2015-07-17 20:57:03

by Dave Gerlach

[permalink] [raw]
Subject: [PATCH v2 0/3] soc: ti: Introduce wkup_m3_ipc driver

Hi,

This series is version 2 of the code to introduce a wkup_m3_ipc driver
to handle communication between the MPU and Cortex M3 present on TI AM335x
and AM437x SoCs. v1 of this series can be found at [1]. Changes include:

- Buildable as a module
- Added am437x support
- Various cleanups and fixes based on comments on v1
- Ability to use mailbox in noirq mode for cpuidle on am335x

v2 contains an additional patch for the omap mailbox driver now to allow
us to set ti,mbox-send-noirq for the wkup_m3 mailbox to allow us to
support cpuidle on am335x. Although we can rely on interrupts during
the suspend path, we must send a message during the cpuidle path from
noirq context so we must have the ability to do this without using
an interrupt, so we introduce the flag to indicate this. The patch has
been included here with the wkup_m3_ipc patch so that the usage and
context is clear.

This series uses the wkup_m3_rproc driver which is merged as of v4.2-rc1,
but the required dt nodes are not yet merged and can be found here [2].
A full branch containing all necessary PM code for both am335x and am437x
has been pushed here [3] a big picture view of the plan for this series.

This driver relies on the firmware at [4] being present in /lib/firmware
in the rootfs or built in to the kernel.

Regards,
Dave

[1] http://www.spinics.net/lists/arm-kernel/msg387990.html
[2] http://www.spinics.net/lists/linux-omap/msg119973.html
[3] https://github.com/dgerlach/linux-pm/tree/pm-v4.2-rc2-amx3-suspend
[4] https://git.ti.com/ti-cm3-pm-firmware

Dave Gerlach (3):
mailbox/omap: Add ti,mbox-send-noirq quirk to fix AM33xx CPU Idle
Documentation: dt: add bindings for TI Wakeup M3 IPC device
soc: ti: Add wkup_m3_ipc driver

.../devicetree/bindings/mailbox/omap-mailbox.txt | 8 +
.../devicetree/bindings/soc/ti/wkup_m3_ipc.txt | 57 +++
drivers/mailbox/omap-mailbox.c | 49 ++-
drivers/soc/ti/Kconfig | 10 +
drivers/soc/ti/Makefile | 1 +
drivers/soc/ti/wkup_m3_ipc.c | 486 +++++++++++++++++++++
include/linux/wkup_m3_ipc.h | 30 ++
7 files changed, 637 insertions(+), 4 deletions(-)
create mode 100644 Documentation/devicetree/bindings/soc/ti/wkup_m3_ipc.txt
create mode 100644 drivers/soc/ti/wkup_m3_ipc.c
create mode 100644 include/linux/wkup_m3_ipc.h

--
2.4.5


2015-07-17 20:57:15

by Dave Gerlach

[permalink] [raw]
Subject: [PATCH v2 1/3] mailbox/omap: Add ti,mbox-send-noirq quirk to fix AM33xx CPU Idle

The mailbox framework controls the transmission queue and requires
either its controller implementations or clients to run the state
machine for the Tx queue. The OMAP mailbox controller uses a Tx-ready
interrupt as the equivalent of a Tx-done interrupt to run this Tx
queue state-machine.

The WkupM3 processor on AM33xx and AM43xx SoCs is used to offload
certain PM tasks, like doing the necessary operations for Device
PM suspend/resume or for entering lower c-states during cpuidle.

The CPUIdle on AM33xx requires the messages to be sent without
having to trigger the Tx-ready interrupts, as the interrupt
would immediately terminate the CPUIdle operation. Support for
this has been added by introducing a DT quirk, "ti,mbox-send-noirq"
and using it to modify the normal OMAP mailbox controller behavior
on the sub-mailboxes used to communicate with the WkupM3 remote
processor. This also requires the wkup_m3_ipc driver to adjust
its mailbox usage logic to run the Tx state machine.

NOTE:
- AM43xx does not communicate with WkupM3 for CPU Idle, so is
not affected by this behavior. But, it uses the same IPC driver
for PM suspend/resume functionality, so requires the quirk as
well, because of changes to the common wkup_m3_ipc driver.

Signed-off-by: Dave Gerlach <[email protected]>
[[email protected]: revise logic and update comments/patch description]
Signed-off-by: Suman Anna <[email protected]>
---
.../devicetree/bindings/mailbox/omap-mailbox.txt | 8 ++++
drivers/mailbox/omap-mailbox.c | 49 ++++++++++++++++++++--
2 files changed, 53 insertions(+), 4 deletions(-)

diff --git a/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt b/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
index d1a0433..9b40c49 100644
--- a/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
+++ b/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
@@ -75,6 +75,14 @@ data that represent the following:
Cell #3 (usr_id) - mailbox user id for identifying the interrupt line
associated with generating a tx/rx fifo interrupt.

