From: Andrew Morton <akpm@osdl.org>
To: Hugh Dickins <hugh@veritas.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: [PATCH 04/21] mm: zap_pte_range dont dirty anon
Date: Sun, 25 Sep 2005 15:26:30 -0700 [thread overview]
Message-ID: <20050925152630.75560571.akpm@osdl.org> (raw)
In-Reply-To: <Pine.LNX.4.61.0509251649100.3490@goblin.wat.veritas.com>
Hugh Dickins <hugh@veritas.com> wrote:
>
> zap_pte_range already avoids wasting time to mark_page_accessed on anon
> pages: it can also skip anon set_page_dirty - the page only needs to be
> marked dirty if shared with another mm, but that will say pte_dirty too.
>
Are you sure about this?
> --- mm03/mm/memory.c 2005-09-24 19:26:38.000000000 +0100
> +++ mm04/mm/memory.c 2005-09-24 19:27:05.000000000 +0100
> @@ -574,12 +574,14 @@ static void zap_pte_range(struct mmu_gat
> addr) != page->index)
> set_pte_at(tlb->mm, addr, pte,
> pgoff_to_pte(page->index));
> - if (pte_dirty(ptent))
> - set_page_dirty(page);
> if (PageAnon(page))
> dec_mm_counter(tlb->mm, anon_rss);
> - else if (pte_young(ptent))
> - mark_page_accessed(page);
> + else {
> + if (pte_dirty(ptent))
> + set_page_dirty(page);
> + if (pte_young(ptent))
> + mark_page_accessed(page);
> + }
> tlb->freed++;
> page_remove_rmap(page);
> tlb_remove_page(tlb, page);
What is the page is (for example) clean swapcache, having been recently
faulted in. If this pte indicates that this process has modified the page
and we don't run set_page_dirty(), the page could be reclaimed and the
change is lost.
Or what is the page was an anon page resulting from (say) a swapoff, and
it's shared by two mm's and one has modified it and we drop that dirty pte?
Or <other scenarios>.
Need more convincing.
next prev parent reply other threads:[~2005-09-25 22:27 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-09-25 15:46 [PATCH 00/21] mm: page fault scalability prep Hugh Dickins
2005-09-25 15:47 ` [PATCH 01/21] mm: hugetlb truncation fixes Hugh Dickins
2005-09-25 15:48 ` [PATCH 02/21] mm: copy_pte_range progress fix Hugh Dickins
2005-09-25 15:49 ` [PATCH 03/21] mm: msync_pte_range progress Hugh Dickins
2005-09-25 15:49 ` [PATCH 04/21] mm: zap_pte_range dont dirty anon Hugh Dickins
2005-09-25 22:26 ` Andrew Morton [this message]
2005-09-26 6:02 ` Hugh Dickins
2005-09-26 6:14 ` Andrew Morton
2005-09-26 7:20 ` Hugh Dickins
2005-09-25 15:51 ` [PATCH 05/21] mm: anon is already wrprotected Hugh Dickins
2005-09-25 15:52 ` [PATCH 06/21] mm: vm_stat_account unshackled Hugh Dickins
2005-09-25 15:53 ` [PATCH 07/21] mm: remove_vma_list consolidation Hugh Dickins
2005-09-25 15:53 ` [PATCH 08/21] mm: unlink_file_vma, remove_vma Hugh Dickins
2005-09-25 15:54 ` [PATCH 09/21] mm: exit_mmap need not reset Hugh Dickins
2005-09-25 15:56 ` [PATCH 10/21] mm: page fault handlers tidyup Hugh Dickins
2005-09-25 15:57 ` [PATCH 11/21] mm: move_page_tables by extents Hugh Dickins
2005-09-25 15:59 ` [PATCH 12/21] mm: tlb_gather_mmu get_cpu_var Hugh Dickins
2005-09-25 16:01 ` [PATCH 13/21] mm: tlb_is_full_mm was obscure Hugh Dickins
2005-09-25 16:03 ` [PATCH 14/21] mm: tlb_finish_mmu forget rss Hugh Dickins
2005-09-25 16:06 ` [PATCH 15/21] mm: mm_init set_mm_counters Hugh Dickins
2005-09-25 16:07 ` [PATCH 16/21] mm: rss = file_rss + anon_rss Hugh Dickins
2005-09-25 16:08 ` [PATCH 17/21] mm: batch updating mm_counters Hugh Dickins
2005-09-26 7:25 ` Nick Piggin
2005-09-26 8:42 ` Hugh Dickins
2005-09-25 16:09 ` [PATCH 18/21] mm: dup_mmap use oldmm more Hugh Dickins
2005-09-25 16:10 ` [PATCH 19/21] mm: dup_mmap down new mmap_sem Hugh Dickins
2005-09-25 16:11 ` [PATCH 20/21] mm: sh64 hugetlbpage.c Hugh Dickins
2005-09-29 7:00 ` Paul Mundt
2005-09-25 16:15 ` [PATCH 21/21] mm: m68k kill stram swap Hugh Dickins
2005-09-28 0:05 ` [PATCH 00/21] mm: page fault scalability prep Christoph Lameter
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20050925152630.75560571.akpm@osdl.org \
--to=akpm@osdl.org \
--cc=hugh@veritas.com \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®