[patch 00/38] power management / hibernate support for s390

[patch 00/38] power management / hibernate support for s390

[Martin Schwidefsky]

Greetings,
this patch set implements hibernation support for s390. All s390
specific buses grow pm callbacks and all the s390 device drivers
are modified to make use of the callbacks. This allows to do

echo disk > /sys/power/state

to save the state of the running linux image to swap. A normal
ipl (alias boot) of the same kernel will detect the stored image
on swap and resume the save image. Works for guests under z/VM
and for LPARs.

The patches are meant to be applied on top of the s390 tree found
at
	git://git390.marist.edu/pub/scm/linux-2.6.git features

Have fun..
-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 01/38] pm: Move nvs routines into a seperate file.

[Martin Schwidefsky]

[-- Attachment #1: swsusp_split_out_nvs.patch --]
[-- Type: text/plain, Size: 10155 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

The *_nvs_* routines in swsusp.c make use of the io*map()
functions, which are only provided for HAS_IOMEM, thus
breaking compilation if HAS_IOMEM is not set. Fix this
by moving the *_nvs_* routines into nvs.c, which is only
compiled if HAS_IOMEM is set.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 include/linux/suspend.h |   18 ++++--
 kernel/power/Kconfig    |    4 +
 kernel/power/Makefile   |    1 
 kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
 kernel/power/swsusp.c   |  122 --------------------------------------------
 5 files changed, 147 insertions(+), 129 deletions(-)

Index: linux-2.6/kernel/power/nvs.c
===================================================================
--- /dev/null
+++ linux-2.6/kernel/power/nvs.c
@@ -0,0 +1,131 @@
+/*
+ * Routines for NVS memory handling
+ */
+
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/mm.h>
+#include <linux/suspend.h>
+
+/*
+ * Platforms, like ACPI, may want us to save some memory used by them during
+ * hibernation and to restore the contents of this memory during the subsequent
+ * resume.  The code below implements a mechanism allowing us to do that.
+ */
+
+struct nvs_page {
+	unsigned long phys_start;
+	unsigned int size;
+	void *kaddr;
+	void *data;
+	struct list_head node;
+};
+
+static LIST_HEAD(nvs_list);
+
+/**
+ *	hibernate_nvs_register - register platform NVS memory region to save
+ *	@start - physical address of the region
+ *	@size - size of the region
+ *
+ *	The NVS region need not be page-aligned (both ends) and we arrange
+ *	things so that the data from page-aligned addresses in this region will
+ *	be copied into separate RAM pages.
+ */
+int hibernate_nvs_register(unsigned long start, unsigned long size)
+{
+	struct nvs_page *entry, *next;
+
+	while (size > 0) {
+		unsigned int nr_bytes;
+
+		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
+		if (!entry)
+			goto Error;
+
+		list_add_tail(&entry->node, &nvs_list);
+		entry->phys_start = start;
+		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
+		entry->size = (size < nr_bytes) ? size : nr_bytes;
+
+		start += entry->size;
+		size -= entry->size;
+	}
+	return 0;
+
+ Error:
+	list_for_each_entry_safe(entry, next, &nvs_list, node) {
+		list_del(&entry->node);
+		kfree(entry);
+	}
+	return -ENOMEM;
+}
+
+/**
+ *	hibernate_nvs_free - free data pages allocated for saving NVS regions
+ */
+void hibernate_nvs_free(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			free_page((unsigned long)entry->data);
+			entry->data = NULL;
+			if (entry->kaddr) {
+				iounmap(entry->kaddr);
+				entry->kaddr = NULL;
+			}
+		}
+}
+
+/**
+ *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
+ */
+int hibernate_nvs_alloc(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node) {
+		entry->data = (void *)__get_free_page(GFP_KERNEL);
+		if (!entry->data) {
+			hibernate_nvs_free();
+			return -ENOMEM;
+		}
+	}
+	return 0;
+}
+
+/**
+ *	hibernate_nvs_save - save NVS memory regions
+ */
+void hibernate_nvs_save(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Saving platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			entry->kaddr = ioremap(entry->phys_start, entry->size);
+			memcpy(entry->data, entry->kaddr, entry->size);
+		}
+}
+
+/**
+ *	hibernate_nvs_restore - restore NVS memory regions
+ *
+ *	This function is going to be called with interrupts disabled, so it
+ *	cannot iounmap the virtual addresses used to access the NVS region.
+ */
+void hibernate_nvs_restore(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data)
+			memcpy(entry->kaddr, entry->data, entry->size);
+}
Index: linux-2.6/kernel/power/swsusp.c
===================================================================
--- linux-2.6.orig/kernel/power/swsusp.c
+++ linux-2.6/kernel/power/swsusp.c
@@ -262,125 +262,3 @@ int swsusp_shrink_memory(void)
 
 	return 0;
 }
-
-/*
- * Platforms, like ACPI, may want us to save some memory used by them during
- * hibernation and to restore the contents of this memory during the subsequent
- * resume.  The code below implements a mechanism allowing us to do that.
- */
-
-struct nvs_page {
-	unsigned long phys_start;
-	unsigned int size;
-	void *kaddr;
-	void *data;
-	struct list_head node;
-};
-
-static LIST_HEAD(nvs_list);
-
-/**
- *	hibernate_nvs_register - register platform NVS memory region to save
- *	@start - physical address of the region
- *	@size - size of the region
- *
- *	The NVS region need not be page-aligned (both ends) and we arrange
- *	things so that the data from page-aligned addresses in this region will
- *	be copied into separate RAM pages.
- */
-int hibernate_nvs_register(unsigned long start, unsigned long size)
-{
-	struct nvs_page *entry, *next;
-
-	while (size > 0) {
-		unsigned int nr_bytes;
-
-		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
-		if (!entry)
-			goto Error;
-
-		list_add_tail(&entry->node, &nvs_list);
-		entry->phys_start = start;
-		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
-		entry->size = (size < nr_bytes) ? size : nr_bytes;
-
-		start += entry->size;
-		size -= entry->size;
-	}
-	return 0;
-
- Error:
-	list_for_each_entry_safe(entry, next, &nvs_list, node) {
-		list_del(&entry->node);
-		kfree(entry);
-	}
-	return -ENOMEM;
-}
-
-/**
- *	hibernate_nvs_free - free data pages allocated for saving NVS regions
- */
-void hibernate_nvs_free(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			free_page((unsigned long)entry->data);
-			entry->data = NULL;
-			if (entry->kaddr) {
-				iounmap(entry->kaddr);
-				entry->kaddr = NULL;
-			}
-		}
-}
-
-/**
- *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
- */
-int hibernate_nvs_alloc(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node) {
-		entry->data = (void *)__get_free_page(GFP_KERNEL);
-		if (!entry->data) {
-			hibernate_nvs_free();
-			return -ENOMEM;
-		}
-	}
-	return 0;
-}
-
-/**
- *	hibernate_nvs_save - save NVS memory regions
- */
-void hibernate_nvs_save(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Saving platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			entry->kaddr = ioremap(entry->phys_start, entry->size);
-			memcpy(entry->data, entry->kaddr, entry->size);
-		}
-}
-
-/**
- *	hibernate_nvs_restore - restore NVS memory regions
- *
- *	This function is going to be called with interrupts disabled, so it
- *	cannot iounmap the virtual addresses used to access the NVS region.
- */
-void hibernate_nvs_restore(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data)
-			memcpy(entry->kaddr, entry->data, entry->size);
-}
Index: linux-2.6/kernel/power/Kconfig
===================================================================
--- linux-2.6.orig/kernel/power/Kconfig
+++ linux-2.6/kernel/power/Kconfig
@@ -116,9 +116,13 @@ config SUSPEND_FREEZER
 
 	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
 
+config NVS
+	bool
+
 config HIBERNATION
 	bool "Hibernation (aka 'suspend to disk')"
 	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
+	select NVS if HAS_IOMEM
 	---help---
 	  Enable the suspend to disk (STD) functionality, which is usually
 	  called "hibernation" in user interfaces.  STD checkpoints the
Index: linux-2.6/kernel/power/Makefile
===================================================================
--- linux-2.6.orig/kernel/power/Makefile
+++ linux-2.6/kernel/power/Makefile
@@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
 obj-$(CONFIG_PM_SLEEP)		+= console.o
 obj-$(CONFIG_FREEZER)		+= process.o
 obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
+obj-$(CONFIG_NVS)		+= nvs.o
 
 obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
Index: linux-2.6/include/linux/suspend.h
===================================================================
--- linux-2.6.orig/include/linux/suspend.h
+++ linux-2.6/include/linux/suspend.h
@@ -245,11 +245,6 @@ extern unsigned long get_safe_page(gfp_t
 
 extern void hibernation_set_ops(struct platform_hibernation_ops *ops);
 extern int hibernate(void);
-extern int hibernate_nvs_register(unsigned long start, unsigned long size);
-extern int hibernate_nvs_alloc(void);
-extern void hibernate_nvs_free(void);
-extern void hibernate_nvs_save(void);
-extern void hibernate_nvs_restore(void);
 extern bool system_entering_hibernation(void);
 #else /* CONFIG_HIBERNATION */
 static inline int swsusp_page_is_forbidden(struct page *p) { return 0; }
@@ -258,6 +253,16 @@ static inline void swsusp_unset_page_fre
 
 static inline void hibernation_set_ops(struct platform_hibernation_ops *ops) {}
 static inline int hibernate(void) { return -ENOSYS; }
+static inline bool system_entering_hibernation(void) { return false; }
+#endif /* CONFIG_HIBERNATION */
+
+#ifdef CONFIG_NVS
+extern int hibernate_nvs_register(unsigned long start, unsigned long size);
+extern int hibernate_nvs_alloc(void);
+extern void hibernate_nvs_free(void);
+extern void hibernate_nvs_save(void);
+extern void hibernate_nvs_restore(void);
+#else /* CONFIG_NVS */
 static inline int hibernate_nvs_register(unsigned long a, unsigned long b)
 {
 	return 0;
@@ -266,8 +271,7 @@ static inline int hibernate_nvs_alloc(vo
 static inline void hibernate_nvs_free(void) {}
 static inline void hibernate_nvs_save(void) {}
 static inline void hibernate_nvs_restore(void) {}
-static inline bool system_entering_hibernation(void) { return false; }
-#endif /* CONFIG_HIBERNATION */
+#endif /* CONFIG_NVS */
 
 #ifdef CONFIG_PM_SLEEP
 void save_processor_state(void);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 01/38] pm: Move nvs routines into a seperate file.

