[PATCH v2 0/7] MPFS clock fixes required for correct CAN clock modeling

[PATCH v2 0/7] MPFS clock fixes required for correct CAN clock modeling

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

While reviewing a CAN driver internally for MPFS [1], I realised
that the modeling of the MSSPLL such that only one of its outputs could
be used was not correct. The CAN controllers on MPFS take 2 input
clocks - one that is the bus clock, acquired from the main MSSPLL and
a second clock for the AHB interface to the result of the SoC.
Currently the binding for the CAN controllers and the represetnation
of the MSSPLL only allows for one of these clocks.
Modify the binding and devicetree to expect two clocks and rework the
main clock controller driver for MPFS such that it is capable of
providing multiple outputs from the MSSPLL.

Cheers,
Conor.

1 - Hopefully that'll show up on the lists soon, once we are happy with
  it ourselves. (As of v2, that's not happened yet, Christmas etc gettin
  in the way, soonTM).

Changes in v2:
- Swap MSSPLL out for MSSPLL0 in the driver, there's no functional
  change there.
- drop the unneeded maxItems from clocks.

CC: Conor Dooley <conor.dooley@microchip.com>
CC: Daire McNamara <daire.mcnamara@microchip.com>
CC: Wolfgang Grandegger <wg@grandegger.com>
CC: Marc Kleine-Budde <mkl@pengutronix.de>
CC: "David S. Miller" <davem@davemloft.net>
CC: Eric Dumazet <edumazet@google.com>
CC: Jakub Kicinski <kuba@kernel.org>
CC: Paolo Abeni <pabeni@redhat.com>
CC: Rob Herring <robh+dt@kernel.org>
CC: Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>
CC: Paul Walmsley <paul.walmsley@sifive.com>
CC: Palmer Dabbelt <palmer@dabbelt.com>
CC: Albert Ou <aou@eecs.berkeley.edu>
CC: Michael Turquette <mturquette@baylibre.com>
CC: Stephen Boyd <sboyd@kernel.org>
CC: linux-riscv@lists.infradead.org
CC: linux-can@vger.kernel.org
CC: netdev@vger.kernel.org
CC: devicetree@vger.kernel.org
CC: linux-kernel@vger.kernel.org
CC: linux-clk@vger.kernel.org

Conor Dooley (7):
  dt-bindings: clock: mpfs: add more MSSPLL output definitions
  dt-bindings: can: mpfs: add missing required clock
  clk: microchip: mpfs: split MSSPLL in two
  clk: microchip: mpfs: setup for using other mss pll outputs
  clk: microchip: mpfs: add missing MSSPLL outputs
  clk: microchip: mpfs: convert MSSPLL outputs to clk_divider
  riscv: dts: microchip: add missing CAN bus clocks

 .../bindings/net/can/microchip,mpfs-can.yaml  |   6 +-
 arch/riscv/boot/dts/microchip/mpfs.dtsi       |   4 +-
 drivers/clk/microchip/clk-mpfs.c              | 154 ++++++++++--------
 .../dt-bindings/clock/microchip,mpfs-clock.h  |   5 +
 4 files changed, 98 insertions(+), 71 deletions(-)

-- 
2.43.0

[PATCH v2 1/7] dt-bindings: clock: mpfs: add more MSSPLL output definitions

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

There are 3 undocumented outputs of the MSSPLL that are used for the CAN
bus, "user crypto" module and eMMC. Add their clock IDs so that they can
be hooked up in DT.

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 include/dt-bindings/clock/microchip,mpfs-clock.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/dt-bindings/clock/microchip,mpfs-clock.h b/include/dt-bindings/clock/microchip,mpfs-clock.h
index 79775a5134ca..b52f19a2b480 100644
--- a/include/dt-bindings/clock/microchip,mpfs-clock.h
+++ b/include/dt-bindings/clock/microchip,mpfs-clock.h
@@ -44,6 +44,11 @@
 
 #define CLK_RTCREF	33
 #define CLK_MSSPLL	34
+#define CLK_MSSPLL0	34
+#define CLK_MSSPLL1	35
+#define CLK_MSSPLL2	36
+#define CLK_MSSPLL3	37
+/* 38 is reserved for MSS PLL internals */
 
 /* Clock Conditioning Circuitry Clock IDs */
 
-- 
2.43.0

[PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
CAN bus clock. The bus clock was omitted when the binding was written,
but is required for operation. Make up for lost time and add it.

Cautionary tale in adding bindings without having implemented a real
user for them perhaps.

Fixes: c878d518d7b6 ("dt-bindings: can: mpfs: document the mpfs CAN controller")
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 .../devicetree/bindings/net/can/microchip,mpfs-can.yaml     | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
index 45aa3de7cf01..01e4d4a54df6 100644
--- a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
+++ b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
@@ -24,7 +24,9 @@ properties:
     maxItems: 1
 
   clocks:
-    maxItems: 1
+    items:
+      - description: AHB peripheral clock
+      - description: CAN bus clock
 
 required:
   - compatible
@@ -39,7 +41,7 @@ examples:
     can@2010c000 {
         compatible = "microchip,mpfs-can";
         reg = <0x2010c000 0x1000>;
-        clocks = <&clkcfg 17>;
+        clocks = <&clkcfg 17>, <&clkcfg 37>;
         interrupt-parent = <&plic>;
         interrupts = <56>;
     };
-- 
2.43.0

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Marc Kleine-Budde]

[-- Attachment #1: Type: text/plain, Size: 1970 bytes --]

On 22.01.2024 12:19:50, Conor Dooley wrote:
> From: Conor Dooley <conor.dooley@microchip.com>
> 
> The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
> CAN bus clock. The bus clock was omitted when the binding was written,
> but is required for operation. Make up for lost time and add it.
> 
> Cautionary tale in adding bindings without having implemented a real
> user for them perhaps.
> 
> Fixes: c878d518d7b6 ("dt-bindings: can: mpfs: document the mpfs CAN controller")
> Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> ---
>  .../devicetree/bindings/net/can/microchip,mpfs-can.yaml     | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> index 45aa3de7cf01..01e4d4a54df6 100644
> --- a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> +++ b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> @@ -24,7 +24,9 @@ properties:
>      maxItems: 1
>  
>    clocks:
> -    maxItems: 1
> +    items:
> +      - description: AHB peripheral clock
> +      - description: CAN bus clock

What about adding clock-names, so that the order can be checked
automatically?

