[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <trinity-d7a098bf-51a2-4579-93c9-963826ed2895-1576677243636@3c-app-gmx-bap62>
Date: Wed, 18 Dec 2019 14:54:03 +0100
From: "Frank Wunderlich" <frank-w@...lic-files.de>
To: "Masahiro Yamada" <masahiroy@...nel.org>
Cc: "Linux Kbuild mailing list" <linux-kbuild@...r.kernel.org>,
"Michal Marek" <michal.lkml@...kovi.net>,
"Linux Kernel Mailing List" <linux-kernel@...r.kernel.org>,
"Rob Herring" <robh+dt@...nel.org>,
"Frank Rowand" <frowand.list@...il.com>,
DTML <devicetree@...r.kernel.org>
Subject: Aw: Re: [PATCH] kbuild: Add DTC_CPP_FLAGS
Hi,
thank you for your answer... i waited for this argument ;) so here an example (don't wanted to add in patch itself):
i have 2 different Bananapi-r64 boards with different switch-chips (v0.x rtl8367,v1.x xmt7531). for 1 of the switches (mt7531) there are 2 drivers (phy-driver + dsa driver) available and so i have 3 different configs in DTS.
Imho dtsi should contain only config for 1 chip, but the board has only 1 dts. Till now i have copied the whole dts twice, so i have 3 independent dts. if i Patch any other in one like for bluetooth, wifi, which is same on each board i have to do it on all 3 dts files.
so i tried to make the switch-part conditionally, include the right dtsi for switch to 1 bpi-r64 dts based on -D param passed to the build-chain.
#ifdef __SWITCH_RTL8367__
#include "rtl8367.dtsi"
#endif
#ifdef __SWITCH_MT7531_DSA__
#include "mt7531.dtsi"
#endif
imho i can/should not include a dtsi to a dtsi and no dts to dts, so i cannot have a small dts (with the switch-config) including the whole r64 board. Am i right?
and this change should be less interferring than other ideas i'm thinking about.
Another example: I had tried adding the include/generated/autoconf.h to dtsi (same reason) which needs a change in scripts/dtc/include-prefixes/ (add symlink) or an odd include in the dtsi (#include "../../../../../include/generated/autoconf.h"). Both ways not nice...
regards Frank
> Gesendet: Mittwoch, 18. Dezember 2019 um 14:31 Uhr
> Von: "Masahiro Yamada" <masahiroy@...nel.org>
> I think this is questionable
> since DT is supposed to describe hardware.
> Does it depends on #ifdef or some external defines?
>
> If you want to add variations,
> you can add a new DTS which includes other DTS(I) files.
Powered by blists - more mailing lists