Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:7 as permitted sender) client-ip=2620:137:e000::3:7;
Message-ID: <68b18065d8be905c25522bd3f5a9b46dbe3a976d.camel@sipsolutions.net>
Subject: Re: [PATCH] um: time-travel: fix time going backwards
From:   Johannes Berg <johannes@sipsolutions.net>
To:     Vincent Whitchurch <Vincent.Whitchurch@axis.com>,
        "anton.ivanov@cambridgegreys.com" <anton.ivanov@cambridgegreys.com>,
        "richard@nod.at" <richard@nod.at>
Cc:     "linux-um@lists.infradead.org" <linux-um@lists.infradead.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        kernel <kernel@axis.com>
Date:   Wed, 25 Oct 2023 22:02:33 +0200
In-Reply-To: <a2f6e0cc8e498f72dfe1fcd5f2233ada8548a4d1.camel@sipsolutions.net>
References: <20231020-uml-time-backwards-v1-1-90b776fc6dfd@axis.com>
         <4a75c40636be267163dc30b5a6a2442089628e57.camel@axis.com>
         <4dbed8896c94a347dcb58b3a83792c52fdc1c04a.camel@sipsolutions.net>
         <cc7d340e4e06e42cd67b46bfa95da6129b0010cb.camel@axis.com>
         <a2f6e0cc8e498f72dfe1fcd5f2233ada8548a4d1.camel@sipsolutions.net>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.4 (3.48.4-1.fc38) 
MIME-Version: 1.0
Precedence: bulk

On Wed, 2023-10-25 at 21:51 +0200, Johannes Berg wrote:
> On Wed, 2023-10-25 at 11:55 +0000, Vincent Whitchurch wrote:
> > On Mon, 2023-10-23 at 09:33 +0200, Johannes Berg wrote:
> > > Do you have a specific workload that tends to reproduce this?
> >=20
> > I've been seeing it when running roadtest, but it's easily reproducible
> > without that by using the attached config and the following program as
> > init.
> >=20
> >   cp repro.config .config
> >   make ARCH=3Dum olddefconfig all
> >   gcc -Wall -static -o repro repro.c
> >   ./linux time-travel init=3D$PWD/repro rootfstype=3Dhostfs

Ohhh.

Pure "time-travel" is actually something I hardly think about these
days, we mostly use time-travel=3Dinf-cpu (or =3Dext).

That makes sense, here you actually *can* get interrupted. I'll need to
dig into what happens though.

johannes