[PATCH v3 0/2] can: m_can: support out-of-band wakeup capability

[PATCH v3 0/2] can: m_can: support out-of-band wakeup capability

[Kendall Willis]

This series adds support for the out-of-band wakeup capability in the m_can
driver. Some CAN devices are wakeup capable and can wake up the system
while being in a domain that is OFF. Allow the driver to set the
out-of-band wakeup flag during suspend if the "out-band-wakeup" property
is present in the device tree.

A corresponding DTS series enabling this feature on TI K3 AM62A, AM62D,
AM62P, and AM62X-LP SoCs has been submitted separately:
"arm64: dts: ti: add out-band-wakeup property to m_can devices"

Dependency
----------
The DT bindings are dependent on this PR in the dt-schema repository,
which defines the out-band-wakeup property:
https://github.com/devicetree-org/dt-schema/pull/205

Testing
-------
Tested suspend/resume with CAN wakeup on AM62P SK EVM.

Signed-off-by: Kendall Willis <k-willis@ti.com>
---
Changes in v3:
- Remove definition of "out-band-wakeup" property in dt binding.
- Allow "out-band-wakeup" property in dt binding.
- Link to v2: https://lore.kernel.org/r/20260820-temp-v2-0-dd3f8cacc15e@ti.com

Changes in v2:
- Create "out-band-wakeup" property in dt bindings.
- Use "out-band-wakeup" property to set the out_band_wakeup flag upon
  suspend.
- Rebase to v7.2
- Link to v1: https://lore.kernel.org/r/20260213-mcan-out-of-band-v1-1-af68d4c570b3@ti.com

---
Kendall Willis (2):
      dt-bindings: can: m_can: add out-band-wakeup property
      can: m_can: set out-of-band wakeup if wakeup pinctrl exists

 Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
 drivers/net/can/m_can/m_can.c                              | 6 +++++-
 2 files changed, 7 insertions(+), 1 deletion(-)
---
base-commit: 8d3ae59288f1e7d58d76558a6ee96d533bc5019f
change-id: 20260819-temp-93299bee27b7

Best regards,
-- 
Kendall Willis <k-willis@ti.com>

[PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[Kendall Willis]

Add the out-band-wakeup property as a possible property. The property
indicates that the device can wakeup the system even when its power
domain is off.

Signed-off-by: Kendall Willis <k-willis@ti.com>
---
 Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
index 2c9d37975bedd652b3060ab11ba75c37565edaad..0663beaa532bcc4a71cd6ad01520a4dfea294d0c 100644
--- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
+++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
@@ -150,6 +150,8 @@ properties:
     description:
       List of phandles to system idle states in which mcan can wakeup the system.
 
+  out-band-wakeup: true
+
 required:
   - compatible
   - reg

-- 
2.34.1

Re: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The newly added `out-band-wakeup` property lacks a type definition, description, and the required vendor prefix.
--

--- Patch [1]: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property ---
commit 18cc4cb55e5269767b220486da77c10e4826ca15
Author: Kendall Willis <k-willis@ti.com>

dt-bindings: can: m_can: add out-band-wakeup property

This commit adds the out-band-wakeup property to the bosch,m_can device tree
binding. The property indicates that the device can wake up the system even
when its power domain is off.

> diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> index 2c9d37975bedd..0663beaa532bc 100644
> --- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> +++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> @@ -150,6 +150,8 @@ properties:
>      description:
>        List of phandles to system idle states in which mcan can wakeup the system.
>  
> +  out-band-wakeup: true

[Severity: Medium]
Does this property need a type constraint and a vendor prefix?

By defining the property merely as true, schema validation lacks the ability to
enforce a specific data type, meaning it could silently allow arrays or strings
instead of a boolean flag.

Additionally, if this is a custom hardware property rather than a standard one,
should it have a vendor prefix to avoid namespace pollution?

> +
>  required:
>    - compatible
>    - reg

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260821-temp-v3-0-9ac1f8806929@ti.com?part=1

Re: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 1141 bytes --]

On Fri, Aug 21, 2026 at 11:07:00AM -0500, Kendall Willis wrote:
> Add the out-band-wakeup property as a possible property. The property
> indicates that the device can wakeup the system even when its power
> domain is off.
> 
> Signed-off-by: Kendall Willis <k-willis@ti.com>
> ---
>  Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> index 2c9d37975bedd652b3060ab11ba75c37565edaad..0663beaa532bcc4a71cd6ad01520a4dfea294d0c 100644
> --- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> +++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> @@ -150,6 +150,8 @@ properties:
>      description:
>        List of phandles to system idle states in which mcan can wakeup the system.
>  
> +  out-band-wakeup: true

Where is "-wakeup" defined genericly, or out-band-wakeup defined that
you're getting the property definition from?

