[RFC,1/2] Documentation: arm: define DT bindings for system suspend

Commit Message

Sudeep Holla Jan. 21, 2015, 11:35 a.m. UTC

ARM based platforms implement unique ways to enter system suspend
(i.e. Suspend to RAM). The mechanism and the parameters defining the
system state vary on a per-platform basis forcing the OS to handle it
in very platform specific way.

Since ARM 32-bit systems had machine specific code, no attempts to
standardize are being made as it provides easy way to implement suspend
operations in a platform specific manner. However, this approach not
only makes maintainance more difficult as the number of platforms
supported increases but also not feasible for ARM64.

This DT binding aims at standardizing the system suspend for ARM
platforms. ARM64 platforms mandates entry-method property in DT for
this system suspend node.

On system implementing PSCI as an enable-method to enter system suspend,
the PSCI CPU suspend method is used on versions upto v0.2 and requires
the power_state parameter to be passed to the PSCI CPU suspend function.

This parameter is platform specific, therefore must be provided by
firmware to the OS in order to enable proper call sequence.

This ARM system suspend DT bindings rely on a property
(i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
how the PSCI CPU suspend power_state parameter should be defined in DT.

Signed-off-by: Sudeep Holla <[e-mail]>
---
 Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
 .../devicetree/bindings/arm/system-suspend.txt     | 93 ++++++++++++++++++++++
 2 files changed, 104 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/arm/system-suspend.txt

Comments

Jisheng Zhang Jan. 21, 2015, 1:21 p.m. UTC | #1

Dear Sudeep,

On Wed, 21 Jan 2015 03:35:54 -0800
Sudeep Holla <[e-mail]> wrote:

> ARM based platforms implement unique ways to enter system suspend
> (i.e. Suspend to RAM). The mechanism and the parameters defining the
> system state vary on a per-platform basis forcing the OS to handle it
> in very platform specific way.
> 
> Since ARM 32-bit systems had machine specific code, no attempts to
> standardize are being made as it provides easy way to implement suspend
> operations in a platform specific manner. However, this approach not
> only makes maintainance more difficult as the number of platforms
> supported increases but also not feasible for ARM64.
> 
> This DT binding aims at standardizing the system suspend for ARM
> platforms. ARM64 platforms mandates entry-method property in DT for
> this system suspend node.
> 
> On system implementing PSCI as an enable-method to enter system suspend,
> the PSCI CPU suspend method is used on versions upto v0.2 and requires
> the power_state parameter to be passed to the PSCI CPU suspend function.
> 
> This parameter is platform specific, therefore must be provided by
> firmware to the OS in order to enable proper call sequence.
> 
> This ARM system suspend DT bindings rely on a property
> (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> how the PSCI CPU suspend power_state parameter should be defined in DT.
> 
> Signed-off-by: Sudeep Holla <[e-mail]>
> ---
>  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
>  .../devicetree/bindings/arm/system-suspend.txt     | 93
> ++++++++++++++++++++++ 2 files changed, 104 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/arm/system-suspend.txt
> 
> diff --git a/Documentation/devicetree/bindings/arm/psci.txt
> b/Documentation/devicetree/bindings/arm/psci.txt index
> 5aa40ede0e99..bd3977a2a333 100644 ---
> a/Documentation/devicetree/bindings/arm/psci.txt +++
> b/Documentation/devicetree/bindings/arm/psci.txt @@ -61,6 +61,14 @@ Device
> tree nodes that require usage of PSCI CPU_SUSPEND function (ie idle
> Definition: power_state parameter to pass to the PSCI suspend call.
>  
> +PSCI v0.2 and earlier versions don't have explicit operation for system
> +suspend. However, one can implement system suspend using CPU_SUSPEND by
> +ensuring every other core except the one executing the CPU_SUSPEND call
> +has called into PSCI through a CPU_OFF call.

If users explicitly hot-unplug other cores when system load is low to save
power, then we want to suspend at some point, how does the firmware know this
case?

In my private tree, I extend the PSCI spec to tell firmware we want to suspend
the system.

Thanks,
Jisheng

Jisheng Zhang Jan. 21, 2015, 1:35 p.m. UTC | #2

Dear Sudeep,

On Wed, 21 Jan 2015 05:21:39 -0800
Jisheng Zhang <[e-mail]> wrote:

> Dear Sudeep,
> 
> On Wed, 21 Jan 2015 03:35:54 -0800
> Sudeep Holla <[e-mail]> wrote:
> 
> > ARM based platforms implement unique ways to enter system suspend
> > (i.e. Suspend to RAM). The mechanism and the parameters defining the
> > system state vary on a per-platform basis forcing the OS to handle it
> > in very platform specific way.
> > 
> > Since ARM 32-bit systems had machine specific code, no attempts to
> > standardize are being made as it provides easy way to implement suspend
> > operations in a platform specific manner. However, this approach not
> > only makes maintainance more difficult as the number of platforms
> > supported increases but also not feasible for ARM64.
> > 
> > This DT binding aims at standardizing the system suspend for ARM
> > platforms. ARM64 platforms mandates entry-method property in DT for
> > this system suspend node.
> > 
> > On system implementing PSCI as an enable-method to enter system suspend,
> > the PSCI CPU suspend method is used on versions upto v0.2 and requires
> > the power_state parameter to be passed to the PSCI CPU suspend function.
> > 
> > This parameter is platform specific, therefore must be provided by
> > firmware to the OS in order to enable proper call sequence.
> > 
> > This ARM system suspend DT bindings rely on a property
> > (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> > how the PSCI CPU suspend power_state parameter should be defined in DT.
> > 
> > Signed-off-by: Sudeep Holla <[e-mail]>
> > ---
> >  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
> >  .../devicetree/bindings/arm/system-suspend.txt     | 93
> > ++++++++++++++++++++++ 2 files changed, 104 insertions(+)
> >  create mode 100644
> > Documentation/devicetree/bindings/arm/system-suspend.txt
> > 
> > diff --git a/Documentation/devicetree/bindings/arm/psci.txt
> > b/Documentation/devicetree/bindings/arm/psci.txt index
> > 5aa40ede0e99..bd3977a2a333 100644 ---
> > a/Documentation/devicetree/bindings/arm/psci.txt +++
> > b/Documentation/devicetree/bindings/arm/psci.txt @@ -61,6 +61,14 @@ Device
> > tree nodes that require usage of PSCI CPU_SUSPEND function (ie idle
> > Definition: power_state parameter to pass to the PSCI suspend call.
> >  
> > +PSCI v0.2 and earlier versions don't have explicit operation for system
> > +suspend. However, one can implement system suspend using CPU_SUSPEND by
> > +ensuring every other core except the one executing the CPU_SUSPEND call
> > +has called into PSCI through a CPU_OFF call.
> 
> If users explicitly hot-unplug other cores when system load is low to save
> power, then we want to suspend at some point, how does the firmware know
> this case?

Sorry for confusion. I mean 

If users explicitly hot-unplug other cores when system load is low to save
power, then at some point cpuidle want to suspend the cluster, how does the
distinguish this case with suspend the system to ram.

> 
> In my private tree, I extend the PSCI spec to tell firmware we want to
> suspend the system.