+Optional Properties:
+--------------------
+- ti,mbox-send-noirq: Quirk flag to allow the client user of this sub-mailbox
+ to send messages without triggering a Tx ready interrupt,
+ and to control the Tx ticker. Should be used only on
+ sub-mailboxes used to communicate with WkupM3 remote
+ processor on AM33xx/AM43xx SoCs.
+
Mailbox Users:
==============
A device needing to communicate with a target processor device should specify
diff --git a/drivers/mailbox/omap-mailbox.c b/drivers/mailbox/omap-mailbox.c
index a3dbfd9..b7f636f 100644
--- a/drivers/mailbox/omap-mailbox.c
+++ b/drivers/mailbox/omap-mailbox.c
@@ -38,6 +38,8 @@
#include <linux/mailbox_controller.h>
#include <linux/mailbox_client.h>

+#include "mailbox.h"
+
#define MAILBOX_REVISION 0x000
#define MAILBOX_MESSAGE(m) (0x040 + 4 * (m))
#define MAILBOX_FIFOSTATUS(m) (0x080 + 4 * (m))
@@ -106,6 +108,7 @@ struct omap_mbox_fifo_info {
int rx_irq;

const char *name;
+ bool send_no_irq;
};

struct omap_mbox {
@@ -119,6 +122,7 @@ struct omap_mbox {
u32 ctx[OMAP4_MBOX_NR_REGS];
u32 intr_type;
struct mbox_chan *chan;
+ bool send_no_irq;
};

/* global variables for the mailbox devices */
@@ -418,6 +422,9 @@ static int omap_mbox_startup(struct omap_mbox *mbox)
goto fail_request_irq;
}

+ if (mbox->send_no_irq)
+ mbox->chan->txdone_method = TXDONE_BY_ACK;
+
_omap_mbox_enable_irq(mbox, IRQ_RX);

return 0;
@@ -586,13 +593,27 @@ static void omap_mbox_chan_shutdown(struct mbox_chan *chan)
mutex_unlock(&mdev->cfg_lock);
}

-static int omap_mbox_chan_send_data(struct mbox_chan *chan, void *data)
+static int omap_mbox_chan_send_noirq(struct omap_mbox *mbox, void *data)
{
- struct omap_mbox *mbox = mbox_chan_to_omap_mbox(chan);
int ret = -EBUSY;

- if (!mbox)
- return -EINVAL;
+ if (!mbox_fifo_full(mbox)) {
+ _omap_mbox_enable_irq(mbox, IRQ_RX);
+ mbox_fifo_write(mbox, (mbox_msg_t)data);
+ ret = 0;
+ _omap_mbox_disable_irq(mbox, IRQ_RX);
+
+ /* we must read and ack the interrupt directly from here */
+ mbox_fifo_read(mbox);
+ ack_mbox_irq(mbox, IRQ_RX);
+ }
+
+ return ret;
+}
+
+static int omap_mbox_chan_send(struct omap_mbox *mbox, void *data)
+{
+ int ret = -EBUSY;

if (!mbox_fifo_full(mbox)) {
mbox_fifo_write(mbox, (mbox_msg_t)data);
@@ -604,6 +625,22 @@ static int omap_mbox_chan_send_data(struct mbox_chan *chan, void *data)
return ret;
}

+static int omap_mbox_chan_send_data(struct mbox_chan *chan, void *data)
+{
+ struct omap_mbox *mbox = mbox_chan_to_omap_mbox(chan);
+ int ret;
+
+ if (!mbox)
+ return -EINVAL;
+
+ if (mbox->send_no_irq)
+ ret = omap_mbox_chan_send_noirq(mbox, data);
+ else
+ ret = omap_mbox_chan_send(mbox, data);
+
+ return ret;
+}
+
static const struct mbox_chan_ops omap_mbox_chan_ops = {
.startup = omap_mbox_chan_startup,
.send_data = omap_mbox_chan_send_data,
@@ -732,6 +769,9 @@ static int omap_mbox_probe(struct platform_device *pdev)
finfo->rx_usr = tmp[2];

finfo->name = child->name;
+
+ if (of_find_property(child, "ti,mbox-send-noirq", NULL))
+ finfo->send_no_irq = true;
} else {
finfo->tx_id = info->tx_id;
finfo->rx_id = info->rx_id;
@@ -791,6 +831,7 @@ static int omap_mbox_probe(struct platform_device *pdev)
fifo->irqstatus = MAILBOX_IRQSTATUS(intr_type, finfo->rx_usr);
fifo->irqdisable = MAILBOX_IRQDISABLE(intr_type, finfo->rx_usr);

+ mbox->send_no_irq = finfo->send_no_irq;
mbox->intr_type = intr_type;

mbox->parent = mdev;
--
2.4.5

2015-07-17 20:58:10

by Dave Gerlach

[permalink] [raw]
Subject: [PATCH v2 2/3] Documentation: dt: add bindings for TI Wakeup M3 IPC device

Add the device tree bindings document for the TI Wakeup M3 IPC
device on AM33xx and AM43xx SoCs. These devices are used by the
TI wkup_m3_ipc driver, and contain the registers upon which the
IPC protocol to communicate with the Wakeup M3 processor is
implemented.