[Martin Schwidefsky]

[-- Attachment #1: swsusp_split_out_nvs.patch --]
[-- Type: text/plain, Size: 10154 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

The *_nvs_* routines in swsusp.c make use of the io*map()
functions, which are only provided for HAS_IOMEM, thus
breaking compilation if HAS_IOMEM is not set. Fix this
by moving the *_nvs_* routines into nvs.c, which is only
compiled if HAS_IOMEM is set.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 include/linux/suspend.h |   18 ++++--
 kernel/power/Kconfig    |    4 +
 kernel/power/Makefile   |    1 
 kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
 kernel/power/swsusp.c   |  122 --------------------------------------------
 5 files changed, 147 insertions(+), 129 deletions(-)

Index: linux-2.6/kernel/power/nvs.c
===================================================================
--- /dev/null
+++ linux-2.6/kernel/power/nvs.c
@@ -0,0 +1,131 @@
+/*
+ * Routines for NVS memory handling
+ */
+
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/mm.h>
+#include <linux/suspend.h>
+
+/*
+ * Platforms, like ACPI, may want us to save some memory used by them during
+ * hibernation and to restore the contents of this memory during the subsequent
+ * resume.  The code below implements a mechanism allowing us to do that.
+ */
+
+struct nvs_page {
+	unsigned long phys_start;
+	unsigned int size;
+	void *kaddr;
+	void *data;
+	struct list_head node;
+};
+
+static LIST_HEAD(nvs_list);
+
+/**
+ *	hibernate_nvs_register - register platform NVS memory region to save
+ *	@start - physical address of the region
+ *	@size - size of the region
+ *
+ *	The NVS region need not be page-aligned (both ends) and we arrange
+ *	things so that the data from page-aligned addresses in this region will
+ *	be copied into separate RAM pages.
+ */
+int hibernate_nvs_register(unsigned long start, unsigned long size)
+{
+	struct nvs_page *entry, *next;
+
+	while (size > 0) {
+		unsigned int nr_bytes;
+
+		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
+		if (!entry)
+			goto Error;
+
+		list_add_tail(&entry->node, &nvs_list);
+		entry->phys_start = start;
+		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
+		entry->size = (size < nr_bytes) ? size : nr_bytes;
+
+		start += entry->size;
+		size -= entry->size;
+	}
+	return 0;
+
+ Error:
+	list_for_each_entry_safe(entry, next, &nvs_list, node) {
+		list_del(&entry->node);
+		kfree(entry);
+	}
+	return -ENOMEM;
+}
+
+/**
+ *	hibernate_nvs_free - free data pages allocated for saving NVS regions
+ */
+void hibernate_nvs_free(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			free_page((unsigned long)entry->data);
+			entry->data = NULL;
+			if (entry->kaddr) {
+				iounmap(entry->kaddr);
+				entry->kaddr = NULL;
+			}
+		}
+}
+
+/**
+ *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
+ */
+int hibernate_nvs_alloc(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node) {
+		entry->data = (void *)__get_free_page(GFP_KERNEL);
+		if (!entry->data) {
+			hibernate_nvs_free();
+			return -ENOMEM;
+		}
+	}
+	return 0;
+}
+
+/**
+ *	hibernate_nvs_save - save NVS memory regions
+ */
+void hibernate_nvs_save(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Saving platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			entry->kaddr = ioremap(entry->phys_start, entry->size);
+			memcpy(entry->data, entry->kaddr, entry->size);
+		}
+}
+
+/**
+ *	hibernate_nvs_restore - restore NVS memory regions
+ *
+ *	This function is going to be called with interrupts disabled, so it
+ *	cannot iounmap the virtual addresses used to access the NVS region.
+ */
+void hibernate_nvs_restore(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data)
+			memcpy(entry->kaddr, entry->data, entry->size);
+}
Index: linux-2.6/kernel/power/swsusp.c
===================================================================
--- linux-2.6.orig/kernel/power/swsusp.c
+++ linux-2.6/kernel/power/swsusp.c
@@ -262,125 +262,3 @@ int swsusp_shrink_memory(void)
 
 	return 0;
 }
-
-/*
- * Platforms, like ACPI, may want us to save some memory used by them during
- * hibernation and to restore the contents of this memory during the subsequent
- * resume.  The code below implements a mechanism allowing us to do that.
- */
-
-struct nvs_page {
-	unsigned long phys_start;
-	unsigned int size;
-	void *kaddr;
-	void *data;
-	struct list_head node;
-};
-
-static LIST_HEAD(nvs_list);
-
-/**
- *	hibernate_nvs_register - register platform NVS memory region to save
- *	@start - physical address of the region
- *	@size - size of the region
- *
- *	The NVS region need not be page-aligned (both ends) and we arrange
- *	things so that the data from page-aligned addresses in this region will
- *	be copied into separate RAM pages.
- */
-int hibernate_nvs_register(unsigned long start, unsigned long size)
-{
-	struct nvs_page *entry, *next;
-
-	while (size > 0) {
-		unsigned int nr_bytes;
-
-		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
-		if (!entry)
-			goto Error;
-
-		list_add_tail(&entry->node, &nvs_list);
-		entry->phys_start = start;
-		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
-		entry->size = (size < nr_bytes) ? size : nr_bytes;
-
-		start += entry->size;
-		size -= entry->size;
-	}
-	return 0;
-
- Error:
-	list_for_each_entry_safe(entry, next, &nvs_list, node) {
-		list_del(&entry->node);
-		kfree(entry);
-	}
-	return -ENOMEM;
-}
-
-/**
- *	hibernate_nvs_free - free data pages allocated for saving NVS regions
- */
-void hibernate_nvs_free(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			free_page((unsigned long)entry->data);
-			entry->data = NULL;
-			if (entry->kaddr) {
-				iounmap(entry->kaddr);
-				entry->kaddr = NULL;
-			}
-		}
-}
-
-/**
- *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
- */
-int hibernate_nvs_alloc(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node) {
-		entry->data = (void *)__get_free_page(GFP_KERNEL);
-		if (!entry->data) {
-			hibernate_nvs_free();
-			return -ENOMEM;
-		}
-	}
-	return 0;
-}
-
-/**
- *	hibernate_nvs_save - save NVS memory regions
- */
-void hibernate_nvs_save(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Saving platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			entry->kaddr = ioremap(entry->phys_start, entry->size);
-			memcpy(entry->data, entry->kaddr, entry->size);
-		}
-}
-
-/**
- *	hibernate_nvs_restore - restore NVS memory regions
- *
- *	This function is going to be called with interrupts disabled, so it
- *	cannot iounmap the virtual addresses used to access the NVS region.
- */
-void hibernate_nvs_restore(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data)
-			memcpy(entry->kaddr, entry->data, entry->size);
-}
Index: linux-2.6/kernel/power/Kconfig
===================================================================
--- linux-2.6.orig/kernel/power/Kconfig
+++ linux-2.6/kernel/power/Kconfig
@@ -116,9 +116,13 @@ config SUSPEND_FREEZER
 
 	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
 
+config NVS
+	bool
+
 config HIBERNATION
 	bool "Hibernation (aka 'suspend to disk')"
 	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
+	select NVS if HAS_IOMEM
 	---help---
 	  Enable the suspend to disk (STD) functionality, which is usually
 	  called "hibernation" in user interfaces.  STD checkpoints the
Index: linux-2.6/kernel/power/Makefile
===================================================================
--- linux-2.6.orig/kernel/power/Makefile
+++ linux-2.6/kernel/power/Makefile
@@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
 obj-$(CONFIG_PM_SLEEP)		+= console.o
 obj-$(CONFIG_FREEZER)		+= process.o
 obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
+obj-$(CONFIG_NVS)		+= nvs.o
 
 obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
Index: linux-2.6/include/linux/suspend.h
===================================================================
--- linux-2.6.orig/include/linux/suspend.h
+++ linux-2.6/include/linux/suspend.h
@@ -245,11 +245,6 @@ extern unsigned long get_safe_page(gfp_t
 
 extern void hibernation_set_ops(struct platform_hibernation_ops *ops);
 extern int hibernate(void);
-extern int hibernate_nvs_register(unsigned long start, unsigned long size);
-extern int hibernate_nvs_alloc(void);
-extern void hibernate_nvs_free(void);
-extern void hibernate_nvs_save(void);
-extern void hibernate_nvs_restore(void);
 extern bool system_entering_hibernation(void);
 #else /* CONFIG_HIBERNATION */
 static inline int swsusp_page_is_forbidden(struct page *p) { return 0; }
@@ -258,6 +253,16 @@ static inline void swsusp_unset_page_fre
 
 static inline void hibernation_set_ops(struct platform_hibernation_ops *ops) {}
 static inline int hibernate(void) { return -ENOSYS; }
+static inline bool system_entering_hibernation(void) { return false; }
+#endif /* CONFIG_HIBERNATION */
+
+#ifdef CONFIG_NVS
+extern int hibernate_nvs_register(unsigned long start, unsigned long size);
+extern int hibernate_nvs_alloc(void);
+extern void hibernate_nvs_free(void);
+extern void hibernate_nvs_save(void);
+extern void hibernate_nvs_restore(void);
+#else /* CONFIG_NVS */
 static inline int hibernate_nvs_register(unsigned long a, unsigned long b)
 {
 	return 0;
@@ -266,8 +271,7 @@ static inline int hibernate_nvs_alloc(vo
 static inline void hibernate_nvs_free(void) {}
 static inline void hibernate_nvs_save(void) {}
 static inline void hibernate_nvs_restore(void) {}
-static inline bool system_entering_hibernation(void) { return false; }
-#endif /* CONFIG_HIBERNATION */
+#endif /* CONFIG_NVS */
 
 #ifdef CONFIG_PM_SLEEP
 void save_processor_state(void);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 02/38] dasd: forward internal errors to dasd_sleep_on caller

[Martin Schwidefsky]

[-- Attachment #1: dasd-forward-startio-errors.patch --]
[-- Type: text/plain, Size: 4898 bytes --]

From: Stefan Weinhuber <wein@de.ibm.com>

If a DASD requests is started with dasd_sleep_on and fails, then the
calling function may need to know the reason for the failure.
In cases of hardware errors it can inspect the sense data in the irb,
but when the reason is internal (e.g. start_IO failed) then it needs
a meaningfull return code.

