Message ID | 20220510152020.1.I6094f0a935212eabc7d9348d6cb5a2517ccc2514@changeid (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | arm64: dts: mt8173: add aliases for mt8173 mmc device nodes | expand |
Hi, On Tue, May 10, 2022 at 1:21 PM Eizan Miyamoto <eizan@chromium.org> wrote: > > Doing this makes them get enumerated in a deterministic order so that > the assigned name in /dev that the mmc devices get assigned is > consistent across reboots. Arnd, one of the soc maintainers asked folks not to do this: https://lore.kernel.org/all/CAK8P3a3_uO23Y0tyfdh26tmCVcyp6XM+ur0WUd4khGUoBb32Hw@mail.gmail.com/ and subsequently all mmc aliases for Rockchip were moved to the board level. > See ead9f7d7ea9 ("arm64: dts: qcom: sc7180: Assign numbers to eMMC and > SD") for similar work. I don't know why it didn't happen for Qualcomm though. ChenYu > Signed-off-by: Eizan Miyamoto <eizan@chromium.org> > --- > > arch/arm64/boot/dts/mediatek/mt8173.dtsi | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/arch/arm64/boot/dts/mediatek/mt8173.dtsi b/arch/arm64/boot/dts/mediatek/mt8173.dtsi > index 6b02cff605fbe..9f3cd54116de9 100644 > --- a/arch/arm64/boot/dts/mediatek/mt8173.dtsi > +++ b/arch/arm64/boot/dts/mediatek/mt8173.dtsi > @@ -55,6 +55,10 @@ aliases { > serial1 = &uart1; > serial2 = &uart2; > serial3 = &uart3; > + mmc0 = &mmc0; > + mmc1 = &mmc1; > + mmc2 = &mmc2; > + mmc3 = &mmc3; > }; > > cluster0_opp: opp_table0 { > -- > 2.36.0.512.ge40c2bad7a-goog > > > _______________________________________________ > Linux-mediatek mailing list > Linux-mediatek@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/linux-mediatek
On Tue, May 10, 2022 at 9:36 AM Chen-Yu Tsai <wenst@chromium.org> wrote: > On Tue, May 10, 2022 at 1:21 PM Eizan Miyamoto <eizan@chromium.org> wrote: > > > > Doing this makes them get enumerated in a deterministic order so that > > the assigned name in /dev that the mmc devices get assigned is > > consistent across reboots. > > Arnd, one of the soc maintainers asked folks not to do this: > > https://lore.kernel.org/all/CAK8P3a3_uO23Y0tyfdh26tmCVcyp6XM+ur0WUd4khGUoBb32Hw@mail.gmail.com/ > > and subsequently all mmc aliases for Rockchip were moved to the board level. Right, please move the aliases into the per-board dts files, and make sure to only list the ones that are actually connected on a given machine. > > See ead9f7d7ea9 ("arm64: dts: qcom: sc7180: Assign numbers to eMMC and > > SD") for similar work. > > I don't know why it didn't happen for Qualcomm though. It takes a while to get things consistent, need to have another look there. Arnd
diff --git a/arch/arm64/boot/dts/mediatek/mt8173.dtsi b/arch/arm64/boot/dts/mediatek/mt8173.dtsi index 6b02cff605fbe..9f3cd54116de9 100644 --- a/arch/arm64/boot/dts/mediatek/mt8173.dtsi +++ b/arch/arm64/boot/dts/mediatek/mt8173.dtsi @@ -55,6 +55,10 @@ aliases { serial1 = &uart1; serial2 = &uart2; serial3 = &uart3; + mmc0 = &mmc0; + mmc1 = &mmc1; + mmc2 = &mmc2; + mmc3 = &mmc3; }; cluster0_opp: opp_table0 {
Doing this makes them get enumerated in a deterministic order so that the assigned name in /dev that the mmc devices get assigned is consistent across reboots. See ead9f7d7ea9 ("arm64: dts: qcom: sc7180: Assign numbers to eMMC and SD") for similar work. Signed-off-by: Eizan Miyamoto <eizan@chromium.org> --- arch/arm64/boot/dts/mediatek/mt8173.dtsi | 4 ++++ 1 file changed, 4 insertions(+)