| Message ID | 1347266386-16229-7-git-send-email-sebastian.hesselbarth@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 09/10/2012 02:39 AM, Sebastian Hesselbarth wrote: > From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> > > The Armada 370 and XP SoCs have configurable muxing for a certain > number of their pins, controlled through a pinctrl driver. Hmmm. I'd be tempted just to put the entire node definition there; putting in a .dtsi file just to share the reg property doesn't seem especially useful. > The 'compatible' property is defined in the SoC-specific .dtsi files, > since the compatible string identifies the number of pins and other > SoC-specific properties. > diff --git a/arch/arm/boot/dts/armada-370-xp.dtsi b/arch/arm/boot/dts/armada-370-xp.dtsi > + pinctrl@d0018000 { If this is the only pinctrl instance, you'd typically name the node just "pinctrl", since the "@d0018000" isn't needed to get unique node names. > + reg = <0xd0018000 0x38>; > + #address-cells = <1>; > + #size-cells = <1>; > + ranges; What is "ranges" for; this isn't a memory-mapped bus, right? > + }; > }; > };
Le Tue, 11 Sep 2012 16:23:19 -0600, Stephen Warren <swarren@wwwdotorg.org> a écrit : > On 09/10/2012 02:39 AM, Sebastian Hesselbarth wrote: > > From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> > > > > The Armada 370 and XP SoCs have configurable muxing for a certain > > number of their pins, controlled through a pinctrl driver. > > Hmmm. I'd be tempted just to put the entire node definition there; > putting in a .dtsi file just to share the reg property doesn't seem > especially useful. When you say "here" you're mentioning the SoC-specific .dtsi files (i.e the ones in PATCH 7/9 and PATCH 8/9), correct? > > The 'compatible' property is defined in the SoC-specific .dtsi files, > > since the compatible string identifies the number of pins and other > > SoC-specific properties. > > > diff --git a/arch/arm/boot/dts/armada-370-xp.dtsi b/arch/arm/boot/dts/armada-370-xp.dtsi > > > + pinctrl@d0018000 { > > If this is the only pinctrl instance, you'd typically name the node just > "pinctrl", since the "@d0018000" isn't needed to get unique node names. Ack. > > + reg = <0xd0018000 0x38>; > > + #address-cells = <1>; > > + #size-cells = <1>; > > + ranges; > > What is "ranges" for; this isn't a memory-mapped bus, right? Ack. Best regards, Thomas
On 09/12/2012 12:56 AM, Thomas Petazzoni wrote: > Le Tue, 11 Sep 2012 16:23:19 -0600, > Stephen Warren <swarren@wwwdotorg.org> a écrit : > >> On 09/10/2012 02:39 AM, Sebastian Hesselbarth wrote: >>> From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> >>> >>> The Armada 370 and XP SoCs have configurable muxing for a certain >>> number of their pins, controlled through a pinctrl driver. >> >> Hmmm. I'd be tempted just to put the entire node definition there; >> putting in a .dtsi file just to share the reg property doesn't seem >> especially useful. > > When you say "here" you're mentioning the SoC-specific .dtsi files (i.e (s/here/there I assume). > the ones in PATCH 7/9 and PATCH 8/9), correct? Yes, I believe so.
diff --git a/arch/arm/boot/dts/armada-370-xp.dtsi b/arch/arm/boot/dts/armada-370-xp.dtsi index 16cc82c..ff1c7a6 100644 --- a/arch/arm/boot/dts/armada-370-xp.dtsi +++ b/arch/arm/boot/dts/armada-370-xp.dtsi @@ -68,6 +68,13 @@ compatible = "marvell,armada-addr-decoding-controller"; reg = <0xd0020000 0x258>; }; + + pinctrl@d0018000 { + reg = <0xd0018000 0x38>; + #address-cells = <1>; + #size-cells = <1>; + ranges; + }; }; };