[PATCH v13 00/10] cpuidle driver for QCOM SoCs: 8064, 8074, 8084

[PATCH v13 00/10] cpuidle driver for QCOM SoCs: 8064, 8074, 8084

[Lina Iyer]

Dependent patchsets - 
        https://lkml.org/lkml/2014/8/4/767
        http://www.spinics.net/lists/linux-arm-msm/msg10799.html
        http://www.spinics.net/lists/linux-arm-msm/msg10795.html

Changes since v12:
- Minor fixes
- Added Reviewed-by

Changes since v11:
- Address review comments on spm.c
- Commenting style fixes
- Added Reviewed-by

Changes since v10:
[ https://www.mail-archive.com/devicetree@vger.kernel.org/msg51880.html ]
- Address review comments
- Added Acked-by and Reviewed-by

Changes since v9:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11714.html ]
- Address review comments on v9

Changes since v8:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11473.html ]
- Flatten out the file structure - merge pm.c into spm.c after discussions
- Add a new function to set warm boot address, in scm-boot.c
- Support for 8064 (New)
- Tested on 8074, 8084. 8064 was tested with a WIP tree
- Address review comments from v8
- Looking into possiblility of  initializing the cpuidle device for a cpu,
only when the corresponding spm device is probed successfully.

Changes since v7:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11199.html ]
- Address review comments
- Tested on 8974 but not 8084
- WFI renamed to Standby
- Update commit text with original author and link to the downstream tree

Changes since v6:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11012.html ]
- SPM device nodes merged with existing SAW DT nodes
- SPM register information is handled within the driver
- Clean up from using 'msm' to 'qcom'
        - Shorten some enumerations as well
- Review comments from v6 addressed
- New: Support for 8084 SoC
        - Not tested. I do not have a board with this SoC, but the SPM
        configuration should be identical for WFI and SPC

Changes since v5:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10559.html ]
- Merge spm-devices.c and spm.c into one file and one patch
        - Simplify implementation of the driver.
        - Update documentation mapping the DT properties with corresponding
          SPM register information.
- Removed scm-boot changes for quad core warmboot, its has been pulled in.

Changes since v4:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10327.html ]
- Update to the v8 of ARM generic idle states patches
- Use platform device model for cpuidle-qcom
- Clean up msm-pm.c to remove unnecessary include files and functions
- Update commit text and documentation for all idle states
- Remove scm-boot relocate patch from this series, submitted earlier
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10518.html ]

Changes since v3:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10288.html ]
- Fix CONFIG_QCOM_PM Kconfig as bool
- More clean ups in spm.c and spm-devices.c
        - Removed and re-organized data structures to make initialization simple
        - Remove export of sequence flush functions
        - Updated commit text
        - Comments for use of barriers.
- Rebase on top of 3.17-rc1

Changes since v2:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10148.html ]
- Prune all the drivers to support basic WFI and power down cpuidle
  functionality. Remove debug code.
- Integrate KConfig changes into the drivers' patches.
- Use Lorenzo's ARM idle-states patches as the basis for reading cpuidle
  c-states from DT.
  [ http://marc.info/?l=linux-pm&m=140794514812383&w=2 ]
- Incorporate review comments
- Rebase on top of 3.16

Changes since v1/RFC:
[ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10065.html ]
- Remove hotplug from the patch series. Will submit it separately.
- Fix SPM drivers per the review comments
- Modify patch sequence to compile SPM drivers independent of msm-pm, so as to
  allow wfi() calls to use SPM even without SoC interface driver.

8074/8084/8064 like any ARM SoC can do architectural clock gating, that helps save
on power, but not enough of leakage power.  Leakage power of the SoC can be
further reduced by turning off power to the core. To aid this, every core (cpu
and L2) is accompanied by a Sub-system Power Manager (SPM), that can be
configured to indicate the low power mode, the core would be put into and the
SPM programs the peripheral h/w accordingly to enter low power and turn off the
power rail to the core.

The idle invocation hierarchy -

CPUIDLE
        |
        cpuidle-qcom.c [CPUIdle driver]
        |
        ------> spm.c [SPM driver]
                |
                ------> scm-boot.c [SCM interface layer]
                        |
------------------------|--------------------------
(EL)                    Secure Monitor Code
                        |
                        |
                        wfi();
------------------------|--------------------------
(HW)                    [CPU] {clock gate}
                        |
                        -----> [SPM] {statemachine}


The patchset does the following -

- Introduce the SPM driver to control power to the core
- Add device bindings for 8974, 8084, 8064 CPU SPM devices
- Add cpuidle driver for QCOM cpus, using ARM generic idle state definitions.
- Add device bindings for 8974, 8084, 8064 idle-states - WFI and SPC

Thanks,
Lina




Lina Iyer (10):
  qcom: scm: Move scm-boot files to drivers/soc/qcom/ and
    include/soc/qcom
  qcom: scm: Add SCM warmboot support for quad core SoCs
  qcom: spm: Add Subsystem Power Manager driver
  arm: dts: qcom: Add power-controller device node for 8074 Krait CPUs
  arm: dts: qcom: Add power-controller device node for 8084 Krait CPUs
  arm: dts: qcom: Update power-controller device node for 8064 Krait
    CPUs
  qcom: cpuidle: Add cpuidle driver for QCOM cpus
  arm: dts: qcom: Add idle states device nodes for 8074
  arm: dts: qcom: Add idle states device nodes for 8084
  arm: dts: qcom: Add idle state device nodes for 8064

 .../bindings/arm/msm/qcom,idle-state.txt           |  81 +++++
 .../devicetree/bindings/arm/msm/qcom,saw2.txt      |  31 +-
 arch/arm/boot/dts/qcom-apq8064.dtsi                |  36 ++-
 arch/arm/boot/dts/qcom-apq8084.dtsi                |  46 ++-
 arch/arm/boot/dts/qcom-msm8974.dtsi                |  46 ++-
 arch/arm/mach-qcom/Makefile                        |   1 -
 arch/arm/mach-qcom/platsmp.c                       |   2 +-
 drivers/cpuidle/Kconfig.arm                        |   7 +
 drivers/cpuidle/Makefile                           |   1 +
 drivers/cpuidle/cpuidle-qcom.c                     |  78 +++++
 drivers/soc/qcom/Kconfig                           |   8 +
 drivers/soc/qcom/Makefile                          |   3 +-
 .../arm/mach-qcom => drivers/soc/qcom}/scm-boot.c  |  37 ++-
 drivers/soc/qcom/spm.c                             | 327 +++++++++++++++++++++
 include/soc/qcom/pm.h                              |  31 ++
 .../arm/mach-qcom => include/soc/qcom}/scm-boot.h  |   3 +-
 16 files changed, 716 insertions(+), 22 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/msm/qcom,idle-state.txt
 create mode 100644 drivers/cpuidle/cpuidle-qcom.c
 rename {arch/arm/mach-qcom => drivers/soc/qcom}/scm-boot.c (59%)
 create mode 100644 drivers/soc/qcom/spm.c
 create mode 100644 include/soc/qcom/pm.h
 rename {arch/arm/mach-qcom => include/soc/qcom}/scm-boot.h (91%)

-- 
2.1.0

[PATCH v13 01/10] qcom: scm: Move scm-boot files to drivers/soc/qcom/ and include/soc/qcom

[Lina Iyer]

Follow the scm.c and move scm-boot files to drivers/soc/qcom. The
guidance is to clean files out from mach-qcom and move to drivers/soc
area.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 arch/arm/mach-qcom/Makefile                         | 1 -
 arch/arm/mach-qcom/platsmp.c                        | 2 +-
 drivers/soc/qcom/Makefile                           | 2 +-
 {arch/arm/mach-qcom => drivers/soc/qcom}/scm-boot.c | 4 ++--
 {arch/arm/mach-qcom => include/soc/qcom}/scm-boot.h | 0
 5 files changed, 4 insertions(+), 5 deletions(-)
 rename {arch/arm/mach-qcom => drivers/soc/qcom}/scm-boot.c (97%)
 rename {arch/arm/mach-qcom => include/soc/qcom}/scm-boot.h (100%)

diff --git a/arch/arm/mach-qcom/Makefile b/arch/arm/mach-qcom/Makefile
index db41e8c..e324375 100644
--- a/arch/arm/mach-qcom/Makefile
+++ b/arch/arm/mach-qcom/Makefile
@@ -1,3 +1,2 @@
 obj-y			:= board.o
 obj-$(CONFIG_SMP)	+= platsmp.o
-obj-$(CONFIG_QCOM_SCM)	+= scm-boot.o
diff --git a/arch/arm/mach-qcom/platsmp.c b/arch/arm/mach-qcom/platsmp.c
index d690856..a692bcb 100644
--- a/arch/arm/mach-qcom/platsmp.c
+++ b/arch/arm/mach-qcom/platsmp.c
@@ -20,7 +20,7 @@
 
 #include <asm/smp_plat.h>
 
-#include "scm-boot.h"
+#include <soc/qcom/scm-boot.h>
 
 #define VDD_SC1_ARRAY_CLAMP_GFS_CTL	0x35a0
 #define SCSS_CPU1CORE_RESET		0x2d80
diff --git a/drivers/soc/qcom/Makefile b/drivers/soc/qcom/Makefile
index a39446d..70d52ed 100644
--- a/drivers/soc/qcom/Makefile
+++ b/drivers/soc/qcom/Makefile
@@ -1,3 +1,3 @@
 obj-$(CONFIG_QCOM_GSBI)	+=	qcom_gsbi.o
 CFLAGS_scm.o :=$(call as-instr,.arch_extension sec,-DREQUIRES_SEC=1)
-obj-$(CONFIG_QCOM_SCM) += scm.o
+obj-$(CONFIG_QCOM_SCM) += scm.o scm-boot.o
diff --git a/arch/arm/mach-qcom/scm-boot.c b/drivers/soc/qcom/scm-boot.c
similarity index 97%
rename from arch/arm/mach-qcom/scm-boot.c
rename to drivers/soc/qcom/scm-boot.c
index 5add20e..60ff7b4 100644
--- a/arch/arm/mach-qcom/scm-boot.c
+++ b/drivers/soc/qcom/scm-boot.c
@@ -17,9 +17,9 @@
 
 #include <linux/module.h>
 #include <linux/slab.h>
-#include <soc/qcom/scm.h>
 
-#include "scm-boot.h"
+#include <soc/qcom/scm.h>
+#include <soc/qcom/scm-boot.h>
 
 /*
  * Set the cold/warm boot address for one of the CPU cores.
diff --git a/arch/arm/mach-qcom/scm-boot.h b/include/soc/qcom/scm-boot.h
similarity index 100%
rename from arch/arm/mach-qcom/scm-boot.h
rename to include/soc/qcom/scm-boot.h
-- 
2.1.0

[PATCH v13 02/10] qcom: scm: Add SCM warmboot support for quad core SoCs

[Lina Iyer]

Quad core SoCs like APQ8074, APQ8064, APQ8084 need SCM support set up
warm boot addresses in the Secure Monitor. Extend the SCM flags to
support warm boot addresses for secondary cores.

We do not need to export the warmboot flags. Move them into the
implementation file.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 drivers/soc/qcom/scm-boot.c | 35 +++++++++++++++++++++++++++++++++++
 include/soc/qcom/scm-boot.h |  3 +--
 2 files changed, 36 insertions(+), 2 deletions(-)

diff --git a/drivers/soc/qcom/scm-boot.c b/drivers/soc/qcom/scm-boot.c
index 60ff7b4..f653217a 100644
--- a/drivers/soc/qcom/scm-boot.c
+++ b/drivers/soc/qcom/scm-boot.c
@@ -21,6 +21,23 @@
 #include <soc/qcom/scm.h>
 #include <soc/qcom/scm-boot.h>
 
+#define SCM_FLAG_WARMBOOT_CPU0		0x04
+#define SCM_FLAG_WARMBOOT_CPU1		0x02
+#define SCM_FLAG_WARMBOOT_CPU2		0x10
+#define SCM_FLAG_WARMBOOT_CPU3		0x40
+
+struct scm_warmboot {
+	int flag;
+	void *entry;
+};
+
+static struct scm_warmboot scm_flags[] = {
+	{ .flag = SCM_FLAG_WARMBOOT_CPU0 },
+	{ .flag = SCM_FLAG_WARMBOOT_CPU1 },
+	{ .flag = SCM_FLAG_WARMBOOT_CPU2 },
+	{ .flag = SCM_FLAG_WARMBOOT_CPU3 },
+};
+
 /*
  * Set the cold/warm boot address for one of the CPU cores.
  */
@@ -37,3 +54,21 @@ int scm_set_boot_addr(phys_addr_t addr, int flags)
 			&cmd, sizeof(cmd), NULL, 0);
 }
 EXPORT_SYMBOL(scm_set_boot_addr);
+
+int scm_set_warm_boot_addr(void *entry, int cpu)
+{
+	int ret;
+
+	/*
+	 * Reassign only if we are switching from hotplug entry point
+	 * to cpuidle entry point or vice versa.
+	 */
+	if (entry == scm_flags[cpu].entry)
+		return 0;
+
+	ret = scm_set_boot_addr(virt_to_phys(entry), scm_flags[cpu].flag);
+	if (!ret)
+		scm_flags[cpu].entry  = entry;
+
+	return ret;
+}
diff --git a/include/soc/qcom/scm-boot.h b/include/soc/qcom/scm-boot.h
index 6aabb24..529f55a 100644
--- a/include/soc/qcom/scm-boot.h
+++ b/include/soc/qcom/scm-boot.h
@@ -16,9 +16,8 @@
 #define SCM_FLAG_COLDBOOT_CPU1		0x01
 #define SCM_FLAG_COLDBOOT_CPU2		0x08
 #define SCM_FLAG_COLDBOOT_CPU3		0x20
-#define SCM_FLAG_WARMBOOT_CPU0		0x04
-#define SCM_FLAG_WARMBOOT_CPU1		0x02
 
 int scm_set_boot_addr(phys_addr_t addr, int flags);
+int scm_set_warm_boot_addr(void *entry, int cpu);
 
 #endif
-- 
2.1.0

[PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Lina Iyer]

SPM is a hardware block that controls the peripheral logic surrounding
the application cores (cpu/l$). When the core executes WFI instruction,
the SPM takes over the putting the core in low power state as
configured. The wake up for the SPM is an interrupt at the GIC, which
then completes the rest of low power mode sequence and brings the core
out of low power mode.

The SPM has a set of control registers that configure the SPMs
individually based on the type of the core and the runtime conditions.
SPM is a finite state machine block to which a sequence is provided and
it interprets the bytes and executes them in sequence. Each low power
mode that the core can enter into is provided to the SPM as a sequence.

Configure the SPM to set the core (cpu or L2) into its low power mode,
the index of the first command in the sequence is set in the SPM_CTL
register. When the core executes ARM wfi instruction, it triggers the
SPM state machine to start executing from that index. The SPM state
machine waits until the interrupt occurs and starts executing the rest
of the sequence until it hits the end of the sequence. The end of the
sequence jumps the core out of its low power mode.

Add support for an idle driver to set up the SPM to place the core in
Standby or Standalone power collapse mode when the core is idle.