Signed-off-by: Dave Gerlach <[email protected]>
Signed-off-by: Suman Anna <[email protected]>
---
.../devicetree/bindings/soc/ti/wkup_m3_ipc.txt | 57 ++++++++++++++++++++++
1 file changed, 57 insertions(+)
create mode 100644 Documentation/devicetree/bindings/soc/ti/wkup_m3_ipc.txt

diff --git a/Documentation/devicetree/bindings/soc/ti/wkup_m3_ipc.txt b/Documentation/devicetree/bindings/soc/ti/wkup_m3_ipc.txt
new file mode 100644
index 0000000..4015504
--- /dev/null
+++ b/Documentation/devicetree/bindings/soc/ti/wkup_m3_ipc.txt
@@ -0,0 +1,57 @@
+Wakeup M3 IPC Driver
+=====================
+
+The TI AM33xx and AM43xx family of devices use a small Cortex M3 co-processor
+(commonly referred to as Wakeup M3 or CM3) to help with various low power tasks
+that cannot be controlled from the MPU, like suspend/resume and certain deep
+C-states for CPU Idle. Once the wkup_m3_ipc driver uses the wkup_m3_rproc driver
+to boot the wkup_m3, it handles communication with the CM3 using IPC registers
+present in the SoC's control module and a mailbox. The wkup_m3_ipc exposes an
+API to allow the SoC PM code to execute specific PM tasks.
+
+Wkup M3 Device Node:
+====================
+A wkup_m3_ipc device node is used to represent the IPC registers within an
+SoC.
+
+Required properties:
+--------------------
+- compatible: Should be,
+ "ti,am3352-wkup-m3-ipc" for AM33xx SoCs
+ "ti,am4372-wkup-m3-ipc" for AM43xx SoCs
+- reg: Contains the IPC register address space to communicate
+ with the Wakeup M3 processor
+- interrupts: Contains the interrupt information for the wkup_m3
+ interrupt that signals the MPU.
+- ti,rproc: phandle to the wkup_m3 rproc node so the IPC driver
+ can boot it.
+- mboxes: phandles used by IPC framework to get correct mbox
+ channel for communication. Must point to appropriate
+ mbox_wkupm3 child node.
+
+Example:
+--------
+/* AM33xx */
+ l4_wkup: l4_wkup@44c00000 {
+ ...
+
+ scm: scm@210000 {
+ compatible = "ti,am3-scm", "simple-bus";
+ reg = <0x210000 0x2000>;
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges = <0 0x210000 0x2000>;
+
+ ...
+
+ wkup_m3_ipc: wkup_m3_ipc@1324 {
+ compatible = "ti,am3352-wkup-m3-ipc";
+ reg = <0x1324 0x24>;
+ interrupts = <78>;
+ ti,rproc = <&wkup_m3>;
+ mboxes = <&mailbox &mbox_wkupm3>;
+ };
+
+ ...
+ };
+ };
--
2.4.5

2015-07-17 20:57:13

by Dave Gerlach

[permalink] [raw]
Subject: [PATCH v2 3/3] soc: ti: Add wkup_m3_ipc driver

Introduce a wkup_m3_ipc driver to handle communication between the MPU
and Cortex M3 wkup_m3 present on am335x.

This driver is responsible for actually booting the wkup_m3_rproc and
also handling all IPC which is done using the IPC registers in the control
module, a mailbox, and a separate interrupt back from the wkup_m3. A small
API is exposed for executing specific power commands, which include
configuring for low power mode, request a transition to a low power mode,
and status info on a previous transition.

Signed-off-by: Dave Gerlach <[email protected]>
---
drivers/soc/ti/Kconfig | 10 +
drivers/soc/ti/Makefile | 1 +
drivers/soc/ti/wkup_m3_ipc.c | 486 +++++++++++++++++++++++++++++++++++++++++++
include/linux/wkup_m3_ipc.h | 30 +++
4 files changed, 527 insertions(+)
create mode 100644 drivers/soc/ti/wkup_m3_ipc.c
create mode 100644 include/linux/wkup_m3_ipc.h

diff --git a/drivers/soc/ti/Kconfig b/drivers/soc/ti/Kconfig
index 7266b21..3557c5e 100644
--- a/drivers/soc/ti/Kconfig
+++ b/drivers/soc/ti/Kconfig
@@ -28,4 +28,14 @@ config KEYSTONE_NAVIGATOR_DMA

If unsure, say N.