>  
>  required:
>    - compatible
> @@ -39,7 +41,7 @@ examples:
>      can@2010c000 {
>          compatible = "microchip,mpfs-can";
>          reg = <0x2010c000 0x1000>;
> -        clocks = <&clkcfg 17>;
> +        clocks = <&clkcfg 17>, <&clkcfg 37>;
>          interrupt-parent = <&plic>;
>          interrupts = <56>;
>      };
> -- 
> 2.43.0
> 
> 

Marc

-- 
Pengutronix e.K.                 | Marc Kleine-Budde          |
Embedded Linux                   | https://www.pengutronix.de |
Vertretung Nürnberg              | Phone: +49-5121-206917-129 |
Amtsgericht Hildesheim, HRA 2686 | Fax:   +49-5121-206917-9   |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 1566 bytes --]

On Mon, Jan 22, 2024 at 02:13:16PM +0100, Marc Kleine-Budde wrote:
> On 22.01.2024 12:19:50, Conor Dooley wrote:
> > From: Conor Dooley <conor.dooley@microchip.com>
> > 
> > The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
> > CAN bus clock. The bus clock was omitted when the binding was written,
> > but is required for operation. Make up for lost time and add it.
> > 
> > Cautionary tale in adding bindings without having implemented a real
> > user for them perhaps.
> > 
> > Fixes: c878d518d7b6 ("dt-bindings: can: mpfs: document the mpfs CAN controller")
> > Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> > ---
> >  .../devicetree/bindings/net/can/microchip,mpfs-can.yaml     | 6 ++++--
> >  1 file changed, 4 insertions(+), 2 deletions(-)
> > 
> > diff --git a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > index 45aa3de7cf01..01e4d4a54df6 100644
> > --- a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > +++ b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > @@ -24,7 +24,9 @@ properties:
> >      maxItems: 1
> >  
> >    clocks:
> > -    maxItems: 1
> > +    items:
> > +      - description: AHB peripheral clock
> > +      - description: CAN bus clock
> 
> What about adding clock-names, so that the order can be checked
> automatically?

I don't personally care for doing so, but if your heart is set on having
them, then sure.

Cheers,
Conor.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Marc Kleine-Budde]

[-- Attachment #1: Type: text/plain, Size: 2200 bytes --]

On 22.01.2024 14:21:04, Conor Dooley wrote:
> On Mon, Jan 22, 2024 at 02:13:16PM +0100, Marc Kleine-Budde wrote:
> > On 22.01.2024 12:19:50, Conor Dooley wrote:
> > > From: Conor Dooley <conor.dooley@microchip.com>
> > > 
> > > The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
> > > CAN bus clock. The bus clock was omitted when the binding was written,
> > > but is required for operation. Make up for lost time and add it.
> > > 
> > > Cautionary tale in adding bindings without having implemented a real
> > > user for them perhaps.
> > > 
> > > Fixes: c878d518d7b6 ("dt-bindings: can: mpfs: document the mpfs CAN controller")
> > > Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> > > ---
> > >  .../devicetree/bindings/net/can/microchip,mpfs-can.yaml     | 6 ++++--
> > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > index 45aa3de7cf01..01e4d4a54df6 100644
> > > --- a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > +++ b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > @@ -24,7 +24,9 @@ properties:
> > >      maxItems: 1
> > >  
> > >    clocks:
> > > -    maxItems: 1
> > > +    items:
> > > +      - description: AHB peripheral clock
> > > +      - description: CAN bus clock
> > 
> > What about adding clock-names, so that the order can be checked
> > automatically?
> 
> I don't personally care for doing so, but if your heart is set on having
> them, then sure.

Usually the CAN driver needs to have the clock rate of the clocks that
the basis for the CAN bus clock. Looking at the clocks description it's
probably the 2nd one.

With clock-names we can automatically check that the 2nd clock is always
the CAN clock.

Marc

-- 
Pengutronix e.K.                 | Marc Kleine-Budde          |
Embedded Linux                   | https://www.pengutronix.de |
Vertretung Nürnberg              | Phone: +49-5121-206917-129 |
Amtsgericht Hildesheim, HRA 2686 | Fax:   +49-5121-206917-9   |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 2371 bytes --]

On Mon, Jan 22, 2024 at 03:46:04PM +0100, Marc Kleine-Budde wrote:
> On 22.01.2024 14:21:04, Conor Dooley wrote:
> > On Mon, Jan 22, 2024 at 02:13:16PM +0100, Marc Kleine-Budde wrote:
> > > On 22.01.2024 12:19:50, Conor Dooley wrote:
> > > > From: Conor Dooley <conor.dooley@microchip.com>
> > > > 
> > > > The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
> > > > CAN bus clock. The bus clock was omitted when the binding was written,
> > > > but is required for operation. Make up for lost time and add it.
> > > > 
> > > > Cautionary tale in adding bindings without having implemented a real
> > > > user for them perhaps.
> > > > 
> > > > Fixes: c878d518d7b6 ("dt-bindings: can: mpfs: document the mpfs CAN controller")
> > > > Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> > > > ---
> > > >  .../devicetree/bindings/net/can/microchip,mpfs-can.yaml     | 6 ++++--
> > > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > > > 
> > > > diff --git a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > > index 45aa3de7cf01..01e4d4a54df6 100644
> > > > --- a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > > +++ b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > > @@ -24,7 +24,9 @@ properties:
> > > >      maxItems: 1
> > > >  
> > > >    clocks:
> > > > -    maxItems: 1
> > > > +    items:
> > > > +      - description: AHB peripheral clock
> > > > +      - description: CAN bus clock
> > > 
> > > What about adding clock-names, so that the order can be checked
> > > automatically?
> > 
> > I don't personally care for doing so, but if your heart is set on having
> > them, then sure.
> 
> Usually the CAN driver needs to have the clock rate of the clocks that
> the basis for the CAN bus clock. Looking at the clocks description it's
> probably the 2nd one.
> 
> With clock-names we can automatically check that the 2nd clock is always
> the CAN clock.

I think we already had this discussion on v1, where I said that the
binding requires the clocks to be in that order, regardless of whether
or not clock-names is provided. You feel more strongly about it than I
do, so I will add them when I get around to sending a v3.

