Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id <S290103AbSBSSKA>; Tue, 19 Feb 2002 13:10:00 -0500
Received: (majordomo@vger.kernel.org) by vger.kernel.org
	id <S291560AbSBSSJu>; Tue, 19 Feb 2002 13:09:50 -0500
Received: from bay-bridge.veritas.com ([143.127.3.10]:38845 "EHLO
	svldns02.veritas.com") by vger.kernel.org with ESMTP
	id <S290103AbSBSSJn>; Tue, 19 Feb 2002 13:09:43 -0500
Date: Tue, 19 Feb 2002 18:11:26 +0000 (GMT)
From: Hugh Dickins <hugh@veritas.com>
To: Linus Torvalds <torvalds@transmeta.com>
cc: Daniel Phillips <phillips@bonn-fries.net>,
        Rik van Riel <riel@conectiva.com.br>, dmccr@us.ibm.com,
        Kernel Mailing List <linux-kernel@vger.kernel.org>, linux-mm@kvack.org,
        Robert Love <rml@tech9.net>, mingo@redhat.co,
        Andrew Morton <akpm@zip.com.au>, manfred@colorfullife.com,
        wli@holomorphy.com
Subject: Re: [RFC] Page table sharing
In-Reply-To: <Pine.LNX.4.33.0202190923390.26476-100000@home.transmeta.com>
Message-ID: <Pine.LNX.4.21.0202191801430.15103-100000@localhost.localdomain>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
X-Mailing-List: linux-kernel@vger.kernel.org

On Tue, 19 Feb 2002, Linus Torvalds wrote:
> On Tue, 19 Feb 2002, Daniel Phillips wrote:
> > >
> > > At that point you might as well make the TLB shootdown global (ie you keep
> > > track of a mask of CPU's whose TLB's you want to kill, and any pmd that
> > > has count > 1 just makes that mask be "all CPU's").
> >
> > How do we know when to do the global tlb flush?
> 
> See above.
> 
> Basically, the algorithm is:
> 
> 	invalidate_cpu_mask = 0;
> 
> 	.. for each page swapped out ..
> 
> 		pte = ptep_get_and_clear(ptep);
> 		save_pte_and_mm(pte_page(pte));
> 		mask = mm->cpu_vm_mask;
> 		if (page_count(pmd_page) > 1)
> 			mask = ~0UL;
> 		invalidate_cpu_mask |= mask;
> 
> and then at the end you just do
> 
> 	flush_tlb_cpus(invalidate_cpu_mask);
> 	for_each_page_saved() {
> 		free_page(page);
> 	}
> 
> (yeah, yeah, add cache coherency etc).

It's a little worse than this, I think.  Propagating pte_dirty(pte) to
set_page_dirty(page) cannot be done until after the flush_tlb_cpus, if
the ptes are writable: and copy_page_range is not setting "cow", so not
write protecting, when it's a shared writable mapping.  Easy answer is
to scrap "cow" there and always do the write protection; but I doubt
that's the correct answer.  swap_out could keep an array of pointers to
ptes, to propagate dirty after flushing TLB and before freeing pages,
but that's not very pretty.

Hugh

-
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/