Signed-off-by: Stefan Weinhuber <wein@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/dasd.c      |   30 ++++++++++++++++++++++++------
 drivers/s390/block/dasd_diag.c |    1 +
 drivers/s390/block/dasd_eckd.c |    8 +++++---
 drivers/s390/block/dasd_int.h  |    1 +
 4 files changed, 31 insertions(+), 9 deletions(-)

Index: linux-2.6/drivers/s390/block/dasd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd.c
+++ linux-2.6/drivers/s390/block/dasd.c
@@ -851,8 +851,10 @@ int dasd_start_IO(struct dasd_ccw_req *c
 
 	/* Check the cqr */
 	rc = dasd_check_cqr(cqr);
-	if (rc)
+	if (rc) {
+		cqr->intrc = rc;
 		return rc;
+	}
 	device = (struct dasd_device *) cqr->startdev;
 	if (cqr->retries < 0) {
 		/* internal error 14 - start_IO run out of retries */
@@ -915,6 +917,7 @@ int dasd_start_IO(struct dasd_ccw_req *c
 		BUG();
 		break;
 	}
+	cqr->intrc = rc;
 	return rc;
 }
 
@@ -1454,8 +1457,12 @@ int dasd_sleep_on(struct dasd_ccw_req *c
 	dasd_add_request_tail(cqr);
 	wait_event(generic_waitq, _wait_for_wakeup(cqr));
 
-	/* Request status is either done or failed. */
-	rc = (cqr->status == DASD_CQR_DONE) ? 0 : -EIO;
+	if (cqr->status == DASD_CQR_DONE)
+		rc = 0;
+	else if (cqr->intrc)
+		rc = cqr->intrc;
+	else
+		rc = -EIO;
 	return rc;
 }
 
@@ -1477,8 +1484,15 @@ int dasd_sleep_on_interruptible(struct d
 		dasd_cancel_req(cqr);
 		/* wait (non-interruptible) for final status */
 		wait_event(generic_waitq, _wait_for_wakeup(cqr));
+		cqr->intrc = rc;
 	}
-	rc = (cqr->status == DASD_CQR_DONE) ? 0 : -EIO;
+
+	if (cqr->status == DASD_CQR_DONE)
+		rc = 0;
+	else if (cqr->intrc)
+		rc = cqr->intrc;
+	else
+		rc = -EIO;
 	return rc;
 }
 
@@ -1523,8 +1537,12 @@ int dasd_sleep_on_immediatly(struct dasd
 
 	wait_event(generic_waitq, _wait_for_wakeup(cqr));
 
-	/* Request status is either done or failed. */
-	rc = (cqr->status == DASD_CQR_DONE) ? 0 : -EIO;
+	if (cqr->status == DASD_CQR_DONE)
+		rc = 0;
+	else if (cqr->intrc)
+		rc = cqr->intrc;
+	else
+		rc = -EIO;
 	return rc;
 }
 
Index: linux-2.6/drivers/s390/block/dasd_diag.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_diag.c
+++ linux-2.6/drivers/s390/block/dasd_diag.c
@@ -202,6 +202,7 @@ dasd_start_diag(struct dasd_ccw_req * cq
 		rc = -EIO;
 		break;
 	}
+	cqr->intrc = rc;
 	return rc;
 }
 
Index: linux-2.6/drivers/s390/block/dasd_int.h
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_int.h
+++ linux-2.6/drivers/s390/block/dasd_int.h
@@ -173,6 +173,7 @@ struct dasd_ccw_req {
 	void *data;			/* pointer to data area */
 
 	/* these are important for recovering erroneous requests          */
+	int intrc;			/* internal error, e.g. from start_IO */
 	struct irb irb;			/* device status in case of an error */
 	struct dasd_ccw_req *refers;	/* ERP-chain queueing. */
 	void *function; 		/* originating ERP action */
Index: linux-2.6/drivers/s390/block/dasd_eckd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_eckd.c
+++ linux-2.6/drivers/s390/block/dasd_eckd.c
@@ -3017,8 +3017,9 @@ static void dasd_eckd_dump_sense_ccw(str
 		      " I/O status report for device %s:\n",
 		      dev_name(&device->cdev->dev));
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
-		       " in req: %p CS: 0x%02X DS: 0x%02X\n", req,
-		       scsw_cstat(&irb->scsw), scsw_dstat(&irb->scsw));
+		       " in req: %p CS: 0x%02X DS: 0x%02X CC: 0x%02X RC: %d\n",
+		       req, scsw_cstat(&irb->scsw), scsw_dstat(&irb->scsw),
+		       scsw_cc(&irb->scsw), req->intrc);
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
 		       " device %s: Failing CCW: %p\n",
 		       dev_name(&device->cdev->dev),
@@ -3119,9 +3120,10 @@ static void dasd_eckd_dump_sense_tcw(str
 		      " I/O status report for device %s:\n",
 		      dev_name(&device->cdev->dev));
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
-		       " in req: %p CS: 0x%02X DS: 0x%02X "
+		       " in req: %p CS: 0x%02X DS: 0x%02X CC: 0x%02X RC: %d "
 		       "fcxs: 0x%02X schxs: 0x%02X\n", req,
 		       scsw_cstat(&irb->scsw), scsw_dstat(&irb->scsw),
+		       scsw_cc(&irb->scsw), req->intrc,
 		       irb->scsw.tm.fcxs, irb->scsw.tm.schxs);
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
 		       " device %s: Failing TCW: %p\n",

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 02/38] dasd: forward internal errors to dasd_sleep_on caller

[Martin Schwidefsky]

[-- Attachment #1: dasd-forward-startio-errors.patch --]
[-- Type: text/plain, Size: 4897 bytes --]

From: Stefan Weinhuber <wein@de.ibm.com>

If a DASD requests is started with dasd_sleep_on and fails, then the
calling function may need to know the reason for the failure.
In cases of hardware errors it can inspect the sense data in the irb,
but when the reason is internal (e.g. start_IO failed) then it needs
a meaningfull return code.

Signed-off-by: Stefan Weinhuber <wein@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/dasd.c      |   30 ++++++++++++++++++++++++------
 drivers/s390/block/dasd_diag.c |    1 +
 drivers/s390/block/dasd_eckd.c |    8 +++++---
 drivers/s390/block/dasd_int.h  |    1 +
 4 files changed, 31 insertions(+), 9 deletions(-)

Index: linux-2.6/drivers/s390/block/dasd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd.c
+++ linux-2.6/drivers/s390/block/dasd.c
@@ -851,8 +851,10 @@ int dasd_start_IO(struct dasd_ccw_req *c
 
 	/* Check the cqr */
 	rc = dasd_check_cqr(cqr);
-	if (rc)
+	if (rc) {
+		cqr->intrc = rc;
 		return rc;
+	}
 	device = (struct dasd_device *) cqr->startdev;
 	if (cqr->retries < 0) {
 		/* internal error 14 - start_IO run out of retries */
@@ -915,6 +917,7 @@ int dasd_start_IO(struct dasd_ccw_req *c
 		BUG();
 		break;
 	}
+	cqr->intrc = rc;
 	return rc;
 }
 
@@ -1454,8 +1457,12 @@ int dasd_sleep_on(struct dasd_ccw_req *c
 	dasd_add_request_tail(cqr);
 	wait_event(generic_waitq, _wait_for_wakeup(cqr));
 
-	/* Request status is either done or failed. */
-	rc = (cqr->status == DASD_CQR_DONE) ? 0 : -EIO;
+	if (cqr->status == DASD_CQR_DONE)
+		rc = 0;
+	else if (cqr->intrc)
+		rc = cqr->intrc;
+	else
+		rc = -EIO;
 	return rc;
 }
 
@@ -1477,8 +1484,15 @@ int dasd_sleep_on_interruptible(struct d
 		dasd_cancel_req(cqr);
 		/* wait (non-interruptible) for final status */
 		wait_event(generic_waitq, _wait_for_wakeup(cqr));
+		cqr->intrc = rc;
 	}
-	rc = (cqr->status == DASD_CQR_DONE) ? 0 : -EIO;
+
+	if (cqr->status == DASD_CQR_DONE)
+		rc = 0;
+	else if (cqr->intrc)
+		rc = cqr->intrc;
+	else
+		rc = -EIO;
 	return rc;
 }
 
@@ -1523,8 +1537,12 @@ int dasd_sleep_on_immediatly(struct dasd
 
 	wait_event(generic_waitq, _wait_for_wakeup(cqr));
 
-	/* Request status is either done or failed. */
-	rc = (cqr->status == DASD_CQR_DONE) ? 0 : -EIO;
+	if (cqr->status == DASD_CQR_DONE)
+		rc = 0;
+	else if (cqr->intrc)
+		rc = cqr->intrc;
+	else
+		rc = -EIO;
 	return rc;
 }
 
Index: linux-2.6/drivers/s390/block/dasd_diag.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_diag.c
+++ linux-2.6/drivers/s390/block/dasd_diag.c
@@ -202,6 +202,7 @@ dasd_start_diag(struct dasd_ccw_req * cq
 		rc = -EIO;
 		break;
 	}
+	cqr->intrc = rc;
 	return rc;
 }
 
Index: linux-2.6/drivers/s390/block/dasd_int.h
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_int.h
+++ linux-2.6/drivers/s390/block/dasd_int.h
@@ -173,6 +173,7 @@ struct dasd_ccw_req {
 	void *data;			/* pointer to data area */
 
 	/* these are important for recovering erroneous requests          */
+	int intrc;			/* internal error, e.g. from start_IO */
 	struct irb irb;			/* device status in case of an error */
 	struct dasd_ccw_req *refers;	/* ERP-chain queueing. */
 	void *function; 		/* originating ERP action */
Index: linux-2.6/drivers/s390/block/dasd_eckd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_eckd.c
+++ linux-2.6/drivers/s390/block/dasd_eckd.c
@@ -3017,8 +3017,9 @@ static void dasd_eckd_dump_sense_ccw(str
 		      " I/O status report for device %s:\n",
 		      dev_name(&device->cdev->dev));
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
-		       " in req: %p CS: 0x%02X DS: 0x%02X\n", req,
-		       scsw_cstat(&irb->scsw), scsw_dstat(&irb->scsw));
+		       " in req: %p CS: 0x%02X DS: 0x%02X CC: 0x%02X RC: %d\n",
+		       req, scsw_cstat(&irb->scsw), scsw_dstat(&irb->scsw),
+		       scsw_cc(&irb->scsw), req->intrc);
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
 		       " device %s: Failing CCW: %p\n",
 		       dev_name(&device->cdev->dev),
@@ -3119,9 +3120,10 @@ static void dasd_eckd_dump_sense_tcw(str
 		      " I/O status report for device %s:\n",
 		      dev_name(&device->cdev->dev));
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
-		       " in req: %p CS: 0x%02X DS: 0x%02X "
+		       " in req: %p CS: 0x%02X DS: 0x%02X CC: 0x%02X RC: %d "
 		       "fcxs: 0x%02X schxs: 0x%02X\n", req,
 		       scsw_cstat(&irb->scsw), scsw_dstat(&irb->scsw),
+		       scsw_cc(&irb->scsw), req->intrc,
 		       irb->scsw.tm.fcxs, irb->scsw.tm.schxs);
 	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
 		       " device %s: Failing TCW: %p\n",