Cheers,
Conor.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Marc Kleine-Budde]

[-- Attachment #1: Type: text/plain, Size: 2918 bytes --]

On 22.01.2024 14:56:09, Conor Dooley wrote:
> On Mon, Jan 22, 2024 at 03:46:04PM +0100, Marc Kleine-Budde wrote:
> > On 22.01.2024 14:21:04, Conor Dooley wrote:
> > > On Mon, Jan 22, 2024 at 02:13:16PM +0100, Marc Kleine-Budde wrote:
> > > > On 22.01.2024 12:19:50, Conor Dooley wrote:
> > > > > From: Conor Dooley <conor.dooley@microchip.com>
> > > > > 
> > > > > The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
> > > > > CAN bus clock. The bus clock was omitted when the binding was written,
> > > > > but is required for operation. Make up for lost time and add it.
> > > > > 
> > > > > Cautionary tale in adding bindings without having implemented a real
> > > > > user for them perhaps.
> > > > > 
> > > > > Fixes: c878d518d7b6 ("dt-bindings: can: mpfs: document the mpfs CAN controller")
> > > > > Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> > > > > ---
> > > > >  .../devicetree/bindings/net/can/microchip,mpfs-can.yaml     | 6 ++++--
> > > > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > > > > 
> > > > > diff --git a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > > > index 45aa3de7cf01..01e4d4a54df6 100644
> > > > > --- a/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > > > +++ b/Documentation/devicetree/bindings/net/can/microchip,mpfs-can.yaml
> > > > > @@ -24,7 +24,9 @@ properties:
> > > > >      maxItems: 1
> > > > >  
> > > > >    clocks:
> > > > > -    maxItems: 1
> > > > > +    items:
> > > > > +      - description: AHB peripheral clock
> > > > > +      - description: CAN bus clock
> > > > 
> > > > What about adding clock-names, so that the order can be checked
> > > > automatically?
> > > 
> > > I don't personally care for doing so, but if your heart is set on having
> > > them, then sure.
> > 
> > Usually the CAN driver needs to have the clock rate of the clocks that
> > the basis for the CAN bus clock. Looking at the clocks description it's
> > probably the 2nd one.
> > 
> > With clock-names we can automatically check that the 2nd clock is always
> > the CAN clock.
> 
> I think we already had this discussion on v1, where I said that the
> binding requires the clocks to be in that order, regardless of whether
> or not clock-names is provided. You feel more strongly about it than I
> do, so I will add them when I get around to sending a v3.

Yes, this discussion sounded very familiar to me, never mind. Keep it as
is, and let's get this binding and the CAN driver upstream!

regards,
Marc

-- 
Pengutronix e.K.                 | Marc Kleine-Budde          |
Embedded Linux                   | https://www.pengutronix.de |
Vertretung Nürnberg              | Phone: +49-5121-206917-129 |
Amtsgericht Hildesheim, HRA 2686 | Fax:   +49-5121-206917-9   |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 307 bytes --]

On Mon, Jan 22, 2024 at 04:31:32PM +0100, Marc Kleine-Budde wrote:
> let's get this binding and the CAN driver upstream!

FWIW, the driver seems to be in a good state (or as good as it can to
someone unaware of the goings on of CAN). Hopefully the author gets
a chance to send it out soon.

Cheers,
Conor.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 864 bytes --]

On Mon, Jan 22, 2024 at 04:31:32PM +0100, Marc Kleine-Budde wrote:
> On 22.01.2024 14:56:09, Conor Dooley wrote:

> > I think we already had this discussion on v1, where I said that the
> > binding requires the clocks to be in that order, regardless of whether
> > or not clock-names is provided. You feel more strongly about it than I
> > do, so I will add them when I get around to sending a v3.
> 
> Yes, this discussion sounded very familiar to me, never mind. Keep it as
> is, and let's get this binding and the CAN driver upstream!

BTW, I didn't see an ack on this nor do I see it in linux-next (yet).
Are you expecting the patch to go with the rest via the clock tree,
via the DT tree or will you be taking it with CAN stuff via netdev?

I can resend this one patch with a netdev appropriate subject prefix
if you like.

Thanks,
Conor.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Marc Kleine-Budde]

[-- Attachment #1: Type: text/plain, Size: 1258 bytes --]

On 30.01.2024 12:11:44, Conor Dooley wrote:
> On Mon, Jan 22, 2024 at 04:31:32PM +0100, Marc Kleine-Budde wrote:
> > On 22.01.2024 14:56:09, Conor Dooley wrote:
> 
> > > I think we already had this discussion on v1, where I said that the
> > > binding requires the clocks to be in that order, regardless of whether
> > > or not clock-names is provided. You feel more strongly about it than I
> > > do, so I will add them when I get around to sending a v3.
> > 
> > Yes, this discussion sounded very familiar to me, never mind. Keep it as
> > is, and let's get this binding and the CAN driver upstream!
> 
> BTW, I didn't see an ack on this nor do I see it in linux-next (yet).
> Are you expecting the patch to go with the rest via the clock tree,
> via the DT tree or will you be taking it with CAN stuff via netdev?
> 
> I can resend this one patch with a netdev appropriate subject prefix
> if you like.

Feel free to take the whole series via the clock tree.

Marc

-- 
Pengutronix e.K.                 | Marc Kleine-Budde          |
Embedded Linux                   | https://www.pengutronix.de |
Vertretung Nürnberg              | Phone: +49-5121-206917-129 |
Amtsgericht Hildesheim, HRA 2686 | Fax:   +49-5121-206917-9   |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 1211 bytes --]

On Tue, Jan 30, 2024 at 02:01:09PM +0100, Marc Kleine-Budde wrote:
> On 30.01.2024 12:11:44, Conor Dooley wrote:
> > On Mon, Jan 22, 2024 at 04:31:32PM +0100, Marc Kleine-Budde wrote:
> > > On 22.01.2024 14:56:09, Conor Dooley wrote:
> > 
> > > > I think we already had this discussion on v1, where I said that the
> > > > binding requires the clocks to be in that order, regardless of whether
> > > > or not clock-names is provided. You feel more strongly about it than I
> > > > do, so I will add them when I get around to sending a v3.
> > > 
> > > Yes, this discussion sounded very familiar to me, never mind. Keep it as
> > > is, and let's get this binding and the CAN driver upstream!
> > 
> > BTW, I didn't see an ack on this nor do I see it in linux-next (yet).
> > Are you expecting the patch to go with the rest via the clock tree,
> > via the DT tree or will you be taking it with CAN stuff via netdev?
> > 
> > I can resend this one patch with a netdev appropriate subject prefix
> > if you like.
> 
> Feel free to take the whole series via the clock tree.

Okay, I've gone and applied all of the binding and driver patches to the
branch we will send to Stephen. Thanks!



