linux-kernel - Re: [PATCH v2 2/7] drm: ci: Force db410c to host mode

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, 5 Sep 2023 13:00:09 +0200
From:   Maxime Ripard <mripard@...nel.org>
To:     Vignesh Raman <vignesh.raman@...labora.com>
Cc:     Dmitry Baryshkov <dmitry.baryshkov@...aro.org>,
        dri-devel@...ts.freedesktop.org, helen.koike@...labora.com,
        guilherme.gallo@...labora.com, sergi.blanch.torne@...labora.com,
        david.heidelberg@...labora.com, daniels@...labora.com,
        gustavo.padovan@...labora.com, emma@...olt.net,
        robclark@...edesktop.org, robdclark@...gle.com, anholt@...gle.com,
        robdclark@...il.com, airlied@...il.com, daniel@...ll.ch,
        jani.nikula@...ux.intel.com, linux-kernel@...r.kernel.org,
        linux-mediatek@...ts.infradead.org,
        virtualization@...ts.linux-foundation.org,
        linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH v2 2/7] drm: ci: Force db410c to host mode

On Tue, Sep 05, 2023 at 03:39:33PM +0530, Vignesh Raman wrote:
> Hi Dmitry, Maxime,
> 
> On 05/09/23 14:13, Maxime Ripard wrote:
> > Hi,
> > 
> > On Mon, Sep 04, 2023 at 07:59:26PM +0300, Dmitry Baryshkov wrote:
> > > On Mon, 4 Sept 2023 at 19:16, Vignesh Raman <vignesh.raman@...labora.com> wrote:
> > > > 
> > > > Force db410c to host mode to fix network issue which results in failure
> > > > to mount root fs via NFS.
> > > > See https://gitlab.freedesktop.org/gfx-ci/linux/-/commit/cb72a629b8c15c80a54dda510743cefd1c4b65b8
> > > > 
> > > > Use fdtoverlay command to merge base device tree with an overlay
> > > > which contains the fix for USB controllers to work in host mode.
> > > > 
> > > > Signed-off-by: Vignesh Raman <vignesh.raman@...labora.com>
> > > > ---
> > > > 
> > > > v2:
> > > >    - Use fdtoverlay command to merge overlay dtbo with the base dtb instead of modifying the kernel sources
> > > > 
> > > > ---
> > > >   drivers/gpu/drm/ci/build.sh                         |  5 +++++
> > > >   .../gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dts  | 13 +++++++++++++
> > > >   2 files changed, 18 insertions(+)
> > > >   create mode 100644 drivers/gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dts
> > > > 
> > > > diff --git a/drivers/gpu/drm/ci/build.sh b/drivers/gpu/drm/ci/build.sh
> > > > index 7b014287a041..92ffd98cd09e 100644
> > > > --- a/drivers/gpu/drm/ci/build.sh
> > > > +++ b/drivers/gpu/drm/ci/build.sh
> > > > @@ -92,6 +92,11 @@ done
> > > > 
> > > >   if [[ -n ${DEVICE_TREES} ]]; then
> > > >       make dtbs
> > > > +    if [[ -e arch/arm64/boot/dts/qcom/apq8016-sbc.dtb ]]; then
> > > > +        dtc -@ -I dts -O dtb -o drivers/gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dtbo drivers/gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dts
> > > > +        fdtoverlay -i arch/arm64/boot/dts/qcom/apq8016-sbc.dtb -o arch/arm64/boot/dts/qcom/apq8016-sbc-overlay.dtb drivers/gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dtbo
> > > > +        mv arch/arm64/boot/dts/qcom/apq8016-sbc-overlay.dtb arch/arm64/boot/dts/qcom/apq8016-sbc.dtb
> > > > +    fi
> > > >       cp ${DEVICE_TREES} /lava-files/.
> > > >   fi
> > > > 
> > > > diff --git a/drivers/gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dts b/drivers/gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dts
> > > > new file mode 100644
> > > > index 000000000000..57b7604f1c23
> > > > --- /dev/null
> > > > +++ b/drivers/gpu/drm/ci/dt-overlays/apq8016-sbc-overlay.dts
> > > > @@ -0,0 +1,13 @@
> > > > +/dts-v1/;
> > > > +/plugin/;
> > > > +
> > > > +/ {
> > > > +    fragment@0 {
> > > > +        target-path = "/soc@0";
> > > > +        __overlay__ {
> > > > +            usb@...9000 {
> > > > +                dr_mode = "host";
> > > > +            };
> > > > +        };
> > > > +    };
> > > > +};
> > > > --
> > > > 2.40.1
> > > 
> > > Can we use normal dtso syntax here instead of defining fragments manually?
> > 
> > What Dmitry is hinting about is to use the "Sugar Syntax". There a good documentation here:
> > https://source.android.com/docs/core/architecture/dto/syntax
> 
> 
> With the below DTO syntax,
> /dts-v1/;
> /plugin/;
> 
> &usb {
>   usb@...9000 {
>     dr_mode = "host";
>   };
> };
> 
> Decoded dtbo file is,
> /dts-v1/;
> 
> / {
> 
> 	fragment@0 {
> 		target = <0xffffffff>;
> 
> 		__overlay__ {
> 
> 			usb@...9000 {
> 				dr_mode = "host";
> 			};
> 		};
> 	};
> 
> 	__fixups__ {
> 		usb = "/fragment@0:target:0";
> 	};
> };
> 
> With the previous fix using fragment we get,
> / {
> 
> 	fragment@0 {
> 		target-path	 = "/soc@0";
> 
> 		__overlay__ {
> 
> 			usb@...9000 {
> 				dr_mode = "host";
> 			};
> 		};
> 	};
> };
> 
> Decoded apq8016-sbc.dtb file with the fix (setting dr_mode to host) is,
> /dts-v1/;
> / {	
> 	soc@0 {
> 		usb@...9000 {
> 			dr_mode = "host";
> 		};	
> 	};
> };
> 
> How can set the target to "soc@0" using the DTO syntax?

To strictly answer your question, that would be something like

&{/soc@0} {
	usb@...9000 {
		dr_mode = "host";
	};
};

You can simplify this further however by doing:


&{/soc@...sb@...9000} {
	dr_mode = "host";
};

Also, that node actually has a label ("usb"), defined here:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm64/boot/dts/qcom/msm8916.dtsi#n2322

So you can end up with

&usb {
	dr_mode = "host";
};

All of them should be equivalent to the one you had in your patch.

Maxime