+config WKUP_M3_IPC
+ tristate "TI AMx3 Wkup-M3 IPC Driver"
+ depends on WKUP_M3_RPROC
+ depends on OMAP2PLUS_MBOX
+ help
+ TI AM33XX and AM43XX have a Cortex M3, the Wakeup M3, to handle
+ low power transitions. This IPC driver provides the necessary API
+ to communicate and use the Wakeup M3 for PM features like suspend
+ resume and boots it using wkup_m3_rproc driver.
+
endif # SOC_TI
diff --git a/drivers/soc/ti/Makefile b/drivers/soc/ti/Makefile
index 135bdad..48ff3a7 100644
--- a/drivers/soc/ti/Makefile
+++ b/drivers/soc/ti/Makefile
@@ -4,3 +4,4 @@
obj-$(CONFIG_KEYSTONE_NAVIGATOR_QMSS) += knav_qmss.o
knav_qmss-y := knav_qmss_queue.o knav_qmss_acc.o
obj-$(CONFIG_KEYSTONE_NAVIGATOR_DMA) += knav_dma.o
+obj-$(CONFIG_WKUP_M3_IPC) += wkup_m3_ipc.o
diff --git a/drivers/soc/ti/wkup_m3_ipc.c b/drivers/soc/ti/wkup_m3_ipc.c
new file mode 100644
index 0000000..7eb05aa
--- /dev/null
+++ b/drivers/soc/ti/wkup_m3_ipc.c
@@ -0,0 +1,486 @@
+/*
+ * AMx3 Wkup M3 IPC driver
+ *
+ * Copyright (C) 2015 Texas Instruments, Inc.
+ *
+ * Dave Gerlach <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed 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/err.h>
+#include <linux/kernel.h>
+#include <linux/kthread.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/mailbox_client.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/omap-mailbox.h>
+#include <linux/platform_device.h>
+#include <linux/remoteproc.h>
+#include <linux/suspend.h>
+#include <linux/wkup_m3_ipc.h>
+
+#define AM33XX_CTRL_IPC_REG_COUNT 0x8
+#define AM33XX_CTRL_IPC_REG_OFFSET(m) (0x4 + 4 * (m))
+
+/* AM33XX M3_TXEV_EOI register */
+#define AM33XX_CONTROL_M3_TXEV_EOI 0x00
+
+#define AM33XX_M3_TXEV_ACK (0x1 << 0)
+#define AM33XX_M3_TXEV_ENABLE (0x0 << 0)
+
+#define IPC_CMD_DS0 0x4
+#define IPC_CMD_STANDBY 0xc
+#define IPC_CMD_IDLE 0x10
+#define IPC_CMD_RESET 0xe
+#define DS_IPC_DEFAULT 0xffffffff
+#define M3_VERSION_UNKNOWN 0x0000ffff
+#define M3_BASELINE_VERSION 0x191
+#define M3_STATUS_RESP_MASK (0xffff << 16)
+#define M3_FW_VERSION_MASK 0xffff
+
+#define M3_STATE_UNKNOWN 0
+#define M3_STATE_RESET 1
+#define M3_STATE_INITED 2
+#define M3_STATE_MSG_FOR_LP 3
+#define M3_STATE_MSG_FOR_RESET 4
+
+struct wkup_m3_ipc {
+ struct rproc *rproc;
+
+ void __iomem *ipc_mem_base;
+ struct device *dev;
+
+ int mem_type;
+ unsigned long resume_addr;
+ int state;
+
+ struct completion sync_complete;
+ struct mbox_client mbox_client;
+ struct mbox_chan *mbox;
+};
+
+static struct wkup_m3_ipc m3_ipc_state;
+
+static void am33xx_txev_eoi(struct wkup_m3_ipc *m3_ipc)
+{
+ writel(AM33XX_M3_TXEV_ACK,
+ m3_ipc->ipc_mem_base + AM33XX_CONTROL_M3_TXEV_EOI);
+}
+
+static void am33xx_txev_enable(struct wkup_m3_ipc *m3_ipc)
+{
+ writel(AM33XX_M3_TXEV_ENABLE,
+ m3_ipc->ipc_mem_base + AM33XX_CONTROL_M3_TXEV_EOI);
+}
+
+static void wkup_m3_ctrl_ipc_write(struct wkup_m3_ipc *m3_ipc,
+ u32 val, int ipc_reg_num)
+{
+ if (WARN(ipc_reg_num < 0 || ipc_reg_num > AM33XX_CTRL_IPC_REG_COUNT,
+ "ipc register operation out of range"))
+ return;
+
+ writel(val, m3_ipc->ipc_mem_base +
+ AM33XX_CTRL_IPC_REG_OFFSET(ipc_reg_num));
+}
+
+static unsigned int wkup_m3_ctrl_ipc_read(struct wkup_m3_ipc *m3_ipc,
+ int ipc_reg_num)
+{
+ if (WARN(ipc_reg_num < 0 || ipc_reg_num > AM33XX_CTRL_IPC_REG_COUNT,
+ "ipc register operation out of range"))
+ return 0;
+
+ return readl(m3_ipc->ipc_mem_base +