[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 2/7] dt-bindings: can: mpfs: add missing required clock

[Krzysztof Kozlowski]

On 22/01/2024 13:19, Conor Dooley wrote:
> From: Conor Dooley <conor.dooley@microchip.com>
> 
> The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
> CAN bus clock. The bus clock was omitted when the binding was written,
> but is required for operation. Make up for lost time and add it.
> 
> Cautionary tale in adding bindings without having implemented a real
> user for them perhaps.
> 
> Fixes: c878d518d7b6 ("dt-bindings: can: mpfs: document the mpfs CAN controller")
> Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> ---
>  .../devicetree/bindings/net/can/microchip,mpfs-can.yaml     | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

Best regards,
Krzysztof

[PATCH v2 3/7] clk: microchip: mpfs: split MSSPLL in two

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

The MSSPLL is really two stages - there's the PLL itself and 4 outputs,
each with their own divider. The current driver models this as a single
entity, outputting a single clock, used for both the CPU and AHB/AXI
buses. The other 3 outputs are used for the eMMC, "user crypto" and CAN
controller. Split the MSSPLL in two, as a precursor to adding support
for the other 3 outputs, with the PLL itself as one "hw" clock and the
output divider stage as another.

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 drivers/clk/microchip/clk-mpfs.c | 188 ++++++++++++++++++++-----------
 1 file changed, 123 insertions(+), 65 deletions(-)

diff --git a/drivers/clk/microchip/clk-mpfs.c b/drivers/clk/microchip/clk-mpfs.c
index c8ffa755b58d..acf598a32ce0 100644
--- a/drivers/clk/microchip/clk-mpfs.c
+++ b/drivers/clk/microchip/clk-mpfs.c
@@ -30,6 +30,13 @@
 #define MSSPLL_POSTDIV_WIDTH	0x07u
 #define MSSPLL_FIXED_DIV	4u
 
+/*
+ * This clock ID is defined here, rather than the binding headers, as it is an
+ * internal clock only, and therefore has no consumers in other peripheral
+ * blocks.
+ */
+#define CLK_MSSPLL_INTERNAL	38u
+
 struct mpfs_clock_data {
 	struct device *dev;
 	void __iomem *base;
@@ -39,17 +46,27 @@ struct mpfs_clock_data {
 
 struct mpfs_msspll_hw_clock {
 	void __iomem *base;
+	struct clk_hw hw;
+	struct clk_init_data init;
 	unsigned int id;
 	u32 reg_offset;
 	u32 shift;
 	u32 width;
 	u32 flags;
-	struct clk_hw hw;
-	struct clk_init_data init;
 };
 
 #define to_mpfs_msspll_clk(_hw) container_of(_hw, struct mpfs_msspll_hw_clock, hw)
 
+struct mpfs_msspll_out_hw_clock {
+	void __iomem *base;
+	struct clk_hw hw;
+	struct clk_init_data init;
+	unsigned int id;
+	u32 flags;
+};
+
+#define to_mpfs_msspll_out_clk(_hw) container_of(_hw, struct mpfs_msspll_out_hw_clock, hw)
+
 struct mpfs_cfg_hw_clock {
 	struct clk_divider cfg;
 	struct clk_init_data init;
@@ -93,93 +110,40 @@ static const struct clk_div_table mpfs_div_rtcref_table[] = {
 	{ 0, 0 }
 };
 
+/*
+ * MSS PLL internal clock
+ */
+
 static unsigned long mpfs_clk_msspll_recalc_rate(struct clk_hw *hw, unsigned long prate)
-{
-	struct mpfs_msspll_hw_clock *msspll_hw = to_mpfs_msspll_clk(hw);
-	void __iomem *mult_addr = msspll_hw->base + msspll_hw->reg_offset;
-	void __iomem *ref_div_addr = msspll_hw->base + REG_MSSPLL_REF_CR;
-	void __iomem *postdiv_addr = msspll_hw->base + REG_MSSPLL_POSTDIV_CR;
-	u32 mult, ref_div, postdiv;
-
-	mult = readl_relaxed(mult_addr) >> MSSPLL_FBDIV_SHIFT;
-	mult &= clk_div_mask(MSSPLL_FBDIV_WIDTH);
-	ref_div = readl_relaxed(ref_div_addr) >> MSSPLL_REFDIV_SHIFT;
-	ref_div &= clk_div_mask(MSSPLL_REFDIV_WIDTH);
-	postdiv = readl_relaxed(postdiv_addr) >> MSSPLL_POSTDIV_SHIFT;
-	postdiv &= clk_div_mask(MSSPLL_POSTDIV_WIDTH);
-
-	return prate * mult / (ref_div * MSSPLL_FIXED_DIV * postdiv);
-}
-
-static long mpfs_clk_msspll_round_rate(struct clk_hw *hw, unsigned long rate, unsigned long *prate)
 {
 	struct mpfs_msspll_hw_clock *msspll_hw = to_mpfs_msspll_clk(hw);
 	void __iomem *mult_addr = msspll_hw->base + msspll_hw->reg_offset;
 	void __iomem *ref_div_addr = msspll_hw->base + REG_MSSPLL_REF_CR;
 	u32 mult, ref_div;
-	unsigned long rate_before_ctrl;
 
 	mult = readl_relaxed(mult_addr) >> MSSPLL_FBDIV_SHIFT;
 	mult &= clk_div_mask(MSSPLL_FBDIV_WIDTH);
 	ref_div = readl_relaxed(ref_div_addr) >> MSSPLL_REFDIV_SHIFT;
 	ref_div &= clk_div_mask(MSSPLL_REFDIV_WIDTH);
 
-	rate_before_ctrl = rate * (ref_div * MSSPLL_FIXED_DIV) / mult;
-
-	return divider_round_rate(hw, rate_before_ctrl, prate, NULL, MSSPLL_POSTDIV_WIDTH,
-				  msspll_hw->flags);
-}
-
-static int mpfs_clk_msspll_set_rate(struct clk_hw *hw, unsigned long rate, unsigned long prate)
-{
-	struct mpfs_msspll_hw_clock *msspll_hw = to_mpfs_msspll_clk(hw);
-	void __iomem *mult_addr = msspll_hw->base + msspll_hw->reg_offset;
-	void __iomem *ref_div_addr = msspll_hw->base + REG_MSSPLL_REF_CR;
-	void __iomem *postdiv_addr = msspll_hw->base + REG_MSSPLL_POSTDIV_CR;
-	u32 mult, ref_div, postdiv;
-	int divider_setting;
-	unsigned long rate_before_ctrl, flags;
-
-	mult = readl_relaxed(mult_addr) >> MSSPLL_FBDIV_SHIFT;
-	mult &= clk_div_mask(MSSPLL_FBDIV_WIDTH);
-	ref_div = readl_relaxed(ref_div_addr) >> MSSPLL_REFDIV_SHIFT;
-	ref_div &= clk_div_mask(MSSPLL_REFDIV_WIDTH);
-
-	rate_before_ctrl = rate * (ref_div * MSSPLL_FIXED_DIV) / mult;
-	divider_setting = divider_get_val(rate_before_ctrl, prate, NULL, MSSPLL_POSTDIV_WIDTH,
-					  msspll_hw->flags);
-
-	if (divider_setting < 0)
-		return divider_setting;
-
-	spin_lock_irqsave(&mpfs_clk_lock, flags);
-
-	postdiv = readl_relaxed(postdiv_addr);
-	postdiv &= ~(clk_div_mask(MSSPLL_POSTDIV_WIDTH) << MSSPLL_POSTDIV_SHIFT);
-	writel_relaxed(postdiv, postdiv_addr);
-
-	spin_unlock_irqrestore(&mpfs_clk_lock, flags);
-
-	return 0;
+	return prate * mult / (ref_div * MSSPLL_FIXED_DIV);
 }
 
