Skip to contentSkip to conversation

mfd: introduce a driver for LPSS devices on SPT

PATCH Applied Andy Shevchenko 6 participants 35 messages Jul 27, 2015 – Jul 30, 2015

Subsystems (9)
lore
Share & tools
lore

Andy Shevchenko

author reviewer
#0 lore raw

v6 cover mfd: introduce a driver for LPSS devices on SPT

The new coming Intel platforms such as Skylake will contain Sunrisepoint PCH.

The driver is based on MFD framework since the main device, i.e. serial bus controller, contains register space for itself, DMA part, and an additional address space (convergence layer).

The public specification of the register map is avaiable in [1].

This is sixth generation of the patch series to bring support LPSS devices found on Intel Sunrisepoint (Intel Skylake PCH). Previous one can be found here [2].

The series has few logical parts:
- patches 1-3 prepares PM core, ACPI, and driver core (PM) to handle our case
- patches 4-6 introduce unregistering platform devices in MFD in reversed order
- patch 7 implements iDMA 64-bit driver
- patch 8 introduces an MFD driver for LPSS devices

The driver has been tested with SPI and UART on Intel Skylake PCH.

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

[1] https://download.01.org/future-platform-configuration-hub/skylake/register-definitions/332219-002.pdf
[2] http://www.spinics.net/lists/linux-acpi/msg59267.html

Changelog v6:
- rebase on top of v4.2-rc4
- address Vinod's comments regarding to idma64 driver
- fix a bug in idma64 driver found by our Q/A team (use after free)

Changelog v5:
- rebase on top of v4.2-rc1
- patch 3 is modified accordingly to new changes in drivers/base/dd.c
- correct Rafael's email in patch 3

Changelog v4:
- replace patch 3 by Rafael's suggession
- append ACKs from Rafael
- rebase on top of v4.1-rc8

Changelog v3:
- rework patch 3 to be used for all kind of devices (address Alan comment)
- iDMA 64-bit driver improvements:
- fix burst-to-register conversion
- remove excessive locking
- do not expose core part to the user (Kconfig)
- address most of Lee's comments
- append ACKs
- rebase on top of v4.1-rc6

Changelog v2:
- new DMA driver to fully support iDMA 64-bit IP
- patch 3 is added to wake up parent devices when ->probe(), ->remove(), or

    ->shutdown()

- MFD core is unregistering devices in reversed order
- address few Lee's comments on v1
- address Russel's comment, therefore use clkdev_create() helper
- intel-lpss{,-acpi,-pci} are modified regarding to above changes

Andy Shevchenko (5):
klist: implement klist_prev()
driver core: implement device_for_each_child_reverse()
mfd: make mfd_remove_devices() iterate in reverse order
dmaengine: add a driver for Intel integrated DMA 64-bit
mfd: Add support for Intel Sunrisepoint LPSS devices

Mika Westerberg (2):
PM / QoS: Make it possible to expose device latency tolerance to
userspace
ACPI / PM: Attach ACPI power domain only once

Rafael J. Wysocki (1):
Driver core: wakeup the parent device before trying probe

 drivers/acpi/device_pm.c      |   8 +
 drivers/acpi/internal.h       |   2 +
 drivers/acpi/scan.c           |  46 ++-
 drivers/base/core.c           |  43 +++
 drivers/base/dd.c             |  20 ++
 drivers/base/power/power.h    |   2 +
 drivers/base/power/qos.c      |  37 +++
 drivers/base/power/sysfs.c    |  11 +
 drivers/dma/Kconfig           |   8 +
 drivers/dma/Makefile          |   1 +
 drivers/dma/idma64.c          | 710 ++++++++++++++++++++++++++++++++++++++++++
 drivers/dma/idma64.h          | 233 ++++++++++++++
 drivers/mfd/Kconfig           |  23 ++
 drivers/mfd/Makefile          |   3 +
 drivers/mfd/intel-lpss-acpi.c |  84 +++++
 drivers/mfd/intel-lpss-pci.c  | 113 +++++++
 drivers/mfd/intel-lpss.c      | 524 +++++++++++++++++++++++++++++++
 drivers/mfd/intel-lpss.h      |  62 ++++
 drivers/mfd/mfd-core.c        |   2 +-
 include/linux/device.h        |   2 +
 include/linux/klist.h         |   1 +
 include/linux/pm_qos.h        |   5 +
 lib/klist.c                   |  41 +++
 23 files changed, 1964 insertions(+), 17 deletions(-)
 create mode 100644 drivers/dma/idma64.c
 create mode 100644 drivers/dma/idma64.h
 create mode 100644 drivers/mfd/intel-lpss-acpi.c
 create mode 100644 drivers/mfd/intel-lpss-pci.c
 create mode 100644 drivers/mfd/intel-lpss.c
 create mode 100644 drivers/mfd/intel-lpss.h
Signature · 1 line
2.4.6
author reviewer
#1 lore raw

PATCH v6 1/8 PM / QoS: Make it possible to expose device latency tolerance to userspace

From: Mika Westerberg <[email protected]>

Typically when a device is created the bus core it belongs to (for example PCI) does not know if the device supports things like latency tolerance. This is left to the driver that binds to the device in question. However, at that time the device has already been created and there is no way to set its dev->power.set_latency_tolerance anymore.

So follow what has been done for other PM QoS attributes as well and allow drivers to expose and hide latency tolerance from userspace, if the device supports it.

---
 drivers/base/power/power.h |  2 ++
 drivers/base/power/qos.c   | 37 +++++++++++++++++++++++++++++++++++++
 drivers/base/power/sysfs.c | 11 +++++++++++
 include/linux/pm_qos.h     |  5 +++++
 4 files changed, 55 insertions(+)
drivers/base/power/power.h +2 -0
index f1a5d95..998fa6b 100644
@@ -73,6 +73,8 @@ extern int pm_qos_sysfs_add_resume_latency(struct device *dev);
 extern void pm_qos_sysfs_remove_resume_latency(struct device *dev);
 extern int pm_qos_sysfs_add_flags(struct device *dev);
 extern void pm_qos_sysfs_remove_flags(struct device *dev);
+extern int pm_qos_sysfs_add_latency_tolerance(struct device *dev);
+extern void pm_qos_sysfs_remove_latency_tolerance(struct device *dev);
 
 #else /* CONFIG_PM */
 
drivers/base/power/qos.c +37 -0
index e56d538..7f3646e 100644
@@ -883,3 +883,40 @@ int dev_pm_qos_update_user_latency_tolerance(struct device *dev, s32 val)
 	mutex_unlock(&dev_pm_qos_mtx);
 	return ret;
 }
+
+/**
+ * dev_pm_qos_expose_latency_tolerance - Expose latency tolerance to userspace
+ * @dev: Device whose latency tolerance to expose
+ */
+int dev_pm_qos_expose_latency_tolerance(struct device *dev)
+{
+	int ret;
+
+	if (!dev->power.set_latency_tolerance)
+		return -EINVAL;
+
+	mutex_lock(&dev_pm_qos_sysfs_mtx);
+	ret = pm_qos_sysfs_add_latency_tolerance(dev);
+	mutex_unlock(&dev_pm_qos_sysfs_mtx);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(dev_pm_qos_expose_latency_tolerance);
+
+/**
+ * dev_pm_qos_hide_latency_tolerance - Hide latency tolerance from userspace
+ * @dev: Device whose latency tolerance to hide
+ */
+void dev_pm_qos_hide_latency_tolerance(struct device *dev)
+{
+	mutex_lock(&dev_pm_qos_sysfs_mtx);
+	pm_qos_sysfs_remove_latency_tolerance(dev);
+	mutex_unlock(&dev_pm_qos_sysfs_mtx);
+
+	/* Remove the request from user space now */
+	pm_runtime_get_sync(dev);
+	dev_pm_qos_update_user_latency_tolerance(dev,
+		PM_QOS_LATENCY_TOLERANCE_NO_CONSTRAINT);
+	pm_runtime_put(dev);
+}
+EXPORT_SYMBOL_GPL(dev_pm_qos_hide_latency_tolerance);
drivers/base/power/sysfs.c +11 -0
index d2be3f9..a7b4679 100644
@@ -738,6 +738,17 @@ void pm_qos_sysfs_remove_flags(struct device *dev)
 	sysfs_unmerge_group(&dev->kobj, &pm_qos_flags_attr_group);
 }
 
+int pm_qos_sysfs_add_latency_tolerance(struct device *dev)
+{
+	return sysfs_merge_group(&dev->kobj,
+				 &pm_qos_latency_tolerance_attr_group);
+}
+
+void pm_qos_sysfs_remove_latency_tolerance(struct device *dev)
+{
+	sysfs_unmerge_group(&dev->kobj, &pm_qos_latency_tolerance_attr_group);
+}
+
 void rpm_sysfs_remove(struct device *dev)
 {
 	sysfs_unmerge_group(&dev->kobj, &pm_runtime_attr_group);
include/linux/pm_qos.h +5 -0
index 7b3ae0c..0f65d36 100644
@@ -161,6 +161,8 @@ void dev_pm_qos_hide_flags(struct device *dev);
 int dev_pm_qos_update_flags(struct device *dev, s32 mask, bool set);
 s32 dev_pm_qos_get_user_latency_tolerance(struct device *dev);
 int dev_pm_qos_update_user_latency_tolerance(struct device *dev, s32 val);
+int dev_pm_qos_expose_latency_tolerance(struct device *dev);
+void dev_pm_qos_hide_latency_tolerance(struct device *dev);
 
 static inline s32 dev_pm_qos_requested_resume_latency(struct device *dev)
 {
@@ -229,6 +231,9 @@ static inline s32 dev_pm_qos_get_user_latency_tolerance(struct device *dev)
 			{ return PM_QOS_LATENCY_TOLERANCE_NO_CONSTRAINT; }
 static inline int dev_pm_qos_update_user_latency_tolerance(struct device *dev, s32 val)
 			{ return 0; }
+static inline int dev_pm_qos_expose_latency_tolerance(struct device *dev)
+			{ return 0; }
+static inline void dev_pm_qos_hide_latency_tolerance(struct device *dev) {}
 
 static inline s32 dev_pm_qos_requested_resume_latency(struct device *dev) { return 0; }
 static inline s32 dev_pm_qos_requested_flags(struct device *dev) { return 0; }
Signature · 1 line
2.4.6
Applied
#2 lore raw

Re: [PATCH v6 1/8] PM / QoS: Make it possible to expose device latency tolerance to userspace

Quoted 15 lines · Andy Shevchenko

From: Mika Westerberg <[email protected]>

Typically when a device is created the bus core it belongs to (for example PCI) does not know if the device supports things like latency tolerance. This is left to the driver that binds to the device in question. However, at that time the device has already been created and there is no way to set its dev->power.set_latency_tolerance anymore.

So follow what has been done for other PM QoS attributes as well and allow drivers to expose and hide latency tolerance from userspace, if the device supports it.

Quoted 6 lines
---
 drivers/base/power/power.h |  2 ++
 drivers/base/power/qos.c   | 37 +++++++++++++++++++++++++++++++++++++
 drivers/base/power/sysfs.c | 11 +++++++++++
 include/linux/pm_qos.h     |  5 +++++
 4 files changed, 55 insertions(+)

Applied, thanks. Pull request to follow.

Quoted 103 lines
drivers/base/power/power.h +2 -0
index f1a5d95..998fa6b 100644
@@ -73,6 +73,8 @@ extern int pm_qos_sysfs_add_resume_latency(struct device *dev);
 extern void pm_qos_sysfs_remove_resume_latency(struct device *dev);
 extern int pm_qos_sysfs_add_flags(struct device *dev);
 extern void pm_qos_sysfs_remove_flags(struct device *dev);
+extern int pm_qos_sysfs_add_latency_tolerance(struct device *dev);
+extern void pm_qos_sysfs_remove_latency_tolerance(struct device *dev);
 
 #else /* CONFIG_PM */
 
drivers/base/power/qos.c +37 -0
index e56d538..7f3646e 100644
@@ -883,3 +883,40 @@ int dev_pm_qos_update_user_latency_tolerance(struct device *dev, s32 val)
 	mutex_unlock(&dev_pm_qos_mtx);
 	return ret;
 }
+
+/**
+ * dev_pm_qos_expose_latency_tolerance - Expose latency tolerance to userspace
+ * @dev: Device whose latency tolerance to expose
+ */
+int dev_pm_qos_expose_latency_tolerance(struct device *dev)
+{
+	int ret;
+
+	if (!dev->power.set_latency_tolerance)
+		return -EINVAL;
+
+	mutex_lock(&dev_pm_qos_sysfs_mtx);
+	ret = pm_qos_sysfs_add_latency_tolerance(dev);
+	mutex_unlock(&dev_pm_qos_sysfs_mtx);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(dev_pm_qos_expose_latency_tolerance);
+
+/**
+ * dev_pm_qos_hide_latency_tolerance - Hide latency tolerance from userspace
+ * @dev: Device whose latency tolerance to hide
+ */
+void dev_pm_qos_hide_latency_tolerance(struct device *dev)
+{
+	mutex_lock(&dev_pm_qos_sysfs_mtx);
+	pm_qos_sysfs_remove_latency_tolerance(dev);
+	mutex_unlock(&dev_pm_qos_sysfs_mtx);
+
+	/* Remove the request from user space now */
+	pm_runtime_get_sync(dev);
+	dev_pm_qos_update_user_latency_tolerance(dev,
+		PM_QOS_LATENCY_TOLERANCE_NO_CONSTRAINT);
+	pm_runtime_put(dev);
+}
+EXPORT_SYMBOL_GPL(dev_pm_qos_hide_latency_tolerance);
drivers/base/power/sysfs.c +11 -0
index d2be3f9..a7b4679 100644
@@ -738,6 +738,17 @@ void pm_qos_sysfs_remove_flags(struct device *dev)
 	sysfs_unmerge_group(&dev->kobj, &pm_qos_flags_attr_group);
 }
 
+int pm_qos_sysfs_add_latency_tolerance(struct device *dev)
+{
+	return sysfs_merge_group(&dev->kobj,
+				 &pm_qos_latency_tolerance_attr_group);
+}
+
+void pm_qos_sysfs_remove_latency_tolerance(struct device *dev)
+{
+	sysfs_unmerge_group(&dev->kobj, &pm_qos_latency_tolerance_attr_group);
+}
+
 void rpm_sysfs_remove(struct device *dev)
 {
 	sysfs_unmerge_group(&dev->kobj, &pm_runtime_attr_group);
include/linux/pm_qos.h +5 -0
index 7b3ae0c..0f65d36 100644
@@ -161,6 +161,8 @@ void dev_pm_qos_hide_flags(struct device *dev);
 int dev_pm_qos_update_flags(struct device *dev, s32 mask, bool set);
 s32 dev_pm_qos_get_user_latency_tolerance(struct device *dev);
 int dev_pm_qos_update_user_latency_tolerance(struct device *dev, s32 val);
+int dev_pm_qos_expose_latency_tolerance(struct device *dev);
+void dev_pm_qos_hide_latency_tolerance(struct device *dev);
 
 static inline s32 dev_pm_qos_requested_resume_latency(struct device *dev)
 {
@@ -229,6 +231,9 @@ static inline s32 dev_pm_qos_get_user_latency_tolerance(struct device *dev)
 			{ return PM_QOS_LATENCY_TOLERANCE_NO_CONSTRAINT; }
 static inline int dev_pm_qos_update_user_latency_tolerance(struct device *dev, s32 val)
 			{ return 0; }
+static inline int dev_pm_qos_expose_latency_tolerance(struct device *dev)
+			{ return 0; }
+static inline void dev_pm_qos_hide_latency_tolerance(struct device *dev) {}
 
 static inline s32 dev_pm_qos_requested_resume_latency(struct device *dev) { return 0; }
 static inline s32 dev_pm_qos_requested_flags(struct device *dev) { return 0; }
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
author reviewer
#3 lore raw

PATCH v6 2/8 ACPI / PM: Attach ACPI power domain only once

In reply to Andy Shevchenko

From: Mika Westerberg <[email protected]>

Some devices, like MFD subdevices, share a single ACPI companion device so that they are able to access their resources and children. However, currently all these subdevices are attached to the ACPI power domain and this might cause that the power methods for the companion device get called more than once.

In order to solve this we attach the ACPI power domain only to the first physical device that is bound to the ACPI companion device. In case of MFD devices, this is the parent MFD device itself.

---
 drivers/acpi/device_pm.c |  8 ++++++++
 drivers/acpi/internal.h  |  2 ++
 drivers/acpi/scan.c      | 46 ++++++++++++++++++++++++++++++----------------
 3 files changed, 40 insertions(+), 16 deletions(-)
drivers/acpi/device_pm.c +8 -0
index 717afcd..08dc3ec 100644
@@ -1123,6 +1123,14 @@ int acpi_dev_pm_attach(struct device *dev, bool power_on)
 	if (dev->pm_domain)
 		return -EEXIST;
 
+	/*
+	 * Only attach the power domain to the first device if the
+	 * companion is shared by multiple. This is to prevent doing power
+	 * management twice.
+	 */
+	if (!acpi_device_is_first_physical_node(adev, dev))
+		return -EBUSY;
+
 	acpi_add_pm_notifier(adev, dev, acpi_pm_notify_work_func);
 	dev->pm_domain = &acpi_general_pm_domain;
 	if (power_on) {
drivers/acpi/internal.h +2 -0
index 4683a96..f6aefe9 100644
@@ -97,6 +97,8 @@ void acpi_device_add_finalize(struct acpi_device *device);
 void acpi_free_pnp_ids(struct acpi_device_pnp *pnp);
 bool acpi_device_is_present(struct acpi_device *adev);
 bool acpi_device_is_battery(struct acpi_device *adev);
+bool acpi_device_is_first_physical_node(struct acpi_device *adev,
+					const struct device *dev);
 
 /* --------------------------------------------------------------------------
                                   Power Resource
drivers/acpi/scan.c +30 -16

64 lines not on this page · raw message

Signature · 1 line
2.4.6
Applied
#4 lore raw

Re: [PATCH v6 2/8] ACPI / PM: Attach ACPI power domain only once

Quoted 12 lines · Andy Shevchenko

From: Mika Westerberg <[email protected]>

Some devices, like MFD subdevices, share a single ACPI companion device so that they are able to access their resources and children. However, currently all these subdevices are attached to the ACPI power domain and this might cause that the power methods for the companion device get called more than once.

In order to solve this we attach the ACPI power domain only to the first physical device that is bound to the ACPI companion device. In case of MFD devices, this is the parent MFD device itself.

