[PATCH 2/5] arm64, thunder: Add initial dts for Cavium Thunder SoC

[Mark Rutland]

[...]

> >> +     gic0: interrupt-controller at 801000000000 {
> >
> > To make this easier to read, please place a comma between 32-bit
> > portions of the unit address (e.g. here have 8010,00000000).
> 
> Mark, perhaps a dtc or checkpatch.pl check for this?

Sure. Dodgy first atttempt at checkpatch below.

---->8----
diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
index 182be0f..8aee3f5 100755
--- a/scripts/checkpatch.pl
+++ b/scripts/checkpatch.pl
@@ -2136,6 +2136,14 @@ sub process {
                        }
                }
 
+# check for difficult-to-read unit-addresses
+               if (defined $root &&
+                       ($realfile =~ /\.dtsi?$/ && $line =~ /([a-z0-9._\-+]++@([0-9a-f]+))\s*{/gi) &&
+                       (length($2) > 8)) {
+                               WARN("LONG_DT_UNIT_ADDRESS",
+                                    "Consider splitting long unit address \"$2\" with a comma between cells\n" . $herecurr);
+               }
+
 # check we are in a valid source file if not then ignore this hunk
                next if ($realfile !~ /\.(h|c|s|S|pl|sh)$/);
----8<----

It would also be nice to check matching unit-address and reg, but doing
that correctly requires knowing #address-cells, which sounds a little
painful.

I'm not sure where I picked up the comma convention, as it doesn't seem
to be in ePAPR.  It does seem common though, and is my personal
preference:

[mark at leverpostej:~/src/linux]% git grep '@[a-z0-9]\+,[a-b0-9]\+' \                          
	-- arch/arm/boot/dts | wc -l
254
[mark at leverpostej:~/src/linux]% git grep '@[a-z0-9]\+,[a-b0-9]\+' \
	-- arch/powerpc/boot/dts | wc -l
370
[mark at leverpostej:~/src/linux]% git grep '@[a-z0-9]\+,[a-b0-9]\+' \
        -- arch/*/boot/dts | wc -l
631

[...]

> >> +             uaa0: serial at 87e024000000 {
> >> +                     compatible = "arm,pl011", "arm,primecell";
> >> +                     reg = <0x87e0 0x24000000 0x0 0x1000>;
> >> +                     interrupts = <1 21 4>;
> >> +                     clocks = <&refclk50mhz>;
> >> +                     clock-names = "apb_pclk";
> >
> > Is this actually the apb_pclk, or is the the uartclk? I assume it's the
> > latter.
> 
> Shouldn't new bindings have both clocks here? A single clock was a
> mistake I think (mine in fact).

I don't think we fixed it up in the end. It made drivers look a bit messy and
it dropped off my priority queue.

Cheers,
Mark.