> +
>  required:
>    - compatible
>    - reg
> 
> -- 
> 2.34.1
> 

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

Re: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[Kendall Willis]

On 17:45-20260821, Conor Dooley wrote:
> On Fri, Aug 21, 2026 at 11:07:00AM -0500, Kendall Willis wrote:
> > Add the out-band-wakeup property as a possible property. The property
> > indicates that the device can wakeup the system even when its power
> > domain is off.
> > 
> > Signed-off-by: Kendall Willis <k-willis@ti.com>
> > ---
> >  Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > index 2c9d37975bedd652b3060ab11ba75c37565edaad..0663beaa532bcc4a71cd6ad01520a4dfea294d0c 100644
> > --- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > +++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > @@ -150,6 +150,8 @@ properties:
> >      description:
> >        List of phandles to system idle states in which mcan can wakeup the system.
> >  
> > +  out-band-wakeup: true
> 
> Where is "-wakeup" defined genericly, or out-band-wakeup defined that
> you're getting the property definition from?

I sent a PR to define out-band-wakeup in the wakeup-source.yaml binding
in dt-schema repo. I figured it made more sense to be defined in the
wakeup-source.yaml binding to allow other drivers the option to use it.

dt-schema PR:
https://github.com/devicetree-org/dt-schema/pull/205

> 
> > +
> >  required:
> >    - compatible
> >    - reg
> > 
> > -- 
> > 2.34.1
> > 

Re: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 1934 bytes --]

On Fri, Aug 21, 2026 at 02:05:11PM -0500, Kendall Willis wrote:
> On 17:45-20260821, Conor Dooley wrote:
> > On Fri, Aug 21, 2026 at 11:07:00AM -0500, Kendall Willis wrote:
> > > Add the out-band-wakeup property as a possible property. The property
> > > indicates that the device can wakeup the system even when its power
> > > domain is off.
> > > 
> > > Signed-off-by: Kendall Willis <k-willis@ti.com>
> > > ---
> > >  Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
> > >  1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > index 2c9d37975bedd652b3060ab11ba75c37565edaad..0663beaa532bcc4a71cd6ad01520a4dfea294d0c 100644
> > > --- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > +++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > @@ -150,6 +150,8 @@ properties:
> > >      description:
> > >        List of phandles to system idle states in which mcan can wakeup the system.
> > >  
> > > +  out-band-wakeup: true
> > 
> > Where is "-wakeup" defined genericly, or out-band-wakeup defined that
> > you're getting the property definition from?
> 
> I sent a PR to define out-band-wakeup in the wakeup-source.yaml binding
> in dt-schema repo. I figured it made more sense to be defined in the
> wakeup-source.yaml binding to allow other drivers the option to use it.
> 
> dt-schema PR:
> https://github.com/devicetree-org/dt-schema/pull/205

This sounds like something you should be able to determine from
device specific compatibles for these platforms. Not sure why they are
not used for this particular device, but this is an indication to me
that that is a mistake.

Cheers,
Conor.

> 
> > 
> > > +
> > >  required:
> > >    - compatible
> > >    - reg
> > > 
> > > -- 
> > > 2.34.1
> > > 

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

Re: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[Krzysztof Kozlowski]

On Mon, Aug 24, 2026 at 05:50:20PM +0100, Conor Dooley wrote:
> On Fri, Aug 21, 2026 at 02:05:11PM -0500, Kendall Willis wrote:
> > On 17:45-20260821, Conor Dooley wrote:
> > > On Fri, Aug 21, 2026 at 11:07:00AM -0500, Kendall Willis wrote:
> > > > Add the out-band-wakeup property as a possible property. The property
> > > > indicates that the device can wakeup the system even when its power
> > > > domain is off.
> > > > 
> > > > Signed-off-by: Kendall Willis <k-willis@ti.com>
> > > > ---
> > > >  Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
> > > >  1 file changed, 2 insertions(+)
> > > > 
> > > > diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > > index 2c9d37975bedd652b3060ab11ba75c37565edaad..0663beaa532bcc4a71cd6ad01520a4dfea294d0c 100644
> > > > --- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > > +++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > > @@ -150,6 +150,8 @@ properties:
> > > >      description:
> > > >        List of phandles to system idle states in which mcan can wakeup the system.
> > > >  
> > > > +  out-band-wakeup: true
> > > 
> > > Where is "-wakeup" defined genericly, or out-band-wakeup defined that
> > > you're getting the property definition from?
> > 
> > I sent a PR to define out-band-wakeup in the wakeup-source.yaml binding
> > in dt-schema repo. I figured it made more sense to be defined in the
> > wakeup-source.yaml binding to allow other drivers the option to use it.
> > 
> > dt-schema PR:
> > https://github.com/devicetree-org/dt-schema/pull/205
> 
> This sounds like something you should be able to determine from
> device specific compatibles for these platforms. Not sure why they are
> not used for this particular device, but this is an indication to me
> that that is a mistake.