Based on work by: Mahesh Sivasubramanian <msivasub@codeaurora.org>,
Ai Li <ali@codeaurora.org>, Praveen Chidambaram <pchidamb@codeaurora.org>
Original tree available at -
git://codeaurora.org/quic/la/kernel/msm-3.10.git

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
---
 .../devicetree/bindings/arm/msm/qcom,saw2.txt      |  31 +-
 drivers/soc/qcom/Kconfig                           |   8 +
 drivers/soc/qcom/Makefile                          |   1 +
 drivers/soc/qcom/spm.c                             | 327 +++++++++++++++++++++
 include/soc/qcom/pm.h                              |  31 ++
 5 files changed, 392 insertions(+), 6 deletions(-)
 create mode 100644 drivers/soc/qcom/spm.c
 create mode 100644 include/soc/qcom/pm.h

diff --git a/Documentation/devicetree/bindings/arm/msm/qcom,saw2.txt b/Documentation/devicetree/bindings/arm/msm/qcom,saw2.txt
index 1505fb8..690c3c0 100644
--- a/Documentation/devicetree/bindings/arm/msm/qcom,saw2.txt
+++ b/Documentation/devicetree/bindings/arm/msm/qcom,saw2.txt
@@ -2,11 +2,20 @@ SPM AVS Wrapper 2 (SAW2)
 
 The SAW2 is a wrapper around the Subsystem Power Manager (SPM) and the
 Adaptive Voltage Scaling (AVS) hardware. The SPM is a programmable
-micro-controller that transitions a piece of hardware (like a processor or
+power-controller that transitions a piece of hardware (like a processor or
 subsystem) into and out of low power modes via a direct connection to
 the PMIC. It can also be wired up to interact with other processors in the
 system, notifying them when a low power state is entered or exited.
 
+Multiple revisions of the SAW hardware are supported using these Device Nodes.
+SAW2 revisions differ in the register offset and configuration data. Also, the
+same revision of the SAW in different SoCs may have different configuration
+data due the the differences in hardware capabilities. Hence the SoC name, the
+version of the SAW hardware in that SoC and the distinction between cpu (big
+or Little) or cache, may be needed to uniquely identify the SAW register
+configuration and initialization data. The compatible string is used to
+indicate this parameter.
+
 PROPERTIES
 
 - compatible:
@@ -14,10 +23,13 @@ PROPERTIES
 	Value type: <string>
 	Definition: shall contain "qcom,saw2". A more specific value should be
 		    one of:
-			 "qcom,saw2-v1"
-			 "qcom,saw2-v1.1"
-			 "qcom,saw2-v2"
-			 "qcom,saw2-v2.1"
+			"qcom,saw2-v1"
+			"qcom,saw2-v1.1"
+			"qcom,saw2-v2"
+			"qcom,saw2-v2.1"
+			"qcom,apq8064-saw2-v1.1-cpu"
+			"qcom,msm8974-saw2-v2.1-cpu"
+			"qcom,apq8084-saw2-v2.1-cpu"
 
 - reg:
 	Usage: required
@@ -26,10 +38,17 @@ PROPERTIES
 		    the register region. An optional second element specifies
 		    the base address and size of the alias register region.
 
+- regulator:
+	Usage: optional
+	Value type: boolean
+	Definition: Indicates that this SPM device acts as a regulator device
+			device for the core (CPU or Cache) the SPM is attached
+			to.
 
 Example:
 
-	regulator@2099000 {
+	power-controller@2099000 {
 		compatible = "qcom,saw2";
 		reg = <0x02099000 0x1000>, <0x02009000 0x1000>;
+		regulator;
 	};
diff --git a/drivers/soc/qcom/Kconfig b/drivers/soc/qcom/Kconfig
index 7dcd554..012fb37 100644
--- a/drivers/soc/qcom/Kconfig
+++ b/drivers/soc/qcom/Kconfig
@@ -11,3 +11,11 @@ config QCOM_GSBI
 
 config QCOM_SCM
 	bool
+
+config QCOM_PM
+	bool "Qualcomm Power Management"
+	depends on ARCH_QCOM
+	help
+	  QCOM Platform specific power driver to manage cores and L2 low power
+	  modes. It interface with various system drivers to put the cores in
+	  low power modes.
diff --git a/drivers/soc/qcom/Makefile b/drivers/soc/qcom/Makefile
index 70d52ed..20b329f 100644
--- a/drivers/soc/qcom/Makefile
+++ b/drivers/soc/qcom/Makefile
@@ -1,3 +1,4 @@
 obj-$(CONFIG_QCOM_GSBI)	+=	qcom_gsbi.o
+obj-$(CONFIG_QCOM_PM)	+=	spm.o
 CFLAGS_scm.o :=$(call as-instr,.arch_extension sec,-DREQUIRES_SEC=1)
 obj-$(CONFIG_QCOM_SCM) += scm.o scm-boot.o
diff --git a/drivers/soc/qcom/spm.c b/drivers/soc/qcom/spm.c
new file mode 100644
index 0000000..0b4616a
--- /dev/null
+++ b/drivers/soc/qcom/spm.c
@@ -0,0 +1,327 @@
+/*
+ * Copyright (c) 2011-2014, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that 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.
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/slab.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/err.h>
+#include <linux/platform_device.h>
+#include <linux/cpuidle.h>
+#include <linux/cpu_pm.h>
+
+#include <asm/proc-fns.h>
+#include <asm/suspend.h>
+
+#include <soc/qcom/pm.h>
+#include <soc/qcom/pm.h>
+#include <soc/qcom/scm.h>
+#include <soc/qcom/scm-boot.h>
+
+
+#define SCM_CMD_TERMINATE_PC	0x2
+#define SCM_FLUSH_FLAG_MASK	0x3
+#define SCM_L2_ON		0x0
+#define SCM_L2_OFF		0x1
+#define MAX_PMIC_DATA		2
+#define MAX_SEQ_DATA		64
+#define SPM_CTL_INDEX		0x7f
+#define SPM_CTL_INDEX_SHIFT	4
+#define SPM_CTL_EN		BIT(0)
+
+enum spm_reg {
+	SPM_REG_CFG,
+	SPM_REG_SPM_CTL,
+	SPM_REG_DLY,
+	SPM_REG_PMIC_DLY,
+	SPM_REG_PMIC_DATA_0,
+	SPM_REG_PMIC_DATA_1,
+	SPM_REG_VCTL,
+	SPM_REG_SEQ_ENTRY,
+	SPM_REG_SPM_STS,
+	SPM_REG_PMIC_STS,
+	SPM_REG_NR,
+};
+
+struct spm_reg_data {
+	const u8 *reg_offset;
+	u32 spm_cfg;
+	u32 spm_dly;
+	u32 pmic_dly;
+	u32 pmic_data[MAX_PMIC_DATA];
+	u8 seq[MAX_SEQ_DATA];
+	u8 start_index[PM_SLEEP_MODE_NR];
+};
+
+struct spm_driver_data {
+	void __iomem *reg_base;
+	const struct spm_reg_data *reg_data;
+};
+
+static const u8 spm_reg_offset_v2_1[SPM_REG_NR] = {
+	[SPM_REG_CFG]		= 0x08,
+	[SPM_REG_SPM_CTL]	= 0x30,
+	[SPM_REG_DLY]		= 0x34,
+	[SPM_REG_SEQ_ENTRY]	= 0x80,
+};
+
+/* SPM register data for 8974, 8084 */
+static const struct spm_reg_data spm_reg_8974_8084_cpu  = {
+	.reg_offset = spm_reg_offset_v2_1,
+	.spm_cfg = 0x1,
+	.spm_dly = 0x3C102800,
+	.seq = { 0x03, 0x0B, 0x0F, 0x00, 0x20, 0x80, 0x10, 0xE8, 0x5B, 0x03,
+		0x3B, 0xE8, 0x5B, 0x82, 0x10, 0x0B, 0x30, 0x06, 0x26, 0x30,
+		0x0F },
+	.start_index[PM_SLEEP_MODE_STBY] = 0,
+	.start_index[PM_SLEEP_MODE_SPC] = 3,
+};
+
+static const u8 spm_reg_offset_v1_1[SPM_REG_NR] = {
+	[SPM_REG_CFG]		= 0x08,
+	[SPM_REG_SPM_CTL]	= 0x20,
+	[SPM_REG_PMIC_DLY]	= 0x24,
+	[SPM_REG_PMIC_DATA_0]	= 0x28,
+	[SPM_REG_PMIC_DATA_1]	= 0x2C,
+	[SPM_REG_SEQ_ENTRY]	= 0x80,
+};
+
+/* SPM register data for 8064 */
+static const struct spm_reg_data spm_reg_8064_cpu = {
+	.reg_offset = spm_reg_offset_v1_1,
+	.spm_cfg = 0x1F,
+	.pmic_dly = 0x02020004,
+	.pmic_data[0] = 0x0084009C,
+	.pmic_data[1] = 0x00A4001C,
+	.seq = { 0x03, 0x0F, 0x00, 0x24, 0x54, 0x10, 0x09, 0x03, 0x01,
+		0x10, 0x54, 0x30, 0x0C, 0x24, 0x30, 0x0F },
+	.start_index[PM_SLEEP_MODE_STBY] = 0,
+	.start_index[PM_SLEEP_MODE_SPC] = 2,
+};
+
+static DEFINE_PER_CPU_SHARED_ALIGNED(struct spm_driver_data *, cpu_spm_drv);
+
+static inline void spm_register_write(struct spm_driver_data *drv,
+					enum spm_reg reg, u32 val)
+{
+	if (drv->reg_data->reg_offset[reg])
+		writel_relaxed(val, drv->reg_base +
+				drv->reg_data->reg_offset[reg]);
+}
+
+static inline u32 spm_register_read(struct spm_driver_data *drv,
+					enum spm_reg reg)
+{
+	return readl_relaxed(drv->reg_base + drv->reg_data->reg_offset[reg]);
+}
+
+static int spm_set_low_power_mode(enum pm_sleep_mode mode)
+{
+	struct spm_driver_data *drv = per_cpu(cpu_spm_drv,
+						smp_processor_id());
+	u32 start_index;
+	u32 ctl_val;
+
+	if (!drv)
+		return -ENXIO;
+
+	start_index = drv->reg_data->start_index[mode];
+
+	ctl_val = spm_register_read(drv, SPM_REG_SPM_CTL);
+	ctl_val &= ~(SPM_CTL_INDEX << SPM_CTL_INDEX_SHIFT);
+	ctl_val |= start_index << SPM_CTL_INDEX_SHIFT;
+	ctl_val |= SPM_CTL_EN;
+	spm_register_write(drv, SPM_REG_SPM_CTL, ctl_val);
+
+	/* Ensure we have written the start address */
+	wmb();
+
+	return 0;
+}
+
+static int qcom_pm_collapse(unsigned long int unused)
+{
+	int ret;
+	u32 flag;
+	int cpu = smp_processor_id();
+
+	ret = scm_set_warm_boot_addr(cpu_resume, cpu);
+	if (ret)
+		return ret;
+
+	flag = SCM_L2_ON & SCM_FLUSH_FLAG_MASK;
+	scm_call_atomic1(SCM_SVC_BOOT, SCM_CMD_TERMINATE_PC, flag);
+
+	/*
+	 * Returns here only if there was a pending interrupt and we did not
+	 * power down as a result.
+	 */
+	return 0;
+}
+
+static int qcom_cpu_standby(void *unused)
+{
+	int ret;
+
+	ret = spm_set_low_power_mode(PM_SLEEP_MODE_STBY);
+	if (ret)
+		return ret;
+
+	cpu_do_idle();
+
+	return 0;
+}
+
+static int qcom_cpu_spc(void *unused)
+{	int ret;
+
+	ret = spm_set_low_power_mode(PM_SLEEP_MODE_STBY);
+	if (ret)
+		return ret;
+
+	cpu_pm_enter();
+	cpu_suspend(0, qcom_pm_collapse);
+	cpu_pm_exit();
+
+	return 0;
+}
+
+static struct spm_driver_data *spm_get_drv(struct platform_device *pdev,
+		int *spm_cpu)
+{
+	struct spm_driver_data *drv = NULL;
+	struct device_node *cpu_node, *saw_node;
+	int cpu;
+	bool found = false;
+
+	for_each_possible_cpu(cpu) {
+		cpu_node = of_cpu_device_node_get(cpu);
+		if (!cpu_node)
+			continue;
+		saw_node = of_parse_phandle(cpu_node, "qcom,saw", 0);
+		if (saw_node) {
+			if (saw_node == pdev->dev.of_node)
+				found = true;
+			of_node_put(saw_node);
+		}
+		of_node_put(cpu_node);
+		if (found)
+			break;
+	}
+
+	if (found) {
+		drv = devm_kzalloc(&pdev->dev, sizeof(*drv), GFP_KERNEL);
+		if (drv)
+			*spm_cpu = cpu;
+	}
+
+	return drv;
+}
+
+static const struct of_device_id spm_match_table[] = {
+	{ .compatible = "qcom,msm8974-saw2-v2.1-cpu",
+	  .data = &spm_reg_8974_8084_cpu },
+	{ .compatible = "qcom,apq8084-saw2-v2.1-cpu",
+	  .data = &spm_reg_8974_8084_cpu },
+	{ .compatible = "qcom,apq8064-saw2-v1.1-cpu",
+	  .data = &spm_reg_8064_cpu },
+	{ },
+};
+
+static const struct qcom_cpu_pm_ops lpm_ops = {
+	.standby = qcom_cpu_standby,
+	.spc = qcom_cpu_spc,
+};
+
+static int spm_dev_probe(struct platform_device *pdev)
+{
+	struct spm_driver_data *drv;
+	struct resource *res;
+	const struct of_device_id *match_id;
+	void __iomem *addr;
+	int cpu = -EINVAL;
+	static bool cpuidle_drv_init;
+	const struct platform_device_info qcom_cpuidle_info = {
+		.name	= "qcom_cpuidle",
+		.id	= -1,
+		.data = &lpm_ops,
+		.size_data = sizeof(lpm_ops),
+	};
+
+	drv = spm_get_drv(pdev, &cpu);
+	if (!drv || cpu < 0)
+		return -EINVAL;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	drv->reg_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(drv->reg_base))
+		return PTR_ERR(drv->reg_base);
+
+	match_id = of_match_node(spm_match_table, pdev->dev.of_node);
+	if (!match_id)
+		return -ENODEV;
+
+	drv->reg_data = match_id->data;
+
+	/* Write the SPM sequences first.. */
+	addr = drv->reg_base + drv->reg_data->reg_offset[SPM_REG_SEQ_ENTRY];
+	__iowrite32_copy(addr, drv->reg_data->seq,
+			ARRAY_SIZE(drv->reg_data->seq) / 4);
+
+	/*
+	 * ..and then the control registers.
+	 * On some SoC's if the control registers are written first and if the
+	 * CPU was held in reset, the reset signal could trigger the SPM state
+	 * machine, before the sequences are completely written.
+	 */
+	spm_register_write(drv, SPM_REG_CFG, drv->reg_data->spm_cfg);
+	spm_register_write(drv, SPM_REG_DLY, drv->reg_data->spm_dly);
+	spm_register_write(drv, SPM_REG_PMIC_DLY, drv->reg_data->pmic_dly);
+
+	spm_register_write(drv, SPM_REG_PMIC_DATA_0,
+				drv->reg_data->pmic_data[0]);
+	spm_register_write(drv, SPM_REG_PMIC_DATA_1,
+				drv->reg_data->pmic_data[1]);
+
+	/*
+	 * Ensure all observers see the above register writes before the
+	 * cpuidle driver is allowed to use the SPM.
+	 */
+	wmb();
+	per_cpu(cpu_spm_drv, cpu) = drv;
+
+	if (!cpuidle_drv_init) {
+		platform_device_register_full(&qcom_cpuidle_info);
+		cpuidle_drv_init = true;
+	}
+
+	return 0;
+}
+
+static struct platform_driver spm_driver = {
+	.probe = spm_dev_probe,
+	.driver = {
+		.name = "saw",
+		.of_match_table = spm_match_table,
+	},
+};
+
+module_platform_driver(spm_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("SAW power controller driver");
+MODULE_ALIAS("platform:saw");
diff --git a/include/soc/qcom/pm.h b/include/soc/qcom/pm.h
new file mode 100644
index 0000000..d9a56d7
--- /dev/null
+++ b/include/soc/qcom/pm.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright (c) 2009-2014, The Linux Foundation. All rights reserved.
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that 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.
+ *
+ */
+
+#ifndef __QCOM_PM_H
+#define __QCOM_PM_H
+
+enum pm_sleep_mode {
+	PM_SLEEP_MODE_STBY,
+	PM_SLEEP_MODE_RET,
+	PM_SLEEP_MODE_SPC,
+	PM_SLEEP_MODE_PC,
+	PM_SLEEP_MODE_NR,
+};
+
+struct qcom_cpu_pm_ops {
+	int (*standby)(void *data);
+	int (*spc)(void *data);
+};
+
+#endif  /* __QCOM_PM_H */
-- 
2.1.0

