[PATCH v4 00/20] firmware: ARM System Control and Management Interface(SCMI) support

[PATCH v4 00/20] firmware: ARM System Control and Management Interface(SCMI) support

[Sudeep Holla]

Hi all,

ARM System Control and Management Interface(SCMI) is more flexible and
easily extensible than any of the existing interfaces. Many vendors were
involved in the making of this formal specification and is now published[1].

There is a strong trend in the industry to provide micro-controllers in
systems to abstract various power, or other system management tasks.
These controllers usually have similar interfaces, both in terms of the
functions that are provided by them, and in terms of how requests are
communicated to them.

This specification is to standardise and avoid (any further)
fragmentation in the design of such interface by various vendors.

This patch set is intended to get feedback on the design and structure
of the code. This is not complete and not fully tested due to
non-availability of firmware with full feature set at this time.

It currently doesn't support notification, asynchronous/delayed response,
perf/power statistics region and sensor register region to name a few.
I have borrowed some of the ideas of message allocation/management from
TI SCI.

Changes:

v3[5]->v4:
	- Added SCMI protocol bus to enumerate supported protocols as
	  suggested by Arnd
	- Dropped the abstraction to mailbox as it may be optional

v2[4]->v3:
	- Addressed various comments recieved so far(clock, hwmon and
	  cpufreq drivers along with scmi drivers)
	- Hwmon driver now uses core layer to create and manage sysfs
	  attributes
	- Added a shim layer to abstract the mailbox interface to support
	  any custom adaptation required by the controller driver
	- Simple ARM MHU shim layer using newly added abstraction

v1[3]->v2[4]:
	- Additional support for polling based DVFS and per protocol
	  channels
	- Dependent drivers(clock, hwmon, cpufreq and power domains)
	- Various other review comments and issued found during testing
	  addressed
	- Explicit binding for method dropped as even SMC based method
	  are adviertised as mailbox

RFC[2]->v1[3]:
	- Add generic mailbox binding for shared memory(Rob H)
	- Dropped compatibles per protocol(Suggested by Matt S)
	- Dropped lot of unnecessary pointer casting(Arnd B)
	- Dropped packing of structures(Arnd B)
	- Few other changes/additions based initial testing with firmware
	  providing SCMI interface to OSPM

--
Regards,
Sudeep

[1] http://infocenter.arm.com/help/topic/com.arm.doc.den0056a/index.html
[2] https://marc.info/?l=linux-kernel&m=149685193627620&w=2
[3] https://marc.info/?l=linux-arm-kernel&m=149849482623492&w=2
[4] https://marc.info/?l=devicetree&m=150185763105926&w=2
[5] https://marc.info/?l=devicetree&m=150660452015351&w=2

Sudeep Holla (20):
  dt-bindings: mailbox: add support for mailbox client shared memory
  dt-bindings: arm: add support for ARM System Control and Management
    Interface(SCMI) protocol
  firmware: arm_scmi: add basic driver infrastructure for SCMI
  firmware: arm_scmi: add common infrastructure and support for base
    protocol
  firmware: arm_scmi: add scmi protocol bus to enumerate protocol
    devices
  firmware: arm_scmi: add initial support for performance protocol
  firmware: arm_scmi: add initial support for clock protocol
  firmware: arm_scmi: add initial support for power protocol
  firmware: arm_scmi: add initial support for sensor protocol
  firmware: arm_scmi: probe and initialise all the supported protocols
  firmware: arm_scmi: add support for polling based SCMI transfers
  firmware: arm_scmi: add option for polling based performance domain
    operations
  firmware: arm_scmi: refactor in preparation to support per-protocol
    channels
  firmware: arm_scmi: add per-protocol channels support using idr
    objects
  firmware: arm_scmi: add device power domain support using genpd
  clk: add support for clocks provided by SCMI
  hwmon: (core) Add hwmon_max to hwmon_sensor_types enumeration
  hwmon: add support for sensors exported via ARM SCMI
  cpufreq: add support for CPU DVFS based on SCMI message protocol
  cpufreq: scmi: add support for fast frequency switching

 Documentation/devicetree/bindings/arm/arm,scmi.txt | 179 +++++
 .../devicetree/bindings/mailbox/mailbox.txt        |  28 +
 MAINTAINERS                                        |  11 +-
 drivers/clk/Kconfig                                |  10 +
 drivers/clk/Makefile                               |   1 +
 drivers/clk/clk-scmi.c                             | 213 +++++
 drivers/cpufreq/Kconfig.arm                        |  11 +
 drivers/cpufreq/Makefile                           |   1 +
 drivers/cpufreq/scmi-cpufreq.c                     | 290 +++++++
 drivers/firmware/Kconfig                           |  34 +
 drivers/firmware/Makefile                          |   1 +
 drivers/firmware/arm_scmi/Makefile                 |   5 +
 drivers/firmware/arm_scmi/base.c                   | 293 +++++++
 drivers/firmware/arm_scmi/bus.c                    | 232 ++++++
 drivers/firmware/arm_scmi/clock.c                  | 353 ++++++++
 drivers/firmware/arm_scmi/common.h                 | 116 +++
 drivers/firmware/arm_scmi/driver.c                 | 895 +++++++++++++++++++++
 drivers/firmware/arm_scmi/perf.c                   | 530 ++++++++++++
 drivers/firmware/arm_scmi/power.c                  | 255 ++++++
 drivers/firmware/arm_scmi/scmi_pm_domain.c         | 140 ++++
 drivers/firmware/arm_scmi/sensors.c                | 302 +++++++
 drivers/hwmon/Kconfig                              |  12 +
 drivers/hwmon/Makefile                             |   1 +
 drivers/hwmon/scmi-hwmon.c                         | 233 ++++++
 include/linux/hwmon.h                              |   1 +
 include/linux/scmi_protocol.h                      | 272 +++++++
 26 files changed, 4414 insertions(+), 5 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/arm,scmi.txt
 create mode 100644 drivers/clk/clk-scmi.c
 create mode 100644 drivers/cpufreq/scmi-cpufreq.c
 create mode 100644 drivers/firmware/arm_scmi/Makefile
 create mode 100644 drivers/firmware/arm_scmi/base.c
 create mode 100644 drivers/firmware/arm_scmi/bus.c
 create mode 100644 drivers/firmware/arm_scmi/clock.c
 create mode 100644 drivers/firmware/arm_scmi/common.h
 create mode 100644 drivers/firmware/arm_scmi/driver.c
 create mode 100644 drivers/firmware/arm_scmi/perf.c
 create mode 100644 drivers/firmware/arm_scmi/power.c
 create mode 100644 drivers/firmware/arm_scmi/scmi_pm_domain.c
 create mode 100644 drivers/firmware/arm_scmi/sensors.c
 create mode 100644 drivers/hwmon/scmi-hwmon.c
 create mode 100644 include/linux/scmi_protocol.h

-- 
2.7.4

[PATCH v4 01/20] dt-bindings: mailbox: add support for mailbox client shared memory

[Sudeep Holla]

Many users of the mailbox controllers depend on the shared memory
between the two end points to exchange the main data while using simple
doorbell mechanism to alert the end points of the presence of a message.

This patch defines device tree bindings to represent such shared memory
in a generic way.

Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Acked-by: Rob Herring <robh+dt@kernel.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 .../devicetree/bindings/mailbox/mailbox.txt        | 28 ++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/Documentation/devicetree/bindings/mailbox/mailbox.txt b/Documentation/devicetree/bindings/mailbox/mailbox.txt
index be05b9746c69..af8ecee2ac68 100644
--- a/Documentation/devicetree/bindings/mailbox/mailbox.txt
+++ b/Documentation/devicetree/bindings/mailbox/mailbox.txt
@@ -23,6 +23,11 @@ assign appropriate mailbox channel to client drivers.
 
 Optional property:
 - mbox-names: List of identifier strings for each mailbox channel.
+- shmem : List of phandle pointing to the shared memory(SHM) area between the
+	  users of these mailboxes for IPC, one for each mailbox. This shared
+	  memory can be part of any memory reserved for the purpose of this
+	  communication between the mailbox client and the remote.
+
 
 Example:
 	pwr_cntrl: power {
@@ -30,3 +35,26 @@ assign appropriate mailbox channel to client drivers.
 		mbox-names = "pwr-ctrl", "rpc";
 		mboxes = <&mailbox 0 &mailbox 1>;
 	};
+
+Example with shared memory(shmem):
+
+	sram: sram@50000000 {
+		compatible = "mmio-sram";
+		reg = <0x50000000 0x10000>;
+
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges = <0 0x50000000 0x10000>;
+
+		cl_shmem: shmem@0 {
+			compatible = "client-shmem";
+			reg = <0x0 0x200>;
+		};
+	};
+
+	client@2e000000 {
+		...
+		mboxes = <&mailbox 0>;
+		shmem = <&cl_shmem>;
+		..
+	};
-- 
2.7.4

[PATCH v4 02/20] dt-bindings: arm: add support for ARM System Control and Management Interface(SCMI) protocol

[Sudeep Holla]

This patch adds devicetree binding for System Control and Management
Interface (SCMI) Message Protocol used between the Application Cores(AP)
and the System Control Processor(SCP). The MHU peripheral provides a
mechanism for inter-processor communication between SCP's M3 processor
and AP.

SCP offers control and management of the core/cluster power states,
various power domain DVFS including the core/cluster, certain system
clocks configuration, thermal sensors and many others.

SCMI protocol is developed as better replacement to the existing SCPI
which is not flexible and easily extensible.

Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 Documentation/devicetree/bindings/arm/arm,scmi.txt | 179 +++++++++++++++++++++
 MAINTAINERS                                        |   4 +-
 2 files changed, 181 insertions(+), 2 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/arm,scmi.txt

diff --git a/Documentation/devicetree/bindings/arm/arm,scmi.txt b/Documentation/devicetree/bindings/arm/arm,scmi.txt
new file mode 100644
index 000000000000..5f3719ab7075
--- /dev/null
+++ b/Documentation/devicetree/bindings/arm/arm,scmi.txt
@@ -0,0 +1,179 @@
+System Control and Management Interface (SCMI) Message Protocol
+----------------------------------------------------------
+
+The SCMI is intended to allow agents such as OSPM to manage various functions
+that are provided by the hardware platform it is running on, including power
+and performance functions.
+
+This binding is intended to define the interface the firmware implementing
+the SCMI as described in ARM document number ARM DUI 0922B ("ARM System Control
+and Management Interface Platform Design Document")[0] provide for OSPM in
+the device tree.
+
+Required properties:
+
+The scmi node with the following properties shall be under the /firmware/ node.
+
+- compatible : shall be "arm,scmi"
+- mboxes: List of phandle and mailbox channel specifiers. It should contain
+	  exactly one or two mailboxes, one for transmitting messages("tx")
+	  and another optional for receiving the notifications("rx") if
+	  supported.
+- shmem : List of phandle pointing to the shared memory(SHM) area as per
+	  generic mailbox client binding.
+- #address-cells : should be '1' if the device has sub-nodes, maps to
+	  protocol identifier for a given sub-node.
+- #size-cells : should be '0' as 'reg' property doesn't have any size
+	  associated with it.
+
+Optional properties:
+
+- mbox-names: shall be "tx" or "rx" depending on mboxes entries.
+
+See Documentation/devicetree/bindings/mailbox/mailbox.txt for more details
+about the generic mailbox controller and client driver bindings.
+
+The mailbox is the only permitted method of calling the SCMI firmware.
+Mailbox doorbell is used as a mechanism to alert the presence of a
+messages and/or notification.
+
+Each protocol supported shall have a sub-node with corresponding compatible
+as described in the following sections. If the platform supports dedicated
+communication channel for a particular protocol, the 3 properties namely:
+mboxes, mbox-names and shmem shall be present in the sub-node corresponding
+to that protocol.
+
+Clock/Performance bindings for the clocks/OPPs based on SCMI Message Protocol
+------------------------------------------------------------
+
+This binding uses the common clock binding[1].
+
+Required properties:
+- #clock-cells : Should be 1. Contains the Clock ID value used by SCMI commands.
+
+Power domain bindings for the power domains based on SCMI Message Protocol
+------------------------------------------------------------
+
+This binding for the SCMI power domain providers uses the generic power
+domain binding[2].
+
+Required properties:
+ - #power-domain-cells : Should be 1. Contains the device or the power
+			 domain ID value used by SCMI commands.
+
+Sensor bindings for the sensors based on SCMI Message Protocol
+--------------------------------------------------------------
+SCMI provides an API to access the various sensors on the SoC.
+
+Required properties:
+- #thermal-sensor-cells: should be set to 1. This property follows the
+			 thermal device tree bindings[3].
+
+			 Valid cell values are raw identifiers (Sensor ID)
+			 as used by the firmware. Refer to  platform details
+			 for your implementation for the IDs to use.
+
+SRAM and Shared Memory for SCMI
+-------------------------------
+
+A small area of SRAM is reserved for SCMI communication between application
+processors and SCP.
+
+The properties should follow the generic mmio-sram description found in [4]
+
+Each sub-node represents the reserved area for SCMI.
+
+Required sub-node properties:
+- reg : The base offset and size of the reserved area with the SRAM
+- compatible : should be "arm,scmi-shmem" for Non-secure SRAM based
+	       shared memory
+
+[0] http://infocenter.arm.com/help/topic/com.arm.doc.den0056a/index.html
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+[2] Documentation/devicetree/bindings/power/power_domain.txt
+[3] Documentation/devicetree/bindings/thermal/thermal.txt
+[4] Documentation/devicetree/bindings/sram/sram.txt
+
+Example:
+
+sram@50000000 {
+	compatible = "mmio-sram";
+	reg = <0x0 0x50000000 0x0 0x10000>;
+
+	#address-cells = <1>;
+	#size-cells = <1>;
+	ranges = <0 0x0 0x50000000 0x10000>;
+
+	cpu_scp_lpri: scp-shmem@0 {
+		compatible = "arm,scmi-shmem";
+		reg = <0x0 0x200>;
+	};
+
+	cpu_scp_hpri: scp-shmem@200 {
+		compatible = "arm,scmi-shmem";
+		reg = <0x200 0x200>;
+	};
+};
+
+mailbox@40000000 {
+	....
+	#mbox-cells = <1>;
+	reg = <0x0 0x40000000 0x0 0x10000>;
+};
+
+firmware {
+
+	...
+
+	scmi {
+		compatible = "arm,scmi";
+		mboxes = <&mailbox 0 &mailbox 1>;
+		mbox-names = "tx", "rx";
+		shmem = <&cpu_scp_lpri &cpu_scp_hpri>;
+		#address-cells = <1>;
+		#size-cells = <0>;
+
+		scmi_devpd: protocol@11 {
+			reg = <0x11>;
+			#power-domain-cells = <1>;
+		};
+
+		scmi_dvfs: protocol@13 {
+			reg = <0x13>;
+			#clock-cells = <1>;
+		};
+
+		scmi_clk: protocol@14 {
+			reg = <0x14>;
+			#clock-cells = <1>;
+		};
+
+		scmi_sensors0: protocol@15 {
+			reg = <0x15>;
+			#thermal-sensor-cells = <1>;
+		};
+	};
+};
+
+cpu@0 {
+	...
+	reg = <0 0>;
+	clocks = <&scmi_dvfs 0>;
+};
+
+hdlcd@7ff60000 {
+	...
+	reg = <0 0x7ff60000 0 0x1000>;
+	clocks = <&scmi_clk 4>;
+	power-domains = <&scmi_devpd 1>;
+};
+
+thermal-zones {
+	soc_thermal {
+		polling-delay-passive = <100>;
+		polling-delay = <1000>;
+					/* sensor ID */
+		thermal-sensors = <&scmi_sensors0 3>;
+		...
+	};
+};
diff --git a/MAINTAINERS b/MAINTAINERS
index af0cb69f6a3e..25fbf29331c0 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -12949,11 +12949,11 @@ T:	git git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd.git
 S:	Supported
 F:	drivers/mfd/syscon.c
 
-SYSTEM CONTROL & POWER INTERFACE (SCPI) Message Protocol drivers
+SYSTEM CONTROL & POWER/MANAGEMENT INTERFACE (SCPI/SCMI) Message Protocol drivers
 M:	Sudeep Holla <sudeep.holla@arm.com>
 L:	linux-arm-kernel@lists.infradead.org
 S:	Maintained
-F:	Documentation/devicetree/bindings/arm/arm,scpi.txt
+F:	Documentation/devicetree/bindings/arm/arm,sc[mp]i.txt
 F:	drivers/clk/clk-scpi.c
 F:	drivers/cpufreq/scpi-cpufreq.c
 F:	drivers/firmware/arm_scpi.c
-- 
2.7.4

[PATCH v4 03/20] firmware: arm_scmi: add basic driver infrastructure for SCMI

[Sudeep Holla]

The SCMI is intended to allow OSPM to manage various functions that are
provided by the hardware platform it is running on, including power and
performance functions. SCMI provides two levels of abstraction, protocols
and transports. Protocols define individual groups of system control and
management messages. A protocol specification describes the messages
that it supports. Transports describe the method by which protocol
messages are communicated between agents and the platform.

This patch adds basic infrastructure to manage the message allocation,
initialisation, packing/unpacking and shared memory management.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 MAINTAINERS                        |   3 +-
 drivers/firmware/Kconfig           |  21 ++
 drivers/firmware/Makefile          |   1 +
 drivers/firmware/arm_scmi/Makefile |   2 +
 drivers/firmware/arm_scmi/common.h |  77 ++++
 drivers/firmware/arm_scmi/driver.c | 708 +++++++++++++++++++++++++++++++++++++
 include/linux/scmi_protocol.h      |  27 ++
 7 files changed, 838 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/Makefile
 create mode 100644 drivers/firmware/arm_scmi/common.h
 create mode 100644 drivers/firmware/arm_scmi/driver.c
 create mode 100644 include/linux/scmi_protocol.h

diff --git a/MAINTAINERS b/MAINTAINERS
index 25fbf29331c0..8ee63627f1ac 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -12957,7 +12957,8 @@ F:	Documentation/devicetree/bindings/arm/arm,sc[mp]i.txt
 F:	drivers/clk/clk-scpi.c
 F:	drivers/cpufreq/scpi-cpufreq.c
 F:	drivers/firmware/arm_scpi.c
-F:	include/linux/scpi_protocol.h
+F:	drivers/firmware/arm_scmi/
+F:	include/linux/sc[mp]i_protocol.h
 
 SYSTEM RESET/SHUTDOWN DRIVERS
 M:	Sebastian Reichel <sre@kernel.org>
diff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig
index 6e4ed5a9c6fd..86ea9d531587 100644
--- a/drivers/firmware/Kconfig
+++ b/drivers/firmware/Kconfig
@@ -19,6 +19,27 @@ config ARM_PSCI_CHECKER
 	  on and off through hotplug, so for now torture tests and PSCI checker
 	  are mutually exclusive.
 
+config ARM_SCMI_PROTOCOL
+	bool "ARM System Control and Management Interface (SCMI) Message Protocol"
+	depends on ARM || ARM64 || COMPILE_TEST
+	depends on MAILBOX
+	help
+	  ARM System Control and Management Interface (SCMI) protocol is a
+	  set of operating system-independent software interfaces that are
+	  used in system management. SCMI is extensible and currently provides
+	  interfaces for: Discovery and self-description of the interfaces
+	  it supports, Power domain management which is the ability to place
+	  a given device or domain into the various power-saving states that
+	  it supports, Performance management which is the ability to control
+	  the performance of a domain that is composed of compute engines
+	  such as application processors and other accelerators, Clock
+	  management which is the ability to set and inquire rates on platform
+	  managed clocks and Sensor management which is the ability to read
+	  sensor data, and be notified of sensor value.
+
+	  This protocol library provides interface for all the client drivers
+	  making use of the features offered by the SCMI.
+
 config ARM_SCPI_PROTOCOL
 	tristate "ARM System Control and Power Interface (SCPI) Message Protocol"
 	depends on ARM || ARM64 || COMPILE_TEST
diff --git a/drivers/firmware/Makefile b/drivers/firmware/Makefile
index a37f12e8d137..91d3ff62c653 100644
--- a/drivers/firmware/Makefile
+++ b/drivers/firmware/Makefile
@@ -23,6 +23,7 @@ obj-$(CONFIG_QCOM_SCM_32)	+= qcom_scm-32.o
 CFLAGS_qcom_scm-32.o :=$(call as-instr,.arch armv7-a\n.arch_extension sec,-DREQUIRES_SEC=1) -march=armv7-a
 obj-$(CONFIG_TI_SCI_PROTOCOL)	+= ti_sci.o
 