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 03/38] iucv: provide second per-cpu IUCV command parameter block

[Martin Schwidefsky]

[-- Attachment #1: iucv-2ndparm.patch --]
[-- Type: text/plain, Size: 4830 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Some of the IUCV commands can be invoked in interrupt context.
Those commands need a different per-cpu IUCV command parameter block,
otherwise they might overwrite an IUCV command parameter of a not yet
finished IUCV command invocation in process context.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 net/iucv/iucv.c |   41 +++++++++++++++++++++++++++++++----------
 1 file changed, 31 insertions(+), 10 deletions(-)

Index: linux-2.6/net/iucv/iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/iucv.c
+++ linux-2.6/net/iucv/iucv.c
@@ -280,6 +280,7 @@ union iucv_param {
  * Anchor for per-cpu IUCV command parameter block.
  */
 static union iucv_param *iucv_param[NR_CPUS];
+static union iucv_param *iucv_param_irq[NR_CPUS];
 
 /**
  * iucv_call_b2f0
@@ -358,7 +359,7 @@ static void iucv_allow_cpu(void *data)
 	 *	0x10 - Flag to allow priority message completion interrupts
 	 *	0x08 - Flag to allow IUCV control interrupts
 	 */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->set_mask.ipmask = 0xf8;
 	iucv_call_b2f0(IUCV_SETMASK, parm);
@@ -379,7 +380,7 @@ static void iucv_block_cpu(void *data)
 	union iucv_param *parm;
 
 	/* Disable all iucv interrupts. */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	memset(parm, 0, sizeof(union iucv_param));
 	iucv_call_b2f0(IUCV_SETMASK, parm);
 
@@ -403,7 +404,7 @@ static void iucv_declare_cpu(void *data)
 		return;
 
 	/* Declare interrupt buffer. */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->db.ipbfadr1 = virt_to_phys(iucv_irq_data[cpu]);
 	rc = iucv_call_b2f0(IUCV_DECLARE_BUFFER, parm);
@@ -460,7 +461,7 @@ static void iucv_retrieve_cpu(void *data
 	iucv_block_cpu(NULL);
 
 	/* Retrieve interrupt buffer. */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	iucv_call_b2f0(IUCV_RETRIEVE_BUFFER, parm);
 
 	/* Clear indication that an iucv buffer exists for this cpu. */
@@ -574,11 +575,22 @@ static int __cpuinit iucv_cpu_notify(str
 			iucv_irq_data[cpu] = NULL;
 			return NOTIFY_BAD;
 		}
+		iucv_param_irq[cpu] = kmalloc_node(sizeof(union iucv_param),
+					GFP_KERNEL|GFP_DMA, cpu_to_node(cpu));
+		if (!iucv_param_irq[cpu]) {
+			kfree(iucv_param[cpu]);
+			iucv_param[cpu] = NULL;
+			kfree(iucv_irq_data[cpu]);
+			iucv_irq_data[cpu] = NULL;
+			return NOTIFY_BAD;
+		}
 		break;
 	case CPU_UP_CANCELED:
 	case CPU_UP_CANCELED_FROZEN:
 	case CPU_DEAD:
 	case CPU_DEAD_FROZEN:
+		kfree(iucv_param_irq[cpu]);
+		iucv_param_irq[cpu] = NULL;
 		kfree(iucv_param[cpu]);
 		iucv_param[cpu] = NULL;
 		kfree(iucv_irq_data[cpu]);
@@ -625,7 +637,7 @@ static int iucv_sever_pathid(u16 pathid,
 {
 	union iucv_param *parm;
 
-	parm = iucv_param[smp_processor_id()];
+	parm = iucv_param_irq[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (userdata)
 		memcpy(parm->ctrl.ipuser, userdata, sizeof(parm->ctrl.ipuser));
@@ -918,10 +930,8 @@ int iucv_path_sever(struct iucv_path *pa
 	if (iucv_active_cpu != smp_processor_id())
 		spin_lock_bh(&iucv_table_lock);
 	rc = iucv_sever_pathid(path->pathid, userdata);
-	if (!rc) {
-		iucv_path_table[path->pathid] = NULL;
-		list_del_init(&path->list);
-	}
+	iucv_path_table[path->pathid] = NULL;
+	list_del_init(&path->list);
 	if (iucv_active_cpu != smp_processor_id())
 		spin_unlock_bh(&iucv_table_lock);
 	preempt_enable();
@@ -1413,7 +1423,7 @@ static void iucv_path_severed(struct iuc
 	else {
 		iucv_sever_pathid(path->pathid, NULL);
 		iucv_path_table[path->pathid] = NULL;
-		list_del_init(&path->list);
+		list_del(&path->list);
 		iucv_path_free(path);
 	}
 }
@@ -1717,6 +1727,13 @@ static int __init iucv_init(void)
 			rc = -ENOMEM;
 			goto out_free;
 		}
+		iucv_param_irq[cpu] = kmalloc_node(sizeof(union iucv_param),
+				  GFP_KERNEL|GFP_DMA, cpu_to_node(cpu));
+		if (!iucv_param_irq[cpu]) {
+			rc = -ENOMEM;
+			goto out_free;
+		}
+
 	}
 	rc = register_hotcpu_notifier(&iucv_cpu_notifier);
 	if (rc)
@@ -1734,6 +1751,8 @@ out_cpu:
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 out_free:
 	for_each_possible_cpu(cpu) {
+		kfree(iucv_param_irq[cpu]);
+		iucv_param_irq[cpu] = NULL;
 		kfree(iucv_param[cpu]);
 		iucv_param[cpu] = NULL;
 		kfree(iucv_irq_data[cpu]);
@@ -1764,6 +1783,8 @@ static void __exit iucv_exit(void)
 	spin_unlock_irq(&iucv_queue_lock);
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 	for_each_possible_cpu(cpu) {
+		kfree(iucv_param_irq[cpu]);
+		iucv_param_irq[cpu] = NULL;
 		kfree(iucv_param[cpu]);
 		iucv_param[cpu] = NULL;
 		kfree(iucv_irq_data[cpu]);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 03/38] iucv: provide second per-cpu IUCV command parameter block

[Martin Schwidefsky]

[-- Attachment #1: iucv-2ndparm.patch --]
[-- Type: text/plain, Size: 4829 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Some of the IUCV commands can be invoked in interrupt context.
Those commands need a different per-cpu IUCV command parameter block,
otherwise they might overwrite an IUCV command parameter of a not yet
finished IUCV command invocation in process context.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 net/iucv/iucv.c |   41 +++++++++++++++++++++++++++++++----------
 1 file changed, 31 insertions(+), 10 deletions(-)

Index: linux-2.6/net/iucv/iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/iucv.c
+++ linux-2.6/net/iucv/iucv.c
@@ -280,6 +280,7 @@ union iucv_param {
  * Anchor for per-cpu IUCV command parameter block.
  */
 static union iucv_param *iucv_param[NR_CPUS];
+static union iucv_param *iucv_param_irq[NR_CPUS];
 
 /**
  * iucv_call_b2f0
@@ -358,7 +359,7 @@ static void iucv_allow_cpu(void *data)
 	 *	0x10 - Flag to allow priority message completion interrupts
 	 *	0x08 - Flag to allow IUCV control interrupts
 	 */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->set_mask.ipmask = 0xf8;
 	iucv_call_b2f0(IUCV_SETMASK, parm);
@@ -379,7 +380,7 @@ static void iucv_block_cpu(void *data)
 	union iucv_param *parm;
 
 	/* Disable all iucv interrupts. */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	memset(parm, 0, sizeof(union iucv_param));
 	iucv_call_b2f0(IUCV_SETMASK, parm);
 
@@ -403,7 +404,7 @@ static void iucv_declare_cpu(void *data)
 		return;
 
 	/* Declare interrupt buffer. */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->db.ipbfadr1 = virt_to_phys(iucv_irq_data[cpu]);
 	rc = iucv_call_b2f0(IUCV_DECLARE_BUFFER, parm);
@@ -460,7 +461,7 @@ static void iucv_retrieve_cpu(void *data
 	iucv_block_cpu(NULL);
 
 	/* Retrieve interrupt buffer. */
-	parm = iucv_param[cpu];
+	parm = iucv_param_irq[cpu];
 	iucv_call_b2f0(IUCV_RETRIEVE_BUFFER, parm);
 
 	/* Clear indication that an iucv buffer exists for this cpu. */
@@ -574,11 +575,22 @@ static int __cpuinit iucv_cpu_notify(str
 			iucv_irq_data[cpu] = NULL;
 			return NOTIFY_BAD;
 		}
+		iucv_param_irq[cpu] = kmalloc_node(sizeof(union iucv_param),
+					GFP_KERNEL|GFP_DMA, cpu_to_node(cpu));
+		if (!iucv_param_irq[cpu]) {
+			kfree(iucv_param[cpu]);
+			iucv_param[cpu] = NULL;
+			kfree(iucv_irq_data[cpu]);
+			iucv_irq_data[cpu] = NULL;
+			return NOTIFY_BAD;
+		}
 		break;
 	case CPU_UP_CANCELED:
 	case CPU_UP_CANCELED_FROZEN:
 	case CPU_DEAD:
 	case CPU_DEAD_FROZEN:
+		kfree(iucv_param_irq[cpu]);
+		iucv_param_irq[cpu] = NULL;
 		kfree(iucv_param[cpu]);
 		iucv_param[cpu] = NULL;
 		kfree(iucv_irq_data[cpu]);
@@ -625,7 +637,7 @@ static int iucv_sever_pathid(u16 pathid,
 {
 	union iucv_param *parm;
 