[PATCH v13 04/10] arm: dts: qcom: Add power-controller device node for 8074 Krait CPUs

[Lina Iyer]

Each Krait CPU in the QCOM 8074/8974 SoC has an SAW power controller to
regulate the power to the cpu and aide the core in entering idle states.
Reference the SAW instance and associate the instance with the CPU core.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 arch/arm/boot/dts/qcom-msm8974.dtsi | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/arch/arm/boot/dts/qcom-msm8974.dtsi b/arch/arm/boot/dts/qcom-msm8974.dtsi
index e265ec1..5a41f44 100644
--- a/arch/arm/boot/dts/qcom-msm8974.dtsi
+++ b/arch/arm/boot/dts/qcom-msm8974.dtsi
@@ -21,6 +21,7 @@
 			reg = <0>;
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc0>;
+			qcom,saw = <&saw0>;
 		};
 
 		cpu@1 {
@@ -30,6 +31,7 @@
 			reg = <1>;
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc1>;
+			qcom,saw = <&saw1>;
 		};
 
 		cpu@2 {
@@ -39,6 +41,7 @@
 			reg = <2>;
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc2>;
+			qcom,saw = <&saw2>;
 		};
 
 		cpu@3 {
@@ -48,6 +51,7 @@
 			reg = <3>;
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc3>;
+			qcom,saw = <&saw3>;
 		};
 
 		L2: l2-cache {
@@ -144,7 +148,27 @@
 			};
 		};
 
-		saw_l2: regulator@f9012000 {
+		saw0: power-controller@f9089000 {
+			compatible = "qcom,msm8974-saw2-v2.1-cpu";
+			reg = <0xf9089000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw1: power-controller@f9099000 {
+			compatible = "qcom,msm8974-saw2-v2.1-cpu";
+			reg = <0xf9099000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw2: power-controller@f90a9000 {
+			compatible = "qcom,msm8974-saw2-v2.1-cpu";
+			reg = <0xf90a9000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw3: power-controller@f90b9000 {
+			compatible = "qcom,msm8974-saw2-v2.1-cpu";
+			reg = <0xf90b9000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw_l2: power-controller@f9012000 {
 			compatible = "qcom,saw2";
 			reg = <0xf9012000 0x1000>;
 			regulator;
-- 
2.1.0

[PATCH v13 05/10] arm: dts: qcom: Add power-controller device node for 8084 Krait CPUs

[Lina Iyer]

Each Krait CPU in the QCOM 8084 SoC has an SAW power controller to
regulate the power to the cpu and aide the core in entering idle states.
Reference the SAW instance and associate the instance with the CPU core.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 arch/arm/boot/dts/qcom-apq8084.dtsi | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/arch/arm/boot/dts/qcom-apq8084.dtsi b/arch/arm/boot/dts/qcom-apq8084.dtsi
index 1f130bc..71182bf 100644
--- a/arch/arm/boot/dts/qcom-apq8084.dtsi
+++ b/arch/arm/boot/dts/qcom-apq8084.dtsi
@@ -21,6 +21,7 @@
 			enable-method = "qcom,kpss-acc-v2";
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc0>;
+			qcom,saw = <&saw0>;
 		};
 
 		cpu@1 {
@@ -30,6 +31,7 @@
 			enable-method = "qcom,kpss-acc-v2";
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc1>;
+			qcom,saw = <&saw1>;
 		};
 
 		cpu@2 {
@@ -39,6 +41,7 @@
 			enable-method = "qcom,kpss-acc-v2";
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc2>;
+			qcom,saw = <&saw2>;
 		};
 
 		cpu@3 {
@@ -48,6 +51,7 @@
 			enable-method = "qcom,kpss-acc-v2";
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc3>;
+			qcom,saw = <&saw3>;
 		};
 
 		L2: l2-cache {
@@ -144,7 +148,27 @@
 			};
 		};
 
-		saw_l2: regulator@f9012000 {
+		saw0: power-controller@f9089000 {
+			compatible = "qcom,apq8084-saw2-v2.1-cpu";
+			reg = <0xf9089000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw1: power-controller@f9099000 {
+			compatible = "qcom,apq8084-saw2-v2.1-cpu";
+			reg = <0xf9099000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw2: power-controller@f90a9000 {
+			compatible = "qcom,apq8084-saw2-v2.1-cpu";
+			reg = <0xf90a9000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw3: power-controller@f90b9000 {
+			compatible = "qcom,apq8084-saw2-v2.1-cpu";
+			reg = <0xf90b9000 0x1000>, <0xf9009000 0x1000>;
+		};
+
+		saw_l2: power-controller@f9012000 {
 			compatible = "qcom,saw2";
 			reg = <0xf9012000 0x1000>;
 			regulator;
-- 
2.1.0

[PATCH v13 06/10] arm: dts: qcom: Update power-controller device node for 8064 Krait CPUs

[Lina Iyer]

Update the SAW2 DT bindings to add qcom,apq8064-saw2-v1.1-cpu compatible
binding string to configure SPM registers and allow the SPM to put the
core in deeper idle states when the core is idle.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 arch/arm/boot/dts/qcom-apq8064.dtsi | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/arch/arm/boot/dts/qcom-apq8064.dtsi b/arch/arm/boot/dts/qcom-apq8064.dtsi
index b3154c0..9fd24bc 100644
--- a/arch/arm/boot/dts/qcom-apq8064.dtsi
+++ b/arch/arm/boot/dts/qcom-apq8064.dtsi
@@ -139,26 +139,26 @@
 			reg = <0x020b8000 0x1000>, <0x02008000 0x1000>;
 		};
 
-		saw0: regulator@2089000 {
-			compatible = "qcom,saw2";
+		saw0: power-controller@2089000 {
+			compatible = "qcom,apq8064-saw2-v1.1-cpu", "qcom,saw2";
 			reg = <0x02089000 0x1000>, <0x02009000 0x1000>;
 			regulator;
 		};
 
-		saw1: regulator@2099000 {
-			compatible = "qcom,saw2";
+		saw1: power-controller@2099000 {
+			compatible = "qcom,apq8064-saw2-v1.1-cpu", "qcom,saw2";
 			reg = <0x02099000 0x1000>, <0x02009000 0x1000>;
 			regulator;
 		};
 
-		saw2: regulator@20a9000 {
-			compatible = "qcom,saw2";
+		saw2: power-controller@20a9000 {
+			compatible = "qcom,apq8064-saw2-v1.1-cpu", "qcom,saw2";
 			reg = <0x020a9000 0x1000>, <0x02009000 0x1000>;
 			regulator;
 		};
 
-		saw3: regulator@20b9000 {
-			compatible = "qcom,saw2";
+		saw3: power-controller@20b9000 {
+			compatible = "qcom,apq8064-saw2-v1.1-cpu", "qcom,saw2";
 			reg = <0x020b9000 0x1000>, <0x02009000 0x1000>;
 			regulator;
 		};
-- 
2.1.0

[PATCH v13 07/10] qcom: cpuidle: Add cpuidle driver for QCOM cpus

[Lina Iyer]

Add cpuidle driver interface to allow cpus to go into idle states. Use
the cpuidle DT interface, common across ARM architectures, to provide
the idle state information to the cpuidle framework.

Supported modes at this time are Standby and Standalone Power Collapse.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
---
 .../bindings/arm/msm/qcom,idle-state.txt           | 81 ++++++++++++++++++++++
 drivers/cpuidle/Kconfig.arm                        |  7 ++
 drivers/cpuidle/Makefile                           |  1 +
 drivers/cpuidle/cpuidle-qcom.c                     | 78 +++++++++++++++++++++
 4 files changed, 167 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/arm/msm/qcom,idle-state.txt
 create mode 100644 drivers/cpuidle/cpuidle-qcom.c

diff --git a/Documentation/devicetree/bindings/arm/msm/qcom,idle-state.txt b/Documentation/devicetree/bindings/arm/msm/qcom,idle-state.txt
new file mode 100644
index 0000000..ae1b07f
--- /dev/null
+++ b/Documentation/devicetree/bindings/arm/msm/qcom,idle-state.txt
@@ -0,0 +1,81 @@
+QCOM Idle States for cpuidle driver
+
+ARM provides idle-state node to define the cpuidle states, as defined in [1].
+cpuidle-qcom is the cpuidle driver for Qualcomm SoCs and uses these idle
+states. Idle states have different enter/exit latency and residency values.
+The idle states supported by the QCOM SoC are defined as -
+
+    * Standby
+    * Retention
+    * Standalone Power Collapse (Standalone PC or SPC)
+    * Power Collapse (PC)
+
+Standby: Standby does a little more in addition to architectural clock gating.
+When the WFI instruction is executed the ARM core would gate its internal
+clocks. In addition to gating the clocks, QCOM cpus use this instruction as a
+trigger to execute the SPM state machine. The SPM state machine waits for the
+interrupt to trigger the core back in to active. This triggers the cache
+hierarchy to enter standby states, when all cpus are idle. An interrupt brings
+the SPM state machine out of its wait, the next step is to ensure that the
+cache hierarchy is also out of standby, and then the cpu is allowed to resume
+execution.
+
+Retention: Retention is a low power state where the core is clock gated and
+the memory and the registers associated with the core are retained. The
+voltage may be reduced to the minimum value needed to keep the processor
+registers active. The SPM should be configured to execute the retention
+sequence and would wait for interrupt, before restoring the cpu to execution
+state. Retention may have a slightly higher latency than Standby.
+
+Standalone PC: A cpu can power down and warmboot if there is a sufficient time
+between the time it enters idle and the next known wake up. SPC mode is used
+to indicate a core entering a power down state without consulting any other
+cpu or the system resources. This helps save power only on that core.  The SPM
+sequence for this idle state is programmed to power down the supply to the
+core, wait for the interrupt, restore power to the core, and ensure the
+system state including cache hierarchy is ready before allowing core to
+resume. Applying power and resetting the core causes the core to warmboot
+back into Elevation Level (EL) which trampolines the control back to the
+kernel. Entering a power down state for the cpu, needs to be done by trapping
+into a EL. Failing to do so, would result in a crash enforced by the warm boot
+code in the EL for the SoC. On SoCs with write-back L1 cache, the cache has to
+be flushed in s/w, before powering down the core.
+
+Power Collapse: This state is similar to the SPC mode, but distinguishes
+itself in that the cpu acknowledges and permits the SoC to enter deeper sleep
+modes. In a hierarchical power domain SoC, this means L2 and other caches can
+be flushed, system bus, clocks - lowered, and SoC main XO clock gated and
+voltages reduced, provided all cpus enter this state.  Since the span of low
+power modes possible at this state is vast, the exit latency and the residency
+of this low power mode would be considered high even though at a cpu level,
+this essentially is cpu power down. The SPM in this state also may handshake
+with the Resource power manager processor in the SoC to indicate a complete
+application processor subsystem shut down.
+
+The idle-state for QCOM SoCs are distinguished by the compatible property of
+the idle-states device node.
+The devicetree representation of the idle state should be -
+
+Required properties:
+
+- compatible: Must be one of -
+			"qcom,idle-state-stby",
+			"qcom,idle-state-ret",
+			"qcom,idle-state-spc",
+			"qcom,idle-state-pc",
+		and "arm,idle-state".
+
+Other required and optional properties are specified in [1].
+
+Example:
+
+	idle-states {
+		CPU_SPC: spc {
+			compatible = "qcom,idle-state-spc", "arm,idle-state";
+			entry-latency-us = <150>;
+			exit-latency-us = <200>;
+			min-residency-us = <2000>;
+		};
+	};
+
+[1]. Documentation/devicetree/bindings/arm/idle-states.txt
diff --git a/drivers/cpuidle/Kconfig.arm b/drivers/cpuidle/Kconfig.arm
index 8c16ab2..e98993c 100644
--- a/drivers/cpuidle/Kconfig.arm
+++ b/drivers/cpuidle/Kconfig.arm
@@ -63,3 +63,10 @@ config ARM_MVEBU_V7_CPUIDLE
 	depends on ARCH_MVEBU
 	help
 	  Select this to enable cpuidle on Armada 370, 38x and XP processors.
+
+config ARM_QCOM_CPUIDLE
+	bool "CPU Idle drivers for Qualcomm processors"
+	depends on ARCH_QCOM
+	select DT_IDLE_STATES
+	help
+	  Select this to enable cpuidle for QCOM processors
diff --git a/drivers/cpuidle/Makefile b/drivers/cpuidle/Makefile
index 4d177b9..6c222d5 100644
--- a/drivers/cpuidle/Makefile
+++ b/drivers/cpuidle/Makefile
@@ -17,6 +17,7 @@ obj-$(CONFIG_ARM_ZYNQ_CPUIDLE)		+= cpuidle-zynq.o
 obj-$(CONFIG_ARM_U8500_CPUIDLE)         += cpuidle-ux500.o
 obj-$(CONFIG_ARM_AT91_CPUIDLE)          += cpuidle-at91.o
 obj-$(CONFIG_ARM_EXYNOS_CPUIDLE)        += cpuidle-exynos.o
+obj-$(CONFIG_ARM_QCOM_CPUIDLE)		+= cpuidle-qcom.o
 
 ###############################################################################
 # MIPS drivers
diff --git a/drivers/cpuidle/cpuidle-qcom.c b/drivers/cpuidle/cpuidle-qcom.c
new file mode 100644
index 0000000..5e54d9d
--- /dev/null
+++ b/drivers/cpuidle/cpuidle-qcom.c
@@ -0,0 +1,78 @@
+/*
+ * Copyright (c) 2014, Linaro Limited.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that 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.
+ *
+ */
+
+#include <linux/cpuidle.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+
+#include <soc/qcom/pm.h>
+#include "dt_idle_states.h"
+
+static struct qcom_cpu_pm_ops *lpm_ops;
+
+static int qcom_cpu_stby(struct cpuidle_device *dev,
+				struct cpuidle_driver *drv, int index)
+{
+	if (!lpm_ops->standby(NULL))
+		return -1;
+
+	return index;
+}
+
+static int qcom_cpu_spc(struct cpuidle_device *dev,
+				struct cpuidle_driver *drv, int index)
+{
+	if (!lpm_ops->spc(NULL))
+		return -1;
+
+	return index;
+}
+
+static struct cpuidle_driver qcom_cpuidle_driver = {
+	.name	= "qcom_cpuidle",
+};
+
+static const struct of_device_id qcom_idle_state_match[] = {
+	{ .compatible = "qcom,idle-state-stby", .data = qcom_cpu_stby },
+	{ .compatible = "qcom,idle-state-spc", .data = qcom_cpu_spc },
+	{ },
+};
+
+static int qcom_cpuidle_probe(struct platform_device *pdev)
+{
+	struct cpuidle_driver *drv = &qcom_cpuidle_driver;
+	int ret;
+
+	lpm_ops = pdev->dev.platform_data;
+
+	/* Probe for other states, including standby */
+	ret = dt_init_idle_driver(drv, qcom_idle_state_match, 0);
+	if (ret < 0)
+		return ret;
+
+	return cpuidle_register(drv, NULL);
+}
+
+static struct platform_driver qcom_cpuidle_plat_driver = {
+	.probe	= qcom_cpuidle_probe,
+	.driver = {
+		.name = "qcom_cpuidle",
+	},
+};
+
+module_platform_driver(qcom_cpuidle_plat_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("QCOM cpuidle driver");
+MODULE_ALIAS("platform:qcom-cpuidle");
-- 
2.1.0

[PATCH v13 08/10] arm: dts: qcom: Add idle states device nodes for 8074

[Lina Iyer]

Add ARM common idle states device bindings for cpuidle support for APQ
8074/8974.