+obj-$(CONFIG_ARM_SCMI_PROTOCOL)	+= arm_scmi/
 obj-y				+= broadcom/
 obj-y				+= meson/
 obj-$(CONFIG_GOOGLE_FIRMWARE)	+= google/
diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
new file mode 100644
index 000000000000..b2a24ba2b636
--- /dev/null
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -0,0 +1,2 @@
+obj-y	= scmi-driver.o
+scmi-driver-y = driver.o
diff --git a/drivers/firmware/arm_scmi/common.h b/drivers/firmware/arm_scmi/common.h
new file mode 100644
index 000000000000..cdfc43427c3c
--- /dev/null
+++ b/drivers/firmware/arm_scmi/common.h
@@ -0,0 +1,77 @@
+/*
+ * System Control and Management Interface (SCMI) Message Protocol
+ * driver common header file containing some definitions, structures
+ * and function prototypes used in all the different SCMI protocols.
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/completion.h>
+#include <linux/scmi_protocol.h>
+#include <linux/types.h>
+
+/**
+ * struct scmi_msg_hdr - Message(Tx/Rx) header
+ *
+ * @id: The identifier of the command being sent
+ * @protocol_id: The identifier of the protocol used to send @id command
+ * @seq: The token to identify the message. when a message/command returns,
+ *       the platform returns the whole message header unmodified including
+ *	 the token.
+ */
+struct scmi_msg_hdr {
+	u8 id;
+	u8 protocol_id;
+	u16 seq;
+	u32 status;
+	bool poll_completion;
+};
+
+/**
+ * struct scmi_msg - Message(Tx/Rx) structure
+ *
+ * @buf: Buffer pointer
+ * @len: Length of data in the Buffer
+ */
+struct scmi_msg {
+	void *buf;
+	size_t len;
+};
+
+/**
+ * struct scmi_xfer - Structure representing a message flow
+ *
+ * @hdr: Transmit message header
+ * @tx: Transmit message
+ * @rx: Receive message, the buffer should be pre-allocated to store
+ *	message. If request-ACK protocol is used, we can reuse the same
+ *	buffer for the rx path as we use for the tx path.
+ * @done: completion event
+ */
+
+struct scmi_xfer {
+	void *con_priv;
+	struct scmi_msg_hdr hdr;
+	struct scmi_msg tx;
+	struct scmi_msg rx;
+	struct completion done;
+};
+
+void scmi_one_xfer_put(const struct scmi_handle *h, struct scmi_xfer *xfer);
+int scmi_do_xfer(const struct scmi_handle *h, struct scmi_xfer *xfer);
+int scmi_one_xfer_init(const struct scmi_handle *h, u8 msg_id, u8 prot_id,
+		       size_t tx_size, size_t rx_size, struct scmi_xfer **p);
+int scmi_handle_put(const struct scmi_handle *handle);
+struct scmi_handle *scmi_handle_get(struct device *dev);
diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c
new file mode 100644
index 000000000000..58d8f88893e6
--- /dev/null
+++ b/drivers/firmware/arm_scmi/driver.c
@@ -0,0 +1,708 @@
+/*
+ * System Control and Management Interface (SCMI) Message Protocol driver
+ *
+ * SCMI Message Protocol is used between the System Control Processor(SCP)
+ * and the Application Processors(AP). The Message Handling Unit(MHU)
+ * provides a mechanism for inter-processor communication between SCP's
+ * Cortex M3 and AP.
+ *
+ * SCP offers control and management of the core/cluster power states,
+ * various power domain DVFS including the core/cluster, certain system
+ * clocks configuration, thermal sensors and many others.
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/bitmap.h>
+#include <linux/export.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/mailbox_client.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/semaphore.h>
+#include <linux/slab.h>
+
+#include "common.h"
+
+#define MSG_ID_SHIFT		0
+#define MSG_ID_MASK		0xff
+#define MSG_TYPE_SHIFT		8
+#define MSG_TYPE_MASK		0x3
+#define MSG_PROTOCOL_ID_SHIFT	10
+#define MSG_PROTOCOL_ID_MASK	0xff
+#define MSG_TOKEN_ID_SHIFT	18
+#define MSG_TOKEN_ID_MASK	0x3ff
+#define MSG_XTRACT_TOKEN(header)	\
+	(((header) >> MSG_TOKEN_ID_SHIFT) & MSG_TOKEN_ID_MASK)
+
+enum scmi_error_codes {
+	SCMI_SUCCESS = 0,	/* Success */
+	SCMI_ERR_SUPPORT = -1,	/* Not supported */
+	SCMI_ERR_PARAMS = -2,	/* Invalid Parameters */
+	SCMI_ERR_ACCESS = -3,	/* Invalid access/permission denied */
+	SCMI_ERR_ENTRY = -4,	/* Not found */
+	SCMI_ERR_RANGE = -5,	/* Value out of range */
+	SCMI_ERR_BUSY = -6,	/* Device busy */
+	SCMI_ERR_COMMS = -7,	/* Communication Error */
+	SCMI_ERR_GENERIC = -8,	/* Generic Error */
+	SCMI_ERR_HARDWARE = -9,	/* Hardware Error */
+	SCMI_ERR_PROTOCOL = -10,/* Protocol Error */
+	SCMI_ERR_MAX
+};
+
+/* List of all  SCMI devices active in system */
+static LIST_HEAD(scmi_list);
+/* Protection for the entire list */
+static DEFINE_MUTEX(scmi_list_mutex);
+
+/**
+ * struct scmi_xfers_info - Structure to manage transfer information
+ *
+ * @sem_xfer_count: Counting Semaphore for managing max simultaneous
+ *	Messages.
+ * @xfer_block: Preallocated Message array
+ * @xfer_alloc_table: Bitmap table for allocated messages.
+ *	Index of this bitmap table is also used for message
+ *	sequence identifier.
+ * @xfer_lock: Protection for message allocation
+ */
+struct scmi_xfers_info {
+	struct semaphore sem_xfer_count;
+	struct scmi_xfer *xfer_block;
+	unsigned long *xfer_alloc_table;
+	/* protect transfer allocation */
+	spinlock_t xfer_lock;
+};
+
+/**
+ * struct scmi_desc - Description of SoC integration
+ *
+ * @max_rx_timeout_ms: Timeout for communication with SoC (in Milliseconds)
+ * @max_msg: Maximum number of messages that can be pending
+ *	simultaneously in the system
+ * @max_msg_size: Maximum size of data per message that can be handled.
+ */
+struct scmi_desc {
+	int max_rx_timeout_ms;
+	int max_msg;
+	int max_msg_size;
+};
+
+/**
+ * struct scmi_info - Structure representing a  SCMI instance
+ *
+ * @dev: Device pointer
+ * @desc: SoC description for this instance
+ * @handle: Instance of SCMI handle to send to clients
+ * @cl: Mailbox Client
+ * @tx_chan: Transmit mailbox channel
+ * @tx_payload: Transmit mailbox channel payload area
+ * @minfo: Message info
+ * @node: list head
+ * @users: Number of users of this instance
+ */
+struct scmi_info {
+	struct device *dev;
+	const struct scmi_desc *desc;
+	struct scmi_handle handle;
+	struct mbox_client cl;
+	struct mbox_chan *tx_chan;
+	void __iomem *tx_payload;
+	struct scmi_xfers_info minfo;
+	struct list_head node;
+	int users;
+};
+
+#define client_to_scmi_info(c)	container_of(c, struct scmi_info, cl)
+#define handle_to_scmi_info(h)	container_of(h, struct scmi_info, handle)
+
+/*
+ * SCMI specification requires all parameters, message headers, return
+ * arguments or any protocol data to be expressed in little endian
+ * format only.
+ */
+struct scmi_shared_mem {
+	__le32 reserved;
+	__le32 channel_status;
+#define SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR	BIT(1)
+#define SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE	BIT(0)
+	__le32 reserved1[2];
+	__le32 flags;
+#define SCMI_SHMEM_FLAG_INTR_ENABLED	BIT(0)
+	__le32 length;
+	__le32 msg_header;
+	u8 msg_payload[0];
+};
+
+static const int scmi_linux_errmap[] = {
+	/* better than switch case as long as return value is continuous */
+	0,			/* SCMI_SUCCESS */
+	-EOPNOTSUPP,		/* SCMI_ERR_SUPPORT */
+	-EINVAL,		/* SCMI_ERR_PARAM */
+	-EACCES,		/* SCMI_ERR_ACCESS */
+	-ENOENT,		/* SCMI_ERR_ENTRY */
+	-ERANGE,		/* SCMI_ERR_RANGE */
+	-EBUSY,			/* SCMI_ERR_BUSY */
+	-ECOMM,			/* SCMI_ERR_COMMS */
+	-EIO,			/* SCMI_ERR_GENERIC */
+	-EREMOTEIO,		/* SCMI_ERR_HARDWARE */
+	-EPROTO,		/* SCMI_ERR_PROTOCOL */
+};
+
+static inline int scmi_to_linux_errno(int errno)
+{
+	if (errno < SCMI_SUCCESS && errno > SCMI_ERR_MAX)
+		return scmi_linux_errmap[-errno];
+	return -EIO;
+}
+
+/**
+ * scmi_dump_header_dbg() - Helper to dump a message header.
+ *
+ * @dev: Device pointer corresponding to the SCMI entity
+ * @hdr: pointer to header.
+ */
+static inline void scmi_dump_header_dbg(struct device *dev,
+					struct scmi_msg_hdr *hdr)
+{
+	dev_dbg(dev, "Command ID: %x Sequence ID: %x Protocol: %x\n",
+		hdr->id, hdr->seq, hdr->protocol_id);
+}
+
+static void scmi_fetch_response(struct scmi_xfer *xfer,
+				struct scmi_shared_mem __iomem *mem)
+{
+	xfer->hdr.status = ioread32(mem->msg_payload);
+	/* Skip the length of header and statues in payload area i.e 8 bytes*/
+	xfer->rx.len = min_t(size_t, xfer->rx.len, ioread32(&mem->length) - 8);
+
+	/* Take a copy to the rx buffer.. */
+	memcpy_fromio(xfer->rx.buf, mem->msg_payload + 4, xfer->rx.len);
+}
+
+/**
+ * scmi_rx_callback() - mailbox client callback for receive messages
+ *
+ * @cl: client pointer
+ * @m: mailbox message
+ *
+ * Processes one received message to appropriate transfer information and
+ * signals completion of the transfer.
+ *
+ * NOTE: This function will be invoked in IRQ context, hence should be
+ * as optimal as possible.
+ */
+static void scmi_rx_callback(struct mbox_client *cl, void *m)
+{
+	u16 xfer_id;
+	struct scmi_xfer *xfer;
+	struct scmi_info *info = client_to_scmi_info(cl);
+	struct scmi_xfers_info *minfo = &info->minfo;
+	struct device *dev = info->dev;
+	struct scmi_shared_mem __iomem *mem = info->tx_payload;
+
+	xfer_id = MSG_XTRACT_TOKEN(ioread32(&mem->msg_header));
+
+	/*
+	 * Are we even expecting this?
+	 */
+	if (!test_bit(xfer_id, minfo->xfer_alloc_table)) {
+		dev_err(dev, "message for %d is not expected!\n", xfer_id);
+		return;
+	}
+
+	xfer = &minfo->xfer_block[xfer_id];
+
+	scmi_dump_header_dbg(dev, &xfer->hdr);
+	/* Is the message of valid length? */
+	if (xfer->rx.len > info->desc->max_msg_size) {
+		dev_err(dev, "unable to handle %zu xfer(max %d)\n",
+			xfer->rx.len, info->desc->max_msg_size);
+		return;
+	}
+
+	scmi_fetch_response(xfer, mem);
+	complete(&xfer->done);
+}
+
+/**
+ * pack_scmi_header() - packs and returns 32-bit header
+ *
+ * @hdr: pointer to header containing all the information on message id,
+ *	protocol id and sequence id.
+ */
+static inline u32 pack_scmi_header(struct scmi_msg_hdr *hdr)
+{
+	return ((hdr->id & MSG_ID_MASK) << MSG_ID_SHIFT) |
+	   ((hdr->seq & MSG_TOKEN_ID_MASK) << MSG_TOKEN_ID_SHIFT) |
+	   ((hdr->protocol_id & MSG_PROTOCOL_ID_MASK) << MSG_PROTOCOL_ID_SHIFT);
+}
+
+/**
+ * scmi_tx_prepare() - mailbox client callback to prepare for the transfer
+ *
+ * @cl: client pointer
+ * @m: mailbox message
+ *
+ * This function prepares the shared memory which contains the header and the
+ * payload.
+ */
+static void scmi_tx_prepare(struct mbox_client *cl, void *m)
+{
+	struct scmi_xfer *t = m;
+	struct scmi_info *info = client_to_scmi_info(cl);
+	struct scmi_shared_mem __iomem *mem = info->tx_payload;
+
+	/* Mark channel busy + clear error */
+	iowrite32(0x0, &mem->channel_status);
+	iowrite32(t->hdr.poll_completion ? 0 : SCMI_SHMEM_FLAG_INTR_ENABLED,
+		  &mem->flags);
+	iowrite32(sizeof(mem->msg_header) + t->tx.len, &mem->length);
+	iowrite32(pack_scmi_header(&t->hdr), &mem->msg_header);
+	if (t->tx.buf)
+		memcpy_toio(mem->msg_payload, t->tx.buf, t->tx.len);
+}
+
+/**
+ * scmi_one_xfer_get() - Allocate one message
+ *
+ * @handle: SCMI entity handle
+ *
+ * Helper function which is used by various command functions that are
+ * exposed to clients of this driver for allocating a message traffic event.
+ *
+ * This function can sleep depending on pending requests already in the system
+ * for the SCMI entity. Further, this also holds a spinlock to maintain
+ * integrity of internal data structures.
+ *
+ * Return: 0 if all went fine, else corresponding error.
+ */
+static struct scmi_xfer *scmi_one_xfer_get(const struct scmi_handle *handle)
+{
+	u16 xfer_id;
+	int ret, timeout;
+	struct scmi_xfer *xfer;
+	unsigned long flags, bit_pos;
+	struct scmi_info *info = handle_to_scmi_info(handle);
+	struct scmi_xfers_info *minfo = &info->minfo;
+
+	/*
+	 * Ensure we have only controlled number of pending messages.
+	 * Ideally, we might just have to wait a single message, be
+	 * conservative and wait 5 times that..
+	 */
+	timeout = msecs_to_jiffies(info->desc->max_rx_timeout_ms) * 5;
+	ret = down_timeout(&minfo->sem_xfer_count, timeout);
+	if (ret < 0)
+		return ERR_PTR(ret);
+
+	/* Keep the locked section as small as possible */
+	spin_lock_irqsave(&minfo->xfer_lock, flags);
+	bit_pos = find_first_zero_bit(minfo->xfer_alloc_table,
+				      info->desc->max_msg);
+	if (bit_pos == info->desc->max_msg) {
+		spin_unlock_irqrestore(&minfo->xfer_lock, flags);
+		return ERR_PTR(-ENOMEM);
+	}
+	set_bit(bit_pos, minfo->xfer_alloc_table);
+	spin_unlock_irqrestore(&minfo->xfer_lock, flags);
+
+	xfer_id = bit_pos;
+
+	xfer = &minfo->xfer_block[xfer_id];
+	xfer->hdr.seq = xfer_id;
+	reinit_completion(&xfer->done);
+
+	return xfer;
+}
+
+/**
+ * scmi_one_xfer_put() - Release a message
+ *
+ * @minfo: transfer info pointer
+ * @xfer: message that was reserved by scmi_one_xfer_get
+ *
+ * This holds a spinlock to maintain integrity of internal data structures.
+ */
+void scmi_one_xfer_put(const struct scmi_handle *handle, struct scmi_xfer *xfer)
+{
+	unsigned long flags;
+	struct scmi_info *info = handle_to_scmi_info(handle);
+	struct scmi_xfers_info *minfo = &info->minfo;
+
+	/*
+	 * Keep the locked section as small as possible
+	 * NOTE: we might escape with smp_mb and no lock here..
+	 * but just be conservative and symmetric.
+	 */
+	spin_lock_irqsave(&minfo->xfer_lock, flags);
+	clear_bit(xfer->hdr.seq, minfo->xfer_alloc_table);
+	spin_unlock_irqrestore(&minfo->xfer_lock, flags);
+
+	/* Increment the count for the next user to get through */
+	up(&minfo->sem_xfer_count);
+}
+
+/**
+ * scmi_do_xfer() - Do one transfer
+ *
+ * @info: Pointer to SCMI entity information
+ * @xfer: Transfer to initiate and wait for response
+ *
+ * Return: -ETIMEDOUT in case of no response, if transmit error,
+ *   return corresponding error, else if all goes well,
+ *   return 0.
+ */
+int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
+{
+	int ret;
+	int timeout;
+	struct scmi_info *info = handle_to_scmi_info(handle);
+	struct device *dev = info->dev;
+
+	ret = mbox_send_message(info->tx_chan, xfer);
+	if (ret < 0) {
+		dev_dbg(dev, "mbox send fail %d\n", ret);
+		return ret;
+	}
+
+	/* mbox_send_message returns non-negative value on success, so reset */
+	ret = 0;
+
+	/* And we wait for the response. */
+	timeout = msecs_to_jiffies(info->desc->max_rx_timeout_ms);
+	if (!wait_for_completion_timeout(&xfer->done, timeout)) {
+		dev_err(dev, "mbox timed out in resp(caller: %pF)\n",
+			(void *)_RET_IP_);
+		ret = -ETIMEDOUT;
+	} else if (xfer->hdr.status) {
+		ret = scmi_to_linux_errno(xfer->hdr.status);
+	}
+	/*
+	 * NOTE: we might prefer not to need the mailbox ticker to manage the
+	 * transfer queueing since the protocol layer queues things by itself.
+	 * Unfortunately, we have to kick the mailbox framework after we have
+	 * received our message.
+	 */
+	mbox_client_txdone(info->tx_chan, ret);
+
+	return ret;
+}
+
+/**
+ * scmi_one_xfer_init() - Allocate and initialise one message
+ *
+ * @handle: SCMI entity handle
+ * @msg_id: Message identifier
+ * @msg_prot_id: Protocol identifier for the message
+ * @tx_size: transmit message size
+ * @rx_size: receive message size
+ * @p: pointer to the allocated and initialised message
+ *
+ * This function allocates the message using @scmi_one_xfer_get and
+ * initialise the header.
+ *
+ * Return: 0 if all went fine with @p pointing to message, else
+ *	corresponding error.
+ */
+int scmi_one_xfer_init(const struct scmi_handle *handle, u8 msg_id, u8 prot_id,
+		       size_t tx_size, size_t rx_size, struct scmi_xfer **p)
+{
+	int ret;
+	struct scmi_xfer *xfer;
+	struct scmi_info *info = handle_to_scmi_info(handle);
+	struct device *dev = info->dev;
+
+	/* Ensure we have sane transfer sizes */
+	if (rx_size > info->desc->max_msg_size ||
+	    tx_size > info->desc->max_msg_size)
+		return -ERANGE;
+
+	xfer = scmi_one_xfer_get(handle);
+	if (IS_ERR(xfer)) {
+		ret = PTR_ERR(xfer);
+		dev_err(dev, "failed to get free message slot(%d)\n", ret);
+		return ret;
+	}
+
+	xfer->tx.len = tx_size;
+	xfer->rx.len = rx_size ? : info->desc->max_msg_size;
+	xfer->hdr.id = msg_id;
+	xfer->hdr.protocol_id = prot_id;
+	xfer->hdr.poll_completion = false;
+
+	*p = xfer;
+	return 0;
+}
+
+/**
+ * scmi_handle_get() - Get the  SCMI handle for a device
+ *
+ * @dev: pointer to device for which we want SCMI handle
+ *
+ * NOTE: The function does not track individual clients of the framework
+ * and is expected to be maintained by caller of  SCMI protocol library.
+ * scmi_handle_put must be balanced with successful scmi_handle_get
+ *
+ * Return: pointer to handle if successful, NULL on error
+ */
+struct scmi_handle *scmi_handle_get(struct device *dev)
+{
+	struct list_head *p;
+	struct scmi_info *info;
+	struct scmi_handle *handle = NULL;
+
+	mutex_lock(&scmi_list_mutex);
+	list_for_each(p, &scmi_list) {
+		info = list_entry(p, struct scmi_info, node);
+		if (dev->parent == info->dev) {
+			handle = &info->handle;
+			info->users++;
+			break;
+		}
+	}
+	mutex_unlock(&scmi_list_mutex);
+
+	return handle;
+}
+
+/**
+ * scmi_handle_put() - Release the handle acquired by scmi_handle_get
+ *
+ * @handle: handle acquired by scmi_handle_get
+ *
+ * NOTE: The function does not track individual clients of the framework
+ * and is expected to be maintained by caller of  SCMI protocol library.
+ * scmi_handle_put must be balanced with successful scmi_handle_get
+ *
+ * Return: 0 is successfully released
+ *	if null was passed, it returns -EINVAL;
+ */
+int scmi_handle_put(const struct scmi_handle *handle)
+{
+	struct scmi_info *info;
+
+	if (!handle)
+		return -EINVAL;
+
+	info = handle_to_scmi_info(handle);
+	mutex_lock(&scmi_list_mutex);
+	if (!WARN_ON(!info->users))
+		info->users--;
+	mutex_unlock(&scmi_list_mutex);
+
+	return 0;
+}
+
+static const struct scmi_desc scmi_generic_desc = {
+	.max_rx_timeout_ms = 30,	/* we may increase this if required */
+	.max_msg = 20,		/* Limited by MBOX_TX_QUEUE_LEN */
+	.max_msg_size = 128,
+};
+
+/* Each compatible listed below must have descriptor associated with it */
+static const struct of_device_id scmi_of_match[] = {
+	{ .compatible = "arm,scmi", .data = &scmi_generic_desc },
+	{ /* Sentinel */ },
+};
+
+MODULE_DEVICE_TABLE(of, scmi_of_match);
+
+static int scmi_xfer_info_init(struct scmi_info *sinfo)
+{
+	int i;
+	struct scmi_xfer *xfer;
+	struct device *dev = sinfo->dev;
+	const struct scmi_desc *desc = sinfo->desc;
+	struct scmi_xfers_info *info = &sinfo->minfo;
+
+	/* Pre-allocated messages, no more than what hdr.seq can support */
+	if (WARN_ON(desc->max_msg >= (MSG_TOKEN_ID_MASK + 1))) {
+		dev_err(dev, "Maximum message of %d exceeds supported %d\n",
+			desc->max_msg, MSG_TOKEN_ID_MASK + 1);
+		return -EINVAL;
+	}
+
+	info->xfer_block = devm_kcalloc(dev, desc->max_msg,
+					sizeof(*info->xfer_block), GFP_KERNEL);
+	if (!info->xfer_block)
+		return -ENOMEM;
+
+	info->xfer_alloc_table = devm_kcalloc(dev, BITS_TO_LONGS(desc->max_msg),
+					      sizeof(long), GFP_KERNEL);
+	if (!info->xfer_alloc_table)
+		return -ENOMEM;
+
+	bitmap_zero(info->xfer_alloc_table, desc->max_msg);
+
+	/* Pre-initialize the buffer pointer to pre-allocated buffers */
+	for (i = 0, xfer = info->xfer_block; i < desc->max_msg; i++, xfer++) {
+		xfer->rx.buf = devm_kcalloc(dev, sizeof(u8), desc->max_msg_size,
+					    GFP_KERNEL);
+		if (!xfer->rx.buf)
+			return -ENOMEM;
+
+		xfer->tx.buf = xfer->rx.buf;
+		init_completion(&xfer->done);
+	}
+
+	spin_lock_init(&info->xfer_lock);
+
+	sema_init(&info->sem_xfer_count, desc->max_msg);
+
+	return 0;
+}
+
+static int scmi_mailbox_check(struct device_node *np)
+{
+	struct of_phandle_args arg;
+
+	return of_parse_phandle_with_args(np, "mboxes", "#mbox-cells", 0, &arg);
+}
+
+static int scmi_mbox_free_channel(struct scmi_info *info)
+{
+	if (!IS_ERR_OR_NULL(info->tx_chan)) {
+		mbox_free_channel(info->tx_chan);
+		info->tx_chan = NULL;
+	}
+
+	return 0;
+}
+
+static int scmi_remove(struct platform_device *pdev)
+{
+	int ret = 0;
+	struct scmi_info *info = platform_get_drvdata(pdev);
+
+	mutex_lock(&scmi_list_mutex);
+	if (info->users)
+		ret = -EBUSY;
+	else
+		list_del(&info->node);
+	mutex_unlock(&scmi_list_mutex);
+
+	if (!ret)
+		/* Safe to free channels since no more users */
+		return scmi_mbox_free_channel(info);
+
+	return ret;
+}
+
+static inline int scmi_mbox_chan_setup(struct scmi_info *info)
+{
+	int ret;
+	struct resource res;
+	resource_size_t size;
+	struct device *dev = info->dev;
+	struct device_node *shmem, *np = dev->of_node;
+	struct mbox_client *cl;
+
+	cl = &info->cl;
+	cl->dev = dev;
+	cl->rx_callback = scmi_rx_callback;
+	cl->tx_prepare = scmi_tx_prepare;
+	cl->tx_block = false;
+	cl->knows_txdone = true;
+
+	shmem = of_parse_phandle(np, "shmem", 0);
+	ret = of_address_to_resource(shmem, 0, &res);
+	of_node_put(shmem);
+	if (ret) {
+		dev_err(dev, "failed to get SCMI Tx payload mem resource\n");
+		return ret;
+	}
+
+	size = resource_size(&res);
+	info->tx_payload = devm_ioremap(dev, res.start, size);
+	if (!info->tx_payload) {
+		dev_err(dev, "failed to ioremap SCMI Tx payload\n");
+		return -EADDRNOTAVAIL;
+	}
+
+	/* Transmit channel is first entry i.e. index 0 */
+	info->tx_chan = mbox_request_channel(cl, 0);
+	if (IS_ERR(info->tx_chan)) {
+		ret = PTR_ERR(info->tx_chan);
+		if (ret != -EPROBE_DEFER)
+			dev_err(dev, "failed to request SCMI Tx mailbox\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+static int scmi_probe(struct platform_device *pdev)
+{
+	int ret;
+	struct scmi_handle *handle;
+	const struct scmi_desc *desc;
+	struct scmi_info *info;
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev->of_node;
+
+	/* Only mailbox method supported, check for the presence of one */
+	if (scmi_mailbox_check(np)) {
+		dev_err(dev, "no mailbox found in %pOF\n", np);
+		return -EINVAL;
+	}
+
+	desc = of_match_device(scmi_of_match, dev)->data;
+
+	info = devm_kzalloc(dev, sizeof(*info), GFP_KERNEL);
+	if (!info)
+		return -ENOMEM;
+
+	info->dev = dev;
+	info->desc = desc;
+	INIT_LIST_HEAD(&info->node);
+
+	ret = scmi_xfer_info_init(info);
+	if (ret)
+		return ret;
+
+	platform_set_drvdata(pdev, info);
+
+	handle = &info->handle;
+	handle->dev = info->dev;
+
+	ret = scmi_mbox_chan_setup(info);
+	if (ret)
+		return ret;
+
+	mutex_lock(&scmi_list_mutex);
+	list_add_tail(&info->node, &scmi_list);
+	mutex_unlock(&scmi_list_mutex);
+
+	return 0;
+}
+
+static struct platform_driver scmi_driver = {
+	.driver = {
+		   .name = "arm-scmi",
+		   .of_match_table = scmi_of_match,
+		   },
+	.probe = scmi_probe,
+	.remove = scmi_remove,
+};
+
+module_platform_driver(scmi_driver);
+
+MODULE_ALIAS("platform: arm-scmi");
+MODULE_AUTHOR("Sudeep Holla <sudeep.holla@arm.com>");
+MODULE_DESCRIPTION("ARM SCMI protocol driver");
+MODULE_LICENSE("GPL v2");
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
new file mode 100644
index 000000000000..854ed2479993
--- /dev/null
+++ b/include/linux/scmi_protocol.h
@@ -0,0 +1,27 @@
+/*
+ * SCMI Message Protocol driver header
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/types.h>
+
+/**
+ * struct scmi_handle - Handle returned to ARM SCMI clients for usage.
+ *
+ * @dev: pointer to the SCMI device
+ */
+struct scmi_handle {
+	struct device *dev;
+};
-- 
2.7.4

[PATCH v4 04/20] firmware: arm_scmi: add common infrastructure and support for base protocol

[Sudeep Holla]

The base protocol describes the properties of the implementation and
provide generic error management. The base protocol provides commands
to describe protocol version, discover implementation specific
attributes and vendor/sub-vendor identification, list of protocols
implemented and the various agents are in the system including OSPM
and the platform. It also supports registering for notifications of
platform errors.

This protocol is mandatory. This patch adds support for the same along
with some basic infrastructure to add support for other protocols.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/Makefile |   3 +-
 drivers/firmware/arm_scmi/base.c   | 293 +++++++++++++++++++++++++++++++++++++
 drivers/firmware/arm_scmi/common.h |  37 +++++
 drivers/firmware/arm_scmi/driver.c |  53 +++++++
 include/linux/scmi_protocol.h      |  37 +++++
 5 files changed, 422 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/base.c

diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
index b2a24ba2b636..5d9c7ef35f0f 100644
--- a/drivers/firmware/arm_scmi/Makefile
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -1,2 +1,3 @@
-obj-y	= scmi-driver.o
+obj-y	= scmi-driver.o scmi-protocols.o
 scmi-driver-y = driver.o
+scmi-protocols-y = base.o
diff --git a/drivers/firmware/arm_scmi/base.c b/drivers/firmware/arm_scmi/base.c
new file mode 100644
index 000000000000..2b5fbb724899
--- /dev/null
+++ b/drivers/firmware/arm_scmi/base.c
@@ -0,0 +1,293 @@
+/*
+ * System Control and Management Interface (SCMI) Base Protocol
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "common.h"
+
+enum scmi_base_protocol_cmd {
+	BASE_DISCOVER_VENDOR = 0x3,
+	BASE_DISCOVER_SUB_VENDOR = 0x4,
+	BASE_DISCOVER_IMPLEMENT_VERSION = 0x5,
+	BASE_DISCOVER_LIST_PROTOCOLS = 0x6,
+	BASE_DISCOVER_AGENT = 0x7,
+	BASE_NOTIFY_ERRORS = 0x8,
+};
+
+struct scmi_msg_resp_base_attributes {
+	    u8 num_protocols;
+	    u8 num_agents;
+	__le16 reserved;
+};
+
+/**
+ * scmi_base_attributes_get() - gets the implementation details
+ *	that are associated with the base protocol.
+ *
+ * @handle - SCMI entity handle
+ *
+ * Return: 0 on success, else appropriate SCMI error.
+ */
+static int scmi_base_attributes_get(const struct scmi_handle *handle)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_base_attributes *attr_info;
+	struct scmi_revision_info *rev = handle->version;
+
+	ret = scmi_one_xfer_init(handle, PROTOCOL_ATTRIBUTES,
+				 SCMI_PROTOCOL_BASE, 0, sizeof(*attr_info), &t);
+	if (ret)
+		return ret;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		attr_info = t->rx.buf;
+		rev->num_protocols = attr_info->num_protocols;
+		rev->num_agents = attr_info->num_agents;
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+/**
+ * scmi_base_vendor_id_get() - gets vendor/subvendor identifier ASCII string.
+ *
+ * @handle - SCMI entity handle
+ * @sub_vendor - specify true if sub-vendor ID is needed
+ *
+ * Return: 0 on success, else appropriate SCMI error.
+ */
+static int
+scmi_base_vendor_id_get(const struct scmi_handle *handle, bool sub_vendor)
+{
+	u8 cmd;
+	int ret, size;
+	char *vendor_id;
+	struct scmi_xfer *t;
+	struct scmi_revision_info *rev = handle->version;
+
+	if (sub_vendor) {
+		cmd = BASE_DISCOVER_SUB_VENDOR;
+		vendor_id = rev->sub_vendor_id;
+		size = ARRAY_SIZE(rev->sub_vendor_id);
+	} else {
+		cmd = BASE_DISCOVER_VENDOR;
+		vendor_id = rev->vendor_id;
+		size = ARRAY_SIZE(rev->vendor_id);
+	}
+
+	ret = scmi_one_xfer_init(handle, cmd, SCMI_PROTOCOL_BASE, 0, size, &t);
+	if (ret)
+		return ret;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret)
+		memcpy(vendor_id, t->rx.buf, size);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+/**
+ * scmi_base_implementation_version_get() - gets a vendor-specific
+ *	implementation 32-bit version. The format of the version number is
+ *	vendor-specific
+ *
+ * @handle - SCMI entity handle
+ *
+ * Return: 0 on success, else appropriate SCMI error.
+ */
+static int
+scmi_base_implementation_version_get(const struct scmi_handle *handle)
+{
+	int ret;
+	__le32 *impl_ver;
+	struct scmi_xfer *t;
+	struct scmi_revision_info *rev = handle->version;
+
+	ret = scmi_one_xfer_init(handle, BASE_DISCOVER_IMPLEMENT_VERSION,
+				 SCMI_PROTOCOL_BASE, 0, sizeof(*impl_ver), &t);
+	if (ret)
+		return ret;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		impl_ver = t->rx.buf;
+		rev->impl_ver = le32_to_cpu(*impl_ver);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+/**
+ * scmi_base_implementation_list_get() - gets the list of protocols it is
+ *	OSPM is allowed to access
+ *
+ * @handle - SCMI entity handle
+ * @protocols_imp - pointer to hold the list of protocol identifiers
+ *
+ * Return: 0 on success, else appropriate SCMI error.
+ */
+static int scmi_base_implementation_list_get(const struct scmi_handle *handle,
+					     u8 *protocols_imp)
+{
+	u8 *list;
+	int ret, loop;
+	struct scmi_xfer *t;
+	__le32 *num_skip, *num_ret;
+	u32 tot_num_ret = 0, loop_num_ret;
+	struct device *dev = handle->dev;
+
+	ret = scmi_one_xfer_init(handle, BASE_DISCOVER_LIST_PROTOCOLS,
+				 SCMI_PROTOCOL_BASE, sizeof(*num_skip), 0, &t);
+	if (ret)
+		return ret;
+
+	num_skip = t->tx.buf;
+	num_ret = t->rx.buf;
+	list = t->rx.buf + sizeof(*num_ret);
+
+	do {
+		/* Set the number of protocols to be skipped/already read */
+		*num_skip = cpu_to_le32(tot_num_ret);
+
+		ret = scmi_do_xfer(handle, t);
+		if (ret)
+			break;
+
+		loop_num_ret = le32_to_cpu(*num_ret);
+		if (tot_num_ret + loop_num_ret > MAX_PROTOCOLS_IMP) {
+			dev_err(dev, "No. of Protocol > MAX_PROTOCOLS_IMP");
+			break;
+		}
+
+		for (loop = 0; loop < loop_num_ret; loop++)
+			protocols_imp[tot_num_ret + loop] = *(list + loop);
+
+		tot_num_ret += loop_num_ret;
+	} while (loop_num_ret);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+/**
+ * scmi_base_discover_agent_get() - discover the name of an agent
+ *
+ * @handle - SCMI entity handle
+ * @id - Agent identifier
+ * @name - Agent identifier ASCII string
+ *
+ * An agent id of 0 is reserved to identify the platform itself.
+ * Generally operating system is represented as "OSPM"
+ *
+ * Return: 0 on success, else appropriate SCMI error.
+ */
+static int scmi_base_discover_agent_get(const struct scmi_handle *handle,
+					int id, char *name)
+{
+	int ret;
+	struct scmi_xfer *t;
+
+	ret = scmi_one_xfer_init(handle, BASE_DISCOVER_AGENT,
+				 SCMI_PROTOCOL_BASE, sizeof(__le32),
+				 SCMI_MAX_STR_SIZE, &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(id);
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret)
+		memcpy(name, t->rx.buf, SCMI_MAX_STR_SIZE);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+/**
+ * scmi_base_error_notifications_enable() - register/unregister for
+ *	notifications of errors in the platform
+ *
+ * @handle - SCMI entity handle
+ * @enable - Enable/Disable the notification
+ *
+ * Return: 0 on success, else appropriate SCMI error.
+ */
+static int
+scmi_base_error_notifications_enable(const struct scmi_handle *handle, bool en)
+{
+	int ret;
+	struct scmi_xfer *t;
+
+	ret = scmi_one_xfer_init(handle, BASE_NOTIFY_ERRORS, SCMI_PROTOCOL_BASE,
+				 sizeof(__le32), 0, &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(en & BIT(0));
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+int scmi_base_protocol_init(struct scmi_handle *h)
+{
+	int id, ret;
+	u8 *prot_imp;
+	u32 version;
+	char name[SCMI_MAX_STR_SIZE];
+	const struct scmi_handle *handle = h;
+	struct device *dev = handle->dev;
+	struct scmi_revision_info *rev = handle->version;
+
+	ret = scmi_version_get(handle, SCMI_PROTOCOL_BASE, &version);
+	if (ret)
+		return ret;
+
+	prot_imp = devm_kcalloc(dev, MAX_PROTOCOLS_IMP, sizeof(u8), GFP_KERNEL);
+	if (!prot_imp)
+		return -ENOMEM;
+
+	rev->major_ver = PROTOCOL_REV_MAJOR(version),
+	rev->minor_ver = PROTOCOL_REV_MINOR(version);
+
+	scmi_base_attributes_get(handle);
+	scmi_base_vendor_id_get(handle, false);
+	scmi_base_vendor_id_get(handle, true);
+	scmi_base_implementation_version_get(handle);
+	scmi_base_implementation_list_get(handle, prot_imp);
+	scmi_base_error_notifications_enable(handle, true);
+	scmi_setup_protocol_implemented(handle, prot_imp);
+
+	dev_info(dev, "SCMI Protocol v%d.%d '%s:%s' Firmware version 0x%x\n",
+		 rev->major_ver, rev->minor_ver, rev->vendor_id,
+		 rev->sub_vendor_id, rev->impl_ver);
+	dev_dbg(dev, "Found %d protocol(s) %d agent(s)\n", rev->num_protocols,
+		rev->num_agents);
+
+	for (id = 0; id < rev->num_agents; id++) {
+		scmi_base_discover_agent_get(handle, id, name);
+		dev_dbg(dev, "Agent %d: %s\n", id, name);
+	}
+
+	return 0;
+}
diff --git a/drivers/firmware/arm_scmi/common.h b/drivers/firmware/arm_scmi/common.h
index cdfc43427c3c..bc767f4997e0 100644
--- a/drivers/firmware/arm_scmi/common.h
+++ b/drivers/firmware/arm_scmi/common.h
@@ -19,9 +19,41 @@
  */
 
