[PATCH 0/7] sy7636a: power the panel only during updates

[PATCH 0/7] sy7636a: power the panel only during updates

[Pablo Mazzini]

The SY7636A hwmon driver keeps VCOM on from probe, so a display driver
can not power the panel only during updates.

Read the temperature in the regulator driver instead, as the TPS65185
and FP9931 drivers do, and remove the hwmon driver. Boards that relied
on VCOM staying on mark it always-on. The regulator also switches VCOM
with VCOM_EN, gains VCOM voltage control and binds to the I2C device
directly.

Tested on a PocketBook Verse (B288).

Pablo Mazzini (7):
  ARM: dts: imx: keep the SY7636A VCOM on
  regulator: sy7636a: fix the chip name
  regulator: sy7636a: add VCOM voltage control
  regulator: sy7636a: wait for power good, then switch VCOM_EN
  regulator: sy7636a: report the temperature
  regulator: sy7636a: bind to the I2C device directly
  ARM: imx_v6_v7_defconfig: drop the SY7636A MFD and hwmon options

 Documentation/hwmon/index.rst                 |   1 -
 Documentation/hwmon/sy7636a-hwmon.rst         |  26 --
 arch/arm/boot/dts/nxp/imx/e70k02.dtsi         |   1 +
 .../boot/dts/nxp/imx/imx7d-remarkable2.dts    |   1 +
 arch/arm/configs/imx_v6_v7_defconfig          |   2 -
 drivers/hwmon/Kconfig                         |  10 -
 drivers/hwmon/Makefile                        |   1 -
 drivers/hwmon/sy7636a-hwmon.c                 | 102 --------
 drivers/mfd/Kconfig                           |  10 -
 drivers/mfd/simple-mfd-i2c.c                  |  11 -
 drivers/regulator/Kconfig                     |   6 +-
 drivers/regulator/sy7636a-regulator.c         | 240 +++++++++++++++---
 include/linux/mfd/sy7636a.h                   |  34 ---
 13 files changed, 205 insertions(+), 240 deletions(-)
 delete mode 100644 Documentation/hwmon/sy7636a-hwmon.rst
 delete mode 100644 drivers/hwmon/sy7636a-hwmon.c
 delete mode 100644 include/linux/mfd/sy7636a.h


base-commit: aac26bee2287c88af5be5a5ff96d783b19a28790
-- 
2.52.0

[PATCH 1/7] ARM: dts: imx: keep the SY7636A VCOM on

[Pablo Mazzini]

The SY7636A hwmon driver turns VCOM on at probe. The reMarkable 2, Kobo
Libra H2O and Tolino Vision 5 have had the rails on so far. Mark VCOM
always-on before the driver stops doing it.

Signed-off-by: Pablo Mazzini <pmazzini@gmail.com>
---
 arch/arm/boot/dts/nxp/imx/e70k02.dtsi           | 1 +
 arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts | 1 +
 2 files changed, 2 insertions(+)

diff --git a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
index dc7ff108aa85..bd9c6c6fa117 100644
--- a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
+++ b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
@@ -154,6 +154,7 @@ sy7636: pmic@62 {
 		regulators {
 			reg_epdpmic: vcom {
 				regulator-name = "vcom";
+				regulator-always-on;
 			};
 		};
 	};
diff --git a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
index ff9d50942884..6cf5603c0752 100644
--- a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
+++ b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
@@ -327,6 +327,7 @@ sy7636a: pmic@62 {
 		regulators {
 			reg_epdpmic: vcom {
 				regulator-name = "vcom";
+				regulator-always-on;
 				regulator-boot-on;
 			};
 		};
-- 
2.52.0

Re: [PATCH 1/7] ARM: dts: imx: keep the SY7636A VCOM on

[Andreas Kemnade]

On Mon, 26 Oct 2026 00:39:12 +0000
Pablo Mazzini <pmazzini@gmail.com> wrote:

> The SY7636A hwmon driver turns VCOM on at probe. The reMarkable 2, Kobo
> Libra H2O and Tolino Vision 5 have had the rails on so far. Mark VCOM
> always-on before the driver stops doing it.
> 
> Signed-off-by: Pablo Mazzini <pmazzini@gmail.com>
> ---
>  arch/arm/boot/dts/nxp/imx/e70k02.dtsi           | 1 +
>  arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts | 1 +
>  2 files changed, 2 insertions(+)
> 
> diff --git a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> index dc7ff108aa85..bd9c6c6fa117 100644
> --- a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> +++ b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> @@ -154,6 +154,7 @@ sy7636: pmic@62 {
>  		regulators {
>  			reg_epdpmic: vcom {
>  				regulator-name = "vcom";
> +				regulator-always-on;
>  			};
>  		};
>  	};
No, please not. That regulator is in practice used with a not yet upstreamed
driver for the EPDC on that platform used to deal with the current kernels
fp9931 / tps65185 drivers. So it is very welcome to have it not always turned
on.


> diff --git a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> index ff9d50942884..6cf5603c0752 100644
> --- a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> +++ b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> @@ -327,6 +327,7 @@ sy7636a: pmic@62 {
>  		regulators {
>  			reg_epdpmic: vcom {
>  				regulator-name = "vcom";
> +				regulator-always-on;
>  				regulator-boot-on;
>  			};
>  		};


Here the situation is a bit different. If I remember correctly, there
is some special EPD logic in userspace on top of some ordinary kernel
display driver. But thesse things are missing is devicetree?! But
that might need to learn to deal with regulator enabling/disabling.

Regulator always on is probably not desired here, either.

Regards,
Andreas

Re: [PATCH 1/7] ARM: dts: imx: keep the SY7636A VCOM on

[Pablo Mazzini]

Thanks, I will drop this patch in v2, so VCOM stays off until a consumer
enables it.

On Mon, Oct 26, 2026 at 7:31 AM Andreas Kemnade <andreas@kemnade.info> wrote:
>
> On Mon, 26 Oct 2026 00:39:12 +0000
> Pablo Mazzini <pmazzini@gmail.com> wrote:
>
> > The SY7636A hwmon driver turns VCOM on at probe. The reMarkable 2, Kobo
> > Libra H2O and Tolino Vision 5 have had the rails on so far. Mark VCOM
> > always-on before the driver stops doing it.
> >
> > Signed-off-by: Pablo Mazzini <pmazzini@gmail.com>
> > ---
> >  arch/arm/boot/dts/nxp/imx/e70k02.dtsi           | 1 +
> >  arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts | 1 +
> >  2 files changed, 2 insertions(+)
> >
> > diff --git a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > index dc7ff108aa85..bd9c6c6fa117 100644
> > --- a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > +++ b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > @@ -154,6 +154,7 @@ sy7636: pmic@62 {
> >               regulators {
> >                       reg_epdpmic: vcom {
> >                               regulator-name = "vcom";
> > +                             regulator-always-on;
> >                       };
> >               };
> >       };
> No, please not. That regulator is in practice used with a not yet upstreamed
> driver for the EPDC on that platform used to deal with the current kernels
> fp9931 / tps65185 drivers. So it is very welcome to have it not always turned
> on.
>
>
> > diff --git a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > index ff9d50942884..6cf5603c0752 100644
> > --- a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > +++ b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > @@ -327,6 +327,7 @@ sy7636a: pmic@62 {
> >               regulators {
> >                       reg_epdpmic: vcom {
> >                               regulator-name = "vcom";
> > +                             regulator-always-on;
> >                               regulator-boot-on;
> >                       };
> >               };
>
>
> Here the situation is a bit different. If I remember correctly, there
> is some special EPD logic in userspace on top of some ordinary kernel
> display driver. But thesse things are missing is devicetree?! But
> that might need to learn to deal with regulator enabling/disabling.
>
> Regulator always on is probably not desired here, either.
>
> Regards,
> Andreas

Re: [PATCH 1/7] ARM: dts: imx: keep the SY7636A VCOM on

[Alistair]