Quoted 8 lines
---
 drivers/acpi/device_pm.c |  8 ++++++++
 drivers/acpi/internal.h  |  2 ++
 drivers/acpi/scan.c      | 46 ++++++++++++++++++++++++++++++----------------
 3 files changed, 40 insertions(+), 16 deletions(-)

Applied, thanks. Pull request to follow.

Quoted 103 lines
drivers/acpi/device_pm.c +8 -0
index 717afcd..08dc3ec 100644
@@ -1123,6 +1123,14 @@ int acpi_dev_pm_attach(struct device *dev, bool power_on)
 	if (dev->pm_domain)
 		return -EEXIST;
 
+	/*
+	 * Only attach the power domain to the first device if the
+	 * companion is shared by multiple. This is to prevent doing power
+	 * management twice.
+	 */
+	if (!acpi_device_is_first_physical_node(adev, dev))
+		return -EBUSY;
+
 	acpi_add_pm_notifier(adev, dev, acpi_pm_notify_work_func);
 	dev->pm_domain = &acpi_general_pm_domain;
 	if (power_on) {
drivers/acpi/internal.h +2 -0
index 4683a96..f6aefe9 100644
@@ -97,6 +97,8 @@ void acpi_device_add_finalize(struct acpi_device *device);
 void acpi_free_pnp_ids(struct acpi_device_pnp *pnp);
 bool acpi_device_is_present(struct acpi_device *adev);
 bool acpi_device_is_battery(struct acpi_device *adev);
+bool acpi_device_is_first_physical_node(struct acpi_device *adev,
+					const struct device *dev);
 
 /* --------------------------------------------------------------------------
                                   Power Resource
drivers/acpi/scan.c +30 -16
index ec25635..89ff6d2 100644
@@ -226,6 +226,35 @@ static int create_of_modalias(struct acpi_device *acpi_dev, char *modalias,
 	return len;
 }
 
+/**
+ * acpi_device_is_first_physical_node - Is given dev first physical node
+ * @adev: ACPI companion device
+ * @dev: Physical device to check
+ *
+ * Function checks if given @dev is the first physical devices attached to
+ * the ACPI companion device. This distinction is needed in some cases
+ * where the same companion device is shared between many physical devices.
+ *
+ * Note that the caller have to provide valid @adev pointer.
+ */
+bool acpi_device_is_first_physical_node(struct acpi_device *adev,
+					const struct device *dev)
+{
+	bool ret = false;
+
+	mutex_lock(&adev->physical_node_lock);
+	if (!list_empty(&adev->physical_node_list)) {
+		const struct acpi_device_physical_node *node;
+
+		node = list_first_entry(&adev->physical_node_list,
+					struct acpi_device_physical_node, node);
+		ret = node->dev == dev;
+	}
+	mutex_unlock(&adev->physical_node_lock);
+
+	return ret;
+}
+
 /*
  * acpi_companion_match() - Can we match via ACPI companion device
  * @dev: Device in question
@@ -250,7 +279,6 @@ static int create_of_modalias(struct acpi_device *acpi_dev, char *modalias,
 static struct acpi_device *acpi_companion_match(const struct device *dev)
 {
 	struct acpi_device *adev;
-	struct mutex *physical_node_lock;
 
 	adev = ACPI_COMPANION(dev);
 	if (!adev)
@@ -259,21 +287,7 @@ static struct acpi_device *acpi_companion_match(const struct device *dev)
 	if (list_empty(&adev->pnp.ids))
 		return NULL;
 
-	physical_node_lock = &adev->physical_node_lock;
-	mutex_lock(physical_node_lock);
-	if (list_empty(&adev->physical_node_list)) {
-		adev = NULL;
-	} else {
-		const struct acpi_device_physical_node *node;
-
-		node = list_first_entry(&adev->physical_node_list,
-					struct acpi_device_physical_node, node);
-		if (node->dev != dev)
-			adev = NULL;
-	}
-	mutex_unlock(physical_node_lock);
-
-	return adev;
+	return acpi_device_is_first_physical_node(adev, dev) ? adev : NULL;
 }
 
 static int __acpi_device_uevent_modalias(struct acpi_device *adev,
Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
author reviewer
#5 lore raw

PATCH v6 3/8 Driver core: wakeup the parent device before trying probe

In reply to Andy Shevchenko

From: "Rafael J. Wysocki" <[email protected]>

If the parent is still suspended when driver probe is attempted, the result may be failure.

For example, if the parent is a PCI MFD device that has been suspended when we try to probe our device, any register reads will return 0xffffffff.

To fix the problem, making sure the parent is always awake before attempting driver probe.

---
 drivers/base/dd.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)
drivers/base/dd.c +20 -0
index a638bbb..2d6df1d 100644
@@ -399,6 +399,8 @@ EXPORT_SYMBOL_GPL(wait_for_device_probe);
  *
  * This function must be called with @dev lock held.  When called for a
  * USB interface, @dev->parent lock must be held as well.
+ *
+ * If the device has a parent, runtime-resume the parent before driver probing.
  */
 int driver_probe_device(struct device_driver *drv, struct device *dev)
 {
@@ -410,10 +412,16 @@ int driver_probe_device(struct device_driver *drv, struct device *dev)
 	pr_debug("bus: '%s': %s: matched device %s with driver %s\n",
 		 drv->bus->name, __func__, dev_name(dev), drv->name);
 
+	if (dev->parent)
+		pm_runtime_get_sync(dev->parent);
+
 	pm_runtime_barrier(dev);
 	ret = really_probe(dev, drv);
 	pm_request_idle(dev);
 
+	if (dev->parent)
+		pm_runtime_put(dev->parent);
+
 	return ret;
 }
 
@@ -507,11 +515,17 @@ static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
 
 	device_lock(dev);
 
+	if (dev->parent)
+		pm_runtime_get_sync(dev->parent);
+
 	bus_for_each_drv(dev->bus, NULL, &data, __device_attach_driver);
 	dev_dbg(dev, "async probe completed\n");
 
 	pm_request_idle(dev);
 
+	if (dev->parent)
+		pm_runtime_put(dev->parent);
+
 	device_unlock(dev);
 
 	put_device(dev);
@@ -541,6 +555,9 @@ static int __device_attach(struct device *dev, bool allow_async)
 			.want_async = false,
 		};
 
+		if (dev->parent)
+			pm_runtime_get_sync(dev->parent);
+
 		ret = bus_for_each_drv(dev->bus, NULL, &data,
 					__device_attach_driver);
 		if (!ret && allow_async && data.have_async) {
@@ -557,6 +574,9 @@ static int __device_attach(struct device *dev, bool allow_async)
 		} else {
 			pm_request_idle(dev);
 		}
+
+		if (dev->parent)
+			pm_runtime_put(dev->parent);
 	}
 out_unlock:
 	device_unlock(dev);
Signature · 1 line
2.4.6
Applied
#6 lore raw

Re: [PATCH v6 3/8] Driver core: wakeup the parent device before trying probe

Quoted 12 lines · Andy Shevchenko

From: "Rafael J. Wysocki" <[email protected]>

If the parent is still suspended when driver probe is attempted, the result may be failure.

For example, if the parent is a PCI MFD device that has been suspended when we try to probe our device, any register reads will return 0xffffffff.

To fix the problem, making sure the parent is always awake before attempting driver probe.

Quoted 6 lines
---
 drivers/base/dd.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

Applied, thanks. Pull request to follow.

Quoted 68 lines
drivers/base/dd.c +20 -0
index a638bbb..2d6df1d 100644
@@ -399,6 +399,8 @@ EXPORT_SYMBOL_GPL(wait_for_device_probe);
  *
  * This function must be called with @dev lock held.  When called for a
  * USB interface, @dev->parent lock must be held as well.
+ *
+ * If the device has a parent, runtime-resume the parent before driver probing.
  */
 int driver_probe_device(struct device_driver *drv, struct device *dev)
 {
@@ -410,10 +412,16 @@ int driver_probe_device(struct device_driver *drv, struct device *dev)
 	pr_debug("bus: '%s': %s: matched device %s with driver %s\n",
 		 drv->bus->name, __func__, dev_name(dev), drv->name);
 
+	if (dev->parent)
+		pm_runtime_get_sync(dev->parent);
+
 	pm_runtime_barrier(dev);
 	ret = really_probe(dev, drv);
 	pm_request_idle(dev);
 
+	if (dev->parent)
+		pm_runtime_put(dev->parent);
+
 	return ret;
 }
 
@@ -507,11 +515,17 @@ static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
 
 	device_lock(dev);
 
+	if (dev->parent)
+		pm_runtime_get_sync(dev->parent);
+
 	bus_for_each_drv(dev->bus, NULL, &data, __device_attach_driver);
 	dev_dbg(dev, "async probe completed\n");
 
 	pm_request_idle(dev);
 
+	if (dev->parent)
+		pm_runtime_put(dev->parent);
+
 	device_unlock(dev);
 
 	put_device(dev);
@@ -541,6 +555,9 @@ static int __device_attach(struct device *dev, bool allow_async)
 			.want_async = false,
 		};
 
+		if (dev->parent)
+			pm_runtime_get_sync(dev->parent);
+
 		ret = bus_for_each_drv(dev->bus, NULL, &data,
 					__device_attach_driver);
 		if (!ret && allow_async && data.have_async) {
@@ -557,6 +574,9 @@ static int __device_attach(struct device *dev, bool allow_async)
 		} else {
 			pm_request_idle(dev);
 		}
+
+		if (dev->parent)
+			pm_runtime_put(dev->parent);
 	}
 out_unlock:
 	device_unlock(dev);
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
author reviewer
#7 lore raw

PATCH v6 4/8 klist: implement klist_prev()

In reply to Andy Shevchenko

klist_prev() gets the previous element in the list. It is useful to traverse through the list in reverse order, for example, to provide LIFO (last in first out) variant of access.

---
 include/linux/klist.h |  1 +
 lib/klist.c           | 41 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 42 insertions(+)
include/linux/klist.h +1 -0
index 61e5b72..953f283 100644
@@ -63,6 +63,7 @@ extern void klist_iter_init(struct klist *k, struct klist_iter *i);
 extern void klist_iter_init_node(struct klist *k, struct klist_iter *i,
 				 struct klist_node *n);
 extern void klist_iter_exit(struct klist_iter *i);
+extern struct klist_node *klist_prev(struct klist_iter *i);
 extern struct klist_node *klist_next(struct klist_iter *i);
 
 #endif
lib/klist.c +41 -0

47 lines not on this page · raw message

Signature · 1 line
2.4.6
Applied
#8 lore raw

Re: [PATCH v6 4/8] klist: implement klist_prev()

Quoted 10 lines · Andy Shevchenko

klist_prev() gets the previous element in the list. It is useful to traverse through the list in reverse order, for example, to provide LIFO (last in first out) variant of access.

---
 include/linux/klist.h |  1 +
 lib/klist.c           | 41 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 42 insertions(+)

Applied, thanks. Pull request to follow.

Quoted 64 lines
include/linux/klist.h +1 -0
index 61e5b72..953f283 100644
@@ -63,6 +63,7 @@ extern void klist_iter_init(struct klist *k, struct klist_iter *i);
 extern void klist_iter_init_node(struct klist *k, struct klist_iter *i,
 				 struct klist_node *n);
 extern void klist_iter_exit(struct klist_iter *i);
+extern struct klist_node *klist_prev(struct klist_iter *i);
 extern struct klist_node *klist_next(struct klist_iter *i);
 
 #endif
lib/klist.c +41 -0
index 89b485a..d74cf7a 100644
@@ -324,6 +324,47 @@ static struct klist_node *to_klist_node(struct list_head *n)
 }
 
 /**
+ * klist_prev - Ante up prev node in list.
+ * @i: Iterator structure.
+ *
+ * First grab list lock. Decrement the reference count of the previous
+ * node, if there was one. Grab the prev node, increment its reference
+ * count, drop the lock, and return that prev node.
+ */
+struct klist_node *klist_prev(struct klist_iter *i)
+{
+	void (*put)(struct klist_node *) = i->i_klist->put;
+	struct klist_node *last = i->i_cur;
+	struct klist_node *prev;
+
+	spin_lock(&i->i_klist->k_lock);
+
+	if (last) {
+		prev = to_klist_node(last->n_node.prev);
+		if (!klist_dec_and_del(last))
+			put = NULL;
+	} else
+		prev = to_klist_node(i->i_klist->k_list.prev);
+
+	i->i_cur = NULL;
+	while (prev != to_klist_node(&i->i_klist->k_list)) {
+		if (likely(!knode_dead(prev))) {
+			kref_get(&prev->n_ref);
+			i->i_cur = prev;
+			break;
+		}
+		prev = to_klist_node(prev->n_node.prev);
+	}
+
+	spin_unlock(&i->i_klist->k_lock);
+
+	if (put && last)
+		put(last);
+	return i->i_cur;
+}
+EXPORT_SYMBOL_GPL(klist_prev);
+
+/**
  * klist_next - Ante up next node in list.
  * @i: Iterator structure.
  *
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
author reviewer
#9 lore raw

PATCH v6 5/8 driver core: implement device_for_each_child_reverse()

In reply to Andy Shevchenko

The new function device_for_each_child_reverse() is helpful to traverse the registered devices in a reversed order, e.g. in the case when an operation on each device should be done first on the last added device, then on one before last and so on.

---
 drivers/base/core.c    | 43 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/device.h |  2 ++
 2 files changed, 45 insertions(+)
drivers/base/core.c +43 -0
index dafae6d..7d62795 100644
@@ -1252,6 +1252,19 @@ void device_unregister(struct device *dev)
 }
 EXPORT_SYMBOL_GPL(device_unregister);
 
+static struct device *prev_device(struct klist_iter *i)
+{
+	struct klist_node *n = klist_prev(i);
+	struct device *dev = NULL;
+	struct device_private *p;
+
+	if (n) {
+		p = to_device_private_parent(n);
+		dev = p->device;
+	}
+	return dev;
+}
+
 static struct device *next_device(struct klist_iter *i)
 {
 	struct klist_node *n = klist_next(i);
@@ -1341,6 +1354,36 @@ int device_for_each_child(struct device *parent, void *data,
 EXPORT_SYMBOL_GPL(device_for_each_child);
 
 /**
+ * device_for_each_child_reverse - device child iterator in reversed order.
+ * @parent: parent struct device.
+ * @fn: function to be called for each device.
+ * @data: data for the callback.
+ *
+ * Iterate over @parent's child devices, and call @fn for each,
+ * passing it @data.
+ *
+ * We check the return of @fn each time. If it returns anything
+ * other than 0, we break out and return that value.
+ */
+int device_for_each_child_reverse(struct device *parent, void *data,
+				  int (*fn)(struct device *dev, void *data))
+{
+	struct klist_iter i;
+	struct device *child;
+	int error = 0;
+
+	if (!parent->p)
+		return 0;
+
+	klist_iter_init(&parent->p->klist_children, &i);
+	while ((child = prev_device(&i)) && !error)
+		error = fn(child, data);
+	klist_iter_exit(&i);
+	return error;
+}
+EXPORT_SYMBOL_GPL(device_for_each_child_reverse);
+
+/**
  * device_find_child - device iterator for locating a particular device.
  * @parent: parent struct device
  * @match: Callback function to check device
include/linux/device.h +2 -0
index a2b4ea7..9d212fe 100644
@@ -959,6 +959,8 @@ extern int __must_check device_add(struct device *dev);
 extern void device_del(struct device *dev);
 extern int device_for_each_child(struct device *dev, void *data,
 		     int (*fn)(struct device *dev, void *data));
+extern int device_for_each_child_reverse(struct device *dev, void *data,
+		     int (*fn)(struct device *dev, void *data));
 extern struct device *device_find_child(struct device *dev, void *data,
 				int (*match)(struct device *dev, void *data));
 extern int device_rename(struct device *dev, const char *new_name);
Signature · 1 line
2.4.6
Applied
#10 lore raw

Re: [PATCH v6 5/8] driver core: implement device_for_each_child_reverse()

Quoted 11 lines · Andy Shevchenko

The new function device_for_each_child_reverse() is helpful to traverse the registered devices in a reversed order, e.g. in the case when an operation on each device should be done first on the last added device, then on one before last and so on.

---
 drivers/base/core.c    | 43 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/device.h |  2 ++
 2 files changed, 45 insertions(+)

Applied, thanks. Pull request to follow.

Quoted 74 lines
drivers/base/core.c +43 -0
index dafae6d..7d62795 100644
@@ -1252,6 +1252,19 @@ void device_unregister(struct device *dev)
 }
 EXPORT_SYMBOL_GPL(device_unregister);
 
+static struct device *prev_device(struct klist_iter *i)
+{
+	struct klist_node *n = klist_prev(i);
+	struct device *dev = NULL;
+	struct device_private *p;
+
+	if (n) {
+		p = to_device_private_parent(n);
+		dev = p->device;
+	}
+	return dev;
+}
+
 static struct device *next_device(struct klist_iter *i)
 {
 	struct klist_node *n = klist_next(i);
@@ -1341,6 +1354,36 @@ int device_for_each_child(struct device *parent, void *data,
 EXPORT_SYMBOL_GPL(device_for_each_child);
 
 /**
+ * device_for_each_child_reverse - device child iterator in reversed order.
+ * @parent: parent struct device.
+ * @fn: function to be called for each device.
+ * @data: data for the callback.
+ *
+ * Iterate over @parent's child devices, and call @fn for each,
+ * passing it @data.
+ *
+ * We check the return of @fn each time. If it returns anything
+ * other than 0, we break out and return that value.
+ */
+int device_for_each_child_reverse(struct device *parent, void *data,
+				  int (*fn)(struct device *dev, void *data))
+{
+	struct klist_iter i;
+	struct device *child;
+	int error = 0;
+
+	if (!parent->p)
+		return 0;
+
+	klist_iter_init(&parent->p->klist_children, &i);
+	while ((child = prev_device(&i)) && !error)
+		error = fn(child, data);
+	klist_iter_exit(&i);
+	return error;
+}
+EXPORT_SYMBOL_GPL(device_for_each_child_reverse);
+
+/**
  * device_find_child - device iterator for locating a particular device.
  * @parent: parent struct device
  * @match: Callback function to check device
include/linux/device.h +2 -0
index a2b4ea7..9d212fe 100644
@@ -959,6 +959,8 @@ extern int __must_check device_add(struct device *dev);
 extern void device_del(struct device *dev);
 extern int device_for_each_child(struct device *dev, void *data,
 		     int (*fn)(struct device *dev, void *data));
+extern int device_for_each_child_reverse(struct device *dev, void *data,
+		     int (*fn)(struct device *dev, void *data));
 extern struct device *device_find_child(struct device *dev, void *data,
 				int (*match)(struct device *dev, void *data));
 extern int device_rename(struct device *dev, const char *new_name);
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
author reviewer
#11 lore raw

PATCH v6 6/8 mfd: make mfd_remove_devices() iterate in reverse order

In reply to Andy Shevchenko

The newly introduced device_for_each_child_reverse() would be used when MFD core removes the device.