 #include <linux/completion.h>
+#include <linux/device.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
 #include <linux/scmi_protocol.h>
 #include <linux/types.h>
 
+#define PROTOCOL_REV_MINOR_BITS	16
+#define PROTOCOL_REV_MINOR_MASK	((1U << PROTOCOL_REV_MINOR_BITS) - 1)
+#define PROTOCOL_REV_MAJOR(x)	((x) >> PROTOCOL_REV_MINOR_BITS)
+#define PROTOCOL_REV_MINOR(x)	((x) & PROTOCOL_REV_MINOR_MASK)
+#define MAX_PROTOCOLS_IMP	16
+
+enum scmi_common_cmd {
+	PROTOCOL_VERSION = 0x0,
+	PROTOCOL_ATTRIBUTES = 0x1,
+	PROTOCOL_MESSAGE_ATTRIBUTES = 0x2,
+};
+
+/**
+ * struct scmi_msg_resp_prot_version - Response for a message
+ *
+ * @major_version: Major version of the ABI that firmware supports
+ * @minor_version: Minor version of the ABI that firmware supports
+ *
+ * In general, ABI version changes follow the rule that minor version increments
+ * are backward compatible. Major revision changes in ABI may not be
+ * backward compatible.
+ *
+ * Response to a generic message with message type SCMI_MSG_VERSION
+ */
+struct scmi_msg_resp_prot_version {
+	__le16 minor_version;
+	__le16 major_version;
+};
+
 /**
  * struct scmi_msg_hdr - Message(Tx/Rx) header
  *
@@ -75,3 +107,8 @@ int scmi_one_xfer_init(const struct scmi_handle *h, u8 msg_id, u8 prot_id,
 		       size_t tx_size, size_t rx_size, struct scmi_xfer **p);
 int scmi_handle_put(const struct scmi_handle *handle);
 struct scmi_handle *scmi_handle_get(struct device *dev);
+int scmi_version_get(const struct scmi_handle *h, u8 protocol, u32 *version);
+void scmi_setup_protocol_implemented(const struct scmi_handle *handle,
+				     u8 *prot_imp);
+
+int scmi_base_protocol_init(struct scmi_handle *h);
diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c
index 58d8f88893e6..20d083be474c 100644
--- a/drivers/firmware/arm_scmi/driver.c
+++ b/drivers/firmware/arm_scmi/driver.c
@@ -108,21 +108,27 @@ struct scmi_desc {
  * @dev: Device pointer
  * @desc: SoC description for this instance
  * @handle: Instance of SCMI handle to send to clients
+ * @version: SCMI revision information containing protocol version,
+ *	implementation version and (sub-)vendor identification.
  * @cl: Mailbox Client
  * @tx_chan: Transmit mailbox channel
  * @tx_payload: Transmit mailbox channel payload area
  * @minfo: Message info
+ * @protocols_imp: list of protocols implemented, currently maximum of
+ *	MAX_PROTOCOLS_IMP elements allocated by the base protocol
  * @node: list head
  * @users: Number of users of this instance
  */
 struct scmi_info {
 	struct device *dev;
 	const struct scmi_desc *desc;
+	struct scmi_revision_info version;
 	struct scmi_handle handle;
 	struct mbox_client cl;
 	struct mbox_chan *tx_chan;
 	void __iomem *tx_payload;
 	struct scmi_xfers_info minfo;
+	u8 *protocols_imp;
 	struct list_head node;
 	int users;
 };
@@ -450,6 +456,45 @@ int scmi_one_xfer_init(const struct scmi_handle *handle, u8 msg_id, u8 prot_id,
 }
 
 /**
+ * scmi_version_get() - command to get the revision of the SCMI entity
+ *
+ * @handle: Handle to SCMI entity information
+ *
+ * Updates the SCMI information in the internal data structure.
+ *
+ * Return: 0 if all went fine, else return appropriate error.
+ */
+int scmi_version_get(const struct scmi_handle *handle, u8 protocol,
+		     u32 *version)
+{
+	int ret;
+	__le32 *rev_info;
+	struct scmi_xfer *t;
+
+	ret = scmi_one_xfer_init(handle, PROTOCOL_VERSION, protocol, 0,
+				 sizeof(*version), &t);
+	if (ret)
+		return ret;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		rev_info = t->rx.buf;
+		*version = le32_to_cpu(*rev_info);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+void scmi_setup_protocol_implemented(const struct scmi_handle *handle,
+				     u8 *prot_imp)
+{
+	struct scmi_info *info = handle_to_scmi_info(handle);
+
+	info->protocols_imp = prot_imp;
+}
+
+/**
  * scmi_handle_get() - Get the  SCMI handle for a device
  *
  * @dev: pointer to device for which we want SCMI handle
@@ -679,11 +724,19 @@ static int scmi_probe(struct platform_device *pdev)
 
 	handle = &info->handle;
 	handle->dev = info->dev;
+	handle->version = &info->version;
 
 	ret = scmi_mbox_chan_setup(info);
 	if (ret)
 		return ret;
 
+	ret = scmi_base_protocol_init(handle);
+	if (ret) {
+		dev_err(dev, "unable to communicate with SCMI(%d)\n", ret);
+		scmi_mbox_free_channel(info);
+		return ret;
+	}
+
 	mutex_lock(&scmi_list_mutex);
 	list_add_tail(&info->node, &scmi_list);
 	mutex_unlock(&scmi_list_mutex);
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
index 854ed2479993..664da3d763f2 100644
--- a/include/linux/scmi_protocol.h
+++ b/include/linux/scmi_protocol.h
@@ -17,11 +17,48 @@
  */
 #include <linux/types.h>
 
+#define SCMI_MAX_STR_SIZE	16
+
+/**
+ * struct scmi_revision_info - version information structure
+ *
+ * @major_ver: Major ABI version. Change here implies risk of backward
+ *	compatibility break.
+ * @minor_ver: Minor ABI version. Change here implies new feature addition,
+ *	or compatible change in ABI.
+ * @num_protocols: Number of protocols that are implemented, excluding the
+ *	base protocol.
+ * @num_agents: Number of agents in the system.
+ * @impl_ver: A vendor-specific implementation version.
+ * @vendor_id: A vendor identifier(Null terminated ASCII string)
+ * @sub_vendor_id: A sub-vendor identifier(Null terminated ASCII string)
+ */
+struct scmi_revision_info {
+	u16 major_ver;
+	u16 minor_ver;
+	u8 num_protocols;
+	u8 num_agents;
+	u32 impl_ver;
+	char vendor_id[SCMI_MAX_STR_SIZE];
+	char sub_vendor_id[SCMI_MAX_STR_SIZE];
+};
+
 /**
  * struct scmi_handle - Handle returned to ARM SCMI clients for usage.
  *
  * @dev: pointer to the SCMI device
+ * @version: pointer to the structure containing SCMI version information
  */
 struct scmi_handle {
 	struct device *dev;
+	struct scmi_revision_info *version;
+};
+
+enum scmi_std_protocol {
+	SCMI_PROTOCOL_BASE = 0x10,
+	SCMI_PROTOCOL_POWER = 0x11,
+	SCMI_PROTOCOL_SYSTEM = 0x12,
+	SCMI_PROTOCOL_PERF = 0x13,
+	SCMI_PROTOCOL_CLOCK = 0x14,
+	SCMI_PROTOCOL_SENSOR = 0x15,
 };
-- 
2.7.4

[PATCH v4 05/20] firmware: arm_scmi: add scmi protocol bus to enumerate protocol devices

[Sudeep Holla]

The SCMI specification encompasses various protocols. However, not every
protocol has to be present on a given platform/implementation as not
every protocol is relevant for it.

Furthermore, the platform chooses which protocols it exposes to a given
agent. The only protocol that must be implemented is the base protocol.
The base protocol is used by an agent to discover which protocols are
available to it.

In order to enumerate the discovered implemented protocols, this patch
adds support for a separate scmi protocol bus. It also adds mechanism to
register support for different protocols.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/Makefile |   3 +-
 drivers/firmware/arm_scmi/bus.c    | 232 +++++++++++++++++++++++++++++++++++++
 drivers/firmware/arm_scmi/common.h |   1 +
 include/linux/scmi_protocol.h      |  62 ++++++++++
 4 files changed, 297 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/bus.c

diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
index 5d9c7ef35f0f..5f4ec2613db6 100644
--- a/drivers/firmware/arm_scmi/Makefile
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -1,3 +1,4 @@
-obj-y	= scmi-driver.o scmi-protocols.o
+obj-y	= scmi-bus.o scmi-driver.o scmi-protocols.o
+scmi-bus-y = bus.o
 scmi-driver-y = driver.o
 scmi-protocols-y = base.o
diff --git a/drivers/firmware/arm_scmi/bus.c b/drivers/firmware/arm_scmi/bus.c
new file mode 100644
index 000000000000..40c72a35fc99
--- /dev/null
+++ b/drivers/firmware/arm_scmi/bus.c
@@ -0,0 +1,232 @@
+/*
+ * System Control and Management Interface (SCMI) Message Protocol bus layer
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+
+#include "common.h"
+
+static DEFINE_IDA(scmi_bus_id);
+static DEFINE_IDR(scmi_protocols);
+static DEFINE_SPINLOCK(protocol_lock);
+
+static const struct scmi_device_id *
+scmi_dev_match_id(struct scmi_device *scmi_dev, struct scmi_driver *scmi_drv)
+{
+	const struct scmi_device_id *id = scmi_drv->id_table;
+
+	if (!id)
+		return NULL;
+
+	for (; id->protocol_id; id++)
+		if (id->protocol_id == scmi_dev->protocol_id)
+			return id;
+
+	return NULL;
+}
+
+static int scmi_dev_match(struct device *dev, struct device_driver *drv)
+{
+	struct scmi_driver *scmi_drv = to_scmi_driver(drv);
+	struct scmi_device *scmi_dev = to_scmi_dev(dev);
+	const struct scmi_device_id *id;
+
+	id = scmi_dev_match_id(scmi_dev, scmi_drv);
+	if (id)
+		return 1;
+
+	return 0;
+}
+
+static int scmi_protocol_init(int protocol_id, struct scmi_handle *handle)
+{
+	scmi_prot_init_fn_t fn = idr_find(&scmi_protocols, protocol_id);
+
+	if (unlikely(!fn))
+		return -EINVAL;
+	return fn(handle);
+}
+
+static int scmi_dev_probe(struct device *dev)
+{
+	struct scmi_driver *scmi_drv = to_scmi_driver(dev->driver);
+	struct scmi_device *scmi_dev = to_scmi_dev(dev);
+	const struct scmi_device_id *id;
+	int ret;
+
+	id = scmi_dev_match_id(scmi_dev, scmi_drv);
+	if (!id)
+		return -ENODEV;
+
+	if (!scmi_dev->handle)
+		return -EPROBE_DEFER;
+
+	ret = scmi_protocol_init(scmi_dev->protocol_id, scmi_dev->handle);
+	if (ret)
+		return ret;
+
+	return scmi_drv->probe(scmi_dev);
+}
+
+static int scmi_dev_remove(struct device *dev)
+{
+	struct scmi_driver *scmi_drv = to_scmi_driver(dev->driver);
+	struct scmi_device *scmi_dev = to_scmi_dev(dev);
+
+	if (scmi_drv->remove)
+		scmi_drv->remove(scmi_dev);
+
+	return 0;
+}
+
+static struct bus_type scmi_bus_type = {
+	.name =	"scmi_protocol",
+	.match = scmi_dev_match,
+	.probe = scmi_dev_probe,
+	.remove = scmi_dev_remove,
+};
+
+int scmi_driver_register(struct scmi_driver *driver, struct module *owner,
+			 const char *mod_name)
+{
+	int retval;
+
+	driver->driver.bus = &scmi_bus_type;
+	driver->driver.name = driver->name;
+	driver->driver.owner = owner;
+	driver->driver.mod_name = mod_name;
+
+	retval = driver_register(&driver->driver);
+	if (!retval)
+		pr_debug("registered new scmi driver %s\n", driver->name);
+
+	return retval;
+}
+EXPORT_SYMBOL_GPL(scmi_driver_register);
+
+void scmi_driver_unregister(struct scmi_driver *driver)
+{
+	driver_unregister(&driver->driver);
+}
+EXPORT_SYMBOL_GPL(scmi_driver_unregister);
+
+struct scmi_device *
+scmi_device_create(struct device_node *np, struct device *parent, int protocol)
+{
+	int id, retval;
+	struct scmi_device *scmi_dev;
+
+	id = ida_simple_get(&scmi_bus_id, 1, 0, GFP_KERNEL);
+	if (id < 0)
+		return NULL;
+
+	scmi_dev = kzalloc(sizeof(*scmi_dev), GFP_KERNEL);
+	if (!scmi_dev)
+		goto no_mem;
+
+	scmi_dev->id = id;
+	scmi_dev->protocol_id = protocol;
+	scmi_dev->dev.parent = parent;
+	scmi_dev->dev.of_node = np;
+	scmi_dev->dev.bus = &scmi_bus_type;
+	dev_set_name(&scmi_dev->dev, "scmi_dev.%d", id);
+
+	retval = device_register(&scmi_dev->dev);
+	if (!retval)
+		return scmi_dev;
+
+	put_device(&scmi_dev->dev);
+	kfree(scmi_dev);
+no_mem:
+	ida_simple_remove(&scmi_bus_id, id);
+	return NULL;
+}
+
+void scmi_device_destroy(struct scmi_device *scmi_dev)
+{
+	scmi_handle_put(scmi_dev->handle);
+	device_unregister(&scmi_dev->dev);
+	ida_simple_remove(&scmi_bus_id, scmi_dev->id);
+	kfree(scmi_dev);
+}
+
+void scmi_set_handle(struct scmi_device *scmi_dev)
+{
+	scmi_dev->handle = scmi_handle_get(&scmi_dev->dev);
+}
+
+int scmi_protocol_register(int protocol_id, scmi_prot_init_fn_t fn)
+{
+	int ret;
+
+	spin_lock(&protocol_lock);
+	ret = idr_alloc(&scmi_protocols, fn, protocol_id, protocol_id + 1,
+			GFP_KERNEL);
+	if (ret != protocol_id)
+		pr_err("unable to allocate SCMI idr slot, err %d\n", ret);
+	spin_unlock(&protocol_lock);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(scmi_protocol_register);
+
+void scmi_protocol_unregister(int protocol_id)
+{
+	spin_lock(&protocol_lock);
+	idr_remove(&scmi_protocols, protocol_id);
+	spin_unlock(&protocol_lock);
+}
+EXPORT_SYMBOL_GPL(scmi_protocol_unregister);
+
+static int __scmi_devices_unregister(struct device *dev, void *data)
+{
+	struct scmi_device *scmi_dev = to_scmi_dev(dev);
+
+	scmi_device_destroy(scmi_dev);
+	return 0;
+}
+
+static void scmi_devices_unregister(void)
+{
+	bus_for_each_dev(&scmi_bus_type, NULL, NULL, __scmi_devices_unregister);
+}
+
+static int __init scmi_bus_init(void)
+{
+	int retval;
+
+	retval = bus_register(&scmi_bus_type);
+	if (retval)
+		pr_err("scmi protocol bus register failed (%d)\n", retval);
+
+	return retval;
+}
+subsys_initcall(scmi_bus_init);
+
+void __exit scmi_bus_exit(void)
+{
+	scmi_devices_unregister();
+	bus_unregister(&scmi_bus_type);
+	ida_destroy(&scmi_bus_id);
+}
+module_exit(scmi_bus_exit);
diff --git a/drivers/firmware/arm_scmi/common.h b/drivers/firmware/arm_scmi/common.h
index bc767f4997e0..f1eeacaef57b 100644
--- a/drivers/firmware/arm_scmi/common.h
+++ b/drivers/firmware/arm_scmi/common.h
@@ -107,6 +107,7 @@ int scmi_one_xfer_init(const struct scmi_handle *h, u8 msg_id, u8 prot_id,
 		       size_t tx_size, size_t rx_size, struct scmi_xfer **p);
 int scmi_handle_put(const struct scmi_handle *handle);
 struct scmi_handle *scmi_handle_get(struct device *dev);
+void scmi_set_handle(struct scmi_device *scmi_dev);
 int scmi_version_get(const struct scmi_handle *h, u8 protocol, u32 *version);
 void scmi_setup_protocol_implemented(const struct scmi_handle *handle,
 				     u8 *prot_imp);
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
index 664da3d763f2..7b72c75cb560 100644
--- a/include/linux/scmi_protocol.h
+++ b/include/linux/scmi_protocol.h
@@ -15,6 +15,7 @@
  * You should have received a copy of the GNU General Public License along with
  * this program. If not, see <http://www.gnu.org/licenses/>.
  */
+#include <linux/device.h>
 #include <linux/types.h>
 
 #define SCMI_MAX_STR_SIZE	16
@@ -62,3 +63,64 @@ enum scmi_std_protocol {
 	SCMI_PROTOCOL_CLOCK = 0x14,
 	SCMI_PROTOCOL_SENSOR = 0x15,
 };
+
+struct scmi_device {
+	u32 id;
+	u8 protocol_id;
+	struct device dev;
+	struct scmi_handle *handle;
+};
+
+#define to_scmi_dev(d) container_of(d, struct scmi_device, dev)
+
+struct scmi_device *
+scmi_device_create(struct device_node *np, struct device *parent, int protocol);
+void scmi_device_destroy(struct scmi_device *scmi_dev);
+
+struct scmi_device_id {
+	u8 protocol_id;
+};
+
+struct scmi_driver {
+	const char *name;
+	int (*probe)(struct scmi_device *);
+	void (*remove)(struct scmi_device *);
+	const struct scmi_device_id *id_table;
+
+	struct device_driver driver;
+};
+
+#define to_scmi_driver(d) container_of(d, struct scmi_driver, driver)
+
+#ifdef CONFIG_ARM_SCMI_PROTOCOL
+int scmi_driver_register(struct scmi_driver *driver,
+			 struct module *owner, const char *mod_name);
+void scmi_driver_unregister(struct scmi_driver *driver);
+#else
+static int scmi_driver_register(struct scmi_driver *driver,
+				struct module *owner, const char *mod_name)
+{
+	return -EINVAL;
+}
+static void scmi_driver_unregister(struct scmi_driver *driver) {}
+#endif /* CONFIG_ARM_SCMI_PROTOCOL */
+
+#define scmi_register(driver) \
+	scmi_driver_register(driver, THIS_MODULE, KBUILD_MODNAME)
+#define scmi_unregister(driver) \
+	scmi_driver_unregister(driver)
+
+/**
+ * module_scmi_driver() - Helper macro for registering a scmi driver
+ * @__scmi_driver: scmi_driver structure
+ *
+ * Helper macro for scmi drivers to set up proper module init / exit
+ * functions.  Replaces module_init() and module_exit() and keeps people from
+ * printing pointless things to the kernel log when their driver is loaded.
+ */
+#define module_scmi_driver(__scmi_driver)	\
+	module_driver(__scmi_driver, scmi_register, scmi_unregister)
+
+typedef int (*scmi_prot_init_fn_t)(struct scmi_handle *);
+int scmi_protocol_register(int protocol_id, scmi_prot_init_fn_t fn);
+void scmi_protocol_unregister(int protocol_id);
-- 
2.7.4

[PATCH v4 06/20] firmware: arm_scmi: add initial support for performance protocol

[Sudeep Holla]

The performance protocol is intended for the performance management of
group(s) of device(s) that run in the same performance domain. It
includes even the CPUs. A performance domain is defined by a set of
devices that always have to run at the same performance level.
For example, a set of CPUs that share a voltage domain, and have a
common frequency control, is said to be in the same performance domain.

The commands in this protocol provide functionality to describe the
protocol version, describe various attribute flags, set and get the
performance level of a domain. It also supports discovery of the list
of performance levels supported by a performance domain, and the
properties of each performance level.

This patch adds basic support for the performance protocol.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/Makefile |   2 +-
 drivers/firmware/arm_scmi/common.h |   1 +
 drivers/firmware/arm_scmi/perf.c   | 527 +++++++++++++++++++++++++++++++++++++
 include/linux/scmi_protocol.h      |  34 +++
 4 files changed, 563 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/perf.c

diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
index 5f4ec2613db6..687cbbfb3af6 100644
--- a/drivers/firmware/arm_scmi/Makefile
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -1,4 +1,4 @@
 obj-y	= scmi-bus.o scmi-driver.o scmi-protocols.o
 scmi-bus-y = bus.o
 scmi-driver-y = driver.o
-scmi-protocols-y = base.o
+scmi-protocols-y = base.o perf.o
diff --git a/drivers/firmware/arm_scmi/common.h b/drivers/firmware/arm_scmi/common.h
index f1eeacaef57b..b045a3f268e7 100644
--- a/drivers/firmware/arm_scmi/common.h
+++ b/drivers/firmware/arm_scmi/common.h
@@ -30,6 +30,7 @@
 #define PROTOCOL_REV_MAJOR(x)	((x) >> PROTOCOL_REV_MINOR_BITS)
 #define PROTOCOL_REV_MINOR(x)	((x) & PROTOCOL_REV_MINOR_MASK)
 #define MAX_PROTOCOLS_IMP	16
+#define MAX_OPPS		16
 