I extend the PSCI suspend function to help firmware distinguish suspend system
to ram and other suspend cores/clusters.

> 
> Thanks,
> Jisheng
> 
> _______________________________________________
> linux-arm-kernel mailing list
> [e-mail]
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Lorenzo Pieralisi Jan. 21, 2015, 1:56 p.m. UTC | #3

On Wed, Jan 21, 2015 at 01:35:07PM +0000, Jisheng Zhang wrote:
> Dear Sudeep,
> 
> On Wed, 21 Jan 2015 05:21:39 -0800
> Jisheng Zhang <[e-mail]> wrote:
> 
> > Dear Sudeep,
> > 
> > On Wed, 21 Jan 2015 03:35:54 -0800
> > Sudeep Holla <[e-mail]> wrote:
> > 
> > > ARM based platforms implement unique ways to enter system suspend
> > > (i.e. Suspend to RAM). The mechanism and the parameters defining the
> > > system state vary on a per-platform basis forcing the OS to handle it
> > > in very platform specific way.
> > > 
> > > Since ARM 32-bit systems had machine specific code, no attempts to
> > > standardize are being made as it provides easy way to implement suspend
> > > operations in a platform specific manner. However, this approach not
> > > only makes maintainance more difficult as the number of platforms
> > > supported increases but also not feasible for ARM64.
> > > 
> > > This DT binding aims at standardizing the system suspend for ARM
> > > platforms. ARM64 platforms mandates entry-method property in DT for
> > > this system suspend node.
> > > 
> > > On system implementing PSCI as an enable-method to enter system suspend,
> > > the PSCI CPU suspend method is used on versions upto v0.2 and requires
> > > the power_state parameter to be passed to the PSCI CPU suspend function.
> > > 
> > > This parameter is platform specific, therefore must be provided by
> > > firmware to the OS in order to enable proper call sequence.
> > > 
> > > This ARM system suspend DT bindings rely on a property
> > > (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> > > how the PSCI CPU suspend power_state parameter should be defined in DT.
> > > 
> > > Signed-off-by: Sudeep Holla <[e-mail]>
> > > ---
> > >  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
> > >  .../devicetree/bindings/arm/system-suspend.txt     | 93
> > > ++++++++++++++++++++++ 2 files changed, 104 insertions(+)
> > >  create mode 100644
> > > Documentation/devicetree/bindings/arm/system-suspend.txt
> > > 
> > > diff --git a/Documentation/devicetree/bindings/arm/psci.txt
> > > b/Documentation/devicetree/bindings/arm/psci.txt index
> > > 5aa40ede0e99..bd3977a2a333 100644 ---
> > > a/Documentation/devicetree/bindings/arm/psci.txt +++
> > > b/Documentation/devicetree/bindings/arm/psci.txt @@ -61,6 +61,14 @@ Device
> > > tree nodes that require usage of PSCI CPU_SUSPEND function (ie idle
> > > Definition: power_state parameter to pass to the PSCI suspend call.
> > >  
> > > +PSCI v0.2 and earlier versions don't have explicit operation for system
> > > +suspend. However, one can implement system suspend using CPU_SUSPEND by
> > > +ensuring every other core except the one executing the CPU_SUSPEND call
> > > +has called into PSCI through a CPU_OFF call.
> > 
> > If users explicitly hot-unplug other cores when system load is low to save
> > power, then we want to suspend at some point, how does the firmware know
> > this case?
> 
> Sorry for confusion. I mean 
> 
> If users explicitly hot-unplug other cores when system load is low to save
> power, then at some point cpuidle want to suspend the cluster, how does the
> distinguish this case with suspend the system to ram.

Through the arm,psci-suspend-param DT property, ie PSCI CPU_SUSPEND
power_state parameter.

Did you read the patch :) ?

> > In my private tree, I extend the PSCI spec to tell firmware we want to
> > suspend the system.
> 
> I extend the PSCI suspend function to help firmware distinguish suspend system
> to ram and other suspend cores/clusters.

See above.

Thanks,
Lorenzo

Jisheng Zhang Jan. 22, 2015, 4:33 a.m. UTC | #4

Dear Lorenzo,

On Wed, 21 Jan 2015 05:56:11 -0800
Lorenzo Pieralisi <[e-mail]> wrote:

> On Wed, Jan 21, 2015 at 01:35:07PM +0000, Jisheng Zhang wrote:
> > Dear Sudeep,
> > 
> > On Wed, 21 Jan 2015 05:21:39 -0800
> > Jisheng Zhang <[e-mail]> wrote:
> > 
> > > Dear Sudeep,
> > > 
> > > On Wed, 21 Jan 2015 03:35:54 -0800
> > > Sudeep Holla <[e-mail]> wrote:
> > > 
> > > > ARM based platforms implement unique ways to enter system suspend
> > > > (i.e. Suspend to RAM). The mechanism and the parameters defining the
> > > > system state vary on a per-platform basis forcing the OS to handle it
> > > > in very platform specific way.
> > > > 
> > > > Since ARM 32-bit systems had machine specific code, no attempts to
> > > > standardize are being made as it provides easy way to implement
> > > > suspend operations in a platform specific manner. However, this
> > > > approach not only makes maintainance more difficult as the number of
> > > > platforms supported increases but also not feasible for ARM64.
> > > > 
> > > > This DT binding aims at standardizing the system suspend for ARM
> > > > platforms. ARM64 platforms mandates entry-method property in DT for
> > > > this system suspend node.
> > > > 
> > > > On system implementing PSCI as an enable-method to enter system
> > > > suspend, the PSCI CPU suspend method is used on versions upto v0.2
> > > > and requires the power_state parameter to be passed to the PSCI CPU
> > > > suspend function.
> > > > 
> > > > This parameter is platform specific, therefore must be provided by
> > > > firmware to the OS in order to enable proper call sequence.
> > > > 
> > > > This ARM system suspend DT bindings rely on a property
> > > > (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> > > > how the PSCI CPU suspend power_state parameter should be defined in
> > > > DT.
> > > > 
> > > > Signed-off-by: Sudeep Holla <[e-mail]>
> > > > ---
> > > >  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
> > > >  .../devicetree/bindings/arm/system-suspend.txt     | 93
> > > > ++++++++++++++++++++++ 2 files changed, 104 insertions(+)
> > > >  create mode 100644
> > > > Documentation/devicetree/bindings/arm/system-suspend.txt
> > > > 
> > > > diff --git a/Documentation/devicetree/bindings/arm/psci.txt
> > > > b/Documentation/devicetree/bindings/arm/psci.txt index
> > > > 5aa40ede0e99..bd3977a2a333 100644 ---
> > > > a/Documentation/devicetree/bindings/arm/psci.txt +++
> > > > b/Documentation/devicetree/bindings/arm/psci.txt @@ -61,6 +61,14 @@
> > > > Device tree nodes that require usage of PSCI CPU_SUSPEND function (ie
> > > > idle Definition: power_state parameter to pass to the PSCI suspend
> > > > call. 
> > > > +PSCI v0.2 and earlier versions don't have explicit operation for
> > > > system +suspend. However, one can implement system suspend using
> > > > CPU_SUSPEND by +ensuring every other core except the one executing
> > > > the CPU_SUSPEND call +has called into PSCI through a CPU_OFF call.
> > > 
> > > If users explicitly hot-unplug other cores when system load is low to
> > > save power, then we want to suspend at some point, how does the
> > > firmware know this case?
> > 
> > Sorry for confusion. I mean 
> > 
> > If users explicitly hot-unplug other cores when system load is low to save
> > power, then at some point cpuidle want to suspend the cluster, how does
> > the distinguish this case with suspend the system to ram.
> 
> Through the arm,psci-suspend-param DT property, ie PSCI CPU_SUSPEND
> power_state parameter.
> 
> Did you read the patch :) ?