After this patch applied the devices will be removed in a reversed order. This behaviour is useful when devices have implicit dependency on order, i.e. consider MFD device with serial bus controller, such as SPI, and DMA IP that is attached to serial bus controller: before remove the DMA driver we have to be ensured that no DMA transfers is ongoing and the requested channel are unused.

---
 drivers/mfd/mfd-core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mfd/mfd-core.c +1 -1
index 14fd5cb..c17635d 100644
@@ -302,7 +302,7 @@ void mfd_remove_devices(struct device *parent)
 {
 	atomic_t *cnts = NULL;
 
-	device_for_each_child(parent, &cnts, mfd_remove_devices_fn);
+	device_for_each_child_reverse(parent, &cnts, mfd_remove_devices_fn);
 	kfree(cnts);
 }
 EXPORT_SYMBOL(mfd_remove_devices);
Signature · 1 line
2.4.6
Applied
#12 lore raw

Re: [PATCH v6 6/8] mfd: make mfd_remove_devices() iterate in reverse order

Quoted 9 lines · Andy Shevchenko

The newly introduced device_for_each_child_reverse() would be used when MFD core removes the device.

After this patch applied the devices will be removed in a reversed order. This behaviour is useful when devices have implicit dependency on order, i.e. consider MFD device with serial bus controller, such as SPI, and DMA IP that is attached to serial bus controller: before remove the DMA driver we have to be ensured that no DMA transfers is ongoing and the requested channel are unused.

Quoted 4 lines
---
 drivers/mfd/mfd-core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Applied, thanks. Pull request to follow.

Quoted 13 lines
drivers/mfd/mfd-core.c +1 -1
index 14fd5cb..c17635d 100644
@@ -302,7 +302,7 @@ void mfd_remove_devices(struct device *parent)
 {
 	atomic_t *cnts = NULL;
 
-	device_for_each_child(parent, &cnts, mfd_remove_devices_fn);
+	device_for_each_child_reverse(parent, &cnts, mfd_remove_devices_fn);
 	kfree(cnts);
 }
 EXPORT_SYMBOL(mfd_remove_devices);
Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
author reviewer
#13 lore raw

PATCH v6 7/8 dmaengine: add a driver for Intel integrated DMA 64-bit

In reply to Andy Shevchenko

Intel integrated DMA (iDMA) 64-bit is a specific IP that is used as a part of LPSS devices such as HSUART or SPI. The iDMA IP is attached for private usage on each host controller independently.

While it has similarities with Synopsys DesignWare DMA, the following distinctions doesn't allow to use the existing driver:
- 64-bit mode with corresponding changes in Hardware Linked List data structure
- many slight differences in the channel registers

Moreover this driver is based on the DMA virtual channels framework that helps to make the driver cleaner and easy to understand.

---
 drivers/dma/Kconfig  |   8 +
 drivers/dma/Makefile |   1 +
 drivers/dma/idma64.c | 710 +++++++++++++++++++++++++++++++++++++++++++++++++++
 drivers/dma/idma64.h | 233 +++++++++++++++++
 4 files changed, 952 insertions(+)
 create mode 100644 drivers/dma/idma64.c
 create mode 100644 drivers/dma/idma64.h
drivers/dma/Kconfig +8 -0
index 88d474b..bdbbe5b 100644
@@ -85,6 +85,14 @@ config INTEL_IOP_ADMA
 	help
 	  Enable support for the Intel(R) IOP Series RAID engines.
 
+config IDMA64
+	tristate "Intel integrated DMA 64-bit support"
+	select DMA_ENGINE
+	select DMA_VIRTUAL_CHANNELS
+	help
+	  Enable DMA support for Intel Low Power Subsystem such as found on
+	  Intel Skylake PCH.
+
 source "drivers/dma/dw/Kconfig"
 
 config AT_HDMAC
drivers/dma/Makefile +1 -0
index 6a4d6f2..56ff8c7 100644
@@ -14,6 +14,7 @@ obj-$(CONFIG_HSU_DMA) += hsu/
 obj-$(CONFIG_MPC512X_DMA) += mpc512x_dma.o
 obj-$(CONFIG_PPC_BESTCOMM) += bestcomm/
 obj-$(CONFIG_MV_XOR) += mv_xor.o
+obj-$(CONFIG_IDMA64) += idma64.o
 obj-$(CONFIG_DW_DMAC_CORE) += dw/
 obj-$(CONFIG_AT_HDMAC) += at_hdmac.o
 obj-$(CONFIG_AT_XDMAC) += at_xdmac.o
drivers/dma/idma64.c new file +710 -0

710 lines not on this page · raw message

drivers/dma/idma64.h new file +233 -0

233 lines not on this page · raw message

Signature · 1 line
2.4.6
Applied
#14 lore raw

Re: [PATCH v6 7/8] dmaengine: add a driver for Intel integrated DMA 64-bit

Quoted 13 lines · Andy Shevchenko

Intel integrated DMA (iDMA) 64-bit is a specific IP that is used as a part of LPSS devices such as HSUART or SPI. The iDMA IP is attached for private usage on each host controller independently.

While it has similarities with Synopsys DesignWare DMA, the following distinctions doesn't allow to use the existing driver:
- 64-bit mode with corresponding changes in Hardware Linked List data structure
- many slight differences in the channel registers

Moreover this driver is based on the DMA virtual channels framework that helps to make the driver cleaner and easy to understand.

Quoted 8 lines
---
 drivers/dma/Kconfig  |   8 +
 drivers/dma/Makefile |   1 +
 drivers/dma/idma64.c | 710 +++++++++++++++++++++++++++++++++++++++++++++++++++
 drivers/dma/idma64.h | 233 +++++++++++++++++
 4 files changed, 952 insertions(+)
 create mode 100644 drivers/dma/idma64.c
 create mode 100644 drivers/dma/idma64.h

Applied, thanks. Pull request to follow.

Quoted 986 lines
drivers/dma/Kconfig +8 -0
index 88d474b..bdbbe5b 100644
@@ -85,6 +85,14 @@ config INTEL_IOP_ADMA
 	help
 	  Enable support for the Intel(R) IOP Series RAID engines.
 
+config IDMA64
+	tristate "Intel integrated DMA 64-bit support"
+	select DMA_ENGINE
+	select DMA_VIRTUAL_CHANNELS
+	help
+	  Enable DMA support for Intel Low Power Subsystem such as found on
+	  Intel Skylake PCH.
+
 source "drivers/dma/dw/Kconfig"
 
