Re: [2.6 patch] linux/swap.h must #include <linux/pagemap.h>

[Andrew Morton <akpm@linux-foundation.org>]

On Sat, 26 Jul 2008 21:15:47 +0100 (BST) Hugh Dickins <hugh@veritas.com> wrote:

> On Sat, 26 Jul 2008, Andrew Morton wrote:
> > On Sat, 26 Jul 2008 19:13:43 +0100 (BST) Hugh Dickins <hugh@veritas.com> wrote:
> > 
> > > > /home/bunk/linux/kernel-2.6/git/linux-2.6/include/asm-generic/tlb.h: In function 'tlb_flush_mmu':
> > > > /home/bunk/linux/kernel-2.6/git/linux-2.6/include/asm-generic/tlb.h:76: error: implicit declaration of function 'release_pages'
> > > > /home/bunk/linux/kernel-2.6/git/linux-2.6/include/asm-generic/tlb.h: In function 'tlb_remove_page':
> > > > /home/bunk/linux/kernel-2.6/git/linux-2.6/include/asm-generic/tlb.h:105: error: implicit declaration of function 'page_cache_release'
> > 
> > tlb_finish_mmu() inlines tlb_flush_mmu() and ends up being quite large.
> > 
> > tlb_finish_mmu() has five callsites in core kernel.
> > 
> > The above error is God's way of telling us that it's all too large to
> > be inlined.
> > 
> > uninlining these things would be a bit of a pita because of the various
> > arch overrides.  But it really is the right thing to do.
> 
> Yes, that's a good point too.  They could be put in arch/*/mm/somewhere.c.
> 
> But I'd rather put that off even further into the future, because there's
> quite a bit of rationalization to be done (more arch differences than
> necessary), and Ben's the one currently holding our reorganize-mmu_gather
> ball.  (I keep pointing to you, Ben, please don't think I'm accusing at
> all: I'm the guilty one who held it and did nothing with it before.)
> 
> I'm still in favour the quick fs/exec.c build fixup for now.

Well, safe-simple-and-duck-the-issue has its merits.  My inbox awaits thee :)