 static const struct clk_ops mpfs_clk_msspll_ops = {
 	.recalc_rate = mpfs_clk_msspll_recalc_rate,
-	.round_rate = mpfs_clk_msspll_round_rate,
-	.set_rate = mpfs_clk_msspll_set_rate,
 };
 
 #define CLK_PLL(_id, _name, _parent, _shift, _width, _flags, _offset) {			\
 	.id = _id,									\
+	.flags = _flags,								\
 	.shift = _shift,								\
 	.width = _width,								\
 	.reg_offset = _offset,								\
-	.flags = _flags,								\
 	.hw.init = CLK_HW_INIT_PARENTS_DATA(_name, _parent, &mpfs_clk_msspll_ops, 0),	\
 }
 
 static struct mpfs_msspll_hw_clock mpfs_msspll_clks[] = {
-	CLK_PLL(CLK_MSSPLL, "clk_msspll", mpfs_ext_ref, MSSPLL_FBDIV_SHIFT,
+	CLK_PLL(CLK_MSSPLL_INTERNAL, "clk_msspll_internal", mpfs_ext_ref, MSSPLL_FBDIV_SHIFT,
 		MSSPLL_FBDIV_WIDTH, 0, REG_MSSPLL_SSCG_2_CR),
 };
 
@@ -196,7 +160,7 @@ static int mpfs_clk_register_mssplls(struct device *dev, struct mpfs_msspll_hw_c
 		ret = devm_clk_hw_register(dev, &msspll_hw->hw);
 		if (ret)
 			return dev_err_probe(dev, ret, "failed to register msspll id: %d\n",
-					     CLK_MSSPLL);
+					     CLK_MSSPLL_INTERNAL);
 
 		data->hw_data.hws[msspll_hw->id] = &msspll_hw->hw;
 	}
@@ -204,6 +168,94 @@ static int mpfs_clk_register_mssplls(struct device *dev, struct mpfs_msspll_hw_c
 	return 0;
 }
 
+/*
+ * MSS PLL output clocks
+ */
+
+static unsigned long mpfs_clk_msspll_out_recalc_rate(struct clk_hw *hw, unsigned long prate)
+{
+	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
+	void __iomem *postdiv_addr = msspll_out_hw->base + REG_MSSPLL_POSTDIV_CR;
+	u32 postdiv;
+
+	postdiv = readl_relaxed(postdiv_addr) >> MSSPLL_POSTDIV_SHIFT;
+	postdiv &= clk_div_mask(MSSPLL_POSTDIV_WIDTH);
+
+	return prate / postdiv;
+}
+
+static long mpfs_clk_msspll_out_round_rate(struct clk_hw *hw, unsigned long rate,
+					   unsigned long *prate)
+{
+	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
+
+	return divider_round_rate(hw, rate, prate, NULL, MSSPLL_POSTDIV_WIDTH,
+				  msspll_out_hw->flags);
+}
+
+static int mpfs_clk_msspll_out_set_rate(struct clk_hw *hw, unsigned long rate, unsigned long prate)
+{
+	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
+	void __iomem *postdiv_addr = msspll_out_hw->base + REG_MSSPLL_POSTDIV_CR;
+	u32 postdiv;
+	int divider_setting;
+	unsigned long flags;
+
+	divider_setting = divider_get_val(rate, prate, NULL, MSSPLL_POSTDIV_WIDTH,
+					  msspll_out_hw->flags);
+
+	if (divider_setting < 0)
+		return divider_setting;
+
+	spin_lock_irqsave(&mpfs_clk_lock, flags);
+
+	postdiv = readl_relaxed(postdiv_addr);
+	postdiv &= ~(clk_div_mask(MSSPLL_POSTDIV_WIDTH) << MSSPLL_POSTDIV_SHIFT);
+	writel_relaxed(postdiv, postdiv_addr);
+
+	spin_unlock_irqrestore(&mpfs_clk_lock, flags);
+
+	return 0;
+}
+
+static const struct clk_ops mpfs_clk_msspll_out_ops = {
+	.recalc_rate = mpfs_clk_msspll_out_recalc_rate,
+	.round_rate = mpfs_clk_msspll_out_round_rate,
+	.set_rate = mpfs_clk_msspll_out_set_rate,
+};
+
+#define CLK_PLL_OUT(_id, _name, _parent, _flags) {				\
+	.id = _id,								\
+	.flags = _flags,							\
+	.hw.init = CLK_HW_INIT(_name, _parent, &mpfs_clk_msspll_out_ops, 0),	\
+}
+
+static struct mpfs_msspll_out_hw_clock mpfs_msspll_out_clks[] = {
+	CLK_PLL_OUT(CLK_MSSPLL0, "clk_msspll", "clk_msspll_internal", 0),
+};
+
+static int mpfs_clk_register_msspll_outs(struct device *dev,
+					 struct mpfs_msspll_out_hw_clock *msspll_out_hws,
+					 unsigned int num_clks, struct mpfs_clock_data *data)
+{
+	unsigned int i;
+	int ret;
+
+	for (i = 0; i < num_clks; i++) {
+		struct mpfs_msspll_out_hw_clock *msspll_out_hw = &msspll_out_hws[i];
+
+		msspll_out_hw->base = data->msspll_base;
+		ret = devm_clk_hw_register(dev, &msspll_out_hw->hw);
+		if (ret)
+			return dev_err_probe(dev, ret, "failed to register msspll out id: %d\n",
+					     msspll_out_hw->id);
+
+		data->hw_data.hws[msspll_out_hw->id] = &msspll_out_hw->hw;
+	}
+
+	return 0;
+}
+
 /*
  * "CFG" clocks
  */
@@ -442,8 +494,8 @@ static int mpfs_clk_probe(struct platform_device *pdev)
 	int ret;
 