Support Standby and Standalone power collapse (power down that does not
affect any SoC idle states) for each cpu.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 arch/arm/boot/dts/qcom-msm8974.dtsi | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/arch/arm/boot/dts/qcom-msm8974.dtsi b/arch/arm/boot/dts/qcom-msm8974.dtsi
index 5a41f44..6942b46 100644
--- a/arch/arm/boot/dts/qcom-msm8974.dtsi
+++ b/arch/arm/boot/dts/qcom-msm8974.dtsi
@@ -22,6 +22,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc0>;
 			qcom,saw = <&saw0>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@1 {
@@ -32,6 +33,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc1>;
 			qcom,saw = <&saw1>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@2 {
@@ -42,6 +44,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc2>;
 			qcom,saw = <&saw2>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@3 {
@@ -52,6 +55,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc3>;
 			qcom,saw = <&saw3>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		L2: l2-cache {
@@ -59,6 +63,22 @@
 			cache-level = <2>;
 			qcom,saw = <&saw_l2>;
 		};
+
+		idle-states {
+			CPU_STBY: standby {
+				compatible = "qcom,idle-state-stby", "arm,idle-state";
+				entry-latency-us = <1>;
+				exit-latency-us = <1>;
+				min-residency-us = <2>;
+			};
+
+			CPU_SPC: spc {
+				compatible = "qcom,idle-state-spc", "arm,idle-state";
+				entry-latency-us = <150>;
+				exit-latency-us = <200>;
+				min-residency-us = <2000>;
+			};
+		};
 	};
 
 	cpu-pmu {
-- 
2.1.0

[PATCH v13 09/10] arm: dts: qcom: Add idle states device nodes for 8084

[Lina Iyer]

Add ARM common idle states device bindings for cpuidle support for APQ
8084.

Support standby and standalone power collapse (power down that does not
affect any SoC idle states) for each cpu.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 arch/arm/boot/dts/qcom-apq8084.dtsi | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/arch/arm/boot/dts/qcom-apq8084.dtsi b/arch/arm/boot/dts/qcom-apq8084.dtsi
index 71182bf..7172edf 100644
--- a/arch/arm/boot/dts/qcom-apq8084.dtsi
+++ b/arch/arm/boot/dts/qcom-apq8084.dtsi
@@ -22,6 +22,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc0>;
 			qcom,saw = <&saw0>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@1 {
@@ -32,6 +33,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc1>;
 			qcom,saw = <&saw1>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@2 {
@@ -42,6 +44,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc2>;
 			qcom,saw = <&saw2>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@3 {
@@ -52,6 +55,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc3>;
 			qcom,saw = <&saw3>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		L2: l2-cache {
@@ -59,6 +63,22 @@
 			cache-level = <2>;
 			qcom,saw = <&saw_l2>;
 		};
+
+		idle-states {
+			CPU_STBY: standby {
+				compatible = "qcom,idle-state-stby", "arm,idle-state";
+				entry-latency-us = <1>;
+				exit-latency-us = <1>;
+				min-residency-us = <2>;
+			};
+
+			CPU_SPC: spc {
+				compatible = "qcom,idle-state-spc", "arm,idle-state";
+				entry-latency-us = <150>;
+				exit-latency-us = <200>;
+				min-residency-us = <2000>;
+			};
+		};
 	};
 
 	cpu-pmu {
-- 
2.1.0

[PATCH v13 10/10] arm: dts: qcom: Add idle state device nodes for 8064

[Lina Iyer]

Add ARM common idle state device bindings for cpuidle support for APQ
8064.

Support Standby and Standalone power collapse (power down that does not
affect any SoC idle states) for each cpu.

Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
---
 arch/arm/boot/dts/qcom-apq8064.dtsi | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/arch/arm/boot/dts/qcom-apq8064.dtsi b/arch/arm/boot/dts/qcom-apq8064.dtsi
index 9fd24bc..ab21dba 100644
--- a/arch/arm/boot/dts/qcom-apq8064.dtsi
+++ b/arch/arm/boot/dts/qcom-apq8064.dtsi
@@ -23,6 +23,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc0>;
 			qcom,saw = <&saw0>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@1 {
@@ -33,6 +34,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc1>;
 			qcom,saw = <&saw1>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@2 {
@@ -43,6 +45,7 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc2>;
 			qcom,saw = <&saw2>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		cpu@3 {
@@ -53,12 +56,29 @@
 			next-level-cache = <&L2>;
 			qcom,acc = <&acc3>;
 			qcom,saw = <&saw3>;
+			cpu-idle-states = <&CPU_STBY &CPU_SPC>;
 		};
 
 		L2: l2-cache {
 			compatible = "cache";
 			cache-level = <2>;
 		};
+
+		idle-states {
+			CPU_STBY: standby {
+				compatible = "qcom,idle-state-stby", "arm,idle-state";
+				entry-latency-us = <1>;
+				exit-latency-us = <1>;
+				min-residency-us = <2>;
+			};
+
+			CPU_SPC: spc {
+				compatible = "qcom,idle-state-spc", "arm,idle-state";
+				entry-latency-us = <400>;
+				exit-latency-us = <900>;
+				min-residency-us = <3000>;
+			};
+		};
 	};
 
 	cpu-pmu {
-- 
2.1.0

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Ivan T. Ivanov]

Hi, 

On Wed, 2014-11-26 at 22:24 -0700, Lina Iyer wrote:
> 

>  - compatible:
> @@ -14,10 +23,13 @@ PROPERTIES
>         Value type: <string>
>         Definition: shall contain "qcom,saw2". A more specific value should be
>                                                 one of:

Which driver is supposed to handle this property?

> -                               "qcom,saw2-v1"
> -                               "qcom,saw2-v1.1"
> -                               "qcom,saw2-v2"
> -                               "qcom,saw2-v2.1"
> +                       "qcom,saw2-v1"
> +                       "qcom,saw2-v1.1"
> +                       "qcom,saw2-v2"
> +                       "qcom,saw2-v2.1"

Doesn't seems to be used or supported?

> +                       "qcom,apq8064-saw2-v1.1-cpu"
> +                       "qcom,msm8974-saw2-v2.1-cpu"
> +                       "qcom,apq8084-saw2-v2.1-cpu"
> 

<snip>

> -       regulator@2099000 {
> +       power-controller@2099000 {
>                 compatible = "qcom,saw2";
>                 reg = <0x02099000 0x1000>, <0x02009000 0x1000>;
> +               regulator;
> 

<snip>

> +static const struct of_device_id spm_match_table[] = {
> +       { .compatible = "qcom,msm8974-saw2-v2.1-cpu",
> +                       .data = &spm_reg_8974_8084_cpu },
> +       { .compatible = "qcom,apq8084-saw2-v2.1-cpu",
> +                       .data = &spm_reg_8974_8084_cpu },
> +       { .compatible = "qcom,apq8064-saw2-v1.1-cpu",
> +                       .data = &spm_reg_8064_cpu },
> +       { },
> +};
> +

Regards,
Ivan

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Daniel Lezcano]

On 11/27/2014 06:24 AM, Lina Iyer wrote:
> SPM is a hardware block that controls the peripheral logic surrounding
> the application cores (cpu/l$). When the core executes WFI instruction,
> the SPM takes over the putting the core in low power state as
> configured. The wake up for the SPM is an interrupt at the GIC, which
> then completes the rest of low power mode sequence and brings the core
> out of low power mode.
>
> The SPM has a set of control registers that configure the SPMs
> individually based on the type of the core and the runtime conditions.
> SPM is a finite state machine block to which a sequence is provided and
> it interprets the bytes and executes them in sequence. Each low power
> mode that the core can enter into is provided to the SPM as a sequence.
>
> Configure the SPM to set the core (cpu or L2) into its low power mode,
> the index of the first command in the sequence is set in the SPM_CTL
> register. When the core executes ARM wfi instruction, it triggers the
> SPM state machine to start executing from that index. The SPM state
> machine waits until the interrupt occurs and starts executing the rest
> of the sequence until it hits the end of the sequence. The end of the
> sequence jumps the core out of its low power mode.
>
> Add support for an idle driver to set up the SPM to place the core in
> Standby or Standalone power collapse mode when the core is idle.
>
> Based on work by: Mahesh Sivasubramanian <msivasub@codeaurora.org>,
> Ai Li <ali@codeaurora.org>, Praveen Chidambaram <pchidamb@codeaurora.org>
> Original tree available at -
> git://codeaurora.org/quic/la/kernel/msm-3.10.git
>
> Signed-off-by: Lina Iyer <lina.iyer@linaro.org>
> Reviewed-by: Kevin Hilman <khilman@linaro.org>

[ ... ]

> +static struct spm_driver_data *spm_get_drv(struct platform_device *pdev,
> +		int *spm_cpu)
> +{
> +	struct spm_driver_data *drv = NULL;
> +	struct device_node *cpu_node, *saw_node;
> +	int cpu;
> +	bool found = false;
> +
> +	for_each_possible_cpu(cpu) {
> +		cpu_node = of_cpu_device_node_get(cpu);
> +		if (!cpu_node)
> +			continue;
> +		saw_node = of_parse_phandle(cpu_node, "qcom,saw", 0);
> +		if (saw_node) {
> +			if (saw_node == pdev->dev.of_node)
> +				found = true;
> +			of_node_put(saw_node);
> +		}
> +		of_node_put(cpu_node);
> +		if (found)
> +			break;
> +	}
> +
> +	if (found) {
> +		drv = devm_kzalloc(&pdev->dev, sizeof(*drv), GFP_KERNEL);
> +		if (drv)
> +			*spm_cpu = cpu;
> +	}
> +
> +	return drv;
> +}
> +
> +static const struct of_device_id spm_match_table[] = {
> +	{ .compatible = "qcom,msm8974-saw2-v2.1-cpu",
> +	  .data = &spm_reg_8974_8084_cpu },
> +	{ .compatible = "qcom,apq8084-saw2-v2.1-cpu",
> +	  .data = &spm_reg_8974_8084_cpu },
> +	{ .compatible = "qcom,apq8064-saw2-v1.1-cpu",
> +	  .data = &spm_reg_8064_cpu },
> +	{ },
> +};
> +
> +static const struct qcom_cpu_pm_ops lpm_ops = {
> +	.standby = qcom_cpu_standby,
> +	.spc = qcom_cpu_spc,
> +};
> +
> +static int spm_dev_probe(struct platform_device *pdev)
> +{
> +	struct spm_driver_data *drv;
> +	struct resource *res;
> +	const struct of_device_id *match_id;
> +	void __iomem *addr;
> +	int cpu = -EINVAL;
> +	static bool cpuidle_drv_init;

         ^^^^^^^^^

As already said in a previous comment, please find a way to remove that.

> +	const struct platform_device_info qcom_cpuidle_info = {
> +		.name	= "qcom_cpuidle",
> +		.id	= -1,
> +		.data = &lpm_ops,
> +		.size_data = sizeof(lpm_ops),
> +	};
> +
> +	drv = spm_get_drv(pdev, &cpu);
> +	if (!drv || cpu < 0)
> +		return -EINVAL;

As already said in a previous comment, it is not possible to have "cpu < 
0" with "drv != NULL", so except I am missing something the test should be:

	if (!drv)
		return -EINVAL;

There is something wrong with the init sequence. Don't you find weird 
you have to backward search for the cpu belonging to the pdev each time 
the probe function is called ?


> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	drv->reg_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(drv->reg_base))
> +		return PTR_ERR(drv->reg_base);
> +
> +	match_id = of_match_node(spm_match_table, pdev->dev.of_node);
> +	if (!match_id)
> +		return -ENODEV;
> +
> +	drv->reg_data = match_id->data;
> +
> +	/* Write the SPM sequences first.. */
> +	addr = drv->reg_base + drv->reg_data->reg_offset[SPM_REG_SEQ_ENTRY];
> +	__iowrite32_copy(addr, drv->reg_data->seq,
> +			ARRAY_SIZE(drv->reg_data->seq) / 4);
> +
> +	/*
> +	 * ..and then the control registers.
> +	 * On some SoC's if the control registers are written first and if the
> +	 * CPU was held in reset, the reset signal could trigger the SPM state
> +	 * machine, before the sequences are completely written.
> +	 */
> +	spm_register_write(drv, SPM_REG_CFG, drv->reg_data->spm_cfg);
> +	spm_register_write(drv, SPM_REG_DLY, drv->reg_data->spm_dly);
> +	spm_register_write(drv, SPM_REG_PMIC_DLY, drv->reg_data->pmic_dly);
> +
> +	spm_register_write(drv, SPM_REG_PMIC_DATA_0,
> +				drv->reg_data->pmic_data[0]);
> +	spm_register_write(drv, SPM_REG_PMIC_DATA_1,
> +				drv->reg_data->pmic_data[1]);
> +
> +	/*
> +	 * Ensure all observers see the above register writes before the
> +	 * cpuidle driver is allowed to use the SPM.
> +	 */
> +	wmb();
> +	per_cpu(cpu_spm_drv, cpu) = drv;
> +
> +	if (!cpuidle_drv_init) {
> +		platform_device_register_full(&qcom_cpuidle_info);
> +		cpuidle_drv_init = true;
> +	}
> +
> +	return 0;
> +}
> +
> +static struct platform_driver spm_driver = {
> +	.probe = spm_dev_probe,
> +	.driver = {
> +		.name = "saw",
> +		.of_match_table = spm_match_table,
> +	},
> +};
> +
> +module_platform_driver(spm_driver);
> +
> +MODULE_LICENSE("GPL v2");
> +MODULE_DESCRIPTION("SAW power controller driver");
> +MODULE_ALIAS("platform:saw");
> diff --git a/include/soc/qcom/pm.h b/include/soc/qcom/pm.h
> new file mode 100644
> index 0000000..d9a56d7
> --- /dev/null
> +++ b/include/soc/qcom/pm.h
> @@ -0,0 +1,31 @@
> +/*
> + * Copyright (c) 2009-2014, The Linux Foundation. All rights reserved.
> + *
> + * This software is licensed under the terms of the GNU General Public
> + * License version 2, as published by the Free Software Foundation, and
> + * may be copied, distributed, and modified under those terms.
> + *
> + * This program is distributed in the hope that 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.
> + *
> + */
> +
> +#ifndef __QCOM_PM_H
> +#define __QCOM_PM_H
> +
> +enum pm_sleep_mode {
> +	PM_SLEEP_MODE_STBY,
> +	PM_SLEEP_MODE_RET,
> +	PM_SLEEP_MODE_SPC,
> +	PM_SLEEP_MODE_PC,
> +	PM_SLEEP_MODE_NR,
> +};
> +
> +struct qcom_cpu_pm_ops {
> +	int (*standby)(void *data);
> +	int (*spc)(void *data);
> +};
> +
> +#endif  /* __QCOM_PM_H */
>


-- 
  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

Re: [PATCH v13 00/10] cpuidle driver for QCOM SoCs: 8064, 8074, 8084

[Daniel Lezcano]

Lina,

a night for me has passed and I have in the meantime 3 new versions of 
the patchset obviously done in the hurry and not tested.