-	parm = iucv_param[smp_processor_id()];
+	parm = iucv_param_irq[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (userdata)
 		memcpy(parm->ctrl.ipuser, userdata, sizeof(parm->ctrl.ipuser));
@@ -918,10 +930,8 @@ int iucv_path_sever(struct iucv_path *pa
 	if (iucv_active_cpu != smp_processor_id())
 		spin_lock_bh(&iucv_table_lock);
 	rc = iucv_sever_pathid(path->pathid, userdata);
-	if (!rc) {
-		iucv_path_table[path->pathid] = NULL;
-		list_del_init(&path->list);
-	}
+	iucv_path_table[path->pathid] = NULL;
+	list_del_init(&path->list);
 	if (iucv_active_cpu != smp_processor_id())
 		spin_unlock_bh(&iucv_table_lock);
 	preempt_enable();
@@ -1413,7 +1423,7 @@ static void iucv_path_severed(struct iuc
 	else {
 		iucv_sever_pathid(path->pathid, NULL);
 		iucv_path_table[path->pathid] = NULL;
-		list_del_init(&path->list);
+		list_del(&path->list);
 		iucv_path_free(path);
 	}
 }
@@ -1717,6 +1727,13 @@ static int __init iucv_init(void)
 			rc = -ENOMEM;
 			goto out_free;
 		}
+		iucv_param_irq[cpu] = kmalloc_node(sizeof(union iucv_param),
+				  GFP_KERNEL|GFP_DMA, cpu_to_node(cpu));
+		if (!iucv_param_irq[cpu]) {
+			rc = -ENOMEM;
+			goto out_free;
+		}
+
 	}
 	rc = register_hotcpu_notifier(&iucv_cpu_notifier);
 	if (rc)
@@ -1734,6 +1751,8 @@ out_cpu:
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 out_free:
 	for_each_possible_cpu(cpu) {
+		kfree(iucv_param_irq[cpu]);
+		iucv_param_irq[cpu] = NULL;
 		kfree(iucv_param[cpu]);
 		iucv_param[cpu] = NULL;
 		kfree(iucv_irq_data[cpu]);
@@ -1764,6 +1783,8 @@ static void __exit iucv_exit(void)
 	spin_unlock_irq(&iucv_queue_lock);
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 	for_each_possible_cpu(cpu) {
+		kfree(iucv_param_irq[cpu]);
+		iucv_param_irq[cpu] = NULL;
 		kfree(iucv_param[cpu]);
 		iucv_param[cpu] = NULL;
 		kfree(iucv_irq_data[cpu]);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 04/38] device irq power management

[Martin Schwidefsky]

[-- Attachment #1: pm_genirq_buildfix.patch --]
[-- Type: text/plain, Size: 1338 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

On architectures without GENERIC_HARDIRQS there are no device irqs
to suspend or resume. Add an #ifdef to replace suspend_device_irqs()
and resume_device_irqs() with an empty inline function in that case.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 include/linux/interrupt.h |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

Index: linux-2.6/include/linux/interrupt.h
===================================================================
--- linux-2.6.orig/include/linux/interrupt.h
+++ linux-2.6/include/linux/interrupt.h
@@ -183,9 +183,15 @@ extern void disable_irq(unsigned int irq
 extern void enable_irq(unsigned int irq);
 
 /* The following three functions are for the core kernel use only. */
+#ifdef CONFIG_GENERIC_HARDIRQS
 extern void suspend_device_irqs(void);
 extern void resume_device_irqs(void);
-#ifdef CONFIG_PM_SLEEP
+#else
+static inline void suspend_device_irqs(void) { }
+static inline void resume_device_irqs(void) { }
+#endif
+
+#if defined(CONFIG_PM_SLEEP) && defined(CONFIG_GENERIC_HARDIRQS)
 extern int check_wakeup_irqs(void);
 #else
 static inline int check_wakeup_irqs(void) { return 0; }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 04/38] device irq power management

[Martin Schwidefsky]

[-- Attachment #1: pm_genirq_buildfix.patch --]
[-- Type: text/plain, Size: 1337 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

On architectures without GENERIC_HARDIRQS there are no device irqs
to suspend or resume. Add an #ifdef to replace suspend_device_irqs()
and resume_device_irqs() with an empty inline function in that case.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 include/linux/interrupt.h |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

Index: linux-2.6/include/linux/interrupt.h
===================================================================
--- linux-2.6.orig/include/linux/interrupt.h
+++ linux-2.6/include/linux/interrupt.h
@@ -183,9 +183,15 @@ extern void disable_irq(unsigned int irq
 extern void enable_irq(unsigned int irq);
 
 /* The following three functions are for the core kernel use only. */
+#ifdef CONFIG_GENERIC_HARDIRQS
 extern void suspend_device_irqs(void);
 extern void resume_device_irqs(void);
-#ifdef CONFIG_PM_SLEEP
+#else
+static inline void suspend_device_irqs(void) { }
+static inline void resume_device_irqs(void) { }
+#endif
+
+#if defined(CONFIG_PM_SLEEP) && defined(CONFIG_GENERIC_HARDIRQS)
 extern int check_wakeup_irqs(void);
 #else
 static inline int check_wakeup_irqs(void) { return 0; }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 05/38] s390: hibernation support for s390

[Martin Schwidefsky]

[-- Attachment #1: pm_backend.patch --]
[-- Type: text/plain, Size: 15351 bytes --]

From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>

This patch introduces the hibernation backend support to the 
s390 architecture. Now it is possible to suspend a mainframe Linux 
guest using the following command:

echo disk > /sys/power/state

Signed-off-by: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/Kconfig               |    9 +
 arch/s390/Makefile              |    4 
 arch/s390/include/asm/suspend.h |    5 +
 arch/s390/include/asm/system.h  |   22 +++-
 arch/s390/kernel/early.c        |    6 -
 arch/s390/kernel/mem_detect.c   |   19 ---
 arch/s390/kernel/smp.c          |   38 +++++++
 arch/s390/power/Makefile        |    8 +
 arch/s390/power/suspend.c       |   40 ++++++++
 arch/s390/power/swsusp.c        |   31 ++++++
 arch/s390/power/swsusp_64.c     |   17 +++
 arch/s390/power/swsusp_asm64.S  |  199 ++++++++++++++++++++++++++++++++++++++++
 12 files changed, 371 insertions(+), 27 deletions(-)

Index: linux-2.6/arch/s390/include/asm/suspend.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/suspend.h
+++ linux-2.6/arch/s390/include/asm/suspend.h
@@ -1,5 +1,10 @@
 #ifndef __ASM_S390_SUSPEND_H
 #define __ASM_S390_SUSPEND_H
 
+static inline int arch_prepare_suspend(void)
+{
+	return 0;
+}
+
 #endif
 
Index: linux-2.6/arch/s390/Kconfig
===================================================================
--- linux-2.6.orig/arch/s390/Kconfig
+++ linux-2.6/arch/s390/Kconfig
@@ -343,6 +343,9 @@ config ARCH_ENABLE_MEMORY_HOTPLUG
 config ARCH_ENABLE_MEMORY_HOTREMOVE
 	def_bool y
 
+config ARCH_HIBERNATION_POSSIBLE
+       def_bool y
+
 source "mm/Kconfig"
 
 comment "I/O subsystem configuration"
@@ -587,6 +590,12 @@ config SECCOMP
 
 endmenu
 
+menu "Power Management"
+
+source "kernel/power/Kconfig"
+
+endmenu
+
 source "net/Kconfig"
 
 config PCMCIA
Index: linux-2.6/arch/s390/Makefile
===================================================================
--- linux-2.6.orig/arch/s390/Makefile
+++ linux-2.6/arch/s390/Makefile
@@ -88,7 +88,9 @@ LDFLAGS_vmlinux := -e start
 head-y		:= arch/s390/kernel/head.o arch/s390/kernel/init_task.o
 