On Mon, 2026-10-26 at 08:30 +0100, Andreas Kemnade wrote:
> On Mon, 26 Oct 2026 00:39:12 +0000
> Pablo Mazzini <pmazzini@gmail.com> wrote:
> 
> > The SY7636A hwmon driver turns VCOM on at probe. The reMarkable 2,
> > Kobo
> > Libra H2O and Tolino Vision 5 have had the rails on so far. Mark
> > VCOM
> > always-on before the driver stops doing it.
> > 
> > Signed-off-by: Pablo Mazzini <pmazzini@gmail.com>
> > ---
> >  arch/arm/boot/dts/nxp/imx/e70k02.dtsi           | 1 +
> >  arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts | 1 +
> >  2 files changed, 2 insertions(+)
> > 
> > diff --git a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > index dc7ff108aa85..bd9c6c6fa117 100644
> > --- a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > +++ b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > @@ -154,6 +154,7 @@ sy7636: pmic@62 {
> >  		regulators {
> >  			reg_epdpmic: vcom {
> >  				regulator-name = "vcom";
> > +				regulator-always-on;
> >  			};
> >  		};
> >  	};
> No, please not. That regulator is in practice used with a not yet
> upstreamed
> driver for the EPDC on that platform used to deal with the current
> kernels
> fp9931 / tps65185 drivers. So it is very welcome to have it not
> always turned
> on.
> 
> 
> > diff --git a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > index ff9d50942884..6cf5603c0752 100644
> > --- a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > +++ b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > @@ -327,6 +327,7 @@ sy7636a: pmic@62 {
> >  		regulators {
> >  			reg_epdpmic: vcom {
> >  				regulator-name = "vcom";
> > +				regulator-always-on;
> >  				regulator-boot-on;
> >  			};
> >  		};
> 
> 
> Here the situation is a bit different. If I remember correctly, there
> is some special EPD logic in userspace on top of some ordinary kernel
> display driver. But thesse things are missing is devicetree?! But

Yeah, the kernel just exposes a framebuffer and a special secret sauce
userspace program writes to that.

> that might need to learn to deal with regulator enabling/disabling.
> 
> Regulator always on is probably not desired here, either.

Yeah, I **think** we don't need that

Alistair

> 
> Regards,
> Andreas

Re: [PATCH 1/7] ARM: dts: imx: keep the SY7636A VCOM on

[Andreas Kemnade]

On Tue, 10 Nov 2026 15:23:03 +1000
Alistair <alistair@alistair23.me> wrote:

> On Mon, 2026-10-26 at 08:30 +0100, Andreas Kemnade wrote:
> > On Mon, 26 Oct 2026 00:39:12 +0000
> > Pablo Mazzini <pmazzini@gmail.com> wrote:
> >   
> > > The SY7636A hwmon driver turns VCOM on at probe. The reMarkable 2,
> > > Kobo
> > > Libra H2O and Tolino Vision 5 have had the rails on so far. Mark
> > > VCOM
> > > always-on before the driver stops doing it.
> > > 
> > > Signed-off-by: Pablo Mazzini <pmazzini@gmail.com>
> > > ---
> > >  arch/arm/boot/dts/nxp/imx/e70k02.dtsi           | 1 +
> > >  arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts | 1 +
> > >  2 files changed, 2 insertions(+)
> > > 
> > > diff --git a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > index dc7ff108aa85..bd9c6c6fa117 100644
> > > --- a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > +++ b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > @@ -154,6 +154,7 @@ sy7636: pmic@62 {
> > >  		regulators {
> > >  			reg_epdpmic: vcom {
> > >  				regulator-name = "vcom";
> > > +				regulator-always-on;
> > >  			};
> > >  		};
> > >  	};  
> > No, please not. That regulator is in practice used with a not yet
> > upstreamed
> > driver for the EPDC on that platform used to deal with the current
> > kernels
> > fp9931 / tps65185 drivers. So it is very welcome to have it not
> > always turned
> > on.
> > 
> >   
> > > diff --git a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > index ff9d50942884..6cf5603c0752 100644
> > > --- a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > +++ b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > @@ -327,6 +327,7 @@ sy7636a: pmic@62 {
> > >  		regulators {
> > >  			reg_epdpmic: vcom {
> > >  				regulator-name = "vcom";
> > > +				regulator-always-on;
> > >  				regulator-boot-on;
> > >  			};
> > >  		};  
> > 
> > 
> > Here the situation is a bit different. If I remember correctly, there
> > is some special EPD logic in userspace on top of some ordinary kernel
> > display driver. But thesse things are missing is devicetree?! But  
> 
> Yeah, the kernel just exposes a framebuffer and a special secret sauce
> userspace program writes to that.

