linux-kernel - Re: "mm: use early_pfn_to_nid in page_ext

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]

Date:   Tue, 4 Jul 2017 12:53:42 +0200
From:   Michal Hocko <mhocko@...nel.org>
To:     Vlastimil Babka <vbabka@...e.cz>
Cc:     Joonsoo Kim <js1304@...il.com>, Yang Shi <yang.shi@...aro.org>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: "mm: use early_pfn_to_nid in page_ext_init" broken on some
 configurations?

On Tue 04-07-17 11:39:57, Vlastimil Babka wrote:
> On 07/04/2017 07:23 AM, Joonsoo Kim wrote:
> > On Mon, Jul 03, 2017 at 04:18:01PM +0200, Vlastimil Babka wrote:
> >> allocated" looks much more sane there. But there's a warning nevertheless.
> > 
> > Warning would comes from the fact that drain_all_pages() is called
> > before mm_percpu_wq is initialised. We could remove WARN_ON_ONCE() and add
> > drain_local_page(zone) to fix the problem.
> 
> Wouldn't that still leave some period during boot where kernel already
> runs on multiple CPU's, but mm_percpu_wq is not yet initialized and
> somebody tries to use it? We want to catch such cases, right?

I haven't checked the boot sequence but if we know that we need
mm_percpu_wq initialized earlier than moving it should be not a big deal
I guess.
-- 
Michal Hocko
SUSE Labs