2014-08-18 15:27 GMT+02:00 Carlo Caione [off-list ref]:

On Sun, Aug 17, 2014 at 04:21:15PM +0200, Maxime Ripard wrote:

quoted

Hi Carlo,

Hi Maxime,
thank you for reviewing also these patches, I'll keep you in CC for the
next revisions if you are interested.

quoted

quoted

+static __init void meson_init_machine_devicetree(void)
+{
+   of_platform_populate(NULL, of_default_bus_match_table, NULL, NULL);
+}

This is the default, you can just remove it

Agree

quoted

quoted

+static const char * const m6_common_board_compat[] = {
+   "amlogic,8726_mx",
+   "amlogic,8726_mxs",
+   "amlogic,8726_mxl",
+   "amlogic,meson6",

Why are all those compatibles needed? Usually, you add a single one
per SoC (which would be the last in your case I guess.

They are taken from the weird DTS in the original Amlogic sources but I
guess you are right. I am actually more inclined to just leave "8726_mx"
and "meson6" since online you can find equally both the versions for
exactly the same SoCs.

quoted

quoted

+   NULL,
+};
+
+DT_MACHINE_START(AML8726_MX, "Amlogic Meson6 platform")
+   .init_machine   = meson_init_machine_devicetree,

And since you don't need the init machine, you can just use the
generic machine support. I'm not sure what's been decided on this,
should we remove such empty machines?

I can get rid of the .init_machine but what about the .dt_compat field?

AFIAK there is no consensus about that.
For now I would leave the .dt_compat field. There are proposals which
try to get rid of the name_str
in the DT_MACHINE_START [0], but from what I know they are not yet merged.

Cheers,
Matthias

[0] http://lists.infradead.org/pipermail/linux-arm-kernel/2013-November/208878.html

Thanks,

--
Carlo Caione

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

-- 
motzblog.wordpress.com