[PATCH v2 1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP

[PATCH v2 1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP

[Mikko Rapeli]

On rk3399 based rockpi4b, mounting rootfs from mmc fails unless
ROCKCHIP_PM_DOMAINS is enabled. Accoriding to
Heiko Stübner <heiko@sntech.de> all SoCs since 2012 have power
domains so the support should be enabled by default.

Failing boot without CONFIG_ROCKCHIP_PM_DOMAINS=y:

https://ledge.validation.linaro.org/scheduler/job/119268

/dev/disk/by-partuuid/1d48ffd8-f2a7-4a33-b52f-186089b3c85e: Can't lookup blockdev
/dev/disk/by-partuuid/1d48ffd8-f2a7-4a33-b52f-186089b3c85e: Can't lookup blockdev
/dev/disk/by-partuuid/1d48ffd8-f2a7-4a33-b52f-186089b3c85e: Can't lookup blockdev
/dev/disk/by-partuuid/1d48ffd8-f2a7-4a33-b52f-186089b3c85e: Can't lookup blockdev
/dev/disk/by-partuuid/1d48ffd8-f2a7-4a33-b52f-186089b3c85e: Can't lookup blockdev
/dev/disk/by-partuuid/1d48ffd8-f2a7-4a33-b52f-186089b3c85e: Can't lookup blockdev
dw-apb-uart ff1a0000.serial: forbid DMA for kernel console
root '/dev/disk/by-partuuid/1d48ffd8-f2a7-4a33-b52f-186089b3c85e' doesn't exist or does not contain a /dev.
rk_gmac-dwmac fe300000.ethernet: deferred probe timeout, ignoring dependency
rk_gmac-dwmac fe300000.ethernet: probe with driver rk_gmac-dwmac failed with error -110
rk_iommu ff650800.iommu: deferred probe timeout, ignoring dependency
rk_iommu ff650800.iommu: probe with driver rk_iommu failed with error -110
dwmmc_rockchip fe320000.mmc: deferred probe timeout, ignoring dependency
rockchip-typec-phy ff7c0000.phy: deferred probe timeout, ignoring dependency
dwmmc_rockchip fe320000.mmc: probe with driver dwmmc_rockchip failed with error -110
rockchip-typec-phy ff7c0000.phy: probe with driver rockchip-typec-phy failed with error -110
rockchip-typec-phy ff800000.phy: deferred probe timeout, ignoring dependency
rockchip-typec-phy ff800000.phy: probe with driver rockchip-typec-phy failed with error -110
rk_iommu ff660480.iommu: deferred probe timeout, ignoring dependency
rk_iommu ff660480.iommu: probe with driver rk_iommu failed with error -110
rk_iommu ff8f3f00.iommu: deferred probe timeout, ignoring dependency
rk_iommu ff8f3f00.iommu: probe with driver rk_iommu failed with error -110
rk_iommu ff903f00.iommu: deferred probe timeout, ignoring dependency
rk_iommu ff903f00.iommu: probe with driver rk_iommu failed with error -110
rk_iommu ff914000.iommu: deferred probe timeout, ignoring dependency
rk_iommu ff914000.iommu: probe with driver rk_iommu failed with error -110
rk_iommu ff924000.iommu: deferred probe timeout, ignoring dependency
rk_iommu ff924000.iommu: probe with driver rk_iommu failed with error -110
platform fe800000.usb: deferred probe pending: platform: wait for supplier /phy@ff7c0000/usb3-port
sdhci-arasan fe330000.mmc: deferred probe timeout, ignoring dependency
platform fe900000.usb: deferred probe pending: platform: wait for supplier /phy@ff800000/usb3-port
sdhci-arasan fe330000.mmc: probe with driver sdhci-arasan failed with error -110
platform ff1d0000.spi: deferred probe pending: (reason unknown)
platform hdmi-sound: deferred probe pending: asoc-simple-card: parse error

