linux-kernel - Re: [PATCH] mm: migrate: Fix THP's mapcount on isolation

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]

Message-ID: <Y35FLfDzZjMlwgF4@casper.infradead.org>
Date:   Wed, 23 Nov 2022 16:07:09 +0000
From:   Matthew Wilcox <willy@...radead.org>
To:     David Hildenbrand <david@...hat.com>
Cc:     Hugh Dickins <hughd@...gle.com>, Gavin Shan <gshan@...hat.com>,
        linux-mm@...ck.org, linux-kernel@...r.kernel.org,
        akpm@...ux-foundation.org, william.kucharski@...cle.com,
        ziy@...dia.com, kirill.shutemov@...ux.intel.com,
        zhenyzha@...hat.com, shan.gavin@...il.com, riel@...riel.com,
        apopple@...dia.com
Subject: Re: [PATCH] mm: migrate: Fix THP's mapcount on isolation

On Wed, Nov 23, 2022 at 09:56:38AM +0100, David Hildenbrand wrote:
> But we do have an even better helper in place already:
> mm/huge_memory.c:can_split_folio()
> 
> Which cares about
> 
> a) Swapcache for THP: each subpage could be in the swapcache
> b) Requires the caller to hold one reference to be safe
> 
> But I am a bit confused about the "extra_pins" for !anon. Where do the
> folio_nr_pages() references come from?

When we add a folio to the page cache, we increment its refcount by
folio_nr_pages() instead of by 1.  I suspect this is no longer needed
(if it was ever needed) and it could be changed.  See
__filemap_add_folio():

        long nr = 1;
        if (!huge) {
                nr = folio_nr_pages(folio);
        folio_ref_add(folio, nr);

> So *maybe* it makes sense to factor out can_split_folio() and call it
> something like: "folio_maybe_additionally_referenced"  [to clearly
> distinguish it from "folio_maybe_dma_pinned" that cares about actual page
> pinning (read/write page content)].
> 
> Such a function could return false positives/negatives due to races and the
> caller would have to hold one reference and be able to deal with the
> semantics.

I don't like the 'pextra_pins' parameter to a generic function ...