linux-kernel - Re: [PATCH] rpmsg: virtio_rpmsg_bus: use dev_warn

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: <CA+U=DsrEuOdzn=YhQ_AwTFyaJesqHNC9vycDa9TyQMSaEwaEWw@mail.gmail.com>
Date:   Mon, 11 Oct 2021 11:59:01 +0300
From:   Alexandru Ardelean <ardeleanalex@...il.com>
To:     Bjorn Andersson <bjorn.andersson@...aro.org>
Cc:     linux-remoteproc@...r.kernel.org,
        LKML <linux-kernel@...r.kernel.org>,
        Ohad Ben Cohen <ohad@...ery.com>,
        Mathieu Poirier <mathieu.poirier@...aro.org>
Subject: Re: [PATCH] rpmsg: virtio_rpmsg_bus: use dev_warn_ratelimited for msg
 with no recipient

On Wed, Sep 29, 2021 at 12:40 PM Alexandru Ardelean
<ardeleanalex@...il.com> wrote:
>
> On Wed, Sep 29, 2021 at 12:33 AM Bjorn Andersson
> <bjorn.andersson@...aro.org> wrote:
> >
> > On Tue 28 Sep 08:29 CDT 2021, Alexandru Ardelean wrote:
> >
> > > From: Alexandru Ardelean <ardeleanalex@...il.com>
> > >
> > > Even though it may be user-space's fault for this error (some application
> > > terminated or crashed without cleaning up it's endpoint), the rpmsg
> > > communication should not overflow the syslog with too many messages.
> > >
> > > A dev_warn_ratelimited() seems like a good alternative in case this can
> > > occur.
> > >
> >
> > Is there anything a user could/should do when they see this entry in
> > their log?
>
> Not really, no.
> The userspace application would need to respawn, or some systemd (or
> similar process manager) would need to respawn the application it
> should recover the state, and communication should resume normally.
> I think this message is good mostly as informative.
>
> >
> > It doesn't look very actionable to me, should we perhaps degrade it
> > further to just a dev_dbg()?
>
> It's not actionable unfortunately.
> But I feel it is useful to have this message, until the application recovers.
> Mostly to be informative.
> A more robust mechanism would be to setup some counters, where we
> count the number of missed messages.
> And then access this counter via sysfs or something.
>
> The problem is that a high-rate of dev_warn() (during failure),
> temporarily increases system CPU usage & load-average, making the
> recovery a bit slower, because systemd-journald is processing these
> messages from the kernel.
> So, dev_dbg() would definitely help, but would also require us to bump
> the system log-level to see the messages.
> And if they occur and we don't see them, it causes more questions and
> debugging, because people won't know for sure what the issue is.
>
> Ultimately, dev_dbg() or dev_warn_rate_limited() are both fine.
> The goal is to avoid the temporary increase in CPU load.
> I just wanted to state my arguments for dev_warn_ratelimite() :)
>

Ping on this :)

> Thank you
> Alex
>
> >
> > Regards,
> > Bjorn
> >
> > > Signed-off-by: Alexandru Ardelean <ardeleanalex@...il.com>
> > > ---
> > >  drivers/rpmsg/virtio_rpmsg_bus.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/rpmsg/virtio_rpmsg_bus.c b/drivers/rpmsg/virtio_rpmsg_bus.c
> > > index 8e49a3bacfc7..546f0fb66f1d 100644
> > > --- a/drivers/rpmsg/virtio_rpmsg_bus.c
> > > +++ b/drivers/rpmsg/virtio_rpmsg_bus.c
> > > @@ -749,7 +749,7 @@ static int rpmsg_recv_single(struct virtproc_info *vrp, struct device *dev,
> > >               /* farewell, ept, we don't need you anymore */
> > >               kref_put(&ept->refcount, __ept_release);
> > >       } else
> > > -             dev_warn(dev, "msg received with no recipient\n");
> > > +             dev_warn_ratelimited(dev, "msg received with no recipient\n");
> > >
> > >       /* publish the real size of the buffer */
> > >       rpmsg_sg_init(&sg, msg, vrp->buf_size);
> > > --
> > > 2.31.1
> > >