 enum scmi_common_cmd {
 	PROTOCOL_VERSION = 0x0,
diff --git a/drivers/firmware/arm_scmi/perf.c b/drivers/firmware/arm_scmi/perf.c
new file mode 100644
index 000000000000..a1f5cf136748
--- /dev/null
+++ b/drivers/firmware/arm_scmi/perf.c
@@ -0,0 +1,527 @@
+/*
+ * System Control and Management Interface (SCMI) Performance Protocol
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_opp.h>
+#include <linux/sort.h>
+
+#include "common.h"
+
+enum scmi_performance_protocol_cmd {
+	PERF_DOMAIN_ATTRIBUTES = 0x3,
+	PERF_DESCRIBE_LEVELS = 0x4,
+	PERF_LIMITS_SET = 0x5,
+	PERF_LIMITS_GET = 0x6,
+	PERF_LEVEL_SET = 0x7,
+	PERF_LEVEL_GET = 0x8,
+	PERF_NOTIFY_LIMITS = 0x9,
+	PERF_NOTIFY_LEVEL = 0xa,
+};
+
+struct scmi_opp {
+	u32 perf;
+	u32 power;
+	u32 trans_latency_us;
+};
+
+struct scmi_msg_resp_perf_attributes {
+	__le16 num_domains;
+	__le16 flags;
+#define POWER_SCALE_IN_MILLIWATT(x)	((x) & BIT(0))
+	__le32 stats_addr_low;
+	__le32 stats_addr_high;
+	__le32 stats_size;
+};
+
+struct scmi_msg_resp_perf_domain_attributes {
+	__le32 flags;
+#define SUPPORTS_SET_LIMITS(x)		((x) & BIT(31))
+#define SUPPORTS_SET_PERF_LVL(x)	((x) & BIT(30))
+#define SUPPORTS_PERF_LIMIT_NOTIFY(x)	((x) & BIT(29))
+#define SUPPORTS_PERF_LEVEL_NOTIFY(x)	((x) & BIT(28))
+	__le32 rate_limit_us;
+	__le32 sustained_freq_khz;
+	__le32 sustained_perf_level;
+	    u8 name[SCMI_MAX_STR_SIZE];
+};
+
+struct scmi_msg_perf_describe_levels {
+	__le32 domain;
+	__le32 level_index;
+};
+
+struct scmi_perf_set_limits {
+	__le32 domain;
+	__le32 max_level;
+	__le32 min_level;
+};
+
+struct scmi_perf_get_limits {
+	__le32 max_level;
+	__le32 min_level;
+};
+
+struct scmi_perf_set_level {
+	__le32 domain;
+	__le32 level;
+};
+
+struct scmi_perf_notify_level_or_limits {
+	__le32 domain;
+	__le32 notify_enable;
+};
+
+struct scmi_msg_resp_perf_describe_levels {
+	__le16 num_returned;
+	__le16 num_remaining;
+	struct {
+		__le32 perf_val;
+		__le32 power;
+		__le16 transition_latency_us;
+		__le16 reserved;
+	} opp[0];
+};
+
+struct perf_dom_info {
+	bool set_limits;
+	bool set_perf;
+	bool perf_limit_notify;
+	bool perf_level_notify;
+	u32 opp_count;
+	u32 sustained_freq_khz;
+	u32 sustained_perf_level;
+	u32 mult_factor;
+	char name[SCMI_MAX_STR_SIZE];
+	struct scmi_opp opp[MAX_OPPS];
+};
+
+struct scmi_perf_info {
+	int num_domains;
+	bool power_scale_mw;
+	u64 stats_addr;
+	u32 stats_size;
+	struct perf_dom_info *dom_info;
+};
+
+static int scmi_perf_attributes_get(const struct scmi_handle *handle,
+				    struct scmi_perf_info *pi)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_perf_attributes *attr;
+
+	ret = scmi_one_xfer_init(handle, PROTOCOL_ATTRIBUTES,
+				 SCMI_PROTOCOL_PERF, 0, sizeof(*attr), &t);
+	if (ret)
+		return ret;
+
+	attr = t->rx.buf;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		u16 flags = le16_to_cpu(attr->flags);
+
+		pi->num_domains = le16_to_cpu(attr->num_domains);
+		pi->power_scale_mw = POWER_SCALE_IN_MILLIWATT(flags);
+		pi->stats_addr = le32_to_cpu(attr->stats_addr_low) |
+				(u64)le32_to_cpu(attr->stats_addr_high) << 32;
+		pi->stats_size = le32_to_cpu(attr->stats_size);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_perf_domain_attributes_get(const struct scmi_handle *handle, u32 domain,
+				struct perf_dom_info *dom_info)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_perf_domain_attributes *attr;
+
+	ret = scmi_one_xfer_init(handle, PERF_DOMAIN_ATTRIBUTES,
+				 SCMI_PROTOCOL_PERF, sizeof(domain),
+				 sizeof(*attr), &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(domain);
+	attr = t->rx.buf;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		u32 flags = le32_to_cpu(attr->flags);
+
+		dom_info->set_limits = SUPPORTS_SET_LIMITS(flags);
+		dom_info->set_perf = SUPPORTS_SET_PERF_LVL(flags);
+		dom_info->perf_limit_notify = SUPPORTS_PERF_LIMIT_NOTIFY(flags);
+		dom_info->perf_level_notify = SUPPORTS_PERF_LEVEL_NOTIFY(flags);
+		dom_info->sustained_freq_khz =
+					le32_to_cpu(attr->sustained_freq_khz);
+		dom_info->sustained_perf_level =
+					le32_to_cpu(attr->sustained_perf_level);
+		dom_info->mult_factor =	(dom_info->sustained_freq_khz * 1000) /
+					dom_info->sustained_perf_level;
+		memcpy(dom_info->name, attr->name, SCMI_MAX_STR_SIZE);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int opp_cmp_func(const void *opp1, const void *opp2)
+{
+	const struct scmi_opp *t1 = opp1, *t2 = opp2;
+
+	return t1->perf - t2->perf;
+}
+
+static int
+scmi_perf_describe_levels_get(const struct scmi_handle *handle, u32 domain,
+			      struct perf_dom_info *perf_dom)
+{
+	int ret, cnt;
+	u32 tot_opp_cnt = 0;
+	u16 num_returned, num_remaining;
+	struct scmi_xfer *t;
+	struct scmi_opp *opp;
+	struct scmi_msg_perf_describe_levels *dom_info;
+	struct scmi_msg_resp_perf_describe_levels *level_info;
+
+	ret = scmi_one_xfer_init(handle, PERF_DESCRIBE_LEVELS,
+				 SCMI_PROTOCOL_PERF, sizeof(*dom_info), 0, &t);
+	if (ret)
+		return ret;
+
+	dom_info = t->tx.buf;
+	level_info = t->rx.buf;
+
+	do {
+		dom_info->domain = cpu_to_le32(domain);
+		/* Set the number of OPPs to be skipped/already read */
+		dom_info->level_index = cpu_to_le32(tot_opp_cnt);
+
+		ret = scmi_do_xfer(handle, t);
+		if (ret)
+			break;
+
+		num_returned = le16_to_cpu(level_info->num_returned);
+		num_remaining = le16_to_cpu(level_info->num_remaining);
+		if (tot_opp_cnt + num_returned > MAX_OPPS) {
+			dev_err(handle->dev, "No. of OPPs exceeded MAX_OPPS");
+			break;
+		}
+
+		opp = &perf_dom->opp[tot_opp_cnt];
+		for (cnt = 0; cnt < num_returned; cnt++, opp++) {
+			opp->perf = le32_to_cpu(level_info->opp[cnt].perf_val);
+			opp->power = le32_to_cpu(level_info->opp[cnt].power);
+			opp->trans_latency_us = le16_to_cpu(
+				level_info->opp[cnt].transition_latency_us);
+
+			dev_dbg(handle->dev, "Level %d Power %d Latency %dus\n",
+				opp->perf, opp->power, opp->trans_latency_us);
+		}
+
+		tot_opp_cnt += num_returned;
+		/*
+		 * check for both returned and remaining to avoid infinite
+		 * loop due to buggy firmware
+		 */
+	} while (num_returned && num_remaining);
+
+	perf_dom->opp_count = tot_opp_cnt;
+	scmi_one_xfer_put(handle, t);
+
+	sort(perf_dom->opp, tot_opp_cnt, sizeof(*opp), opp_cmp_func, NULL);
+	return ret;
+}
+
+static int scmi_perf_limits_set(const struct scmi_handle *handle, u32 domain,
+				u32 max_perf, u32 min_perf)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_perf_set_limits *limits;
+
+	ret = scmi_one_xfer_init(handle, PERF_LIMITS_SET, SCMI_PROTOCOL_PERF,
+				 sizeof(*limits), 0, &t);
+	if (ret)
+		return ret;
+
+	limits = t->tx.buf;
+	limits->domain = cpu_to_le32(domain);
+	limits->max_level = cpu_to_le32(max_perf);
+	limits->min_level = cpu_to_le32(min_perf);
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_perf_limits_get(const struct scmi_handle *handle, u32 domain,
+				u32 *max_perf, u32 *min_perf)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_perf_get_limits *limits;
+
+	ret = scmi_one_xfer_init(handle, PERF_LIMITS_GET, SCMI_PROTOCOL_PERF,
+				 sizeof(__le32), 0, &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(domain);
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		limits = t->rx.buf;
+
+		*max_perf = le32_to_cpu(limits->max_level);
+		*min_perf = le32_to_cpu(limits->min_level);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_perf_level_set(const struct scmi_handle *handle, u32 domain, u32 level)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_perf_set_level *lvl;
+
+	ret = scmi_one_xfer_init(handle, PERF_LEVEL_SET, SCMI_PROTOCOL_PERF,
+				 sizeof(*lvl), 0, &t);
+	if (ret)
+		return ret;
+
+	lvl = t->tx.buf;
+	lvl->domain = cpu_to_le32(domain);
+	lvl->level = cpu_to_le32(level);
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_perf_level_get(const struct scmi_handle *handle, u32 domain, u32 *level)
+{
+	int ret;
+	struct scmi_xfer *t;
+
+	ret = scmi_one_xfer_init(handle, PERF_LEVEL_GET, SCMI_PROTOCOL_PERF,
+				 sizeof(u32), sizeof(u32), &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(domain);
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret)
+		*level = le32_to_cpu(*(__le32 *)t->rx.buf);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int __scmi_perf_notify_enable(const struct scmi_handle *handle, u32 cmd,
+				     u32 domain, bool enable)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_perf_notify_level_or_limits *notify;
+
+	ret = scmi_one_xfer_init(handle, cmd, SCMI_PROTOCOL_PERF,
+				 sizeof(*notify), 0, &t);
+	if (ret)
+		return ret;
+
+	notify = t->tx.buf;
+	notify->domain = cpu_to_le32(domain);
+	notify->notify_enable = cpu_to_le32(enable & BIT(0));
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_perf_limits_notify_enable(const struct scmi_handle *handle,
+					  u32 domain, bool enable)
+{
+	return __scmi_perf_notify_enable(handle, PERF_NOTIFY_LIMITS,
+					 domain, enable);
+}
+
+static int scmi_perf_level_notify_enable(const struct scmi_handle *handle,
+					 u32 domain, bool enable)
+{
+	return __scmi_perf_notify_enable(handle, PERF_NOTIFY_LEVEL,
+					 domain, enable);
+}
+
+/* Device specific ops */
+static int scmi_dev_domain_id(struct device *dev)
+{
+	struct of_phandle_args clkspec;
+
+	if (of_parse_phandle_with_args(dev->of_node, "clocks", "#clock-cells",
+				       0, &clkspec))
+		return -EINVAL;
+
+	return clkspec.args[0];
+}
+
+static int scmi_dvfs_add_opps_to_device(const struct scmi_handle *handle,
+					struct device *dev)
+{
+	int idx, ret, domain;
+	unsigned long freq;
+	struct scmi_opp *opp;
+	struct perf_dom_info *dom;
+	struct scmi_perf_info *pi = handle->perf_priv;
+
+	domain = scmi_dev_domain_id(dev);
+	if (domain < 0)
+		return domain;
+
+	dom = pi->dom_info + domain;
+	if (!dom)
+		return -EIO;
+
+	for (opp = dom->opp, idx = 0; idx < dom->opp_count; idx++, opp++) {
+		freq = opp->perf * dom->mult_factor;
+
+		ret = dev_pm_opp_add(dev, freq, 0);
+		if (ret) {
+			dev_warn(dev, "failed to add opp %luHz\n", freq);
+
+			while (idx-- > 0) {
+				freq = (--opp)->perf * dom->mult_factor;
+				dev_pm_opp_remove(dev, freq);
+			}
+			return ret;
+		}
+	}
+	return 0;
+}
+
+static int scmi_dvfs_get_transition_latency(const struct scmi_handle *handle,
+					    struct device *dev)
+{
+	struct perf_dom_info *dom;
+	struct scmi_perf_info *pi = handle->perf_priv;
+	int domain = scmi_dev_domain_id(dev);
+
+	if (domain < 0)
+		return domain;
+
+	dom = pi->dom_info + domain;
+	if (!dom)
+		return -EIO;
+
+	/* uS to nS */
+	return dom->opp[dom->opp_count - 1].trans_latency_us * 1000;
+}
+
+static int scmi_dvfs_freq_set(const struct scmi_handle *handle, u32 domain,
+			      unsigned long freq)
+{
+	struct scmi_perf_info *pi = handle->perf_priv;
+	struct perf_dom_info *dom = pi->dom_info + domain;
+
+	return scmi_perf_level_set(handle, domain, freq / dom->mult_factor);
+}
+
+static int scmi_dvfs_freq_get(const struct scmi_handle *handle, u32 domain,
+			      unsigned long *freq)
+{
+	int ret;
+	u32 level;
+	struct scmi_perf_info *pi = handle->perf_priv;
+	struct perf_dom_info *dom = pi->dom_info + domain;
+
+	ret = scmi_perf_level_get(handle, domain, &level);
+	if (!ret)
+		*freq = level * dom->mult_factor;
+
+	return ret;
+}
+
+static struct scmi_perf_ops perf_ops = {
+	.limits_set = scmi_perf_limits_set,
+	.limits_get = scmi_perf_limits_get,
+	.level_set = scmi_perf_level_set,
+	.level_get = scmi_perf_level_get,
+	.limits_notify_enable = scmi_perf_limits_notify_enable,
+	.level_notify_enable = scmi_perf_level_notify_enable,
+	.device_domain_id = scmi_dev_domain_id,
+	.get_transition_latency = scmi_dvfs_get_transition_latency,
+	.add_opps_to_device = scmi_dvfs_add_opps_to_device,
+	.freq_set = scmi_dvfs_freq_set,
+	.freq_get = scmi_dvfs_freq_get,
+};
+
+static int scmi_perf_protocol_init(struct scmi_handle *handle)
+{
+	int domain;
+	u32 version;
+	struct scmi_perf_info *pinfo;
+
+	scmi_version_get(handle, SCMI_PROTOCOL_PERF, &version);
+
+	dev_dbg(handle->dev, "Performance Version %d.%d\n",
+		PROTOCOL_REV_MAJOR(version), PROTOCOL_REV_MINOR(version));
+
+	pinfo = devm_kzalloc(handle->dev, sizeof(*pinfo), GFP_KERNEL);
+	if (!pinfo)
+		return -ENOMEM;
+
+	scmi_perf_attributes_get(handle, pinfo);
+
+	pinfo->dom_info = devm_kcalloc(handle->dev, pinfo->num_domains,
+				       sizeof(*pinfo->dom_info), GFP_KERNEL);
+	if (!pinfo->dom_info)
+		return -ENOMEM;
+
+	for (domain = 0; domain < pinfo->num_domains; domain++) {
+		struct perf_dom_info *dom = pinfo->dom_info + domain;
+
+		scmi_perf_domain_attributes_get(handle, domain, dom);
+		scmi_perf_describe_levels_get(handle, domain, dom);
+	}
+
+	handle->perf_ops = &perf_ops;
+	handle->perf_priv = pinfo;
+
+	return 0;
+}
+
+static int __init scmi_perf_init(void)
+{
+	return scmi_protocol_register(SCMI_PROTOCOL_PERF,
+				      &scmi_perf_protocol_init);
+}
+subsys_initcall(scmi_perf_init);
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
index 7b72c75cb560..8fa189ca0fed 100644
--- a/include/linux/scmi_protocol.h
+++ b/include/linux/scmi_protocol.h
@@ -44,15 +44,49 @@ struct scmi_revision_info {
 	char sub_vendor_id[SCMI_MAX_STR_SIZE];
 };
 
+struct scmi_handle;
+
+/**
+ * struct scmi_perf_ops - represents the various operations provided
+ *	by SCMI Performance Protocol
+ *
+ * @limits_set: sets limits on the performance level of a domain
+ * @limits_get: gets limits on the performance level of a domain
+ * @level_set: sets the performance level of a domain
+ * @level_get: gets the performance level of a domain
+ * @limits_notify_enable: requests notifications from the platform for changes
+ *	in the allowed maximum and minimum performance levels
+ * @level_notify_enable: requests notifications from the platform when the
+ *	performance level for a domain changes in value
+ */
+struct scmi_perf_ops {
+	int (*limits_set)(const struct scmi_handle *, u32, u32, u32);
+	int (*limits_get)(const struct scmi_handle *, u32, u32 *, u32 *);
+	int (*level_set)(const struct scmi_handle *, u32, u32);
+	int (*level_get)(const struct scmi_handle *, u32, u32 *);
+	int (*limits_notify_enable)(const struct scmi_handle *, u32, bool);
+	int (*level_notify_enable)(const struct scmi_handle *, u32, bool);
+	int (*device_domain_id)(struct device *);
+	int (*get_transition_latency)(const struct scmi_handle *,
+				      struct device *);
+	int (*add_opps_to_device)(const struct scmi_handle *, struct device *);
+	int (*freq_set)(const struct scmi_handle *, u32, unsigned long);
+	int (*freq_get)(const struct scmi_handle *, u32, unsigned long *);
+};
+
 /**
  * struct scmi_handle - Handle returned to ARM SCMI clients for usage.
  *
  * @dev: pointer to the SCMI device
  * @version: pointer to the structure containing SCMI version information
+ * @perf_ops: pointer to set of performance protocol operations
  */
 struct scmi_handle {
 	struct device *dev;
 	struct scmi_revision_info *version;
+	struct scmi_perf_ops *perf_ops;
+	/* for protocol internal use */
+	void *perf_priv;
 };
 
 enum scmi_std_protocol {
-- 
2.7.4

[PATCH v4 07/20] firmware: arm_scmi: add initial support for clock protocol

[Sudeep Holla]

The clock protocol is intended for management of clocks. It is used to
enable or disable clocks, and to set and get the clock rates. This
protocol provides commands to describe the protocol version, discover
various implementation specific attributes, describe a clock, enable
and disable a clock and get/set the rate of the clock synchronously or
asynchronously.

This patch adds initial support for the clock protocol.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/Makefile |   2 +-
 drivers/firmware/arm_scmi/clock.c  | 353 +++++++++++++++++++++++++++++++++++++
 include/linux/scmi_protocol.h      |  41 +++++
 3 files changed, 395 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/clock.c

diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
index 687cbbfb3af6..2130ee9ac825 100644
--- a/drivers/firmware/arm_scmi/Makefile
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -1,4 +1,4 @@
 obj-y	= scmi-bus.o scmi-driver.o scmi-protocols.o
 scmi-bus-y = bus.o
 scmi-driver-y = driver.o
-scmi-protocols-y = base.o perf.o
+scmi-protocols-y = base.o clock.o perf.o
diff --git a/drivers/firmware/arm_scmi/clock.c b/drivers/firmware/arm_scmi/clock.c
new file mode 100644
index 000000000000..0fac1d4b377b
--- /dev/null
+++ b/drivers/firmware/arm_scmi/clock.c
@@ -0,0 +1,353 @@
+/*
+ * System Control and Management Interface (SCMI) Clock Protocol
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "common.h"
+
+enum scmi_clock_protocol_cmd {
+	CLOCK_ATTRIBUTES = 0x3,
+	CLOCK_DESCRIBE_RATES = 0x4,
+	CLOCK_RATE_SET = 0x5,
+	CLOCK_RATE_GET = 0x6,
+	CLOCK_CONFIG_SET = 0x7,
+};
+
+struct scmi_msg_resp_clock_protocol_attributes {
+	__le16 num_clocks;
+	    u8 max_async_req;
+	    u8 reserved;
+};
+
+struct scmi_msg_resp_clock_attributes {
+	__le32 attributes;
+#define	CLOCK_ENABLE	BIT(0)
+	    u8 name[SCMI_MAX_STR_SIZE];
+};
+
+struct scmi_clock_set_config {
+	__le32 id;
+	__le32 attributes;
+};
+
+struct scmi_msg_clock_describe_rates {
+	__le32 id;
+	__le32 rate_index;
+};
+
+struct scmi_msg_resp_clock_describe_rates {
+	__le32 num_rates_flags;
+#define NUM_RETURNED(x)		((x) & 0xfff)
+#define RATE_DISCRETE(x)	!((x) & BIT(12))
+#define NUM_REMAINING(x)	((x) >> 16)
+	struct {
+		__le32 value_low;
+		__le32 value_high;
+	} rate[0];
+#define RATE_TO_U64(X)		\
+({				\
+	typeof(X) x = (X);	\
+	le32_to_cpu((x).value_low) | (u64)le32_to_cpu((x).value_high) << 32; \
+})
+};
+
+struct scmi_clock_set_rate {
+	__le32 flags;
+#define CLOCK_SET_ASYNC		BIT(0)
+#define CLOCK_SET_DELAYED	BIT(1)
+#define CLOCK_SET_ROUND_UP	BIT(2)
+#define CLOCK_SET_ROUND_AUTO	BIT(3)
+	__le32 id;
+	__le32 value_low;
+	__le32 value_high;
+};
+
+struct clock_info {
+	int num_clocks;
+	int max_async_req;
+	struct scmi_clock_info *clk;
+};
+
+static int scmi_clock_protocol_attributes_get(const struct scmi_handle *handle,
+					      struct clock_info *ci)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_clock_protocol_attributes *attr;
+
+	ret = scmi_one_xfer_init(handle, PROTOCOL_ATTRIBUTES,
+				 SCMI_PROTOCOL_CLOCK, 0, sizeof(*attr), &t);
+	if (ret)
+		return ret;
+
+	attr = t->rx.buf;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		ci->num_clocks = le16_to_cpu(attr->num_clocks);
+		ci->max_async_req = attr->max_async_req;
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_clock_attributes_get(const struct scmi_handle *handle,
+				     u32 clk_id, struct scmi_clock_info *clk)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_clock_attributes *attr;
+
+	ret = scmi_one_xfer_init(handle, CLOCK_ATTRIBUTES, SCMI_PROTOCOL_CLOCK,
+				 sizeof(clk_id), sizeof(*attr), &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(clk_id);
+	attr = t->rx.buf;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret)
+		memcpy(clk->name, attr->name, SCMI_MAX_STR_SIZE);
+	else
+		clk->name[0] = '\0';
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_clock_describe_rates_get(const struct scmi_handle *handle, u32 clk_id,
+			      struct scmi_clock_info *clk)
+{
+	u64 *rate;
+	int ret, cnt;
+	bool rate_discrete;
+	u32 tot_rate_cnt = 0, rates_flag;
+	u16 num_returned, num_remaining;
+	struct scmi_xfer *t;
+	struct scmi_msg_clock_describe_rates *clk_desc;
+	struct scmi_msg_resp_clock_describe_rates *rlist;
+
+	ret = scmi_one_xfer_init(handle, CLOCK_DESCRIBE_RATES,
+				 SCMI_PROTOCOL_CLOCK, sizeof(*clk_desc), 0, &t);
+	if (ret)
+		return ret;
+
+	clk_desc = t->tx.buf;
+	rlist = t->rx.buf;
+
+	do {
+		clk_desc->id = cpu_to_le32(clk_id);
+		/* Set the number of rates to be skipped/already read */
+		clk_desc->rate_index = cpu_to_le32(tot_rate_cnt);
+
+		ret = scmi_do_xfer(handle, t);
+		if (ret)
+			break;
+
+		rates_flag = le32_to_cpu(rlist->num_rates_flags);
+		num_remaining = NUM_REMAINING(rates_flag);
+		rate_discrete = RATE_DISCRETE(rates_flag);
+		num_returned = NUM_RETURNED(rates_flag);
+
+		if (tot_rate_cnt + num_returned > SCMI_MAX_NUM_RATES) {
+			dev_err(handle->dev, "No. of rates > MAX_NUM_RATES");
+			break;
+		}
+
+		if (!rate_discrete) {
+			clk->range.min_rate = RATE_TO_U64(rlist->rate[0]);
+			clk->range.max_rate = RATE_TO_U64(rlist->rate[1]);
+			clk->range.step_size = RATE_TO_U64(rlist->rate[2]);
+			dev_dbg(handle->dev, "Min %llu Max %llu Step %llu Hz\n",
+				clk->range.min_rate, clk->range.max_rate,
+				clk->range.step_size);
+			break;
+		}
+
+		rate = &clk->list.rates[tot_rate_cnt];
+		for (cnt = 0; cnt < num_returned; cnt++, rate++) {
+			*rate = RATE_TO_U64(rlist->rate[cnt]);
+			dev_dbg(handle->dev, "Rate %llu Hz\n", *rate);
+		}
+
+		tot_rate_cnt += num_returned;
+		/*
+		 * check for both returned and remaining to avoid infinite
+		 * loop due to buggy firmware
+		 */
+	} while (num_returned && num_remaining);
+
+	if (rate_discrete)
+		clk->list.num_rates = tot_rate_cnt;
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_clock_rate_get(const struct scmi_handle *handle, u32 clk_id, u64 *value)
+{
+	int ret;
+	struct scmi_xfer *t;
+
+	ret = scmi_one_xfer_init(handle, CLOCK_RATE_GET, SCMI_PROTOCOL_CLOCK,
+				 sizeof(__le32), sizeof(u64), &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(clk_id);
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		__le32 *pval = t->rx.buf;
+
+		*value = le32_to_cpu(*pval);
+		*value |= (u64)le32_to_cpu(*(pval + 1)) << 32;
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_clock_rate_set(const struct scmi_handle *handle, u32 clk_id,
+			       u32 config, u64 rate)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_clock_set_rate *cfg;
+
+	ret = scmi_one_xfer_init(handle, CLOCK_RATE_SET, SCMI_PROTOCOL_CLOCK,
+				 sizeof(*cfg), 0, &t);
+	if (ret)
+		return ret;
+
+	cfg = t->tx.buf;
+	cfg->flags = cpu_to_le32(config);
+	cfg->id = cpu_to_le32(clk_id);
+	cfg->value_low = cpu_to_le32(rate & 0xffffffff);
+	cfg->value_high = cpu_to_le32(rate >> 32);
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_clock_config_set(const struct scmi_handle *handle, u32 clk_id, u32 config)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_clock_set_config *cfg;
+
+	ret = scmi_one_xfer_init(handle, CLOCK_CONFIG_SET, SCMI_PROTOCOL_CLOCK,
+				 sizeof(*cfg), 0, &t);
+	if (ret)
+		return ret;
+
+	cfg = t->tx.buf;
+	cfg->id = cpu_to_le32(clk_id);
+	cfg->attributes = cpu_to_le32(config);
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_clock_enable(const struct scmi_handle *handle, u32 clk_id)
+{
+	return scmi_clock_config_set(handle, clk_id, CLOCK_ENABLE);
+}
+
+static int scmi_clock_disable(const struct scmi_handle *handle, u32 clk_id)
+{
+	return scmi_clock_config_set(handle, clk_id, 0);
+}
+
+static int scmi_clock_count_get(const struct scmi_handle *handle)
+{
+	struct clock_info *ci = handle->clk_priv;
+
+	return ci->num_clocks;
+}
+
+static const struct scmi_clock_info *
+scmi_clock_info_get(const struct scmi_handle *handle, u32 clk_id)
+{
+	struct clock_info *ci = handle->clk_priv;
+	struct scmi_clock_info *clk = ci->clk + clk_id;
+
+	if (!clk->name || !clk->name[0])
+		return NULL;
+
+	return clk;
+}
+
+static struct scmi_clk_ops clk_ops = {
+	.count_get = scmi_clock_count_get,
+	.info_get = scmi_clock_info_get,
+	.rate_get = scmi_clock_rate_get,
+	.rate_set = scmi_clock_rate_set,
+	.enable = scmi_clock_enable,
+	.disable = scmi_clock_disable,
+};
+
+static int scmi_clock_protocol_init(struct scmi_handle *handle)
+{
+	u32 version;
+	int clkid, ret;
+	struct clock_info *cinfo;
+
+	scmi_version_get(handle, SCMI_PROTOCOL_CLOCK, &version);
+
+	dev_dbg(handle->dev, "Clock Version %d.%d\n",
+		PROTOCOL_REV_MAJOR(version), PROTOCOL_REV_MINOR(version));
+
+	cinfo = devm_kzalloc(handle->dev, sizeof(*cinfo), GFP_KERNEL);
+	if (!cinfo)
+		return -ENOMEM;
+
+	scmi_clock_protocol_attributes_get(handle, cinfo);
+
+	cinfo->clk = devm_kcalloc(handle->dev, cinfo->num_clocks,
+				  sizeof(*cinfo->clk), GFP_KERNEL);
+	if (!cinfo->clk)
+		return -ENOMEM;
+
+	for (clkid = 0; clkid < cinfo->num_clocks; clkid++) {
+		struct scmi_clock_info *clk = cinfo->clk + clkid;
+
+		ret = scmi_clock_attributes_get(handle, clkid, clk);
+		if (!ret)
+			scmi_clock_describe_rates_get(handle, clkid, clk);
+	}
+
+	handle->clk_ops = &clk_ops;
+	handle->clk_priv = cinfo;
+
+	return 0;
+}
+
+static int __init scmi_clock_init(void)
+{
+	return scmi_protocol_register(SCMI_PROTOCOL_CLOCK,
+				      &scmi_clock_protocol_init);
+}
+subsys_initcall(scmi_clock_init);
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
index 8fa189ca0fed..0c2fd3ab464c 100644
--- a/include/linux/scmi_protocol.h
+++ b/include/linux/scmi_protocol.h
@@ -19,6 +19,7 @@
 #include <linux/types.h>
 
 #define SCMI_MAX_STR_SIZE	16
+#define SCMI_MAX_NUM_RATES	16
 
 /**
  * struct scmi_revision_info - version information structure
@@ -44,9 +45,46 @@ struct scmi_revision_info {
 	char sub_vendor_id[SCMI_MAX_STR_SIZE];
 };
 
+struct scmi_clock_info {
+	char name[SCMI_MAX_STR_SIZE];
+	bool rate_discrete;
+	union {
+		struct {
+			int num_rates;
+			u64 rates[SCMI_MAX_NUM_RATES];
+		} list;
+		struct {
+			u64 min_rate;
+			u64 max_rate;
+			u64 step_size;
+		} range;
+	};
+};
+
 struct scmi_handle;
 
 /**
+ * struct scmi_clk_ops - represents the various operations provided
+ *	by SCMI Clock Protocol
+ *
+ * @count_get: get the count of clocks provided by SCMI
+ * @info_get: get the information of the specified clock
+ * @rate_get: request the current clock rate of a clock
+ * @rate_set: set the clock rate of a clock
+ * @enable: enables the specified clock
+ * @disable: disables the specified clock
+ */
+struct scmi_clk_ops {
+	int (*count_get)(const struct scmi_handle *);
+	const struct scmi_clock_info *(*info_get)(const struct scmi_handle *,
+						  u32);
+	int (*rate_get)(const struct scmi_handle *, u32, u64*);
+	int (*rate_set)(const struct scmi_handle *, u32, u32, u64);
+	int (*enable)(const struct scmi_handle *, u32);
+	int (*disable)(const struct scmi_handle *, u32);
+};
+
+/**
  * struct scmi_perf_ops - represents the various operations provided
  *	by SCMI Performance Protocol
  *
@@ -80,13 +118,16 @@ struct scmi_perf_ops {
  * @dev: pointer to the SCMI device
  * @version: pointer to the structure containing SCMI version information
  * @perf_ops: pointer to set of performance protocol operations
+ * @clk_ops: pointer to set of clock protocol operations
  */
 struct scmi_handle {
 	struct device *dev;
 	struct scmi_revision_info *version;
 	struct scmi_perf_ops *perf_ops;
+	struct scmi_clk_ops *clk_ops;
 	/* for protocol internal use */
 	void *perf_priv;
+	void *clk_priv;
 };
 
 enum scmi_std_protocol {
-- 
2.7.4

[PATCH v4 08/20] firmware: arm_scmi: add initial support for power protocol

[Sudeep Holla]

The power protocol is intended for management of power states of various
power domains. The power domain management protocol provides commands to
describe the protocol version, discover the implementation specific
attributes, set and get the power state of a domain.

