linux-kernel - Re: [PATCH] mm: use clear_user_(high)page() for arch with special user folio handling

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] [day] [month] [year] [list]

Message-ID: <bc3b86dd-9bb6-4c6e-b6df-35c687375cde@kernel.org>
Date: Fri, 6 Dec 2024 09:31:27 -0800
From: Vineet Gupta <vgupta@...nel.org>
To: Zi Yan <ziy@...dia.com>, Matthew Wilcox <willy@...radead.org>
Cc: linux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>,
 Geert Uytterhoeven <geert@...ux-m68k.org>, Vlastimil Babka <vbabka@...e.cz>,
 David Hildenbrand <david@...hat.com>, Miaohe Lin <linmiaohe@...wei.com>,
 Kefeng Wang <wangkefeng.wang@...wei.com>, John Hubbard
 <jhubbard@...dia.com>, "Huang, Ying" <ying.huang@...el.com>,
 Ryan Roberts <ryan.roberts@....com>, Alexander Potapenko
 <glider@...gle.com>, Kees Cook <keescook@...omium.org>,
 linux-kernel@...r.kernel.org, Geert Uytterhoeven <geert+renesas@...der.be>,
 Vineet Gupta <vgupta@...nel.org>,
 Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
Subject: Re: [PATCH] mm: use clear_user_(high)page() for arch with special
 user folio handling

On 12/6/24 09:08, Zi Yan wrote:
> On 6 Dec 2024, at 11:42, Matthew Wilcox wrote:
>
>> On Fri, Dec 06, 2024 at 09:19:38AM -0500, Zi Yan wrote:
>>> For architectures setting ARCH_HAS_CPU_CACHE_ALIASING, which requires
>>> flushing cache and arc, which sets folio->flags after clearing a user
>>> folio
>> I think arc just has some legacy code left that needs to be ripped out.
>> See commit d5272aaa8257 which took away ARCH_HAS_CPU_CACHE_ALIASING
>> from arc.
> The PG_dc_clean bit seems not to be related to cache aliasing.
> See commit eacd0e950dc2 ("ARC: [mm] Lazy D-cache flush (non aliasing VIPT)")

Indeed it is not related to aliasing.

>> Looking at 8690bbcf3b70, I wonder if you want to use
>> cpu_dcache_is_aliasing() here?
> Yeah, this is more precise. Will use this in my next version. Thanks.

Commit 8690bbcf3b70 was wrong for ARC as it assumed aliasing is true. The
aliasing D$ code is all gone and cpu_dcache_is_aliasing() is false for ARC now.

-vineet