 	/* CLK_RESERVED is not part of clock arrays, so add 1 */
-	num_clks = ARRAY_SIZE(mpfs_msspll_clks) + ARRAY_SIZE(mpfs_cfg_clks)
-		   + ARRAY_SIZE(mpfs_periph_clks) + 1;
+	num_clks = ARRAY_SIZE(mpfs_msspll_clks) + ARRAY_SIZE(mpfs_msspll_out_clks)
+		   + ARRAY_SIZE(mpfs_cfg_clks)  + ARRAY_SIZE(mpfs_periph_clks) + 1;
 
 	clk_data = devm_kzalloc(dev, struct_size(clk_data, hw_data.hws, num_clks), GFP_KERNEL);
 	if (!clk_data)
@@ -466,6 +518,12 @@ static int mpfs_clk_probe(struct platform_device *pdev)
 	if (ret)
 		return ret;
 
+	ret = mpfs_clk_register_msspll_outs(dev, mpfs_msspll_out_clks,
+					    ARRAY_SIZE(mpfs_msspll_out_clks),
+					    clk_data);
+	if (ret)
+		return ret;
+
 	ret = mpfs_clk_register_cfgs(dev, mpfs_cfg_clks, ARRAY_SIZE(mpfs_cfg_clks), clk_data);
 	if (ret)
 		return ret;
-- 
2.43.0

[PATCH v2 4/7] clk: microchip: mpfs: setup for using other mss pll outputs

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

Now that the MSSPLL is split, and the "postdiv" divider of the
cpu/AHB/AXI bus clock is represented by its own "hw" struct, make the
shifts, register offset and width a parameter of the initialisation
macro, rather than using defines that only work for one of the four
outputs.
Configuring this at initialisaion paves the way for using the other
three output clocks, where the register offset, and the bit shift
within that register, will differ.

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 drivers/clk/microchip/clk-mpfs.c | 30 +++++++++++++++++++-----------
 1 file changed, 19 insertions(+), 11 deletions(-)

diff --git a/drivers/clk/microchip/clk-mpfs.c b/drivers/clk/microchip/clk-mpfs.c
index acf598a32ce0..911905d0433d 100644
--- a/drivers/clk/microchip/clk-mpfs.c
+++ b/drivers/clk/microchip/clk-mpfs.c
@@ -15,7 +15,8 @@
 
 /* address offset of control registers */
 #define REG_MSSPLL_REF_CR	0x08u
-#define REG_MSSPLL_POSTDIV_CR	0x10u
+#define REG_MSSPLL_POSTDIV01_CR	0x10u
+#define REG_MSSPLL_POSTDIV23_CR	0x14u
 #define REG_MSSPLL_SSCG_2_CR	0x2Cu
 #define REG_CLOCK_CONFIG_CR	0x08u
 #define REG_RTC_CLOCK_CR	0x0Cu
@@ -26,7 +27,7 @@
 #define MSSPLL_FBDIV_WIDTH	0x0Cu
 #define MSSPLL_REFDIV_SHIFT	0x08u
 #define MSSPLL_REFDIV_WIDTH	0x06u
-#define MSSPLL_POSTDIV_SHIFT	0x08u
+#define MSSPLL_POSTDIV02_SHIFT	0x08u
 #define MSSPLL_POSTDIV_WIDTH	0x07u
 #define MSSPLL_FIXED_DIV	4u
 
@@ -62,6 +63,9 @@ struct mpfs_msspll_out_hw_clock {
 	struct clk_hw hw;
 	struct clk_init_data init;
 	unsigned int id;
+	u32 reg_offset;
+	u32 shift;
+	u32 width;
 	u32 flags;
 };
 
@@ -175,11 +179,11 @@ static int mpfs_clk_register_mssplls(struct device *dev, struct mpfs_msspll_hw_c
 static unsigned long mpfs_clk_msspll_out_recalc_rate(struct clk_hw *hw, unsigned long prate)
 {
 	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
-	void __iomem *postdiv_addr = msspll_out_hw->base + REG_MSSPLL_POSTDIV_CR;
+	void __iomem *postdiv_addr = msspll_out_hw->base + msspll_out_hw->reg_offset;
 	u32 postdiv;
 
-	postdiv = readl_relaxed(postdiv_addr) >> MSSPLL_POSTDIV_SHIFT;
-	postdiv &= clk_div_mask(MSSPLL_POSTDIV_WIDTH);
+	postdiv = readl_relaxed(postdiv_addr) >> msspll_out_hw->shift;
+	postdiv &= clk_div_mask(msspll_out_hw->width);
 
 	return prate / postdiv;
 }
@@ -189,19 +193,19 @@ static long mpfs_clk_msspll_out_round_rate(struct clk_hw *hw, unsigned long rate
 {
 	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
 
-	return divider_round_rate(hw, rate, prate, NULL, MSSPLL_POSTDIV_WIDTH,
+	return divider_round_rate(hw, rate, prate, NULL, msspll_out_hw->width,
 				  msspll_out_hw->flags);
 }
 
 static int mpfs_clk_msspll_out_set_rate(struct clk_hw *hw, unsigned long rate, unsigned long prate)
 {
 	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
-	void __iomem *postdiv_addr = msspll_out_hw->base + REG_MSSPLL_POSTDIV_CR;
+	void __iomem *postdiv_addr = msspll_out_hw->base + msspll_out_hw->reg_offset;
 	u32 postdiv;
 	int divider_setting;
 	unsigned long flags;
 
-	divider_setting = divider_get_val(rate, prate, NULL, MSSPLL_POSTDIV_WIDTH,
+	divider_setting = divider_get_val(rate, prate, NULL, msspll_out_hw->width,
 					  msspll_out_hw->flags);
 
 	if (divider_setting < 0)
@@ -210,7 +214,7 @@ static int mpfs_clk_msspll_out_set_rate(struct clk_hw *hw, unsigned long rate, u
 	spin_lock_irqsave(&mpfs_clk_lock, flags);
 
 	postdiv = readl_relaxed(postdiv_addr);
-	postdiv &= ~(clk_div_mask(MSSPLL_POSTDIV_WIDTH) << MSSPLL_POSTDIV_SHIFT);
+	postdiv &= ~(clk_div_mask(msspll_out_hw->width) << msspll_out_hw->shift);
 	writel_relaxed(postdiv, postdiv_addr);
 
 	spin_unlock_irqrestore(&mpfs_clk_lock, flags);
@@ -224,14 +228,18 @@ static const struct clk_ops mpfs_clk_msspll_out_ops = {
 	.set_rate = mpfs_clk_msspll_out_set_rate,
 };
 