 core-y		+= arch/s390/mm/ arch/s390/kernel/ arch/s390/crypto/ \
-		   arch/s390/appldata/ arch/s390/hypfs/ arch/s390/kvm/
+		   arch/s390/appldata/ arch/s390/hypfs/ arch/s390/kvm/ \
+		   arch/s390/power/
+
 libs-y		+= arch/s390/lib/
 drivers-y	+= drivers/s390/
 drivers-$(CONFIG_MATHEMU) += arch/s390/math-emu/
Index: linux-2.6/arch/s390/power/Makefile
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/Makefile
@@ -0,0 +1,8 @@
+#
+# Makefile for s390 PM support
+#
+
+obj-$(CONFIG_HIBERNATION) += suspend.o
+obj-$(CONFIG_HIBERNATION) += swsusp.o
+obj-$(CONFIG_HIBERNATION) += swsusp_64.o
+obj-$(CONFIG_HIBERNATION) += swsusp_asm64.o
Index: linux-2.6/arch/s390/power/suspend.c
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/suspend.c
@@ -0,0 +1,40 @@
+/*
+ * Suspend support specific for s390.
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ */
+
+#include <linux/mm.h>
+#include <linux/suspend.h>
+#include <linux/reboot.h>
+#include <linux/pfn.h>
+#include <asm/sections.h>
+#include <asm/ipl.h>
+
+/*
+ * References to section boundaries
+ */
+extern const void __nosave_begin, __nosave_end;
+
+/*
+ *  check if given pfn is in the 'nosave' or in the read only NSS section
+ */
+int pfn_is_nosave(unsigned long pfn)
+{
+	unsigned long nosave_begin_pfn = __pa(&__nosave_begin) >> PAGE_SHIFT;
+	unsigned long nosave_end_pfn = PAGE_ALIGN(__pa(&__nosave_end))
+					>> PAGE_SHIFT;
+	unsigned long eshared_pfn = PFN_DOWN(__pa(&_eshared)) - 1;
+	unsigned long stext_pfn = PFN_DOWN(__pa(&_stext));
+
+	if (pfn >= nosave_begin_pfn && pfn < nosave_end_pfn)
+		return 1;
+	if (pfn >= stext_pfn && pfn <= eshared_pfn) {
+		if (ipl_info.type == IPL_TYPE_NSS)
+			return 1;
+	} else if ((tprot(pfn * PAGE_SIZE) && pfn > 0))
+		return 1;
+	return 0;
+}
Index: linux-2.6/arch/s390/power/swsusp_64.c
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/swsusp_64.c
@@ -0,0 +1,17 @@
+/*
+ * Support for suspend and resume on s390
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ *
+ */
+
+#include <asm/system.h>
+#include <linux/interrupt.h>
+
+void do_after_copyback(void)
+{
+	mb();
+}
+
Index: linux-2.6/arch/s390/power/swsusp_asm64.S
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/swsusp_asm64.S
@@ -0,0 +1,199 @@
+/*
+ * S390 64-bit swsusp implementation
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ *            Michael Holzheu <holzheu@linux.vnet.ibm.com>
+ */
+
+#include <asm/page.h>
+#include <asm/ptrace.h>
+#include <asm/asm-offsets.h>
+
+/*
+ * Save register context in absolute 0 lowcore and call swsusp_save() to
+ * create in-memory kernel image. The context is saved in the designated
+ * "store status" memory locations (see POP).
+ * We return from this function twice. The first time during the suspend to
+ * disk process. The second time via the swsusp_arch_resume() function
+ * (see below) in the resume process.
+ * This function runs with disabled interrupts.
+ */
+	.section .text
+	.align  2
+	.globl swsusp_arch_suspend
+swsusp_arch_suspend:
+	stmg	%r6,%r15,__SF_GPRS(%r15)
+	lgr	%r1,%r15
+	aghi	%r15,-STACK_FRAME_OVERHEAD
+	stg	%r1,__SF_BACKCHAIN(%r15)
+
+	/* Deactivate DAT */
+ 	stnsm	__SF_EMPTY(%r15),0xfb
+
+	/* Switch off lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	ni	__SF_EMPTY+4(%r15),0xef
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Store prefix register on stack */
+	stpx	__SF_EMPTY(%r15)
+
+	/* Setup base register for lowcore (absolute 0) */
+	llgf	%r1,__SF_EMPTY(%r15)
+
+	/* Get pointer to save area */
+	aghi	%r1,0x1000
+
+	/* Store registers */
+	mvc	0x318(4,%r1),__SF_EMPTY(%r15)	/* move prefix to lowcore */
+	stfpc	0x31c(%r1)			/* store fpu control */
+	std	0,0x200(%r1)			/* store f0 */
+	std	1,0x208(%r1)			/* store f1 */
+	std	2,0x210(%r1)			/* store f2 */
+	std	3,0x218(%r1)			/* store f3 */
+	std	4,0x220(%r1)			/* store f4 */
+	std	5,0x228(%r1)			/* store f5 */
+	std	6,0x230(%r1)			/* store f6 */
+	std	7,0x238(%r1)			/* store f7 */
+	std	8,0x240(%r1)			/* store f8 */
+	std	9,0x248(%r1)			/* store f9 */
+	std	10,0x250(%r1)			/* store f10 */
+	std	11,0x258(%r1)			/* store f11 */
+	std	12,0x260(%r1)			/* store f12 */
+	std	13,0x268(%r1)			/* store f13 */
+	std	14,0x270(%r1)			/* store f14 */
+	std	15,0x278(%r1)			/* store f15 */
+	stam	%a0,%a15,0x340(%r1)		/* store access registers */
+	stctg	%c0,%c15,0x380(%r1)		/* store control registers */
+	stmg	%r0,%r15,0x280(%r1)		/* store general registers */
+
+	stpt	0x328(%r1)			/* store timer */
+	stckc	0x330(%r1)			/* store clock comparator */
+
+	/* Activate DAT */
+	stosm	__SF_EMPTY(%r15),0x04
+
+	/* Set prefix page to zero */
+	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
+	spx	__SF_EMPTY(%r15)
+
+	/* Setup lowcore */
+	brasl	%r14,setup_lowcore_early
+
+	/* Save image */
+	brasl	%r14,swsusp_save
+
+	/* Switch on lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	oi	__SF_EMPTY+4(%r15),0x10
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Restore prefix register and return */
+	lghi	%r1,0x1000
+	spx	0x318(%r1)
+	lmg	%r6,%r15,STACK_FRAME_OVERHEAD + __SF_GPRS(%r15)
+	lghi	%r2,0
+	br	%r14
+
+/*
+ * Restore saved memory image to correct place and restore register context.
+ * Then we return to the function that called swsusp_arch_suspend().
+ * swsusp_arch_resume() runs with disabled interrupts.
+ */
+	.globl swsusp_arch_resume
+swsusp_arch_resume:
+	stmg	%r6,%r15,__SF_GPRS(%r15)
+	lgr	%r1,%r15
+	aghi	%r15,-STACK_FRAME_OVERHEAD
+	stg	%r1,__SF_BACKCHAIN(%r15)
+
+	/* Save boot cpu number */
+	brasl	%r14,smp_get_phys_cpu_id
+	lgr	%r10,%r2
+
+	/* Deactivate DAT */
+ 	stnsm	__SF_EMPTY(%r15),0xfb
+
+	/* Switch off lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	ni	__SF_EMPTY+4(%r15),0xef
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Set prefix page to zero */
+	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
+	spx	__SF_EMPTY(%r15)
+
+	/* Restore saved image */
+	larl	%r1,restore_pblist
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	2f
+0:
+	lg	%r2,8(%r1)
+	lg	%r4,0(%r1)
+	lghi	%r3,PAGE_SIZE
+	lghi	%r5,PAGE_SIZE
+1:
+	mvcle	%r2,%r4,0
+	jo	1b
+	lg	%r1,16(%r1)
+	ltgr	%r1,%r1
+	jnz	0b
+2:
+	ptlb				/* flush tlb */
+
+	/* Restore registers */
+	lghi	%r13,0x1000		/* %r1 = pointer to save arae */
+
+	spt	0x328(%r13)		/* reprogram timer */
+	//sckc	0x330(%r13)		/* set clock comparator */
+
+	lctlg	%c0,%c15,0x380(%r13)	/* load control registers */
+	lam	%a0,%a15,0x340(%r13)	/* load access registers */
+
+	lfpc	0x31c(%r13)		/* load fpu control */
+	ld	0,0x200(%r13)		/* load f0 */
+	ld	1,0x208(%r13)		/* load f1 */
+	ld	2,0x210(%r13)		/* load f2 */
+	ld	3,0x218(%r13)		/* load f3 */
+	ld	4,0x220(%r13)		/* load f4 */
+	ld	5,0x228(%r13)		/* load f5 */
+	ld	6,0x230(%r13)		/* load f6 */
+	ld	7,0x238(%r13)		/* load f7 */
+	ld	8,0x240(%r13)		/* load f8 */
+	ld	9,0x248(%r13)		/* load f9 */
+	ld	10,0x250(%r13)		/* load f10 */
+	ld	11,0x258(%r13)		/* load f11 */
+	ld	12,0x260(%r13)		/* load f12 */
+	ld	13,0x268(%r13)		/* load f13 */
+	ld	14,0x270(%r13)		/* load f14 */
+	ld	15,0x278(%r13)		/* load f15 */
+
+	/* Load old stack */
+	lg	%r15,0x2f8(%r13)
+
+	/* Pointer to save arae */
+	lghi	%r13,0x1000
+
+	/* Switch CPUs */
+	lgr	%r2,%r10		/* get cpu id */
+	llgf	%r3,0x318(%r13)
+	brasl	%r14,smp_switch_boot_cpu_in_resume
+
+	/* Restore prefix register */
+	spx	0x318(%r13)
+
+	/* Switch on lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	oi	__SF_EMPTY+4(%r15),0x10
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Activate DAT */
+	stosm	__SF_EMPTY(%r15),0x04
+
+	/* Return 0 */
+	lmg	%r6,%r15,STACK_FRAME_OVERHEAD + __SF_GPRS(%r15)
+	lghi	%r2,0
+	br	%r14
Index: linux-2.6/arch/s390/power/swsusp.c
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/swsusp.c
@@ -0,0 +1,31 @@
+/*
+ * Support for suspend and resume on s390
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ *
+ */
+
+#include <linux/suspend.h>
+#include <linux/gfp.h>
+#include <linux/cpu.h>
+#include <linux/smp.h>
+#include <linux/cpumask.h>
+#include <asm/ipl.h>
+#include <asm/processor.h>
+
+/*
+ * save CPU registers before creating a hibernation image and before
+ * restoring the memory state from it
+ */
+void save_processor_state(void)
+{
+}
+
+/*
+ * restore the contents of CPU registers
+ */
+void restore_processor_state(void)
+{
+}
Index: linux-2.6/arch/s390/kernel/early.c
===================================================================
--- linux-2.6.orig/arch/s390/kernel/early.c
+++ linux-2.6/arch/s390/kernel/early.c
@@ -1,7 +1,7 @@
 /*
  *  arch/s390/kernel/early.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Hongjie Yang <hongjie@us.ibm.com>,
  *		 Heiko Carstens <heiko.carstens@de.ibm.com>
  */
@@ -209,7 +209,7 @@ static noinline __init void detect_machi
 		machine_flags |= MACHINE_FLAG_VM;
 }
 
-static __init void early_pgm_check_handler(void)
+static void early_pgm_check_handler(void)
 {
 	unsigned long addr;
 	const struct exception_table_entry *fixup;
@@ -221,7 +221,7 @@ static __init void early_pgm_check_handl
 	S390_lowcore.program_old_psw.addr = fixup->fixup | PSW_ADDR_AMODE;
 }
 
-static noinline __init void setup_lowcore_early(void)
+void setup_lowcore_early(void)
 {
 	psw_t psw;
 
Index: linux-2.6/arch/s390/kernel/smp.c
===================================================================
--- linux-2.6.orig/arch/s390/kernel/smp.c
+++ linux-2.6/arch/s390/kernel/smp.c
@@ -1,7 +1,7 @@
 /*
  *  arch/s390/kernel/smp.c
  *
- *    Copyright IBM Corp. 1999,2007
+ *    Copyright IBM Corp. 1999, 2009
  *    Author(s): Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
  *		 Martin Schwidefsky (schwidefsky@de.ibm.com)
  *		 Heiko Carstens (heiko.carstens@de.ibm.com)
@@ -1030,6 +1030,42 @@ out:
 static SYSDEV_CLASS_ATTR(dispatching, 0644, dispatching_show,
 			 dispatching_store);
 
+/*
+ * If the resume kernel runs on another cpu than the suspended kernel,
+ * we have to switch the cpu IDs in the logical map.
+ */
+void smp_switch_boot_cpu_in_resume(u32 resume_phys_cpu_id,
+				   struct _lowcore *suspend_lowcore)
+{
+	int cpu, suspend_cpu_id, resume_cpu_id;
+	u32 suspend_phys_cpu_id;
+
+	suspend_phys_cpu_id = __cpu_logical_map[suspend_lowcore->cpu_nr];
+	suspend_cpu_id = suspend_lowcore->cpu_nr;
+
+	for_each_present_cpu(cpu) {
+		if (__cpu_logical_map[cpu] == resume_phys_cpu_id) {
+			resume_cpu_id = cpu;
+			goto found;
+		}
+	}
+	panic("Could not find resume cpu in logical map.\n");
+
+found:
+	printk("Resume  cpu ID: %i/%i\n", resume_phys_cpu_id, resume_cpu_id);
+	printk("Suspend cpu ID: %i/%i\n", suspend_phys_cpu_id, suspend_cpu_id);
+
+	__cpu_logical_map[resume_cpu_id] = suspend_phys_cpu_id;
+	__cpu_logical_map[suspend_cpu_id] = resume_phys_cpu_id;
+
+	lowcore_ptr[suspend_cpu_id]->cpu_addr = resume_phys_cpu_id;
+}
+
+u32 smp_get_phys_cpu_id(void)
+{
+	return __cpu_logical_map[smp_processor_id()];
+}
+
 static int __init topology_init(void)
 {
 	int cpu;
Index: linux-2.6/arch/s390/include/asm/system.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/system.h
+++ linux-2.6/arch/s390/include/asm/system.h
@@ -1,11 +1,7 @@
 /*
- *  include/asm-s390/system.h
+ * Copyright IBM Corp. 1999, 2009
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
- *
- *  Derived from "include/asm-i386/system.h"
+ * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #ifndef __ASM_SYSTEM_H
@@ -469,6 +465,20 @@ extern psw_t sysc_restore_trace_psw;
 extern psw_t io_restore_trace_psw;
 #endif
 