+ AM33XX_CTRL_IPC_REG_OFFSET(ipc_reg_num));
+}
+
+static int wkup_m3_fw_version_read(struct wkup_m3_ipc *m3_ipc)
+{
+ int val;
+
+ val = wkup_m3_ctrl_ipc_read(m3_ipc, 2);
+
+ return val & M3_FW_VERSION_MASK;
+}
+
+static irqreturn_t wkup_m3_txev_handler(int irq, void *ipc_data)
+{
+ struct wkup_m3_ipc *m3_ipc = ipc_data;
+ struct device *dev = m3_ipc->dev;
+ int ver = 0;
+
+ am33xx_txev_eoi(m3_ipc);
+
+ switch (m3_ipc->state) {
+ case M3_STATE_RESET:
+ ver = wkup_m3_fw_version_read(m3_ipc);
+
+ if (ver == M3_VERSION_UNKNOWN ||
+ ver < M3_BASELINE_VERSION) {
+ dev_warn(dev, "CM3 Firmware Version %x not supported\n",
+ ver);
+ } else {
+ dev_info(dev, "CM3 Firmware Version = 0x%x\n", ver);
+ }
+
+ m3_ipc->state = M3_STATE_INITED;
+ complete(&m3_ipc->sync_complete);
+ break;
+ case M3_STATE_MSG_FOR_RESET:
+ m3_ipc->state = M3_STATE_INITED;
+ complete(&m3_ipc->sync_complete);
+ break;
+ case M3_STATE_MSG_FOR_LP:
+ complete(&m3_ipc->sync_complete);
+ break;
+ case M3_STATE_UNKNOWN:
+ dev_warn(dev, "Unknown CM3 State\n");
+ }
+
+ am33xx_txev_enable(m3_ipc);
+
+ return IRQ_HANDLED;
+}
+
+static int wkup_m3_ping(struct wkup_m3_ipc *m3_ipc)
+{
+ struct device *dev = m3_ipc->dev;
+ mbox_msg_t dummy_msg = 0;
+ int ret;
+
+ if (!m3_ipc->mbox) {
+ dev_err(dev,
+ "No IPC channel to communicate with wkup_m3!\n");
+ return -EIO;
+ }
+
+ /*
+ * Write a dummy message to the mailbox in order to trigger the RX
+ * interrupt to alert the M3 that data is available in the IPC
+ * registers. We must enable the IRQ here and disable it after in
+ * the RX callback to avoid multiple interrupts being received
+ * by the CM3.
+ */
+ ret = mbox_send_message(m3_ipc->mbox, &dummy_msg);
+ if (ret < 0) {
+ dev_err(dev, "%s: mbox_send_message() failed: %d\n",
+ __func__, ret);
+ return ret;
+ }
+
+ ret = wait_for_completion_timeout(&m3_ipc->sync_complete,
+ msecs_to_jiffies(500));
+ if (!ret) {
+ dev_err(dev, "MPU<->CM3 sync failure\n");
+ m3_ipc->state = M3_STATE_UNKNOWN;
+ return -EIO;
+ }
+
+ mbox_client_txdone(m3_ipc->mbox, 0);
+ return 0;
+}
+
+static int wkup_m3_ping_noirq(struct wkup_m3_ipc *m3_ipc)
+{
+ struct device *dev = m3_ipc->dev;
+ mbox_msg_t dummy_msg = 0;
+ int ret;
+
+ if (!m3_ipc->mbox) {
+ dev_err(dev,
+ "No IPC channel to communicate with wkup_m3!\n");
+ return -EIO;
+ }
+
+ ret = mbox_send_message(m3_ipc->mbox, &dummy_msg);
+ if (ret < 0) {
+ dev_err(dev, "%s: mbox_send_message() failed: %d\n",
+ __func__, ret);
+ return ret;
+ }
+
+ mbox_client_txdone(m3_ipc->mbox, 0);
+ return 0;
+}
+
+static int wkup_m3_is_available(struct wkup_m3_ipc *m3_ipc)
+{
+ return ((m3_ipc->state != M3_STATE_RESET) &&
+ (m3_ipc->state != M3_STATE_UNKNOWN));
+}
+
+/* Public functions */
+/**
+ * wkup_m3_set_mem_type - Pass wkup_m3 which type of memory is in use
+ * @mem_type: memory type value read directly from emif
+ *
+ * wkup_m3 must know what memory type is in use to properly suspend
+ * and resume.
+ */
+void wkup_m3_set_mem_type(int mem_type)
+{
+ m3_ipc_state.mem_type = mem_type;
+}
+EXPORT_SYMBOL_GPL(wkup_m3_set_mem_type);
+
+/**
+ * wkup_m3_set_resume_address - Pass wkup_m3 resume address
+ * @addr: Physical address from which resume code should execute
+ */
+void wkup_m3_set_resume_address(void *addr)
+{
+ m3_ipc_state.resume_addr = (unsigned long)addr;
+}
+EXPORT_SYMBOL_GPL(wkup_m3_set_resume_address);
+
+/**
+ * wkup_m3_request_pm_status - Retrieve wkup_m3 status code after suspend
