
>>pruned by the linker except inside lib/?
> 
> 
> If they're static and unreferenced then the linker will remove them.  Of
> course, usually humans remove these because they generate warnings, unless
> special-things happen.  We deliberately do special-things with
> register_cpu_notifier() so the notifier-block and the handler go away if
> !CONFIG_HOTPLUG_CPU.
> 
> If the functions are non-static then yes, I expect we end up with them in
> vmlinux.  There are tricks we can play with -ffunction-sections to fix
> that, but we don't.  

They aren't static. But anyway, AFAIKS, all the CONFIG_ stuff is
there anyway, so we can not only avoid the linking but also the
compiling by using them.

-- 
SUSE Labs, Novell Inc.