-#define CLK_PLL_OUT(_id, _name, _parent, _flags) {				\
+#define CLK_PLL_OUT(_id, _name, _parent, _flags, _shift, _width, _offset) {	\
 	.id = _id,								\
+	.shift = _shift,							\
+	.width = _width,							\
+	.reg_offset = _offset,							\
 	.flags = _flags,							\
 	.hw.init = CLK_HW_INIT(_name, _parent, &mpfs_clk_msspll_out_ops, 0),	\
 }
 
 static struct mpfs_msspll_out_hw_clock mpfs_msspll_out_clks[] = {
-	CLK_PLL_OUT(CLK_MSSPLL0, "clk_msspll", "clk_msspll_internal", 0),
+	CLK_PLL_OUT(CLK_MSSPLL0, "clk_msspll", "clk_msspll_internal", 0,
+		    MSSPLL_POSTDIV02_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV01_CR),
 };
 
 static int mpfs_clk_register_msspll_outs(struct device *dev,
-- 
2.43.0

[PATCH v2 5/7] clk: microchip: mpfs: add missing MSSPLL outputs

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

The MSSPLL has 4 outputs, of which only the cpu/axi/ahb clock parent is
currently implemented.
Add the CAN clock too, as that'll be needed by the driver for the CAN
controller and uses output 3.
While we are here, the other two missing clocks, used by the eMMC/SD
controller and by the "user crypto".

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 drivers/clk/microchip/clk-mpfs.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/clk/microchip/clk-mpfs.c b/drivers/clk/microchip/clk-mpfs.c
index 911905d0433d..bed6adbb8a70 100644
--- a/drivers/clk/microchip/clk-mpfs.c
+++ b/drivers/clk/microchip/clk-mpfs.c
@@ -28,6 +28,7 @@
 #define MSSPLL_REFDIV_SHIFT	0x08u
 #define MSSPLL_REFDIV_WIDTH	0x06u
 #define MSSPLL_POSTDIV02_SHIFT	0x08u
+#define MSSPLL_POSTDIV13_SHIFT	0x18u
 #define MSSPLL_POSTDIV_WIDTH	0x07u
 #define MSSPLL_FIXED_DIV	4u
 
@@ -240,6 +241,12 @@ static const struct clk_ops mpfs_clk_msspll_out_ops = {
 static struct mpfs_msspll_out_hw_clock mpfs_msspll_out_clks[] = {
 	CLK_PLL_OUT(CLK_MSSPLL0, "clk_msspll", "clk_msspll_internal", 0,
 		    MSSPLL_POSTDIV02_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV01_CR),
+	CLK_PLL_OUT(CLK_MSSPLL1, "clk_msspll1", "clk_msspll_internal", 0,
+		    MSSPLL_POSTDIV13_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV01_CR),
+	CLK_PLL_OUT(CLK_MSSPLL2, "clk_msspll2", "clk_msspll_internal", 0,
+		    MSSPLL_POSTDIV02_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV23_CR),
+	CLK_PLL_OUT(CLK_MSSPLL3, "clk_msspll3", "clk_msspll_internal", 0,
+		    MSSPLL_POSTDIV13_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV23_CR),
 };
 
 static int mpfs_clk_register_msspll_outs(struct device *dev,
-- 
2.43.0

[PATCH v2 6/7] clk: microchip: mpfs: convert MSSPLL outputs to clk_divider

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

After splitting the MSSPLL in two, the PLL outputs have become
open-coded versions of clk_divider. Drop the custom clk ops structs, and
instead use the generic clk_divider_ops.

Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 drivers/clk/microchip/clk-mpfs.c | 81 ++++++--------------------------
 1 file changed, 14 insertions(+), 67 deletions(-)

diff --git a/drivers/clk/microchip/clk-mpfs.c b/drivers/clk/microchip/clk-mpfs.c
index bed6adbb8a70..22eab91a6712 100644
--- a/drivers/clk/microchip/clk-mpfs.c
+++ b/drivers/clk/microchip/clk-mpfs.c
@@ -61,13 +61,10 @@ struct mpfs_msspll_hw_clock {
 
 struct mpfs_msspll_out_hw_clock {
 	void __iomem *base;
-	struct clk_hw hw;
+	struct clk_divider output;
 	struct clk_init_data init;
 	unsigned int id;
 	u32 reg_offset;
-	u32 shift;
-	u32 width;
-	u32 flags;
 };
 
 #define to_mpfs_msspll_out_clk(_hw) container_of(_hw, struct mpfs_msspll_out_hw_clock, hw)
@@ -177,75 +174,25 @@ static int mpfs_clk_register_mssplls(struct device *dev, struct mpfs_msspll_hw_c
  * MSS PLL output clocks
  */
 
-static unsigned long mpfs_clk_msspll_out_recalc_rate(struct clk_hw *hw, unsigned long prate)
-{
-	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
-	void __iomem *postdiv_addr = msspll_out_hw->base + msspll_out_hw->reg_offset;
-	u32 postdiv;
-
-	postdiv = readl_relaxed(postdiv_addr) >> msspll_out_hw->shift;
-	postdiv &= clk_div_mask(msspll_out_hw->width);
-
-	return prate / postdiv;
-}
-
-static long mpfs_clk_msspll_out_round_rate(struct clk_hw *hw, unsigned long rate,
-					   unsigned long *prate)
-{
-	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
-
-	return divider_round_rate(hw, rate, prate, NULL, msspll_out_hw->width,
-				  msspll_out_hw->flags);
-}
-
-static int mpfs_clk_msspll_out_set_rate(struct clk_hw *hw, unsigned long rate, unsigned long prate)
-{
-	struct mpfs_msspll_out_hw_clock *msspll_out_hw = to_mpfs_msspll_out_clk(hw);
-	void __iomem *postdiv_addr = msspll_out_hw->base + msspll_out_hw->reg_offset;
-	u32 postdiv;
-	int divider_setting;
-	unsigned long flags;
-
-	divider_setting = divider_get_val(rate, prate, NULL, msspll_out_hw->width,
-					  msspll_out_hw->flags);
-
-	if (divider_setting < 0)
-		return divider_setting;
-
-	spin_lock_irqsave(&mpfs_clk_lock, flags);
-
-	postdiv = readl_relaxed(postdiv_addr);
-	postdiv &= ~(clk_div_mask(msspll_out_hw->width) << msspll_out_hw->shift);
-	writel_relaxed(postdiv, postdiv_addr);
-
-	spin_unlock_irqrestore(&mpfs_clk_lock, flags);
-
-	return 0;
-}
-
-static const struct clk_ops mpfs_clk_msspll_out_ops = {
-	.recalc_rate = mpfs_clk_msspll_out_recalc_rate,
-	.round_rate = mpfs_clk_msspll_out_round_rate,
-	.set_rate = mpfs_clk_msspll_out_set_rate,
-};
-
 #define CLK_PLL_OUT(_id, _name, _parent, _flags, _shift, _width, _offset) {	\
 	.id = _id,								\
-	.shift = _shift,							\
-	.width = _width,							\
+	.output.shift = _shift,							\
+	.output.width = _width,							\
+	.output.table = NULL,							\
 	.reg_offset = _offset,							\
-	.flags = _flags,							\
-	.hw.init = CLK_HW_INIT(_name, _parent, &mpfs_clk_msspll_out_ops, 0),	\
+	.output.flags = _flags,							\
+	.output.hw.init = CLK_HW_INIT(_name, _parent, &clk_divider_ops, 0),	\
+	.output.lock = &mpfs_clk_lock,						\
 }
 
