Date: Wed, 30 Jul 2008 20:39:44 +0300
From: Adrian Bunk <bunk@kernel.org>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>, lguest <lguest@ozlabs.org>,
       linux-kernel@vger.kernel.org, Nick Piggin <nickpiggin@yahoo.com.au>
Subject: Re: [PATCH 1/2] Export get_user_pages_fast
Message-ID: <20080730173944.GC28604@cs181140183.pp.htv.fi>
References: <200807301438.55941.rusty@rustcorp.com.au> <20080730033523.0f38f7e4.akpm@linux-foundation.org> <20080730154736.GC21597@cs181140183.pp.htv.fi> <20080730100355.f1402dcf.akpm@linux-foundation.org> <20080730170939.GA28604@cs181140183.pp.htv.fi> <20080730101438.e1484440.akpm@linux-foundation.org> <20080730171845.GB28604@cs181140183.pp.htv.fi> <20080730102817.c74a9e33.akpm@linux-foundation.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20080730102817.c74a9e33.akpm@linux-foundation.org>
User-Agent: Mutt/1.5.18 (2008-05-17)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1910
Lines: 48

On Wed, Jul 30, 2008 at 10:28:17AM -0700, Andrew Morton wrote:
> On Wed, 30 Jul 2008 20:18:45 +0300 Adrian Bunk <bunk@kernel.org> wrote:
> 
> > > > The export would still only be available on x86 since no other arch 
> > > > would define CONFIG_HAVE_GET_USER_PAGES_FAST.
> > > > 
> > > 
> > > That's presently true in Linus mainline.  The powerpc implementation is
> > > in -mm, but apprarently needs a bit more work.
> > 
> > I still don't get the advantage of your suggestion.
> > 
> > With the powerpc implementation included your suggestion doesn't change 
> > the fact that the export is available only on x86 and powerpc, and if a 
> > driver author mistakenly uses it and only tests it's driver on x86 it 
> > will still break on architectures other than x86 or powerpc.
> 
> On non-CONFIG_HAVE_GET_USER_PAGES_FAST architectures a
> get_user_pages_fast() call will in fact call get_user_pages().
> 
> On CONFIG_HAVE_GET_USER_PAGES_FAST architectures, get_user_pages_fast()
> is a real function.  All architectures which implement that function
> should export it to modules.  Hence it should be exported from within
> generic code.

Are we talking at cross purposes?

I thought you were talking an error due to a missing function.

If you think about an error where the function is present but only the 
EXPORT_SYMBOL is missing then I get your point (although I do not 
consider it a huge problem).

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

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