 config AT_HDMAC
drivers/dma/Makefile +1 -0
index 6a4d6f2..56ff8c7 100644
@@ -14,6 +14,7 @@ obj-$(CONFIG_HSU_DMA) += hsu/
 obj-$(CONFIG_MPC512X_DMA) += mpc512x_dma.o
 obj-$(CONFIG_PPC_BESTCOMM) += bestcomm/
 obj-$(CONFIG_MV_XOR) += mv_xor.o
+obj-$(CONFIG_IDMA64) += idma64.o
 obj-$(CONFIG_DW_DMAC_CORE) += dw/
 obj-$(CONFIG_AT_HDMAC) += at_hdmac.o
 obj-$(CONFIG_AT_XDMAC) += at_xdmac.o
drivers/dma/idma64.c new file +710 -0
new file mode 100644
index 0000000..18c14e1
@@ -0,0 +1,710 @@
+/*
+ * Core driver for the Intel integrated DMA 64-bit
+ *
+ * Copyright (C) 2015 Intel Corporation
+ * Author: Andy Shevchenko <[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.
+ */
+
+#include <linux/bitops.h>
+#include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmapool.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include "idma64.h"
+
+/* Platform driver name */
+#define DRV_NAME		"idma64"
+
+/* For now we support only two channels */
+#define IDMA64_NR_CHAN		2
+
+/* ---------------------------------------------------------------------- */
+
+static struct device *chan2dev(struct dma_chan *chan)
+{
+	return &chan->dev->device;
+}
+
+/* ---------------------------------------------------------------------- */
+
+static void idma64_off(struct idma64 *idma64)
+{
+	unsigned short count = 100;
+
+	dma_writel(idma64, CFG, 0);
+
+	channel_clear_bit(idma64, MASK(XFER), idma64->all_chan_mask);
+	channel_clear_bit(idma64, MASK(BLOCK), idma64->all_chan_mask);
+	channel_clear_bit(idma64, MASK(SRC_TRAN), idma64->all_chan_mask);
+	channel_clear_bit(idma64, MASK(DST_TRAN), idma64->all_chan_mask);
+	channel_clear_bit(idma64, MASK(ERROR), idma64->all_chan_mask);
+
+	do {
+		cpu_relax();
+	} while (dma_readl(idma64, CFG) & IDMA64_CFG_DMA_EN && --count);
+}
+
+static void idma64_on(struct idma64 *idma64)
+{
+	dma_writel(idma64, CFG, IDMA64_CFG_DMA_EN);
+}
+
+/* ---------------------------------------------------------------------- */
+
+static void idma64_chan_init(struct idma64 *idma64, struct idma64_chan *idma64c)
+{
+	u32 cfghi = IDMA64C_CFGH_SRC_PER(1) | IDMA64C_CFGH_DST_PER(0);
+	u32 cfglo = 0;
+
+	/* Enforce FIFO drain when channel is suspended */
+	cfglo |= IDMA64C_CFGL_CH_DRAIN;
+
+	/* Set default burst alignment */
+	cfglo |= IDMA64C_CFGL_DST_BURST_ALIGN | IDMA64C_CFGL_SRC_BURST_ALIGN;
+
+	channel_writel(idma64c, CFG_LO, cfglo);
+	channel_writel(idma64c, CFG_HI, cfghi);
+
+	/* Enable interrupts */
+	channel_set_bit(idma64, MASK(XFER), idma64c->mask);
+	channel_set_bit(idma64, MASK(ERROR), idma64c->mask);
+
+	/*
+	 * Enforce the controller to be turned on.
+	 *
+	 * The iDMA is turned off in ->probe() and looses context during system
+	 * suspend / resume cycle. That's why we have to enable it each time we
+	 * use it.
+	 */
+	idma64_on(idma64);
+}
+
+static void idma64_chan_stop(struct idma64 *idma64, struct idma64_chan *idma64c)
+{
+	channel_clear_bit(idma64, CH_EN, idma64c->mask);
+}
+
+static void idma64_chan_start(struct idma64 *idma64, struct idma64_chan *idma64c)
+{
+	struct idma64_desc *desc = idma64c->desc;
+	struct idma64_hw_desc *hw = &desc->hw[0];
+
+	channel_writeq(idma64c, SAR, 0);
+	channel_writeq(idma64c, DAR, 0);
+
+	channel_writel(idma64c, CTL_HI, IDMA64C_CTLH_BLOCK_TS(~0UL));
+	channel_writel(idma64c, CTL_LO, IDMA64C_CTLL_LLP_S_EN | IDMA64C_CTLL_LLP_D_EN);
+
+	channel_writeq(idma64c, LLP, hw->llp);
+
+	channel_set_bit(idma64, CH_EN, idma64c->mask);
+}
+
+static void idma64_stop_transfer(struct idma64_chan *idma64c)
+{
+	struct idma64 *idma64 = to_idma64(idma64c->vchan.chan.device);
+
+	idma64_chan_stop(idma64, idma64c);
+}
+
+static void idma64_start_transfer(struct idma64_chan *idma64c)
+{
+	struct idma64 *idma64 = to_idma64(idma64c->vchan.chan.device);
+	struct virt_dma_desc *vdesc;
+
+	/* Get the next descriptor */
+	vdesc = vchan_next_desc(&idma64c->vchan);
+	if (!vdesc) {
+		idma64c->desc = NULL;
+		return;
+	}
+
+	list_del(&vdesc->node);
+	idma64c->desc = to_idma64_desc(vdesc);
+
+	/* Configure the channel */
+	idma64_chan_init(idma64, idma64c);
+
+	/* Start the channel with a new descriptor */
+	idma64_chan_start(idma64, idma64c);
+}
+
+/* ---------------------------------------------------------------------- */
+
+static void idma64_chan_irq(struct idma64 *idma64, unsigned short c,
+		u32 status_err, u32 status_xfer)
+{
+	struct idma64_chan *idma64c = &idma64->chan[c];
+	struct idma64_desc *desc;
+	unsigned long flags;
+
+	spin_lock_irqsave(&idma64c->vchan.lock, flags);
+	desc = idma64c->desc;
+	if (desc) {
+		if (status_err & (1 << c)) {
+			dma_writel(idma64, CLEAR(ERROR), idma64c->mask);
+			desc->status = DMA_ERROR;
+		} else if (status_xfer & (1 << c)) {
+			dma_writel(idma64, CLEAR(XFER), idma64c->mask);
+			desc->status = DMA_COMPLETE;
+			vchan_cookie_complete(&desc->vdesc);
+			idma64_start_transfer(idma64c);
+		}
+
+		/* idma64_start_transfer() updates idma64c->desc */
+		if (idma64c->desc == NULL || desc->status == DMA_ERROR)
+			idma64_stop_transfer(idma64c);
+	}
+	spin_unlock_irqrestore(&idma64c->vchan.lock, flags);
+}
+
+static irqreturn_t idma64_irq(int irq, void *dev)
+{
+	struct idma64 *idma64 = dev;
+	u32 status = dma_readl(idma64, STATUS_INT);
+	u32 status_xfer;
+	u32 status_err;
+	unsigned short i;
+
+	dev_vdbg(idma64->dma.dev, "%s: status=%#x\n", __func__, status);
+
+	/* Check if we have any interrupt from the DMA controller */
+	if (!status)
+		return IRQ_NONE;
+
+	/* Disable interrupts */
+	channel_clear_bit(idma64, MASK(XFER), idma64->all_chan_mask);
+	channel_clear_bit(idma64, MASK(ERROR), idma64->all_chan_mask);
+
+	status_xfer = dma_readl(idma64, RAW(XFER));
+	status_err = dma_readl(idma64, RAW(ERROR));
+
+	for (i = 0; i < idma64->dma.chancnt; i++)
+		idma64_chan_irq(idma64, i, status_err, status_xfer);
+
+	/* Re-enable interrupts */
+	channel_set_bit(idma64, MASK(XFER), idma64->all_chan_mask);
+	channel_set_bit(idma64, MASK(ERROR), idma64->all_chan_mask);
+
+	return IRQ_HANDLED;
+}
+
+/* ---------------------------------------------------------------------- */
+
+static struct idma64_desc *idma64_alloc_desc(unsigned int ndesc)
+{
+	struct idma64_desc *desc;
+
+	desc = kzalloc(sizeof(*desc), GFP_NOWAIT);
+	if (!desc)
+		return NULL;
+
+	desc->hw = kcalloc(ndesc, sizeof(*desc->hw), GFP_NOWAIT);
+	if (!desc->hw) {
+		kfree(desc);
+		return NULL;
+	}
+
+	return desc;
+}
+
+static void idma64_desc_free(struct idma64_chan *idma64c,
+		struct idma64_desc *desc)
+{
+	struct idma64_hw_desc *hw;
+
+	if (desc->ndesc) {
+		unsigned int i = desc->ndesc;
+
+		do {
+			hw = &desc->hw[--i];
+			dma_pool_free(idma64c->pool, hw->lli, hw->llp);
+		} while (i);
+	}
+
+	kfree(desc->hw);
+	kfree(desc);
+}
+
+static void idma64_vdesc_free(struct virt_dma_desc *vdesc)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(vdesc->tx.chan);
+
+	idma64_desc_free(idma64c, to_idma64_desc(vdesc));
+}
+
+static u64 idma64_hw_desc_fill(struct idma64_hw_desc *hw,
+		struct dma_slave_config *config,
+		enum dma_transfer_direction direction, u64 llp)
+{
+	struct idma64_lli *lli = hw->lli;
+	u64 sar, dar;
+	u32 ctlhi = IDMA64C_CTLH_BLOCK_TS(hw->len);
+	u32 ctllo = IDMA64C_CTLL_LLP_S_EN | IDMA64C_CTLL_LLP_D_EN;
+	u32 src_width, dst_width;
+
+	if (direction == DMA_MEM_TO_DEV) {
+		sar = hw->phys;
+		dar = config->dst_addr;
+		ctllo |= IDMA64C_CTLL_DST_FIX | IDMA64C_CTLL_SRC_INC |
+			 IDMA64C_CTLL_FC_M2P;
+		src_width = min_t(u32, 2, __fls(sar | hw->len));
+		dst_width = __fls(config->dst_addr_width);
+	} else {	/* DMA_DEV_TO_MEM */
+		sar = config->src_addr;
+		dar = hw->phys;
+		ctllo |= IDMA64C_CTLL_DST_INC | IDMA64C_CTLL_SRC_FIX |
+			 IDMA64C_CTLL_FC_P2M;
+		src_width = __fls(config->src_addr_width);
+		dst_width = min_t(u32, 2, __fls(dar | hw->len));
+	}
+
+	lli->sar = sar;
+	lli->dar = dar;
+
+	lli->ctlhi = ctlhi;
+	lli->ctllo = ctllo |
+		     IDMA64C_CTLL_SRC_MSIZE(config->src_maxburst) |
+		     IDMA64C_CTLL_DST_MSIZE(config->dst_maxburst) |
+		     IDMA64C_CTLL_DST_WIDTH(dst_width) |
+		     IDMA64C_CTLL_SRC_WIDTH(src_width);
+
+	lli->llp = llp;
+	return hw->llp;
+}
+
+static void idma64_desc_fill(struct idma64_chan *idma64c,
+		struct idma64_desc *desc)
+{
+	struct dma_slave_config *config = &idma64c->config;
+	struct idma64_hw_desc *hw = &desc->hw[desc->ndesc - 1];
+	struct idma64_lli *lli = hw->lli;
+	u64 llp = 0;
+	unsigned int i = desc->ndesc;
+
+	/* Fill the hardware descriptors and link them to a list */
+	do {
+		hw = &desc->hw[--i];
+		llp = idma64_hw_desc_fill(hw, config, desc->direction, llp);
+		desc->length += hw->len;
+	} while (i);
+
+	/* Trigger interrupt after last block */
+	lli->ctllo |= IDMA64C_CTLL_INT_EN;
+}
+
+static struct dma_async_tx_descriptor *idma64_prep_slave_sg(
+		struct dma_chan *chan, struct scatterlist *sgl,
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+	struct idma64_desc *desc;
+	struct scatterlist *sg;
+	unsigned int i;
+
+	desc = idma64_alloc_desc(sg_len);
+	if (!desc)
+		return NULL;
+
+	for_each_sg(sgl, sg, sg_len, i) {
+		struct idma64_hw_desc *hw = &desc->hw[i];
+
+		/* Allocate DMA capable memory for hardware descriptor */
+		hw->lli = dma_pool_alloc(idma64c->pool, GFP_NOWAIT, &hw->llp);
+		if (!hw->lli) {
+			desc->ndesc = i;
+			idma64_desc_free(idma64c, desc);
+			return NULL;
+		}
+
+		hw->phys = sg_dma_address(sg);
+		hw->len = sg_dma_len(sg);
+	}
+
+	desc->ndesc = sg_len;
+	desc->direction = direction;
+	desc->status = DMA_IN_PROGRESS;
+
+	idma64_desc_fill(idma64c, desc);
+	return vchan_tx_prep(&idma64c->vchan, &desc->vdesc, flags);
+}
+
+static void idma64_issue_pending(struct dma_chan *chan)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&idma64c->vchan.lock, flags);
+	if (vchan_issue_pending(&idma64c->vchan) && !idma64c->desc)
+		idma64_start_transfer(idma64c);
+	spin_unlock_irqrestore(&idma64c->vchan.lock, flags);
+}
+
+static size_t idma64_active_desc_size(struct idma64_chan *idma64c)
+{
+	struct idma64_desc *desc = idma64c->desc;
+	struct idma64_hw_desc *hw;
+	size_t bytes = desc->length;
+	u64 llp;
+	u32 ctlhi;
+	unsigned int i = 0;
+
+	llp = channel_readq(idma64c, LLP);
+	do {
+		hw = &desc->hw[i];
+	} while ((hw->llp != llp) && (++i < desc->ndesc));
+
+	if (!i)
+		return bytes;
+
+	do {
+		bytes -= desc->hw[--i].len;
+	} while (i);
+
+	ctlhi = channel_readl(idma64c, CTL_HI);
+	return bytes - IDMA64C_CTLH_BLOCK_TS(ctlhi);
+}
+
+static enum dma_status idma64_tx_status(struct dma_chan *chan,
+		dma_cookie_t cookie, struct dma_tx_state *state)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+	struct virt_dma_desc *vdesc;
+	enum dma_status status;
+	size_t bytes;
+	unsigned long flags;
+
+	status = dma_cookie_status(chan, cookie, state);
+	if (status == DMA_COMPLETE)
+		return status;
+
+	spin_lock_irqsave(&idma64c->vchan.lock, flags);
+	vdesc = vchan_find_desc(&idma64c->vchan, cookie);
+	if (idma64c->desc && cookie == idma64c->desc->vdesc.tx.cookie) {
+		bytes = idma64_active_desc_size(idma64c);
+		dma_set_residue(state, bytes);
+		status = idma64c->desc->status;
+	} else if (vdesc) {
+		bytes = to_idma64_desc(vdesc)->length;
+		dma_set_residue(state, bytes);
+	}
+	spin_unlock_irqrestore(&idma64c->vchan.lock, flags);
+
+	return status;
+}
+
+static void convert_burst(u32 *maxburst)
+{
+	if (*maxburst)
+		*maxburst = __fls(*maxburst);
+	else
+		*maxburst = 0;
+}
+
+static int idma64_slave_config(struct dma_chan *chan,
+		struct dma_slave_config *config)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+
+	/* Check if chan will be configured for slave transfers */
+	if (!is_slave_direction(config->direction))
+		return -EINVAL;
+
+	memcpy(&idma64c->config, config, sizeof(idma64c->config));
+
+	convert_burst(&idma64c->config.src_maxburst);
+	convert_burst(&idma64c->config.dst_maxburst);
+
+	return 0;
+}
+
+static void idma64_chan_deactivate(struct idma64_chan *idma64c)
+{
+	unsigned short count = 100;
+	u32 cfglo;
+
+	cfglo = channel_readl(idma64c, CFG_LO);
+	channel_writel(idma64c, CFG_LO, cfglo | IDMA64C_CFGL_CH_SUSP);
+	do {
+		udelay(1);
+		cfglo = channel_readl(idma64c, CFG_LO);
+	} while (!(cfglo & IDMA64C_CFGL_FIFO_EMPTY) && --count);
+}
+
+static void idma64_chan_activate(struct idma64_chan *idma64c)
+{
+	u32 cfglo;
+
+	cfglo = channel_readl(idma64c, CFG_LO);
+	channel_writel(idma64c, CFG_LO, cfglo & ~IDMA64C_CFGL_CH_SUSP);
+}
+
+static int idma64_pause(struct dma_chan *chan)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&idma64c->vchan.lock, flags);
+	if (idma64c->desc && idma64c->desc->status == DMA_IN_PROGRESS) {
+		idma64_chan_deactivate(idma64c);
+		idma64c->desc->status = DMA_PAUSED;
+	}
+	spin_unlock_irqrestore(&idma64c->vchan.lock, flags);
+
+	return 0;
+}
+
+static int idma64_resume(struct dma_chan *chan)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&idma64c->vchan.lock, flags);
+	if (idma64c->desc && idma64c->desc->status == DMA_PAUSED) {
+		idma64c->desc->status = DMA_IN_PROGRESS;
+		idma64_chan_activate(idma64c);
+	}
+	spin_unlock_irqrestore(&idma64c->vchan.lock, flags);
+
+	return 0;
+}
+
+static int idma64_terminate_all(struct dma_chan *chan)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+	unsigned long flags;
+	LIST_HEAD(head);
+
+	spin_lock_irqsave(&idma64c->vchan.lock, flags);
+	idma64_chan_deactivate(idma64c);
+	idma64_stop_transfer(idma64c);
+	if (idma64c->desc) {
+		idma64_vdesc_free(&idma64c->desc->vdesc);
+		idma64c->desc = NULL;
+	}
+	vchan_get_all_descriptors(&idma64c->vchan, &head);
+	spin_unlock_irqrestore(&idma64c->vchan.lock, flags);
+
+	vchan_dma_desc_free_list(&idma64c->vchan, &head);
+	return 0;
+}
+
+static int idma64_alloc_chan_resources(struct dma_chan *chan)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+
+	/* Create a pool of consistent memory blocks for hardware descriptors */
+	idma64c->pool = dma_pool_create(dev_name(chan2dev(chan)),
+					chan->device->dev,
+					sizeof(struct idma64_lli), 8, 0);
+	if (!idma64c->pool) {
+		dev_err(chan2dev(chan), "No memory for descriptors\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void idma64_free_chan_resources(struct dma_chan *chan)
+{
+	struct idma64_chan *idma64c = to_idma64_chan(chan);
+
+	vchan_free_chan_resources(to_virt_chan(chan));
+	dma_pool_destroy(idma64c->pool);
+	idma64c->pool = NULL;
+}
+
+/* ---------------------------------------------------------------------- */
+
+#define IDMA64_BUSWIDTHS				\
+	BIT(DMA_SLAVE_BUSWIDTH_1_BYTE)		|	\
+	BIT(DMA_SLAVE_BUSWIDTH_2_BYTES)		|	\
+	BIT(DMA_SLAVE_BUSWIDTH_4_BYTES)
+
+static int idma64_probe(struct idma64_chip *chip)
+{
+	struct idma64 *idma64;
+	unsigned short nr_chan = IDMA64_NR_CHAN;
+	unsigned short i;
+	int ret;
+
+	idma64 = devm_kzalloc(chip->dev, sizeof(*idma64), GFP_KERNEL);
+	if (!idma64)
+		return -ENOMEM;
+
+	idma64->regs = chip->regs;
+	chip->idma64 = idma64;
+
+	idma64->chan = devm_kcalloc(chip->dev, nr_chan, sizeof(*idma64->chan),
+				    GFP_KERNEL);
+	if (!idma64->chan)
+		return -ENOMEM;
+
+	idma64->all_chan_mask = (1 << nr_chan) - 1;
+
+	/* Turn off iDMA controller */
+	idma64_off(idma64);
+
+	ret = devm_request_irq(chip->dev, chip->irq, idma64_irq, IRQF_SHARED,
+			       dev_name(chip->dev), idma64);
+	if (ret)
+		return ret;
+
+	INIT_LIST_HEAD(&idma64->dma.channels);
+	for (i = 0; i < nr_chan; i++) {
+		struct idma64_chan *idma64c = &idma64->chan[i];
+
+		idma64c->vchan.desc_free = idma64_vdesc_free;
+		vchan_init(&idma64c->vchan, &idma64->dma);
+
+		idma64c->regs = idma64->regs + i * IDMA64_CH_LENGTH;
+		idma64c->mask = BIT(i);
+	}
+
+	dma_cap_set(DMA_SLAVE, idma64->dma.cap_mask);
+	dma_cap_set(DMA_PRIVATE, idma64->dma.cap_mask);
+
+	idma64->dma.device_alloc_chan_resources = idma64_alloc_chan_resources;
+	idma64->dma.device_free_chan_resources = idma64_free_chan_resources;
+
+	idma64->dma.device_prep_slave_sg = idma64_prep_slave_sg;
+
+	idma64->dma.device_issue_pending = idma64_issue_pending;
+	idma64->dma.device_tx_status = idma64_tx_status;
+
+	idma64->dma.device_config = idma64_slave_config;
+	idma64->dma.device_pause = idma64_pause;
+	idma64->dma.device_resume = idma64_resume;
+	idma64->dma.device_terminate_all = idma64_terminate_all;
+
+	idma64->dma.src_addr_widths = IDMA64_BUSWIDTHS;
+	idma64->dma.dst_addr_widths = IDMA64_BUSWIDTHS;
+	idma64->dma.directions = BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV);
+	idma64->dma.residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
+
+	idma64->dma.dev = chip->dev;
+
+	ret = dma_async_device_register(&idma64->dma);
+	if (ret)
+		return ret;
+
+	dev_info(chip->dev, "Found Intel integrated DMA 64-bit\n");
+	return 0;
+}
+
+static int idma64_remove(struct idma64_chip *chip)
+{
+	struct idma64 *idma64 = chip->idma64;
+	unsigned short i;
+
+	dma_async_device_unregister(&idma64->dma);
+
+	/*
+	 * Explicitly call devm_request_irq() to avoid the side effects with
+	 * the scheduled tasklets.
+	 */
+	devm_free_irq(chip->dev, chip->irq, idma64);
+
+	for (i = 0; i < idma64->dma.chancnt; i++) {
+		struct idma64_chan *idma64c = &idma64->chan[i];
+
+		tasklet_kill(&idma64c->vchan.task);
+	}
+
+	return 0;
+}
+
+/* ---------------------------------------------------------------------- */
+
+static int idma64_platform_probe(struct platform_device *pdev)
+{
+	struct idma64_chip *chip;
+	struct device *dev = &pdev->dev;
+	struct resource *mem;
+	int ret;
+
+	chip = devm_kzalloc(dev, sizeof(*chip), GFP_KERNEL);
+	if (!chip)
+		return -ENOMEM;
+
+	chip->irq = platform_get_irq(pdev, 0);
+	if (chip->irq < 0)
+		return chip->irq;
+
+	mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	chip->regs = devm_ioremap_resource(dev, mem);
+	if (IS_ERR(chip->regs))
+		return PTR_ERR(chip->regs);
+
+	ret = dma_coerce_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
+	if (ret)
+		return ret;
+
+	chip->dev = dev;
+
+	ret = idma64_probe(chip);
+	if (ret)
+		return ret;
+
+	platform_set_drvdata(pdev, chip);
+	return 0;
+}
+
+static int idma64_platform_remove(struct platform_device *pdev)
+{
+	struct idma64_chip *chip = platform_get_drvdata(pdev);
+
+	return idma64_remove(chip);
+}
+
+#ifdef CONFIG_PM_SLEEP
+
+static int idma64_pm_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct idma64_chip *chip = platform_get_drvdata(pdev);
+
+	idma64_off(chip->idma64);
+	return 0;
+}
+
+static int idma64_pm_resume(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct idma64_chip *chip = platform_get_drvdata(pdev);
+
+	idma64_on(chip->idma64);
+	return 0;
+}
+
+#endif /* CONFIG_PM_SLEEP */
+
+static const struct dev_pm_ops idma64_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(idma64_pm_suspend, idma64_pm_resume)
+};
+
+static struct platform_driver idma64_platform_driver = {
+	.probe		= idma64_platform_probe,
+	.remove		= idma64_platform_remove,
+	.driver = {
+		.name	= DRV_NAME,
+		.pm	= &idma64_dev_pm_ops,
+	},
+};
+
+module_platform_driver(idma64_platform_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("iDMA64 core driver");
+MODULE_AUTHOR("Andy Shevchenko <[email protected]>");
+MODULE_ALIAS("platform:" DRV_NAME);
drivers/dma/idma64.h new file +233 -0
new file mode 100644
index 0000000..a4d9968
@@ -0,0 +1,233 @@
+/*
+ * Driver for the Intel integrated DMA 64-bit
+ *
+ * Copyright (C) 2015 Intel Corporation
+ *
+ * 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.
+ */
+
+#ifndef __DMA_IDMA64_H__
+#define __DMA_IDMA64_H__
+
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/spinlock.h>
+#include <linux/types.h>
+
+#include "virt-dma.h"
+
+/* Channel registers */
+
+#define IDMA64_CH_SAR		0x00	/* Source Address Register */
+#define IDMA64_CH_DAR		0x08	/* Destination Address Register */
+#define IDMA64_CH_LLP		0x10	/* Linked List Pointer */
+#define IDMA64_CH_CTL_LO	0x18	/* Control Register Low */
+#define IDMA64_CH_CTL_HI	0x1c	/* Control Register High */
+#define IDMA64_CH_SSTAT		0x20
+#define IDMA64_CH_DSTAT		0x28
+#define IDMA64_CH_SSTATAR	0x30
+#define IDMA64_CH_DSTATAR	0x38
+#define IDMA64_CH_CFG_LO	0x40	/* Configuration Register Low */
+#define IDMA64_CH_CFG_HI	0x44	/* Configuration Register High */
+#define IDMA64_CH_SGR		0x48
+#define IDMA64_CH_DSR		0x50
+
+#define IDMA64_CH_LENGTH	0x58
+
+/* Bitfields in CTL_LO */
+#define IDMA64C_CTLL_INT_EN		(1 << 0)	/* irqs enabled? */
+#define IDMA64C_CTLL_DST_WIDTH(x)	((x) << 1)	/* bytes per element */
+#define IDMA64C_CTLL_SRC_WIDTH(x)	((x) << 4)
+#define IDMA64C_CTLL_DST_INC		(0 << 8)	/* DAR update/not */
+#define IDMA64C_CTLL_DST_FIX		(1 << 8)
+#define IDMA64C_CTLL_SRC_INC		(0 << 10)	/* SAR update/not */
+#define IDMA64C_CTLL_SRC_FIX		(1 << 10)
+#define IDMA64C_CTLL_DST_MSIZE(x)	((x) << 11)	/* burst, #elements */
+#define IDMA64C_CTLL_SRC_MSIZE(x)	((x) << 14)
+#define IDMA64C_CTLL_FC_M2P		(1 << 20)	/* mem-to-periph */
+#define IDMA64C_CTLL_FC_P2M		(2 << 20)	/* periph-to-mem */
+#define IDMA64C_CTLL_LLP_D_EN		(1 << 27)	/* dest block chain */
+#define IDMA64C_CTLL_LLP_S_EN		(1 << 28)	/* src block chain */
+
+/* Bitfields in CTL_HI */
+#define IDMA64C_CTLH_BLOCK_TS(x)	((x) & ((1 << 17) - 1))
+#define IDMA64C_CTLH_DONE		(1 << 17)
+
+/* Bitfields in CFG_LO */
+#define IDMA64C_CFGL_DST_BURST_ALIGN	(1 << 0)	/* dst burst align */
+#define IDMA64C_CFGL_SRC_BURST_ALIGN	(1 << 1)	/* src burst align */
+#define IDMA64C_CFGL_CH_SUSP		(1 << 8)
+#define IDMA64C_CFGL_FIFO_EMPTY		(1 << 9)
+#define IDMA64C_CFGL_CH_DRAIN		(1 << 10)	/* drain FIFO */
+#define IDMA64C_CFGL_DST_OPT_BL		(1 << 20)	/* optimize dst burst length */
+#define IDMA64C_CFGL_SRC_OPT_BL		(1 << 21)	/* optimize src burst length */
+
+/* Bitfields in CFG_HI */
+#define IDMA64C_CFGH_SRC_PER(x)		((x) << 0)	/* src peripheral */
+#define IDMA64C_CFGH_DST_PER(x)		((x) << 4)	/* dst peripheral */
+#define IDMA64C_CFGH_RD_ISSUE_THD(x)	((x) << 8)
+#define IDMA64C_CFGH_RW_ISSUE_THD(x)	((x) << 18)
+
+/* Interrupt registers */
+
+#define IDMA64_INT_XFER		0x00
+#define IDMA64_INT_BLOCK	0x08
+#define IDMA64_INT_SRC_TRAN	0x10
+#define IDMA64_INT_DST_TRAN	0x18
+#define IDMA64_INT_ERROR	0x20
+
+#define IDMA64_RAW(x)		(0x2c0 + IDMA64_INT_##x)	/* r */
+#define IDMA64_STATUS(x)	(0x2e8 + IDMA64_INT_##x)	/* r (raw & mask) */
+#define IDMA64_MASK(x)		(0x310 + IDMA64_INT_##x)	/* rw (set = irq enabled) */
+#define IDMA64_CLEAR(x)		(0x338 + IDMA64_INT_##x)	/* w (ack, affects "raw") */
+
+/* Common registers */
+
+#define IDMA64_STATUS_INT	0x360	/* r */
+#define IDMA64_CFG		0x398
+#define IDMA64_CH_EN		0x3a0
+
+/* Bitfields in CFG */
+#define IDMA64_CFG_DMA_EN		(1 << 0)
+
+/* Hardware descriptor for Linked LIst transfers */
+struct idma64_lli {
+	u64		sar;
+	u64		dar;
+	u64		llp;
+	u32		ctllo;
+	u32		ctlhi;
+	u32		sstat;
+	u32		dstat;
+};
+
+struct idma64_hw_desc {
+	struct idma64_lli *lli;
+	dma_addr_t llp;
+	dma_addr_t phys;
+	unsigned int len;
+};
+
+struct idma64_desc {
+	struct virt_dma_desc vdesc;
+	enum dma_transfer_direction direction;
+	struct idma64_hw_desc *hw;
+	unsigned int ndesc;
+	size_t length;
+	enum dma_status status;
+};
+
+static inline struct idma64_desc *to_idma64_desc(struct virt_dma_desc *vdesc)
+{
+	return container_of(vdesc, struct idma64_desc, vdesc);
+}
+
+struct idma64_chan {
+	struct virt_dma_chan vchan;
+
+	void __iomem *regs;
+
+	/* hardware configuration */
+	enum dma_transfer_direction direction;
+	unsigned int mask;
+	struct dma_slave_config config;
+
+	void *pool;
+	struct idma64_desc *desc;
+};
+
+static inline struct idma64_chan *to_idma64_chan(struct dma_chan *chan)
+{
+	return container_of(chan, struct idma64_chan, vchan.chan);
+}
+
+#define channel_set_bit(idma64, reg, mask)	\
+	dma_writel(idma64, reg, ((mask) << 8) | (mask))
+#define channel_clear_bit(idma64, reg, mask)	\
+	dma_writel(idma64, reg, ((mask) << 8) | 0)
+
+static inline u32 idma64c_readl(struct idma64_chan *idma64c, int offset)
+{
+	return readl(idma64c->regs + offset);
+}
+
+static inline void idma64c_writel(struct idma64_chan *idma64c, int offset,
+				  u32 value)
+{
+	writel(value, idma64c->regs + offset);
+}
+
+#define channel_readl(idma64c, reg)		\
+	idma64c_readl(idma64c, IDMA64_CH_##reg)
+#define channel_writel(idma64c, reg, value)	\
+	idma64c_writel(idma64c, IDMA64_CH_##reg, (value))
+
+static inline u64 idma64c_readq(struct idma64_chan *idma64c, int offset)
+{
+	u64 l, h;
+
+	l = idma64c_readl(idma64c, offset);
+	h = idma64c_readl(idma64c, offset + 4);
+
+	return l | (h << 32);
+}
+
+static inline void idma64c_writeq(struct idma64_chan *idma64c, int offset,
+				  u64 value)
+{
+	idma64c_writel(idma64c, offset, value);
+	idma64c_writel(idma64c, offset + 4, value >> 32);
+}
+
+#define channel_readq(idma64c, reg)		\
+	idma64c_readq(idma64c, IDMA64_CH_##reg)
+#define channel_writeq(idma64c, reg, value)	\
+	idma64c_writeq(idma64c, IDMA64_CH_##reg, (value))
+
+struct idma64 {
+	struct dma_device dma;
+
+	void __iomem *regs;
+
+	/* channels */
+	unsigned short all_chan_mask;
+	struct idma64_chan *chan;
+};
+
+static inline struct idma64 *to_idma64(struct dma_device *ddev)
+{
+	return container_of(ddev, struct idma64, dma);
+}
+
+static inline u32 idma64_readl(struct idma64 *idma64, int offset)
+{
+	return readl(idma64->regs + offset);
+}
+
+static inline void idma64_writel(struct idma64 *idma64, int offset, u32 value)
+{
+	writel(value, idma64->regs + offset);
+}
+
+#define dma_readl(idma64, reg)			\
+	idma64_readl(idma64, IDMA64_##reg)
+#define dma_writel(idma64, reg, value)		\
+	idma64_writel(idma64, IDMA64_##reg, (value))
+
+/**
+ * struct idma64_chip - representation of DesignWare DMA controller hardware
+ * @dev:		struct device of the DMA controller
+ * @irq:		irq line
+ * @regs:		memory mapped I/O space
+ * @idma64:		struct idma64 that is filed by idma64_probe()
+ */
+struct idma64_chip {
+	struct device	*dev;
+	int		irq;
+	void __iomem	*regs;
+	struct idma64	*idma64;
+};
+
+#endif /* __DMA_IDMA64_H__ */
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#15 lore raw