Yeah, it looks a bit too much replicating Linux PM detail. I can imagine
that some devices on some board are capable of waking up the system even
when power domain is off, but this does not look like a capability of
the actual device, but power domain. Device does not have different
wakeups. (by device I mean this individual schema - Bosch CAN)

Happy to see some bigger picture through DTS or board layouts. Actually
DTS, showing same Bosch CAN devices which are different on a board -
some are wakeups and some out of band wakeups - would illustrate it
better.

Best regards,
Krzysztof

Re: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 3047 bytes --]

On Fri, Aug 28, 2026 at 11:07:29AM +0200, Krzysztof Kozlowski wrote:
> On Mon, Aug 24, 2026 at 05:50:20PM +0100, Conor Dooley wrote:
> > On Fri, Aug 21, 2026 at 02:05:11PM -0500, Kendall Willis wrote:
> > > On 17:45-20260821, Conor Dooley wrote:
> > > > On Fri, Aug 21, 2026 at 11:07:00AM -0500, Kendall Willis wrote:
> > > > > Add the out-band-wakeup property as a possible property. The property
> > > > > indicates that the device can wakeup the system even when its power
> > > > > domain is off.
> > > > > 
> > > > > Signed-off-by: Kendall Willis <k-willis@ti.com>
> > > > > ---
> > > > >  Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
> > > > >  1 file changed, 2 insertions(+)
> > > > > 
> > > > > diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > > > index 2c9d37975bedd652b3060ab11ba75c37565edaad..0663beaa532bcc4a71cd6ad01520a4dfea294d0c 100644
> > > > > --- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > > > +++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
> > > > > @@ -150,6 +150,8 @@ properties:
> > > > >      description:
> > > > >        List of phandles to system idle states in which mcan can wakeup the system.
> > > > >  
> > > > > +  out-band-wakeup: true
> > > > 
> > > > Where is "-wakeup" defined genericly, or out-band-wakeup defined that
> > > > you're getting the property definition from?
> > > 
> > > I sent a PR to define out-band-wakeup in the wakeup-source.yaml binding
> > > in dt-schema repo. I figured it made more sense to be defined in the
> > > wakeup-source.yaml binding to allow other drivers the option to use it.
> > > 
> > > dt-schema PR:
> > > https://github.com/devicetree-org/dt-schema/pull/205
> > 
> > This sounds like something you should be able to determine from
> > device specific compatibles for these platforms. Not sure why they are
> > not used for this particular device, but this is an indication to me
> > that that is a mistake.
> 
> Yeah, it looks a bit too much replicating Linux PM detail. I can imagine
> that some devices on some board are capable of waking up the system even
> when power domain is off, but this does not look like a capability of
> the actual device, but power domain. Device does not have different
> wakeups. (by device I mean this individual schema - Bosch CAN)
> 
> Happy to see some bigger picture through DTS or board layouts. Actually
> DTS, showing same Bosch CAN devices which are different on a board -
> some are wakeups and some out of band wakeups - would illustrate it
> better.

As far as I could make out when I looked at the patch, "Bosch CAN" is an
IP that can be integrated on an SoC rather than a device on the board,
so I'm doubting that there's any difference in capability to wake up the
device caused by the board at all. If there is, it's between different
power domains on the SoC, as you say.

> 
> Best regards,
> Krzysztof
> 

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

Re: [PATCH v3 1/2] dt-bindings: can: m_can: add out-band-wakeup property

[Markus Schneider-Pargmann]

[-- Attachment #1: Type: text/plain, Size: 3504 bytes --]