Looking at that processor. It seems to have an EPDC like the i.MX6 has.
Looking even further. The Kobo factory kernel for i.MX6SLL devices specifies
fsl,imx7d-epdc as compatible for that.
So if your display is just connected to that EPDC, then you could probably
wire up things on top of my EPDC driver with DRM interface.

So you still too need to run your own fork for the display? 

> 
> > that might need to learn to deal with regulator enabling/disabling.
> > 
> > Regulator always on is probably not desired here, either.  
> 
> Yeah, I **think** we don't need that
> 
You just need to add logic to enable/disable the regulator (if it is not
wired up yet). But it is nothing from the perspective of mainline to worry
about.

Regards,
Andreas

Re: [PATCH 1/7] ARM: dts: imx: keep the SY7636A VCOM on

[Alistair]

On Tue, 2026-11-10 at 10:34 +0100, Andreas Kemnade wrote:
> On Tue, 10 Nov 2026 15:23:03 +1000
> Alistair <alistair@alistair23.me> wrote:
> 
> > On Mon, 2026-10-26 at 08:30 +0100, Andreas Kemnade wrote:
> > > On Mon, 26 Oct 2026 00:39:12 +0000
> > > Pablo Mazzini <pmazzini@gmail.com> wrote:
> > >   
> > > > The SY7636A hwmon driver turns VCOM on at probe. The reMarkable
> > > > 2,
> > > > Kobo
> > > > Libra H2O and Tolino Vision 5 have had the rails on so far.
> > > > Mark
> > > > VCOM
> > > > always-on before the driver stops doing it.
> > > > 
> > > > Signed-off-by: Pablo Mazzini <pmazzini@gmail.com>
> > > > ---
> > > >  arch/arm/boot/dts/nxp/imx/e70k02.dtsi           | 1 +
> > > >  arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts | 1 +
> > > >  2 files changed, 2 insertions(+)
> > > > 
> > > > diff --git a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > > b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > > index dc7ff108aa85..bd9c6c6fa117 100644
> > > > --- a/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > > +++ b/arch/arm/boot/dts/nxp/imx/e70k02.dtsi
> > > > @@ -154,6 +154,7 @@ sy7636: pmic@62 {
> > > >  		regulators {
> > > >  			reg_epdpmic: vcom {
> > > >  				regulator-name = "vcom";
> > > > +				regulator-always-on;
> > > >  			};
> > > >  		};
> > > >  	};  
> > > No, please not. That regulator is in practice used with a not yet
> > > upstreamed
> > > driver for the EPDC on that platform used to deal with the
> > > current
> > > kernels
> > > fp9931 / tps65185 drivers. So it is very welcome to have it not
> > > always turned
> > > on.
> > > 
> > >   
> > > > diff --git a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > > b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > > index ff9d50942884..6cf5603c0752 100644
> > > > --- a/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > > +++ b/arch/arm/boot/dts/nxp/imx/imx7d-remarkable2.dts
> > > > @@ -327,6 +327,7 @@ sy7636a: pmic@62 {
> > > >  		regulators {
> > > >  			reg_epdpmic: vcom {
> > > >  				regulator-name = "vcom";
> > > > +				regulator-always-on;
> > > >  				regulator-boot-on;
> > > >  			};
> > > >  		};  
> > > 
> > > 
> > > Here the situation is a bit different. If I remember correctly,
> > > there
> > > is some special EPD logic in userspace on top of some ordinary
> > > kernel
> > > display driver. But thesse things are missing is devicetree?!
> > > But  
> > 
> > Yeah, the kernel just exposes a framebuffer and a special secret
> > sauce
> > userspace program writes to that.
> 
> Looking at that processor. It seems to have an EPDC like the i.MX6
> has.