Re: [PATCH v6 7/8] dmaengine: add a driver for Intel integrated DMA 64-bit

FAO Vinod,

Quoted 24 lines · Lee Jones
Quoted 13 lines · Andy Shevchenko

Intel integrated DMA (iDMA) 64-bit is a specific IP that is used as a part of LPSS devices such as HSUART or SPI. The iDMA IP is attached for private usage on each host controller independently.

While it has similarities with Synopsys DesignWare DMA, the following distinctions doesn't allow to use the existing driver:
- 64-bit mode with corresponding changes in Hardware Linked List data structure
- many slight differences in the channel registers

Moreover this driver is based on the DMA virtual channels framework that helps to make the driver cleaner and easy to understand.

Quoted 8 lines
---
 drivers/dma/Kconfig  |   8 +
 drivers/dma/Makefile |   1 +
 drivers/dma/idma64.c | 710 +++++++++++++++++++++++++++++++++++++++++++++++++++
 drivers/dma/idma64.h | 233 +++++++++++++++++
 4 files changed, 952 insertions(+)
 create mode 100644 drivers/dma/idma64.c
 create mode 100644 drivers/dma/idma64.h
Quoted 1 line

Applied, thanks. Pull request to follow.

Will send out a pull request when I have Vinod's Ack.

[...]

Patches will still be applied to -next for soak testing though.

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
author reviewer
#16 lore raw

Re: [PATCH v6 7/8] dmaengine: add a driver for Intel integrated DMA 64-bit

Quoted 34 lines · Lee Jones

FAO Vinod,

Quoted 31 lines · Lee Jones
Quoted 28 lines · Andy Shevchenko

Intel integrated DMA (iDMA) 64-bit is a specific IP that is used as a part of
LPSS devices such as HSUART or SPI. The iDMA IP is attached for private
usage on each host controller independently.

While it has similarities with Synopsys DesignWare DMA, the following
distinctions doesn't allow to use the existing driver:
- 64-bit mode with corresponding changes in Hardware Linked List data structure
- many slight differences in the channel registers

Moreover this driver is based on the DMA virtual channels framework that helps
to make the driver cleaner and easy to understand.

Quoted 1 line
---
 drivers/dma/Kconfig  |   8 +
 drivers/dma/Makefile |   1 +
 drivers/dma/idma64.c | 710 
+++++++++++++++++++++++++++++++++++++++++++++++++++
 drivers/dma/idma64.h | 233 +++++++++++++++++
 4 files changed, 952 insertions(+)
 create mode 100644 drivers/dma/idma64.c
 create mode 100644 drivers/dma/idma64.h
Quoted 3 lines
Quoted 1 line

Applied, thanks. Pull request to follow.

Will send out a pull request when I have Vinod's Ack.

Vinod, could you have a look at this once again?

Quoted 4 lines

[...]

Patches will still be applied to -next for soak testing though.

Thanks!

Signature · 2 lines
Andy Shevchenko <[email protected]>
Intel Finland Oy
Acked-by on v6 7/8
#17 lore raw

Re: [PATCH v6 7/8] dmaengine: add a driver for Intel integrated DMA 64-bit

In reply to Andy Shevchenko
Quoted 11 lines · Andy Shevchenko

Intel integrated DMA (iDMA) 64-bit is a specific IP that is used as a part of LPSS devices such as HSUART or SPI. The iDMA IP is attached for private usage on each host controller independently.

While it has similarities with Synopsys DesignWare DMA, the following distinctions doesn't allow to use the existing driver:
- 64-bit mode with corresponding changes in Hardware Linked List data structure
- many slight differences in the channel registers

Moreover this driver is based on the DMA virtual channels framework that helps to make the driver cleaner and easy to understand.

Signature · 1 line
~Vinod
author reviewer
#18 lore raw

PATCH v6 8/8 mfd: Add support for Intel Sunrisepoint LPSS devices

In reply to Andy Shevchenko

The new coming Intel platforms such as Skylake will contain Sunrisepoint PCH. The main difference to the previous platforms is that the LPSS devices are compound devices where usually main (SPI, HSUART, or I2C) and DMA IPs are present.

This patch brings the driver for such devices found on Sunrisepoint PCH.

---
 drivers/mfd/Kconfig           |  23 ++
 drivers/mfd/Makefile          |   3 +
 drivers/mfd/intel-lpss-acpi.c |  84 +++++++
 drivers/mfd/intel-lpss-pci.c  | 113 +++++++++
 drivers/mfd/intel-lpss.c      | 524 ++++++++++++++++++++++++++++++++++++++++++
 drivers/mfd/intel-lpss.h      |  62 +++++
 6 files changed, 809 insertions(+)
 create mode 100644 drivers/mfd/intel-lpss-acpi.c
 create mode 100644 drivers/mfd/intel-lpss-pci.c
 create mode 100644 drivers/mfd/intel-lpss.c
 create mode 100644 drivers/mfd/intel-lpss.h
drivers/mfd/Kconfig +23 -0
index 6538159..268b6dd 100644
@@ -328,6 +328,29 @@ config INTEL_SOC_PMIC
 	  thermal, charger and related power management functions
 	  on these systems.
 
+config MFD_INTEL_LPSS
+	tristate
+	select COMMON_CLK
+	select MFD_CORE
+
+config MFD_INTEL_LPSS_ACPI
+	tristate "Intel Low Power Subsystem support in ACPI mode"
+	select MFD_INTEL_LPSS
+	depends on X86 && ACPI
+	help
+	  This driver supports Intel Low Power Subsystem (LPSS) devices such as
+	  I2C, SPI and HS-UART starting from Intel Sunrisepoint (Intel Skylake
+	  PCH) in ACPI mode.
+
+config MFD_INTEL_LPSS_PCI
+	tristate "Intel Low Power Subsystem support in PCI mode"
+	select MFD_INTEL_LPSS
+	depends on X86 && PCI
+	help
+	  This driver supports Intel Low Power Subsystem (LPSS) devices such as
+	  I2C, SPI and HS-UART starting from Intel Sunrisepoint (Intel Skylake
+	  PCH) in PCI mode.
+
 config MFD_INTEL_MSIC
 	bool "Intel MSIC"
 	depends on INTEL_SCU_IPC
drivers/mfd/Makefile +3 -0
index ea40e07..9d730a2 100644
@@ -161,6 +161,9 @@ obj-$(CONFIG_TPS65911_COMPARATOR)	+= tps65911-comparator.o
 obj-$(CONFIG_MFD_TPS65090)	+= tps65090.o
 obj-$(CONFIG_MFD_AAT2870_CORE)	+= aat2870-core.o
 obj-$(CONFIG_MFD_ATMEL_HLCDC)	+= atmel-hlcdc.o
+obj-$(CONFIG_MFD_INTEL_LPSS)	+= intel-lpss.o
+obj-$(CONFIG_MFD_INTEL_LPSS_PCI)	+= intel-lpss-pci.o
+obj-$(CONFIG_MFD_INTEL_LPSS_ACPI)	+= intel-lpss-acpi.o
 obj-$(CONFIG_MFD_INTEL_MSIC)	+= intel_msic.o
 obj-$(CONFIG_MFD_PALMAS)	+= palmas.o
 obj-$(CONFIG_MFD_VIPERBOARD)    += viperboard.o
drivers/mfd/intel-lpss-acpi.c new file +84 -0

84 lines not on this page · raw message

drivers/mfd/intel-lpss-pci.c new file +113 -0

113 lines not on this page · raw message

drivers/mfd/intel-lpss.c new file +524 -0

524 lines not on this page · raw message

drivers/mfd/intel-lpss.h new file +62 -0

62 lines not on this page · raw message

Signature · 1 line
2.4.6
Applied
#19 lore raw

Re: [PATCH v6 8/8] mfd: Add support for Intel Sunrisepoint LPSS devices

Quoted 9 lines · Andy Shevchenko

The new coming Intel platforms such as Skylake will contain Sunrisepoint PCH. The main difference to the previous platforms is that the LPSS devices are compound devices where usually main (SPI, HSUART, or I2C) and DMA IPs are present.

This patch brings the driver for such devices found on Sunrisepoint PCH.

Quoted 12 lines
---
 drivers/mfd/Kconfig           |  23 ++
 drivers/mfd/Makefile          |   3 +
 drivers/mfd/intel-lpss-acpi.c |  84 +++++++
 drivers/mfd/intel-lpss-pci.c  | 113 +++++++++
 drivers/mfd/intel-lpss.c      | 524 ++++++++++++++++++++++++++++++++++++++++++
 drivers/mfd/intel-lpss.h      |  62 +++++
 6 files changed, 809 insertions(+)
 create mode 100644 drivers/mfd/intel-lpss-acpi.c
 create mode 100644 drivers/mfd/intel-lpss-pci.c
 create mode 100644 drivers/mfd/intel-lpss.c
 create mode 100644 drivers/mfd/intel-lpss.h

Applied, thanks. Pull request to follow.

Quoted 855 lines
drivers/mfd/Kconfig +23 -0
index 6538159..268b6dd 100644
@@ -328,6 +328,29 @@ config INTEL_SOC_PMIC
 	  thermal, charger and related power management functions
 	  on these systems.
 
