linux-kernel - Re: [PATCH 5/6] introduce pmd_to_pte

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 16 Feb 2012 04:02:42 -0500
From:	Naoya Horiguchi <n-horiguchi@...jp.nec.com>
To:	Andrew Morton <akpm@...ux-foundation.org>
Cc:	Naoya Horiguchi <n-horiguchi@...jp.nec.com>, linux-mm@...ck.org,
	David Rientjes <rientjes@...gle.com>,
	Andi Kleen <andi@...stfloor.org>,
	Wu Fengguang <fengguang.wu@...el.com>,
	Andrea Arcangeli <aarcange@...hat.com>,
	KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>,
	KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 5/6] introduce pmd_to_pte_t()

On Wed, Feb 15, 2012 at 04:54:08PM -0800, Andrew Morton wrote:
> On Wed,  8 Feb 2012 10:51:41 -0500
> Naoya Horiguchi <n-horiguchi@...jp.nec.com> wrote:
> 
> > Casting pmd into pte_t to handle thp is strongly architecture dependent.
> > This patch introduces a new function to separate this dependency from
> > independent part.
> > 
> >
> > ...
> >
> > --- 3.3-rc2.orig/include/asm-generic/pgtable.h
> > +++ 3.3-rc2/include/asm-generic/pgtable.h
> > @@ -434,6 +434,10 @@ static inline int pmd_trans_splitting(pmd_t pmd)
> >  {
> >  	return 0;
> >  }
> > +static inline pte_t pmd_to_pte_t(pmd_t *pmd)
> > +{
> > +	return 0;
> > +}
> 
> This doesn't compile.

Sorry for my failing to make sure of compile testing.
The return value should be cast to pte_t to pass the complie.

> And I can't think of a sensible way of generating a stub for this
> operation - if you have a pmd_t and want to convert it to a pte_t then
> just convert it, dammit.  And there's no rationality behind making that
> conversion unavailable or inoperative if CONFIG_TRANSPARENT_HUGEPAGE=n?
> 
> Shudder.  I'll drop the patch.  Rethink, please.

OK for dropping it.
This patch is not enough to solve the problem of isolating arch dependency.

Although it's not clear from the name, the intension of this function was
to get the lowest level of entry in page table hierarchy which points to
a hugepage.  It is pmd for x86_64, but pte for powerpc64 for example.
So I thought it's useful to introduce a stub like above.
But the callers of this function assume that pmd points to hugepage,
so arch dependency in arch independent code still remains.
We need to work on it when thp supports other archs,
but anyway, removing this patch is not critical for others of this series,
so I'm ok about it.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/