Furthermore the change log fails to give the details, "Address review 
comments on spm.c" is just a clue and when I look at the spm.c code all 
the comments where not addressed.

Please, have a look at the "static bool cpuidle_drv_init;", remove it 
and look closely at the idle callbacks error handling in the cpuidle driver.

Thanks

   -- Daniel




On 11/27/2014 06:24 AM, Lina Iyer wrote:
> Dependent patchsets -
>          https://lkml.org/lkml/2014/8/4/767
>          http://www.spinics.net/lists/linux-arm-msm/msg10799.html
>          http://www.spinics.net/lists/linux-arm-msm/msg10795.html
>
> Changes since v12:
> - Minor fixes
> - Added Reviewed-by
>
> Changes since v11:
> - Address review comments on spm.c
> - Commenting style fixes
> - Added Reviewed-by
>
> Changes since v10:
> [ https://www.mail-archive.com/devicetree@vger.kernel.org/msg51880.html ]
> - Address review comments
> - Added Acked-by and Reviewed-by
>
> Changes since v9:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11714.html ]
> - Address review comments on v9
>
> Changes since v8:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11473.html ]
> - Flatten out the file structure - merge pm.c into spm.c after discussions
> - Add a new function to set warm boot address, in scm-boot.c
> - Support for 8064 (New)
> - Tested on 8074, 8084. 8064 was tested with a WIP tree
> - Address review comments from v8
> - Looking into possiblility of  initializing the cpuidle device for a cpu,
> only when the corresponding spm device is probed successfully.
>
> Changes since v7:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11199.html ]
> - Address review comments
> - Tested on 8974 but not 8084
> - WFI renamed to Standby
> - Update commit text with original author and link to the downstream tree
>
> Changes since v6:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg11012.html ]
> - SPM device nodes merged with existing SAW DT nodes
> - SPM register information is handled within the driver
> - Clean up from using 'msm' to 'qcom'
>          - Shorten some enumerations as well
> - Review comments from v6 addressed
> - New: Support for 8084 SoC
>          - Not tested. I do not have a board with this SoC, but the SPM
>          configuration should be identical for WFI and SPC
>
> Changes since v5:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10559.html ]
> - Merge spm-devices.c and spm.c into one file and one patch
>          - Simplify implementation of the driver.
>          - Update documentation mapping the DT properties with corresponding
>            SPM register information.
> - Removed scm-boot changes for quad core warmboot, its has been pulled in.
>
> Changes since v4:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10327.html ]
> - Update to the v8 of ARM generic idle states patches
> - Use platform device model for cpuidle-qcom
> - Clean up msm-pm.c to remove unnecessary include files and functions
> - Update commit text and documentation for all idle states
> - Remove scm-boot relocate patch from this series, submitted earlier
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10518.html ]
>
> Changes since v3:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10288.html ]
> - Fix CONFIG_QCOM_PM Kconfig as bool
> - More clean ups in spm.c and spm-devices.c
>          - Removed and re-organized data structures to make initialization simple
>          - Remove export of sequence flush functions
>          - Updated commit text
>          - Comments for use of barriers.
> - Rebase on top of 3.17-rc1
>
> Changes since v2:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10148.html ]
> - Prune all the drivers to support basic WFI and power down cpuidle
>    functionality. Remove debug code.
> - Integrate KConfig changes into the drivers' patches.
> - Use Lorenzo's ARM idle-states patches as the basis for reading cpuidle
>    c-states from DT.
>    [ http://marc.info/?l=linux-pm&m=140794514812383&w=2 ]
> - Incorporate review comments
> - Rebase on top of 3.16
>
> Changes since v1/RFC:
> [ https://www.mail-archive.com/linux-arm-msm@vger.kernel.org/msg10065.html ]
> - Remove hotplug from the patch series. Will submit it separately.
> - Fix SPM drivers per the review comments
> - Modify patch sequence to compile SPM drivers independent of msm-pm, so as to
>    allow wfi() calls to use SPM even without SoC interface driver.
>
> 8074/8084/8064 like any ARM SoC can do architectural clock gating, that helps save
> on power, but not enough of leakage power.  Leakage power of the SoC can be
> further reduced by turning off power to the core. To aid this, every core (cpu
> and L2) is accompanied by a Sub-system Power Manager (SPM), that can be
> configured to indicate the low power mode, the core would be put into and the
> SPM programs the peripheral h/w accordingly to enter low power and turn off the
> power rail to the core.
>
> The idle invocation hierarchy -
>
> CPUIDLE
>          |
>          cpuidle-qcom.c [CPUIdle driver]
>          |
>          ------> spm.c [SPM driver]
>                  |
>                  ------> scm-boot.c [SCM interface layer]
>                          |
> ------------------------|--------------------------
> (EL)                    Secure Monitor Code
>                          |
>                          |
>                          wfi();
> ------------------------|--------------------------
> (HW)                    [CPU] {clock gate}
>                          |
>                          -----> [SPM] {statemachine}
>
>
> The patchset does the following -
>
> - Introduce the SPM driver to control power to the core
> - Add device bindings for 8974, 8084, 8064 CPU SPM devices
> - Add cpuidle driver for QCOM cpus, using ARM generic idle state definitions.
> - Add device bindings for 8974, 8084, 8064 idle-states - WFI and SPC
>
> Thanks,
> Lina
>
>
>
>
> Lina Iyer (10):
>    qcom: scm: Move scm-boot files to drivers/soc/qcom/ and
>      include/soc/qcom
>    qcom: scm: Add SCM warmboot support for quad core SoCs
>    qcom: spm: Add Subsystem Power Manager driver
>    arm: dts: qcom: Add power-controller device node for 8074 Krait CPUs
>    arm: dts: qcom: Add power-controller device node for 8084 Krait CPUs
>    arm: dts: qcom: Update power-controller device node for 8064 Krait
>      CPUs
>    qcom: cpuidle: Add cpuidle driver for QCOM cpus
>    arm: dts: qcom: Add idle states device nodes for 8074
>    arm: dts: qcom: Add idle states device nodes for 8084
>    arm: dts: qcom: Add idle state device nodes for 8064
>
>   .../bindings/arm/msm/qcom,idle-state.txt           |  81 +++++
>   .../devicetree/bindings/arm/msm/qcom,saw2.txt      |  31 +-
>   arch/arm/boot/dts/qcom-apq8064.dtsi                |  36 ++-
>   arch/arm/boot/dts/qcom-apq8084.dtsi                |  46 ++-
>   arch/arm/boot/dts/qcom-msm8974.dtsi                |  46 ++-
>   arch/arm/mach-qcom/Makefile                        |   1 -
>   arch/arm/mach-qcom/platsmp.c                       |   2 +-
>   drivers/cpuidle/Kconfig.arm                        |   7 +
>   drivers/cpuidle/Makefile                           |   1 +
>   drivers/cpuidle/cpuidle-qcom.c                     |  78 +++++
>   drivers/soc/qcom/Kconfig                           |   8 +
>   drivers/soc/qcom/Makefile                          |   3 +-
>   .../arm/mach-qcom => drivers/soc/qcom}/scm-boot.c  |  37 ++-
>   drivers/soc/qcom/spm.c                             | 327 +++++++++++++++++++++
>   include/soc/qcom/pm.h                              |  31 ++
>   .../arm/mach-qcom => include/soc/qcom}/scm-boot.h  |   3 +-
>   16 files changed, 716 insertions(+), 22 deletions(-)
>   create mode 100644 Documentation/devicetree/bindings/arm/msm/qcom,idle-state.txt
>   create mode 100644 drivers/cpuidle/cpuidle-qcom.c
>   rename {arch/arm/mach-qcom => drivers/soc/qcom}/scm-boot.c (59%)
>   create mode 100644 drivers/soc/qcom/spm.c
>   create mode 100644 include/soc/qcom/pm.h
>   rename {arch/arm/mach-qcom => include/soc/qcom}/scm-boot.h (91%)
>


-- 
  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Lorenzo Pieralisi]

On Thu, Nov 27, 2014 at 05:24:07AM +0000, Lina Iyer wrote:

[...]

> +static int spm_set_low_power_mode(enum pm_sleep_mode mode)
> +{
> +       struct spm_driver_data *drv = per_cpu(cpu_spm_drv,
> +                                               smp_processor_id());
> +       u32 start_index;
> +       u32 ctl_val;
> +
> +       if (!drv)
> +               return -ENXIO;
> +
> +       start_index = drv->reg_data->start_index[mode];
> +
> +       ctl_val = spm_register_read(drv, SPM_REG_SPM_CTL);
> +       ctl_val &= ~(SPM_CTL_INDEX << SPM_CTL_INDEX_SHIFT);
> +       ctl_val |= start_index << SPM_CTL_INDEX_SHIFT;
> +       ctl_val |= SPM_CTL_EN;
> +       spm_register_write(drv, SPM_REG_SPM_CTL, ctl_val);
> +
> +       /* Ensure we have written the start address */
> +       wmb();

Can you explain please what this wmb is meant to achieve ? If it is there
to make sure the write made it to the SPM that's not a proper way
to achieve it, the barrier ensures ordering, not write completion.

> +
> +       return 0;
> +}
> +
> +static int qcom_pm_collapse(unsigned long int unused)
> +{
> +       int ret;
> +       u32 flag;
> +       int cpu = smp_processor_id();
> +
> +       ret = scm_set_warm_boot_addr(cpu_resume, cpu);
> +       if (ret)
> +               return ret;
> +
> +       flag = SCM_L2_ON & SCM_FLUSH_FLAG_MASK;
> +       scm_call_atomic1(SCM_SVC_BOOT, SCM_CMD_TERMINATE_PC, flag);
> +
> +       /*
> +        * Returns here only if there was a pending interrupt and we did not
> +        * power down as a result.
> +        */
> +       return 0;

I know the return value is changed by the cpu_suspend implementation to
signal error when not returning from cpu_resume, but you should not
return 0 anyway.

> +}
> +
> +static int qcom_cpu_standby(void *unused)
> +{
> +       int ret;
> +
> +       ret = spm_set_low_power_mode(PM_SLEEP_MODE_STBY);
> +       if (ret)
> +               return ret;
> +
> +       cpu_do_idle();
> +
> +       return 0;
> +}
> +
> +static int qcom_cpu_spc(void *unused)
> +{      int ret;
> +
> +       ret = spm_set_low_power_mode(PM_SLEEP_MODE_STBY);
> +       if (ret)
> +               return ret;
> +
> +       cpu_pm_enter();
> +       cpu_suspend(0, qcom_pm_collapse);

You must stash the return value and propagate it.

> +       cpu_pm_exit();
> +
> +       return 0;
> +}
> +
> +static struct spm_driver_data *spm_get_drv(struct platform_device *pdev,
> +               int *spm_cpu)
> +{
> +       struct spm_driver_data *drv = NULL;
> +       struct device_node *cpu_node, *saw_node;
> +       int cpu;
> +       bool found = false;
> +
> +       for_each_possible_cpu(cpu) {
> +               cpu_node = of_cpu_device_node_get(cpu);
> +               if (!cpu_node)
> +                       continue;
> +               saw_node = of_parse_phandle(cpu_node, "qcom,saw", 0);

+               if (saw_node) {
+                       if (saw_node == pdev->dev.of_node)
+                               found = true;
+                       of_node_put(saw_node);
+               }

can be:
		found = (saw_node == pdev->dev.of_node);
		of_node_put(saw_node);