 static struct mpfs_msspll_out_hw_clock mpfs_msspll_out_clks[] = {
-	CLK_PLL_OUT(CLK_MSSPLL0, "clk_msspll", "clk_msspll_internal", 0,
+	CLK_PLL_OUT(CLK_MSSPLL0, "clk_msspll", "clk_msspll_internal", CLK_DIVIDER_ONE_BASED,
 		    MSSPLL_POSTDIV02_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV01_CR),
-	CLK_PLL_OUT(CLK_MSSPLL1, "clk_msspll1", "clk_msspll_internal", 0,
+	CLK_PLL_OUT(CLK_MSSPLL1, "clk_msspll1", "clk_msspll_internal", CLK_DIVIDER_ONE_BASED,
 		    MSSPLL_POSTDIV13_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV01_CR),
-	CLK_PLL_OUT(CLK_MSSPLL2, "clk_msspll2", "clk_msspll_internal", 0,
+	CLK_PLL_OUT(CLK_MSSPLL2, "clk_msspll2", "clk_msspll_internal", CLK_DIVIDER_ONE_BASED,
 		    MSSPLL_POSTDIV02_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV23_CR),
-	CLK_PLL_OUT(CLK_MSSPLL3, "clk_msspll3", "clk_msspll_internal", 0,
+	CLK_PLL_OUT(CLK_MSSPLL3, "clk_msspll3", "clk_msspll_internal", CLK_DIVIDER_ONE_BASED,
 		    MSSPLL_POSTDIV13_SHIFT, MSSPLL_POSTDIV_WIDTH, REG_MSSPLL_POSTDIV23_CR),
 };
 
@@ -259,13 +206,13 @@ static int mpfs_clk_register_msspll_outs(struct device *dev,
 	for (i = 0; i < num_clks; i++) {
 		struct mpfs_msspll_out_hw_clock *msspll_out_hw = &msspll_out_hws[i];
 
-		msspll_out_hw->base = data->msspll_base;
-		ret = devm_clk_hw_register(dev, &msspll_out_hw->hw);
+		msspll_out_hw->output.reg = data->msspll_base + msspll_out_hw->reg_offset;
+		ret = devm_clk_hw_register(dev, &msspll_out_hw->output.hw);
 		if (ret)
 			return dev_err_probe(dev, ret, "failed to register msspll out id: %d\n",
 					     msspll_out_hw->id);
 
-		data->hw_data.hws[msspll_out_hw->id] = &msspll_out_hw->hw;
+		data->hw_data.hws[msspll_out_hw->id] = &msspll_out_hw->output.hw;
 	}
 
 	return 0;
-- 
2.43.0

[PATCH v2 7/7] riscv: dts: microchip: add missing CAN bus clocks

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

The CAN controller on PolarFire SoC has an AHB peripheral clock _and_ a
CAN bus clock. The bus clock was omitted when the binding was written,
but is required for operation. Make up for lost time and add to the DT.

Fixes: 38a71fc04895 ("riscv: dts: microchip: add mpfs's CAN controllers")
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 arch/riscv/boot/dts/microchip/mpfs.dtsi | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/riscv/boot/dts/microchip/mpfs.dtsi b/arch/riscv/boot/dts/microchip/mpfs.dtsi
index 266489d43912..4d70df0f908c 100644
--- a/arch/riscv/boot/dts/microchip/mpfs.dtsi
+++ b/arch/riscv/boot/dts/microchip/mpfs.dtsi
@@ -416,7 +416,7 @@ i2c1: i2c@2010b000 {
 		can0: can@2010c000 {
 			compatible = "microchip,mpfs-can";
 			reg = <0x0 0x2010c000 0x0 0x1000>;
-			clocks = <&clkcfg CLK_CAN0>;
+			clocks = <&clkcfg CLK_CAN0>, <&clkcfg CLK_MSSPLL3>;
 			interrupt-parent = <&plic>;
 			interrupts = <56>;
 			status = "disabled";
@@ -425,7 +425,7 @@ can0: can@2010c000 {
 		can1: can@2010d000 {
 			compatible = "microchip,mpfs-can";
 			reg = <0x0 0x2010d000 0x0 0x1000>;
-			clocks = <&clkcfg CLK_CAN1>;
+			clocks = <&clkcfg CLK_CAN1>, <&clkcfg CLK_MSSPLL3>;
 			interrupt-parent = <&plic>;
 			interrupts = <57>;
 			status = "disabled";
-- 
2.43.0

Re: (subset) [PATCH v2 0/7] MPFS clock fixes required for correct CAN clock modeling

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

On Mon, 22 Jan 2024 12:19:48 +0000, Conor Dooley wrote:
> From: Conor Dooley <conor.dooley@microchip.com>
> 
> While reviewing a CAN driver internally for MPFS [1], I realised
> that the modeling of the MSSPLL such that only one of its outputs could
> be used was not correct. The CAN controllers on MPFS take 2 input
> clocks - one that is the bus clock, acquired from the main MSSPLL and
> a second clock for the AHB interface to the result of the SoC.
> Currently the binding for the CAN controllers and the represetnation
> of the MSSPLL only allows for one of these clocks.
> Modify the binding and devicetree to expect two clocks and rework the
> main clock controller driver for MPFS such that it is capable of
> providing multiple outputs from the MSSPLL.
> 
> [...]

And this one is applied to riscv-dt-for-next. I don't think sending this
for the -rcs is needed as there's no impact until the CAN driver shows up.

[7/7] riscv: dts: microchip: add missing CAN bus clocks
      https://git.kernel.org/conor/c/6c7353836a91

Thanks,
Conor.