Hi,

On Fri Aug 28, 2026 at 5:46 PM CEST, Conor Dooley wrote:
> On Fri, Aug 28, 2026 at 11:07:29AM +0200, Krzysztof Kozlowski wrote:
>> On Mon, Aug 24, 2026 at 05:50:20PM +0100, Conor Dooley wrote:
>> > On Fri, Aug 21, 2026 at 02:05:11PM -0500, Kendall Willis wrote:
>> > > On 17:45-20260821, Conor Dooley wrote:
>> > > > On Fri, Aug 21, 2026 at 11:07:00AM -0500, Kendall Willis wrote:
>> > > > > Add the out-band-wakeup property as a possible property. The property
>> > > > > indicates that the device can wakeup the system even when its power
>> > > > > domain is off.
>> > > > > 
>> > > > > Signed-off-by: Kendall Willis <k-willis@ti.com>
>> > > > > ---
>> > > > >  Documentation/devicetree/bindings/net/can/bosch,m_can.yaml | 2 ++
>> > > > >  1 file changed, 2 insertions(+)
>> > > > > 
>> > > > > diff --git a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
>> > > > > index 2c9d37975bedd652b3060ab11ba75c37565edaad..0663beaa532bcc4a71cd6ad01520a4dfea294d0c 100644
>> > > > > --- a/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
>> > > > > +++ b/Documentation/devicetree/bindings/net/can/bosch,m_can.yaml
>> > > > > @@ -150,6 +150,8 @@ properties:
>> > > > >      description:
>> > > > >        List of phandles to system idle states in which mcan can wakeup the system.
>> > > > >  
>> > > > > +  out-band-wakeup: true
>> > > > 
>> > > > Where is "-wakeup" defined genericly, or out-band-wakeup defined that
>> > > > you're getting the property definition from?
>> > > 
>> > > I sent a PR to define out-band-wakeup in the wakeup-source.yaml binding
>> > > in dt-schema repo. I figured it made more sense to be defined in the
>> > > wakeup-source.yaml binding to allow other drivers the option to use it.
>> > > 
>> > > dt-schema PR:
>> > > https://github.com/devicetree-org/dt-schema/pull/205
>> > 
>> > This sounds like something you should be able to determine from
>> > device specific compatibles for these platforms. Not sure why they are
>> > not used for this particular device, but this is an indication to me
>> > that that is a mistake.
>> 
>> Yeah, it looks a bit too much replicating Linux PM detail. I can imagine
>> that some devices on some board are capable of waking up the system even
>> when power domain is off, but this does not look like a capability of
>> the actual device, but power domain. Device does not have different
>> wakeups. (by device I mean this individual schema - Bosch CAN)
>> 
>> Happy to see some bigger picture through DTS or board layouts. Actually
>> DTS, showing same Bosch CAN devices which are different on a board -
>> some are wakeups and some out of band wakeups - would illustrate it
>> better.
>
> As far as I could make out when I looked at the patch, "Bosch CAN" is an
> IP that can be integrated on an SoC rather than a device on the board,
> so I'm doubting that there's any difference in capability to wake up the
> device caused by the board at all. If there is, it's between different
> power domains on the SoC, as you say.

m_can is an IP that can be integrated on an SoC. But it is also
built into external chips, like the TI tcan chips that connect via SPI.
There is also a PCI connected one. Also some SoC integrated m_can
devices are coupled to the mcu and not the main processor and can't be
fully accessed from the main processor, e.g. they are missing
interrupts.

Best
Markus

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 289 bytes --]

[PATCH v3 2/2] can: m_can: set out-of-band wakeup if wakeup pinctrl exists

[Kendall Willis]

In TI AM62X, AM62A, and AM62P SoCs, the m_can pins can act as a wakeup
source in the deepest low power states. However, the m_can pins are a part
of the MCU domain which is OFF in deeper low power states. Since the m_can
pins continue to be ON even if the MCU domain is turned off, set the
out-of-band wakeup flag if the CAN device may wakeup, has the
"out-band-wakeup" property in the device tree node, and has the "wakeup"
pinctrl state.

Signed-off-by: Kendall Willis <k-willis@ti.com>
---
 drivers/net/can/m_can/m_can.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
index eb856547ae7df27a844b236a0c1d4498cbb8b60f..e745660fb3eb311de70635fcfc016a44e62d5602 100644
--- a/drivers/net/can/m_can/m_can.c
+++ b/drivers/net/can/m_can/m_can.c
@@ -2622,8 +2622,12 @@ int m_can_class_suspend(struct device *dev)
 		cdev->can.state = CAN_STATE_SLEEPING;
 	}
 
-	if (!m_can_class_wakeup_pinctrl_enabled(cdev))
+	if (m_can_class_wakeup_pinctrl_enabled(cdev)) {
+		if (device_property_read_bool(dev, "out-band-wakeup"))
+			device_set_out_band_wakeup(dev);
+	} else {
 		pinctrl_pm_select_sleep_state(dev);
+	}
 
 	return ret;
 }

-- 
2.34.1