+ *
+ * Returns code representing the status of a low power mode transition.
+ * 0 - Successful transition
+ * 1 - Failure to transition to low power state
+ */
+int wkup_m3_request_pm_status(void)
+{
+ struct wkup_m3_ipc *m3_ipc = &m3_ipc_state;
+ unsigned int i;
+ int val;
+
+ val = wkup_m3_ctrl_ipc_read(m3_ipc, 1);
+
+ i = M3_STATUS_RESP_MASK & val;
+ i >>= __ffs(M3_STATUS_RESP_MASK);
+
+ return i;
+}
+EXPORT_SYMBOL_GPL(wkup_m3_request_pm_status);
+
+/**
+ * wkup_m3_prepare_low_power - Request preparation for transition to
+ * low power state
+ * @state: A kernel suspend state to enter, either MEM or STANDBY
+ *
+ * Returns 0 if preparation was successful, otherwise returns error code
+ */
+int wkup_m3_prepare_low_power(int state)
+{
+ struct wkup_m3_ipc *m3_ipc = &m3_ipc_state;
+ struct device *dev = m3_ipc_state.dev;
+ int m3_power_state;
+ int ret = 0;
+
+ if (!wkup_m3_is_available(m3_ipc))
+ return -ENODEV;
+
+ switch (state) {
+ case WKUP_M3_DEEPSLEEP:
+ m3_power_state = IPC_CMD_DS0;
+ break;
+ case WKUP_M3_STANDBY:
+ m3_power_state = IPC_CMD_STANDBY;
+ break;
+ case WKUP_M3_IDLE:
+ m3_power_state = IPC_CMD_IDLE;
+ break;
+ default:
+ return 1;
+ }
+
+ /* Program each required IPC register then write defaults to others */
+ wkup_m3_ctrl_ipc_write(m3_ipc, m3_ipc_state.resume_addr, 0);
+ wkup_m3_ctrl_ipc_write(m3_ipc, m3_power_state, 1);
+ wkup_m3_ctrl_ipc_write(m3_ipc, m3_ipc_state.mem_type, 4);
+
+ wkup_m3_ctrl_ipc_write(m3_ipc, DS_IPC_DEFAULT, 2);
+ wkup_m3_ctrl_ipc_write(m3_ipc, DS_IPC_DEFAULT, 3);
+ wkup_m3_ctrl_ipc_write(m3_ipc, DS_IPC_DEFAULT, 5);
+ wkup_m3_ctrl_ipc_write(m3_ipc, DS_IPC_DEFAULT, 6);
+ wkup_m3_ctrl_ipc_write(m3_ipc, DS_IPC_DEFAULT, 7);
+
+ m3_ipc->state = M3_STATE_MSG_FOR_LP;
+
+ if (state == WKUP_M3_IDLE)
+ ret = wkup_m3_ping_noirq(m3_ipc);
+ else
+ ret = wkup_m3_ping(m3_ipc);
+
+ if (ret) {
+ dev_err(dev, "Unable to ping CM3\n");
+ return ret;
+ }
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(wkup_m3_prepare_low_power);
+
+/**
+ * wkup_m3_finish_low_power - Return m3 to reset state
+ *
+ * Returns 0 if reset was successful, otherwise returns error code
+ */
+int wkup_m3_finish_low_power(void)
+{
+ struct wkup_m3_ipc *m3_ipc = &m3_ipc_state;
+ struct device *dev = m3_ipc_state.dev;
+ int ret = 0;
+
+ if (!wkup_m3_is_available(m3_ipc))
+ return -ENODEV;
+
+ wkup_m3_ctrl_ipc_write(m3_ipc, IPC_CMD_RESET, 1);
+ wkup_m3_ctrl_ipc_write(m3_ipc, DS_IPC_DEFAULT, 2);
+
+ m3_ipc->state = M3_STATE_MSG_FOR_RESET;
+
+ ret = wkup_m3_ping(m3_ipc);
+ if (ret) {
+ dev_err(dev, "Unable to ping CM3\n");
+ return ret;
+ }
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(wkup_m3_finish_low_power);
+
+static void wkup_m3_rproc_boot_thread(struct rproc *rproc)
+{
+ struct device *dev = &rproc->dev;
+ int ret;
+
+ wait_for_completion(&rproc->firmware_loading_complete);
+
+ init_completion(&m3_ipc_state.sync_complete);
+
+ ret = rproc_boot(rproc);
+ if (ret)
+ dev_err(dev, "rproc_boot failed\n");
+
+ do_exit(0);
+}
+
+static int wkup_m3_ipc_probe(struct platform_device *pdev)
+{
+ struct device *dev = &pdev->dev;
+ int irq, ret;
+ phandle rproc_phandle;
+ struct rproc *m3_rproc;
+ struct resource *res;
+ struct task_struct *task;
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ m3_ipc_state.ipc_mem_base = devm_ioremap_resource(dev, res);