Working boot with CONFIG_ROCKCHIP_PM_DOMAINS=y:

https://ledge.validation.linaro.org/scheduler/job/119272

dwmmc_rockchip fe320000.mmc: IDMAC supports 32-bit address mode.
dwmmc_rockchip fe320000.mmc: Using internal DMA controller.
dwmmc_rockchip fe320000.mmc: Version ID is 270a
dwmmc_rockchip fe320000.mmc: DW MMC controller at irq 45,32 bit host data width,256 deep fifo
dwmmc_rockchip fe320000.mmc: Got CD GPIO
ff1a0000.serial: ttyS2 at MMIO 0xff1a0000 (irq = 44, base_baud = 1500000) is a 16550A
printk: legacy console [ttyS2] enabled
mmc_host mmc1: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
dw_wdt ff848000.watchdog: No valid TOPs array specified
mmc_host mmc1: Bus speed (slot 0) = 50000000Hz (slot req 50000000Hz, actual 50000000HZ div = 0)
mmc0: CQHCI version 5.10
rk_gmac-dwmac fe300000.ethernet: IRQ eth_wake_irq not found
mmc1: new high speed SDHC card at address aaaa
fan53555-regulator 0-0040: FAN53555 Option[8] Rev[1] Detected!
fan53555-regulator 0-0041: FAN53555 Option[8] Rev[1] Detected!
rk_gmac-dwmac fe300000.ethernet: IRQ eth_lpi not found
mmcblk1: mmc1:aaaa SC16G 14.8 GiB
rk_gmac-dwmac fe300000.ethernet: IRQ sfty not found
GPT:Primary header thinks Alt. header is not at the end of the disk.
rk_gmac-dwmac fe300000.ethernet: Deprecated MDIO bus assumption used
GPT:1978417 != 31116287
rk_gmac-dwmac fe300000.ethernet: PTP uses main clock
GPT:Alternate GPT header not at the end of the disk.
rk_gmac-dwmac fe300000.ethernet: clock input or output? (input).
GPT:1978417 != 31116287
rk_gmac-dwmac fe300000.ethernet: TX delay(0x28).
GPT: Use GNU Parted to correct GPT errors.
rk_gmac-dwmac fe300000.ethernet: RX delay(0x11).
 mmcblk1: p1 p2 p3 p4 p5 p6 p7 p8

Suggested-by: Ulf Hansson <ulf.hansson@linaro.org>
Suggested-by: Heiko Stübner <heiko@sntech.de>
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
---
 arch/arm64/Kconfig.platforms | 1 +
 1 file changed, 1 insertion(+)

v2: changed from "depend on" in MMC driver to "select" as default
    on ARCH_ROCKCHIP as suggested by Ulf and Heiko

v1: https://lore.kernel.org/linux-mmc/20250911144313.2774171-1-mikko.rapeli@linaro.org/

diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms
index a88f5ad9328c2..04b3601486290 100644
--- a/arch/arm64/Kconfig.platforms
+++ b/arch/arm64/Kconfig.platforms
@@ -306,6 +306,7 @@ config ARCH_ROCKCHIP
 	select ARCH_HAS_RESET_CONTROLLER
 	select PINCTRL
 	select PM
+	select ROCKCHIP_PM_DOMAINS
 	select ROCKCHIP_TIMER
 	help
 	  This enables support for the ARMv8 based Rockchip chipsets,
-- 
2.34.1

[PATCH v2 2/2] ARM: rockchip: select ROCKCHIP_PM_DOMAINS

[Mikko Rapeli]

According to Heiko Stuebner <heiko@sntech.de> all SoCs
since 2012 have power domain support in HW and
ROCKCHIP_PM_DOMAINS should always be enabled.