+config MFD_INTEL_LPSS
+	tristate
+	select COMMON_CLK
+	select MFD_CORE
+
+config MFD_INTEL_LPSS_ACPI
+	tristate "Intel Low Power Subsystem support in ACPI mode"
+	select MFD_INTEL_LPSS
+	depends on X86 && ACPI
+	help
+	  This driver supports Intel Low Power Subsystem (LPSS) devices such as
+	  I2C, SPI and HS-UART starting from Intel Sunrisepoint (Intel Skylake
+	  PCH) in ACPI mode.
+
+config MFD_INTEL_LPSS_PCI
+	tristate "Intel Low Power Subsystem support in PCI mode"
+	select MFD_INTEL_LPSS
+	depends on X86 && PCI
+	help
+	  This driver supports Intel Low Power Subsystem (LPSS) devices such as
+	  I2C, SPI and HS-UART starting from Intel Sunrisepoint (Intel Skylake
+	  PCH) in PCI mode.
+
 config MFD_INTEL_MSIC
 	bool "Intel MSIC"
 	depends on INTEL_SCU_IPC
drivers/mfd/Makefile +3 -0
index ea40e07..9d730a2 100644
@@ -161,6 +161,9 @@ obj-$(CONFIG_TPS65911_COMPARATOR)	+= tps65911-comparator.o
 obj-$(CONFIG_MFD_TPS65090)	+= tps65090.o
 obj-$(CONFIG_MFD_AAT2870_CORE)	+= aat2870-core.o
 obj-$(CONFIG_MFD_ATMEL_HLCDC)	+= atmel-hlcdc.o
+obj-$(CONFIG_MFD_INTEL_LPSS)	+= intel-lpss.o
+obj-$(CONFIG_MFD_INTEL_LPSS_PCI)	+= intel-lpss-pci.o
+obj-$(CONFIG_MFD_INTEL_LPSS_ACPI)	+= intel-lpss-acpi.o
 obj-$(CONFIG_MFD_INTEL_MSIC)	+= intel_msic.o
 obj-$(CONFIG_MFD_PALMAS)	+= palmas.o
 obj-$(CONFIG_MFD_VIPERBOARD)    += viperboard.o
drivers/mfd/intel-lpss-acpi.c new file +84 -0
new file mode 100644
index 0000000..0d92d73
@@ -0,0 +1,84 @@
+/*
+ * Intel LPSS ACPI support.
+ *
+ * Copyright (C) 2015, Intel Corporation
+ *
+ * Authors: Andy Shevchenko <[email protected]>
+ *          Mika Westerberg <[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.
+ */
+
+#include <linux/acpi.h>
+#include <linux/ioport.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/platform_device.h>
+
+#include "intel-lpss.h"
+
+static const struct intel_lpss_platform_info spt_info = {
+	.clk_rate = 120000000,
+};
+
+static const struct acpi_device_id intel_lpss_acpi_ids[] = {
+	/* SPT */
+	{ "INT3446", (kernel_ulong_t)&spt_info },
+	{ "INT3447", (kernel_ulong_t)&spt_info },
+	{ }
+};
+MODULE_DEVICE_TABLE(acpi, intel_lpss_acpi_ids);
+
+static int intel_lpss_acpi_probe(struct platform_device *pdev)
+{
+	struct intel_lpss_platform_info *info;
+	const struct acpi_device_id *id;
+
+	id = acpi_match_device(intel_lpss_acpi_ids, &pdev->dev);
+	if (!id)
+		return -ENODEV;
+
+	info = devm_kmemdup(&pdev->dev, (void *)id->driver_data, sizeof(*info),
+			    GFP_KERNEL);
+	if (!info)
+		return -ENOMEM;
+
+	info->mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	info->irq = platform_get_irq(pdev, 0);
+
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+
+	return intel_lpss_probe(&pdev->dev, info);
+}
+
+static int intel_lpss_acpi_remove(struct platform_device *pdev)
+{
+	intel_lpss_remove(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
+	return 0;
+}
+
+static INTEL_LPSS_PM_OPS(intel_lpss_acpi_pm_ops);
+
+static struct platform_driver intel_lpss_acpi_driver = {
+	.probe = intel_lpss_acpi_probe,
+	.remove = intel_lpss_acpi_remove,
+	.driver = {
+		.name = "intel-lpss",
+		.acpi_match_table = intel_lpss_acpi_ids,
+		.pm = &intel_lpss_acpi_pm_ops,
+	},
+};
+
+module_platform_driver(intel_lpss_acpi_driver);
+
+MODULE_AUTHOR("Andy Shevchenko <[email protected]>");
+MODULE_AUTHOR("Mika Westerberg <[email protected]>");
+MODULE_DESCRIPTION("Intel LPSS ACPI driver");
+MODULE_LICENSE("GPL v2");
drivers/mfd/intel-lpss-pci.c new file +113 -0
new file mode 100644
index 0000000..9236dff
@@ -0,0 +1,113 @@
+/*
+ * Intel LPSS PCI support.
+ *
+ * Copyright (C) 2015, Intel Corporation
+ *
+ * Authors: Andy Shevchenko <[email protected]>
+ *          Mika Westerberg <[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.
+ */
+
+#include <linux/ioport.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+
+#include "intel-lpss.h"
+
+static int intel_lpss_pci_probe(struct pci_dev *pdev,
+				const struct pci_device_id *id)
+{
+	struct intel_lpss_platform_info *info;
+	int ret;
+
+	ret = pcim_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	info = devm_kmemdup(&pdev->dev, (void *)id->driver_data, sizeof(*info),
+			    GFP_KERNEL);
+	if (!info)
+		return -ENOMEM;
+
+	info->mem = &pdev->resource[0];
+	info->irq = pdev->irq;
+
+	/* Probably it is enough to set this for iDMA capable devices only */
+	pci_set_master(pdev);
+
+	ret = intel_lpss_probe(&pdev->dev, info);
+	if (ret)
+		return ret;
+
+	pm_runtime_put(&pdev->dev);
+	pm_runtime_allow(&pdev->dev);
+
+	return 0;
+}
+
+static void intel_lpss_pci_remove(struct pci_dev *pdev)
+{
+	pm_runtime_forbid(&pdev->dev);
+	pm_runtime_get_sync(&pdev->dev);
+
+	intel_lpss_remove(&pdev->dev);
+}
+
+static INTEL_LPSS_PM_OPS(intel_lpss_pci_pm_ops);
+
+static const struct intel_lpss_platform_info spt_info = {
+	.clk_rate = 120000000,
+};
+
+static const struct intel_lpss_platform_info spt_uart_info = {
+	.clk_rate = 120000000,
+	.clk_con_id = "baudclk",
+};
+
+static const struct pci_device_id intel_lpss_pci_ids[] = {
+	/* SPT-LP */
+	{ PCI_VDEVICE(INTEL, 0x9d27), (kernel_ulong_t)&spt_uart_info },
+	{ PCI_VDEVICE(INTEL, 0x9d28), (kernel_ulong_t)&spt_uart_info },
+	{ PCI_VDEVICE(INTEL, 0x9d29), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d2a), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d60), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d61), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d62), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d63), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d64), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d65), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0x9d66), (kernel_ulong_t)&spt_uart_info },
+	/* SPT-H */
+	{ PCI_VDEVICE(INTEL, 0xa127), (kernel_ulong_t)&spt_uart_info },
+	{ PCI_VDEVICE(INTEL, 0xa128), (kernel_ulong_t)&spt_uart_info },
+	{ PCI_VDEVICE(INTEL, 0xa129), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0xa12a), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0xa160), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0xa161), (kernel_ulong_t)&spt_info },
+	{ PCI_VDEVICE(INTEL, 0xa166), (kernel_ulong_t)&spt_uart_info },
+	{ }
+};
+MODULE_DEVICE_TABLE(pci, intel_lpss_pci_ids);
+
+static struct pci_driver intel_lpss_pci_driver = {
+	.name = "intel-lpss",
+	.id_table = intel_lpss_pci_ids,
+	.probe = intel_lpss_pci_probe,
+	.remove = intel_lpss_pci_remove,
+	.driver = {
+		.pm = &intel_lpss_pci_pm_ops,
+	},
+};
+
+module_pci_driver(intel_lpss_pci_driver);
+
+MODULE_AUTHOR("Andy Shevchenko <[email protected]>");
+MODULE_AUTHOR("Mika Westerberg <[email protected]>");
+MODULE_DESCRIPTION("Intel LPSS PCI driver");
+MODULE_LICENSE("GPL v2");
drivers/mfd/intel-lpss.c new file +524 -0
new file mode 100644
index 0000000..fdf4d5c
@@ -0,0 +1,524 @@
+/*
+ * Intel Sunrisepoint LPSS core support.
+ *
+ * Copyright (C) 2015, Intel Corporation
+ *
+ * Authors: Andy Shevchenko <[email protected]>
+ *          Mika Westerberg <[email protected]>
+ *          Heikki Krogerus <[email protected]>
+ *          Jarkko Nikula <[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.
+ */
+
+#include <linux/clk.h>
+#include <linux/clkdev.h>
+#include <linux/clk-provider.h>
+#include <linux/debugfs.h>
+#include <linux/idr.h>
+#include <linux/ioport.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/mfd/core.h>
+#include <linux/pm_qos.h>
+#include <linux/pm_runtime.h>
+#include <linux/seq_file.h>
+
+#include "intel-lpss.h"
+
+#define LPSS_DEV_OFFSET		0x000
+#define LPSS_DEV_SIZE		0x200
+#define LPSS_PRIV_OFFSET	0x200
+#define LPSS_PRIV_SIZE		0x100
+#define LPSS_IDMA64_OFFSET	0x800
+#define LPSS_IDMA64_SIZE	0x800
+
+/* Offsets from lpss->priv */
+#define LPSS_PRIV_RESETS		0x04
+#define LPSS_PRIV_RESETS_FUNC		BIT(2)
+#define LPSS_PRIV_RESETS_IDMA		0x3
+
+#define LPSS_PRIV_ACTIVELTR		0x10
+#define LPSS_PRIV_IDLELTR		0x14
+
+#define LPSS_PRIV_LTR_REQ		BIT(15)
+#define LPSS_PRIV_LTR_SCALE_MASK	0xc00
+#define LPSS_PRIV_LTR_SCALE_1US		0x800
+#define LPSS_PRIV_LTR_SCALE_32US	0xc00
+#define LPSS_PRIV_LTR_VALUE_MASK	0x3ff
+
+#define LPSS_PRIV_SSP_REG		0x20
+#define LPSS_PRIV_SSP_REG_DIS_DMA_FIN	BIT(0)
+
+#define LPSS_PRIV_REMAP_ADDR_LO		0x40
+#define LPSS_PRIV_REMAP_ADDR_HI		0x44
+
+#define LPSS_PRIV_CAPS			0xfc
+#define LPSS_PRIV_CAPS_NO_IDMA		BIT(8)
+#define LPSS_PRIV_CAPS_TYPE_SHIFT	4
+#define LPSS_PRIV_CAPS_TYPE_MASK	(0xf << LPSS_PRIV_CAPS_TYPE_SHIFT)
+
+/* This matches the type field in CAPS register */
+enum intel_lpss_dev_type {
+	LPSS_DEV_I2C = 0,
+	LPSS_DEV_UART,
+	LPSS_DEV_SPI,
+};
+
+struct intel_lpss {
+	const struct intel_lpss_platform_info *info;
+	enum intel_lpss_dev_type type;
+	struct clk *clk;
+	struct clk_lookup *clock;
+	const struct mfd_cell *cell;
+	struct device *dev;
+	void __iomem *priv;
+	int devid;
+	u32 caps;
+	u32 active_ltr;
+	u32 idle_ltr;
+	struct dentry *debugfs;
+};
+
+static const struct resource intel_lpss_dev_resources[] = {
+	DEFINE_RES_MEM_NAMED(LPSS_DEV_OFFSET, LPSS_DEV_SIZE, "lpss_dev"),
+	DEFINE_RES_MEM_NAMED(LPSS_PRIV_OFFSET, LPSS_PRIV_SIZE, "lpss_priv"),
+	DEFINE_RES_IRQ(0),
+};
+
+static const struct resource intel_lpss_idma64_resources[] = {
+	DEFINE_RES_MEM(LPSS_IDMA64_OFFSET, LPSS_IDMA64_SIZE),
+	DEFINE_RES_IRQ(0),
+};
+
+#define LPSS_IDMA64_DRIVER_NAME		"idma64"
+
+/*
+ * Cells needs to be ordered so that the iDMA is created first. This is
+ * because we need to be sure the DMA is available when the host controller
+ * driver is probed.
+ */
+static const struct mfd_cell intel_lpss_idma64_cell = {
+	.name = LPSS_IDMA64_DRIVER_NAME,
+	.num_resources = ARRAY_SIZE(intel_lpss_idma64_resources),
+	.resources = intel_lpss_idma64_resources,
+};
+
+static const struct mfd_cell intel_lpss_i2c_cell = {
+	.name = "i2c_designware",
+	.num_resources = ARRAY_SIZE(intel_lpss_dev_resources),
+	.resources = intel_lpss_dev_resources,
+};
+
+static const struct mfd_cell intel_lpss_uart_cell = {
+	.name = "dw-apb-uart",
+	.num_resources = ARRAY_SIZE(intel_lpss_dev_resources),
+	.resources = intel_lpss_dev_resources,
+};
+
+static const struct mfd_cell intel_lpss_spi_cell = {
+	.name = "pxa2xx-spi",
+	.num_resources = ARRAY_SIZE(intel_lpss_dev_resources),
+	.resources = intel_lpss_dev_resources,
+};
+
+static DEFINE_IDA(intel_lpss_devid_ida);
+static struct dentry *intel_lpss_debugfs;
+
+static int intel_lpss_request_dma_module(const char *name)
+{
+	static bool intel_lpss_dma_requested;
+
+	if (intel_lpss_dma_requested)
+		return 0;
+
+	intel_lpss_dma_requested = true;
+	return request_module("%s", name);
+}
+
+static void intel_lpss_cache_ltr(struct intel_lpss *lpss)
+{
+	lpss->active_ltr = readl(lpss->priv + LPSS_PRIV_ACTIVELTR);
+	lpss->idle_ltr = readl(lpss->priv + LPSS_PRIV_IDLELTR);
+}
+
+static int intel_lpss_debugfs_add(struct intel_lpss *lpss)
+{
+	struct dentry *dir;
+
+	dir = debugfs_create_dir(dev_name(lpss->dev), intel_lpss_debugfs);
+	if (IS_ERR(dir))
+		return PTR_ERR(dir);
+
+	/* Cache the values into lpss structure */
+	intel_lpss_cache_ltr(lpss);
+
+	debugfs_create_x32("capabilities", S_IRUGO, dir, &lpss->caps);
+	debugfs_create_x32("active_ltr", S_IRUGO, dir, &lpss->active_ltr);
+	debugfs_create_x32("idle_ltr", S_IRUGO, dir, &lpss->idle_ltr);
+
+	lpss->debugfs = dir;
+	return 0;
+}
+
+static void intel_lpss_debugfs_remove(struct intel_lpss *lpss)
+{
+	debugfs_remove_recursive(lpss->debugfs);
+}
+
+static void intel_lpss_ltr_set(struct device *dev, s32 val)
+{
+	struct intel_lpss *lpss = dev_get_drvdata(dev);
+	u32 ltr;
+
+	/*
+	 * Program latency tolerance (LTR) accordingly what has been asked
+	 * by the PM QoS layer or disable it in case we were passed
+	 * negative value or PM_QOS_LATENCY_ANY.
+	 */
+	ltr = readl(lpss->priv + LPSS_PRIV_ACTIVELTR);
+
+	if (val == PM_QOS_LATENCY_ANY || val < 0) {
+		ltr &= ~LPSS_PRIV_LTR_REQ;
+	} else {
+		ltr |= LPSS_PRIV_LTR_REQ;
+		ltr &= ~LPSS_PRIV_LTR_SCALE_MASK;
+		ltr &= ~LPSS_PRIV_LTR_VALUE_MASK;
+
+		if (val > LPSS_PRIV_LTR_VALUE_MASK)
+			ltr |= LPSS_PRIV_LTR_SCALE_32US | val >> 5;
+		else
+			ltr |= LPSS_PRIV_LTR_SCALE_1US | val;
+	}
+
+	if (ltr == lpss->active_ltr)
+		return;
+
+	writel(ltr, lpss->priv + LPSS_PRIV_ACTIVELTR);
+	writel(ltr, lpss->priv + LPSS_PRIV_IDLELTR);
+
+	/* Cache the values into lpss structure */
+	intel_lpss_cache_ltr(lpss);
+}
+
+static void intel_lpss_ltr_expose(struct intel_lpss *lpss)
+{
+	lpss->dev->power.set_latency_tolerance = intel_lpss_ltr_set;
+	dev_pm_qos_expose_latency_tolerance(lpss->dev);
+}
+
+static void intel_lpss_ltr_hide(struct intel_lpss *lpss)
+{
+	dev_pm_qos_hide_latency_tolerance(lpss->dev);
+	lpss->dev->power.set_latency_tolerance = NULL;
+}
+
+static int intel_lpss_assign_devs(struct intel_lpss *lpss)
+{
+	unsigned int type;
+
+	type = lpss->caps & LPSS_PRIV_CAPS_TYPE_MASK;
+	type >>= LPSS_PRIV_CAPS_TYPE_SHIFT;
+
+	switch (type) {
+	case LPSS_DEV_I2C:
+		lpss->cell = &intel_lpss_i2c_cell;
+		break;
+	case LPSS_DEV_UART:
+		lpss->cell = &intel_lpss_uart_cell;
+		break;
+	case LPSS_DEV_SPI:
+		lpss->cell = &intel_lpss_spi_cell;
+		break;
+	default:
+		return -ENODEV;
+	}
+
+	lpss->type = type;
+
+	return 0;
+}
+
+static bool intel_lpss_has_idma(const struct intel_lpss *lpss)
+{
+	return (lpss->caps & LPSS_PRIV_CAPS_NO_IDMA) == 0;
+}
+
+static void intel_lpss_set_remap_addr(const struct intel_lpss *lpss)
+{
+	resource_size_t addr = lpss->info->mem->start;
+
+	writel(addr, lpss->priv + LPSS_PRIV_REMAP_ADDR_LO);
+#if BITS_PER_LONG > 32
+	writel(addr >> 32, lpss->priv + LPSS_PRIV_REMAP_ADDR_HI);
+#else
+	writel(0, lpss->priv + LPSS_PRIV_REMAP_ADDR_HI);
+#endif
+}
+
+static void intel_lpss_deassert_reset(const struct intel_lpss *lpss)
+{
+	u32 value = LPSS_PRIV_RESETS_FUNC | LPSS_PRIV_RESETS_IDMA;
+
+	/* Bring out the device from reset */
+	writel(value, lpss->priv + LPSS_PRIV_RESETS);
+}
+
+static void intel_lpss_init_dev(const struct intel_lpss *lpss)
+{
+	u32 value = LPSS_PRIV_SSP_REG_DIS_DMA_FIN;
+
+	intel_lpss_deassert_reset(lpss);
+
+	if (!intel_lpss_has_idma(lpss))
+		return;
+
+	intel_lpss_set_remap_addr(lpss);
+
+	/* Make sure that SPI multiblock DMA transfers are re-enabled */
+	if (lpss->type == LPSS_DEV_SPI)
+		writel(value, lpss->priv + LPSS_PRIV_SSP_REG);
+}
+
+static void intel_lpss_unregister_clock_tree(struct clk *clk)
+{
+	struct clk *parent;
+
+	while (clk) {
+		parent = clk_get_parent(clk);
+		clk_unregister(clk);
+		clk = parent;
+	}
+}
+
+static int intel_lpss_register_clock_divider(struct intel_lpss *lpss,
+					     const char *devname,
+					     struct clk **clk)
+{
+	char name[32];
+	struct clk *tmp = *clk;
+
+	snprintf(name, sizeof(name), "%s-enable", devname);
+	tmp = clk_register_gate(NULL, name, __clk_get_name(tmp), 0,
+				lpss->priv, 0, 0, NULL);
+	if (IS_ERR(tmp))
+		return PTR_ERR(tmp);
+
+	snprintf(name, sizeof(name), "%s-div", devname);
+	tmp = clk_register_fractional_divider(NULL, name, __clk_get_name(tmp),
+					      0, lpss->priv, 1, 15, 16, 15, 0,
+					      NULL);
+	if (IS_ERR(tmp))
+		return PTR_ERR(tmp);
+	*clk = tmp;
+
+	snprintf(name, sizeof(name), "%s-update", devname);
+	tmp = clk_register_gate(NULL, name, __clk_get_name(tmp),
+				CLK_SET_RATE_PARENT, lpss->priv, 31, 0, NULL);
+	if (IS_ERR(tmp))
+		return PTR_ERR(tmp);
+	*clk = tmp;
+
+	return 0;
+}
+
+static int intel_lpss_register_clock(struct intel_lpss *lpss)
+{
+	const struct mfd_cell *cell = lpss->cell;
+	struct clk *clk;
+	char devname[24];
+	int ret;
+
+	if (!lpss->info->clk_rate)
+		return 0;
+
+	/* Root clock */
+	clk = clk_register_fixed_rate(NULL, dev_name(lpss->dev), NULL,
+				      CLK_IS_ROOT, lpss->info->clk_rate);
+	if (IS_ERR(clk))
+		return PTR_ERR(clk);
+
+	snprintf(devname, sizeof(devname), "%s.%d", cell->name, lpss->devid);
+
+	/*
+	 * Support for clock divider only if it has some preset value.
+	 * Otherwise we assume that the divider is not used.
+	 */
+	if (lpss->type != LPSS_DEV_I2C) {
+		ret = intel_lpss_register_clock_divider(lpss, devname, &clk);
+		if (ret)
+			goto err_clk_register;
+	}
+
+	ret = -ENOMEM;
+
+	/* Clock for the host controller */
+	lpss->clock = clkdev_create(clk, lpss->info->clk_con_id, "%s", devname);
+	if (!lpss->clock)
+		goto err_clk_register;
+
+	lpss->clk = clk;
+
+	return 0;
+
+err_clk_register:
+	intel_lpss_unregister_clock_tree(clk);
+
+	return ret;
+}
+
+static void intel_lpss_unregister_clock(struct intel_lpss *lpss)
+{
+	if (IS_ERR_OR_NULL(lpss->clk))
+		return;
+
+	clkdev_drop(lpss->clock);
+	intel_lpss_unregister_clock_tree(lpss->clk);
+}
+
+int intel_lpss_probe(struct device *dev,
+		     const struct intel_lpss_platform_info *info)
+{
+	struct intel_lpss *lpss;
+	int ret;
+
+	if (!info || !info->mem || info->irq <= 0)
+		return -EINVAL;
+
+	lpss = devm_kzalloc(dev, sizeof(*lpss), GFP_KERNEL);
+	if (!lpss)
+		return -ENOMEM;
+
+	lpss->priv = devm_ioremap(dev, info->mem->start + LPSS_PRIV_OFFSET,
+				  LPSS_PRIV_SIZE);
+	if (!lpss->priv)
+		return -ENOMEM;
+
+	lpss->info = info;
+	lpss->dev = dev;
+	lpss->caps = readl(lpss->priv + LPSS_PRIV_CAPS);
+
+	dev_set_drvdata(dev, lpss);
+
+	ret = intel_lpss_assign_devs(lpss);
+	if (ret)
+		return ret;
+
+	intel_lpss_init_dev(lpss);
+
+	lpss->devid = ida_simple_get(&intel_lpss_devid_ida, 0, 0, GFP_KERNEL);
+	if (lpss->devid < 0)
+		return lpss->devid;
+
+	ret = intel_lpss_register_clock(lpss);
+	if (ret)
+		goto err_clk_register;
+
+	intel_lpss_ltr_expose(lpss);
+
+	ret = intel_lpss_debugfs_add(lpss);
+	if (ret)
+		dev_warn(dev, "Failed to create debugfs entries\n");
+
+	if (intel_lpss_has_idma(lpss)) {
+		/*
+		 * Ensure the DMA driver is loaded before the host
+		 * controller device appears, so that the host controller
+		 * driver can request its DMA channels as early as
+		 * possible.
+		 *
+		 * If the DMA module is not there that's OK as well.
+		 */
+		intel_lpss_request_dma_module(LPSS_IDMA64_DRIVER_NAME);
+
+		ret = mfd_add_devices(dev, lpss->devid, &intel_lpss_idma64_cell,
+				      1, info->mem, info->irq, NULL);
+		if (ret)
+			dev_warn(dev, "Failed to add %s, fallback to PIO\n",
+				 LPSS_IDMA64_DRIVER_NAME);
+	}
+
+	ret = mfd_add_devices(dev, lpss->devid, lpss->cell,
+			      1, info->mem, info->irq, NULL);
+	if (ret)
+		goto err_remove_ltr;
+
+	return 0;
+
+err_remove_ltr:
+	intel_lpss_debugfs_remove(lpss);
+	intel_lpss_ltr_hide(lpss);
+
+err_clk_register:
+	ida_simple_remove(&intel_lpss_devid_ida, lpss->devid);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(intel_lpss_probe);
+
+void intel_lpss_remove(struct device *dev)
+{
+	struct intel_lpss *lpss = dev_get_drvdata(dev);
+
+	mfd_remove_devices(dev);
+	intel_lpss_debugfs_remove(lpss);
+	intel_lpss_ltr_hide(lpss);
+	intel_lpss_unregister_clock(lpss);
+	ida_simple_remove(&intel_lpss_devid_ida, lpss->devid);
+}
+EXPORT_SYMBOL_GPL(intel_lpss_remove);
+
+static int resume_lpss_device(struct device *dev, void *data)
+{
+	pm_runtime_resume(dev);
+	return 0;
+}
+
+int intel_lpss_prepare(struct device *dev)
+{
+	/*
+	 * Resume both child devices before entering system sleep. This
+	 * ensures that they are in proper state before they get suspended.
+	 */
+	device_for_each_child_reverse(dev, NULL, resume_lpss_device);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(intel_lpss_prepare);
+
+int intel_lpss_suspend(struct device *dev)
+{
+	return 0;
+}
+EXPORT_SYMBOL_GPL(intel_lpss_suspend);
+
+int intel_lpss_resume(struct device *dev)
+{
+	struct intel_lpss *lpss = dev_get_drvdata(dev);
+
+	intel_lpss_init_dev(lpss);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(intel_lpss_resume);
+
+static int __init intel_lpss_init(void)
+{
+	intel_lpss_debugfs = debugfs_create_dir("intel_lpss", NULL);
+	return 0;
+}
+module_init(intel_lpss_init);
+
+static void __exit intel_lpss_exit(void)
+{
+	debugfs_remove(intel_lpss_debugfs);
+}
+module_exit(intel_lpss_exit);
+
+MODULE_AUTHOR("Andy Shevchenko <[email protected]>");
+MODULE_AUTHOR("Mika Westerberg <[email protected]>");
+MODULE_AUTHOR("Heikki Krogerus <[email protected]>");
+MODULE_AUTHOR("Jarkko Nikula <[email protected]>");
+MODULE_DESCRIPTION("Intel LPSS core driver");
+MODULE_LICENSE("GPL v2");
drivers/mfd/intel-lpss.h new file +62 -0
new file mode 100644
index 0000000..f28cb28a
@@ -0,0 +1,62 @@
+/*
+ * Intel LPSS core support.
+ *
+ * Copyright (C) 2015, Intel Corporation
+ *
+ * Authors: Andy Shevchenko <[email protected]>
+ *          Mika Westerberg <[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.
+ */
+
+#ifndef __MFD_INTEL_LPSS_H
+#define __MFD_INTEL_LPSS_H
+
+struct device;
+struct resource;
+
+struct intel_lpss_platform_info {
+	struct resource *mem;
+	int irq;
+	unsigned long clk_rate;
+	const char *clk_con_id;
+};
+
+int intel_lpss_probe(struct device *dev,
+		     const struct intel_lpss_platform_info *info);
+void intel_lpss_remove(struct device *dev);
+
+#ifdef CONFIG_PM
+int intel_lpss_prepare(struct device *dev);
+int intel_lpss_suspend(struct device *dev);
+int intel_lpss_resume(struct device *dev);
+
+#ifdef CONFIG_PM_SLEEP
+#define INTEL_LPSS_SLEEP_PM_OPS			\
+	.prepare = intel_lpss_prepare,		\
+	.suspend = intel_lpss_suspend,		\
+	.resume = intel_lpss_resume,		\
+	.freeze = intel_lpss_suspend,		\
+	.thaw = intel_lpss_resume,		\
+	.poweroff = intel_lpss_suspend,		\
+	.restore = intel_lpss_resume,
+#endif
+
+#define INTEL_LPSS_RUNTIME_PM_OPS		\
+	.runtime_suspend = intel_lpss_suspend,	\
+	.runtime_resume = intel_lpss_resume,
+
+#else /* !CONFIG_PM */
+#define INTEL_LPSS_SLEEP_PM_OPS
+#define INTEL_LPSS_RUNTIME_PM_OPS
+#endif /* CONFIG_PM */
+
+#define INTEL_LPSS_PM_OPS(name)			\
+const struct dev_pm_ops name = {		\
+	INTEL_LPSS_SLEEP_PM_OPS			\
+	INTEL_LPSS_RUNTIME_PM_OPS		\
+}
+
+#endif /* __MFD_INTEL_LPSS_H */
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#20 lore raw