Yep, I do read the patch ;) To be honest, I implemented the s2ram similar as
the patch does. But according to PSCI v0.2, "arm,psci-suspend-param = <0x1010000>"
means suspend the cluster. I'm not sure I understand it correctly, "can implement
system suspend using CPU_SUSPEND by ensuring every other core except the one
executing the CPU_SUSPEND call has called into PSCI through a CPU_OFF call" intend
to ask firmware to 

suspend the system if other cores has called into PSCI through a CPU_OFF

or

suspend the cpu cluster if other cores are not CPU_OFF.


I extend the PSCI CPU_SUSPEND function's to use power_state bit[26] to tell firmware
whether suspend to ram or not.

Could you please correct me if I misunderstand something?

Thank you very much,
Jisheng

> 
> > > In my private tree, I extend the PSCI spec to tell firmware we want to
> > > suspend the system.
> > 
> > I extend the PSCI suspend function to help firmware distinguish suspend
> > system to ram and other suspend cores/clusters.
> 
> See above.
> 
> Thanks,
> Lorenzo
>

Jisheng Zhang Jan. 22, 2015, 6:29 a.m. UTC | #5

Dear Lorenzo and Sudeep,

On Wed, 21 Jan 2015 20:33:14 -0800
Jisheng Zhang <[e-mail]> wrote:

> Dear Lorenzo,
> 
> On Wed, 21 Jan 2015 05:56:11 -0800
> Lorenzo Pieralisi <[e-mail]> wrote:
> 
> > On Wed, Jan 21, 2015 at 01:35:07PM +0000, Jisheng Zhang wrote:
> > > Dear Sudeep,
> > > 
> > > On Wed, 21 Jan 2015 05:21:39 -0800
> > > Jisheng Zhang <[e-mail]> wrote:
> > > 
> > > > Dear Sudeep,
> > > > 
> > > > On Wed, 21 Jan 2015 03:35:54 -0800
> > > > Sudeep Holla <[e-mail]> wrote:
> > > > 
> > > > > ARM based platforms implement unique ways to enter system suspend
> > > > > (i.e. Suspend to RAM). The mechanism and the parameters defining the
> > > > > system state vary on a per-platform basis forcing the OS to handle
> > > > > it in very platform specific way.
> > > > > 
> > > > > Since ARM 32-bit systems had machine specific code, no attempts to
> > > > > standardize are being made as it provides easy way to implement
> > > > > suspend operations in a platform specific manner. However, this
> > > > > approach not only makes maintainance more difficult as the number of
> > > > > platforms supported increases but also not feasible for ARM64.
> > > > > 
> > > > > This DT binding aims at standardizing the system suspend for ARM
> > > > > platforms. ARM64 platforms mandates entry-method property in DT for
> > > > > this system suspend node.
> > > > > 
> > > > > On system implementing PSCI as an enable-method to enter system
> > > > > suspend, the PSCI CPU suspend method is used on versions upto v0.2
> > > > > and requires the power_state parameter to be passed to the PSCI CPU
> > > > > suspend function.
> > > > > 
> > > > > This parameter is platform specific, therefore must be provided by
> > > > > firmware to the OS in order to enable proper call sequence.
> > > > > 
> > > > > This ARM system suspend DT bindings rely on a property
> > > > > (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> > > > > how the PSCI CPU suspend power_state parameter should be defined in
> > > > > DT.
> > > > > 
> > > > > Signed-off-by: Sudeep Holla <[e-mail]>
> > > > > ---
> > > > >  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
> > > > >  .../devicetree/bindings/arm/system-suspend.txt     | 93
> > > > > ++++++++++++++++++++++ 2 files changed, 104 insertions(+)
> > > > >  create mode 100644
> > > > > Documentation/devicetree/bindings/arm/system-suspend.txt
> > > > > 
> > > > > diff --git a/Documentation/devicetree/bindings/arm/psci.txt
> > > > > b/Documentation/devicetree/bindings/arm/psci.txt index
> > > > > 5aa40ede0e99..bd3977a2a333 100644 ---
> > > > > a/Documentation/devicetree/bindings/arm/psci.txt +++
> > > > > b/Documentation/devicetree/bindings/arm/psci.txt @@ -61,6 +61,14 @@
> > > > > Device tree nodes that require usage of PSCI CPU_SUSPEND function
> > > > > (ie idle Definition: power_state parameter to pass to the PSCI
> > > > > suspend call. 
> > > > > +PSCI v0.2 and earlier versions don't have explicit operation for
> > > > > system +suspend. However, one can implement system suspend using
> > > > > CPU_SUSPEND by +ensuring every other core except the one executing
> > > > > the CPU_SUSPEND call +has called into PSCI through a CPU_OFF call.
> > > > 
> > > > If users explicitly hot-unplug other cores when system load is low to
> > > > save power, then we want to suspend at some point, how does the
> > > > firmware know this case?
> > > 
> > > Sorry for confusion. I mean 
> > > 
> > > If users explicitly hot-unplug other cores when system load is low to
> > > save power, then at some point cpuidle want to suspend the cluster, how
> > > does the distinguish this case with suspend the system to ram.
> > 
> > Through the arm,psci-suspend-param DT property, ie PSCI CPU_SUSPEND
> > power_state parameter.
> > 
> > Did you read the patch :) ?
> 
> Yep, I do read the patch ;) To be honest, I implemented the s2ram similar as
> the patch does. But according to PSCI v0.2, "arm,psci-suspend-param =
> <0x1010000>" means suspend the cluster. I'm not sure I understand it
> correctly, "can implement system suspend using CPU_SUSPEND by ensuring
> every other core except the one executing the CPU_SUSPEND call has called
> into PSCI through a CPU_OFF call" intend to ask firmware to 
> 
> suspend the system if other cores has called into PSCI through a CPU_OFF
> 
> or
> 
> suspend the cpu cluster if other cores are not CPU_OFF.
> 
> 
> I extend the PSCI CPU_SUSPEND function's to use power_state bit[26] to tell
> firmware whether suspend to ram or not.
> 

I read the PSCI spec again, power_state bit[0:15] is "platform specific ID",
Is one of these bits used for suspend system?