+ if (IS_ERR(m3_ipc_state.ipc_mem_base)) {
+ dev_err(dev, "could not ioremap ipc_mem\n");
+ ret = PTR_ERR(m3_ipc_state.ipc_mem_base);
+ return ret;
+ }
+
+ irq = platform_get_irq(pdev, 0);
+ if (!irq) {
+ dev_err(&pdev->dev, "no irq resource\n");
+ ret = -ENXIO;
+ return ret;
+ }
+
+ ret = devm_request_irq(dev, irq, wkup_m3_txev_handler,
+ 0, "wkup_m3_txev", &m3_ipc_state);
+ if (ret) {
+ dev_err(dev, "request_irq failed\n");
+ return ret;
+ }
+
+ m3_ipc_state.mbox_client.dev = dev;
+ m3_ipc_state.mbox_client.tx_done = NULL;
+ m3_ipc_state.mbox_client.rx_callback = NULL;
+ m3_ipc_state.mbox_client.tx_block = false;
+ m3_ipc_state.mbox_client.knows_txdone = false;
+
+ m3_ipc_state.mbox = mbox_request_channel(&m3_ipc_state.mbox_client, 0);
+ if (IS_ERR(m3_ipc_state.mbox)) {
+ dev_err(dev, "IPC Request for A8->M3 Channel failed! %ld\n",
+ PTR_ERR(m3_ipc_state.mbox));
+ ret = PTR_ERR(m3_ipc_state.mbox);
+ m3_ipc_state.mbox = NULL;
+ return ret;
+ }
+
+ if (of_property_read_u32(dev->of_node, "ti,rproc", &rproc_phandle)) {
+ dev_err(&pdev->dev, "could not get rproc phandle\n");
+ ret = -ENODEV;
+ goto err;
+ }
+
+ m3_rproc = rproc_get_by_phandle(rproc_phandle);
+ if (!m3_rproc) {
+ dev_err(&pdev->dev, "could not get rproc handle\n");
+ ret = -EPROBE_DEFER;
+ goto err;
+ }
+
+ m3_ipc_state.rproc = m3_rproc;
+ m3_ipc_state.dev = dev;
+ m3_ipc_state.state = M3_STATE_RESET;
+
+ /*
+ * Wait for firmware loading completion in a thread so we
+ * can boot the wkup_m3 as soon as it's ready without holding
+ * up kernel boot
+ */
+ task = kthread_run((void *)wkup_m3_rproc_boot_thread, m3_rproc,
+ "wkup_m3_rproc_loader");
+
+ if (IS_ERR(task)) {
+ dev_err(dev, "can't create rproc_boot thread\n");
+ goto err_put_rproc;
+ }
+
+ return 0;
+
+err_put_rproc:
+ rproc_put(m3_rproc);
+err:
+ mbox_free_channel(m3_ipc_state.mbox);
+ return ret;
+}
+
+static int wkup_m3_ipc_remove(struct platform_device *pdev)
+{
+ mbox_free_channel(m3_ipc_state.mbox);
+
+ rproc_shutdown(m3_ipc_state.rproc);
+ rproc_put(m3_ipc_state.rproc);
+
+ return 0;
+}
+
+static const struct of_device_id wkup_m3_ipc_of_match[] = {
+ { .compatible = "ti,am3352-wkup-m3-ipc", },
+ { .compatible = "ti,am4372-wkup-m3-ipc", },
+ {},
+};
+
+static struct platform_driver wkup_m3_ipc_driver = {
+ .probe = wkup_m3_ipc_probe,
+ .remove = wkup_m3_ipc_remove,
+ .driver = {
+ .name = "wkup_m3_ipc",
+ .of_match_table = wkup_m3_ipc_of_match,
+ },
+};
+
+module_platform_driver(wkup_m3_ipc_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("wkup m3 remote processor ipc driver");
+MODULE_AUTHOR("Dave Gerlach <[email protected]>");
diff --git a/include/linux/wkup_m3_ipc.h b/include/linux/wkup_m3_ipc.h
new file mode 100644
index 0000000..4eda969
--- /dev/null
+++ b/include/linux/wkup_m3_ipc.h
@@ -0,0 +1,30 @@
+/*
+ * TI Wakeup M3 for AMx3 SoCs Power Management Routines
+ *
+ * Copyright (C) 2015 Texas Instruments Incorporated - http://www.ti.com/
+ * Dave Gerlach <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _LINUX_WKUP_M3_IPC_H
+#define _LINUX_WKUP_M3_IPC_H
+
+#define WKUP_M3_DEEPSLEEP 1
+#define WKUP_M3_STANDBY 2
+#define WKUP_M3_IDLE 3
+
+void wkup_m3_set_mem_type(int mem_type);
+void wkup_m3_set_resume_address(void *addr);
+int wkup_m3_prepare_low_power(int state);
+int wkup_m3_finish_low_power(void);
+int wkup_m3_request_pm_status(void);
+
+#endif /* _LINUX_WKUP_M3_IPC_H */
--
2.4.5

