linux-kernel - Re: [PATCH v5 0/7] mm/mprotect: Fix dax puds

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: <ZruEcHS3Yk1O9ovI@x1n>
Date: Tue, 13 Aug 2024 12:06:08 -0400
From: Peter Xu <peterx@...hat.com>
To: Michael Ellerman <mpe@...erman.id.au>
Cc: linux-kernel@...r.kernel.org, linux-mm@...ck.org,
	"Kirill A . Shutemov" <kirill@...temov.name>,
	Nicholas Piggin <npiggin@...il.com>,
	David Hildenbrand <david@...hat.com>,
	Matthew Wilcox <willy@...radead.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	James Houghton <jthoughton@...gle.com>,
	Huang Ying <ying.huang@...el.com>,
	"Aneesh Kumar K . V" <aneesh.kumar@...ux.ibm.com>,
	Vlastimil Babka <vbabka@...e.cz>,
	Rick P Edgecombe <rick.p.edgecombe@...el.com>,
	Hugh Dickins <hughd@...gle.com>, Borislav Petkov <bp@...en8.de>,
	Christophe Leroy <christophe.leroy@...roup.eu>,
	Rik van Riel <riel@...riel.com>,
	Dan Williams <dan.j.williams@...el.com>,
	Mel Gorman <mgorman@...hsingularity.net>, x86@...nel.org,
	Ingo Molnar <mingo@...hat.com>, linuxppc-dev@...ts.ozlabs.org,
	Dave Hansen <dave.hansen@...ux.intel.com>,
	Dave Jiang <dave.jiang@...el.com>,
	Oscar Salvador <osalvador@...e.de>,
	Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH v5 0/7] mm/mprotect: Fix dax puds

On Tue, Aug 13, 2024 at 10:50:04PM +1000, Michael Ellerman wrote:
> > - Power, as no hardware on hand.
> 
> Does it need some specific configuration, or just any Power machine will do?

I am not familiar with both dax and power to be sure on the hardware
implications here, sorry.  I think as long as one can create some /dev/dax
with 1g mapping alignment (I suppose normally with ndctl) then it should be
able to hit the mprotect() path.

One can verify first without this series that it could trigger a bad pud,
then it'll be away after this series applied.  Meanwhile the mprotect()
should apply on the 1g dax range properly.

Thanks,

-- 
Peter Xu