Thanks,
Jisheng

Lorenzo Pieralisi Jan. 22, 2015, 11:59 a.m. UTC | #6

On Thu, Jan 22, 2015 at 06:29:49AM +0000, Jisheng Zhang wrote:
> Dear Lorenzo and Sudeep,
> 
> On Wed, 21 Jan 2015 20:33:14 -0800
> Jisheng Zhang <[e-mail]> wrote:
> 
> > Dear Lorenzo,
> > 
> > On Wed, 21 Jan 2015 05:56:11 -0800
> > Lorenzo Pieralisi <[e-mail]> wrote:
> > 
> > > On Wed, Jan 21, 2015 at 01:35:07PM +0000, Jisheng Zhang wrote:
> > > > Dear Sudeep,
> > > > 
> > > > On Wed, 21 Jan 2015 05:21:39 -0800
> > > > Jisheng Zhang <[e-mail]> wrote:
> > > > 
> > > > > Dear Sudeep,
> > > > > 
> > > > > On Wed, 21 Jan 2015 03:35:54 -0800
> > > > > Sudeep Holla <[e-mail]> wrote:
> > > > > 
> > > > > > ARM based platforms implement unique ways to enter system suspend
> > > > > > (i.e. Suspend to RAM). The mechanism and the parameters defining the
> > > > > > system state vary on a per-platform basis forcing the OS to handle
> > > > > > it in very platform specific way.
> > > > > > 
> > > > > > Since ARM 32-bit systems had machine specific code, no attempts to
> > > > > > standardize are being made as it provides easy way to implement
> > > > > > suspend operations in a platform specific manner. However, this
> > > > > > approach not only makes maintainance more difficult as the number of
> > > > > > platforms supported increases but also not feasible for ARM64.
> > > > > > 
> > > > > > This DT binding aims at standardizing the system suspend for ARM
> > > > > > platforms. ARM64 platforms mandates entry-method property in DT for
> > > > > > this system suspend node.
> > > > > > 
> > > > > > On system implementing PSCI as an enable-method to enter system
> > > > > > suspend, the PSCI CPU suspend method is used on versions upto v0.2
> > > > > > and requires the power_state parameter to be passed to the PSCI CPU
> > > > > > suspend function.
> > > > > > 
> > > > > > This parameter is platform specific, therefore must be provided by
> > > > > > firmware to the OS in order to enable proper call sequence.
> > > > > > 
> > > > > > This ARM system suspend DT bindings rely on a property
> > > > > > (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> > > > > > how the PSCI CPU suspend power_state parameter should be defined in
> > > > > > DT.
> > > > > > 
> > > > > > Signed-off-by: Sudeep Holla <[e-mail]>
> > > > > > ---
> > > > > >  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
> > > > > >  .../devicetree/bindings/arm/system-suspend.txt     | 93
> > > > > > ++++++++++++++++++++++ 2 files changed, 104 insertions(+)
> > > > > >  create mode 100644
> > > > > > Documentation/devicetree/bindings/arm/system-suspend.txt
> > > > > > 
> > > > > > diff --git a/Documentation/devicetree/bindings/arm/psci.txt
> > > > > > b/Documentation/devicetree/bindings/arm/psci.txt index
> > > > > > 5aa40ede0e99..bd3977a2a333 100644 ---
> > > > > > a/Documentation/devicetree/bindings/arm/psci.txt +++
> > > > > > b/Documentation/devicetree/bindings/arm/psci.txt @@ -61,6 +61,14 @@
> > > > > > Device tree nodes that require usage of PSCI CPU_SUSPEND function
> > > > > > (ie idle Definition: power_state parameter to pass to the PSCI
> > > > > > suspend call. 
> > > > > > +PSCI v0.2 and earlier versions don't have explicit operation for
> > > > > > system +suspend. However, one can implement system suspend using
> > > > > > CPU_SUSPEND by +ensuring every other core except the one executing
> > > > > > the CPU_SUSPEND call +has called into PSCI through a CPU_OFF call.
> > > > > 
> > > > > If users explicitly hot-unplug other cores when system load is low to
> > > > > save power, then we want to suspend at some point, how does the
> > > > > firmware know this case?
> > > > 
> > > > Sorry for confusion. I mean 
> > > > 
> > > > If users explicitly hot-unplug other cores when system load is low to
> > > > save power, then at some point cpuidle want to suspend the cluster, how
> > > > does the distinguish this case with suspend the system to ram.
> > > 
> > > Through the arm,psci-suspend-param DT property, ie PSCI CPU_SUSPEND
> > > power_state parameter.
> > > 
> > > Did you read the patch :) ?
> > 
> > Yep, I do read the patch ;) To be honest, I implemented the s2ram similar as
> > the patch does. But according to PSCI v0.2, "arm,psci-suspend-param =
> > <0x1010000>" means suspend the cluster. I'm not sure I understand it
> > correctly, "can implement system suspend using CPU_SUSPEND by ensuring
> > every other core except the one executing the CPU_SUSPEND call has called
> > into PSCI through a CPU_OFF call" intend to ask firmware to 
> > 
> > suspend the system if other cores has called into PSCI through a CPU_OFF
> > 
> > or
> > 
> > suspend the cpu cluster if other cores are not CPU_OFF.
> > 
> > 
> > I extend the PSCI CPU_SUSPEND function's to use power_state bit[26] to tell
> > firmware whether suspend to ram or not.
> > 

And that's what the arm,psci-suspend-param stands for in the
system-state node.

Since system-suspend corresponds supposedly to the highest level of
affinity in the system, I would rather say power_state = 0x3010000
can be used for system suspend (affinity bits[25:24] = 0x3), but we did
not want to force it, probably that's what we should do.

Yes, there is also a platform specific component in power_state
param and you can use that too, we wanted to leave flexibility
to platforms.

PSCI v1.0 will introduce a different separate call for system
suspend, this patch copes with "legacy" versions, as the patch
logs describe.

I agree that the value 0x1010000 was a bad choice for the example, it
is confusing, but it does not mean you _have_ to use that value, is it
clear ?

> I read the PSCI spec again, power_state bit[0:15] is "platform specific ID",
> Is one of these bits used for suspend system?

It is platform specific, you define that :) ! That's the reason why
firmware has to tell the OS what parameter triggers the system-state,
it is platform specific, and we provide a binding to define it and provide
the OS with the correct value to use.

Lorenzo

Jisheng Zhang Jan. 22, 2015, 12:09 p.m. UTC | #7

Dear Lorenzo,

On Thu, 22 Jan 2015 03:59:06 -0800
Lorenzo Pieralisi <[e-mail]> wrote:

> On Thu, Jan 22, 2015 at 06:29:49AM +0000, Jisheng Zhang wrote:
> > Dear Lorenzo and Sudeep,
> > 
> > On Wed, 21 Jan 2015 20:33:14 -0800
> > Jisheng Zhang <[e-mail]> wrote:
> > 
> > > Dear Lorenzo,
> > > 
> > > On Wed, 21 Jan 2015 05:56:11 -0800
> > > Lorenzo Pieralisi <[e-mail]> wrote:
> > > 
> > > > On Wed, Jan 21, 2015 at 01:35:07PM +0000, Jisheng Zhang wrote:
> > > > > Dear Sudeep,
> > > > > 
> > > > > On Wed, 21 Jan 2015 05:21:39 -0800
> > > > > Jisheng Zhang <[e-mail]> wrote:
> > > > > 
> > > > > > Dear Sudeep,
> > > > > > 
> > > > > > On Wed, 21 Jan 2015 03:35:54 -0800
> > > > > > Sudeep Holla <[e-mail]> wrote:
> > > > > > 
> > > > > > > ARM based platforms implement unique ways to enter system
> > > > > > > suspend (i.e. Suspend to RAM). The mechanism and the parameters
> > > > > > > defining the system state vary on a per-platform basis forcing
> > > > > > > the OS to handle it in very platform specific way.
> > > > > > > 
> > > > > > > Since ARM 32-bit systems had machine specific code, no attempts
> > > > > > > to standardize are being made as it provides easy way to
> > > > > > > implement suspend operations in a platform specific manner.
> > > > > > > However, this approach not only makes maintainance more
> > > > > > > difficult as the number of platforms supported increases but
> > > > > > > also not feasible for ARM64.
> > > > > > > 
> > > > > > > This DT binding aims at standardizing the system suspend for ARM
> > > > > > > platforms. ARM64 platforms mandates entry-method property in DT
> > > > > > > for this system suspend node.
> > > > > > > 
> > > > > > > On system implementing PSCI as an enable-method to enter system
> > > > > > > suspend, the PSCI CPU suspend method is used on versions upto
> > > > > > > v0.2 and requires the power_state parameter to be passed to the
> > > > > > > PSCI CPU suspend function.
> > > > > > > 
> > > > > > > This parameter is platform specific, therefore must be provided
> > > > > > > by firmware to the OS in order to enable proper call sequence.
> > > > > > > 
> > > > > > > This ARM system suspend DT bindings rely on a property
> > > > > > > (i.e. arm,psci-suspend-param) in the PSCI DT bindings that
> > > > > > > describes how the PSCI CPU suspend power_state parameter should
> > > > > > > be defined in DT.
> > > > > > > 
> > > > > > > Signed-off-by: Sudeep Holla <[e-mail]>
> > > > > > > ---
> > > > > > >  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
> > > > > > >  .../devicetree/bindings/arm/system-suspend.txt     | 93
> > > > > > > ++++++++++++++++++++++ 2 files changed, 104 insertions(+)
> > > > > > >  create mode 100644
> > > > > > > Documentation/devicetree/bindings/arm/system-suspend.txt
> > > > > > > 
> > > > > > > diff --git a/Documentation/devicetree/bindings/arm/psci.txt
> > > > > > > b/Documentation/devicetree/bindings/arm/psci.txt index
> > > > > > > 5aa40ede0e99..bd3977a2a333 100644 ---
> > > > > > > a/Documentation/devicetree/bindings/arm/psci.txt +++
> > > > > > > b/Documentation/devicetree/bindings/arm/psci.txt @@ -61,6
> > > > > > > +61,14 @@ Device tree nodes that require usage of PSCI
> > > > > > > CPU_SUSPEND function (ie idle Definition: power_state parameter
> > > > > > > to pass to the PSCI suspend call. 
> > > > > > > +PSCI v0.2 and earlier versions don't have explicit operation
> > > > > > > for system +suspend. However, one can implement system suspend
> > > > > > > using CPU_SUSPEND by +ensuring every other core except the one
> > > > > > > executing the CPU_SUSPEND call +has called into PSCI through a
> > > > > > > CPU_OFF call.
> > > > > > 
> > > > > > If users explicitly hot-unplug other cores when system load is
> > > > > > low to save power, then we want to suspend at some point, how
> > > > > > does the firmware know this case?
> > > > > 
> > > > > Sorry for confusion. I mean 
> > > > > 
> > > > > If users explicitly hot-unplug other cores when system load is low
> > > > > to save power, then at some point cpuidle want to suspend the
> > > > > cluster, how does the distinguish this case with suspend the system
> > > > > to ram.
> > > > 
> > > > Through the arm,psci-suspend-param DT property, ie PSCI CPU_SUSPEND
> > > > power_state parameter.
> > > > 
> > > > Did you read the patch :) ?
> > > 
> > > Yep, I do read the patch ;) To be honest, I implemented the s2ram
> > > similar as the patch does. But according to PSCI v0.2,
> > > "arm,psci-suspend-param = <0x1010000>" means suspend the cluster. I'm
> > > not sure I understand it correctly, "can implement system suspend using
> > > CPU_SUSPEND by ensuring every other core except the one executing the
> > > CPU_SUSPEND call has called into PSCI through a CPU_OFF call" intend to
> > > ask firmware to 
> > > 
> > > suspend the system if other cores has called into PSCI through a CPU_OFF
> > > 
> > > or
> > > 
> > > suspend the cpu cluster if other cores are not CPU_OFF.
> > > 
> > > 
> > > I extend the PSCI CPU_SUSPEND function's to use power_state bit[26] to
> > > tell firmware whether suspend to ram or not.
> > > 
> 
> And that's what the arm,psci-suspend-param stands for in the
> system-state node.
> 
> Since system-suspend corresponds supposedly to the highest level of
> affinity in the system, I would rather say power_state = 0x3010000
> can be used for system suspend (affinity bits[25:24] = 0x3), but we did
> not want to force it, probably that's what we should do.
> 
> Yes, there is also a platform specific component in power_state
> param and you can use that too, we wanted to leave flexibility
> to platforms.
> 
> PSCI v1.0 will introduce a different separate call for system
> suspend, this patch copes with "legacy" versions, as the patch
> logs describe.
> 
> I agree that the value 0x1010000 was a bad choice for the example, it
> is confusing, but it does not mean you _have_ to use that value, is it
> clear ?
> 
> > I read the PSCI spec again, power_state bit[0:15] is "platform specific
> > ID", Is one of these bits used for suspend system?
> 
> It is platform specific, you define that :) ! That's the reason why
> firmware has to tell the OS what parameter triggers the system-state,
> it is platform specific, and we provide a binding to define it and provide
> the OS with the correct value to use.
> 
> Lorenzo

Thank you for detailed explanations. Now I got your and the patches' points.
I were just confused by the 0x1010000.

I'll reuse this patch for arm64 suspend system.

Thanks,
Jisheng

Mark Rutland Feb. 4, 2015, 4:10 p.m. UTC | #8

Hi Sudeep,

On Wed, Jan 21, 2015 at 11:35:54AM +0000, Sudeep Holla wrote:
> ARM based platforms implement unique ways to enter system suspend
> (i.e. Suspend to RAM). The mechanism and the parameters defining the
> system state vary on a per-platform basis forcing the OS to handle it
> in very platform specific way.
> 
> Since ARM 32-bit systems had machine specific code, no attempts to
> standardize are being made as it provides easy way to implement suspend
> operations in a platform specific manner. However, this approach not
> only makes maintainance more difficult as the number of platforms
> supported increases but also not feasible for ARM64.
> 
> This DT binding aims at standardizing the system suspend for ARM
> platforms. ARM64 platforms mandates entry-method property in DT for
> this system suspend node.
> 
> On system implementing PSCI as an enable-method to enter system suspend,
> the PSCI CPU suspend method is used on versions upto v0.2 and requires
> the power_state parameter to be passed to the PSCI CPU suspend function.
> 
> This parameter is platform specific, therefore must be provided by
> firmware to the OS in order to enable proper call sequence.
> 
> This ARM system suspend DT bindings rely on a property
> (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> how the PSCI CPU suspend power_state parameter should be defined in DT.

