linux-kernel - Re: arm64/efi handling of persistent memory

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]

Date:	Wed, 20 Jan 2016 14:23:32 +0000
From:	Mark Rutland <mark.rutland@....com>
To:	Ard Biesheuvel <ard.biesheuvel@...aro.org>
Cc:	Leif Lindholm <leif.lindholm@...aro.org>,
	"Elliott, Robert (Persistent Memory)" <elliott@....com>,
	"catalin.marinas@....com" <catalin.marinas@....com>,
	"will.deacon@....com" <will.deacon@....com>,
	"matt@...eblueprint.co.uk" <matt@...eblueprint.co.uk>,
	"dan.j.williams@...el.com" <dan.j.williams@...el.com>,
	"Kani, Toshimitsu" <toshi.kani@....com>,
	"Knippers, Linda" <linda.knippers@....com>,
	"linux-arm-kernel@...ts.infradead.org" 
	<linux-arm-kernel@...ts.infradead.org>,
	"linux-nvdimm@...ts.01.org" <linux-nvdimm@...ts.01.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Matthew Wilcox <willy@...ux.intel.com>,
	Peter Zijlstra <peterz@...radead.org>,
	Theodore Ts'o <tytso@....edu>
Subject: Re: arm64/efi handling of persistent memory

Hi,

For those newly Cc'd, the initially reported problem is that arm64 Linux
currently treats persistent memory as with any other memory (happily clobbering
it), per [1].

On Fri, Dec 18, 2015 at 02:07:10PM +0100, Ard Biesheuvel wrote:
> Before we start hacking away at this at the arm64/EFI level, do we
> have any documentation and/or consensus regarding how persistent
> memory should be treated in the first place? Should it be covered by
> memblock? Should it be covered by the linear mapping? Should it be
> memblock_reserve()'d?

I'm hoping that the lack of replies has more to do with the recent
holiday than a lack of opinion...

I think that it's sensible to say that at minimum we need to ensure that we
don't treat it as available RAM (i.e. we don't clobber it with random data) for
now.

Per [2] it's not clear to me what the consensus is on memblock, the linear
mapping, and the use of struct page, though that's months old so perhaps that's
been figured out since. I've Cc'd some of the attendees in case they can
clarify the situation.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2015-December/394707.html
[2] https://lwn.net/Articles/636096/

Thanks,
Mark.