Hi > Gesendet: Montag, 11. September 2023 um 23:06 Uhr > Von: "Rob Herring" <robh+dt@xxxxxxxxxx> > An: frank-w@xxxxxxxxxxxxxxx > Cc: "Frank Wunderlich" <linux@xxxxxxxxx>, linux-mediatek@xxxxxxxxxxxxxxxxxxx, "Matthias Brugger" <matthias.bgg@xxxxxxxxx>, "AngeloGioacchino Del Regno" <angelogioacchino.delregno@xxxxxxxxxxxxx>, "Krzysztof Kozlowski" <krzysztof.kozlowski+dt@xxxxxxxxxx>, "Conor Dooley" <conor+dt@xxxxxxxxxx>, linux-kernel@xxxxxxxxxxxxxxx, linux-arm-kernel@xxxxxxxxxxxxxxxxxxx, devicetree@xxxxxxxxxxxxxxx > Betreff: Re: [PATCH v1] arm64: dts: mt7986: add overlay for SATA power socket on BPI-R3 > > On Mon, Sep 11, 2023 at 3:11 PM Frank Wunderlich > <frank-w@xxxxxxxxxxxxxxx> wrote: > > > > Am 11. September 2023 21:51:42 MESZ schrieb Rob Herring <robh+dt@xxxxxxxxxx>: > > >On Fri, Sep 1, 2023 at 2:27 AM Frank Wunderlich <linux@xxxxxxxxx> wrote: > > >> > > >> From: Frank Wunderlich <frank-w@xxxxxxxxxxxxxxx> > > >> > > >> Bananapi R3 has a Power socket entended for using external SATA drives. > > >> This Socket is off by default but can be switched with gpio 8. > > >> > > >> Add an overlay to activate it. > > >> > > >> Signed-off-by: Frank Wunderlich <frank-w@xxxxxxxxxxxxxxx> > > >> --- > > >> arch/arm64/boot/dts/mediatek/Makefile | 1 + > > >> .../mt7986a-bananapi-bpi-r3-sata.dtso | 39 +++++++++++++++++++ > > >> 2 files changed, 40 insertions(+) > > >> create mode 100644 arch/arm64/boot/dts/mediatek/mt7986a-bananapi-bpi-r3-sata.dtso > > >> > > >> diff --git a/arch/arm64/boot/dts/mediatek/Makefile b/arch/arm64/boot/dts/mediatek/Makefile > > >> index c99c3372a4b5..822d3e36d3df 100644 > > >> --- a/arch/arm64/boot/dts/mediatek/Makefile > > >> +++ b/arch/arm64/boot/dts/mediatek/Makefile > > >> @@ -13,6 +13,7 @@ dtb-$(CONFIG_ARCH_MEDIATEK) += mt7986a-bananapi-bpi-r3-emmc.dtbo > > >> dtb-$(CONFIG_ARCH_MEDIATEK) += mt7986a-bananapi-bpi-r3-nand.dtbo > > >> dtb-$(CONFIG_ARCH_MEDIATEK) += mt7986a-bananapi-bpi-r3-nor.dtbo > > >> dtb-$(CONFIG_ARCH_MEDIATEK) += mt7986a-bananapi-bpi-r3-sd.dtbo > > >> +dtb-$(CONFIG_ARCH_MEDIATEK) += mt7986a-bananapi-bpi-r3-sata.dtbo > > > > > >The requirement for overlays is they have a target base dt in tree and > > >that you apply the overlay to it. All these existing overlays have the > > >same problem which I pointed out when you submitted them. Please fix > > >the existing ones before adding more. > > > > > > > > >Rob > > > > Hi Rob, > > > > i do not understand the problem as there is a target base dt...mt7986a-bananapi-bpi-r3.dtb. > > > > Do you mean that overlays should be merged with basedt at compiletime? > > Yes, because we want to know an overlay can actually apply to what's upstream. so you want this only as compile-time-check or to provide dtbs for users in all possible combinations? for compiletime-check i would add only these 4 lines: mt7986a-bananapi-bpi-r3-sd-nand-dtbs := mt7986a-bananapi-bpi-r3.dtb mt7986a-bananapi-bpi-r3-sd.dtbo mt7986a-bananapi-bpi-r3-nand.dtbo dtb-$(CONFIG_ARCH_MEDIATEK) += mt7986a-bananapi-bpi-r3-sd-nand.dtb mt7986a-bananapi-bpi-r3-emmc-nor-dtbs := mt7986a-bananapi-bpi-r3.dtb mt7986a-bananapi-bpi-r3-emmc.dtbo mt7986a-bananapi-bpi-r3-nor.dtbo dtb-$(CONFIG_ARCH_MEDIATEK) += mt7986a-bananapi-bpi-r3-emmc-nor.dtb for all possible situations we need 8 lines as i wrote in the thread you've linked (sd-nand, sd-nor, emmc-nand, emmc-nor) and this will double the lines in makefile with each additional overlay (e.g. the sata one here resulting in 16 lines total). if the basic syntax is ok (maybe naming the dtbs differently?) and i know compiletest vs. all possibilities i can send patch for correcting this. > > We pack the base-dt and all overlays into one fit image and let uboot dynamically select the combination to apply. E.g. check if emmc is available, if yes apply emmc overlay else the sd one,same for nand/nor. > > That's irrelevant. > > Though do you really want to wait until that point to find out you > made some mistake in your overlay that it doesn't apply when you could > find out much sooner? > > > Else i do not know *how* to fix as i don't see it as broken. > > You need what I pointed out last time[1]. > > Rob > > [1] https://lore.kernel.org/all/CAL_JsqK_3xxD0DFwipXO85P=q=EYjUdjE1_8g1MKtvw3vVzx5A@xxxxxxxxxxxxxx/ >