A short while ago (after this posting), the PSCI 1.0 spec [1] was
released, featuring the new (optional) SYSTEM_SUSPEND mechanism intended
for suspend to RAM. This has a standard ID, and its presence can be
detected via the new standard PSCI_FEATURES call.

The fundamental mechanism is identical. We would hot unplug all but one
CPU, and from this remaining CPU we would make a SYSTEM_SUSPEND call.
The major differences are that SYSTEM_SUSPEND can be detected via
PSCI_FEATURES, and doesn't need a state parameter.

Given that the only mandatory addition in PSCI 1.0 over PSCI 0.2 is the
PSCI_FEATURES call (used to detect the presence of SYSTEM_SUSPEND), I do
not believe that implementing this should be a signficant overhead
compared to implementing the CPU_SUSPEND based approach with PSCI 0.2.

So I'd very much prefer that we require a minimal PSCI 1.0 with
SYSTEM_SUSPEND rather than extending CPU_SUSPEND in this manner. Is
anyone attempting to implement suspend to RAM with PSCI 0.1?

Thanks,
Mark.

[1] http://infocenter.arm.com/help/topic/com.arm.doc.den0022c/index.html

> 
> Signed-off-by: Sudeep Holla <[e-mail]>
> ---
>  Documentation/devicetree/bindings/arm/psci.txt     | 11 +++
>  .../devicetree/bindings/arm/system-suspend.txt     | 93 ++++++++++++++++++++++
>  2 files changed, 104 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/arm/system-suspend.txt
> 
> diff --git a/Documentation/devicetree/bindings/arm/psci.txt b/Documentation/devicetree/bindings/arm/psci.txt
> index 5aa40ede0e99..bd3977a2a333 100644
> --- a/Documentation/devicetree/bindings/arm/psci.txt
> +++ b/Documentation/devicetree/bindings/arm/psci.txt
> @@ -61,6 +61,14 @@ Device tree nodes that require usage of PSCI CPU_SUSPEND function (ie idle
>  		Definition: power_state parameter to pass to the PSCI
>  			    suspend call.
>  
> +PSCI v0.2 and earlier versions don't have explicit operation for system
> +suspend. However, one can implement system suspend using CPU_SUSPEND by
> +ensuring every other core except the one executing the CPU_SUSPEND call
> +has called into PSCI through a CPU_OFF call.
> +
> +In such cases, device tree nodes representing system suspend as per the
> +bindings in [2] must specify the above "arm,psci-suspend-param" property.
> +
>  Example:
>  
>  Case 1: PSCI v0.1 only.
> @@ -100,3 +108,6 @@ Case 3: PSCI v0.2 and PSCI v0.1.
>  
>  [1] Kernel documentation - ARM idle states bindings
>      Documentation/devicetree/bindings/arm/idle-states.txt
> +
> +[2] Kernel documentation - ARM system suspend bindings
> +    Documentation/devicetree/bindings/arm/system-suspend.txt
> diff --git a/Documentation/devicetree/bindings/arm/system-suspend.txt b/Documentation/devicetree/bindings/arm/system-suspend.txt
> new file mode 100644
> index 000000000000..15cac4c7fe92
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/arm/system-suspend.txt
> @@ -0,0 +1,93 @@
> +==========================================
> +ARM system suspend binding description
> +==========================================
> +
> +==========================================
> +1 - Introduction
> +==========================================
> +
> +System Suspend(commonly known as Suspend to RAM) is a method to remove
> +power from most parts of the machine aside from the RAM, which is required
> +to restore the machine's state. Because of the large power savings, it is
> +widely used in mobile systems like laptops, tablets and smartphones.
> +
> +Usually most mobile systems enter system suspend state aggressively when
> +they are idle even for short time(in seconds) while others systems like
> +laptops automatically enter this mode when they running on batteries and
> +the lid is closed (and/or the user is inactive for some time(in minutes)).
> +
> +Most of the devices in the system are deactivated. Non-volatile memory
> +(like disk drives, flash, memory card), graphics chips and even the CPU
> +are usually deactivated. Only the RAM is powered to keep its contents. On
> +resume, only those individual devices/CPUs need to be reinitialized and
> +work continues relatively fast.
> +
> +It is highly hardware specific especially on ARM platforms. Hence we need
> +device tree binding definition for ARM system suspend state which is the
> +subject of this document in order to provide generic solution.
> +
> +===========================================
> +2 - system suspend node
> +===========================================
> +
> +The system suspend node represents the description of the mechanism to
> +enter system suspend state and must be defined as follows:
> +
> +	- compatible
> +		Usage: Required
> +		Value type: <stringlist>
> +		Definition: Must be "arm,system-suspend";
> +
> +	- entry-method
> +		Value type: <stringlist>
> +		Usage and definition depend on ARM architecture version.
> +			# On ARM v8 64-bit this property is required and must
> +			  be one of:
> +			   - "arm,psci" (see bindings in [1])
> +
> +	- status:
> +		Usage: Optional
> +		Value type: <string>
> +		Definition: A standard device tree property [2] that indicates
> +			    the operational status of system suspend.
> +			    If present, it shall be:
> +			    "okay": to indicate it is operational.
> +			    "disabled": to indicate that it has been disabled
> +			                in firmware so it is not operational.
> +			    By default, it's always enabled if not explicitly
> +			    disabled.
> +
> +	In addition to the properties listed above, it may require additional
> +	properties specifics to the entry-method defined, please refer to the
> +	corresponding entry-method bindings documentation for details.
> +	In the below example using "arm,psci" entry method,
> +	"arm,psci-suspend-param" is a PSCI specific property.
> +
> +	The system suspend node's parent node must be the 'cpus' node.
> +
> +===========================================
> +3 - Examples
> +===========================================
> +
> +Example:
> +cpus {
> +	/* cpu-map, cpu and idle-states nodes */
> +	....
> +
> +	system-suspend {
> +		compatible = "arm,system-suspend";
> +		entry-method = "arm,psci";
> +		arm,psci-suspend-param = <0x1010000>;
> +	};
> +
> +	....
> +};
> +===========================================
> +4 - References
> +===========================================
> +
> +[1] ARM Linux Kernel documentation - PSCI bindings
> +    Documentation/devicetree/bindings/arm/psci.txt
> +
> +[2] ePAPR standard
> +    https://www.power.org/documentation/epapr-version-1-1/
> -- 
> 1.9.1
> 
>

