From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sebastian Hesselbarth Subject: Re: [PATCH 2/2] ARM: dts: mvebu: split SolidRun CuBox into variants Date: Tue, 27 May 2014 23:50:29 +0200 Message-ID: <538508A5.2010809@gmail.com> References: <1401140009-31505-1-git-send-email-sebastian.hesselbarth@gmail.com> <1401140009-31505-2-git-send-email-sebastian.hesselbarth@gmail.com> <20140527161143.GS8664@titan.lakedaemon.net> <5384CB29.1090204@gmail.com> <20140527213549.GU8664@titan.lakedaemon.net> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20140527213549.GU8664-u4khhh1J0LxI1Ri9qeTfzeTW4wlIGRCZ@public.gmane.org> Sender: devicetree-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Jason Cooper Cc: Rob Herring , Pawel Moll , Mark Rutland , Ian Campbell , Kumar Gala , Russell King , Andrew Lunn , Gregory Clement , devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: devicetree@vger.kernel.org On 05/27/2014 11:35 PM, Jason Cooper wrote: > On Tue, May 27, 2014 at 07:28:09PM +0200, Sebastian Hesselbarth wrote: >> On 05/27/2014 06:11 PM, Jason Cooper wrote: >>> On Mon, May 26, 2014 at 11:33:29PM +0200, Sebastian Hesselbarth wrote: >>>> As Mainlining effort for SolidRun CuBox has been carried out on the >>>> Engineering Sample, the board DTS was reflecting this. Actually, >>>> SolidRun CuBox comes in three different variants: Engineering Sample (ES), >>>> production with 1GB RAM (1G), and production with 2GB RAM (2G). >>>> >>>> Therefore, we split the current dove-cubox.dts into a common board include >>>> and one board dts for each of the above variants. >>>> >>>> Signed-off-by: Sebastian Hesselbarth >>>> --- >> [...] >>>> --- >>>> arch/arm/boot/dts/Makefile | 4 +++- >>>> arch/arm/boot/dts/dove-cubox-1g.dts | 17 ++++++++++++++++ >>>> arch/arm/boot/dts/dove-cubox-2g.dts | 17 ++++++++++++++++ >>>> arch/arm/boot/dts/dove-cubox-es.dts | 23 ++++++++++++++++++++++ >>>> .../boot/dts/{dove-cubox.dts => dove-cubox.dtsi} | 17 ---------------- >>>> 5 files changed, 60 insertions(+), 18 deletions(-) >>>> create mode 100644 arch/arm/boot/dts/dove-cubox-1g.dts >>>> create mode 100644 arch/arm/boot/dts/dove-cubox-2g.dts >>>> create mode 100644 arch/arm/boot/dts/dove-cubox-es.dts >>>> rename arch/arm/boot/dts/{dove-cubox.dts => dove-cubox.dtsi} (86%) >>>> >> [...] >>>> diff --git a/arch/arm/boot/dts/dove-cubox-2g.dts b/arch/arm/boot/dts/dove-cubox-2g.dts >>>> new file mode 100644 >>>> index 000000000000..513b6a68eba3 >>>> --- /dev/null >>>> +++ b/arch/arm/boot/dts/dove-cubox-2g.dts >>>> @@ -0,0 +1,17 @@ >>>> +/dts-v1/; >>>> + >>>> +#include "dove-cubox.dtsi" >>>> + >>>> +/ { >>>> + model = "SolidRun CuBox (2G)"; >>>> + compatible = "solidrun,cubox-2g", "solidrun,cubox", "marvell,dove"; >>>> + >>>> + memory { >>>> + device_type = "memory"; >>>> + reg = <0x00000000 0x80000000>; >>> >>> Do you anticipate any other differences between the 1G and the 2G? >>> Otherwise, I'm inclined to just have a "solidrun,cubox". The bootloader >>> should be setting the amount of RAM at boottime anyway. >> >> Given the minor differences between ES and production, instead of >> >> dove-cubox-common.dtsi >> +--> dove-cubox.dts (production) >> +--> dove-cubos-es.dts (engineering sample) >> >> we could also just have an "overlay" for the ES like >> >> dove-cubox.dts (production) >> +--> dove-cubox-es.dts (engineering sample) >> >> It is not used commonly until now, maybe just a matter of taste. >> >> Is there any version you prefer? > > iiuc, overlays were intended for daughterboard (capes, etc) specific Oh, ok. I guess "overlay" was misleading here. I did not mean dynamic loading/unloading of dtb but including a dts from another dts. > info. It may be useful here, but I'd like to hear from the DT > maintainers how they want it used. eg: most popular first, like you > have it, or oldest first > > dove-cubox-es.dts > +--> dove-cubox.dts In the cubox case, this is not possible. ES has a misrouted card-detection for sdhci, this requires an additional property. There is no way to remove a property once it is written down in any of the files included. But you know about that already. > There's also what to do with the older files using #include... > > In short, I'd prefer to stick to the old method until we have a good > reason to move to overlays and a recommended way to execute that.* Ok, the old method is straight forward and I keep that in mind. I'll send a v2 of this using the approach we just talked about to eliminate any misinterpretations. Just have a look and feel free to request an "old-method" v3 immediately :P Sebastian -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html