It has been awhile since I looked at this, but from memory the display
isn't connected to the EPDC. It's connected like a normal (non-eInk)
display. Which is why it needs special userspace software

> Looking even further. The Kobo factory kernel for i.MX6SLL devices
> specifies
> fsl,imx7d-epdc as compatible for that.
> So if your display is just connected to that EPDC, then you could
> probably
> wire up things on top of my EPDC driver with DRM interface.

I remember that it isn't that simple

> 
> So you still too need to run your own fork for the display? 

Yep, it needs a custom kernel (with the FB display driver in the
kernel) and then a custom userspace.

At the time there didn't seem to be anyway around that. I haven't
looked in a few years, but short of reverse engineering their userspace
(which would be easier to do today) there isn't much to do.

Alistair

> 
> > 
> > > that might need to learn to deal with regulator
> > > enabling/disabling.
> > > 
> > > Regulator always on is probably not desired here, either.  
> > 
> > Yeah, I **think** we don't need that
> > 
> You just need to add logic to enable/disable the regulator (if it is
> not
> wired up yet). But it is nothing from the perspective of mainline to
> worry
> about.
> 
> Regards,
> Andreas

Re: [PATCH 0/7] sy7636a: power the panel only during updates

[Andreas Kemnade]

On Mon, 26 Oct 2026 00:39:11 +0000
Pablo Mazzini <pmazzini@gmail.com> wrote:

> The SY7636A hwmon driver keeps VCOM on from probe, so a display driver
> can not power the panel only during updates.
> 
Yes, that is the common use case, power things on, tune update process
to the temperature, update, power off. So it makes completely sense.

> Read the temperature in the regulator driver instead, as the TPS65185
> and FP9931 drivers do, and remove the hwmon driver. Boards that relied
> on VCOM staying on mark it always-on. The regulator also switches VCOM
> with VCOM_EN, gains VCOM voltage control and binds to the I2C device
> directly.
> 
> Tested on a PocketBook Verse (B288).
> 
Which kernel do you use? Just curious, maybe some further upstreaming
actions might be coordinated or cross-reviewed.
There was discussion about in-kernel hwmon readout interface like it
exists for iio.

Regards,
Andreas

Re: [PATCH 0/7] sy7636a: power the panel only during updates

[Pablo Mazzini]

Mainline, on top of linux-next. I started looking at the hwmon IIO
bridge. Coordinating and cross-reviewing further work sounds good.

On Mon, Oct 26, 2026 at 9:37 AM Andreas Kemnade <andreas@kemnade.info> wrote:
>
> On Mon, 26 Oct 2026 00:39:11 +0000
> Pablo Mazzini <pmazzini@gmail.com> wrote:
>
> > The SY7636A hwmon driver keeps VCOM on from probe, so a display driver
> > can not power the panel only during updates.
> >
> Yes, that is the common use case, power things on, tune update process
> to the temperature, update, power off. So it makes completely sense.
>
> > Read the temperature in the regulator driver instead, as the TPS65185
> > and FP9931 drivers do, and remove the hwmon driver. Boards that relied
> > on VCOM staying on mark it always-on. The regulator also switches VCOM
> > with VCOM_EN, gains VCOM voltage control and binds to the I2C device
> > directly.
> >
> > Tested on a PocketBook Verse (B288).
> >
> Which kernel do you use? Just curious, maybe some further upstreaming
> actions might be coordinated or cross-reviewed.
> There was discussion about in-kernel hwmon readout interface like it
> exists for iio.
>
> Regards,
> Andreas