2015-07-20 06:16:06

by Tony Lindgren

[permalink] [raw]
Subject: Re: [PATCH v2 3/3] soc: ti: Add wkup_m3_ipc driver

* Dave Gerlach <[email protected]> [150717 13:59]:
> +
> +/* Public functions */
...
> +EXPORT_SYMBOL_GPL(wkup_m3_set_mem_type);
> +EXPORT_SYMBOL_GPL(wkup_m3_set_resume_address);
> +EXPORT_SYMBOL_GPL(wkup_m3_request_pm_status);
> +EXPORT_SYMBOL_GPL(wkup_m3_prepare_low_power);
> +EXPORT_SYMBOL_GPL(wkup_m3_finish_low_power);

I think you can avoid exporting these SoC specific functions
by just exporting wkup_m3_request() and wkup_m3_free() type
functions with a data structure containing the necessary
function pointers.

Regards,

Tony

2015-08-04 20:59:54

by Dave Gerlach

[permalink] [raw]
Subject: Re: [PATCH v2 3/3] soc: ti: Add wkup_m3_ipc driver

On 07/20/2015 01:16 AM, Tony Lindgren wrote:
> * Dave Gerlach <[email protected]> [150717 13:59]:
>> +
>> +/* Public functions */
> ...
>> +EXPORT_SYMBOL_GPL(wkup_m3_set_mem_type);
>> +EXPORT_SYMBOL_GPL(wkup_m3_set_resume_address);
>> +EXPORT_SYMBOL_GPL(wkup_m3_request_pm_status);
>> +EXPORT_SYMBOL_GPL(wkup_m3_prepare_low_power);
>> +EXPORT_SYMBOL_GPL(wkup_m3_finish_low_power);
>
> I think you can avoid exporting these SoC specific functions
> by just exporting wkup_m3_request() and wkup_m3_free() type
> functions with a data structure containing the necessary
> function pointers.

Ok thanks for the comment, I can try that change out, I agree it
probably isn't necessary to export so much.

Regards,
Dave

>
> Regards,
>
> Tony
>

2015-08-05 10:28:46

by Tony Lindgren

[permalink] [raw]
Subject: Re: [PATCH v2 1/3] mailbox/omap: Add ti,mbox-send-noirq quirk to fix AM33xx CPU Idle

* Dave Gerlach <[email protected]> [150717 13:59]:
> --- a/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
> +++ b/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
> @@ -75,6 +75,14 @@ data that represent the following:
> Cell #3 (usr_id) - mailbox user id for identifying the interrupt line
> associated with generating a tx/rx fifo interrupt.
>
> +Optional Properties:
> +--------------------
> +- ti,mbox-send-noirq: Quirk flag to allow the client user of this sub-mailbox
> + to send messages without triggering a Tx ready interrupt,
> + and to control the Tx ticker. Should be used only on
> + sub-mailboxes used to communicate with WkupM3 remote
> + processor on AM33xx/AM43xx SoCs.
> +

Hmm can't you do this just by setting a flag in the device driver
based on the compatible string?

Regards,

Tony

2015-08-05 15:33:07

by Suman Anna

[permalink] [raw]
Subject: Re: [PATCH v2 1/3] mailbox/omap: Add ti,mbox-send-noirq quirk to fix AM33xx CPU Idle

Hi Tony,

On 08/05/2015 05:28 AM, Tony Lindgren wrote:
> * Dave Gerlach <[email protected]> [150717 13:59]:
>> --- a/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
>> +++ b/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
>> @@ -75,6 +75,14 @@ data that represent the following:
>> Cell #3 (usr_id) - mailbox user id for identifying the interrupt line
>> associated with generating a tx/rx fifo interrupt.
>>
>> +Optional Properties:
>> +--------------------
>> +- ti,mbox-send-noirq: Quirk flag to allow the client user of this sub-mailbox
>> + to send messages without triggering a Tx ready interrupt,
>> + and to control the Tx ticker. Should be used only on
>> + sub-mailboxes used to communicate with WkupM3 remote
>> + processor on AM33xx/AM43xx SoCs.
>> +
>
> Hmm can't you do this just by setting a flag in the device driver
> based on the compatible string?

We can't because there can be other users like PRUSS which will use a
sub-mailbox in a regular fashion. The compatible serves the IP and there
can be multiple sub-mailboxes underneath it, and this quirk is needed
only on the one that's used by the WkupM3.

Jassi,
Do you have any comments on this?

regards
Suman

2015-08-06 06:29:28

by Tony Lindgren

[permalink] [raw]
Subject: Re: [PATCH v2 1/3] mailbox/omap: Add ti,mbox-send-noirq quirk to fix AM33xx CPU Idle

* Suman Anna <[email protected]> [150805 08:35]:
> Hi Tony,
>
> On 08/05/2015 05:28 AM, Tony Lindgren wrote:
> > * Dave Gerlach <[email protected]> [150717 13:59]:
> >> --- a/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
> >> +++ b/Documentation/devicetree/bindings/mailbox/omap-mailbox.txt
> >> @@ -75,6 +75,14 @@ data that represent the following:
> >> Cell #3 (usr_id) - mailbox user id for identifying the interrupt line
> >> associated with generating a tx/rx fifo interrupt.
> >>
> >> +Optional Properties:
> >> +--------------------
> >> +- ti,mbox-send-noirq: Quirk flag to allow the client user of this sub-mailbox
> >> + to send messages without triggering a Tx ready interrupt,
> >> + and to control the Tx ticker. Should be used only on
> >> + sub-mailboxes used to communicate with WkupM3 remote
> >> + processor on AM33xx/AM43xx SoCs.
> >> +
> >
> > Hmm can't you do this just by setting a flag in the device driver
> > based on the compatible string?
>
> We can't because there can be other users like PRUSS which will use a
> sub-mailbox in a regular fashion. The compatible serves the IP and there
> can be multiple sub-mailboxes underneath it, and this quirk is needed
> only on the one that's used by the WkupM3.

OK

> Do you have any comments on this?

Up to you guys then to figure out what property is used for it then.

Regards,

Tony