Suggested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
---
 arch/arm/mach-rockchip/Kconfig | 1 +
 1 file changed, 1 insertion(+)

v2: added "select" for ROCKCHIP_PM_DOMAINS also on arm
    mach-rockchip as suggested by Heiko

v1: https://lore.kernel.org/linux-mmc/20250911144313.2774171-1-mikko.rapeli@linaro.org/

diff --git a/arch/arm/mach-rockchip/Kconfig b/arch/arm/mach-rockchip/Kconfig
index b7855cc665e94..37f1cf38a010b 100644
--- a/arch/arm/mach-rockchip/Kconfig
+++ b/arch/arm/mach-rockchip/Kconfig
@@ -14,6 +14,7 @@ config ARCH_ROCKCHIP
 	select HAVE_ARM_TWD if SMP
 	select DW_APB_TIMER_OF
 	select REGULATOR if PM
+	select ROCKCHIP_PM_DOMAINS
 	select ROCKCHIP_TIMER
 	select ARM_GLOBAL_TIMER
 	select CLKSRC_ARM_GLOBAL_TIMER_SCHED_CLOCK
-- 
2.34.1

Re: [PATCH v2 2/2] ARM: rockchip: select ROCKCHIP_PM_DOMAINS

[Geert Uytterhoeven]

Hi Mikko,

On Fri, 12 Sept 2025 at 10:41, Mikko Rapeli <mikko.rapeli@linaro.org> wrote:
> According to Heiko Stuebner <heiko@sntech.de> all SoCs
> since 2012 have power domain support in HW and
> ROCKCHIP_PM_DOMAINS should always be enabled.
>
> Suggested-by: Heiko Stuebner <heiko@sntech.de>
> Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>

Thanks for your patch!

> --- a/arch/arm/mach-rockchip/Kconfig
> +++ b/arch/arm/mach-rockchip/Kconfig
> @@ -14,6 +14,7 @@ config ARCH_ROCKCHIP
>         select HAVE_ARM_TWD if SMP
>         select DW_APB_TIMER_OF
>         select REGULATOR if PM
> +       select ROCKCHIP_PM_DOMAINS

This requires "select PM". Oh, that is done below (out-of-context).
So the "if PM" above can be dropped.

>         select ROCKCHIP_TIMER
>         select ARM_GLOBAL_TIMER
>         select CLKSRC_ARM_GLOBAL_TIMER_SCHED_CLOCK

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH v2 2/2] ARM: rockchip: select ROCKCHIP_PM_DOMAINS

[Mikko Rapeli]

Hi,

On Fri, Sep 12, 2025 at 10:48:47AM +0200, Geert Uytterhoeven wrote:
> Hi Mikko,
> 
> On Fri, 12 Sept 2025 at 10:41, Mikko Rapeli <mikko.rapeli@linaro.org> wrote:
> > According to Heiko Stuebner <heiko@sntech.de> all SoCs
> > since 2012 have power domain support in HW and
> > ROCKCHIP_PM_DOMAINS should always be enabled.
> >
> > Suggested-by: Heiko Stuebner <heiko@sntech.de>
> > Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
> 
> Thanks for your patch!
> 
> > --- a/arch/arm/mach-rockchip/Kconfig
> > +++ b/arch/arm/mach-rockchip/Kconfig
> > @@ -14,6 +14,7 @@ config ARCH_ROCKCHIP
> >         select HAVE_ARM_TWD if SMP
> >         select DW_APB_TIMER_OF
> >         select REGULATOR if PM
> > +       select ROCKCHIP_PM_DOMAINS
> 
> This requires "select PM". Oh, that is done below (out-of-context).
> So the "if PM" above can be dropped.

Yes, it's selected just after the context. I can drop it from
"select REGULATOR" in v3.

I was not sure if the PM dependency should be there or not
also when it's explicitly selected.

Cheers,

-Mikko

Re: [PATCH v2 1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP

[Heiko Stuebner]

On Fri, 12 Sep 2025 11:41:11 +0300, Mikko Rapeli wrote:
> On rk3399 based rockpi4b, mounting rootfs from mmc fails unless
> ROCKCHIP_PM_DOMAINS is enabled. Accoriding to
> Heiko Stübner <heiko@sntech.de> all SoCs since 2012 have power
> domains so the support should be enabled by default.
> 
> Failing boot without CONFIG_ROCKCHIP_PM_DOMAINS=y:
> 
> [...]

Applied, thanks!

[1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP
      commit: aff3de3b9a597763bec5d1c56c11b56fb2613085
[2/2] ARM: rockchip: select ROCKCHIP_PM_DOMAINS
      commit: 47c9bfdf7376e8b1842699b064ae3b052e2c3c1b

Best regards,
-- 
Heiko Stuebner <heiko@sntech.de>

Re: [PATCH v2 1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP

[Ulf Hansson]

On Thu, 18 Sept 2025 at 23:09, Heiko Stuebner <heiko@sntech.de> wrote:
>
>
> On Fri, 12 Sep 2025 11:41:11 +0300, Mikko Rapeli wrote:
> > On rk3399 based rockpi4b, mounting rootfs from mmc fails unless
> > ROCKCHIP_PM_DOMAINS is enabled. Accoriding to
> > Heiko Stübner <heiko@sntech.de> all SoCs since 2012 have power
> > domains so the support should be enabled by default.
> >
> > Failing boot without CONFIG_ROCKCHIP_PM_DOMAINS=y:
> >
> > [...]
>
> Applied, thanks!
>
> [1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP
>       commit: aff3de3b9a597763bec5d1c56c11b56fb2613085
> [2/2] ARM: rockchip: select ROCKCHIP_PM_DOMAINS
>       commit: 47c9bfdf7376e8b1842699b064ae3b052e2c3c1b

Looks like you picked the old version of the patches. Did you intend
to pick the v4? Then it's the one below you want.
https://lore.kernel.org/all/20250915083317.2885761-5-mikko.rapeli@linaro.org/

Note, I have picked the other patches from the v4 via my pmdomain/mmc trees.

Kind regards
Uffe

Re: [PATCH v2 1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP

[Heiko Stübner]

Hi Ulf,

Am Freitag, 19. September 2025, 12:51:51 Mitteleuropäische Sommerzeit schrieb Ulf Hansson:
> On Thu, 18 Sept 2025 at 23:09, Heiko Stuebner <heiko@sntech.de> wrote:
> >
> >
> > On Fri, 12 Sep 2025 11:41:11 +0300, Mikko Rapeli wrote:
> > > On rk3399 based rockpi4b, mounting rootfs from mmc fails unless
> > > ROCKCHIP_PM_DOMAINS is enabled. Accoriding to
> > > Heiko Stübner <heiko@sntech.de> all SoCs since 2012 have power
> > > domains so the support should be enabled by default.
> > >
> > > Failing boot without CONFIG_ROCKCHIP_PM_DOMAINS=y:
> > >
> > > [...]
> >
> > Applied, thanks!
> >
> > [1/2] arm64: select ROCKCHIP_PM_DOMAINS with ARCH_ROCKCHIP
> >       commit: aff3de3b9a597763bec5d1c56c11b56fb2613085
> > [2/2] ARM: rockchip: select ROCKCHIP_PM_DOMAINS
> >       commit: 47c9bfdf7376e8b1842699b064ae3b052e2c3c1b
> 
> Looks like you picked the old version of the patches. Did you intend
> to pick the v4? Then it's the one below you want.
> https://lore.kernel.org/all/20250915083317.2885761-5-mikko.rapeli@linaro.org/
> 
> Note, I have picked the other patches from the v4 via my pmdomain/mmc trees.

oh ... you're right ... these patches from here are superseeded,
because we don't want to depend on on them in the arch part, but
instead to the default=y if Rockchip from that v4 .

I did pick up the patch4 from over there too. I'll drop these 2 from here.

Thank you sooo much for noticing this.
Heiko