linux-kernel - Re: [PATCH] drm/msm/mdp5: Use drm_device for creating gem address space

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

Date:   Wed, 3 Jul 2019 09:09:57 -0600
From:   Jeffrey Hugo <jeffrey.l.hugo@...il.com>
To:     Rob Clark <robdclark@...il.com>
Cc:     Bjorn Andersson <bjorn.andersson@...aro.org>,
        Sean Paul <sean@...rly.run>, David Airlie <airlied@...ux.ie>,
        Daniel Vetter <daniel@...ll.ch>,
        linux-arm-msm <linux-arm-msm@...r.kernel.org>,
        dri-devel <dri-devel@...ts.freedesktop.org>,
        freedreno <freedreno@...ts.freedesktop.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] drm/msm/mdp5: Use drm_device for creating gem address space

On Wed, Jul 3, 2019 at 6:25 AM Rob Clark <robdclark@...il.com> wrote:
>
> On Tue, Jul 2, 2019 at 9:08 PM Bjorn Andersson
> <bjorn.andersson@...aro.org> wrote:
> >
> > On Mon 01 Jul 10:39 PDT 2019, Jeffrey Hugo wrote:
> >
> > > Creating the msm gem address space requires a reference to the dev where
> > > the iommu is located.  The driver currently assumes this is the same as
> > > the platform device, which breaks when the iommu is outside of the
> > > platform device.  Use the drm_device instead, which happens to always have
> > > a reference to the proper device.
> > >
> > > Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@...il.com>
> >
> > Sorry, but on db820c this patch results in:
> >
> > [   64.803263] msm_mdp 901000.mdp: [drm:mdp5_kms_init [msm]] *ERROR* failed to attach iommu: -19
> >
> > Followed by 3 oopses as we're trying to fail the initialization.
>
> yeah, that is kinda what I suspected would happen.  I guess to deal
> with how things are hooked up on 8998, perhaps the best thing is to
> first try &pdev->dev, and then if that fails try dev->dev

Thanks for the test feedback Bjorn.  Its unfortunate this solution
didn't work as I expected.  I'll give Rob's suggestion a shot and spin
another version.