+static inline int tprot(unsigned long addr)
+{
+	int rc = -EFAULT;
+
+	asm volatile(
+		"	tprot	0(%1),0\n"
+		"0:	ipm	%0\n"
+		"	srl	%0,28\n"
+		"1:\n"
+		EX_TABLE(0b,1b)
+		: "+d" (rc) : "a" (addr) : "cc");
+	return rc;
+}
+
 #endif /* __KERNEL__ */
 
 #endif
Index: linux-2.6/arch/s390/kernel/mem_detect.c
===================================================================
--- linux-2.6.orig/arch/s390/kernel/mem_detect.c
+++ linux-2.6/arch/s390/kernel/mem_detect.c
@@ -1,6 +1,7 @@
 /*
- *    Copyright IBM Corp. 2008
- *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ * Copyright IBM Corp. 2008, 2009
+ *
+ * Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
  */
 
 #include <linux/kernel.h>
@@ -9,20 +10,6 @@
 #include <asm/sclp.h>
 #include <asm/setup.h>
 
-static inline int tprot(unsigned long addr)
-{
-	int rc = -EFAULT;
-
-	asm volatile(
-		"	tprot	0(%1),0\n"
-		"0:	ipm	%0\n"
-		"	srl	%0,28\n"
-		"1:\n"
-		EX_TABLE(0b,1b)
-		: "+d" (rc) : "a" (addr) : "cc");
-	return rc;
-}
-
 #define ADDR2G (1ULL << 31)
 
 static void find_memory_chunks(struct mem_chunk chunk[])

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 05/38] s390: hibernation support for s390

[Martin Schwidefsky]

[-- Attachment #1: pm_backend.patch --]
[-- Type: text/plain, Size: 15350 bytes --]

From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>

This patch introduces the hibernation backend support to the 
s390 architecture. Now it is possible to suspend a mainframe Linux 
guest using the following command:

echo disk > /sys/power/state

Signed-off-by: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/Kconfig               |    9 +
 arch/s390/Makefile              |    4 
 arch/s390/include/asm/suspend.h |    5 +
 arch/s390/include/asm/system.h  |   22 +++-
 arch/s390/kernel/early.c        |    6 -
 arch/s390/kernel/mem_detect.c   |   19 ---
 arch/s390/kernel/smp.c          |   38 +++++++
 arch/s390/power/Makefile        |    8 +
 arch/s390/power/suspend.c       |   40 ++++++++
 arch/s390/power/swsusp.c        |   31 ++++++
 arch/s390/power/swsusp_64.c     |   17 +++
 arch/s390/power/swsusp_asm64.S  |  199 ++++++++++++++++++++++++++++++++++++++++
 12 files changed, 371 insertions(+), 27 deletions(-)

Index: linux-2.6/arch/s390/include/asm/suspend.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/suspend.h
+++ linux-2.6/arch/s390/include/asm/suspend.h
@@ -1,5 +1,10 @@
 #ifndef __ASM_S390_SUSPEND_H
 #define __ASM_S390_SUSPEND_H
 
+static inline int arch_prepare_suspend(void)
+{
+	return 0;
+}
+
 #endif
 
Index: linux-2.6/arch/s390/Kconfig
===================================================================
--- linux-2.6.orig/arch/s390/Kconfig
+++ linux-2.6/arch/s390/Kconfig
@@ -343,6 +343,9 @@ config ARCH_ENABLE_MEMORY_HOTPLUG
 config ARCH_ENABLE_MEMORY_HOTREMOVE
 	def_bool y
 
+config ARCH_HIBERNATION_POSSIBLE
+       def_bool y
+
 source "mm/Kconfig"
 
 comment "I/O subsystem configuration"
@@ -587,6 +590,12 @@ config SECCOMP
 
 endmenu
 
+menu "Power Management"
+
+source "kernel/power/Kconfig"
+
+endmenu
+
 source "net/Kconfig"
 
 config PCMCIA
Index: linux-2.6/arch/s390/Makefile
===================================================================
--- linux-2.6.orig/arch/s390/Makefile
+++ linux-2.6/arch/s390/Makefile
@@ -88,7 +88,9 @@ LDFLAGS_vmlinux := -e start
 head-y		:= arch/s390/kernel/head.o arch/s390/kernel/init_task.o
 