Thanks,
Lorenzo

> +               of_node_put(cpu_node);
> +               if (found)
> +                       break;
> +       }
> +
> +       if (found) {
> +               drv = devm_kzalloc(&pdev->dev, sizeof(*drv), GFP_KERNEL);
> +               if (drv)
> +                       *spm_cpu = cpu;
> +       }
> +
> +       return drv;
> +}
> +
> +static const struct of_device_id spm_match_table[] = {
> +       { .compatible = "qcom,msm8974-saw2-v2.1-cpu",
> +         .data = &spm_reg_8974_8084_cpu },
> +       { .compatible = "qcom,apq8084-saw2-v2.1-cpu",
> +         .data = &spm_reg_8974_8084_cpu },
> +       { .compatible = "qcom,apq8064-saw2-v1.1-cpu",
> +         .data = &spm_reg_8064_cpu },
> +       { },
> +};
> +
> +static const struct qcom_cpu_pm_ops lpm_ops = {
> +       .standby = qcom_cpu_standby,
> +       .spc = qcom_cpu_spc,
> +};
> +
> +static int spm_dev_probe(struct platform_device *pdev)
> +{
> +       struct spm_driver_data *drv;
> +       struct resource *res;
> +       const struct of_device_id *match_id;
> +       void __iomem *addr;
> +       int cpu = -EINVAL;
> +       static bool cpuidle_drv_init;
> +       const struct platform_device_info qcom_cpuidle_info = {
> +               .name   = "qcom_cpuidle",
> +               .id     = -1,
> +               .data = &lpm_ops,
> +               .size_data = sizeof(lpm_ops),
> +       };
> +
> +       drv = spm_get_drv(pdev, &cpu);
> +       if (!drv || cpu < 0)
> +               return -EINVAL;
> +
> +       res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +       drv->reg_base = devm_ioremap_resource(&pdev->dev, res);
> +       if (IS_ERR(drv->reg_base))
> +               return PTR_ERR(drv->reg_base);
> +
> +       match_id = of_match_node(spm_match_table, pdev->dev.of_node);
> +       if (!match_id)
> +               return -ENODEV;
> +
> +       drv->reg_data = match_id->data;
> +
> +       /* Write the SPM sequences first.. */
> +       addr = drv->reg_base + drv->reg_data->reg_offset[SPM_REG_SEQ_ENTRY];
> +       __iowrite32_copy(addr, drv->reg_data->seq,
> +                       ARRAY_SIZE(drv->reg_data->seq) / 4);
> +
> +       /*
> +        * ..and then the control registers.
> +        * On some SoC's if the control registers are written first and if the
> +        * CPU was held in reset, the reset signal could trigger the SPM state
> +        * machine, before the sequences are completely written.
> +        */
> +       spm_register_write(drv, SPM_REG_CFG, drv->reg_data->spm_cfg);
> +       spm_register_write(drv, SPM_REG_DLY, drv->reg_data->spm_dly);
> +       spm_register_write(drv, SPM_REG_PMIC_DLY, drv->reg_data->pmic_dly);
> +
> +       spm_register_write(drv, SPM_REG_PMIC_DATA_0,
> +                               drv->reg_data->pmic_data[0]);
> +       spm_register_write(drv, SPM_REG_PMIC_DATA_1,
> +                               drv->reg_data->pmic_data[1]);
> +
> +       /*
> +        * Ensure all observers see the above register writes before the
> +        * cpuidle driver is allowed to use the SPM.
> +        */
> +       wmb();
> +       per_cpu(cpu_spm_drv, cpu) = drv;
> +
> +       if (!cpuidle_drv_init) {
> +               platform_device_register_full(&qcom_cpuidle_info);
> +               cpuidle_drv_init = true;
> +       }
> +
> +       return 0;
> +}
> +
> +static struct platform_driver spm_driver = {
> +       .probe = spm_dev_probe,
> +       .driver = {
> +               .name = "saw",
> +               .of_match_table = spm_match_table,
> +       },
> +};
> +
> +module_platform_driver(spm_driver);
> +
> +MODULE_LICENSE("GPL v2");
> +MODULE_DESCRIPTION("SAW power controller driver");
> +MODULE_ALIAS("platform:saw");
> diff --git a/include/soc/qcom/pm.h b/include/soc/qcom/pm.h
> new file mode 100644
> index 0000000..d9a56d7
> --- /dev/null
> +++ b/include/soc/qcom/pm.h
> @@ -0,0 +1,31 @@
> +/*
> + * Copyright (c) 2009-2014, The Linux Foundation. All rights reserved.
> + *
> + * This software is licensed under the terms of the GNU General Public
> + * License version 2, as published by the Free Software Foundation, and
> + * may be copied, distributed, and modified under those terms.
> + *
> + * This program is distributed in the hope that 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.
> + *
> + */
> +
> +#ifndef __QCOM_PM_H
> +#define __QCOM_PM_H
> +
> +enum pm_sleep_mode {
> +       PM_SLEEP_MODE_STBY,
> +       PM_SLEEP_MODE_RET,
> +       PM_SLEEP_MODE_SPC,
> +       PM_SLEEP_MODE_PC,
> +       PM_SLEEP_MODE_NR,
> +};
> +
> +struct qcom_cpu_pm_ops {
> +       int (*standby)(void *data);
> +       int (*spc)(void *data);
> +};
> +
> +#endif  /* __QCOM_PM_H */
> --
> 2.1.0
> 
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> 

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Lina Iyer]

Hi Ivan,

On Thu, Nov 27 2014 at 01:43 -0700, Ivan T. Ivanov wrote:
>
>Hi,
>
>On Wed, 2014-11-26 at 22:24 -0700, Lina Iyer wrote:
>>
>
>>  - compatible:
>> @@ -14,10 +23,13 @@ PROPERTIES
>>         Value type: <string>
>>         Definition: shall contain "qcom,saw2". A more specific value should be
>>                                                 one of:
>
>Which driver is supposed to handle this property?
>
drivers/soc/qcom/spm.c - SAW power controller driver.
>> -                               "qcom,saw2-v1"
>> -                               "qcom,saw2-v1.1"
>> -                               "qcom,saw2-v2"
>> -                               "qcom,saw2-v2.1"
>> +                       "qcom,saw2-v1"
>> +                       "qcom,saw2-v1.1"
>> +                       "qcom,saw2-v2"
>> +                       "qcom,saw2-v2.1"
>
>Doesn't seems to be used or supported?
>
Used for identify the version of the SAW that does voltage changes.  Any
SAW that can change the voltage would use one of the above compatible
strings. We only need a generic version number for regulator
functionality as opposed to the ones used for idle control (like below),
which needs to correctly identify the version number (saw2-v1.1), cpu/L2
(-cpu_variant of the SAW on a particlar SoC (qcom,apq/msmXXXX) to set
the correct register values for that SoC.

L2 SAW is currently not upstream or part of this series. Will follow
suit, when we have the power domains to control cluster low power modes.
The regulator functionality is currently not available upstream, as well
for 8084 and 8074.

Thanks,
Lina

>> +                       "qcom,apq8064-saw2-v1.1-cpu"
>> +                       "qcom,msm8974-saw2-v2.1-cpu"
>> +                       "qcom,apq8084-saw2-v2.1-cpu"
>>
>
><snip>
>
>> -       regulator@2099000 {
>> +       power-controller@2099000 {
>>                 compatible = "qcom,saw2";
>>                 reg = <0x02099000 0x1000>, <0x02009000 0x1000>;
>> +               regulator;
>>
>
><snip>
>
>> +static const struct of_device_id spm_match_table[] = {
>> +       { .compatible = "qcom,msm8974-saw2-v2.1-cpu",
>> +                       .data = &spm_reg_8974_8084_cpu },
>> +       { .compatible = "qcom,apq8084-saw2-v2.1-cpu",
>> +                       .data = &spm_reg_8974_8084_cpu },
>> +       { .compatible = "qcom,apq8064-saw2-v1.1-cpu",
>> +                       .data = &spm_reg_8064_cpu },
>> +       { },
>> +};
>> +
>
>Regards,
>Ivan

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Lina Iyer]

On Thu, Nov 27 2014 at 01:52 -0700, Daniel Lezcano wrote:
>On 11/27/2014 06:24 AM, Lina Iyer wrote:

>+	static bool cpuidle_drv_init;
>
>        ^^^^^^^^^
>
>As already said in a previous comment, please find a way to remove that.
>
I will look into it. Stephen and I wanted the cpuidle driver to be
probed only after any of the SPMs are ready. And possibly, only for that
cpu, for which the SPM has been probed.

To achieve the SPM -> CPUIDLE Device relation, I havent found a good way
to do that. Without using CPUIDLE_MULTIPLE_DRIVERS, initializing each
cpuidle device, separate from the cpuidle driver, requires that I update 
the cpuidle_driver->cpumask after probing each SPM device, to allow for
only one driver and cpuidle devices only for the probed cpus. Using the
cpuidle_register_driver(), resets the drv->refcnt in
__cpuidle_driver_init.

I may need something like this in the else clause of CPUIDLE_MULTIPLE_DRIVERS -

static inline struct cpuidle_driver *__cpuidle_get_cpu_driver(int cpu)
{
        struct cpuidle_driver *drv = cpuidle_curr_driver;

        if (drv && !cpumask_test_cpu(cpu, drv->cpumask))
                 drv = NULL;

        return drv;
}

void cpuidle_update_cpumask(struct cpumask *mask) {
        struct cpuidle_driver *drv;

        spin_lock(&cpuidle_driver_lock);
        drv = cpuidle_get_driver();
	if (drv)
		drv->cpumask = mask ?: cpu_possible_mask;
        spin_unlock(&cpuidle_driver_lock);
}

With that, I could register cpuidle driver the first time when the mask changes
from empty and consequent updates would just update the cpumask. (I am not
sure, if I missed anything in this change). It just seemed far too invasive at
this time, in lieu of the static bool.

>>+	const struct platform_device_info qcom_cpuidle_info = {
>>+		.name	= "qcom_cpuidle",
>>+		.id	= -1,
>>+		.data = &lpm_ops,
>>+		.size_data = sizeof(lpm_ops),
>>+	};
>>+
>>+	drv = spm_get_drv(pdev, &cpu);
>>+	if (!drv || cpu < 0)
>>+		return -EINVAL;
>
>As already said in a previous comment, it is not possible to have "cpu 
>< 0" with "drv != NULL", so except I am missing something the test 
>should be:
>
>	if (!drv)
>		return -EINVAL;
>
Sorry, done.

>There is something wrong with the init sequence. Don't you find weird 
>you have to backward search for the cpu belonging to the pdev each 
>time the probe function is called ?
>
>
Well, it was that or the SPM device node pointing to the CPU that it
references. It seems more common to have an iterator than the doubly
linked device nodes. I dont have a strong preference either way, just
chose the way that made device nodes easier.

>>+
>>+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>>+	drv->reg_base = devm_ioremap_resource(&pdev->dev, res);
>>+	if (IS_ERR(drv->reg_base))
>>+		return PTR_ERR(drv->reg_base);
>>+
>>+	match_id = of_match_node(spm_match_table, pdev->dev.of_node);
>>+	if (!match_id)
>>+		return -ENODEV;
>>+
>>+	drv->reg_data = match_id->data;
>>+
>>+	/* Write the SPM sequences first.. */
>>+	addr = drv->reg_base + drv->reg_data->reg_offset[SPM_REG_SEQ_ENTRY];
>>+	__iowrite32_copy(addr, drv->reg_data->seq,
>>+			ARRAY_SIZE(drv->reg_data->seq) / 4);
>>+
>>+	/*
>>+	 * ..and then the control registers.
>>+	 * On some SoC's if the control registers are written first and if the
>>+	 * CPU was held in reset, the reset signal could trigger the SPM state
>>+	 * machine, before the sequences are completely written.
>>+	 */
>>+	spm_register_write(drv, SPM_REG_CFG, drv->reg_data->spm_cfg);
>>+	spm_register_write(drv, SPM_REG_DLY, drv->reg_data->spm_dly);
>>+	spm_register_write(drv, SPM_REG_PMIC_DLY, drv->reg_data->pmic_dly);
>>+
>>+	spm_register_write(drv, SPM_REG_PMIC_DATA_0,
>>+				drv->reg_data->pmic_data[0]);
>>+	spm_register_write(drv, SPM_REG_PMIC_DATA_1,
>>+				drv->reg_data->pmic_data[1]);
>>+
>>+	/*
>>+	 * Ensure all observers see the above register writes before the
>>+	 * cpuidle driver is allowed to use the SPM.
>>+	 */
>>+	wmb();
>>+	per_cpu(cpu_spm_drv, cpu) = drv;
>>+
>>+	if (!cpuidle_drv_init) {
>>+		platform_device_register_full(&qcom_cpuidle_info);
>>+		cpuidle_drv_init = true;
>>+	}
>>+
>>+	return 0;
>>+}
>>+
>>+static struct platform_driver spm_driver = {
>>+	.probe = spm_dev_probe,
>>+	.driver = {
>>+		.name = "saw",
>>+		.of_match_table = spm_match_table,
>>+	},
>>+};
>>+
>>+module_platform_driver(spm_driver);
>>+
>>+MODULE_LICENSE("GPL v2");
>>+MODULE_DESCRIPTION("SAW power controller driver");
>>+MODULE_ALIAS("platform:saw");
>>diff --git a/include/soc/qcom/pm.h b/include/soc/qcom/pm.h
>>new file mode 100644
>>index 0000000..d9a56d7
>>--- /dev/null
>>+++ b/include/soc/qcom/pm.h
>>@@ -0,0 +1,31 @@
>>+/*
>>+ * Copyright (c) 2009-2014, The Linux Foundation. All rights reserved.
>>+ *
>>+ * This software is licensed under the terms of the GNU General Public
>>+ * License version 2, as published by the Free Software Foundation, and
>>+ * may be copied, distributed, and modified under those terms.
>>+ *
>>+ * This program is distributed in the hope that 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.
>>+ *
>>+ */
>>+
>>+#ifndef __QCOM_PM_H
>>+#define __QCOM_PM_H
>>+
>>+enum pm_sleep_mode {
>>+	PM_SLEEP_MODE_STBY,
>>+	PM_SLEEP_MODE_RET,
>>+	PM_SLEEP_MODE_SPC,
>>+	PM_SLEEP_MODE_PC,
>>+	PM_SLEEP_MODE_NR,
>>+};
>>+
>>+struct qcom_cpu_pm_ops {
>>+	int (*standby)(void *data);
>>+	int (*spc)(void *data);
>>+};
>>+
>>+#endif  /* __QCOM_PM_H */
>>
>
>
>-- 
> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
>
>Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
><http://twitter.com/#!/linaroorg> Twitter |
><http://www.linaro.org/linaro-blog/> Blog
>

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Lina Iyer]

On Thu, Nov 27 2014 at 08:01 -0700, Lorenzo Pieralisi wrote:
>On Thu, Nov 27, 2014 at 05:24:07AM +0000, Lina Iyer wrote:
>
>[...]
>
>> +static int spm_set_low_power_mode(enum pm_sleep_mode mode)
>> +{
>> +       struct spm_driver_data *drv = per_cpu(cpu_spm_drv,
>> +                                               smp_processor_id());
>> +       u32 start_index;
>> +       u32 ctl_val;
>> +
>> +       if (!drv)
>> +               return -ENXIO;
>> +
>> +       start_index = drv->reg_data->start_index[mode];
>> +
>> +       ctl_val = spm_register_read(drv, SPM_REG_SPM_CTL);
>> +       ctl_val &= ~(SPM_CTL_INDEX << SPM_CTL_INDEX_SHIFT);
>> +       ctl_val |= start_index << SPM_CTL_INDEX_SHIFT;
>> +       ctl_val |= SPM_CTL_EN;
>> +       spm_register_write(drv, SPM_REG_SPM_CTL, ctl_val);
>> +
>> +       /* Ensure we have written the start address */
>> +       wmb();
>
>Can you explain please what this wmb is meant to achieve ? If it is there
>to make sure the write made it to the SPM that's not a proper way
>to achieve it, the barrier ensures ordering, not write completion.
>
A write completion would be great, but assuming that the context switch
using SCM or WFI could ensure the flush, then ordering is probably good
enough for this.

Is my understanding incorrect?