This patch adds support for the above mention features of the protocol.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
--
 drivers/firmware/arm_scmi/Makefile |   2 +-
 drivers/firmware/arm_scmi/power.c  | 242 +++++++++++++++++++++++++++++++++++++
 include/linux/scmi_protocol.h      |  28 +++++
 3 files changed, 271 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/power.c
---
 drivers/firmware/arm_scmi/Makefile |   2 +-
 drivers/firmware/arm_scmi/power.c  | 255 +++++++++++++++++++++++++++++++++++++
 include/linux/scmi_protocol.h      |  29 +++++
 3 files changed, 285 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/power.c

diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
index 2130ee9ac825..420c761ced94 100644
--- a/drivers/firmware/arm_scmi/Makefile
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -1,4 +1,4 @@
 obj-y	= scmi-bus.o scmi-driver.o scmi-protocols.o
 scmi-bus-y = bus.o
 scmi-driver-y = driver.o
-scmi-protocols-y = base.o clock.o perf.o
+scmi-protocols-y = base.o clock.o perf.o power.o
diff --git a/drivers/firmware/arm_scmi/power.c b/drivers/firmware/arm_scmi/power.c
new file mode 100644
index 000000000000..50195bd6bbd2
--- /dev/null
+++ b/drivers/firmware/arm_scmi/power.c
@@ -0,0 +1,255 @@
+/*
+ * System Control and Management Interface (SCMI) Power Protocol
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "common.h"
+
+enum scmi_power_protocol_cmd {
+	POWER_DOMAIN_ATTRIBUTES = 0x3,
+	POWER_STATE_SET = 0x4,
+	POWER_STATE_GET = 0x5,
+	POWER_STATE_NOTIFY = 0x6,
+};
+
+struct scmi_msg_resp_power_attributes {
+	__le16 num_domains;
+	__le16 reserved;
+	__le32 stats_addr_low;
+	__le32 stats_addr_high;
+	__le32 stats_size;
+};
+
+struct scmi_msg_resp_power_domain_attributes {
+	__le32 flags;
+#define SUPPORTS_STATE_SET_NOTIFY(x)	((x) & BIT(31))
+#define SUPPORTS_STATE_SET_ASYNC(x)	((x) & BIT(30))
+#define SUPPORTS_STATE_SET_SYNC(x)	((x) & BIT(29))
+	    u8 name[SCMI_MAX_STR_SIZE];
+};
+
+struct scmi_power_set_state {
+	__le32 flags;
+#define STATE_SET_ASYNC		BIT(0)
+	__le32 domain;
+	__le32 state;
+};
+
+struct scmi_power_state_notify {
+	__le32 domain;
+	__le32 notify_enable;
+};
+
+struct power_dom_info {
+	bool state_set_sync;
+	bool state_set_async;
+	bool state_set_notify;
+	char name[SCMI_MAX_STR_SIZE];
+};
+
+struct scmi_power_info {
+	int num_domains;
+	u64 stats_addr;
+	u32 stats_size;
+	struct power_dom_info *dom_info;
+};
+
+static int scmi_power_attributes_get(const struct scmi_handle *handle,
+				     struct scmi_power_info *pi)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_power_attributes *attr;
+
+	ret = scmi_one_xfer_init(handle, PROTOCOL_ATTRIBUTES,
+				 SCMI_PROTOCOL_POWER, 0, sizeof(*attr), &t);
+	if (ret)
+		return ret;
+
+	attr = t->rx.buf;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		pi->num_domains = le16_to_cpu(attr->num_domains);
+		pi->stats_addr = le32_to_cpu(attr->stats_addr_low) |
+				(u64)le32_to_cpu(attr->stats_addr_high) << 32;
+		pi->stats_size = le32_to_cpu(attr->stats_size);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_power_domain_attributes_get(const struct scmi_handle *handle, u32 domain,
+				 struct power_dom_info *dom_info)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_power_domain_attributes *attr;
+
+	ret = scmi_one_xfer_init(handle, POWER_DOMAIN_ATTRIBUTES,
+				 SCMI_PROTOCOL_POWER, sizeof(domain),
+				 sizeof(*attr), &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(domain);
+	attr = t->rx.buf;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		u32 flags = le32_to_cpu(attr->flags);
+
+		dom_info->state_set_notify = SUPPORTS_STATE_SET_NOTIFY(flags);
+		dom_info->state_set_async = SUPPORTS_STATE_SET_ASYNC(flags);
+		dom_info->state_set_sync = SUPPORTS_STATE_SET_SYNC(flags);
+		memcpy(dom_info->name, attr->name, SCMI_MAX_STR_SIZE);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_power_state_set(const struct scmi_handle *handle, u32 domain, u32 state)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_power_set_state *st;
+
+	ret = scmi_one_xfer_init(handle, POWER_STATE_SET, SCMI_PROTOCOL_POWER,
+				 sizeof(*st), 0, &t);
+	if (ret)
+		return ret;
+
+	st = t->tx.buf;
+	st->flags = cpu_to_le32(0);
+	st->domain = cpu_to_le32(domain);
+	st->state = cpu_to_le32(state);
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_power_state_get(const struct scmi_handle *handle, u32 domain, u32 *state)
+{
+	int ret;
+	struct scmi_xfer *t;
+
+	ret = scmi_one_xfer_init(handle, POWER_STATE_GET, SCMI_PROTOCOL_POWER,
+				 sizeof(u32), sizeof(u32), &t);
+	if (ret)
+		return ret;
+
+	*(__le32 *)t->tx.buf = cpu_to_le32(domain);
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret)
+		*state = le32_to_cpu(*(__le32 *)t->rx.buf);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_power_state_notify_enable(const struct scmi_handle *handle,
+					  u32 domain, bool enable)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_power_state_notify *notify;
+
+	ret = scmi_one_xfer_init(handle, POWER_STATE_NOTIFY,
+				 SCMI_PROTOCOL_POWER, sizeof(*notify), 0, &t);
+	if (ret)
+		return ret;
+
+	notify = t->tx.buf;
+	notify->domain = cpu_to_le32(domain);
+	notify->notify_enable = cpu_to_le32(enable & BIT(0));
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_power_num_domains_get(const struct scmi_handle *handle)
+{
+	struct scmi_power_info *pi = handle->power_priv;
+
+	return pi->num_domains;
+}
+
+static char *scmi_power_name_get(const struct scmi_handle *handle, u32 domain)
+{
+	struct scmi_power_info *pi = handle->power_priv;
+	struct power_dom_info *dom = pi->dom_info + domain;
+
+	return dom->name;
+}
+
+static struct scmi_power_ops power_ops = {
+	.num_domains_get = scmi_power_num_domains_get,
+	.name_get = scmi_power_name_get,
+	.state_set = scmi_power_state_set,
+	.state_get = scmi_power_state_get,
+	.state_notify_enable = scmi_power_state_notify_enable,
+};
+
+static int scmi_power_protocol_init(struct scmi_handle *handle)
+{
+	int domain;
+	u32 version;
+	struct scmi_power_info *pinfo;
+
+	scmi_version_get(handle, SCMI_PROTOCOL_POWER, &version);
+
+	dev_dbg(handle->dev, "Power Version %d.%d\n",
+		PROTOCOL_REV_MAJOR(version), PROTOCOL_REV_MINOR(version));
+
+	pinfo = devm_kzalloc(handle->dev, sizeof(*pinfo), GFP_KERNEL);
+	if (!pinfo)
+		return -ENOMEM;
+
+	scmi_power_attributes_get(handle, pinfo);
+
+	pinfo->dom_info = devm_kcalloc(handle->dev, pinfo->num_domains,
+				       sizeof(*pinfo->dom_info), GFP_KERNEL);
+	if (!pinfo->dom_info)
+		return -ENOMEM;
+
+	for (domain = 0; domain < pinfo->num_domains; domain++) {
+		struct power_dom_info *dom = pinfo->dom_info + domain;
+
+		scmi_power_domain_attributes_get(handle, domain, dom);
+	}
+
+	handle->power_ops = &power_ops;
+	handle->power_priv = pinfo;
+
+	return 0;
+}
+
+static int __init scmi_power_init(void)
+{
+	return scmi_protocol_register(SCMI_PROTOCOL_POWER,
+				      &scmi_power_protocol_init);
+}
+subsys_initcall(scmi_power_init);
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
index 0c2fd3ab464c..95a645f9039d 100644
--- a/include/linux/scmi_protocol.h
+++ b/include/linux/scmi_protocol.h
@@ -113,10 +113,37 @@ struct scmi_perf_ops {
 };
 
 /**
+ * struct scmi_power_ops - represents the various operations provided
+ *	by SCMI Power Protocol
+ *
+ * @num_domains_get: get the count of power domains provided by SCMI
+ * @name_get: gets the name of a power domain
+ * @state_set: sets the power state of a power domain
+ * @state_get: gets the power state of a power domain
+ * @state_notify_enable: request notifications from the platform for
+ *	state changes in a specific power domain
+ */
+struct scmi_power_ops {
+	int (*num_domains_get)(const struct scmi_handle *);
+	char *(*name_get)(const struct scmi_handle *, u32);
+#define SCMI_POWER_STATE_TYPE_SHIFT	30
+#define SCMI_POWER_STATE_ID_MASK	(BIT(28) - 1)
+#define SCMI_POWER_STATE_PARAM(type, id) \
+	((((type) & BIT(0)) << SCMI_POWER_STATE_TYPE_SHIFT) | \
+		((id) & SCMI_POWER_STATE_ID_MASK))
+#define SCMI_POWER_STATE_GENERIC_ON	SCMI_POWER_STATE_PARAM(0, 0)
+#define SCMI_POWER_STATE_GENERIC_OFF	SCMI_POWER_STATE_PARAM(1, 0)
+	int (*state_set)(const struct scmi_handle *, u32, u32);
+	int (*state_get)(const struct scmi_handle *, u32, u32 *);
+	int (*state_notify_enable)(const struct scmi_handle *, u32, bool);
+};
+
+/**
  * struct scmi_handle - Handle returned to ARM SCMI clients for usage.
  *
  * @dev: pointer to the SCMI device
  * @version: pointer to the structure containing SCMI version information
+ * @power_ops: pointer to set of power protocol operations
  * @perf_ops: pointer to set of performance protocol operations
  * @clk_ops: pointer to set of clock protocol operations
  */
@@ -125,9 +152,11 @@ struct scmi_handle {
 	struct scmi_revision_info *version;
 	struct scmi_perf_ops *perf_ops;
 	struct scmi_clk_ops *clk_ops;
+	struct scmi_power_ops *power_ops;
 	/* for protocol internal use */
 	void *perf_priv;
 	void *clk_priv;
+	void *power_priv;
 };
 
 enum scmi_std_protocol {
-- 
2.7.4

[PATCH v4 09/20] firmware: arm_scmi: add initial support for sensor protocol

[Sudeep Holla]

The sensor protocol provides functions to manage platform sensors, and
provides the commands to describe the protocol version and the various
attribute flags. It also provides commands to discover various sensors
implemented and managed by the platform, read any sensor synchronously
or asynchronously as allowed by the platform, program sensor attributes
and/or configurations, if applicable.

This patch adds support for most of the above features.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/Makefile  |   2 +-
 drivers/firmware/arm_scmi/sensors.c | 302 ++++++++++++++++++++++++++++++++++++
 include/linux/scmi_protocol.h       |  42 +++++
 3 files changed, 345 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/arm_scmi/sensors.c

diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
index 420c761ced94..3236890905b9 100644
--- a/drivers/firmware/arm_scmi/Makefile
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -1,4 +1,4 @@
 obj-y	= scmi-bus.o scmi-driver.o scmi-protocols.o
 scmi-bus-y = bus.o
 scmi-driver-y = driver.o
-scmi-protocols-y = base.o clock.o perf.o power.o
+scmi-protocols-y = base.o clock.o perf.o power.o sensors.o
diff --git a/drivers/firmware/arm_scmi/sensors.c b/drivers/firmware/arm_scmi/sensors.c
new file mode 100644
index 000000000000..aee38d3d1c2d
--- /dev/null
+++ b/drivers/firmware/arm_scmi/sensors.c
@@ -0,0 +1,302 @@
+/*
+ * System Control and Management Interface (SCMI) Sensor Protocol
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "common.h"
+
+enum scmi_sensor_protocol_cmd {
+	SENSOR_DESCRIPTION_GET = 0x3,
+	SENSOR_CONFIG_SET = 0x4,
+	SENSOR_TRIP_POINT_SET = 0x5,
+	SENSOR_READING_GET = 0x6,
+};
+
+struct scmi_msg_resp_sensor_attributes {
+	__le16 num_sensors;
+	    u8 max_requests;
+	    u8 reserved;
+	__le32 reg_addr_low;
+	__le32 reg_addr_high;
+	__le32 reg_size;
+};
+
+struct scmi_msg_resp_sensor_description {
+	__le16 num_returned;
+	__le16 num_remaining;
+	struct {
+		__le32 id;
+		__le32 attributes_low;
+#define SUPPORTS_ASYNC_READ(x)	((x) & BIT(31))
+#define NUM_TRIP_POINTS(x)	(((x) >> 4) & 0xff)
+		__le32 attributes_high;
+#define SENSOR_TYPE(x)		((x) & 0xff)
+#define SENSOR_SCALE(x)		(((x) >> 11) & 0x3f)
+#define SENSOR_UPDATE_SCALE(x)	(((x) >> 22) & 0x1f)
+#define SENSOR_UPDATE_BASE(x)	(((x) >> 27) & 0x1f)
+		    u8 name[SCMI_MAX_STR_SIZE];
+	} desc[0];
+};
+
+struct scmi_msg_set_sensor_config {
+	__le32 id;
+	__le32 event_control;
+};
+
+struct scmi_msg_set_sensor_trip_point {
+	__le32 id;
+	__le32 event_control;
+#define SENSOR_TP_EVENT_MASK	(0x3)
+#define SENSOR_TP_DISABLED	0x0
+#define SENSOR_TP_POSITIVE	0x1
+#define SENSOR_TP_NEGATIVE	0x2
+#define SENSOR_TP_BOTH		0x3
+#define SENSOR_TP_ID(x)		(((x) & 0xff) << 4)
+	__le32 value_low;
+	__le32 value_high;
+};
+
+struct scmi_msg_sensor_reading_get {
+	__le32 id;
+	__le32 flags;
+#define SENSOR_READ_ASYNC	BIT(0)
+};
+
+struct sensors_info {
+	int num_sensors;
+	int max_requests;
+	u64 reg_addr;
+	u32 reg_size;
+	struct scmi_sensor_info *sensors;
+};
+
+static int scmi_sensor_attributes_get(const struct scmi_handle *handle,
+				      struct sensors_info *si)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_sensor_attributes *attr;
+
+	ret = scmi_one_xfer_init(handle, PROTOCOL_ATTRIBUTES,
+				 SCMI_PROTOCOL_SENSOR, 0, sizeof(*attr), &t);
+	if (ret)
+		return ret;
+
+	attr = t->rx.buf;
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		si->num_sensors = le16_to_cpu(attr->num_sensors);
+		si->max_requests = attr->max_requests;
+		si->reg_addr = le32_to_cpu(attr->reg_addr_low) |
+				(u64)le32_to_cpu(attr->reg_addr_high) << 32;
+		si->reg_size = le32_to_cpu(attr->reg_size);
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_sensor_description_get(const struct scmi_handle *handle,
+				       struct sensors_info *si)
+{
+	int ret, cnt;
+	u32 desc_index = 0;
+	u16 num_returned, num_remaining;
+	struct scmi_xfer *t;
+	struct scmi_msg_resp_sensor_description *buf;
+
+	ret = scmi_one_xfer_init(handle, SENSOR_DESCRIPTION_GET,
+				 SCMI_PROTOCOL_SENSOR, sizeof(__le32), 0, &t);
+	if (ret)
+		return ret;
+
+	buf = t->rx.buf;
+
+	do {
+		/* Set the number of sensors to be skipped/already read */
+		*(__le32 *)t->tx.buf = cpu_to_le32(desc_index);
+
+		ret = scmi_do_xfer(handle, t);
+		if (ret)
+			break;
+
+		num_returned = le16_to_cpu(buf->num_returned);
+		num_remaining = le16_to_cpu(buf->num_remaining);
+
+		if (desc_index + num_returned > si->num_sensors) {
+			dev_err(handle->dev, "No. of sensors can't exceed %d",
+				si->num_sensors);
+			break;
+		}
+
+		for (cnt = 0; cnt < num_returned; cnt++) {
+			u32 attrh;
+			struct scmi_sensor_info *s;
+
+			attrh = le32_to_cpu(buf->desc[cnt].attributes_high);
+			s = &si->sensors[desc_index + cnt];
+			s->id = le32_to_cpu(buf->desc[cnt].id);
+			s->type = SENSOR_TYPE(attrh);
+			memcpy(s->name, buf->desc[cnt].name, SCMI_MAX_STR_SIZE);
+		}
+
+		desc_index += num_returned;
+		/*
+		 * check for both returned and remaining to avoid infinite
+		 * loop due to buggy firmware
+		 */
+	} while (num_returned && num_remaining);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int
+scmi_sensor_configuration_set(const struct scmi_handle *handle, u32 sensor_id)
+{
+	int ret;
+	u32 evt_cntl = BIT(0);
+	struct scmi_xfer *t;
+	struct scmi_msg_set_sensor_config *cfg;
+
+	ret = scmi_one_xfer_init(handle, SENSOR_CONFIG_SET,
+				 SCMI_PROTOCOL_SENSOR, sizeof(*cfg), 0, &t);
+	if (ret)
+		return ret;
+
+	cfg = t->tx.buf;
+	cfg->id = cpu_to_le32(sensor_id);
+	cfg->event_control = cpu_to_le32(evt_cntl);
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_sensor_trip_point_set(const struct scmi_handle *handle,
+				      u32 sensor_id, u8 trip_id, u64 trip_value)
+{
+	int ret;
+	u32 evt_cntl = SENSOR_TP_BOTH;
+	struct scmi_xfer *t;
+	struct scmi_msg_set_sensor_trip_point *trip;
+
+	ret = scmi_one_xfer_init(handle, SENSOR_TRIP_POINT_SET,
+				 SCMI_PROTOCOL_SENSOR, sizeof(*trip), 0, &t);
+	if (ret)
+		return ret;
+
+	trip = t->tx.buf;
+	trip->id = cpu_to_le32(sensor_id);
+	trip->event_control = cpu_to_le32(evt_cntl | SENSOR_TP_ID(trip_id));
+	trip->value_low = cpu_to_le32(trip_value & 0xffffffff);
+	trip->value_high = cpu_to_le32(trip_value >> 32);
+
+	ret = scmi_do_xfer(handle, t);
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static int scmi_sensor_reading_get(const struct scmi_handle *handle,
+				   u32 sensor_id, bool async, u64 *value)
+{
+	int ret;
+	struct scmi_xfer *t;
+	struct scmi_msg_sensor_reading_get *sensor;
+
+	ret = scmi_one_xfer_init(handle, SENSOR_READING_GET,
+				 SCMI_PROTOCOL_SENSOR, sizeof(*sensor),
+				 sizeof(u64), &t);
+	if (ret)
+		return ret;
+
+	sensor = t->tx.buf;
+	sensor->id = cpu_to_le32(sensor_id);
+	sensor->flags = cpu_to_le32(async ? SENSOR_READ_ASYNC : 0);
+
+	ret = scmi_do_xfer(handle, t);
+	if (!ret) {
+		__le32 *pval = t->rx.buf;
+
+		*value = le32_to_cpu(*pval);
+		*value |= (u64)le32_to_cpu(*(pval + 1)) << 32;
+	}
+
+	scmi_one_xfer_put(handle, t);
+	return ret;
+}
+
+static const struct scmi_sensor_info *
+scmi_sensor_info_get(const struct scmi_handle *handle, u32 sensor_id)
+{
+	struct sensors_info *si = handle->sensor_priv;
+
+	return si->sensors + sensor_id;
+}
+
+static int scmi_sensor_count_get(const struct scmi_handle *handle)
+{
+	struct sensors_info *si = handle->sensor_priv;
+
+	return si->num_sensors;
+}
+
+static struct scmi_sensor_ops sensor_ops = {
+	.count_get = scmi_sensor_count_get,
+	.info_get = scmi_sensor_info_get,
+	.configuration_set = scmi_sensor_configuration_set,
+	.trip_point_set = scmi_sensor_trip_point_set,
+	.reading_get = scmi_sensor_reading_get,
+};
+
+static int scmi_sensors_protocol_init(struct scmi_handle *handle)
+{
+	u32 version;
+	struct sensors_info *sinfo;
+
+	scmi_version_get(handle, SCMI_PROTOCOL_SENSOR, &version);
+
+	dev_dbg(handle->dev, "Sensor Version %d.%d\n",
+		PROTOCOL_REV_MAJOR(version), PROTOCOL_REV_MINOR(version));
+
+	sinfo = devm_kzalloc(handle->dev, sizeof(*sinfo), GFP_KERNEL);
+	if (!sinfo)
+		return -ENOMEM;
+
+	scmi_sensor_attributes_get(handle, sinfo);
+
+	sinfo->sensors = devm_kcalloc(handle->dev, sinfo->num_sensors,
+				      sizeof(*sinfo->sensors), GFP_KERNEL);
+	if (!sinfo->sensors)
+		return -ENOMEM;
+
+	scmi_sensor_description_get(handle, sinfo);
+
+	handle->sensor_ops = &sensor_ops;
+	handle->sensor_priv = sinfo;
+
+	return 0;
+}
+
+static int __init scmi_sensors_init(void)
+{
+	return scmi_protocol_register(SCMI_PROTOCOL_SENSOR,
+				      &scmi_sensors_protocol_init);
+}
+subsys_initcall(scmi_sensors_init);
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
index 95a645f9039d..70e2993f6d48 100644
--- a/include/linux/scmi_protocol.h
+++ b/include/linux/scmi_protocol.h
@@ -138,6 +138,45 @@ struct scmi_power_ops {
 	int (*state_notify_enable)(const struct scmi_handle *, u32, bool);
 };
 
+struct scmi_sensor_info {
+	u32 id;
+	u8 type;
+	char name[SCMI_MAX_STR_SIZE];
+};
+
+/*
+ * Partial list from Distributed Management Task Force (DMTF) specification:
+ * DSP0249 (Platform Level Data Model specification)
+ */
+enum scmi_sensor_class {
+	NONE = 0x0,
+	TEMPERATURE_C = 0x2,
+	VOLTAGE = 0x5,
+	CURRENT = 0x6,
+	POWER = 0x7,
+	ENERGY = 0x8,
+};
+
+/**
+ * struct scmi_sensor_ops - represents the various operations provided
+ *	by SCMI Sensor Protocol
+ *
+ * @count_get: get the count of sensors provided by SCMI
+ * @info_get: get the information of the specified sensor
+ * @configuration_set: control notifications on cross-over events for
+ *	the trip-points
+ * @trip_point_set: selects and configures a trip-point of interest
+ * @reading_get: gets the current value of the sensor
+ */
+struct scmi_sensor_ops {
+	int (*count_get)(const struct scmi_handle *);
+	const struct scmi_sensor_info *(*info_get)(const struct scmi_handle *,
+						   u32);
+	int (*configuration_set)(const struct scmi_handle *, u32);
+	int (*trip_point_set)(const struct scmi_handle *, u32, u8, u64);
+	int (*reading_get)(const struct scmi_handle *, u32, bool, u64 *);
+};
+
 /**
  * struct scmi_handle - Handle returned to ARM SCMI clients for usage.
  *
@@ -146,6 +185,7 @@ struct scmi_power_ops {
  * @power_ops: pointer to set of power protocol operations
  * @perf_ops: pointer to set of performance protocol operations
  * @clk_ops: pointer to set of clock protocol operations
+ * @sensor_ops: pointer to set of sensor protocol operations
  */
 struct scmi_handle {
 	struct device *dev;
@@ -153,10 +193,12 @@ struct scmi_handle {
 	struct scmi_perf_ops *perf_ops;
 	struct scmi_clk_ops *clk_ops;
 	struct scmi_power_ops *power_ops;
+	struct scmi_sensor_ops *sensor_ops;
 	/* for protocol internal use */
 	void *perf_priv;
 	void *clk_priv;
 	void *power_priv;
+	void *sensor_priv;
 };
 
 enum scmi_std_protocol {
-- 
2.7.4

[PATCH v4 10/20] firmware: arm_scmi: probe and initialise all the supported protocols

[Sudeep Holla]

Now that we have basic support for all the protocols in the
specification, let's probe them individually and initialise them.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/driver.c | 51 +++++++++++++++++++++++++++++++++++++-
 1 file changed, 50 insertions(+), 1 deletion(-)

diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c
index 20d083be474c..11c18ac9816f 100644
--- a/drivers/firmware/arm_scmi/driver.c
+++ b/drivers/firmware/arm_scmi/driver.c
@@ -494,6 +494,21 @@ void scmi_setup_protocol_implemented(const struct scmi_handle *handle,
 	info->protocols_imp = prot_imp;
 }
 
+static bool
+scmi_is_protocol_implemented(const struct scmi_handle *handle, u8 prot_id)
+{
+	int i;
+	struct scmi_info *info = handle_to_scmi_info(handle);
+
+	if (!info->protocols_imp)
+		return false;
+
+	for (i = 0; i < MAX_PROTOCOLS_IMP; i++)
+		if (info->protocols_imp[i] == prot_id)
+			return true;
+	return false;
+}
+
 /**
  * scmi_handle_get() - Get the  SCMI handle for a device
  *
@@ -691,6 +706,23 @@ static inline int scmi_mbox_chan_setup(struct scmi_info *info)
 	return 0;
 }
 
+static inline void
+scmi_create_protocol_device(struct device_node *np, struct scmi_info *info,
+			    int prot_id)
+{
+	struct scmi_device *sdev;
+
+	sdev = scmi_device_create(np, info->dev, prot_id);
+	if (!sdev) {
+		dev_err(info->dev, "failed to create %d protocol device\n",
+			prot_id);
+		return;
+	}
+
+	/* setup handle now as the transport is ready */
+	scmi_set_handle(sdev);
+}
+
 static int scmi_probe(struct platform_device *pdev)
 {
 	int ret;
@@ -698,7 +730,7 @@ static int scmi_probe(struct platform_device *pdev)
 	const struct scmi_desc *desc;
 	struct scmi_info *info;
 	struct device *dev = &pdev->dev;
-	struct device_node *np = dev->of_node;
+	struct device_node *child, *np = dev->of_node;
 
 	/* Only mailbox method supported, check for the presence of one */
 	if (scmi_mailbox_check(np)) {
@@ -741,6 +773,23 @@ static int scmi_probe(struct platform_device *pdev)
 	list_add_tail(&info->node, &scmi_list);
 	mutex_unlock(&scmi_list_mutex);
 
+	for_each_available_child_of_node(np, child) {
+		u32 prot_id;
+
+		if (of_property_read_u32(child, "reg", &prot_id))
+			continue;
+
+		prot_id &= MSG_PROTOCOL_ID_MASK;
+
+		if (!scmi_is_protocol_implemented(handle, prot_id)) {
+			dev_err(dev, "SCMI protocol %d not implemented\n",
+				prot_id);
+			continue;
+		}
+
+		scmi_create_protocol_device(child, info, prot_id);
+	}
+
 	return 0;
 }
 
-- 
2.7.4

[PATCH v4 11/20] firmware: arm_scmi: add support for polling based SCMI transfers

[Sudeep Holla]

It would be useful to have options to perform some SCMI transfers
atomically by polling for the completion flag instead of interrupt
driven. The SCMI specification has option to disable the interrupt and
poll for the completion flag in the shared memory.

