From mboxrd@z Thu Jan 1 00:00:00 1970 From: xuejiancheng Subject: Re: [PATCH v2 4/9] ARM: dts: add dts files for hi3519-demb board Date: Mon, 7 Dec 2015 14:37:06 +0800 Message-ID: <56652912.80308@huawei.com> References: <1449110668-23647-1-git-send-email-xuejiancheng@huawei.com> <1728470.0OiiXMcl88@wuerfel> <5660FA2E.6040601@huawei.com> <3047121.2z9SCS5Au2@wuerfel> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <3047121.2z9SCS5Au2@wuerfel> Sender: linux-kernel-owner@vger.kernel.org To: Arnd Bergmann Cc: linux-arm-kernel@lists.infradead.org, robh+dt@kernel.org, pawel.moll@arm.com, mark.rutland@arm.com, ijc+devicetree@hellion.org.uk, galak@codeaurora.org, linux@arm.linux.org.uk, khilman@linaro.org, olof@lixom.net, xuwei5@hisilicon.com, haojian.zhuang@linaro.org, zhangfei.gao@linaro.org, bintian.wang@huawei.com, suwenping@hisilicon.com, devicetree@vger.kernel.org, linux-kernel@vger.kernel.org, yanhaifeng@hisilicon.com, gaofei@hisilicon.com, ml.yang@hisilicon.com, yanghongwei@hisilicon.com List-Id: devicetree@vger.kernel.org On 2015/12/4 18:49, Arnd Bergmann wrote: > On Friday 04 December 2015 10:27:58 xuejiancheng wrote: >>> >>>> + sysctrl: system-controller@12020000 { >>>> + compatible =3D "hisilicon,sysctrl"; >>>> + reg =3D <0x12020000 0x1000>; >>>> + reboot-offset =3D <0x4>; >>>> + }; >>> >>> Is this one identical to the one in hip04? >>> >>> If not, pick a new unique compatible string >> >> Yes. It's compatible with the one in hip04. >=20 > Ok, we should add a compatible string for that then, as the hip04 app= arently > has a slightly different layout from hip01. >=20 > Can you add a separate patch to clarify the existing hisilicon,sysctr= l nodes? >=20 > It look like hi3620 accidentally uses the same string as hip04 alread= y > but is actually a bit different. >=20 > Maybe split out the sysctrl binding from > Documentation/devicetree/bindings/arm/hisilicon/hisilicon.txt, as it = has > you already have a couple of those, and it's not clear how they relat= e > to one another. >=20 > If we introduce a string for all hip04 compatible sysctrl devices, we= should > document that and use it consistently, so hi3519 becomes >=20 > compatible =3D "hisilicon,hi3519-sysctrl", "hisilicon,hip04-sysctrl"= , "hisilicon,sysctrl"; >=20 > but I'd clarify in the binding documentation that "hisilicon,sysctrl"= should > only be used for hip04 and hi3519 but not the others. >=20 > As this seems to be a standard part, we can also think about making a > high-level driver for in in drivers/soc rather than relying on the sy= scon > driver which we tend to use more for one-off devices with random regi= ster > layouts. >=20 Sorry. I didn't understand your meaning well and maybe I gave you a = wrong description. Please allow me to clarify it again. The "sysctrl" nodes here is just used for the "reboot" function. It = is corresponding to the driver "drivers/power/reset/hisi-reboot.c". The compatible string i= n the driver is "hisilicon,sysctrl". The layout of this block is also different from the one in HiP04. >>>> + >>>> + crg: crg@12010000 { >>>> + compatible =3D "hisilicon,hi3519-crg"; >>> >>> >>> what is a "crg"? Is there a standard name for these? >> >> The "crg" means Clock and Reset Generator. It's a block which suppli= es clock >> and reset signals for other modules in the chip. I used "hi3519-cloc= k" >> in last version patch. Rob Herring suggested that it's better to use= "hi3519-crg" >> as the compatible string if it's a whole block. >> >> what about writing like this=EF=BC=9F >> >> crg: clock-reset-controller@12010000 { >> compatible =3D "hisilicon,hi3519-crg"; >> } >> >=20 > Ok, that's better. >=20 > Arnd >=20 > . >=20