>> +
>> +       return 0;
>> +}
>> +
>> +static int qcom_pm_collapse(unsigned long int unused)
>> +{
>> +       int ret;
>> +       u32 flag;
>> +       int cpu = smp_processor_id();
>> +
>> +       ret = scm_set_warm_boot_addr(cpu_resume, cpu);
>> +       if (ret)
>> +               return ret;
>> +
>> +       flag = SCM_L2_ON & SCM_FLUSH_FLAG_MASK;
>> +       scm_call_atomic1(SCM_SVC_BOOT, SCM_CMD_TERMINATE_PC, flag);
>> +
>> +       /*
>> +        * Returns here only if there was a pending interrupt and we did not
>> +        * power down as a result.
>> +        */
>> +       return 0;
>
>I know the return value is changed by the cpu_suspend implementation to
>signal error when not returning from cpu_resume, but you should not
>return 0 anyway.
>
Sure. Will do.
>> +}
>> +
>> +static int qcom_cpu_standby(void *unused)
>> +{
>> +       int ret;
>> +
>> +       ret = spm_set_low_power_mode(PM_SLEEP_MODE_STBY);
>> +       if (ret)
>> +               return ret;
>> +
>> +       cpu_do_idle();
>> +
>> +       return 0;
>> +}
>> +
>> +static int qcom_cpu_spc(void *unused)
>> +{      int ret;
>> +
>> +       ret = spm_set_low_power_mode(PM_SLEEP_MODE_STBY);
>> +       if (ret)
>> +               return ret;
>> +
>> +       cpu_pm_enter();
>> +       cpu_suspend(0, qcom_pm_collapse);
>
>You must stash the return value and propagate it.
>
Sure. Will do.
>> +       cpu_pm_exit();
>> +
>> +       return 0;
>> +}
>> +
>> +static struct spm_driver_data *spm_get_drv(struct platform_device *pdev,
>> +               int *spm_cpu)
>> +{
>> +       struct spm_driver_data *drv = NULL;
>> +       struct device_node *cpu_node, *saw_node;
>> +       int cpu;
>> +       bool found = false;
>> +
>> +       for_each_possible_cpu(cpu) {
>> +               cpu_node = of_cpu_device_node_get(cpu);
>> +               if (!cpu_node)
>> +                       continue;
>> +               saw_node = of_parse_phandle(cpu_node, "qcom,saw", 0);
>
>+               if (saw_node) {
>+                       if (saw_node == pdev->dev.of_node)
>+                               found = true;
>+                       of_node_put(saw_node);
>+               }
>
>can be:
>		found = (saw_node == pdev->dev.of_node);
>		of_node_put(saw_node);
>
Sure.

>Thanks,
>Lorenzo
>
>> +               of_node_put(cpu_node);
>> +               if (found)
>> +                       break;
>> +       }
>> +
>> +       if (found) {
>> +               drv = devm_kzalloc(&pdev->dev, sizeof(*drv), GFP_KERNEL);
>> +               if (drv)
>> +                       *spm_cpu = cpu;
>> +       }
>> +
>> +       return drv;
>> +}
>> +
>> +static const struct of_device_id spm_match_table[] = {
>> +       { .compatible = "qcom,msm8974-saw2-v2.1-cpu",
>> +         .data = &spm_reg_8974_8084_cpu },
>> +       { .compatible = "qcom,apq8084-saw2-v2.1-cpu",
>> +         .data = &spm_reg_8974_8084_cpu },
>> +       { .compatible = "qcom,apq8064-saw2-v1.1-cpu",
>> +         .data = &spm_reg_8064_cpu },
>> +       { },
>> +};
>> +
>> +static const struct qcom_cpu_pm_ops lpm_ops = {
>> +       .standby = qcom_cpu_standby,
>> +       .spc = qcom_cpu_spc,
>> +};
>> +
>> +static int spm_dev_probe(struct platform_device *pdev)
>> +{
>> +       struct spm_driver_data *drv;
>> +       struct resource *res;
>> +       const struct of_device_id *match_id;
>> +       void __iomem *addr;
>> +       int cpu = -EINVAL;
>> +       static bool cpuidle_drv_init;
>> +       const struct platform_device_info qcom_cpuidle_info = {
>> +               .name   = "qcom_cpuidle",
>> +               .id     = -1,
>> +               .data = &lpm_ops,
>> +               .size_data = sizeof(lpm_ops),
>> +       };
>> +
>> +       drv = spm_get_drv(pdev, &cpu);
>> +       if (!drv || cpu < 0)
>> +               return -EINVAL;
>> +
>> +       res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +       drv->reg_base = devm_ioremap_resource(&pdev->dev, res);
>> +       if (IS_ERR(drv->reg_base))
>> +               return PTR_ERR(drv->reg_base);
>> +
>> +       match_id = of_match_node(spm_match_table, pdev->dev.of_node);
>> +       if (!match_id)
>> +               return -ENODEV;
>> +
>> +       drv->reg_data = match_id->data;
>> +
>> +       /* Write the SPM sequences first.. */
>> +       addr = drv->reg_base + drv->reg_data->reg_offset[SPM_REG_SEQ_ENTRY];
>> +       __iowrite32_copy(addr, drv->reg_data->seq,
>> +                       ARRAY_SIZE(drv->reg_data->seq) / 4);
>> +
>> +       /*
>> +        * ..and then the control registers.
>> +        * On some SoC's if the control registers are written first and if the
>> +        * CPU was held in reset, the reset signal could trigger the SPM state
>> +        * machine, before the sequences are completely written.
>> +        */
>> +       spm_register_write(drv, SPM_REG_CFG, drv->reg_data->spm_cfg);
>> +       spm_register_write(drv, SPM_REG_DLY, drv->reg_data->spm_dly);
>> +       spm_register_write(drv, SPM_REG_PMIC_DLY, drv->reg_data->pmic_dly);
>> +
>> +       spm_register_write(drv, SPM_REG_PMIC_DATA_0,
>> +                               drv->reg_data->pmic_data[0]);
>> +       spm_register_write(drv, SPM_REG_PMIC_DATA_1,
>> +                               drv->reg_data->pmic_data[1]);
>> +
>> +       /*
>> +        * Ensure all observers see the above register writes before the
>> +        * cpuidle driver is allowed to use the SPM.
>> +        */
>> +       wmb();
>> +       per_cpu(cpu_spm_drv, cpu) = drv;
>> +
>> +       if (!cpuidle_drv_init) {
>> +               platform_device_register_full(&qcom_cpuidle_info);
>> +               cpuidle_drv_init = true;
>> +       }
>> +
>> +       return 0;
>> +}
>> +
>> +static struct platform_driver spm_driver = {
>> +       .probe = spm_dev_probe,
>> +       .driver = {
>> +               .name = "saw",
>> +               .of_match_table = spm_match_table,
>> +       },
>> +};
>> +
>> +module_platform_driver(spm_driver);
>> +
>> +MODULE_LICENSE("GPL v2");
>> +MODULE_DESCRIPTION("SAW power controller driver");
>> +MODULE_ALIAS("platform:saw");
>> diff --git a/include/soc/qcom/pm.h b/include/soc/qcom/pm.h
>> new file mode 100644
>> index 0000000..d9a56d7
>> --- /dev/null
>> +++ b/include/soc/qcom/pm.h
>> @@ -0,0 +1,31 @@
>> +/*
>> + * Copyright (c) 2009-2014, The Linux Foundation. All rights reserved.
>> + *
>> + * This software is licensed under the terms of the GNU General Public
>> + * License version 2, as published by the Free Software Foundation, and
>> + * may be copied, distributed, and modified under those terms.
>> + *
>> + * This program is distributed in the hope that 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.
>> + *
>> + */
>> +
>> +#ifndef __QCOM_PM_H
>> +#define __QCOM_PM_H
>> +
>> +enum pm_sleep_mode {
>> +       PM_SLEEP_MODE_STBY,
>> +       PM_SLEEP_MODE_RET,
>> +       PM_SLEEP_MODE_SPC,
>> +       PM_SLEEP_MODE_PC,
>> +       PM_SLEEP_MODE_NR,
>> +};
>> +
>> +struct qcom_cpu_pm_ops {
>> +       int (*standby)(void *data);
>> +       int (*spc)(void *data);
>> +};
>> +
>> +#endif  /* __QCOM_PM_H */
>> --
>> 2.1.0
>>
>>
>> _______________________________________________
>> linux-arm-kernel mailing list
>> linux-arm-kernel@lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>>
>--
>To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Daniel Lezcano]

On 12/01/2014 07:50 PM, Lina Iyer wrote:
> On Thu, Nov 27 2014 at 01:52 -0700, Daniel Lezcano wrote:
>> On 11/27/2014 06:24 AM, Lina Iyer wrote:
>
>> +    static bool cpuidle_drv_init;
>>
>>        ^^^^^^^^^
>>
>> As already said in a previous comment, please find a way to remove that.
>>
> I will look into it. Stephen and I wanted the cpuidle driver to be
> probed only after any of the SPMs are ready. And possibly, only for that
> cpu, for which the SPM has been probed.
>
> To achieve the SPM -> CPUIDLE Device relation, I havent found a good way
> to do that. Without using CPUIDLE_MULTIPLE_DRIVERS, initializing each
> cpuidle device, separate from the cpuidle driver, requires that I update
> the cpuidle_driver->cpumask after probing each SPM device, to allow for
> only one driver and cpuidle devices only for the probed cpus. Using the
> cpuidle_register_driver(), resets the drv->refcnt in
> __cpuidle_driver_init.
>
> I may need something like this in the else clause of
> CPUIDLE_MULTIPLE_DRIVERS -
>
> static inline struct cpuidle_driver *__cpuidle_get_cpu_driver(int cpu)
> {
>         struct cpuidle_driver *drv = cpuidle_curr_driver;
>
>         if (drv && !cpumask_test_cpu(cpu, drv->cpumask))
>                  drv = NULL;
>
>         return drv;
> }

You don't have to deal with the drv->cpumask and 
CPUIDLE_MULTIPLE_DRIVERS. This field gives the list of the cpus the 
driver will have to potentially handle.

Just register the driver at the first place with the platform device by 
using cpuidle_register_driver(drv), I suggest somewhere else than the 
spm code.

The underlying code will do:

         if (!drv->cpumask)
                 drv->cpumask = (struct cpumask *)cpu_possible_mask;

Until the cpuidle device is not initialized for a specific cpu, the 
cpuidle code will have no effect for this cpu. If you register the 
driver but without registering the cpuidle devices that will result in 
nothing more than a structure initialized but inoperative.

For each SPM being probed:

struct cpuidle_device *dev = &per_cpu(cpuidle_dev, cpu);
dev->cpu = cpu;
cpuidle_register_device(dev);

Note cpuidle_dev is exported via cpuidle.h, so you don't have to 
redeclare a per cpu cpuidle device by your own.

So rephrasing all that:

(1) in cpuidle-qcom register the driver
(2) in the spm code register the device for each spm probe

(1) being called before (2).

If after that in the code you still have a "static bool 
cpuidle_drv_init", then I guess we have a problem in the init sequence 
somewhere else.

> void cpuidle_update_cpumask(struct cpumask *mask) {
>         struct cpuidle_driver *drv;
>
>         spin_lock(&cpuidle_driver_lock);
>         drv = cpuidle_get_driver();
>      if (drv)
>          drv->cpumask = mask ?: cpu_possible_mask;
>         spin_unlock(&cpuidle_driver_lock);
> }
>
> With that, I could register cpuidle driver the first time when the mask
> changes
> from empty and consequent updates would just update the cpumask. (I am not
> sure, if I missed anything in this change). It just seemed far too
> invasive at
> this time, in lieu of the static bool.
>
>>> +    const struct platform_device_info qcom_cpuidle_info = {
>>> +        .name    = "qcom_cpuidle",
>>> +        .id    = -1,
>>> +        .data = &lpm_ops,
>>> +        .size_data = sizeof(lpm_ops),
>>> +    };
>>> +
>>> +    drv = spm_get_drv(pdev, &cpu);
>>> +    if (!drv || cpu < 0)
>>> +        return -EINVAL;
>>
>> As already said in a previous comment, it is not possible to have "cpu
>> < 0" with "drv != NULL", so except I am missing something the test
>> should be:
>>
>>     if (!drv)
>>         return -EINVAL;
>>
> Sorry, done.
>
>> There is something wrong with the init sequence. Don't you find weird
>> you have to backward search for the cpu belonging to the pdev each
>> time the probe function is called ?
>>
>>
> Well, it was that or the SPM device node pointing to the CPU that it
> references. It seems more common to have an iterator than the doubly
> linked device nodes. I dont have a strong preference either way, just
> chose the way that made device nodes easier.
>
>>> +
>>> +    res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>>> +    drv->reg_base = devm_ioremap_resource(&pdev->dev, res);
>>> +    if (IS_ERR(drv->reg_base))
>>> +        return PTR_ERR(drv->reg_base);
>>> +
>>> +    match_id = of_match_node(spm_match_table, pdev->dev.of_node);
>>> +    if (!match_id)
>>> +        return -ENODEV;
>>> +
>>> +    drv->reg_data = match_id->data;
>>> +
>>> +    /* Write the SPM sequences first.. */
>>> +    addr = drv->reg_base +
>>> drv->reg_data->reg_offset[SPM_REG_SEQ_ENTRY];
>>> +    __iowrite32_copy(addr, drv->reg_data->seq,
>>> +            ARRAY_SIZE(drv->reg_data->seq) / 4);
>>> +
>>> +    /*
>>> +     * ..and then the control registers.
>>> +     * On some SoC's if the control registers are written first and
>>> if the
>>> +     * CPU was held in reset, the reset signal could trigger the SPM
>>> state
>>> +     * machine, before the sequences are completely written.
>>> +     */
>>> +    spm_register_write(drv, SPM_REG_CFG, drv->reg_data->spm_cfg);
>>> +    spm_register_write(drv, SPM_REG_DLY, drv->reg_data->spm_dly);
>>> +    spm_register_write(drv, SPM_REG_PMIC_DLY, drv->reg_data->pmic_dly);
>>> +
>>> +    spm_register_write(drv, SPM_REG_PMIC_DATA_0,
>>> +                drv->reg_data->pmic_data[0]);
>>> +    spm_register_write(drv, SPM_REG_PMIC_DATA_1,
>>> +                drv->reg_data->pmic_data[1]);
>>> +
>>> +    /*
>>> +     * Ensure all observers see the above register writes before the
>>> +     * cpuidle driver is allowed to use the SPM.
>>> +     */
>>> +    wmb();
>>> +    per_cpu(cpu_spm_drv, cpu) = drv;
>>> +
>>> +    if (!cpuidle_drv_init) {
>>> +        platform_device_register_full(&qcom_cpuidle_info);
>>> +        cpuidle_drv_init = true;
>>> +    }
>>> +
>>> +    return 0;
>>> +}
>>> +
>>> +static struct platform_driver spm_driver = {
>>> +    .probe = spm_dev_probe,
>>> +    .driver = {
>>> +        .name = "saw",
>>> +        .of_match_table = spm_match_table,
>>> +    },
>>> +};
>>> +
>>> +module_platform_driver(spm_driver);
>>> +
>>> +MODULE_LICENSE("GPL v2");
>>> +MODULE_DESCRIPTION("SAW power controller driver");
>>> +MODULE_ALIAS("platform:saw");
>>> diff --git a/include/soc/qcom/pm.h b/include/soc/qcom/pm.h
>>> new file mode 100644
>>> index 0000000..d9a56d7
>>> --- /dev/null
>>> +++ b/include/soc/qcom/pm.h
>>> @@ -0,0 +1,31 @@
>>> +/*
>>> + * Copyright (c) 2009-2014, The Linux Foundation. All rights reserved.
>>> + *
>>> + * This software is licensed under the terms of the GNU General Public
>>> + * License version 2, as published by the Free Software Foundation, and
>>> + * may be copied, distributed, and modified under those terms.
>>> + *
>>> + * This program is distributed in the hope that 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.
>>> + *
>>> + */
>>> +
>>> +#ifndef __QCOM_PM_H
>>> +#define __QCOM_PM_H
>>> +
>>> +enum pm_sleep_mode {
>>> +    PM_SLEEP_MODE_STBY,
>>> +    PM_SLEEP_MODE_RET,
>>> +    PM_SLEEP_MODE_SPC,
>>> +    PM_SLEEP_MODE_PC,
>>> +    PM_SLEEP_MODE_NR,
>>> +};
>>> +
>>> +struct qcom_cpu_pm_ops {
>>> +    int (*standby)(void *data);
>>> +    int (*spc)(void *data);
>>> +};
>>> +
>>> +#endif  /* __QCOM_PM_H */
>>>
>>
>>
>> --
>> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
>>
>> Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
>> <http://twitter.com/#!/linaroorg> Twitter |
>> <http://www.linaro.org/linaro-blog/> Blog
>>


-- 
  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Catalin Marinas]

On Mon, Dec 01, 2014 at 06:57:05PM +0000, Lina Iyer wrote:
> On Thu, Nov 27 2014 at 08:01 -0700, Lorenzo Pieralisi wrote:
> >On Thu, Nov 27, 2014 at 05:24:07AM +0000, Lina Iyer wrote:
> >
> >[...]
> >
> >> +static int spm_set_low_power_mode(enum pm_sleep_mode mode)
> >> +{
> >> +       struct spm_driver_data *drv = per_cpu(cpu_spm_drv,
> >> +                                               smp_processor_id());
> >> +       u32 start_index;
> >> +       u32 ctl_val;
> >> +
> >> +       if (!drv)
> >> +               return -ENXIO;
> >> +
> >> +       start_index = drv->reg_data->start_index[mode];
> >> +
> >> +       ctl_val = spm_register_read(drv, SPM_REG_SPM_CTL);
> >> +       ctl_val &= ~(SPM_CTL_INDEX << SPM_CTL_INDEX_SHIFT);
> >> +       ctl_val |= start_index << SPM_CTL_INDEX_SHIFT;
> >> +       ctl_val |= SPM_CTL_EN;
> >> +       spm_register_write(drv, SPM_REG_SPM_CTL, ctl_val);
> >> +
> >> +       /* Ensure we have written the start address */
> >> +       wmb();
> >
> >Can you explain please what this wmb is meant to achieve ? If it is there
> >to make sure the write made it to the SPM that's not a proper way
> >to achieve it, the barrier ensures ordering, not write completion.
>
> A write completion would be great, but assuming that the context switch
> using SCM or WFI could ensure the flush, then ordering is probably good
> enough for this.