This patch adds support for polling based SCMI transfers using that
option. This might be used for uninterrupted/atomic DVFS operations
from the scheduler context.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/driver.c | 49 +++++++++++++++++++++++++++++++-------
 1 file changed, 41 insertions(+), 8 deletions(-)

diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c
index 11c18ac9816f..0c9dda72f10c 100644
--- a/drivers/firmware/arm_scmi/driver.c
+++ b/drivers/firmware/arm_scmi/driver.c
@@ -26,9 +26,11 @@
  */
 
 #include <linux/bitmap.h>
+#include <linux/delay.h>
 #include <linux/export.h>
 #include <linux/io.h>
 #include <linux/kernel.h>
+#include <linux/ktime.h>
 #include <linux/mailbox_client.h>
 #include <linux/module.h>
 #include <linux/of_address.h>
@@ -363,6 +365,21 @@ void scmi_one_xfer_put(const struct scmi_handle *handle, struct scmi_xfer *xfer)
 	up(&minfo->sem_xfer_count);
 }
 
+static bool
+scmi_xfer_poll_done(const struct scmi_info *info, struct scmi_xfer *xfer)
+{
+	struct scmi_shared_mem *mem = info->tx_payload;
+	u16 xfer_id = MSG_XTRACT_TOKEN(le32_to_cpu(mem->msg_header));
+
+	if (xfer->hdr.seq != xfer_id)
+		return false;
+
+	return le32_to_cpu(mem->channel_status) &
+		(SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR |
+		SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE);
+}
+
+#define SCMI_MAX_POLLING_TIMEOUT_NS	(100 * NSEC_PER_USEC)
 /**
  * scmi_do_xfer() - Do one transfer
  *
@@ -389,14 +406,30 @@ int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
 	/* mbox_send_message returns non-negative value on success, so reset */
 	ret = 0;
 
-	/* And we wait for the response. */
-	timeout = msecs_to_jiffies(info->desc->max_rx_timeout_ms);
-	if (!wait_for_completion_timeout(&xfer->done, timeout)) {
-		dev_err(dev, "mbox timed out in resp(caller: %pF)\n",
-			(void *)_RET_IP_);
-		ret = -ETIMEDOUT;
-	} else if (xfer->hdr.status) {
-		ret = scmi_to_linux_errno(xfer->hdr.status);
+	if (xfer->hdr.poll_completion) {
+		ktime_t stop, cur;
+
+		stop = ktime_add_ns(ktime_get(), SCMI_MAX_POLLING_TIMEOUT_NS);
+		do {
+			udelay(5);
+			cur = ktime_get();
+		} while (!scmi_xfer_poll_done(info, xfer) &&
+			 ktime_before(cur, stop));
+
+		if (ktime_before(cur, stop))
+			scmi_fetch_response(xfer, info->tx_payload);
+		else
+			ret = -ETIMEDOUT;
+	} else {
+		/* And we wait for the response. */
+		timeout = msecs_to_jiffies(info->desc->max_rx_timeout_ms);
+		if (!wait_for_completion_timeout(&xfer->done, timeout)) {
+			dev_err(dev, "mbox timed out in resp(caller: %pF)\n",
+				(void *)_RET_IP_);
+			ret = -ETIMEDOUT;
+		} else if (xfer->hdr.status) {
+			ret = scmi_to_linux_errno(xfer->hdr.status);
+		}
 	}
 	/*
 	 * NOTE: we might prefer not to need the mailbox ticker to manage the
-- 
2.7.4

[PATCH v4 12/20] firmware: arm_scmi: add option for polling based performance domain operations

[Sudeep Holla]

In order to implement fast CPU DVFS switching, we need to perform all
DVFS operations atomically. Since SCMI transfer already provide option
to choose between pooling vs interrupt driven(default), we can opt for
polling based transfers for set,get performance domain operations.

This patch adds option to choose between polling vs interrupt driven
SCMI transfers for set,get performance level operations.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/perf.c | 19 +++++++++++--------
 include/linux/scmi_protocol.h    |  8 ++++----
 2 files changed, 15 insertions(+), 12 deletions(-)

diff --git a/drivers/firmware/arm_scmi/perf.c b/drivers/firmware/arm_scmi/perf.c
index a1f5cf136748..ad73ef6b8d7d 100644
--- a/drivers/firmware/arm_scmi/perf.c
+++ b/drivers/firmware/arm_scmi/perf.c
@@ -303,8 +303,8 @@ static int scmi_perf_limits_get(const struct scmi_handle *handle, u32 domain,
 	return ret;
 }
 
-static int
-scmi_perf_level_set(const struct scmi_handle *handle, u32 domain, u32 level)
+static int scmi_perf_level_set(const struct scmi_handle *handle, u32 domain,
+			       u32 level, bool poll)
 {
 	int ret;
 	struct scmi_xfer *t;
@@ -315,6 +315,7 @@ scmi_perf_level_set(const struct scmi_handle *handle, u32 domain, u32 level)
 	if (ret)
 		return ret;
 
+	t->hdr.poll_completion = poll;
 	lvl = t->tx.buf;
 	lvl->domain = cpu_to_le32(domain);
 	lvl->level = cpu_to_le32(level);
@@ -325,8 +326,8 @@ scmi_perf_level_set(const struct scmi_handle *handle, u32 domain, u32 level)
 	return ret;
 }
 
-static int
-scmi_perf_level_get(const struct scmi_handle *handle, u32 domain, u32 *level)
+static int scmi_perf_level_get(const struct scmi_handle *handle, u32 domain,
+			       u32 *level, bool poll)
 {
 	int ret;
 	struct scmi_xfer *t;
@@ -336,6 +337,7 @@ scmi_perf_level_get(const struct scmi_handle *handle, u32 domain, u32 *level)
 	if (ret)
 		return ret;
 
+	t->hdr.poll_completion = poll;
 	*(__le32 *)t->tx.buf = cpu_to_le32(domain);
 
 	ret = scmi_do_xfer(handle, t);
@@ -447,23 +449,24 @@ static int scmi_dvfs_get_transition_latency(const struct scmi_handle *handle,
 }
 
 static int scmi_dvfs_freq_set(const struct scmi_handle *handle, u32 domain,
-			      unsigned long freq)
+			      unsigned long freq, bool poll)
 {
 	struct scmi_perf_info *pi = handle->perf_priv;
 	struct perf_dom_info *dom = pi->dom_info + domain;
 
-	return scmi_perf_level_set(handle, domain, freq / dom->mult_factor);
+	return scmi_perf_level_set(handle, domain, freq / dom->mult_factor,
+				   poll);
 }
 
 static int scmi_dvfs_freq_get(const struct scmi_handle *handle, u32 domain,
-			      unsigned long *freq)
+			      unsigned long *freq, bool poll)
 {
 	int ret;
 	u32 level;
 	struct scmi_perf_info *pi = handle->perf_priv;
 	struct perf_dom_info *dom = pi->dom_info + domain;
 
-	ret = scmi_perf_level_get(handle, domain, &level);
+	ret = scmi_perf_level_get(handle, domain, &level, poll);
 	if (!ret)
 		*freq = level * dom->mult_factor;
 
diff --git a/include/linux/scmi_protocol.h b/include/linux/scmi_protocol.h
index 70e2993f6d48..3b378af1badd 100644
--- a/include/linux/scmi_protocol.h
+++ b/include/linux/scmi_protocol.h
@@ -100,16 +100,16 @@ struct scmi_clk_ops {
 struct scmi_perf_ops {
 	int (*limits_set)(const struct scmi_handle *, u32, u32, u32);
 	int (*limits_get)(const struct scmi_handle *, u32, u32 *, u32 *);
-	int (*level_set)(const struct scmi_handle *, u32, u32);
-	int (*level_get)(const struct scmi_handle *, u32, u32 *);
+	int (*level_set)(const struct scmi_handle *, u32, u32, bool);
+	int (*level_get)(const struct scmi_handle *, u32, u32 *, bool);
 	int (*limits_notify_enable)(const struct scmi_handle *, u32, bool);
 	int (*level_notify_enable)(const struct scmi_handle *, u32, bool);
 	int (*device_domain_id)(struct device *);
 	int (*get_transition_latency)(const struct scmi_handle *,
 				      struct device *);
 	int (*add_opps_to_device)(const struct scmi_handle *, struct device *);
-	int (*freq_set)(const struct scmi_handle *, u32, unsigned long);
-	int (*freq_get)(const struct scmi_handle *, u32, unsigned long *);
+	int (*freq_set)(const struct scmi_handle *, u32, unsigned long, bool);
+	int (*freq_get)(const struct scmi_handle *, u32, unsigned long *, bool);
 };
 
 /**
-- 
2.7.4

[PATCH v4 13/20] firmware: arm_scmi: refactor in preparation to support per-protocol channels

[Sudeep Holla]

In order to support per-protocol channels if available, we need to
factor out all the mailbox channel information(Tx/Rx payload and
channel handle) out of the main SCMI instance information structure.

This patch refactors the existing channel information into a separate
chan_info structure.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/driver.c | 86 ++++++++++++++++++++++++--------------
 1 file changed, 54 insertions(+), 32 deletions(-)

diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c
index 0c9dda72f10c..24acb421208c 100644
--- a/drivers/firmware/arm_scmi/driver.c
+++ b/drivers/firmware/arm_scmi/driver.c
@@ -105,6 +105,22 @@ struct scmi_desc {
 };
 
 /**
+ * struct scmi_chan_info - Structure representing a SCMI channel informfation
+ *
+ * @cl: Mailbox Client
+ * @chan: Transmit/Receive mailbox channel
+ * @payload: Transmit/Receive mailbox channel payload area
+ * @dev: Reference to device in the SCMI hierarchy corresponding to this
+ *	 channel
+ */
+struct scmi_chan_info {
+	struct mbox_client cl;
+	struct mbox_chan *chan;
+	void __iomem *payload;
+	struct device *dev;
+};
+
+/**
  * struct scmi_info - Structure representing a  SCMI instance
  *
  * @dev: Device pointer
@@ -112,10 +128,8 @@ struct scmi_desc {
  * @handle: Instance of SCMI handle to send to clients
  * @version: SCMI revision information containing protocol version,
  *	implementation version and (sub-)vendor identification.
- * @cl: Mailbox Client
- * @tx_chan: Transmit mailbox channel
- * @tx_payload: Transmit mailbox channel payload area
  * @minfo: Message info
+ * @tx_cinfo: Reference to SCMI channel information
  * @protocols_imp: list of protocols implemented, currently maximum of
  *	MAX_PROTOCOLS_IMP elements allocated by the base protocol
  * @node: list head
@@ -126,16 +140,14 @@ struct scmi_info {
 	const struct scmi_desc *desc;
 	struct scmi_revision_info version;
 	struct scmi_handle handle;
-	struct mbox_client cl;
-	struct mbox_chan *tx_chan;
-	void __iomem *tx_payload;
 	struct scmi_xfers_info minfo;
+	struct scmi_chan_info *tx_cinfo;
 	u8 *protocols_imp;
 	struct list_head node;
 	int users;
 };
 
-#define client_to_scmi_info(c)	container_of(c, struct scmi_info, cl)
+#define client_to_scmi_chan_info(c) container_of(c, struct scmi_chan_info, cl)
 #define handle_to_scmi_info(h)	container_of(h, struct scmi_info, handle)
 
 /*
@@ -218,10 +230,11 @@ static void scmi_rx_callback(struct mbox_client *cl, void *m)
 {
 	u16 xfer_id;
 	struct scmi_xfer *xfer;
-	struct scmi_info *info = client_to_scmi_info(cl);
+	struct scmi_chan_info *cinfo = client_to_scmi_chan_info(cl);
+	struct device *dev = cinfo->dev;
+	struct scmi_info *info = dev_get_drvdata(dev);
 	struct scmi_xfers_info *minfo = &info->minfo;
-	struct device *dev = info->dev;
-	struct scmi_shared_mem __iomem *mem = info->tx_payload;
+	struct scmi_shared_mem __iomem *mem = cinfo->payload;
 
 	xfer_id = MSG_XTRACT_TOKEN(ioread32(&mem->msg_header));
 
@@ -272,8 +285,8 @@ static inline u32 pack_scmi_header(struct scmi_msg_hdr *hdr)
 static void scmi_tx_prepare(struct mbox_client *cl, void *m)
 {
 	struct scmi_xfer *t = m;
-	struct scmi_info *info = client_to_scmi_info(cl);
-	struct scmi_shared_mem __iomem *mem = info->tx_payload;
+	struct scmi_chan_info *cinfo = client_to_scmi_chan_info(cl);
+	struct scmi_shared_mem __iomem *mem = cinfo->payload;
 
 	/* Mark channel busy + clear error */
 	iowrite32(0x0, &mem->channel_status);
@@ -366,15 +379,15 @@ void scmi_one_xfer_put(const struct scmi_handle *handle, struct scmi_xfer *xfer)
 }
 
 static bool
-scmi_xfer_poll_done(const struct scmi_info *info, struct scmi_xfer *xfer)
+scmi_xfer_poll_done(const struct scmi_chan_info *cinfo, struct scmi_xfer *xfer)
 {
-	struct scmi_shared_mem *mem = info->tx_payload;
-	u16 xfer_id = MSG_XTRACT_TOKEN(le32_to_cpu(mem->msg_header));
+	struct scmi_shared_mem __iomem *mem = cinfo->payload;
+	u16 xfer_id = MSG_XTRACT_TOKEN(ioread32(&mem->msg_header));
 
 	if (xfer->hdr.seq != xfer_id)
 		return false;
 
-	return le32_to_cpu(mem->channel_status) &
+	return ioread32(&mem->channel_status) &
 		(SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR |
 		SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE);
 }
@@ -396,8 +409,9 @@ int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
 	int timeout;
 	struct scmi_info *info = handle_to_scmi_info(handle);
 	struct device *dev = info->dev;
+	struct scmi_chan_info *cinfo = info->tx_cinfo;
 
-	ret = mbox_send_message(info->tx_chan, xfer);
+	ret = mbox_send_message(cinfo->chan, xfer);
 	if (ret < 0) {
 		dev_dbg(dev, "mbox send fail %d\n", ret);
 		return ret;
@@ -413,11 +427,11 @@ int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
 		do {
 			udelay(5);
 			cur = ktime_get();
-		} while (!scmi_xfer_poll_done(info, xfer) &&
+		} while (!scmi_xfer_poll_done(cinfo, xfer) &&
 			 ktime_before(cur, stop));
 
 		if (ktime_before(cur, stop))
-			scmi_fetch_response(xfer, info->tx_payload);
+			scmi_fetch_response(xfer, cinfo->payload);
 		else
 			ret = -ETIMEDOUT;
 	} else {
@@ -437,7 +451,7 @@ int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
 	 * Unfortunately, we have to kick the mailbox framework after we have
 	 * received our message.
 	 */
-	mbox_client_txdone(info->tx_chan, ret);
+	mbox_client_txdone(cinfo->chan, ret);
 
 	return ret;
 }
@@ -667,11 +681,11 @@ static int scmi_mailbox_check(struct device_node *np)
 	return of_parse_phandle_with_args(np, "mboxes", "#mbox-cells", 0, &arg);
 }
 
-static int scmi_mbox_free_channel(struct scmi_info *info)
+static int scmi_mbox_free_channel(struct scmi_chan_info *cinfo)
 {
-	if (!IS_ERR_OR_NULL(info->tx_chan)) {
-		mbox_free_channel(info->tx_chan);
-		info->tx_chan = NULL;
+	if (!IS_ERR_OR_NULL(cinfo->chan)) {
+		mbox_free_channel(cinfo->chan);
+		cinfo->chan = NULL;
 	}
 
 	return 0;
@@ -691,7 +705,7 @@ static int scmi_remove(struct platform_device *pdev)
 
 	if (!ret)
 		/* Safe to free channels since no more users */
-		return scmi_mbox_free_channel(info);
+		return scmi_mbox_free_channel(info->tx_cinfo);
 
 	return ret;
 }
@@ -703,9 +717,17 @@ static inline int scmi_mbox_chan_setup(struct scmi_info *info)
 	resource_size_t size;
 	struct device *dev = info->dev;
 	struct device_node *shmem, *np = dev->of_node;
+	struct scmi_chan_info *cinfo;
 	struct mbox_client *cl;
 
-	cl = &info->cl;
+	cinfo = devm_kzalloc(info->dev, sizeof(*cinfo), GFP_KERNEL);
+	if (!cinfo)
+		return -ENOMEM;
+
+	info->tx_cinfo = cinfo;
+	cinfo->dev = dev;
+
+	cl = &cinfo->cl;
 	cl->dev = dev;
 	cl->rx_callback = scmi_rx_callback;
 	cl->tx_prepare = scmi_tx_prepare;
@@ -721,16 +743,16 @@ static inline int scmi_mbox_chan_setup(struct scmi_info *info)
 	}
 
 	size = resource_size(&res);
-	info->tx_payload = devm_ioremap(dev, res.start, size);
-	if (!info->tx_payload) {
+	cinfo->payload = devm_ioremap(info->dev, res.start, size);
+	if (!cinfo->payload) {
 		dev_err(dev, "failed to ioremap SCMI Tx payload\n");
 		return -EADDRNOTAVAIL;
 	}
 
 	/* Transmit channel is first entry i.e. index 0 */
-	info->tx_chan = mbox_request_channel(cl, 0);
-	if (IS_ERR(info->tx_chan)) {
-		ret = PTR_ERR(info->tx_chan);
+	cinfo->chan = mbox_request_channel(cl, 0);
+	if (IS_ERR(cinfo->chan)) {
+		ret = PTR_ERR(cinfo->chan);
 		if (ret != -EPROBE_DEFER)
 			dev_err(dev, "failed to request SCMI Tx mailbox\n");
 		return ret;
@@ -798,7 +820,7 @@ static int scmi_probe(struct platform_device *pdev)
 	ret = scmi_base_protocol_init(handle);
 	if (ret) {
 		dev_err(dev, "unable to communicate with SCMI(%d)\n", ret);
-		scmi_mbox_free_channel(info);
+		scmi_mbox_free_channel(info->tx_cinfo);
 		return ret;
 	}
 
-- 
2.7.4

[PATCH v4 14/20] firmware: arm_scmi: add per-protocol channels support using idr objects

[Sudeep Holla]

In order to maintain the channel information per protocol, we need
some sort of list or hashtable to hold all this information. IDR
provides sparse array mapping of small integer ID numbers onto arbitrary
pointers. In this case the arbitrary pointers can be pointers to the
channel information.

This patch adds support for per-protocol channels using those idr
objects.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/arm_scmi/driver.c | 54 +++++++++++++++++++++++++++++---------
 1 file changed, 42 insertions(+), 12 deletions(-)

diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c
index 24acb421208c..6734a035bcc6 100644
--- a/drivers/firmware/arm_scmi/driver.c
+++ b/drivers/firmware/arm_scmi/driver.c
@@ -118,6 +118,7 @@ struct scmi_chan_info {
 	struct mbox_chan *chan;
 	void __iomem *payload;
 	struct device *dev;
+	struct scmi_handle *handle;
 };
 
 /**
@@ -129,7 +130,7 @@ struct scmi_chan_info {
  * @version: SCMI revision information containing protocol version,
  *	implementation version and (sub-)vendor identification.
  * @minfo: Message info
- * @tx_cinfo: Reference to SCMI channel information
+ * @tx_idr: IDR object to map protocol id to channel info pointer
  * @protocols_imp: list of protocols implemented, currently maximum of
  *	MAX_PROTOCOLS_IMP elements allocated by the base protocol
  * @node: list head
@@ -141,7 +142,7 @@ struct scmi_info {
 	struct scmi_revision_info version;
 	struct scmi_handle handle;
 	struct scmi_xfers_info minfo;
-	struct scmi_chan_info *tx_cinfo;
+	struct idr tx_idr;
 	u8 *protocols_imp;
 	struct list_head node;
 	int users;
@@ -232,7 +233,7 @@ static void scmi_rx_callback(struct mbox_client *cl, void *m)
 	struct scmi_xfer *xfer;
 	struct scmi_chan_info *cinfo = client_to_scmi_chan_info(cl);
 	struct device *dev = cinfo->dev;
-	struct scmi_info *info = dev_get_drvdata(dev);
+	struct scmi_info *info = handle_to_scmi_info(cinfo->handle);
 	struct scmi_xfers_info *minfo = &info->minfo;
 	struct scmi_shared_mem __iomem *mem = cinfo->payload;
 
@@ -409,7 +410,11 @@ int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
 	int timeout;
 	struct scmi_info *info = handle_to_scmi_info(handle);
 	struct device *dev = info->dev;
-	struct scmi_chan_info *cinfo = info->tx_cinfo;
+	struct scmi_chan_info *cinfo;
+
+	cinfo = idr_find(&info->tx_idr, xfer->hdr.protocol_id);
+	if (unlikely(!cinfo))
+		return -EINVAL;
 
 	ret = mbox_send_message(cinfo->chan, xfer);
 	if (ret < 0) {
@@ -681,13 +686,18 @@ static int scmi_mailbox_check(struct device_node *np)
 	return of_parse_phandle_with_args(np, "mboxes", "#mbox-cells", 0, &arg);
 }
 
-static int scmi_mbox_free_channel(struct scmi_chan_info *cinfo)
+static int scmi_mbox_free_channel(int id, void *p, void *data)
 {
+	struct scmi_chan_info *cinfo = p;
+	struct idr *idr = data;
+
 	if (!IS_ERR_OR_NULL(cinfo->chan)) {
 		mbox_free_channel(cinfo->chan);
 		cinfo->chan = NULL;
 	}
 
+	idr_remove(idr, id);
+
 	return 0;
 }
 
@@ -695,6 +705,7 @@ static int scmi_remove(struct platform_device *pdev)
 {
 	int ret = 0;
 	struct scmi_info *info = platform_get_drvdata(pdev);
+	struct idr *idr = &info->tx_idr;
 
 	mutex_lock(&scmi_list_mutex);
 	if (info->users)
@@ -703,28 +714,34 @@ static int scmi_remove(struct platform_device *pdev)
 		list_del(&info->node);
 	mutex_unlock(&scmi_list_mutex);
 
-	if (!ret)
+	if (!ret) {
 		/* Safe to free channels since no more users */
-		return scmi_mbox_free_channel(info->tx_cinfo);
+		ret = idr_for_each(idr, scmi_mbox_free_channel, idr);
+		idr_destroy(&info->tx_idr);
+	}
 
 	return ret;
 }
 
-static inline int scmi_mbox_chan_setup(struct scmi_info *info)
+static inline int
+scmi_mbox_chan_setup(struct scmi_info *info, struct device *dev, int prot_id)
 {
 	int ret;
 	struct resource res;
 	resource_size_t size;
-	struct device *dev = info->dev;
 	struct device_node *shmem, *np = dev->of_node;
 	struct scmi_chan_info *cinfo;
 	struct mbox_client *cl;
 
+	if (scmi_mailbox_check(np)) {
+		cinfo = idr_find(&info->tx_idr, SCMI_PROTOCOL_BASE);
+		goto idr_alloc;
+	}
+
 	cinfo = devm_kzalloc(info->dev, sizeof(*cinfo), GFP_KERNEL);
 	if (!cinfo)
 		return -ENOMEM;
 
-	info->tx_cinfo = cinfo;
 	cinfo->dev = dev;
 
 	cl = &cinfo->cl;
@@ -758,6 +775,14 @@ static inline int scmi_mbox_chan_setup(struct scmi_info *info)
 		return ret;
 	}
 
+idr_alloc:
+	ret = idr_alloc(&info->tx_idr, cinfo, prot_id, prot_id + 1, GFP_KERNEL);
+	if (ret != prot_id) {
+		dev_err(dev, "unable to allocate SCMI idr slot err %d\n", ret);
+		return ret;
+	}
+
+	cinfo->handle = &info->handle;
 	return 0;
 }
 
@@ -774,6 +799,11 @@ scmi_create_protocol_device(struct device_node *np, struct scmi_info *info,
 		return;
 	}
 
+	if (scmi_mbox_chan_setup(info, &sdev->dev, prot_id)) {
+		dev_err(&sdev->dev, "failed to setup transport\n");
+		scmi_device_destroy(sdev);
+	}
+
 	/* setup handle now as the transport is ready */
 	scmi_set_handle(sdev);
 }
@@ -808,19 +838,19 @@ static int scmi_probe(struct platform_device *pdev)
 		return ret;
 
 	platform_set_drvdata(pdev, info);
+	idr_init(&info->tx_idr);
 
 	handle = &info->handle;
 	handle->dev = info->dev;
 	handle->version = &info->version;
 
-	ret = scmi_mbox_chan_setup(info);
+	ret = scmi_mbox_chan_setup(info, dev, SCMI_PROTOCOL_BASE);
 	if (ret)
 		return ret;
 
 	ret = scmi_base_protocol_init(handle);
 	if (ret) {
 		dev_err(dev, "unable to communicate with SCMI(%d)\n", ret);
-		scmi_mbox_free_channel(info->tx_cinfo);
 		return ret;
 	}
 
-- 
2.7.4

[PATCH v4 15/20] firmware: arm_scmi: add device power domain support using genpd

[Sudeep Holla]

This patch hooks up the support for device power domain provided by
SCMI using the Linux generic power domain infrastructure.

Cc: Kevin Hilman <khilman@baylibre.com>
Cc: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/firmware/Kconfig                   |  13 +++
 drivers/firmware/arm_scmi/Makefile         |   1 +
 drivers/firmware/arm_scmi/scmi_pm_domain.c | 140 +++++++++++++++++++++++++++++
 3 files changed, 154 insertions(+)
 create mode 100644 drivers/firmware/arm_scmi/scmi_pm_domain.c

diff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig
index 86ea9d531587..86d2901ad87a 100644
--- a/drivers/firmware/Kconfig
+++ b/drivers/firmware/Kconfig
@@ -40,6 +40,19 @@ config ARM_SCMI_PROTOCOL
 	  This protocol library provides interface for all the client drivers
 	  making use of the features offered by the SCMI.
 
+config ARM_SCMI_POWER_DOMAIN
+	tristate "SCMI power domain driver"
+	depends on ARM_SCMI_PROTOCOL || (COMPILE_TEST && OF)
+	default y
+	select PM_GENERIC_DOMAINS if PM
+	help
+	  This enables support for the SCMI power domains which can be
+	  enabled or disabled via the SCP firmware
+
+	  This driver can also be built as a module.  If so, the module
+	  will be called scmi_pm_domain. Note this may needed early in boot
+	  before rootfs may be available.
+
 config ARM_SCPI_PROTOCOL
 	tristate "ARM System Control and Power Interface (SCPI) Message Protocol"
 	depends on ARM || ARM64 || COMPILE_TEST
diff --git a/drivers/firmware/arm_scmi/Makefile b/drivers/firmware/arm_scmi/Makefile
index 3236890905b9..99e36c580fbc 100644
--- a/drivers/firmware/arm_scmi/Makefile
+++ b/drivers/firmware/arm_scmi/Makefile
@@ -2,3 +2,4 @@ obj-y	= scmi-bus.o scmi-driver.o scmi-protocols.o
 scmi-bus-y = bus.o
 scmi-driver-y = driver.o
 scmi-protocols-y = base.o clock.o perf.o power.o sensors.o
+obj-$(CONFIG_ARM_SCMI_POWER_DOMAIN) += scmi_pm_domain.o
diff --git a/drivers/firmware/arm_scmi/scmi_pm_domain.c b/drivers/firmware/arm_scmi/scmi_pm_domain.c
new file mode 100644
index 000000000000..d8c9aa6960cf
--- /dev/null
+++ b/drivers/firmware/arm_scmi/scmi_pm_domain.c
@@ -0,0 +1,140 @@
+/*
+ * SCMI Generic power domain support.
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/pm_domain.h>
+#include <linux/scmi_protocol.h>
+
+struct scmi_pm_domain {
+	struct generic_pm_domain genpd;
+	const struct scmi_handle *handle;
+	const char *name;
+	u32 domain;
+};
+
+#define to_scmi_pd(gpd) container_of(gpd, struct scmi_pm_domain, genpd)
+
+static int scmi_pd_power(struct generic_pm_domain *domain, bool power_on)
+{
+	int ret;
+	u32 state, ret_state;
+	struct scmi_pm_domain *pd = to_scmi_pd(domain);
+	const struct scmi_power_ops *ops = pd->handle->power_ops;
+
+	if (power_on)
+		state = SCMI_POWER_STATE_GENERIC_ON;
+	else
+		state = SCMI_POWER_STATE_GENERIC_OFF;
+
+	ret = ops->state_set(pd->handle, pd->domain, state);
+	if (!ret)
+		ret = ops->state_get(pd->handle, pd->domain, &ret_state);
+	if (!ret && state != ret_state)
+		return -EIO;
+
+	return ret;
+}
+
+static int scmi_pd_power_on(struct generic_pm_domain *domain)
+{
+	return scmi_pd_power(domain, true);
+}
+
+static int scmi_pd_power_off(struct generic_pm_domain *domain)
+{
+	return scmi_pd_power(domain, false);
+}
+
+static int scmi_pm_domain_probe(struct scmi_device *sdev)
+{
+	int num_domains, i;
+	struct device *dev = &sdev->dev;
+	struct device_node *np = dev->of_node;
+	struct scmi_pm_domain *scmi_pd;
+	struct genpd_onecell_data *scmi_pd_data;
+	struct generic_pm_domain **domains;
+	const struct scmi_handle *handle = sdev->handle;
+
+	if (!handle || !handle->power_ops)
+		return -ENODEV;
+
+	num_domains = handle->power_ops->num_domains_get(handle);
+	if (num_domains < 0) {
+		dev_err(dev, "number of domains not found\n");
+		return num_domains;
+	}
+
+	scmi_pd = devm_kcalloc(dev, num_domains, sizeof(*scmi_pd), GFP_KERNEL);
+	if (!scmi_pd)
+		return -ENOMEM;
+
+	scmi_pd_data = devm_kzalloc(dev, sizeof(*scmi_pd_data), GFP_KERNEL);
+	if (!scmi_pd_data)
+		return -ENOMEM;
+
+	domains = devm_kcalloc(dev, num_domains, sizeof(*domains), GFP_KERNEL);
+	if (!domains)
+		return -ENOMEM;
+
+	for (i = 0; i < num_domains; i++, scmi_pd++) {
+		u32 state;
+
+		domains[i] = &scmi_pd->genpd;
+
+		scmi_pd->domain = i;
+		scmi_pd->handle = handle;
+		scmi_pd->name = handle->power_ops->name_get(handle, i);
+		scmi_pd->genpd.name = scmi_pd->name;
+		scmi_pd->genpd.power_off = scmi_pd_power_off;
+		scmi_pd->genpd.power_on = scmi_pd_power_on;
+
+		if (handle->power_ops->state_get(handle, i, &state)) {
+			dev_dbg(dev, "failed to get state for domain %d\n", i);
+			continue;
+		}
+
+		pm_genpd_init(&scmi_pd->genpd, NULL,
+			      state == SCMI_POWER_STATE_GENERIC_OFF);
+	}
+
+	scmi_pd_data->domains = domains;
+	scmi_pd_data->num_domains = num_domains;
+
+	of_genpd_add_provider_onecell(np, scmi_pd_data);
+
+	return 0;
+}
+
+static const struct scmi_device_id scmi_id_table[] = {
+	{ SCMI_PROTOCOL_POWER },
+	{ },
+};
+MODULE_DEVICE_TABLE(scmi, scmi_id_table);
+
+static struct scmi_driver scmi_power_domain_driver = {
+	.name = "scmi-power-domain",
+	.probe = scmi_pm_domain_probe,
+	.id_table = scmi_id_table,
+};
+module_scmi_driver(scmi_power_domain_driver);
+
+MODULE_AUTHOR("Sudeep Holla <sudeep.holla@arm.com>");
+MODULE_DESCRIPTION("ARM SCMI power domain driver");
+MODULE_LICENSE("GPL v2");
-- 
2.7.4

[PATCH v4 16/20] clk: add support for clocks provided by SCMI

[Sudeep Holla]

On some ARM based systems, a separate Cortex-M based System Control
Processor(SCP) provides the overall power, clock, reset and system
control. System Control and Management Interface(SCMI) Message Protocol
is defined for the communication between the Application Cores(AP)
and the SCP.