Re: [PATCH v6 8/8] mfd: Add support for Intel Sunrisepoint LPSS devices

In reply to Andy Shevchenko

Hi Andy,

Is there a data sheet or reference manual publicly available for the LPSS?

Quoting Andy Shevchenko (2015-07-27 08:04:03)

Quoted 13 lines
drivers/mfd/intel-lpss.c new file +7 -0
new file mode 100644
index 0000000..fdf4d5c
@@ -0,0 +1,524 @@
+/*
+ * Intel Sunrisepoint LPSS core support.
+ *
+ * Copyright (C) 2015, Intel Corporation
+ *
+ * Authors: Andy Shevchenko <[email protected]>
+ *          Mika Westerberg <[email protected]>

drivers/mfd/intel-lpss.c line 17

Quoted 10 lines
drivers/mfd/intel-lpss.c new file +10 -0
+ *          Heikki Krogerus <[email protected]>
+ *          Jarkko Nikula <[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.
+ */
+
+#include <linux/clk.h>
+#include <linux/clkdev.h>

I hope that we can get rid of the above two headers. See more below.

drivers/mfd/intel-lpss.c line 74

Quoted 5 lines
drivers/mfd/intel-lpss.c new file +5 -0
+struct intel_lpss {
+       const struct intel_lpss_platform_info *info;
+       enum intel_lpss_dev_type type;
+       struct clk *clk;
+       struct clk_lookup *clock;

Same here. I don't see a use for the clk_lookup?

drivers/mfd/intel-lpss.c line 293

Quoted 9 lines
drivers/mfd/intel-lpss.c new file +9 -0
+static void intel_lpss_unregister_clock_tree(struct clk *clk)
+{
+       struct clk *parent;
+
+       while (clk) {
+               parent = clk_get_parent(clk);
+               clk_unregister(clk);
+               clk = parent;
+       }

clk_unregister is for clock provider drivers that might be unloaded at run-time, so it is nice of you to support it. However this implementation isn't fully safe.

For instance, what happens if one day you plug in another clock provider on top of this one? Your root clock here no longer be a root clock; it would have a parent, and possibly a whole tree of clocks up above. If the above happens then you'll unregister all of the parents up until you hit a root clock!

A better alternative is for your clock provider driver (i.e. _this_ driver) to keep track of which clocks its registers and then unregister them later.

In fact we have a nice function to help with that: devm_clk_register! There are plenty of examples of how that is used. It doesn't work with the "basic" clock types you are using here, but I'm not convinced that you should be using those anyways...

Quoted 16 lines
drivers/mfd/intel-lpss.c new file +16 -0
+}
+
+static int intel_lpss_register_clock_divider(struct intel_lpss *lpss,
+                                            const char *devname,
+                                            struct clk **clk)
+{
+       char name[32];
+       struct clk *tmp = *clk;
+
+       snprintf(name, sizeof(name), "%s-enable", devname);
+       tmp = clk_register_gate(NULL, name, __clk_get_name(tmp), 0,
+                               lpss->priv, 0, 0, NULL);
+       if (IS_ERR(tmp))
+               return PTR_ERR(tmp);
+
+       snprintf(name, sizeof(name), "%s-div", devname);

drivers/mfd/intel-lpss.c line 319

Quoted 10 lines
drivers/mfd/intel-lpss.c new file +10 -0
+       tmp = clk_register_fractional_divider(NULL, name, __clk_get_name(tmp),
+                                             0, lpss->priv, 1, 15, 16, 15, 0,
+                                             NULL);
+       if (IS_ERR(tmp))
+               return PTR_ERR(tmp);
+       *clk = tmp;
+
+       snprintf(name, sizeof(name), "%s-update", devname);
+       tmp = clk_register_gate(NULL, name, __clk_get_name(tmp),
+                               CLK_SET_RATE_PARENT, lpss->priv, 31, 0, NULL);

A gate, then a fixed divider, and then another gate? Seems weird. Do you have other clock signals that are not enumerated here?

Where did those clock names come from? Are they like that in the data sheet or just made up for the sake of writing this driver?

Why are the names generated dynamically? Some static data might be better here. Do you anticipate having more than one LPSS in a running system? I really hate __clk_get_name and would like to remove it some day. If we can avoid using it here then I would be grateful.

Also, why do you register these clocks with clkdev? Do other drivers clk_get them? When are the gates ever enabled/disabled?

Quoted 24 lines
drivers/mfd/intel-lpss.c new file +24 -0
+       if (IS_ERR(tmp))
+               return PTR_ERR(tmp);
+       *clk = tmp;
+
+       return 0;
+}
+
+static int intel_lpss_register_clock(struct intel_lpss *lpss)
+{
+       const struct mfd_cell *cell = lpss->cell;
+       struct clk *clk;
+       char devname[24];
+       int ret;
+
+       if (!lpss->info->clk_rate)
+               return 0;
+
+       /* Root clock */
+       clk = clk_register_fixed_rate(NULL, dev_name(lpss->dev), NULL,
+                                     CLK_IS_ROOT, lpss->info->clk_rate);
+       if (IS_ERR(clk))
+               return PTR_ERR(clk);
+
+       snprintf(devname, sizeof(devname), "%s.%d", cell->name, lpss->devid);

drivers/mfd/intel-lpss.c line 353

Quoted 10 lines
drivers/mfd/intel-lpss.c new file +10 -0
+
+       /*
+        * Support for clock divider only if it has some preset value.
+        * Otherwise we assume that the divider is not used.
+        */
+       if (lpss->type != LPSS_DEV_I2C) {
+               ret = intel_lpss_register_clock_divider(lpss, devname, &clk);
+               if (ret)
+                       goto err_clk_register;
+       }

Do ACPI tables encode any meaningful data about clock signals? I gave some advice on this topic a couple of years ago to folks hacking on ACPI stuff in Linux, but I haven't kept up with the topic.

By comparison, most ARM-based platforms use Devicetree to link up a clock provider to a clock consumer. I do not think similar infrastructure exists yet for ACPI-based designs, but I wonder if the clock data exists in newer ACPI tables to at least make it possible?

Regards,
Mike

#21 lore raw

Re: [PATCH v6 8/8] mfd: Add support for Intel Sunrisepoint LPSS devices

Quoted 1 line · Michael Turquette

Hi Andy,

[cut]

Quoted 8 lines

Do ACPI tables encode any meaningful data about clock signals? I gave some advice on this topic a couple of years ago to folks hacking on ACPI stuff in Linux, but I haven't kept up with the topic.

By comparison, most ARM-based platforms use Devicetree to link up a clock provider to a clock consumer. I do not think similar infrastructure exists yet for ACPI-based designs, but I wonder if the clock data exists in newer ACPI tables to at least make it possible?

No, it doesn't.

Thanks,
Rafael

author reviewer
#22 lore raw

Re: [PATCH v6 8/8] mfd: Add support for Intel Sunrisepoint LPSS devices

In reply to Michael Turquette
Quoted 4 lines · Michael Turquette

Hi Andy,

Is there a data sheet or reference manual publicly available for the LPSS?

Yes, I mentioned it in cover letter. For your convenience I just copy'n'paste it here:

https://download.01.org/future-platform-configuration
-hub/skylake/register-definitions/332219-002.pdf

Quoted 30 lines

Quoting Andy Shevchenko (2015-07-27 08:04:03)

Quoted 22 lines
drivers/mfd/intel-lpss.c new file +11 -0
new file mode 100644
index 0000000..fdf4d5c
@@ -0,0 +1,524 @@
+/*
+ * Intel Sunrisepoint LPSS core support.
+ *
+ * Copyright (C) 2015, Intel Corporation
+ *
+ * Authors: Andy Shevchenko <[email protected]>
+ *          Mika Westerberg <[email protected]>
+ *          Heikki Krogerus <[email protected]>
+ *          Jarkko Nikula <[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.
+ */

drivers/mfd/intel-lpss.c line 17

Quoted 3 lines
drivers/mfd/intel-lpss.c new file +3 -0
+
+#include <linux/clk.h>
+#include <linux/clkdev.h>

I hope that we can get rid of the above two headers. See more below.

Quoted 7 lines

drivers/mfd/intel-lpss.c line 74

Quoted 5 lines
drivers/mfd/intel-lpss.c new file +5 -0
+struct intel_lpss {
+       const struct intel_lpss_platform_info *info;
+       enum intel_lpss_dev_type type;
+       struct clk *clk;
+       struct clk_lookup *clock;

Same here. I don't see a use for the clk_lookup?

clkdev_create() returns it.

Quoted 25 lines

drivers/mfd/intel-lpss.c line 293

Quoted 10 lines
drivers/mfd/intel-lpss.c new file +9 -0
+static void intel_lpss_unregister_clock_tree(struct clk *clk)
+{
+       struct clk *parent;
+
+       while (clk) {
+               parent = clk_get_parent(clk);
+               clk_unregister(clk);
+               clk = parent;
+       }

clk_unregister is for clock provider drivers that might be unloaded at
run-time, so it is nice of you to support it. However this implementation isn't fully safe.

For instance, what happens if one day you plug in another clock provider
on top of this one? Your root clock here no longer be a root clock; it
would have a parent, and possibly a whole tree of clocks up above. If the above happens then you'll unregister all of the parents up until you
hit a root clock!

Quoted 7 lines

A better alternative is for your clock provider driver (i.e. _this_ driver) to keep track of which clocks its registers and then unregister
them later.

In fact we have a nice function to help with that: devm_clk_register!

I will look at this.

Quoted 41 lines

There are plenty of examples of how that is used. It doesn't work with
the "basic" clock types you are using here, but I'm not convinced that
you should be using those anyways...

Quoted 29 lines
drivers/mfd/intel-lpss.c new file +3 -0
+}
+
+static int intel_lpss_register_clock_divider(struct intel_lpss 

*lpss,

drivers/mfd/intel-lpss.c new file +14 -0
+                                            const char *devname,
+                                            struct clk **clk)
+{
+       char name[32];
+       struct clk *tmp = *clk;
+
+       snprintf(name, sizeof(name), "%s-enable", devname);
+       tmp = clk_register_gate(NULL, name, __clk_get_name(tmp), 0,
+                               lpss->priv, 0, 0, NULL);
+       if (IS_ERR(tmp))
+               return PTR_ERR(tmp);
+
+       snprintf(name, sizeof(name), "%s-div", devname);
+       tmp = clk_register_fractional_divider(NULL, name, 

__clk_get_name(tmp),
+ 0, lpss->priv, 1, 15,
16, 15, 0,

drivers/mfd/intel-lpss.c new file +8 -0
+                                             NULL);
+       if (IS_ERR(tmp))
+               return PTR_ERR(tmp);
+       *clk = tmp;
+
+       snprintf(name, sizeof(name), "%s-update", devname);
+       tmp = clk_register_gate(NULL, name, __clk_get_name(tmp),
+                               CLK_SET_RATE_PARENT, lpss->priv, 
Quoted 1 line

31, 0, NULL);

A gate, then a fixed divider, and then another gate? Seems weird. Do you
have other clock signals that are not enumerated here?

Quoted 2 lines

Where did those clock names come from? Are they like that in the data sheet or just made up for the sake of writing this driver?

For example, chapter 1.0, section 1.1 describes clock gating and fractional divider for UART. Also the high bit updates the actual divider after we supply m and n values.

Since we have the clock implementation which enables one bit per device we have to do this in a way of clock tree.

If you have any better suggestion we would improve our driver.

Quoted 4 lines

Why are the names generated dynamically? Some static data might be better here. Do you anticipate having more than one LPSS in a running system?

What do you mean under LPSS? We mean the entire IP block consists of several devices (we only can say which and in what configuration dynamically). devname here corresponds to the each device in the LPSS IP block.

As an example
clock enable_cnt prepare_cnt rate
accuracy phase
-----------------------------------------------------------------------
-----------------
0000:00:1e.3 0 0 120000000

       0 0  
    pxa2xx-spi.6-enable                   0            0   120000000   
       0 0  
       pxa2xx-spi.6-div                   0            0   120000000   
       0 0  
          pxa2xx-spi.6-update             0            0   120000000   
       0 0  
 0000:00:1e.0                             0            0   120000000   
       0 0  
    dw-apb-uart.5-enable                  0            0   120000000   
       0 0  
       dw-apb-uart.5-div                  0            0     1843200   
       0 0  
          dw-apb-uart.5-update            0            0     1843200   
       0 0  
 0000:00:19.0                             0            0   120000000   
       0 0  
    dw-apb-uart.4-enable                  0            0   120000000   
       0 0  
       dw-apb-uart.4-div                  0            0     1843200   
       0 0  
          dw-apb-uart.4-update            0            0     1843200   
       0 0  
Quoted 5 lines

I really hate __clk_get_name and would like to remove it some day. If we can avoid using it here then I would be grateful.

