| Message ID | 20190920153622.14616-1-robh@kernel.org (mailing list archive) |
|---|---|
| State | Mainlined |
| Commit | 40d39c1a7cacd8db29f298f14e4e0617e8ca65bd |
| Headers | show |
| Series | [RFC] arm: dts: integratorap: Remove top level dma-ranges | expand |
On Fri, Sep 20, 2019 at 5:36 PM Rob Herring <robh@kernel.org> wrote: > 'dma-ranges' at the top level doesn't make sense. 'dma-ranges' implies > there is a parent bus node with '#address-cells' and '#size-cells' which > is impossible here. > > Likely this translation needs to be moved down to sub-nodes that need > it. > > Cc: Linus Walleij <linus.walleij@linaro.org> > Cc: linux-arm-kernel@lists.infradead.org > Signed-off-by: Rob Herring <robh@kernel.org> Patch applied! I'll get this to ARM SoC with some other stuff. Yours, Linus Walleij
diff --git a/arch/arm/boot/dts/integratorap.dts b/arch/arm/boot/dts/integratorap.dts index 94d2ff9836d0..198d66181c50 100644 --- a/arch/arm/boot/dts/integratorap.dts +++ b/arch/arm/boot/dts/integratorap.dts @@ -9,7 +9,6 @@ / { model = "ARM Integrator/AP"; compatible = "arm,integrator-ap"; - dma-ranges = <0x80000000 0x0 0x80000000>; cpus { #address-cells = <1>;
'dma-ranges' at the top level doesn't make sense. 'dma-ranges' implies there is a parent bus node with '#address-cells' and '#size-cells' which is impossible here. Likely this translation needs to be moved down to sub-nodes that need it. Cc: Linus Walleij <linus.walleij@linaro.org> Cc: linux-arm-kernel@lists.infradead.org Signed-off-by: Rob Herring <robh@kernel.org> --- This probably worked due to bugs in dma-ranges address translations (or lack of). AFAICT, only PCI needs dma-ranges. The TRM says inbound PCI addresses are 1:1. So a translation from child address 0x80000000 to cpu address 0x0 seems wrong. The v3 PCI driver also fails to look at parent nodes for dma-ranges. arch/arm/boot/dts/integratorap.dts | 1 - 1 file changed, 1 deletion(-)