RE: [PATCH 1/2] power: act8945a: add charger driver for ACT8945A

Hello Krzysztof,

-----Original Message-----
From: Krzysztof Kozlowski [mailto:k.kozlowski@samsung.com]
Sent: 2016年1月8日 15:03
To: Yang, Wenyou <redacted>
Cc: Sebastian Reichel <sre@kernel.org>; Dmitry Eremin-Solenikov
[off-list ref]; David Woodhouse [off-list ref]; Rob
Herring [off-list ref]; Pawel Moll [off-list ref]; Mark
Rutland [off-list ref]; Ian Campbell [off-list ref];
Kumar Gala [off-list ref]; Ferre, Nicolas
[off-list ref]; linux-arm-kernel@lists.infradead.org; linux-
kernel@vger.kernel.org; linux-pm@vger.kernel.org
Subject: Re: [PATCH 1/2] power: act8945a: add charger driver for ACT8945A

On 08.01.2016 15:43, Yang, Wenyou wrote:

[...]

quoted

quoted

quoted

+static struct platform_driver act8945a_charger_driver = {
+       .driver = {
+               .name = "act8945a-charger",
+       },
+       .probe  = act8945a_charger_probe,
+       .remove = act8945a_charger_remove, };
+
+static int __init act8945a_charger_init(void) {
+       return platform_driver_register(&act8945a_charger_driver);
+}
+subsys_initcall(act8945a_charger_init);

Why subsys_initcall? This should be regular module_platform_driver.

I want it to register early enough as MFD.  Maybe I am wrong.

No, you shouldn't manually order the probing by initcalls. This should be
registered as usual and, if needed, support deferred probing. In this case I even
can't find any reason to register it earlier than usual.

OK, I will change it.

Thanks.

Best Regards,
Wenyou Yang