Also, why do you register these clocks with clkdev? Do other drivers clk_get them?

Yes. The actual consumers do that: spi-pxa2xx, 8250_dw, i2c-designware.

Quoted 1 line

When are the gates ever enabled/disabled?

In the mentioned drivers. For example, for UART is a clock called 'baudclk' which is changed whenever user changes a baud rate.

Signature · 2 lines
Andy Shevchenko <[email protected]>
Intel Finland Oy
#23 lore raw
In reply to Andy Shevchenko

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#24 lore raw
Quoted 10 lines · Lee Jones

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Lee, can you suggest an alternative then?

Why we are doing it like this is that number of different LPSS devices changes from SoC to SoC. In addition to that the device (called "slice") might have iDMA block or not.

Since the drivers in question (pxa2xx-spi, i2c-designware and 8250_dw) use standard clk framework to request their clocks the Linux device must have clock registered which matches the device in advance.

Because we add the host controller device dynamically (from the MFD driver) based on how many devices are actually present, we need somehow predict what would be the correct name and instance number for that device to get the clock for it. That's the reason we use IDA here along with the cell name (or driver name).

#25 lore raw
Quoted 22 lines · Mika Westerberg
Quoted 10 lines · Lee Jones

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Lee, can you suggest an alternative then?

Why we are doing it like this is that number of different LPSS devices changes from SoC to SoC. In addition to that the device (called "slice") might have iDMA block or not.

Since the drivers in question (pxa2xx-spi, i2c-designware and 8250_dw) use standard clk framework to request their clocks the Linux device must have clock registered which matches the device in advance.

Quoted 5 lines

Because we add the host controller device dynamically (from the MFD driver) based on how many devices are actually present, we need somehow predict what would be the correct name and instance number for that device to get the clock for it. That's the reason we use IDA here along with the cell name (or driver name).

I'm sure there are perfectly viable reasons for you doing this. And I don't know the CCF well enough to know whether it's the best idea or not, or else I would have made a suggestion rather than waiting all this time.

It's for this reason that I needed Mike (now Stephen) to take a look and give me either an Ack, to say it's the best solution, or to provide a better alternative.

Until that happens, I'm stuck!

Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#26 lore raw
Quoted 35 lines · Lee Jones
Quoted 22 lines · Mika Westerberg
Quoted 10 lines · Lee Jones

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Lee, can you suggest an alternative then?

Why we are doing it like this is that number of different LPSS devices changes from SoC to SoC. In addition to that the device (called "slice") might have iDMA block or not.

Since the drivers in question (pxa2xx-spi, i2c-designware and 8250_dw) use standard clk framework to request their clocks the Linux device must have clock registered which matches the device in advance.

Quoted 5 lines

Because we add the host controller device dynamically (from the MFD driver) based on how many devices are actually present, we need somehow predict what would be the correct name and instance number for that device to get the clock for it. That's the reason we use IDA here along with the cell name (or driver name).

I'm sure there are perfectly viable reasons for you doing this. And I don't know the CCF well enough to know whether it's the best idea or not, or else I would have made a suggestion rather than waiting all this time.

Quoted 5 lines

It's for this reason that I needed Mike (now Stephen) to take a look and give me either an Ack, to say it's the best solution, or to provide a better alternative.

Until that happens, I'm stuck!

Well, what if we had no one at hand to review that code? Would that mean it would not be applicable forever?

Thanks,
Rafael

#27 lore raw
Quoted 36 lines · Rafael J. Wysocki
Quoted 35 lines · Lee Jones
Quoted 22 lines · Mika Westerberg
Quoted 10 lines · Lee Jones

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Lee, can you suggest an alternative then?

Why we are doing it like this is that number of different LPSS devices changes from SoC to SoC. In addition to that the device (called "slice") might have iDMA block or not.

Since the drivers in question (pxa2xx-spi, i2c-designware and 8250_dw) use standard clk framework to request their clocks the Linux device must have clock registered which matches the device in advance.

Quoted 5 lines

Because we add the host controller device dynamically (from the MFD driver) based on how many devices are actually present, we need somehow predict what would be the correct name and instance number for that device to get the clock for it. That's the reason we use IDA here along with the cell name (or driver name).

I'm sure there are perfectly viable reasons for you doing this. And I don't know the CCF well enough to know whether it's the best idea or not, or else I would have made a suggestion rather than waiting all this time.

Quoted 8 lines
Quoted 5 lines

It's for this reason that I needed Mike (now Stephen) to take a look and give me either an Ack, to say it's the best solution, or to provide a better alternative.

Until that happens, I'm stuck!

Well, what if we had no one at hand to review that code? Would that mean it would not be applicable forever?

No, but that's not the case is it?

I don't understand why Mike and Stephen aren't helping!

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#28 lore raw
Quoted 42 lines · Lee Jones
Quoted 40 lines · Rafael J. Wysocki
Quoted 39 lines · Lee Jones
Quoted 22 lines · Mika Westerberg
Quoted 10 lines · Lee Jones

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Lee, can you suggest an alternative then?

Why we are doing it like this is that number of different LPSS devices changes from SoC to SoC. In addition to that the device (called "slice") might have iDMA block or not.

Since the drivers in question (pxa2xx-spi, i2c-designware and 8250_dw) use standard clk framework to request their clocks the Linux device must have clock registered which matches the device in advance.

Quoted 5 lines

Because we add the host controller device dynamically (from the MFD driver) based on how many devices are actually present, we need somehow predict what would be the correct name and instance number for that device to get the clock for it. That's the reason we use IDA here along with the cell name (or driver name).

I'm sure there are perfectly viable reasons for you doing this. And I don't know the CCF well enough to know whether it's the best idea or not, or else I would have made a suggestion rather than waiting all this time.

It's for this reason that I needed Mike (now Stephen) to take a look and give me either an Ack, to say it's the best solution, or to provide a better alternative.

Quoted 8 lines
Quoted 4 lines
Quoted 1 line

Until that happens, I'm stuck!

Well, what if we had no one at hand to review that code? Would that mean it would not be applicable forever?

No, but that's not the case is it?

I don't understand why Mike and Stephen aren't helping!

I'll wait until tomorrow and if we haven't heard anything I'll make a decision.

Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#29 lore raw
Quoted 49 lines · Lee Jones
Quoted 47 lines · Lee Jones
Quoted 44 lines · Rafael J. Wysocki
Quoted 35 lines · Lee Jones
Quoted 22 lines · Mika Westerberg
Quoted 10 lines · Lee Jones

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Lee, can you suggest an alternative then?

Why we are doing it like this is that number of different LPSS devices changes from SoC to SoC. In addition to that the device (called "slice") might have iDMA block or not.

Since the drivers in question (pxa2xx-spi, i2c-designware and 8250_dw) use standard clk framework to request their clocks the Linux device must have clock registered which matches the device in advance.

Quoted 5 lines

Because we add the host controller device dynamically (from the MFD driver) based on how many devices are actually present, we need somehow predict what would be the correct name and instance number for that device to get the clock for it. That's the reason we use IDA here along with the cell name (or driver name).

I'm sure there are perfectly viable reasons for you doing this. And I don't know the CCF well enough to know whether it's the best idea or not, or else I would have made a suggestion rather than waiting all this time.

Quoted 5 lines

It's for this reason that I needed Mike (now Stephen) to take a look and give me either an Ack, to say it's the best solution, or to provide a better alternative.

Until that happens, I'm stuck!

Well, what if we had no one at hand to review that code? Would that mean it would not be applicable forever?

Quoted 6 lines
Quoted 3 lines

No, but that's not the case is it?

I don't understand why Mike and Stephen aren't helping!

I'll wait until tomorrow and if we haven't heard anything I'll make a decision.

OK, thanks!

BTW, I don't have the time to review every single patch using ACPI or one of the PM frameworks. If people who use them make mistakes, it is their burden to fix those mistakes when they show up in testing.

What's happening here is that Andy and Mika are taking the responsibility for fixing the new code if it turns out to be buggy and so it's their problem if it happens to be broken.

And you can still revert commits that introduce bugs as a last resort.

#30 lore raw
Quoted 63 lines · Rafael J. Wysocki
Quoted 49 lines · Lee Jones
Quoted 47 lines · Lee Jones
Quoted 44 lines · Rafael J. Wysocki
Quoted 35 lines · Lee Jones
Quoted 22 lines · Mika Westerberg
Quoted 10 lines · Lee Jones

FAO Stephen Boyd,

Quoted 2 lines

Stephen, can you, please, have a look into patch 8 regarding to clock name matching and other stuff Lee asked?

Patch 8:

"Can you review the clock implementation please? It looks fragile to me as it relies heavily on device names constructed of MFD cell names and IDA numbers cat'ed together!"

Lee, can you suggest an alternative then?

Why we are doing it like this is that number of different LPSS devices changes from SoC to SoC. In addition to that the device (called "slice") might have iDMA block or not.

Since the drivers in question (pxa2xx-spi, i2c-designware and 8250_dw) use standard clk framework to request their clocks the Linux device must have clock registered which matches the device in advance.

Quoted 5 lines

Because we add the host controller device dynamically (from the MFD driver) based on how many devices are actually present, we need somehow predict what would be the correct name and instance number for that device to get the clock for it. That's the reason we use IDA here along with the cell name (or driver name).

I'm sure there are perfectly viable reasons for you doing this. And I don't know the CCF well enough to know whether it's the best idea or not, or else I would have made a suggestion rather than waiting all this time.

Quoted 5 lines

It's for this reason that I needed Mike (now Stephen) to take a look and give me either an Ack, to say it's the best solution, or to provide a better alternative.

Until that happens, I'm stuck!

Well, what if we had no one at hand to review that code? Would that mean it would not be applicable forever?

Quoted 6 lines
Quoted 3 lines

No, but that's not the case is it?

I don't understand why Mike and Stephen aren't helping!

I'll wait until tomorrow and if we haven't heard anything I'll make a decision.

OK, thanks!

BTW, I don't have the time to review every single patch using ACPI or one of the PM frameworks. If people who use them make mistakes, it is their burden to fix those mistakes when they show up in testing.

Quoted 5 lines

What's happening here is that Andy and Mika are taking the responsibility for fixing the new code if it turns out to be buggy and so it's their problem if it happens to be broken.

And you can still revert commits that introduce bugs as a last resort.

I'm fine with that in principle. My issue here was that it looks wrong to me. I just don't know enough about the inner workings of the CCF to be able to say that for sure, or to provide a suitable alternative. I think, probably the correct thing to do is to have an accompanying clock driver, but who knows (I guess Stephen and Mike to, but are seemingly unwilling to help).

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#31 lore raw
In reply to Andy Shevchenko

Enjoy,

The following changes since commit bc0195aad0daa2ad5b0d76cce22b167bc3435590:

Linux 4.2-rc2 (2015-07-12 15:10:30 -0700)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd.git tags/ib-mfd-base-acpi-dma-v4.3

for you to fetch changes up to 4b45efe8526359a11ca60a299bef3aebf413fd77:

mfd: Add support for Intel Sunrisepoint LPSS devices (2015-07-28 09:56:47 +0100)

----------------------------------------------------------------
Immutable branch between MFD, Base, ACPI and DMA due for v4.3

----------------------------------------------------------------
Andy Shevchenko (5):
klist: implement klist_prev()
driver core: implement device_for_each_child_reverse()
mfd: make mfd_remove_devices() iterate in reverse order
dmaengine: add a driver for Intel integrated DMA 64-bit
mfd: Add support for Intel Sunrisepoint LPSS devices

Mika Westerberg (2):
PM / QoS: Make it possible to expose device latency tolerance to userspace ACPI / PM: Attach ACPI power domain only once

Rafael J. Wysocki (1):
Driver core: wakeup the parent device before trying probe

 drivers/acpi/device_pm.c      |   8 +
 drivers/acpi/internal.h       |   2 +
 drivers/acpi/scan.c           |  46 ++-
 drivers/base/core.c           |  43 +++
 drivers/base/dd.c             |  20 ++
 drivers/base/power/power.h    |   2 +
 drivers/base/power/qos.c      |  37 +++
 drivers/base/power/sysfs.c    |  11 +
 drivers/dma/Kconfig           |   8 +
 drivers/dma/Makefile          |   1 +
 drivers/dma/idma64.c          | 710 ++++++++++++++++++++++++++++++++++++++++++
 drivers/dma/idma64.h          | 233 ++++++++++++++
 drivers/mfd/Kconfig           |  23 ++
 drivers/mfd/Makefile          |   3 +
 drivers/mfd/intel-lpss-acpi.c |  84 +++++
 drivers/mfd/intel-lpss-pci.c  | 113 +++++++
 drivers/mfd/intel-lpss.c      | 524 +++++++++++++++++++++++++++++++
 drivers/mfd/intel-lpss.h      |  62 ++++
 drivers/mfd/mfd-core.c        |   2 +-
 include/linux/device.h        |   2 +
 include/linux/klist.h         |   1 +
 include/linux/pm_qos.h        |   5 +
 lib/klist.c                   |  41 +++
 23 files changed, 1964 insertions(+), 17 deletions(-)
 create mode 100644 drivers/dma/idma64.c
 create mode 100644 drivers/dma/idma64.h
 create mode 100644 drivers/mfd/intel-lpss-acpi.c
 create mode 100644 drivers/mfd/intel-lpss-pci.c
 create mode 100644 drivers/mfd/intel-lpss.c
 create mode 100644 drivers/mfd/intel-lpss.h
Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#32 lore raw

[GIT PULL] mfd: Immutable branch between MFD, Base, ACPI and DMA

With more explicit $SUBJECT line this time.

Sorry for the noise.

Quoted 63 lines

Enjoy,

The following changes since commit bc0195aad0daa2ad5b0d76cce22b167bc3435590:

Linux 4.2-rc2 (2015-07-12 15:10:30 -0700)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd.git tags/ib-mfd-base-acpi-dma-v4.3

for you to fetch changes up to 4b45efe8526359a11ca60a299bef3aebf413fd77:

mfd: Add support for Intel Sunrisepoint LPSS devices (2015-07-28 09:56:47 +0100)

----------------------------------------------------------------
Immutable branch between MFD, Base, ACPI and DMA due for v4.3

----------------------------------------------------------------
Andy Shevchenko (5):
klist: implement klist_prev()
driver core: implement device_for_each_child_reverse()
mfd: make mfd_remove_devices() iterate in reverse order
dmaengine: add a driver for Intel integrated DMA 64-bit
mfd: Add support for Intel Sunrisepoint LPSS devices

Mika Westerberg (2):
PM / QoS: Make it possible to expose device latency tolerance to userspace ACPI / PM: Attach ACPI power domain only once

Rafael J. Wysocki (1):
Driver core: wakeup the parent device before trying probe

 drivers/acpi/device_pm.c      |   8 +
 drivers/acpi/internal.h       |   2 +
 drivers/acpi/scan.c           |  46 ++-
 drivers/base/core.c           |  43 +++
 drivers/base/dd.c             |  20 ++
 drivers/base/power/power.h    |   2 +
 drivers/base/power/qos.c      |  37 +++
 drivers/base/power/sysfs.c    |  11 +
 drivers/dma/Kconfig           |   8 +
 drivers/dma/Makefile          |   1 +
 drivers/dma/idma64.c          | 710 ++++++++++++++++++++++++++++++++++++++++++
 drivers/dma/idma64.h          | 233 ++++++++++++++
 drivers/mfd/Kconfig           |  23 ++
 drivers/mfd/Makefile          |   3 +
 drivers/mfd/intel-lpss-acpi.c |  84 +++++
 drivers/mfd/intel-lpss-pci.c  | 113 +++++++
 drivers/mfd/intel-lpss.c      | 524 +++++++++++++++++++++++++++++++
 drivers/mfd/intel-lpss.h      |  62 ++++
 drivers/mfd/mfd-core.c        |   2 +-
 include/linux/device.h        |   2 +
 include/linux/klist.h         |   1 +
 include/linux/pm_qos.h        |   5 +
 lib/klist.c                   |  41 +++
 23 files changed, 1964 insertions(+), 17 deletions(-)
 create mode 100644 drivers/dma/idma64.c
 create mode 100644 drivers/dma/idma64.h
 create mode 100644 drivers/mfd/intel-lpss-acpi.c
 create mode 100644 drivers/mfd/intel-lpss-pci.c
 create mode 100644 drivers/mfd/intel-lpss.c
 create mode 100644 drivers/mfd/intel-lpss.h
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#33 lore raw
In reply to Lee Jones
Quoted 25 lines · Lee Jones

Enjoy,

The following changes since commit bc0195aad0daa2ad5b0d76cce22b167bc3435590:

Linux 4.2-rc2 (2015-07-12 15:10:30 -0700)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd.git tags/ib-mfd-base-acpi-dma-v4.3

for you to fetch changes up to 4b45efe8526359a11ca60a299bef3aebf413fd77:

mfd: Add support for Intel Sunrisepoint LPSS devices (2015-07-28 09:56:47 +0100)

----------------------------------------------------------------
Immutable branch between MFD, Base, ACPI and DMA due for v4.3

----------------------------------------------------------------
Andy Shevchenko (5):
klist: implement klist_prev()
driver core: implement device_for_each_child_reverse()
mfd: make mfd_remove_devices() iterate in reverse order
dmaengine: add a driver for Intel integrated DMA 64-bit
mfd: Add support for Intel Sunrisepoint LPSS devices

Quoted 6 lines

Mika Westerberg (2):
PM / QoS: Make it possible to expose device latency tolerance to userspace ACPI / PM: Attach ACPI power domain only once

Rafael J. Wysocki (1):
Driver core: wakeup the parent device before trying probe

Thank you :-)

#34 lore raw
In reply to Lee Jones
Quoted 8 lines · Lee Jones

Enjoy,

The following changes since commit bc0195aad0daa2ad5b0d76cce22b167bc3435590:

Linux 4.2-rc2 (2015-07-12 15:10:30 -0700)

are available in the git repository at:

Quoted 8 lines

git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd.git tags/ib-mfd-base-acpi-dma-v4.3

for you to fetch changes up to 4b45efe8526359a11ca60a299bef3aebf413fd77:

mfd: Add support for Intel Sunrisepoint LPSS devices (2015-07-28 09:56:47 +0100)

----------------------------------------------------------------
Immutable branch between MFD, Base, ACPI and DMA due for v4.3

Thanks, much appreciated!

Rafael