linux-kernel - Re: [PATCH] tools/vm/page_owner_sort.c: add switch between culling by stacktrace and txt

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

Message-Id: <20220324161515.19e74f062210861d71ac414b@linux-foundation.org>
Date:   Thu, 24 Mar 2022 16:15:15 -0700
From:   Andrew Morton <akpm@...ux-foundation.org>
To:     Yinan Zhang <zhangyinan2019@...il.szu.edu.cn>
Cc:     Sean Anderson <seanga2@...il.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] tools/vm/page_owner_sort.c: add switch between culling
 by stacktrace and txt

On Thu, 24 Mar 2022 08:54:34 +0800 Yinan Zhang <zhangyinan2019@...il.szu.edu.cn> wrote:

> on 2022/3/24 3:08, Andrew Morton wrote:
> > On Wed, 23 Mar 2022 10:05:28 +0800 Yinan Zhang<zhangyinan2019@...il.szu.edu.cn>  wrote:
> >
> >>>> *(&list[i].txt+offset)) != 0) {
> >>>>                list2[count++] = list[i];
> >>>>            } else {
> >>>>                list2[count-1].num += list[i].num;
> >>>>
> >>>
> >> You are right. Due to timestamp etc, it's almost impossible to
> >>
> >> cull by txt. And I did observe some blocks differ in head and same in
> >>
> >> stack trace on my linux. The examples I mentioned in patch are
> >>
> >> part of them.
> >>
> >>
> >> Besides, --cull has been added in patch:
> >>
> >> tools/vm/page_owner_sort.c: support for user-defined culling rules
> >>
> > So should I drop this patch?
> 
> you decide.

I'd rather not.

And dropping this patch makes a huge mess of the series.  So I either
ask for the whole series to be redone and we might miss the merge
window or I send it all in as-is.

Sigh.  I'll send it all in as-is.  Please continue to work on this
issue and if we decide to drop the -c option, please send along a patch
to remove it over the next week or two.