Sudeep Holla Feb. 5, 2015, 1:28 p.m. UTC | #9

On 04/02/15 16:10, Mark Rutland wrote:
> Hi Sudeep,
>
> On Wed, Jan 21, 2015 at 11:35:54AM +0000, Sudeep Holla wrote:
>> ARM based platforms implement unique ways to enter system suspend
>> (i.e. Suspend to RAM). The mechanism and the parameters defining the
>> system state vary on a per-platform basis forcing the OS to handle it
>> in very platform specific way.
>>
>> Since ARM 32-bit systems had machine specific code, no attempts to
>> standardize are being made as it provides easy way to implement suspend
>> operations in a platform specific manner. However, this approach not
>> only makes maintainance more difficult as the number of platforms
>> supported increases but also not feasible for ARM64.
>>
>> This DT binding aims at standardizing the system suspend for ARM
>> platforms. ARM64 platforms mandates entry-method property in DT for
>> this system suspend node.
>>
>> On system implementing PSCI as an enable-method to enter system suspend,
>> the PSCI CPU suspend method is used on versions upto v0.2 and requires
>> the power_state parameter to be passed to the PSCI CPU suspend function.
>>
>> This parameter is platform specific, therefore must be provided by
>> firmware to the OS in order to enable proper call sequence.
>>
>> This ARM system suspend DT bindings rely on a property
>> (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
>> how the PSCI CPU suspend power_state parameter should be defined in DT.
>
> A short while ago (after this posting), the PSCI 1.0 spec [1] was
> released, featuring the new (optional) SYSTEM_SUSPEND mechanism intended
> for suspend to RAM. This has a standard ID, and its presence can be
> detected via the new standard PSCI_FEATURES call.
>
> The fundamental mechanism is identical. We would hot unplug all but one
> CPU, and from this remaining CPU we would make a SYSTEM_SUSPEND call.
> The major differences are that SYSTEM_SUSPEND can be detected via
> PSCI_FEATURES, and doesn't need a state parameter.
>
> Given that the only mandatory addition in PSCI 1.0 over PSCI 0.2 is the
> PSCI_FEATURES call (used to detect the presence of SYSTEM_SUSPEND), I do
> not believe that implementing this should be a signficant overhead
> compared to implementing the CPU_SUSPEND based approach with PSCI 0.2.
>
> So I'd very much prefer that we require a minimal PSCI 1.0 with
> SYSTEM_SUSPEND rather than extending CPU_SUSPEND in this manner. Is
> anyone attempting to implement suspend to RAM with PSCI 0.1?
>

I too prefer have PSCI v1.0 for SYSTEM SUSPEND support, which eliminates
the need for this binding. But the question is: are silicon vendors
ready to upgrade their firmware to PSCI v1.0 for system system feature
especially since it's one of the fundamental feature needed in Android
systems.

Regards,
Sudeep

Mark Rutland Feb. 5, 2015, 1:32 p.m. UTC | #10

On Thu, Feb 05, 2015 at 01:28:32PM +0000, Sudeep Holla wrote:
> 
> 
> On 04/02/15 16:10, Mark Rutland wrote:
> > Hi Sudeep,
> >
> > On Wed, Jan 21, 2015 at 11:35:54AM +0000, Sudeep Holla wrote:
> >> ARM based platforms implement unique ways to enter system suspend
> >> (i.e. Suspend to RAM). The mechanism and the parameters defining the
> >> system state vary on a per-platform basis forcing the OS to handle it
> >> in very platform specific way.
> >>
> >> Since ARM 32-bit systems had machine specific code, no attempts to
> >> standardize are being made as it provides easy way to implement suspend
> >> operations in a platform specific manner. However, this approach not
> >> only makes maintainance more difficult as the number of platforms
> >> supported increases but also not feasible for ARM64.
> >>
> >> This DT binding aims at standardizing the system suspend for ARM
> >> platforms. ARM64 platforms mandates entry-method property in DT for
> >> this system suspend node.
> >>
> >> On system implementing PSCI as an enable-method to enter system suspend,
> >> the PSCI CPU suspend method is used on versions upto v0.2 and requires
> >> the power_state parameter to be passed to the PSCI CPU suspend function.
> >>
> >> This parameter is platform specific, therefore must be provided by
> >> firmware to the OS in order to enable proper call sequence.
> >>
> >> This ARM system suspend DT bindings rely on a property
> >> (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
> >> how the PSCI CPU suspend power_state parameter should be defined in DT.
> >
> > A short while ago (after this posting), the PSCI 1.0 spec [1] was
> > released, featuring the new (optional) SYSTEM_SUSPEND mechanism intended
> > for suspend to RAM. This has a standard ID, and its presence can be
> > detected via the new standard PSCI_FEATURES call.
> >
> > The fundamental mechanism is identical. We would hot unplug all but one
> > CPU, and from this remaining CPU we would make a SYSTEM_SUSPEND call.
> > The major differences are that SYSTEM_SUSPEND can be detected via
> > PSCI_FEATURES, and doesn't need a state parameter.
> >
> > Given that the only mandatory addition in PSCI 1.0 over PSCI 0.2 is the
> > PSCI_FEATURES call (used to detect the presence of SYSTEM_SUSPEND), I do
> > not believe that implementing this should be a signficant overhead
> > compared to implementing the CPU_SUSPEND based approach with PSCI 0.2.
> >
> > So I'd very much prefer that we require a minimal PSCI 1.0 with
> > SYSTEM_SUSPEND rather than extending CPU_SUSPEND in this manner. Is
> > anyone attempting to implement suspend to RAM with PSCI 0.1?
> >
> 
> I too prefer have PSCI v1.0 for SYSTEM SUSPEND support, which eliminates
> the need for this binding. But the question is: are silicon vendors
> ready to upgrade their firmware to PSCI v1.0 for system system feature
> especially since it's one of the fundamental feature needed in Android
> systems.

Sure. That's the question I'd like to know the answer to.

If they're bringing up PSCIv0.2 now, the delta to PSCIv0.1 is not large.

If they already have an implementation baked, then that's a different
scenario.

Regardless, what constitutes a wakeup device is a fundamental question
to answer, along with what in the system (e.g.  peripherals) the kernel
must save/restore the state of.

Thanks,
Mark.

