regulator node names and unit-addresses? (Re: [PATCH v2 1/5] ARM: shmobile: kzm9d: Remove spurious regulator base addresses)

Hi Simon,

(pulling in regulator people and DT list)

On Wed, Sep 24, 2014 at 3:13 AM, Simon Horman [off-list ref] wrote:

On Wed, Sep 17, 2014 at 09:33:49AM +0200, Geert Uytterhoeven wrote:

quoted

On Wed, Sep 17, 2014 at 2:48 AM, Simon Horman
[off-list ref] wrote:

quoted

As there is no reg property the regulator nodes should
not include @...

You missed an important part of my comment:

"However, in the absence of a unit-address, the node name must be unique."

quoted

Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Simon Horman <redacted>

---
v2
* First post
---
 arch/arm/boot/dts/emev2-kzm9d.dts | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/arm/boot/dts/emev2-kzm9d.dts b/arch/arm/boot/dts/emev2-kzm9d.dts
index 50ccd15..f27fcf0 100644
--- a/arch/arm/boot/dts/emev2-kzm9d.dts
+++ b/arch/arm/boot/dts/emev2-kzm9d.dts

@@ -27,7 +27,7 @@
                bootargs = "console=ttyS1,115200n81 ignore_loglevel root=/dev/nfs ip=dhcp";
        };

-       reg_1p8v: regulator at 0 {
+       reg_1p8v: regulator {
                compatible = "regulator-fixed";
                regulator-name = "fixed-1.8V";
                regulator-min-microvolt = <1800000>;

@@ -36,7 +36,7 @@
                regulator-boot-on;
        };

-       reg_3p3v: regulator at 1 {
+       reg_3p3v: regulator {
                compatible = "regulator-fixed";
                regulator-name = "fixed-3.3V";
                regulator-min-microvolt = <3300000>;

Now there are two "regulator" nodes.

Dtc will not complain, but will assume the second contains overrides for
first. "vddvario-supply" and "vdd33a-supply" of node "lan9220 at 20000000"
will both the point to the same node, and the network may not work.

You can see this yourself running

    dtc -I dtb -O dts arch/arm/boot/dts/emev2-kzm9d.dtb

What should we call the nodes?

"regulator0" and "regulator1"?
"regulator1v8" and "regulator3v3"?

Documentation/devicetree/bindings/regulator/regulator.txt doesn't mention
node names or (presence/lack of) unit-adresses, but the examples do have
unit-addresses.

Is this another case where we don't follow the "no reg, no unit-address"
ePAPR rule?

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

On Wed, Sep 24, 2014 at 09:02:23AM +0200, Geert Uytterhoeven wrote:

On Wed, Sep 24, 2014 at 3:13 AM, Simon Horman [off-list ref] wrote:

quoted

What should we call the nodes?

"regulator0" and "regulator1"?
"regulator1v8" and "regulator3v3"?

Documentation/devicetree/bindings/regulator/regulator.txt doesn't mention
node names or (presence/lack of) unit-adresses, but the examples do have
unit-addresses.

Is this another case where we don't follow the "no reg, no unit-address"
ePAPR rule?

As far as the regulator API is concerned the node name is completly
immaterial and all this stuff is just verbiage we're forced to include.
As far as I can tell the requirement that node names be in the form
"regulator" or whatever is just another thing that wasn't terribly well
thought through in ePAPR, they were trying to do classes I think.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20140924/ab0007ee/attachment.sig>