Re: [PATCH] arm64: dts: imx8mn: Fix duplicate node name

[Adam Ford]

On Mon, Dec 7, 2020 at 7:36 AM Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
>
> On Mon, Dec 07, 2020 at 02:21:40PM +0100, Krzysztof Kozlowski wrote:
> > On Mon, Dec 07, 2020 at 02:53:24PM +0800, Shengjiu Wang wrote:
> > > Error log:
> > > sysfs: cannot create duplicate filename '/bus/platform/devices/30000000.bus'
> > >
> > > The spba bus name is duplicate with aips bus name.
> > > Refine spba bus name to fix this issue.
> > >
> > > Fixes: 970406eaef3a ("arm64: dts: imx8mn: Enable Asynchronous Sample Rate Converter")
> > > Signed-off-by: Shengjiu Wang <shengjiu.wang@xxxxxxx>
> > > ---
> > >  arch/arm64/boot/dts/freescale/imx8mn.dtsi | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/arch/arm64/boot/dts/freescale/imx8mn.dtsi b/arch/arm64/boot/dts/freescale/imx8mn.dtsi
> > > index fd669c0f3fe5..30762eb4f0a7 100644
> > > --- a/arch/arm64/boot/dts/freescale/imx8mn.dtsi
> > > +++ b/arch/arm64/boot/dts/freescale/imx8mn.dtsi
> > > @@ -246,7 +246,7 @@ aips1: bus@30000000 {
> > >                     #size-cells = <1>;
> > >                     ranges;
> > >
> > > -                   spba: bus@30000000 {
> > > +                   spba: spba-bus@30000000 {

I originally attempted to call it spba-bus@30000000, but I was told to
switch it to 'bus' but all my testing was with the spba-bus name, so I
didn't even notice the conflict.

> >
> > The proper node name is "bus" so basically you introduce wrong name to
> > other problem.  Introducing wrong names at least requires a comment.
>
> I just noticed that my message was barely understandable... so let me
> fix it:
>
> The proper node name is "bus" so basically you introduce wrong name to
> _fix_ other problem.  Introducing wrong names at least requires a comment.
>
> > However the actual problem here is not in node names but in addresses:
> >
> >       aips1: bus@30000000 {
> >               spba: bus@30000000 {
> >
> > You have to devices with the same unit address. How do you share the
> > address space?
> >
> > I think this should be rather fixed.
>
> And again, hungry keyboard ate a letter, so:
>
> You have _two_ devices with the same unit address. How do you share the
> address space?
> I think this should be rather fixed.

I am no expert on this driver, but from what I can tell, the SDMA
driver searches for the memory range assigned to the bus associated
with the spba so the DMA has an idea of which peripheral memory ranges
are shared.
from drivers/dma/imx-sdma.c:

    spba_bus = of_find_compatible_node(NULL, NULL, "fsl,spba-bus");
    ret = of_address_to_resource(spba_bus, 0, &spba_res);
    if (!ret) {
        sdma->spba_start_addr = spba_res.start;
        sdma->spba_end_addr = spba_res.end;
    }

Table 2-3 of the reference manual for the i.MX8M Nano shows the memory
range from 3000_0000 - 300F_FFFF is defined as AIPS-1 (s_b_1, via
SPBA) Glob. Module Enable which implies to me that the SPBA needs to
start at the same place as the AIPS1, but AIPS1 goes from 3000_0000 to
303F_FFFF, so the spba-bus is a subset of the AIPS1.

NXP/Freescale has a variety of SoC's that use the SPBA and in every
instance, they do it this way.  Having the spba-bus with the same
starting address of its parent aips bus isn't unique to this SoC:

imx6q.dtsi
imx6sl.dtsi
imx31.dtsi
imx7s.dtsi
imx6sll.dtsi
imx51.dtsi
imx6qdl.dtsi
imx35.dtsi
imx50.dtsi
imx25.dtsi
imx6sx.dtsi
imx6ul.dtsi
imx53.dtsi

However, in each instance, the bus associated to the SPBA is called
spba-bus and not just 'bus'

adam
>
> Best regards,
> Krzysztof
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel