[v1] RISC-V: Add eMMC support for TH1520 boards

[PATCH 1/6] dt-bindings: mmc: sdhci-of-dwcmhsc: Add T-Head TH1520 support

Posted by Drew Fustini 2 years, 4 months ago

Add compatible value for the T-Head TH1520 dwcmshc controller and add
thead,phy-pull-up property.

Signed-off-by: Drew Fustini <dfustini@baylibre.com>
---
 Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
index a43eb837f8da..46b768d46712 100644
--- a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
+++ b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
@@ -19,6 +19,7 @@ properties:
       - rockchip,rk3568-dwcmshc
       - rockchip,rk3588-dwcmshc
       - snps,dwcmshc-sdhci
+      - thead,th1520-dwcmshc
 
   reg:
     maxItems: 1
@@ -60,6 +61,9 @@ properties:
     description: Specify the number of delay for tx sampling.
     $ref: /schemas/types.yaml#/definitions/uint8
 
+  thead,phy-pull-up:
+    description: Enable weak pull-up on PHY pads
+    type: boolean
 
 required:
   - compatible

-- 
2.34.1

Re: [PATCH 1/6] dt-bindings: mmc: sdhci-of-dwcmhsc: Add T-Head TH1520 support

Posted by Conor Dooley 2 years, 4 months ago

Hey Drew,

On Thu, Sep 21, 2023 at 06:49:48PM -0700, Drew Fustini wrote:
> Add compatible value for the T-Head TH1520 dwcmshc controller and add
> thead,phy-pull-up property.
> 
> Signed-off-by: Drew Fustini <dfustini@baylibre.com>
> ---
>  Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> index a43eb837f8da..46b768d46712 100644
> --- a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> +++ b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> @@ -19,6 +19,7 @@ properties:
>        - rockchip,rk3568-dwcmshc
>        - rockchip,rk3588-dwcmshc
>        - snps,dwcmshc-sdhci
> +      - thead,th1520-dwcmshc
>  
>    reg:
>      maxItems: 1
> @@ -60,6 +61,9 @@ properties:
>      description: Specify the number of delay for tx sampling.
>      $ref: /schemas/types.yaml#/definitions/uint8
>  
> +  thead,phy-pull-up:
> +    description: Enable weak pull-up on PHY pads
> +    type: boolean

Why is the weak pull-up required? How would the dts author know if they
need to use this property?

Thanks,
Conor.

>  
>  required:
>    - compatible
> 
> -- 
> 2.34.1
>

Re: [PATCH 1/6] dt-bindings: mmc: sdhci-of-dwcmhsc: Add T-Head TH1520 support

Posted by Drew Fustini 2 years, 4 months ago

On Fri, Sep 22, 2023 at 10:57:36AM +0100, Conor Dooley wrote:
> Hey Drew,
> 
> On Thu, Sep 21, 2023 at 06:49:48PM -0700, Drew Fustini wrote:
> > Add compatible value for the T-Head TH1520 dwcmshc controller and add
> > thead,phy-pull-up property.
> > 
> > Signed-off-by: Drew Fustini <dfustini@baylibre.com>
> > ---
> >  Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml | 4 ++++
> >  1 file changed, 4 insertions(+)
> > 
> > diff --git a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > index a43eb837f8da..46b768d46712 100644
> > --- a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > +++ b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > @@ -19,6 +19,7 @@ properties:
> >        - rockchip,rk3568-dwcmshc
> >        - rockchip,rk3588-dwcmshc
> >        - snps,dwcmshc-sdhci
> > +      - thead,th1520-dwcmshc
> >  
> >    reg:
> >      maxItems: 1
> > @@ -60,6 +61,9 @@ properties:
> >      description: Specify the number of delay for tx sampling.
> >      $ref: /schemas/types.yaml#/definitions/uint8
> >  
> > +  thead,phy-pull-up:
> > +    description: Enable weak pull-up on PHY pads
> > +    type: boolean
> 
> Why is the weak pull-up required? How would the dts author know if they
> need to use this property?

This is a good question, and I don't have a good reason beyond it is
what the vendor SDK was doing.

There are only two boards right now using the TH1520 that I know of.
Both the LPi4a [1] and the Ahead [2] have the pull-up property set on
all the mmc controller nodes their downstream device trees.

Rob suggested on #devicetree that it would be simpler to just enable it
in the driver and disable support when needed. I like this idea as it
will simplify this binding patch and the code in the driver patch.

Thanks,
Drew

[1] https://git.beagleboard.org/beaglev-ahead/BeagleBoard-DeviceTrees/-/blob/v5.10.x-ti-unified/src/riscv/light-beagle-ref.dts
[2] https://github.com/revyos/thead-kernel/blob/lpi4a/arch/riscv/boot/dts/thead/light-lpi4a-ref.dts

Re: [PATCH 1/6] dt-bindings: mmc: sdhci-of-dwcmhsc: Add T-Head TH1520 support

Posted by Guo Ren 2 years, 4 months ago

On Fri, Sep 22, 2023 at 11:18 PM Drew Fustini <dfustini@baylibre.com> wrote:
>
> On Fri, Sep 22, 2023 at 10:57:36AM +0100, Conor Dooley wrote:
> > Hey Drew,
> >
> > On Thu, Sep 21, 2023 at 06:49:48PM -0700, Drew Fustini wrote:
> > > Add compatible value for the T-Head TH1520 dwcmshc controller and add
> > > thead,phy-pull-up property.
> > >
> > > Signed-off-by: Drew Fustini <dfustini@baylibre.com>
> > > ---
> > >  Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml | 4 ++++
> > >  1 file changed, 4 insertions(+)
> > >
> > > diff --git a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > > index a43eb837f8da..46b768d46712 100644
> > > --- a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > > +++ b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > > @@ -19,6 +19,7 @@ properties:
> > >        - rockchip,rk3568-dwcmshc
> > >        - rockchip,rk3588-dwcmshc
> > >        - snps,dwcmshc-sdhci
> > > +      - thead,th1520-dwcmshc
> > >
> > >    reg:
> > >      maxItems: 1
> > > @@ -60,6 +61,9 @@ properties:
> > >      description: Specify the number of delay for tx sampling.
> > >      $ref: /schemas/types.yaml#/definitions/uint8
> > >
> > > +  thead,phy-pull-up:
> > > +    description: Enable weak pull-up on PHY pads
> > > +    type: boolean
> >
> > Why is the weak pull-up required? How would the dts author know if they
> > need to use this property?
>
> This is a good question, and I don't have a good reason beyond it is
> what the vendor SDK was doing.
>
> There are only two boards right now using the TH1520 that I know of.
> Both the LPi4a [1] and the Ahead [2] have the pull-up property set on
> all the mmc controller nodes their downstream device trees.
>
> Rob suggested on #devicetree that it would be simpler to just enable it
> in the driver and disable support when needed. I like this idea as it
> will simplify this binding patch and the code in the driver patch.
It's for the PHY debug. You could directly remove them, or keep it
with no-pull-up flag, then no pull-up flag in the dts for default..

static void sdhci_phy_1_8v_init_no_pull(struct sdhci_host *host)
{
        uint32_t val;
        sdhci_writel(host, 1, DWC_MSHC_PTR_PHY_R);
        sdhci_writeb(host, 1 << 4, PHY_SDCLKDL_CNFG_R);
        sdhci_writeb(host, 0x40, PHY_SDCLKDL_DC_R);

        val = sdhci_readb(host, PHY_SDCLKDL_CNFG_R);
        val &= ~(1 << 4);
        sdhci_writeb(host, val, PHY_SDCLKDL_CNFG_R);


        val = sdhci_readw(host, PHY_CMDPAD_CNFG_R);
        sdhci_writew(host, val | 1, PHY_CMDPAD_CNFG_R);

        val = sdhci_readw(host, PHY_DATAPAD_CNFG_R);
        sdhci_writew(host, val | 1, PHY_DATAPAD_CNFG_R);

        val = sdhci_readw(host, PHY_RSTNPAD_CNFG_R);
        sdhci_writew(host, val | 1, PHY_RSTNPAD_CNFG_R);

        val = sdhci_readw(host, PHY_STBPAD_CNFG_R);
        sdhci_writew(host, val | 1, PHY_STBPAD_CNFG_R);

        val = sdhci_readb(host, PHY_DLL_CTRL_R);
        sdhci_writeb(host, val | 1, PHY_DLL_CTRL_R);
}

static void sdhci_phy_3_3v_init_no_pull(struct sdhci_host *host)
{
        uint32_t val;
        sdhci_writel(host, 1, DWC_MSHC_PTR_PHY_R);
        sdhci_writeb(host, 1 << 4, PHY_SDCLKDL_CNFG_R);
        sdhci_writeb(host, 0x40, PHY_SDCLKDL_DC_R);

        val = sdhci_readb(host, PHY_SDCLKDL_CNFG_R);
        val &= ~(1 << 4);
        sdhci_writeb(host, val, PHY_SDCLKDL_CNFG_R);

        val = sdhci_readw(host, PHY_CMDPAD_CNFG_R);
        sdhci_writew(host, val | 2, PHY_CMDPAD_CNFG_R);

        val = sdhci_readw(host, PHY_DATAPAD_CNFG_R);
        sdhci_writew(host, val | 2, PHY_DATAPAD_CNFG_R);

        val = sdhci_readw(host, PHY_RSTNPAD_CNFG_R);
        sdhci_writew(host, val | 2, PHY_RSTNPAD_CNFG_R);

        val = sdhci_readw(host, PHY_STBPAD_CNFG_R);
        sdhci_writew(host, val | 2, PHY_STBPAD_CNFG_R);

        val = sdhci_readb(host, PHY_DLL_CTRL_R);
        sdhci_writeb(host, val | 1, PHY_DLL_CTRL_R);
}

>
> Thanks,
> Drew
>
> [1] https://git.beagleboard.org/beaglev-ahead/BeagleBoard-DeviceTrees/-/blob/v5.10.x-ti-unified/src/riscv/light-beagle-ref.dts
> [2] https://github.com/revyos/thead-kernel/blob/lpi4a/arch/riscv/boot/dts/thead/light-lpi4a-ref.dts



-- 
Best Regards
 Guo Ren

Re: [PATCH 1/6] dt-bindings: mmc: sdhci-of-dwcmhsc: Add T-Head TH1520 support

Posted by Drew Fustini 2 years, 4 months ago

On Sat, Sep 23, 2023 at 04:10:11PM +0800, Guo Ren wrote:
> On Fri, Sep 22, 2023 at 11:18 PM Drew Fustini <dfustini@baylibre.com> wrote:
> >
> > On Fri, Sep 22, 2023 at 10:57:36AM +0100, Conor Dooley wrote:
> > > Hey Drew,
> > >
> > > On Thu, Sep 21, 2023 at 06:49:48PM -0700, Drew Fustini wrote:
> > > > Add compatible value for the T-Head TH1520 dwcmshc controller and add
> > > > thead,phy-pull-up property.
> > > >
> > > > Signed-off-by: Drew Fustini <dfustini@baylibre.com>
> > > > ---
> > > >  Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml | 4 ++++
> > > >  1 file changed, 4 insertions(+)
> > > >
> > > > diff --git a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > > > index a43eb837f8da..46b768d46712 100644
> > > > --- a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > > > +++ b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
> > > > @@ -19,6 +19,7 @@ properties:
> > > >        - rockchip,rk3568-dwcmshc
> > > >        - rockchip,rk3588-dwcmshc
> > > >        - snps,dwcmshc-sdhci
> > > > +      - thead,th1520-dwcmshc
> > > >
> > > >    reg:
> > > >      maxItems: 1
> > > > @@ -60,6 +61,9 @@ properties:
> > > >      description: Specify the number of delay for tx sampling.
> > > >      $ref: /schemas/types.yaml#/definitions/uint8
> > > >
> > > > +  thead,phy-pull-up:
> > > > +    description: Enable weak pull-up on PHY pads
> > > > +    type: boolean
> > >
> > > Why is the weak pull-up required? How would the dts author know if they
> > > need to use this property?
> >
> > This is a good question, and I don't have a good reason beyond it is
> > what the vendor SDK was doing.
> >
> > There are only two boards right now using the TH1520 that I know of.
> > Both the LPi4a [1] and the Ahead [2] have the pull-up property set on
> > all the mmc controller nodes their downstream device trees.
> >
> > Rob suggested on #devicetree that it would be simpler to just enable it
> > in the driver and disable support when needed. I like this idea as it
> > will simplify this binding patch and the code in the driver patch.
> It's for the PHY debug. You could directly remove them, or keep it
> with no-pull-up flag, then no pull-up flag in the dts for default..

Thank you for explaining that the purpose of the pull-up DT property
was for PHY debug.

I will plan to remove the pull-up DT property in the next version of
this patch series and remove sdhci_phy_1_8v_init_no_pull() and
sdhci_phy_3_3v_init_no_pull() from my sdhci-of-dwcmshc.c patch.

I will make th1520_phy_1_8v_init() and th1520_phy_3_3v_init() always
set WEAKPULL_EN = 1 (PULLUP) for CMDPAD_CNFG, DATAPAD_CNFG, RSTNPAD_CNFG
and set WEAKPULL_EN = 2 (PULLDOWN) for STBPAD_CNFG.

Thanks,
Drew