2019-05-25 19:05:16

by Martin Blumenstingl

[permalink] [raw]
Subject: [PATCH 0/4] ARM: dts: meson8b: add VDDEE / mali-supply

EC-100 and Odroid-C1 use a "copy" of the VCCK regulator as "VDDEE"
regulator. VDDEE supplies the Mali GPU and various other bits within
the SoC.

The VDDEE regulator is not exclusive to the Mali GPU so it must not
change it's voltage. The GPU OPP table has a fixed voltage for all
frequencies of 1.10V. This matches with what u-boot sets on my EC-100
and Odroid-C1.

Dependencies:
- compile time: patch #4 depends on my other patch "ARM: meson8b-mxq:
better support for the TRONFY MXQ" from [0]
- runtime: we don't want the kernel to change the output of the VDDEE
regulator to the maximum value. Thus the PWM driver has to be able
to read the PWM period and duty cycle from u-boot. This is supported
with my series called "pwm-meson: cleanups and improvements" from [1]


[0] https://patchwork.kernel.org/patch/10960283/
[1] https://patchwork.kernel.org/cover/10961073/


Martin Blumenstingl (4):
ARM: dts: meson8b: add the PWM_D output pin
ARM: dts: meson8b: ec100: add the VDDEE regulator
ARM: dts: meson8b: odroidc1: add the VDDEE regulator
ARM: dts: meson8b: mxq: add the VDDEE regulator

arch/arm/boot/dts/meson8b-ec100.dts | 31 +++++++++++++++++++++++---
arch/arm/boot/dts/meson8b-mxq.dts | 26 ++++++++++++++++++---
arch/arm/boot/dts/meson8b-odroidc1.dts | 27 +++++++++++++++++++---
arch/arm/boot/dts/meson8b.dtsi | 8 +++++++
4 files changed, 83 insertions(+), 9 deletions(-)

--
2.21.0


2019-05-25 19:06:02

by Martin Blumenstingl

[permalink] [raw]
Subject: [PATCH 1/4] ARM: dts: meson8b: add the PWM_D output pin

The PWM_D output is used for the VDDEE PWM regulator which supplies for
example the Mali GPU on the EC-100 and Odroid-C1 boards. Add the output
pin the VDDEE regulators can be added.

Signed-off-by: Martin Blumenstingl <[email protected]>
---
arch/arm/boot/dts/meson8b.dtsi | 8 ++++++++
1 file changed, 8 insertions(+)

diff --git a/arch/arm/boot/dts/meson8b.dtsi b/arch/arm/boot/dts/meson8b.dtsi
index e4134c63a48c..1b0106fbb0ac 100644
--- a/arch/arm/boot/dts/meson8b.dtsi
+++ b/arch/arm/boot/dts/meson8b.dtsi
@@ -401,6 +401,14 @@
};
};

+ pwm_d_pins: pwm-d {
+ mux {
+ groups = "pwm_d";
+ function = "pwm_d";
+ bias-disable;
+ };
+ };
+
uart_b0_pins: uart-b0 {
mux {
groups = "uart_tx_b0",
--
2.21.0

2019-06-12 07:02:12

by Kevin Hilman

[permalink] [raw]
Subject: Re: [PATCH 0/4] ARM: dts: meson8b: add VDDEE / mali-supply

Martin Blumenstingl <[email protected]> writes:

> EC-100 and Odroid-C1 use a "copy" of the VCCK regulator as "VDDEE"
> regulator. VDDEE supplies the Mali GPU and various other bits within
> the SoC.
>
> The VDDEE regulator is not exclusive to the Mali GPU so it must not
> change it's voltage. The GPU OPP table has a fixed voltage for all
> frequencies of 1.10V. This matches with what u-boot sets on my EC-100
> and Odroid-C1.
>
> Dependencies:
> - compile time: patch #4 depends on my other patch "ARM: meson8b-mxq:
> better support for the TRONFY MXQ" from [0]
> - runtime: we don't want the kernel to change the output of the VDDEE
> regulator to the maximum value. Thus the PWM driver has to be able
> to read the PWM period and duty cycle from u-boot. This is supported
> with my series called "pwm-meson: cleanups and improvements" from [1]

Just FYI... unless I hear otherwise, I'll wait for the PWM cleanups to
land before queuing this series.

Kevin

2019-06-12 19:10:36

by Martin Blumenstingl

[permalink] [raw]
Subject: Re: [PATCH 0/4] ARM: dts: meson8b: add VDDEE / mali-supply

On Wed, Jun 12, 2019 at 1:32 AM Kevin Hilman <[email protected]> wrote:
>
> Martin Blumenstingl <[email protected]> writes:
>
> > EC-100 and Odroid-C1 use a "copy" of the VCCK regulator as "VDDEE"
> > regulator. VDDEE supplies the Mali GPU and various other bits within
> > the SoC.
> >
> > The VDDEE regulator is not exclusive to the Mali GPU so it must not
> > change it's voltage. The GPU OPP table has a fixed voltage for all
> > frequencies of 1.10V. This matches with what u-boot sets on my EC-100
> > and Odroid-C1.
> >
> > Dependencies:
> > - compile time: patch #4 depends on my other patch "ARM: meson8b-mxq:
> > better support for the TRONFY MXQ" from [0]
> > - runtime: we don't want the kernel to change the output of the VDDEE
> > regulator to the maximum value. Thus the PWM driver has to be able
> > to read the PWM period and duty cycle from u-boot. This is supported
> > with my series called "pwm-meson: cleanups and improvements" from [1]
>
> Just FYI... unless I hear otherwise, I'll wait for the PWM cleanups to
> land before queuing this series.
I'm happy with that because I'm not sure what will happen *without*
the PWM improvements


Martin

2019-06-13 19:44:09

by Kevin Hilman

[permalink] [raw]
Subject: Re: [PATCH 0/4] ARM: dts: meson8b: add VDDEE / mali-supply

Martin Blumenstingl <[email protected]> writes:

> On Wed, Jun 12, 2019 at 1:32 AM Kevin Hilman <[email protected]> wrote:
>>
>> Martin Blumenstingl <[email protected]> writes:
>>
>> > EC-100 and Odroid-C1 use a "copy" of the VCCK regulator as "VDDEE"
>> > regulator. VDDEE supplies the Mali GPU and various other bits within
>> > the SoC.
>> >
>> > The VDDEE regulator is not exclusive to the Mali GPU so it must not
>> > change it's voltage. The GPU OPP table has a fixed voltage for all
>> > frequencies of 1.10V. This matches with what u-boot sets on my EC-100
>> > and Odroid-C1.
>> >
>> > Dependencies:
>> > - compile time: patch #4 depends on my other patch "ARM: meson8b-mxq:
>> > better support for the TRONFY MXQ" from [0]
>> > - runtime: we don't want the kernel to change the output of the VDDEE
>> > regulator to the maximum value. Thus the PWM driver has to be able
>> > to read the PWM period and duty cycle from u-boot. This is supported
>> > with my series called "pwm-meson: cleanups and improvements" from [1]
>>
>> Just FYI... unless I hear otherwise, I'll wait for the PWM cleanups to
>> land before queuing this series.
>
> I'm happy with that because I'm not sure what will happen *without*
> the PWM improvements

OK, then I'll rely on you to ping (or resend) on this series to remind
me when the PWM stuff has landed.

Thanks,

Kevin