Sudeep Holla Feb. 5, 2015, 1:49 p.m. UTC | #11

On 05/02/15 13:32, Mark Rutland wrote:
> On Thu, Feb 05, 2015 at 01:28:32PM +0000, Sudeep Holla wrote:
>>
>>
>> On 04/02/15 16:10, Mark Rutland wrote:
>>> Hi Sudeep,
>>>
>>> On Wed, Jan 21, 2015 at 11:35:54AM +0000, Sudeep Holla wrote:
>>>> ARM based platforms implement unique ways to enter system suspend
>>>> (i.e. Suspend to RAM). The mechanism and the parameters defining the
>>>> system state vary on a per-platform basis forcing the OS to handle it
>>>> in very platform specific way.
>>>>
>>>> Since ARM 32-bit systems had machine specific code, no attempts to
>>>> standardize are being made as it provides easy way to implement suspend
>>>> operations in a platform specific manner. However, this approach not
>>>> only makes maintainance more difficult as the number of platforms
>>>> supported increases but also not feasible for ARM64.
>>>>
>>>> This DT binding aims at standardizing the system suspend for ARM
>>>> platforms. ARM64 platforms mandates entry-method property in DT for
>>>> this system suspend node.
>>>>
>>>> On system implementing PSCI as an enable-method to enter system suspend,
>>>> the PSCI CPU suspend method is used on versions upto v0.2 and requires
>>>> the power_state parameter to be passed to the PSCI CPU suspend function.
>>>>
>>>> This parameter is platform specific, therefore must be provided by
>>>> firmware to the OS in order to enable proper call sequence.
>>>>
>>>> This ARM system suspend DT bindings rely on a property
>>>> (i.e. arm,psci-suspend-param) in the PSCI DT bindings that describes
>>>> how the PSCI CPU suspend power_state parameter should be defined in DT.
>>>
>>> A short while ago (after this posting), the PSCI 1.0 spec [1] was
>>> released, featuring the new (optional) SYSTEM_SUSPEND mechanism intended
>>> for suspend to RAM. This has a standard ID, and its presence can be
>>> detected via the new standard PSCI_FEATURES call.
>>>
>>> The fundamental mechanism is identical. We would hot unplug all but one
>>> CPU, and from this remaining CPU we would make a SYSTEM_SUSPEND call.
>>> The major differences are that SYSTEM_SUSPEND can be detected via
>>> PSCI_FEATURES, and doesn't need a state parameter.
>>>
>>> Given that the only mandatory addition in PSCI 1.0 over PSCI 0.2 is the
>>> PSCI_FEATURES call (used to detect the presence of SYSTEM_SUSPEND), I do
>>> not believe that implementing this should be a signficant overhead
>>> compared to implementing the CPU_SUSPEND based approach with PSCI 0.2.
>>>
>>> So I'd very much prefer that we require a minimal PSCI 1.0 with
>>> SYSTEM_SUSPEND rather than extending CPU_SUSPEND in this manner. Is
>>> anyone attempting to implement suspend to RAM with PSCI 0.1?
>>>
>>
>> I too prefer have PSCI v1.0 for SYSTEM SUSPEND support, which eliminates
>> the need for this binding. But the question is: are silicon vendors
>> ready to upgrade their firmware to PSCI v1.0 for system system feature
>> especially since it's one of the fundamental feature needed in Android
>> systems.
>
> Sure. That's the question I'd like to know the answer to.
>

Right, I too hope to get response here.

> If they're bringing up PSCIv0.2 now, the delta to PSCIv0.1 is not large.
>

That's true.

> If they already have an implementation baked, then that's a different
> scenario.
>
> Regardless, what constitutes a wakeup device is a fundamental question

IIUC individual devices can expose if they are wake-up capable. E.g.
RTC can be wakeup capable and if it's enabled(sysfs entry exists, I did
use it to test on Juno), it's interrupt is kept unmasked in suspend path
while other devices keep their interrupts masked(can be both at GIC and
source).

> to answer, along with what in the system (e.g.  peripherals) the kernel
> must save/restore the state of.
>

Ideally individual drivers need to take care of saving and restoring 
their state. However there will be always exceptions :)

Regards,
Sudeep

Patch

diff --git a/Documentation/devicetree/bindings/arm/psci.txt b/Documentation/devicetree/bindings/arm/psci.txt
index 5aa40ede0e99..bd3977a2a333 100644
--- a/Documentation/devicetree/bindings/arm/psci.txt
+++ b/Documentation/devicetree/bindings/arm/psci.txt
@@ -61,6 +61,14 @@  Device tree nodes that require usage of PSCI CPU_SUSPEND function (ie idle
 		Definition: power_state parameter to pass to the PSCI
 			    suspend call.
 
+PSCI v0.2 and earlier versions don't have explicit operation for system
+suspend. However, one can implement system suspend using CPU_SUSPEND by
+ensuring every other core except the one executing the CPU_SUSPEND call
+has called into PSCI through a CPU_OFF call.
+
+In such cases, device tree nodes representing system suspend as per the
+bindings in [2] must specify the above "arm,psci-suspend-param" property.
+
 Example:
 
 Case 1: PSCI v0.1 only.
@@ -100,3 +108,6 @@  Case 3: PSCI v0.2 and PSCI v0.1.
 
 [1] Kernel documentation - ARM idle states bindings
     Documentation/devicetree/bindings/arm/idle-states.txt
+
+[2] Kernel documentation - ARM system suspend bindings
+    Documentation/devicetree/bindings/arm/system-suspend.txt
diff --git a/Documentation/devicetree/bindings/arm/system-suspend.txt b/Documentation/devicetree/bindings/arm/system-suspend.txt
new file mode 100644
index 000000000000..15cac4c7fe92
--- /dev/null
+++ b/Documentation/devicetree/bindings/arm/system-suspend.txt
@@ -0,0 +1,93 @@ 
+==========================================
+ARM system suspend binding description
+==========================================
+
+==========================================
+1 - Introduction
+==========================================
+
+System Suspend(commonly known as Suspend to RAM) is a method to remove
+power from most parts of the machine aside from the RAM, which is required
+to restore the machine's state. Because of the large power savings, it is
+widely used in mobile systems like laptops, tablets and smartphones.
+
+Usually most mobile systems enter system suspend state aggressively when
+they are idle even for short time(in seconds) while others systems like
+laptops automatically enter this mode when they running on batteries and
+the lid is closed (and/or the user is inactive for some time(in minutes)).
+
+Most of the devices in the system are deactivated. Non-volatile memory
+(like disk drives, flash, memory card), graphics chips and even the CPU
+are usually deactivated. Only the RAM is powered to keep its contents. On
+resume, only those individual devices/CPUs need to be reinitialized and
+work continues relatively fast.
+
+It is highly hardware specific especially on ARM platforms. Hence we need
+device tree binding definition for ARM system suspend state which is the
+subject of this document in order to provide generic solution.
+
+===========================================
+2 - system suspend node
+===========================================
+
+The system suspend node represents the description of the mechanism to
+enter system suspend state and must be defined as follows:
+
+	- compatible
+		Usage: Required
+		Value type: <stringlist>
+		Definition: Must be "arm,system-suspend";
+
+	- entry-method
+		Value type: <stringlist>
+		Usage and definition depend on ARM architecture version.
+			# On ARM v8 64-bit this property is required and must
+			  be one of:
+			   - "arm,psci" (see bindings in [1])
+
+	- status:
+		Usage: Optional
+		Value type: <string>
+		Definition: A standard device tree property [2] that indicates
+			    the operational status of system suspend.
+			    If present, it shall be:
+			    "okay": to indicate it is operational.
+			    "disabled": to indicate that it has been disabled
+			                in firmware so it is not operational.
+			    By default, it's always enabled if not explicitly
+			    disabled.
+
+	In addition to the properties listed above, it may require additional
+	properties specifics to the entry-method defined, please refer to the
+	corresponding entry-method bindings documentation for details.
+	In the below example using "arm,psci" entry method,
+	"arm,psci-suspend-param" is a PSCI specific property.
+
+	The system suspend node's parent node must be the 'cpus' node.
+
+===========================================
+3 - Examples
+===========================================
+
+Example:
+cpus {
+	/* cpu-map, cpu and idle-states nodes */
+	....
+
+	system-suspend {
+		compatible = "arm,system-suspend";
+		entry-method = "arm,psci";
+		arm,psci-suspend-param = <0x1010000>;
+	};
+
+	....
+};
+===========================================
+4 - References
+===========================================
+
+[1] ARM Linux Kernel documentation - PSCI bindings
+    Documentation/devicetree/bindings/arm/psci.txt
+
+[2] ePAPR standard
+    https://www.power.org/documentation/epapr-version-1-1/