linux-kernel - Re: [PATCH] drm/msm: Disable frequency clamping on a630

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, 7 Sep 2021 07:25:07 -0700
From:   Rob Clark <robdclark@...il.com>
To:     Amit Pundir <amit.pundir@...aro.org>
Cc:     John Stultz <john.stultz@...aro.org>,
        Caleb Connolly <caleb.connolly@...aro.org>,
        Rob Clark <robdclark@...omium.org>,
        freedreno <freedreno@...ts.freedesktop.org>,
        Sai Prakash Ranjan <saiprakash.ranjan@...eaurora.org>,
        Jonathan Marek <jonathan@...ek.ca>,
        David Airlie <airlied@...ux.ie>,
        linux-arm-msm <linux-arm-msm@...r.kernel.org>,
        Akhil P Oommen <akhilpo@...eaurora.org>,
        dri-devel <dri-devel@...ts.freedesktop.org>,
        Jordan Crouse <jordan@...micpenguin.net>,
        Stephen Boyd <sboyd@...nel.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        Sean Paul <sean@...rly.run>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] drm/msm: Disable frequency clamping on a630

On Tue, Sep 7, 2021 at 1:25 AM Amit Pundir <amit.pundir@...aro.org> wrote:
>
> On Tue, 7 Sept 2021 at 07:11, Rob Clark <robdclark@...il.com> wrote:
> >
> > One thing I thought of, which would be worth ruling out, is whether
> > this issue only occurs with freq changes immediately after resuming
> > the GPU, vs freq changes in general.  Could you try the below patch.
> > And if it "fixes" the issue, then try reducing the delay until you
> > start seeing GPU hangs again.
>
> It doesn't fix the crash and I can still reproduce it
> https://pastebin.com/raw/bxK4mAhB

Ok, thanks for confirming.  That implies the limitation is about
changing freq in general, rather than immediately after resume.. the
latter could be a new scenario after 9bc95570175a.  But this confirms
that this is an issue that has been there all along.

I'm still not quite sure what the correct fix is

BR,
-R

>
> >
> > ----------
> > diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> > b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> > index 40c9fef457a4..278b85207ea3 100644
> > --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> > +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> > @@ -1513,6 +1513,8 @@ static int a6xx_pm_resume(struct msm_gpu *gpu)
> >   if (ret)
> >   return ret;
> >
> > + msleep(5);
> > +
> >   msm_devfreq_resume(gpu);
> >
> >   a6xx_llc_activate(a6xx_gpu);
> > ----------
> >
> > BR,
> > -R