This patch adds support for the clocks provided by SCP using SCMI
protocol.

Cc: linux-clk@vger.kernel.org
Cc: Michael Turquette <mturquette@baylibre.com>
Acked-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 MAINTAINERS            |   2 +-
 drivers/clk/Kconfig    |  10 +++
 drivers/clk/Makefile   |   1 +
 drivers/clk/clk-scmi.c | 213 +++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 225 insertions(+), 1 deletion(-)
 create mode 100644 drivers/clk/clk-scmi.c

diff --git a/MAINTAINERS b/MAINTAINERS
index 8ee63627f1ac..1f9004d68ad4 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -12954,7 +12954,7 @@ M:	Sudeep Holla <sudeep.holla@arm.com>
 L:	linux-arm-kernel@lists.infradead.org
 S:	Maintained
 F:	Documentation/devicetree/bindings/arm/arm,sc[mp]i.txt
-F:	drivers/clk/clk-scpi.c
+F:	drivers/clk/clk-sc[mp]i.c
 F:	drivers/cpufreq/scpi-cpufreq.c
 F:	drivers/firmware/arm_scpi.c
 F:	drivers/firmware/arm_scmi/
diff --git a/drivers/clk/Kconfig b/drivers/clk/Kconfig
index 1c4e1aa6767e..57c66b22eab8 100644
--- a/drivers/clk/Kconfig
+++ b/drivers/clk/Kconfig
@@ -62,6 +62,16 @@ config COMMON_CLK_HI655X
 	  multi-function device has one fixed-rate oscillator, clocked
 	  at 32KHz.
 
+config COMMON_CLK_SCMI
+	tristate "Clock driver controlled via SCMI interface"
+	depends on ARM_SCMI_PROTOCOL || COMPILE_TEST
+	  ---help---
+	  This driver provides support for clocks that are controlled
+	  by firmware that implements the SCMI interface.
+
+	  This driver uses SCMI Message Protocol to interact with the
+	  firmware providing all the clock controls.
+
 config COMMON_CLK_SCPI
 	tristate "Clock driver controlled via SCPI interface"
 	depends on ARM_SCPI_PROTOCOL || COMPILE_TEST
diff --git a/drivers/clk/Makefile b/drivers/clk/Makefile
index c99f363826f0..46ad2f2b686a 100644
--- a/drivers/clk/Makefile
+++ b/drivers/clk/Makefile
@@ -39,6 +39,7 @@ obj-$(CONFIG_CLK_QORIQ)			+= clk-qoriq.o
 obj-$(CONFIG_COMMON_CLK_RK808)		+= clk-rk808.o
 obj-$(CONFIG_COMMON_CLK_HI655X)		+= clk-hi655x.o
 obj-$(CONFIG_COMMON_CLK_S2MPS11)	+= clk-s2mps11.o
+obj-$(CONFIG_COMMON_CLK_SCMI)           += clk-scmi.o
 obj-$(CONFIG_COMMON_CLK_SCPI)           += clk-scpi.o
 obj-$(CONFIG_COMMON_CLK_SI5351)		+= clk-si5351.o
 obj-$(CONFIG_COMMON_CLK_SI514)		+= clk-si514.o
diff --git a/drivers/clk/clk-scmi.c b/drivers/clk/clk-scmi.c
new file mode 100644
index 000000000000..1e4d7a57779b
--- /dev/null
+++ b/drivers/clk/clk-scmi.c
@@ -0,0 +1,213 @@
+/*
+ * System Control and Power Interface (SCMI) Protocol based clock driver
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/clk-provider.h>
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/of.h>
+#include <linux/module.h>
+#include <linux/scmi_protocol.h>
+#include <asm/div64.h>
+
+struct scmi_clk {
+	u32 id;
+	struct clk_hw hw;
+	const struct scmi_clock_info *info;
+	const struct scmi_handle *handle;
+};
+
+#define to_scmi_clk(clk) container_of(clk, struct scmi_clk, hw)
+
+static unsigned long scmi_clk_recalc_rate(struct clk_hw *hw,
+					  unsigned long parent_rate)
+{
+	int ret;
+	u64 rate;
+	struct scmi_clk *clk = to_scmi_clk(hw);
+
+	ret = clk->handle->clk_ops->rate_get(clk->handle, clk->id, &rate);
+	if (ret)
+		return 0;
+	return rate;
+}
+
+static long scmi_clk_round_rate(struct clk_hw *hw, unsigned long rate,
+				unsigned long *parent_rate)
+{
+	int step;
+	u64 fmin, fmax, ftmp;
+	struct scmi_clk *clk = to_scmi_clk(hw);
+
+	/*
+	 * We can't figure out what rate it will be, so just return the
+	 * rate back to the caller. scmi_clk_recalc_rate() will be called
+	 * after the rate is set and we'll know what rate the clock is
+	 * running at then.
+	 */
+	if (clk->info->rate_discrete)
+		return rate;
+
+	fmin = clk->info->range.min_rate;
+	fmax = clk->info->range.max_rate;
+	if (rate <= fmin)
+		return fmin;
+	else if (rate >= fmax)
+		return fmax;
+
+	ftmp = rate - fmin;
+	ftmp += clk->info->range.step_size - 1; /* to round up */
+	step = do_div(ftmp, clk->info->range.step_size);
+
+	return step * clk->info->range.step_size + fmin;
+}
+
+static int scmi_clk_set_rate(struct clk_hw *hw, unsigned long rate,
+			     unsigned long parent_rate)
+{
+	struct scmi_clk *clk = to_scmi_clk(hw);
+
+	return clk->handle->clk_ops->rate_set(clk->handle, clk->id, 0, rate);
+}
+
+static int scmi_clk_enable(struct clk_hw *hw)
+{
+	struct scmi_clk *clk = to_scmi_clk(hw);
+
+	return clk->handle->clk_ops->enable(clk->handle, clk->id);
+}
+
+static void scmi_clk_disable(struct clk_hw *hw)
+{
+	struct scmi_clk *clk = to_scmi_clk(hw);
+
+	clk->handle->clk_ops->disable(clk->handle, clk->id);
+}
+
+static const struct clk_ops scmi_clk_ops = {
+	.recalc_rate = scmi_clk_recalc_rate,
+	.round_rate = scmi_clk_round_rate,
+	.set_rate = scmi_clk_set_rate,
+	/*
+	 * We can't provide enable/disable callback as we can't perform the same
+	 * in atomic context. Since the clock framework provides standard API
+	 * clk_prepare_enable that helps cases using clk_enable in non-atomic
+	 * context, it should be fine providing prepare/unprepare.
+	 */
+	.prepare = scmi_clk_enable,
+	.unprepare = scmi_clk_disable,
+};
+
+static int scmi_clk_ops_init(struct device *dev, struct scmi_clk *sclk)
+{
+	int ret;
+	struct clk_init_data init = {
+		.flags = CLK_GET_RATE_NOCACHE,
+		.num_parents = 0,
+		.ops = &scmi_clk_ops,
+		.name = sclk->info->name,
+	};
+
+	sclk->hw.init = &init;
+	ret = devm_clk_hw_register(dev, &sclk->hw);
+	if (!ret)
+		clk_hw_set_rate_range(&sclk->hw, sclk->info->range.min_rate,
+				      sclk->info->range.max_rate);
+	return ret;
+}
+
+static int scmi_clocks_probe(struct scmi_device *sdev)
+{
+	int idx, count, err;
+	struct clk_hw **hws;
+	struct clk_hw_onecell_data *clk_data;
+	struct device *dev = &sdev->dev;
+	struct device_node *np = dev->of_node;
+	const struct scmi_handle *handle = sdev->handle;
+
+	if (!handle || !handle->clk_ops)
+		return -ENODEV;
+
+	count = handle->clk_ops->count_get(handle);
+	if (count < 0) {
+		dev_err(dev, "%s: invalid clock output count\n", np->name);
+		return -EINVAL;
+	}
+
+	clk_data = devm_kzalloc(dev, sizeof(*clk_data) +
+				sizeof(*clk_data->hws) * count, GFP_KERNEL);
+	if (!clk_data)
+		return -ENOMEM;
+
+	clk_data->num = count;
+	hws = clk_data->hws;
+
+	for (idx = 0; idx < count; idx++) {
+		struct scmi_clk *sclk;
+
+		sclk = devm_kzalloc(dev, sizeof(*sclk), GFP_KERNEL);
+		if (!sclk)
+			return -ENOMEM;
+
+		sclk->info = handle->clk_ops->info_get(handle, idx);
+		if (!sclk->info) {
+			dev_dbg(dev, "invalid clock info for idx %d\n", idx);
+			continue;
+		}
+
+		sclk->id = idx;
+		sclk->handle = handle;
+
+		err = scmi_clk_ops_init(dev, sclk);
+		if (err) {
+			dev_err(dev, "failed to register clock %d\n", idx);
+			devm_kfree(dev, sclk);
+			hws[idx] = NULL;
+		} else {
+			dev_dbg(dev, "Registered clock:%s\n", sclk->info->name);
+			hws[idx] = &sclk->hw;
+		}
+	}
+
+	return of_clk_add_hw_provider(np, of_clk_hw_onecell_get, clk_data);
+}
+
+static void scmi_clocks_remove(struct scmi_device *sdev)
+{
+	struct device *dev = &sdev->dev;
+	struct device_node *np = dev->of_node;
+
+	of_clk_del_provider(np);
+}
+
+static const struct scmi_device_id scmi_id_table[] = {
+	{ SCMI_PROTOCOL_CLOCK },
+	{ },
+};
+MODULE_DEVICE_TABLE(scmi, scmi_id_table);
+
+static struct scmi_driver scmi_clocks_driver = {
+	.name = "scmi-clocks",
+	.probe = scmi_clocks_probe,
+	.remove = scmi_clocks_remove,
+	.id_table = scmi_id_table,
+};
+module_scmi_driver(scmi_clocks_driver);
+
+MODULE_AUTHOR("Sudeep Holla <sudeep.holla@arm.com>");
+MODULE_DESCRIPTION("ARM SCMI clock driver");
+MODULE_LICENSE("GPL v2");
-- 
2.7.4

[PATCH v4 17/20] hwmon: (core) Add hwmon_max to hwmon_sensor_types enumeration

[Sudeep Holla]

It's useful to know the maximum types of sensor supported by hwmon
framework. It can be used to allocate some data structures when sorting
the monitors based on their type.

This will be used by scmi hwmon support.

Cc: linux-hwmon@vger.kernel.org
Acked-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 include/linux/hwmon.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/include/linux/hwmon.h b/include/linux/hwmon.h
index ceb751987c40..e5fd2707b6df 100644
--- a/include/linux/hwmon.h
+++ b/include/linux/hwmon.h
@@ -29,6 +29,7 @@ enum hwmon_sensor_types {
 	hwmon_humidity,
 	hwmon_fan,
 	hwmon_pwm,
+	hwmon_max,
 };
 
 enum hwmon_chip_attributes {
-- 
2.7.4

[PATCH v4 18/20] hwmon: add support for sensors exported via ARM SCMI

[Sudeep Holla]

Create a driver to add support for SoC sensors exported by the System
Control Processor (SCP) via the System Control and Management Interface
(SCMI). The supported sensor types is one of voltage, temperature,
current, and power.

The sensor labels and values provided by the SCP are exported via the
hwmon sysfs interface.

Cc: linux-hwmon@vger.kernel.org
Acked-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/hwmon/Kconfig      |  12 +++
 drivers/hwmon/Makefile     |   1 +
 drivers/hwmon/scmi-hwmon.c | 233 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 246 insertions(+)
 create mode 100644 drivers/hwmon/scmi-hwmon.c

diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
index d65431417b17..0b75e9a89463 100644
--- a/drivers/hwmon/Kconfig
+++ b/drivers/hwmon/Kconfig
@@ -321,6 +321,18 @@ config SENSORS_APPLESMC
 	  Say Y here if you have an applicable laptop and want to experience
 	  the awesome power of applesmc.
 
+config SENSORS_ARM_SCMI
+	tristate "ARM SCMI Sensors"
+	depends on ARM_SCMI_PROTOCOL
+	depends on THERMAL || !THERMAL_OF
+	help
+	  This driver provides support for temperature, voltage, current
+	  and power sensors available on SCMI based platforms. The actual
+	  number and type of sensors exported depend on the platform.
+
+	  This driver can also be built as a module.  If so, the module
+	  will be called scmi-hwmon.
+
 config SENSORS_ARM_SCPI
 	tristate "ARM SCPI Sensors"
 	depends on ARM_SCPI_PROTOCOL
diff --git a/drivers/hwmon/Makefile b/drivers/hwmon/Makefile
index c84d9784be98..a51c2dcef11c 100644
--- a/drivers/hwmon/Makefile
+++ b/drivers/hwmon/Makefile
@@ -44,6 +44,7 @@ obj-$(CONFIG_SENSORS_ADT7462)	+= adt7462.o
 obj-$(CONFIG_SENSORS_ADT7470)	+= adt7470.o
 obj-$(CONFIG_SENSORS_ADT7475)	+= adt7475.o
 obj-$(CONFIG_SENSORS_APPLESMC)	+= applesmc.o
+obj-$(CONFIG_SENSORS_ARM_SCMI)	+= scmi-hwmon.o
 obj-$(CONFIG_SENSORS_ARM_SCPI)	+= scpi-hwmon.o
 obj-$(CONFIG_SENSORS_ASC7621)	+= asc7621.o
 obj-$(CONFIG_SENSORS_ASPEED)	+= aspeed-pwm-tacho.o
diff --git a/drivers/hwmon/scmi-hwmon.c b/drivers/hwmon/scmi-hwmon.c
new file mode 100644
index 000000000000..7a0ab768d4b5
--- /dev/null
+++ b/drivers/hwmon/scmi-hwmon.c
@@ -0,0 +1,233 @@
+/*
+ * System Control and Management Interface(SCMI) based hwmon sensor driver
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ * Sudeep Holla <sudeep.holla@arm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/hwmon.h>
+#include <linux/module.h>
+#include <linux/scmi_protocol.h>
+#include <linux/slab.h>
+#include <linux/sysfs.h>
+#include <linux/thermal.h>
+
+struct scmi_sensors {
+	const struct scmi_handle *handle;
+	const struct scmi_sensor_info **info[hwmon_max];
+};
+
+static int scmi_hwmon_read(struct device *dev, enum hwmon_sensor_types type,
+			   u32 attr, int channel, long *val)
+{
+	int ret;
+	u64 value;
+	const struct scmi_sensor_info *sensor;
+	struct scmi_sensors *scmi_sensors = dev_get_drvdata(dev);
+	const struct scmi_handle *h = scmi_sensors->handle;
+
+	sensor = *(scmi_sensors->info[type] + channel);
+	ret = h->sensor_ops->reading_get(h, sensor->id, false, &value);
+	if (!ret)
+		*val = value;
+
+	return ret;
+}
+
+static int
+scmi_hwmon_read_string(struct device *dev, enum hwmon_sensor_types type,
+		       u32 attr, int channel, const char **str)
+{
+	const struct scmi_sensor_info *sensor;
+	struct scmi_sensors *scmi_sensors = dev_get_drvdata(dev);
+
+	sensor = *(scmi_sensors->info[type] + channel);
+	*str = sensor->name;
+
+	return 0;
+}
+
+static umode_t
+scmi_hwmon_is_visible(const void *drvdata, enum hwmon_sensor_types type,
+		      u32 attr, int channel)
+{
+	const struct scmi_sensor_info *sensor;
+	const struct scmi_sensors *scmi_sensors = drvdata;
+
+	sensor = *(scmi_sensors->info[type] + channel);
+	if (sensor && sensor->name)
+		return S_IRUGO;
+
+	return 0;
+}
+
+static const struct hwmon_ops scmi_hwmon_ops = {
+	.is_visible = scmi_hwmon_is_visible,
+	.read = scmi_hwmon_read,
+	.read_string = scmi_hwmon_read_string,
+};
+
+static struct hwmon_chip_info scmi_chip_info = {
+	.ops = &scmi_hwmon_ops,
+	.info = NULL,
+};
+
+static int scmi_hwmon_add_chan_info(struct hwmon_channel_info *scmi_hwmon_chan,
+				    struct device *dev, int num,
+				    enum hwmon_sensor_types type, u32 config)
+{
+	int i;
+	u32 *cfg = devm_kcalloc(dev, num + 1, sizeof(*cfg), GFP_KERNEL);
+
+	if (!cfg)
+		return -ENOMEM;
+
+	scmi_hwmon_chan->type = type;
+	scmi_hwmon_chan->config = cfg;
+	for (i = 0; i < num; i++, cfg++)
+		*cfg = config;
+
+	return 0;
+}
+
+static enum hwmon_sensor_types scmi_types[] = {
+	[TEMPERATURE_C] = hwmon_temp,
+	[VOLTAGE] = hwmon_in,
+	[CURRENT] = hwmon_curr,
+	[POWER] = hwmon_power,
+	[ENERGY] = hwmon_energy,
+};
+
+static u32 hwmon_attributes[] = {
+	[hwmon_chip] = HWMON_C_REGISTER_TZ,
+	[hwmon_temp] = HWMON_T_INPUT | HWMON_T_LABEL,
+	[hwmon_in] = HWMON_I_INPUT | HWMON_I_LABEL,
+	[hwmon_curr] = HWMON_C_INPUT | HWMON_C_LABEL,
+	[hwmon_power] = HWMON_P_INPUT | HWMON_P_LABEL,
+	[hwmon_energy] = HWMON_E_INPUT | HWMON_E_LABEL,
+};
+
+static int scmi_hwmon_probe(struct scmi_device *sdev)
+{
+	int i, idx;
+	u16 nr_sensors;
+	enum hwmon_sensor_types type;
+	struct scmi_sensors *scmi_sensors;
+	const struct scmi_sensor_info *sensor;
+	int nr_count[hwmon_max] = {0}, nr_types = 0;
+	const struct hwmon_chip_info *chip_info;
+	struct device *hwdev, *dev = &sdev->dev;
+	struct hwmon_channel_info *scmi_hwmon_chan;
+	const struct hwmon_channel_info **ptr_scmi_ci;
+	const struct scmi_handle *handle = sdev->handle;
+
+	if (!handle || !handle->sensor_ops)
+		return -ENODEV;
+
+	nr_sensors = handle->sensor_ops->count_get(handle);
+	if (!nr_sensors)
+		return -EIO;
+
+	scmi_sensors = devm_kzalloc(dev, sizeof(*scmi_sensors), GFP_KERNEL);
+	if (!scmi_sensors)
+		return -ENOMEM;
+
+	scmi_sensors->handle = handle;
+
+	for (i = 0; i < nr_sensors; i++) {
+		sensor = handle->sensor_ops->info_get(handle, i);
+		if (!sensor)
+			return PTR_ERR(sensor);
+
+		switch (sensor->type) {
+		case TEMPERATURE_C:
+		case VOLTAGE:
+		case CURRENT:
+		case POWER:
+		case ENERGY:
+			type = scmi_types[sensor->type];
+			if (!nr_count[type])
+				nr_types++;
+			nr_count[type]++;
+			break;
+		}
+	}
+
+	if (nr_count[hwmon_temp])
+		nr_count[hwmon_chip]++, nr_types++;
+
+	scmi_hwmon_chan = devm_kcalloc(dev, nr_types, sizeof(*scmi_hwmon_chan),
+				       GFP_KERNEL);
+	if (!scmi_hwmon_chan)
+		return -ENOMEM;
+
+	ptr_scmi_ci = devm_kcalloc(dev, nr_types + 1, sizeof(*ptr_scmi_ci),
+				   GFP_KERNEL);
+	if (!ptr_scmi_ci)
+		return -ENOMEM;
+
+	scmi_chip_info.info = ptr_scmi_ci;
+	chip_info = &scmi_chip_info;
+
+	for (type = 0; type < hwmon_max && nr_count[type]; type++) {
+		scmi_hwmon_add_chan_info(scmi_hwmon_chan, dev, nr_count[type],
+					 type, hwmon_attributes[type]);
+		*ptr_scmi_ci++ = scmi_hwmon_chan++;
+
+		scmi_sensors->info[type] =
+			devm_kcalloc(dev, nr_count[type],
+				     sizeof(*scmi_sensors->info), GFP_KERNEL);
+		if (!scmi_sensors->info[type])
+			return -ENOMEM;
+	}
+
+	for (i = nr_sensors - 1; i >= 0 ; i--) {
+		sensor = handle->sensor_ops->info_get(handle, i);
+		if (!sensor)
+			continue;
+
+		switch (sensor->type) {
+		case TEMPERATURE_C:
+		case VOLTAGE:
+		case CURRENT:
+		case POWER:
+		case ENERGY:
+			type = scmi_types[sensor->type];
+			idx = --nr_count[type];
+			*(scmi_sensors->info[type] + idx) = sensor;
+			break;
+		}
+	}
+
+	hwdev = devm_hwmon_device_register_with_info(dev, "scmi_sensors",
+						     scmi_sensors, chip_info,
+						     NULL);
+
+	return PTR_ERR_OR_ZERO(hwdev);
+}
+
+static const struct scmi_device_id scmi_id_table[] = {
+	{ SCMI_PROTOCOL_SENSOR },
+	{ },
+};
+MODULE_DEVICE_TABLE(scmi, scmi_id_table);
+
+static struct scmi_driver scmi_hwmon_drv = {
+	.name		= "scmi-hwmon",
+	.probe		= scmi_hwmon_probe,
+	.id_table	= scmi_id_table,
+};
+module_scmi_driver(scmi_hwmon_drv);
+
+MODULE_AUTHOR("Sudeep Holla <sudeep.holla@arm.com>");
+MODULE_DESCRIPTION("ARM SCMI HWMON interface driver");
+MODULE_LICENSE("GPL v2");
-- 
2.7.4

[PATCH v4 19/20] cpufreq: add support for CPU DVFS based on SCMI message protocol

[Sudeep Holla]

On some ARM based systems, a separate Cortex-M based System Control
Processor(SCP) provides the overall power, clock, reset and system
control including CPU DVFS. SCMI Message Protocol is used to
communicate with the SCP.

This patch adds a cpufreq driver for such systems using SCMI interface
to drive CPU DVFS.

Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: linux-pm@vger.kernel.org
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 MAINTAINERS                    |   2 +-
 drivers/cpufreq/Kconfig.arm    |  11 ++
 drivers/cpufreq/Makefile       |   1 +
 drivers/cpufreq/scmi-cpufreq.c | 274 +++++++++++++++++++++++++++++++++++++++++
 4 files changed, 287 insertions(+), 1 deletion(-)
 create mode 100644 drivers/cpufreq/scmi-cpufreq.c

diff --git a/MAINTAINERS b/MAINTAINERS
index 1f9004d68ad4..f0a4251461a7 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -12955,7 +12955,7 @@ L:	linux-arm-kernel@lists.infradead.org
 S:	Maintained
 F:	Documentation/devicetree/bindings/arm/arm,sc[mp]i.txt
 F:	drivers/clk/clk-sc[mp]i.c
-F:	drivers/cpufreq/scpi-cpufreq.c
+F:	drivers/cpufreq/sc[mp]i-cpufreq.c
 F:	drivers/firmware/arm_scpi.c
 F:	drivers/firmware/arm_scmi/
 F:	include/linux/sc[mp]i_protocol.h
diff --git a/drivers/cpufreq/Kconfig.arm b/drivers/cpufreq/Kconfig.arm
index bdce4488ded1..e21f84cbd9b4 100644
--- a/drivers/cpufreq/Kconfig.arm
+++ b/drivers/cpufreq/Kconfig.arm
@@ -205,6 +205,17 @@ config ARM_SA1100_CPUFREQ
 config ARM_SA1110_CPUFREQ
 	bool
 
+config ARM_SCMI_CPUFREQ
+	tristate "SCMI based CPUfreq driver"
+	depends on ARM_SCMI_PROTOCOL || COMPILE_TEST
+	select PM_OPP
+	help
+	  This adds the CPUfreq driver support for ARM platforms using SCMI
+	  protocol for CPU power management.
+
+	  This driver uses SCMI Message Protocol driver to interact with the
+	  firmware providing the CPU DVFS functionality.
+
 config ARM_SCPI_CPUFREQ
         tristate "SCPI based CPUfreq driver"
 	depends on ARM_BIG_LITTLE_CPUFREQ && ARM_SCPI_PROTOCOL && COMMON_CLK_SCPI
diff --git a/drivers/cpufreq/Makefile b/drivers/cpufreq/Makefile
index c7af9b2a255e..1206207f9b62 100644
--- a/drivers/cpufreq/Makefile
+++ b/drivers/cpufreq/Makefile
@@ -71,6 +71,7 @@ obj-$(CONFIG_ARM_S3C64XX_CPUFREQ)	+= s3c64xx-cpufreq.o
 obj-$(CONFIG_ARM_S5PV210_CPUFREQ)	+= s5pv210-cpufreq.o
 obj-$(CONFIG_ARM_SA1100_CPUFREQ)	+= sa1100-cpufreq.o
 obj-$(CONFIG_ARM_SA1110_CPUFREQ)	+= sa1110-cpufreq.o
+obj-$(CONFIG_ARM_SCMI_CPUFREQ)		+= scmi-cpufreq.o
 obj-$(CONFIG_ARM_SCPI_CPUFREQ)		+= scpi-cpufreq.o
 obj-$(CONFIG_ARM_SPEAR_CPUFREQ)		+= spear-cpufreq.o
 obj-$(CONFIG_ARM_STI_CPUFREQ)		+= sti-cpufreq.o
diff --git a/drivers/cpufreq/scmi-cpufreq.c b/drivers/cpufreq/scmi-cpufreq.c
new file mode 100644
index 000000000000..b1057a13e6a7
--- /dev/null
+++ b/drivers/cpufreq/scmi-cpufreq.c
@@ -0,0 +1,274 @@
+/*
+ * System Control and Power Interface (SCMI) based CPUFreq Interface driver
+ *
+ * Copyright (C) 2017 ARM Ltd.
+ * Sudeep Holla <sudeep.holla@arm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/cpu.h>
+#include <linux/cpufreq.h>
+#include <linux/cpumask.h>
+#include <linux/cpu_cooling.h>
+#include <linux/export.h>
+#include <linux/module.h>
+#include <linux/pm_opp.h>
+#include <linux/slab.h>
+#include <linux/scmi_protocol.h>
+#include <linux/types.h>
+
+struct scmi_data {
+	int domain_id;
+	struct device *cpu_dev;
+	struct thermal_cooling_device *cdev;
+};
+
+static const struct scmi_handle *handle;
+
+unsigned int scmi_cpufreq_get_rate(unsigned int cpu)
+{
+	struct cpufreq_policy *policy = cpufreq_cpu_get_raw(cpu);
+	struct scmi_perf_ops *perf_ops = handle->perf_ops;
+	struct scmi_data *priv = policy->driver_data;
+	unsigned long rate;
+	int ret;
+
+	ret = perf_ops->freq_get(handle, priv->domain_id, &rate, false);
+	if (ret)
+		return 0;
+	return rate / 1000;
+}
+
+/*
+ * perf_ops->freq_set is not a synchronous, the actual OPP change will
+ * happen asynchronously and can get notified if the events are
+ * subscribed for by the SCMI firmware
+ */
+static int
+scmi_cpufreq_set_target(struct cpufreq_policy *policy, unsigned int index)
+{
+	struct scmi_data *priv = policy->driver_data;
+	struct scmi_perf_ops *perf_ops = handle->perf_ops;
+	u64 freq = policy->freq_table[index].frequency * 1000;
+
+	return perf_ops->freq_set(handle, priv->domain_id, freq, false);
+}
+
+static int
+scmi_get_sharing_cpus(struct device *cpu_dev, struct cpumask *cpumask)
+{
+	int cpu, domain, tdomain;
+	struct device *tcpu_dev;
+
+	domain = handle->perf_ops->device_domain_id(cpu_dev);
+	if (domain < 0)
+		return domain;
+
+	for_each_possible_cpu(cpu) {
+		if (cpu == cpu_dev->id)
+			continue;
+
+		tcpu_dev = get_cpu_device(cpu);
+		if (!tcpu_dev)
+			continue;
+
+		tdomain = handle->perf_ops->device_domain_id(tcpu_dev);
+		if (tdomain == domain)
+			cpumask_set_cpu(cpu, cpumask);
+	}
+
+	return 0;
+}
+
+static int scmi_cpufreq_init(struct cpufreq_policy *policy)
+{
+	int ret;
+	unsigned int latency;
+	struct device *cpu_dev;
+	struct scmi_data *priv;
+	struct cpufreq_frequency_table *freq_table;
+
+	cpu_dev = get_cpu_device(policy->cpu);
+	if (!cpu_dev) {
+		pr_err("failed to get cpu%d device\n", policy->cpu);
+		return -ENODEV;
+	}
+
+	ret = handle->perf_ops->add_opps_to_device(handle, cpu_dev);
+	if (ret) {
+		dev_warn(cpu_dev, "failed to add opps to the device\n");
+		return ret;
+	}
+
+	ret = scmi_get_sharing_cpus(cpu_dev, policy->cpus);
+	if (ret) {
+		dev_warn(cpu_dev, "failed to get sharing cpumask\n");
+		return ret;
+	}
+
+	ret = dev_pm_opp_set_sharing_cpus(cpu_dev, policy->cpus);
+	if (ret) {
+		dev_err(cpu_dev, "%s: failed to mark OPPs as shared: %d\n",
+			__func__, ret);
+		return ret;
+	}
+
+	/*
+	 * We need OPP table to function and since OPPs are added by the
+	 * platform, let's wait for the same.
+	 */
+	ret = dev_pm_opp_get_opp_count(cpu_dev);
+	if (ret <= 0) {
+		dev_dbg(cpu_dev, "OPP table is not ready, deferring probe\n");
+		ret = -EPROBE_DEFER;
+		goto out_free_opp;
+	}
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv) {
+		ret = -ENOMEM;
+		goto out_free_opp;
+	}
+
+	ret = dev_pm_opp_init_cpufreq_table(cpu_dev, &freq_table);
+	if (ret) {
+		dev_err(cpu_dev, "failed to init cpufreq table: %d\n", ret);
+		goto out_free_priv;
+	}
+
+	priv->cpu_dev = cpu_dev;
+	priv->domain_id = handle->perf_ops->device_domain_id(cpu_dev);
+
+	policy->driver_data = priv;
+
+	ret = cpufreq_table_validate_and_show(policy, freq_table);
+	if (ret) {
+		dev_err(cpu_dev, "%s: invalid frequency table: %d\n", __func__,
+			ret);
+		goto out_free_cpufreq_table;
+	}
+
+	/* SCMI allows DVFS request for any domain from any CPU */
+	policy->dvfs_possible_from_any_cpu = true;
+
+	latency = handle->perf_ops->get_transition_latency(handle, cpu_dev);
+	if (!latency)
+		latency = CPUFREQ_ETERNAL;
+
+	policy->cpuinfo.transition_latency = latency;
+
+	return 0;
+
+out_free_cpufreq_table:
+	dev_pm_opp_free_cpufreq_table(cpu_dev, &freq_table);
+out_free_priv:
+	kfree(priv);
+out_free_opp:
+	dev_pm_opp_cpumask_remove_table(policy->cpus);
+
+	return ret;
+}
+
+static int scmi_cpufreq_exit(struct cpufreq_policy *policy)
+{
+	struct scmi_data *priv = policy->driver_data;
+
+	cpufreq_cooling_unregister(priv->cdev);
+	dev_pm_opp_free_cpufreq_table(priv->cpu_dev, &policy->freq_table);
+	kfree(priv);
+	dev_pm_opp_cpumask_remove_table(policy->related_cpus);
+
+	return 0;
+}
+
+static void scmi_cpufreq_ready(struct cpufreq_policy *policy)
+{
+	struct scmi_data *priv = policy->driver_data;
+	struct device_node *np = of_node_get(priv->cpu_dev->of_node);
+
+	if (WARN_ON(!np))
+		return;
+
+	if (of_find_property(np, "#cooling-cells", NULL)) {
+		u32 pcoeff = 0;
+
+		of_property_read_u32(np, "dynamic-power-coefficient",
+				     &pcoeff);
+
+		priv->cdev = of_cpufreq_power_cooling_register(np, policy,
+							       pcoeff, NULL);
+		if (IS_ERR(priv->cdev)) {
+			dev_err(priv->cpu_dev,
+				"running cpufreq without cooling device: %ld\n",
+				PTR_ERR(priv->cdev));
+
+			priv->cdev = NULL;
+		}
+	}
+
+	of_node_put(np);
+}
+
+static struct cpufreq_driver scmi_cpufreq_driver = {
+	.name	= "scmi",
+	.flags	= CPUFREQ_STICKY | CPUFREQ_HAVE_GOVERNOR_PER_POLICY |
+		  CPUFREQ_NEED_INITIAL_FREQ_CHECK,
+	.verify	= cpufreq_generic_frequency_table_verify,
+	.attr	= cpufreq_generic_attr,
+	.target_index	= scmi_cpufreq_set_target,
+	.get	= scmi_cpufreq_get_rate,
+	.init	= scmi_cpufreq_init,
+	.exit	= scmi_cpufreq_exit,
+	.ready	= scmi_cpufreq_ready,
+};
+
+static int scmi_cpufreq_probe(struct scmi_device *sdev)
+{
+	int ret;
+
+	handle = sdev->handle;
+
+	if (!handle || !handle->perf_ops)
+		return -ENODEV;
+
+	ret = cpufreq_register_driver(&scmi_cpufreq_driver);
+	if (ret) {
+		dev_err(&sdev->dev, "%s: registering cpufreq failed, err: %d\n",
+			__func__, ret);
+	}
+
+	return ret;
+}
+
+static void scmi_cpufreq_remove(struct scmi_device *sdev)
+{
+	cpufreq_unregister_driver(&scmi_cpufreq_driver);
+}
+
+static const struct scmi_device_id scmi_id_table[] = {
+	{ SCMI_PROTOCOL_PERF },
+	{ },
+};
+MODULE_DEVICE_TABLE(scmi, scmi_id_table);
+
+static struct scmi_driver scmi_cpufreq_drv = {
+	.name		= "scmi-cpufreq",
+	.probe		= scmi_cpufreq_probe,
+	.remove		= scmi_cpufreq_remove,
+	.id_table	= scmi_id_table,
+};
+module_scmi_driver(scmi_cpufreq_drv);
+
+MODULE_AUTHOR("Sudeep Holla <sudeep.holla@arm.com>");
+MODULE_DESCRIPTION("ARM SCMI CPUFreq interface driver");
+MODULE_LICENSE("GPL v2");
-- 
2.7.4

[PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Sudeep Holla]

The cpufreq core provides option for drivers to implement fast_switch
callback which is invoked for frequency switching from interrupt context.

This patch adds support for fast_switch callback in SCMI cpufreq driver
by making use of polling based SCMI transfer. It also sets the flag
fast_switch_possible.

Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: linux-pm@vger.kernel.org
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 drivers/cpufreq/scmi-cpufreq.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/drivers/cpufreq/scmi-cpufreq.c b/drivers/cpufreq/scmi-cpufreq.c
index b1057a13e6a7..91df5013f7b2 100644
--- a/drivers/cpufreq/scmi-cpufreq.c
+++ b/drivers/cpufreq/scmi-cpufreq.c
@@ -64,6 +64,19 @@ scmi_cpufreq_set_target(struct cpufreq_policy *policy, unsigned int index)
 	return perf_ops->freq_set(handle, priv->domain_id, freq, false);
 }
 
+static unsigned int scmi_cpufreq_fast_switch(struct cpufreq_policy *policy,
+					     unsigned int target_freq)
+{
+	struct scmi_data *priv = policy->driver_data;
+	struct scmi_perf_ops *perf_ops = handle->perf_ops;
+
+	if (!perf_ops->freq_set(handle, priv->domain_id,
+				target_freq * 1000, true))
+		return target_freq;
+
+	return 0;
+}
+
 static int
 scmi_get_sharing_cpus(struct device *cpu_dev, struct cpumask *cpumask)
 {
@@ -167,6 +180,7 @@ static int scmi_cpufreq_init(struct cpufreq_policy *policy)
 
 	policy->cpuinfo.transition_latency = latency;
 
+	policy->fast_switch_possible = true;
 	return 0;
 
 out_free_cpufreq_table:
@@ -183,6 +197,7 @@ static int scmi_cpufreq_exit(struct cpufreq_policy *policy)
 {
 	struct scmi_data *priv = policy->driver_data;
 
+	policy->fast_switch_possible = false;
 	cpufreq_cooling_unregister(priv->cdev);
 	dev_pm_opp_free_cpufreq_table(priv->cpu_dev, &policy->freq_table);
 	kfree(priv);
@@ -226,6 +241,7 @@ static struct cpufreq_driver scmi_cpufreq_driver = {
 	.verify	= cpufreq_generic_frequency_table_verify,
 	.attr	= cpufreq_generic_attr,
 	.target_index	= scmi_cpufreq_set_target,
+	.fast_switch	= scmi_cpufreq_fast_switch,
 	.get	= scmi_cpufreq_get_rate,
 	.init	= scmi_cpufreq_init,
 	.exit	= scmi_cpufreq_exit,
-- 
2.7.4

Re: [PATCH v4 15/20] firmware: arm_scmi: add device power domain support using genpd

[Sudeep Holla]

Hi Ulf,

On 03/11/17 14:47, Sudeep Holla wrote:
> This patch hooks up the support for device power domain provided by
> SCMI using the Linux generic power domain infrastructure.
> 
> Cc: Kevin Hilman <khilman@baylibre.com>
> Cc: Ulf Hansson <ulf.hansson@linaro.org>

Sorry for not adding you Reviewed-by tag along with dev_warn suggestion
you gave. I seem to have lost that change locally somehow. I have fixed
it now.

-- 
Regards,
Sudeep

Re: [PATCH v4 03/20] firmware: arm_scmi: add basic driver infrastructure for SCMI

[Jassi Brar]

On Fri, Nov 3, 2017 at 8:17 PM, Sudeep Holla <sudeep.holla@arm.com> wrote:

.....

> +int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
> +{
> +       int ret;
> +       int timeout;
> +       struct scmi_info *info = handle_to_scmi_info(handle);
> +       struct device *dev = info->dev;
> +
> +       ret = mbox_send_message(info->tx_chan, xfer);
>                                                                           ^^^^^^^^
The call still remains unchanged and broken.
https://lkml.org/lkml/2017/7/24/502

Re: [PATCH v4 02/20] dt-bindings: arm: add support for ARM System Control and Management Interface(SCMI) protocol

[Rob Herring]

On Fri, Nov 03, 2017 at 02:47:39PM +0000, Sudeep Holla wrote:
> This patch adds devicetree binding for System Control and Management
> Interface (SCMI) Message Protocol used between the Application Cores(AP)
> and the System Control Processor(SCP). The MHU peripheral provides a
> mechanism for inter-processor communication between SCP's M3 processor
> and AP.
> 
> SCP offers control and management of the core/cluster power states,
> various power domain DVFS including the core/cluster, certain system
> clocks configuration, thermal sensors and many others.
> 
> SCMI protocol is developed as better replacement to the existing SCPI
> which is not flexible and easily extensible.
> 
> Cc: Rob Herring <robh+dt@kernel.org>
> Cc: Mark Rutland <mark.rutland@arm.com>
> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
> ---
>  Documentation/devicetree/bindings/arm/arm,scmi.txt | 179 +++++++++++++++++++++
>  MAINTAINERS                                        |   4 +-
>  2 files changed, 181 insertions(+), 2 deletions(-)
>  create mode 100644 Documentation/devicetree/bindings/arm/arm,scmi.txt

Acked-by: Rob Herring <robh@kernel.org>

Re: [PATCH v4 19/20] cpufreq: add support for CPU DVFS based on SCMI message protocol

[Viresh Kumar]

On 03-11-17, 14:47, Sudeep Holla wrote:
> On some ARM based systems, a separate Cortex-M based System Control
> Processor(SCP) provides the overall power, clock, reset and system
> control including CPU DVFS. SCMI Message Protocol is used to
> communicate with the SCP.
> 
> This patch adds a cpufreq driver for such systems using SCMI interface
> to drive CPU DVFS.
> 
> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
> Cc: Viresh Kumar <viresh.kumar@linaro.org>
> Cc: linux-pm@vger.kernel.org
> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
> ---
>  MAINTAINERS                    |   2 +-
>  drivers/cpufreq/Kconfig.arm    |  11 ++
>  drivers/cpufreq/Makefile       |   1 +
>  drivers/cpufreq/scmi-cpufreq.c | 274 +++++++++++++++++++++++++++++++++++++++++
>  4 files changed, 287 insertions(+), 1 deletion(-)
>  create mode 100644 drivers/cpufreq/scmi-cpufreq.c

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

-- 
viresh

Re: [PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Viresh Kumar]

On 03-11-17, 14:47, Sudeep Holla wrote:
> The cpufreq core provides option for drivers to implement fast_switch
> callback which is invoked for frequency switching from interrupt context.
> 
> This patch adds support for fast_switch callback in SCMI cpufreq driver
> by making use of polling based SCMI transfer. It also sets the flag
> fast_switch_possible.
> 
> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
> Cc: Viresh Kumar <viresh.kumar@linaro.org>
> Cc: linux-pm@vger.kernel.org
> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
> ---
>  drivers/cpufreq/scmi-cpufreq.c | 16 ++++++++++++++++
>  1 file changed, 16 insertions(+)
> 
> diff --git a/drivers/cpufreq/scmi-cpufreq.c b/drivers/cpufreq/scmi-cpufreq.c
> index b1057a13e6a7..91df5013f7b2 100644
> --- a/drivers/cpufreq/scmi-cpufreq.c
> +++ b/drivers/cpufreq/scmi-cpufreq.c
> @@ -64,6 +64,19 @@ scmi_cpufreq_set_target(struct cpufreq_policy *policy, unsigned int index)
>  	return perf_ops->freq_set(handle, priv->domain_id, freq, false);
>  }
>  
> +static unsigned int scmi_cpufreq_fast_switch(struct cpufreq_policy *policy,
> +					     unsigned int target_freq)
> +{
> +	struct scmi_data *priv = policy->driver_data;
> +	struct scmi_perf_ops *perf_ops = handle->perf_ops;
> +
> +	if (!perf_ops->freq_set(handle, priv->domain_id,
> +				target_freq * 1000, true))
> +		return target_freq;
> +
> +	return 0;
> +}
> +
>  static int
>  scmi_get_sharing_cpus(struct device *cpu_dev, struct cpumask *cpumask)
>  {
> @@ -167,6 +180,7 @@ static int scmi_cpufreq_init(struct cpufreq_policy *policy)
>  
>  	policy->cpuinfo.transition_latency = latency;
>  
> +	policy->fast_switch_possible = true;
>  	return 0;
>  
>  out_free_cpufreq_table:
> @@ -183,6 +197,7 @@ static int scmi_cpufreq_exit(struct cpufreq_policy *policy)
>  {
>  	struct scmi_data *priv = policy->driver_data;
>  
> +	policy->fast_switch_possible = false;

You don't need this.

>  	cpufreq_cooling_unregister(priv->cdev);
>  	dev_pm_opp_free_cpufreq_table(priv->cpu_dev, &policy->freq_table);
>  	kfree(priv);
> @@ -226,6 +241,7 @@ static struct cpufreq_driver scmi_cpufreq_driver = {
>  	.verify	= cpufreq_generic_frequency_table_verify,
>  	.attr	= cpufreq_generic_attr,
>  	.target_index	= scmi_cpufreq_set_target,
> +	.fast_switch	= scmi_cpufreq_fast_switch,
>  	.get	= scmi_cpufreq_get_rate,
>  	.init	= scmi_cpufreq_init,
>  	.exit	= scmi_cpufreq_exit,

Apart from that.

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

-- 
viresh

Re: [PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Sudeep Holla]

On 07/11/17 06:19, Viresh Kumar wrote:
> On 03-11-17, 14:47, Sudeep Holla wrote:
>> The cpufreq core provides option for drivers to implement fast_switch
>> callback which is invoked for frequency switching from interrupt context.
>>
>> This patch adds support for fast_switch callback in SCMI cpufreq driver
>> by making use of polling based SCMI transfer. It also sets the flag
>> fast_switch_possible.
>>
>> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
>> Cc: Viresh Kumar <viresh.kumar@linaro.org>
>> Cc: linux-pm@vger.kernel.org
>> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
>> ---
>>  drivers/cpufreq/scmi-cpufreq.c | 16 ++++++++++++++++
>>  1 file changed, 16 insertions(+)
>>
>> diff --git a/drivers/cpufreq/scmi-cpufreq.c b/drivers/cpufreq/scmi-cpufreq.c
>> index b1057a13e6a7..91df5013f7b2 100644
>> --- a/drivers/cpufreq/scmi-cpufreq.c
>> +++ b/drivers/cpufreq/scmi-cpufreq.c
>> @@ -64,6 +64,19 @@ scmi_cpufreq_set_target(struct cpufreq_policy *policy, unsigned int index)
>>  	return perf_ops->freq_set(handle, priv->domain_id, freq, false);
>>  }
>>  
>> +static unsigned int scmi_cpufreq_fast_switch(struct cpufreq_policy *policy,
>> +					     unsigned int target_freq)
>> +{
>> +	struct scmi_data *priv = policy->driver_data;
>> +	struct scmi_perf_ops *perf_ops = handle->perf_ops;
>> +
>> +	if (!perf_ops->freq_set(handle, priv->domain_id,
>> +				target_freq * 1000, true))
>> +		return target_freq;
>> +
>> +	return 0;
>> +}
>> +
>>  static int
>>  scmi_get_sharing_cpus(struct device *cpu_dev, struct cpumask *cpumask)
>>  {
>> @@ -167,6 +180,7 @@ static int scmi_cpufreq_init(struct cpufreq_policy *policy)
>>  
>>  	policy->cpuinfo.transition_latency = latency;
>>  
>> +	policy->fast_switch_possible = true;
>>  	return 0;
>>  
>>  out_free_cpufreq_table:
>> @@ -183,6 +197,7 @@ static int scmi_cpufreq_exit(struct cpufreq_policy *policy)
>>  {
>>  	struct scmi_data *priv = policy->driver_data;
>>  
>> +	policy->fast_switch_possible = false;
> 
> You don't need this.
> 

OK, dropped now.

>>  	cpufreq_cooling_unregister(priv->cdev);
>>  	dev_pm_opp_free_cpufreq_table(priv->cpu_dev, &policy->freq_table);
>>  	kfree(priv);
>> @@ -226,6 +241,7 @@ static struct cpufreq_driver scmi_cpufreq_driver = {
>>  	.verify	= cpufreq_generic_frequency_table_verify,
>>  	.attr	= cpufreq_generic_attr,
>>  	.target_index	= scmi_cpufreq_set_target,
>> +	.fast_switch	= scmi_cpufreq_fast_switch,
>>  	.get	= scmi_cpufreq_get_rate,
>>  	.init	= scmi_cpufreq_init,
>>  	.exit	= scmi_cpufreq_exit,
> 
> Apart from that.
> 
> Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
> 

Thanks

-- 
Regards,
Sudeep

Re: [PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Rafael J. Wysocki]

On Friday, November 3, 2017 3:47:57 PM CET Sudeep Holla wrote:
> The cpufreq core provides option for drivers to implement fast_switch
> callback which is invoked for frequency switching from interrupt context.
> 
> This patch adds support for fast_switch callback in SCMI cpufreq driver
> by making use of polling based SCMI transfer. It also sets the flag
> fast_switch_possible.
> 
> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
> Cc: Viresh Kumar <viresh.kumar@linaro.org>
> Cc: linux-pm@vger.kernel.org
> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>

I'm assuming that this material will go in along with the rest of the series
through a different tree.

Thanks,
Rafael

Re: [PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Sudeep Holla]

On 08/11/17 00:24, Rafael J. Wysocki wrote:
> On Friday, November 3, 2017 3:47:57 PM CET Sudeep Holla wrote:
>> The cpufreq core provides option for drivers to implement fast_switch
>> callback which is invoked for frequency switching from interrupt context.
>>
>> This patch adds support for fast_switch callback in SCMI cpufreq driver
>> by making use of polling based SCMI transfer. It also sets the flag
>> fast_switch_possible.
>>
>> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
>> Cc: Viresh Kumar <viresh.kumar@linaro.org>
>> Cc: linux-pm@vger.kernel.org
>> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
> 
> I'm assuming that this material will go in along with the rest of the series
> through a different tree.
> 

Indeed, I will ask ARM-SoC guys to take it via their tree when ready. I
assume you are fine with that, please provide ack if that's still the case.

-- 
Regards,
Sudeep

Re: [PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Arnd Bergmann]

On Wed, Nov 8, 2017 at 11:42 AM, Sudeep Holla <sudeep.holla@arm.com> wrote:
>
>
> On 08/11/17 00:24, Rafael J. Wysocki wrote:
>> On Friday, November 3, 2017 3:47:57 PM CET Sudeep Holla wrote:
>>> The cpufreq core provides option for drivers to implement fast_switch
>>> callback which is invoked for frequency switching from interrupt context.
>>>
>>> This patch adds support for fast_switch callback in SCMI cpufreq driver
>>> by making use of polling based SCMI transfer. It also sets the flag
>>> fast_switch_possible.
>>>
>>> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
>>> Cc: Viresh Kumar <viresh.kumar@linaro.org>
>>> Cc: linux-pm@vger.kernel.org
>>> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
>>
>> I'm assuming that this material will go in along with the rest of the series
>> through a different tree.
>>
>
> Indeed, I will ask ARM-SoC guys to take it via their tree when ready. I
> assume you are fine with that, please provide ack if that's still the case.

Yes, that makes sense.

      Arnd

Re: [PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Rafael J. Wysocki]

On Wed, Nov 8, 2017 at 11:42 AM, Sudeep Holla <sudeep.holla@arm.com> wrote:
>
>
> On 08/11/17 00:24, Rafael J. Wysocki wrote:
>> On Friday, November 3, 2017 3:47:57 PM CET Sudeep Holla wrote:
>>> The cpufreq core provides option for drivers to implement fast_switch
>>> callback which is invoked for frequency switching from interrupt context.
>>>
>>> This patch adds support for fast_switch callback in SCMI cpufreq driver
>>> by making use of polling based SCMI transfer. It also sets the flag
>>> fast_switch_possible.
>>>
>>> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
>>> Cc: Viresh Kumar <viresh.kumar@linaro.org>
>>> Cc: linux-pm@vger.kernel.org
>>> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
>>
>> I'm assuming that this material will go in along with the rest of the series
>> through a different tree.
>>
>
> Indeed, I will ask ARM-SoC guys to take it via their tree when ready. I
> assume you are fine with that, please provide ack if that's still the case.

OK

Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

as long as the changes are fine by Viresh.

Re: [PATCH v4 20/20] cpufreq: scmi: add support for fast frequency switching

[Sudeep Holla]

On 08/11/17 11:05, Arnd Bergmann wrote:
> On Wed, Nov 8, 2017 at 11:42 AM, Sudeep Holla <sudeep.holla@arm.com> wrote:
>>
>>
>> On 08/11/17 00:24, Rafael J. Wysocki wrote:
>>> On Friday, November 3, 2017 3:47:57 PM CET Sudeep Holla wrote:
>>>> The cpufreq core provides option for drivers to implement fast_switch
>>>> callback which is invoked for frequency switching from interrupt context.
>>>>
>>>> This patch adds support for fast_switch callback in SCMI cpufreq driver
>>>> by making use of polling based SCMI transfer. It also sets the flag
>>>> fast_switch_possible.
>>>>
>>>> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
>>>> Cc: Viresh Kumar <viresh.kumar@linaro.org>
>>>> Cc: linux-pm@vger.kernel.org
>>>> Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
>>>
>>> I'm assuming that this material will go in along with the rest of the series
>>> through a different tree.
>>>
>>
>> Indeed, I will ask ARM-SoC guys to take it via their tree when ready. I
>> assume you are fine with that, please provide ack if that's still the case.
> 
> Yes, that makes sense.
> 


Thanks Arnd. Thanks Rafael for the ack.

-- 
Regards,
Sudeep

Re: [PATCH v4 03/20] firmware: arm_scmi: add basic driver infrastructure for SCMI

[Sudeep Holla]

Hi Jassi,

On 04/11/17 11:51, Jassi Brar wrote:
> On Fri, Nov 3, 2017 at 8:17 PM, Sudeep Holla <sudeep.holla@arm.com> wrote:
> 
> .....
> 
>> +int scmi_do_xfer(const struct scmi_handle *handle, struct scmi_xfer *xfer)
>> +{
>> +       int ret;
>> +       int timeout;
>> +       struct scmi_info *info = handle_to_scmi_info(handle);
>> +       struct device *dev = info->dev;
>> +
>> +       ret = mbox_send_message(info->tx_chan, xfer);
>>                                                                           ^^^^^^^^
> The call still remains unchanged and broken.
> https://lkml.org/lkml/2017/7/24/502
> 

Yes, since you are in favor of abstraction on top of mailbox while Arnd,
Bjorn, and me prefer to deal with that in mailbox framework itself, I
thought I will keep that discussion separate and get the remaining parts
of SCMI reviewed.

I have not forgotten about it, just trying to keep it separate. Sorry, I
could have been more specific.

-- 
Regards,
Sudeep

Re: [PATCH v4 05/20] firmware: arm_scmi: add scmi protocol bus to enumerate protocol devices

[Sudeep Holla]

Hi Arnd,

On 03/11/17 14:47, Sudeep Holla wrote:
> The SCMI specification encompasses various protocols. However, not every
> protocol has to be present on a given platform/implementation as not
> every protocol is relevant for it.
> 
> Furthermore, the platform chooses which protocols it exposes to a given
> agent. The only protocol that must be implemented is the base protocol.
> The base protocol is used by an agent to discover which protocols are
> available to it.
> 
> In order to enumerate the discovered implemented protocols, this patch
> adds support for a separate scmi protocol bus. It also adds mechanism to
> register support for different protocols.
> 
> Cc: Arnd Bergmann <arnd@arndb.de>
Extremely sorry to both during merge window, but just wanted to check if
you can spare sometime reviewing these SCMI patches. I know it's too
early for v4.16 but with holiday period ahead, thought of checking so
that I won't miss v4.16

-- 
Regards,
Sudeep