I assume you want the write to the device to complete (side-effects are
observed) before the WFI (which would raise a StandbyWFI signal). The
safer way is to read back from the device (or check/poll for the device
state change) rather than a barrier.

-- 
Catalin

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Lina Iyer]

On Tue, Dec 02 2014 at 02:53 -0700, Daniel Lezcano wrote:
>On 12/01/2014 07:50 PM, Lina Iyer wrote:
>>On Thu, Nov 27 2014 at 01:52 -0700, Daniel Lezcano wrote:
>>>On 11/27/2014 06:24 AM, Lina Iyer wrote:
>>
>>>+    static bool cpuidle_drv_init;
>>>
>>>       ^^^^^^^^^
>>>
>>>As already said in a previous comment, please find a way to remove that.
>>>
>>I will look into it. Stephen and I wanted the cpuidle driver to be
>>probed only after any of the SPMs are ready. And possibly, only for that
>>cpu, for which the SPM has been probed.
>>
>>To achieve the SPM -> CPUIDLE Device relation, I havent found a good way
>>to do that. Without using CPUIDLE_MULTIPLE_DRIVERS, initializing each
>>cpuidle device, separate from the cpuidle driver, requires that I update
>>the cpuidle_driver->cpumask after probing each SPM device, to allow for
>>only one driver and cpuidle devices only for the probed cpus. Using the
>>cpuidle_register_driver(), resets the drv->refcnt in
>>__cpuidle_driver_init.
>>
>>I may need something like this in the else clause of
>>CPUIDLE_MULTIPLE_DRIVERS -
>>
>>static inline struct cpuidle_driver *__cpuidle_get_cpu_driver(int cpu)
>>{
>>        struct cpuidle_driver *drv = cpuidle_curr_driver;
>>
>>        if (drv && !cpumask_test_cpu(cpu, drv->cpumask))
>>                 drv = NULL;
>>
>>        return drv;
>>}
>
>You don't have to deal with the drv->cpumask and 
>CPUIDLE_MULTIPLE_DRIVERS. This field gives the list of the cpus the 
>driver will have to potentially handle.
>
>Just register the driver at the first place with the platform device 
>by using cpuidle_register_driver(drv), I suggest somewhere else than 
>the spm code.
>
>The underlying code will do:
>
>        if (!drv->cpumask)
>                drv->cpumask = (struct cpumask *)cpu_possible_mask;
>
>Until the cpuidle device is not initialized for a specific cpu, the 
>cpuidle code will have no effect for this cpu. If you register the 
>driver but without registering the cpuidle devices that will result in 
>nothing more than a structure initialized but inoperative.
>
>For each SPM being probed:
>
>struct cpuidle_device *dev = &per_cpu(cpuidle_dev, cpu);
>dev->cpu = cpu;
>cpuidle_register_device(dev);
>
>Note cpuidle_dev is exported via cpuidle.h, so you don't have to 
>redeclare a per cpu cpuidle device by your own.
>
>So rephrasing all that:
>
>(1) in cpuidle-qcom register the driver
>(2) in the spm code register the device for each spm probe
>
>(1) being called before (2).
>
>If after that in the code you still have a "static bool 
>cpuidle_drv_init", then I guess we have a problem in the init sequence 
>somewhere else.
>
Okay, Thanks. Last night I was able to get a patch working with one
cpuidle_register_driver and multiple cpuidle_register_device. The issue
using an module_init() was -

1. Register a device to probe the cpuidle driver.
2. Register cpuidle devices for every SPM probed.

What used to happen is that the cpuidle driver probe always happens 
after the cpuidle devices were created.

I have a solution where I register for the cpuidle device to probe a
driver, at postcore_initcall() at which point, the cpuidle framework
would also be available.

I am not sure how elegant it looks, but thats the latest initcall I
could get a consistent driver initialization before the SPM probe
(therefore the device registration).

I will send the patch this morning.

>>void cpuidle_update_cpumask(struct cpumask *mask) {
>>        struct cpuidle_driver *drv;
>>
>>        spin_lock(&cpuidle_driver_lock);
>>        drv = cpuidle_get_driver();
>>     if (drv)
>>         drv->cpumask = mask ?: cpu_possible_mask;
>>        spin_unlock(&cpuidle_driver_lock);
>>}
>>
>>With that, I could register cpuidle driver the first time when the mask
>>changes
>>from empty and consequent updates would just update the cpumask. (I am not
>>sure, if I missed anything in this change). It just seemed far too
>>invasive at
>>this time, in lieu of the static bool.
>>

>>>>+    const struct platform_device_info qcom_cpuidle_info = {
>>>>+        .name    = "qcom_cpuidle",
>>>>+        .id    = -1,
>>>>+        .data = &lpm_ops,
>>>>+        .size_data = sizeof(lpm_ops),
>>>>+    };
>>>>+
>>>>+    drv = spm_get_drv(pdev, &cpu);
>>>>+    if (!drv || cpu < 0)
>>>>+        return -EINVAL;
>>>
>>>As already said in a previous comment, it is not possible to have "cpu
>>>< 0" with "drv != NULL", so except I am missing something the test
>>>should be:
>>>
>>>    if (!drv)
>>>        return -EINVAL;
>>>
>>Sorry, done.
>>
>>>There is something wrong with the init sequence. Don't you find weird
>>>you have to backward search for the cpu belonging to the pdev each
>>>time the probe function is called ?
>>>
>>>
>>Well, it was that or the SPM device node pointing to the CPU that it
>>references. It seems more common to have an iterator than the doubly
>>linked device nodes. I dont have a strong preference either way, just
>>chose the way that made device nodes easier.
>>
This one, I havent addressed. How strongly do you feel about this? I
still have to iterate through all the cpus to find the cpu node that
matches the handle specified in the SAW node.

Thanks,
Lina

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Daniel Lezcano]

On 12/02/2014 04:35 PM, Lina Iyer wrote:
> On Tue, Dec 02 2014 at 02:53 -0700, Daniel Lezcano wrote:
>> On 12/01/2014 07:50 PM, Lina Iyer wrote:
>>> On Thu, Nov 27 2014 at 01:52 -0700, Daniel Lezcano wrote:
>>>> On 11/27/2014 06:24 AM, Lina Iyer wrote:
>>>
>>>> +    static bool cpuidle_drv_init;
>>>>
>>>>       ^^^^^^^^^
>>>>
>>>> As already said in a previous comment, please find a way to remove
>>>> that.
>>>>
>>> I will look into it. Stephen and I wanted the cpuidle driver to be
>>> probed only after any of the SPMs are ready. And possibly, only for that
>>> cpu, for which the SPM has been probed.
>>>
>>> To achieve the SPM -> CPUIDLE Device relation, I havent found a good way
>>> to do that. Without using CPUIDLE_MULTIPLE_DRIVERS, initializing each
>>> cpuidle device, separate from the cpuidle driver, requires that I update
>>> the cpuidle_driver->cpumask after probing each SPM device, to allow for
>>> only one driver and cpuidle devices only for the probed cpus. Using the
>>> cpuidle_register_driver(), resets the drv->refcnt in
>>> __cpuidle_driver_init.
>>>
>>> I may need something like this in the else clause of
>>> CPUIDLE_MULTIPLE_DRIVERS -
>>>
>>> static inline struct cpuidle_driver *__cpuidle_get_cpu_driver(int cpu)
>>> {
>>>        struct cpuidle_driver *drv = cpuidle_curr_driver;
>>>
>>>        if (drv && !cpumask_test_cpu(cpu, drv->cpumask))
>>>                 drv = NULL;
>>>
>>>        return drv;
>>> }
>>
>> You don't have to deal with the drv->cpumask and
>> CPUIDLE_MULTIPLE_DRIVERS. This field gives the list of the cpus the
>> driver will have to potentially handle.
>>
>> Just register the driver at the first place with the platform device
>> by using cpuidle_register_driver(drv), I suggest somewhere else than
>> the spm code.
>>
>> The underlying code will do:
>>
>>        if (!drv->cpumask)
>>                drv->cpumask = (struct cpumask *)cpu_possible_mask;
>>
>> Until the cpuidle device is not initialized for a specific cpu, the
>> cpuidle code will have no effect for this cpu. If you register the
>> driver but without registering the cpuidle devices that will result in
>> nothing more than a structure initialized but inoperative.
>>
>> For each SPM being probed:
>>
>> struct cpuidle_device *dev = &per_cpu(cpuidle_dev, cpu);
>> dev->cpu = cpu;
>> cpuidle_register_device(dev);
>>
>> Note cpuidle_dev is exported via cpuidle.h, so you don't have to
>> redeclare a per cpu cpuidle device by your own.
>>
>> So rephrasing all that:
>>
>> (1) in cpuidle-qcom register the driver
>> (2) in the spm code register the device for each spm probe
>>
>> (1) being called before (2).
>>
>> If after that in the code you still have a "static bool
>> cpuidle_drv_init", then I guess we have a problem in the init sequence
>> somewhere else.
>>
> Okay, Thanks. Last night I was able to get a patch working with one
> cpuidle_register_driver and multiple cpuidle_register_device. The issue
> using an module_init() was -
>
> 1. Register a device to probe the cpuidle driver.
> 2. Register cpuidle devices for every SPM probed.

That's it.

> What used to happen is that the cpuidle driver probe always happens
> after the cpuidle devices were created.
>
> I have a solution where I register for the cpuidle device to probe a
> driver, at postcore_initcall() at which point, the cpuidle framework
> would also be available.
>
> I am not sure how elegant it looks, but thats the latest initcall I
> could get a consistent driver initialization before the SPM probe
> (therefore the device registration).

I don't have the code in front of me but we can probably find a nice 
place to probe the cpuidle driver. Is there any pm.c file with an init 
function that can be useful ? In any case, regarding the unusual init 
sequence for the cpuidle driver, I think the probe would deserve a comment.

> I will send the patch this morning.

Ok, cool.

[ ... ]

>>>> There is something wrong with the init sequence. Don't you find weird
>>>> you have to backward search for the cpu belonging to the pdev each
>>>> time the probe function is called ?
>>>>
>>>>
>>> Well, it was that or the SPM device node pointing to the CPU that it
>>> references. It seems more common to have an iterator than the doubly
>>> linked device nodes. I dont have a strong preference either way, just
>>> chose the way that made device nodes easier.
>>>
> This one, I havent addressed. How strongly do you feel about this? I
> still have to iterate through all the cpus to find the cpu node that
> matches the handle specified in the SAW node.

IMO, we can live with that if there is no longer any static variable 
around telling us we initialized something.

Thanks !

   -- Daniel


-- 
  <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

Re: [PATCH v13 03/10] qcom: spm: Add Subsystem Power Manager driver

[Lina Iyer]

On Tue, Dec 02 2014 at 04:10 -0700, Catalin Marinas wrote:
>On Mon, Dec 01, 2014 at 06:57:05PM +0000, Lina Iyer wrote:
>> On Thu, Nov 27 2014 at 08:01 -0700, Lorenzo Pieralisi wrote:
>> >On Thu, Nov 27, 2014 at 05:24:07AM +0000, Lina Iyer wrote:
>> >
>> >[...]
>> >
>> >> +static int spm_set_low_power_mode(enum pm_sleep_mode mode)
>> >> +{
>> >> +       struct spm_driver_data *drv = per_cpu(cpu_spm_drv,
>> >> +                                               smp_processor_id());
>> >> +       u32 start_index;
>> >> +       u32 ctl_val;
>> >> +
>> >> +       if (!drv)
>> >> +               return -ENXIO;
>> >> +
>> >> +       start_index = drv->reg_data->start_index[mode];
>> >> +
>> >> +       ctl_val = spm_register_read(drv, SPM_REG_SPM_CTL);
>> >> +       ctl_val &= ~(SPM_CTL_INDEX << SPM_CTL_INDEX_SHIFT);
>> >> +       ctl_val |= start_index << SPM_CTL_INDEX_SHIFT;
>> >> +       ctl_val |= SPM_CTL_EN;
>> >> +       spm_register_write(drv, SPM_REG_SPM_CTL, ctl_val);
>> >> +
>> >> +       /* Ensure we have written the start address */
>> >> +       wmb();
>> >
>> >Can you explain please what this wmb is meant to achieve ? If it is there
>> >to make sure the write made it to the SPM that's not a proper way
>> >to achieve it, the barrier ensures ordering, not write completion.
>>
>> A write completion would be great, but assuming that the context switch
>> using SCM or WFI could ensure the flush, then ordering is probably good
>> enough for this.
>
>I assume you want the write to the device to complete (side-effects are
>observed) before the WFI (which would raise a StandbyWFI signal). The
>safer way is to read back from the device (or check/poll for the device
>state change) rather than a barrier.

Polling back to see if the register write was indeed successful, seems
a bit more than what I would like on a low latency mode like Standby.
Let me check with the hardware engineers, if a write ordering is
required. It seems to me like I am not getting much out of the wmb()
anyways. I will find out what kind of guarantees I would need to ensure
the SPM write is successful.

Thanks for your time, Catalin.

Thanks,
Lina.