linux-kernel - Re: scheduling while atomic in z3fold

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]

Message-ID: <e1260d25c623f940138230dc2ebaad471c61f116.camel@gmx.de>
Date:   Mon, 30 Nov 2020 20:09:22 +0100
From:   Mike Galbraith <efault@....de>
To:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc:     Oleksandr Natalenko <oleksandr@...alenko.name>,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        linux-rt-users@...r.kernel.org
Subject: Re: scheduling while atomic in z3fold

On Mon, 2020-11-30 at 17:32 +0100, Sebastian Andrzej Siewior wrote:
> On 2020-11-30 17:27:17 [+0100], Mike Galbraith wrote:
> > > This just passed. It however killed my git-gc task which wasn't done.
> > > Let me try tomorrow with your config.
> >
> > FYI, I tried 5.9-rt (after fixing 5.9.11), it exploded in the same way,
> > so (as expected) it's not some devel tree oopsie.
>
> v5.9 has the same missing local-lock init due to merged local-lock
> upstream. I don't remember when this get_cpu_ptr() came it.

Looks like get_cpu_ptr() arrived in 4.14.

I patched up zswap/z3fold in my old 5.1-rt tree (old forward port of
4.19-rt), turned them on with a script, and it all worked fine, so the
RT zswap allergy lies somewhere upstream of there it seems.

If I were to start merging zswap and friends forward, I'd likely
eventually meet whatever RT is allergic to.  I might try that, but
won't be the least surprised if it turns into a god awful mess of
dependencies.

Hopping forward in rt branches first and patching them up will bracket
the little bugger a lot quicker.

	-Mike