 core-y		+= arch/s390/mm/ arch/s390/kernel/ arch/s390/crypto/ \
-		   arch/s390/appldata/ arch/s390/hypfs/ arch/s390/kvm/
+		   arch/s390/appldata/ arch/s390/hypfs/ arch/s390/kvm/ \
+		   arch/s390/power/
+
 libs-y		+= arch/s390/lib/
 drivers-y	+= drivers/s390/
 drivers-$(CONFIG_MATHEMU) += arch/s390/math-emu/
Index: linux-2.6/arch/s390/power/Makefile
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/Makefile
@@ -0,0 +1,8 @@
+#
+# Makefile for s390 PM support
+#
+
+obj-$(CONFIG_HIBERNATION) += suspend.o
+obj-$(CONFIG_HIBERNATION) += swsusp.o
+obj-$(CONFIG_HIBERNATION) += swsusp_64.o
+obj-$(CONFIG_HIBERNATION) += swsusp_asm64.o
Index: linux-2.6/arch/s390/power/suspend.c
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/suspend.c
@@ -0,0 +1,40 @@
+/*
+ * Suspend support specific for s390.
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ */
+
+#include <linux/mm.h>
+#include <linux/suspend.h>
+#include <linux/reboot.h>
+#include <linux/pfn.h>
+#include <asm/sections.h>
+#include <asm/ipl.h>
+
+/*
+ * References to section boundaries
+ */
+extern const void __nosave_begin, __nosave_end;
+
+/*
+ *  check if given pfn is in the 'nosave' or in the read only NSS section
+ */
+int pfn_is_nosave(unsigned long pfn)
+{
+	unsigned long nosave_begin_pfn = __pa(&__nosave_begin) >> PAGE_SHIFT;
+	unsigned long nosave_end_pfn = PAGE_ALIGN(__pa(&__nosave_end))
+					>> PAGE_SHIFT;
+	unsigned long eshared_pfn = PFN_DOWN(__pa(&_eshared)) - 1;
+	unsigned long stext_pfn = PFN_DOWN(__pa(&_stext));
+
+	if (pfn >= nosave_begin_pfn && pfn < nosave_end_pfn)
+		return 1;
+	if (pfn >= stext_pfn && pfn <= eshared_pfn) {
+		if (ipl_info.type == IPL_TYPE_NSS)
+			return 1;
+	} else if ((tprot(pfn * PAGE_SIZE) && pfn > 0))
+		return 1;
+	return 0;
+}
Index: linux-2.6/arch/s390/power/swsusp_64.c
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/swsusp_64.c
@@ -0,0 +1,17 @@
+/*
+ * Support for suspend and resume on s390
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ *
+ */
+
+#include <asm/system.h>
+#include <linux/interrupt.h>
+
+void do_after_copyback(void)
+{
+	mb();
+}
+
Index: linux-2.6/arch/s390/power/swsusp_asm64.S
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/swsusp_asm64.S
@@ -0,0 +1,199 @@
+/*
+ * S390 64-bit swsusp implementation
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ *            Michael Holzheu <holzheu@linux.vnet.ibm.com>
+ */
+
+#include <asm/page.h>
+#include <asm/ptrace.h>
+#include <asm/asm-offsets.h>
+
+/*
+ * Save register context in absolute 0 lowcore and call swsusp_save() to
+ * create in-memory kernel image. The context is saved in the designated
+ * "store status" memory locations (see POP).
+ * We return from this function twice. The first time during the suspend to
+ * disk process. The second time via the swsusp_arch_resume() function
+ * (see below) in the resume process.
+ * This function runs with disabled interrupts.
+ */
+	.section .text
+	.align  2
+	.globl swsusp_arch_suspend
+swsusp_arch_suspend:
+	stmg	%r6,%r15,__SF_GPRS(%r15)
+	lgr	%r1,%r15
+	aghi	%r15,-STACK_FRAME_OVERHEAD
+	stg	%r1,__SF_BACKCHAIN(%r15)
+
+	/* Deactivate DAT */
+ 	stnsm	__SF_EMPTY(%r15),0xfb
+
+	/* Switch off lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	ni	__SF_EMPTY+4(%r15),0xef
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Store prefix register on stack */
+	stpx	__SF_EMPTY(%r15)
+
+	/* Setup base register for lowcore (absolute 0) */
+	llgf	%r1,__SF_EMPTY(%r15)
+
+	/* Get pointer to save area */
+	aghi	%r1,0x1000
+
+	/* Store registers */
+	mvc	0x318(4,%r1),__SF_EMPTY(%r15)	/* move prefix to lowcore */
+	stfpc	0x31c(%r1)			/* store fpu control */
+	std	0,0x200(%r1)			/* store f0 */
+	std	1,0x208(%r1)			/* store f1 */
+	std	2,0x210(%r1)			/* store f2 */
+	std	3,0x218(%r1)			/* store f3 */
+	std	4,0x220(%r1)			/* store f4 */
+	std	5,0x228(%r1)			/* store f5 */
+	std	6,0x230(%r1)			/* store f6 */
+	std	7,0x238(%r1)			/* store f7 */
+	std	8,0x240(%r1)			/* store f8 */
+	std	9,0x248(%r1)			/* store f9 */
+	std	10,0x250(%r1)			/* store f10 */
+	std	11,0x258(%r1)			/* store f11 */
+	std	12,0x260(%r1)			/* store f12 */
+	std	13,0x268(%r1)			/* store f13 */
+	std	14,0x270(%r1)			/* store f14 */
+	std	15,0x278(%r1)			/* store f15 */
+	stam	%a0,%a15,0x340(%r1)		/* store access registers */
+	stctg	%c0,%c15,0x380(%r1)		/* store control registers */
+	stmg	%r0,%r15,0x280(%r1)		/* store general registers */
+
+	stpt	0x328(%r1)			/* store timer */
+	stckc	0x330(%r1)			/* store clock comparator */
+
+	/* Activate DAT */
+	stosm	__SF_EMPTY(%r15),0x04
+
+	/* Set prefix page to zero */
+	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
+	spx	__SF_EMPTY(%r15)
+
+	/* Setup lowcore */
+	brasl	%r14,setup_lowcore_early
+
+	/* Save image */
+	brasl	%r14,swsusp_save
+
+	/* Switch on lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	oi	__SF_EMPTY+4(%r15),0x10
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Restore prefix register and return */
+	lghi	%r1,0x1000
+	spx	0x318(%r1)
+	lmg	%r6,%r15,STACK_FRAME_OVERHEAD + __SF_GPRS(%r15)
+	lghi	%r2,0
+	br	%r14
+
+/*
+ * Restore saved memory image to correct place and restore register context.
+ * Then we return to the function that called swsusp_arch_suspend().
+ * swsusp_arch_resume() runs with disabled interrupts.
+ */
+	.globl swsusp_arch_resume
+swsusp_arch_resume:
+	stmg	%r6,%r15,__SF_GPRS(%r15)
+	lgr	%r1,%r15
+	aghi	%r15,-STACK_FRAME_OVERHEAD
+	stg	%r1,__SF_BACKCHAIN(%r15)
+
+	/* Save boot cpu number */
+	brasl	%r14,smp_get_phys_cpu_id
+	lgr	%r10,%r2
+
+	/* Deactivate DAT */
+ 	stnsm	__SF_EMPTY(%r15),0xfb
+
+	/* Switch off lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	ni	__SF_EMPTY+4(%r15),0xef
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Set prefix page to zero */
+	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
+	spx	__SF_EMPTY(%r15)
+
+	/* Restore saved image */
+	larl	%r1,restore_pblist
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	2f
+0:
+	lg	%r2,8(%r1)
+	lg	%r4,0(%r1)
+	lghi	%r3,PAGE_SIZE
+	lghi	%r5,PAGE_SIZE
+1:
+	mvcle	%r2,%r4,0
+	jo	1b
+	lg	%r1,16(%r1)
+	ltgr	%r1,%r1
+	jnz	0b
+2:
+	ptlb				/* flush tlb */
+
+	/* Restore registers */
+	lghi	%r13,0x1000		/* %r1 = pointer to save arae */
+
+	spt	0x328(%r13)		/* reprogram timer */
+	//sckc	0x330(%r13)		/* set clock comparator */
+
+	lctlg	%c0,%c15,0x380(%r13)	/* load control registers */
+	lam	%a0,%a15,0x340(%r13)	/* load access registers */
+
+	lfpc	0x31c(%r13)		/* load fpu control */
+	ld	0,0x200(%r13)		/* load f0 */
+	ld	1,0x208(%r13)		/* load f1 */
+	ld	2,0x210(%r13)		/* load f2 */
+	ld	3,0x218(%r13)		/* load f3 */
+	ld	4,0x220(%r13)		/* load f4 */
+	ld	5,0x228(%r13)		/* load f5 */
+	ld	6,0x230(%r13)		/* load f6 */
+	ld	7,0x238(%r13)		/* load f7 */
+	ld	8,0x240(%r13)		/* load f8 */
+	ld	9,0x248(%r13)		/* load f9 */
+	ld	10,0x250(%r13)		/* load f10 */
+	ld	11,0x258(%r13)		/* load f11 */
+	ld	12,0x260(%r13)		/* load f12 */
+	ld	13,0x268(%r13)		/* load f13 */
+	ld	14,0x270(%r13)		/* load f14 */
+	ld	15,0x278(%r13)		/* load f15 */
+
+	/* Load old stack */
+	lg	%r15,0x2f8(%r13)
+
+	/* Pointer to save arae */
+	lghi	%r13,0x1000
+
+	/* Switch CPUs */
+	lgr	%r2,%r10		/* get cpu id */
+	llgf	%r3,0x318(%r13)
+	brasl	%r14,smp_switch_boot_cpu_in_resume
+
+	/* Restore prefix register */
+	spx	0x318(%r13)
+
+	/* Switch on lowcore protection */
+	stctg	%c0,%c0,__SF_EMPTY(%r15)
+	oi	__SF_EMPTY+4(%r15),0x10
+	lctlg	%c0,%c0,__SF_EMPTY(%r15)
+
+	/* Activate DAT */
+	stosm	__SF_EMPTY(%r15),0x04
+
+	/* Return 0 */
+	lmg	%r6,%r15,STACK_FRAME_OVERHEAD + __SF_GPRS(%r15)
+	lghi	%r2,0
+	br	%r14
Index: linux-2.6/arch/s390/power/swsusp.c
===================================================================
--- /dev/null
+++ linux-2.6/arch/s390/power/swsusp.c
@@ -0,0 +1,31 @@
+/*
+ * Support for suspend and resume on s390
+ *
+ * Copyright IBM Corp. 2009
+ *
+ * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
+ *
+ */
+
+#include <linux/suspend.h>
+#include <linux/gfp.h>
+#include <linux/cpu.h>
+#include <linux/smp.h>
+#include <linux/cpumask.h>
+#include <asm/ipl.h>
+#include <asm/processor.h>
+
+/*
+ * save CPU registers before creating a hibernation image and before
+ * restoring the memory state from it
+ */
+void save_processor_state(void)
+{
+}
+
+/*
+ * restore the contents of CPU registers
+ */
+void restore_processor_state(void)
+{
+}
Index: linux-2.6/arch/s390/kernel/early.c
===================================================================
--- linux-2.6.orig/arch/s390/kernel/early.c
+++ linux-2.6/arch/s390/kernel/early.c
@@ -1,7 +1,7 @@
 /*
  *  arch/s390/kernel/early.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Hongjie Yang <hongjie@us.ibm.com>,
  *		 Heiko Carstens <heiko.carstens@de.ibm.com>
  */
@@ -209,7 +209,7 @@ static noinline __init void detect_machi
 		machine_flags |= MACHINE_FLAG_VM;
 }
 
-static __init void early_pgm_check_handler(void)
+static void early_pgm_check_handler(void)
 {
 	unsigned long addr;
 	const struct exception_table_entry *fixup;
@@ -221,7 +221,7 @@ static __init void early_pgm_check_handl
 	S390_lowcore.program_old_psw.addr = fixup->fixup | PSW_ADDR_AMODE;
 }
 
-static noinline __init void setup_lowcore_early(void)
+void setup_lowcore_early(void)
 {
 	psw_t psw;
 
Index: linux-2.6/arch/s390/kernel/smp.c
===================================================================
--- linux-2.6.orig/arch/s390/kernel/smp.c
+++ linux-2.6/arch/s390/kernel/smp.c
@@ -1,7 +1,7 @@
 /*
  *  arch/s390/kernel/smp.c
  *
- *    Copyright IBM Corp. 1999,2007
+ *    Copyright IBM Corp. 1999, 2009
  *    Author(s): Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
  *		 Martin Schwidefsky (schwidefsky@de.ibm.com)
  *		 Heiko Carstens (heiko.carstens@de.ibm.com)
@@ -1030,6 +1030,42 @@ out:
 static SYSDEV_CLASS_ATTR(dispatching, 0644, dispatching_show,
 			 dispatching_store);
 
+/*
+ * If the resume kernel runs on another cpu than the suspended kernel,
+ * we have to switch the cpu IDs in the logical map.
+ */
+void smp_switch_boot_cpu_in_resume(u32 resume_phys_cpu_id,
+				   struct _lowcore *suspend_lowcore)
+{
+	int cpu, suspend_cpu_id, resume_cpu_id;
+	u32 suspend_phys_cpu_id;
+
+	suspend_phys_cpu_id = __cpu_logical_map[suspend_lowcore->cpu_nr];
+	suspend_cpu_id = suspend_lowcore->cpu_nr;
+
+	for_each_present_cpu(cpu) {
+		if (__cpu_logical_map[cpu] == resume_phys_cpu_id) {
+			resume_cpu_id = cpu;
+			goto found;
+		}
+	}
+	panic("Could not find resume cpu in logical map.\n");
+
+found:
+	printk("Resume  cpu ID: %i/%i\n", resume_phys_cpu_id, resume_cpu_id);
+	printk("Suspend cpu ID: %i/%i\n", suspend_phys_cpu_id, suspend_cpu_id);
+
+	__cpu_logical_map[resume_cpu_id] = suspend_phys_cpu_id;
+	__cpu_logical_map[suspend_cpu_id] = resume_phys_cpu_id;
+
+	lowcore_ptr[suspend_cpu_id]->cpu_addr = resume_phys_cpu_id;
+}
+
+u32 smp_get_phys_cpu_id(void)
+{
+	return __cpu_logical_map[smp_processor_id()];
+}
+
 static int __init topology_init(void)
 {
 	int cpu;
Index: linux-2.6/arch/s390/include/asm/system.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/system.h
+++ linux-2.6/arch/s390/include/asm/system.h
@@ -1,11 +1,7 @@
 /*
- *  include/asm-s390/system.h
+ * Copyright IBM Corp. 1999, 2009
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
- *
- *  Derived from "include/asm-i386/system.h"
+ * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #ifndef __ASM_SYSTEM_H
@@ -469,6 +465,20 @@ extern psw_t sysc_restore_trace_psw;
 extern psw_t io_restore_trace_psw;
 #endif
 
+static inline int tprot(unsigned long addr)
+{
+	int rc = -EFAULT;
+
+	asm volatile(
+		"	tprot	0(%1),0\n"
+		"0:	ipm	%0\n"
+		"	srl	%0,28\n"
+		"1:\n"
+		EX_TABLE(0b,1b)
+		: "+d" (rc) : "a" (addr) : "cc");
+	return rc;
+}
+
 #endif /* __KERNEL__ */
 
 #endif
Index: linux-2.6/arch/s390/kernel/mem_detect.c
===================================================================
--- linux-2.6.orig/arch/s390/kernel/mem_detect.c
+++ linux-2.6/arch/s390/kernel/mem_detect.c
@@ -1,6 +1,7 @@
 /*
- *    Copyright IBM Corp. 2008
- *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ * Copyright IBM Corp. 2008, 2009
+ *
+ * Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
  */
 
 #include <linux/kernel.h>
@@ -9,20 +10,6 @@
 #include <asm/sclp.h>
 #include <asm/setup.h>
 
-static inline int tprot(unsigned long addr)
-{
-	int rc = -EFAULT;
-
-	asm volatile(
-		"	tprot	0(%1),0\n"
-		"0:	ipm	%0\n"
-		"	srl	%0,28\n"
-		"1:\n"
-		EX_TABLE(0b,1b)
-		: "+d" (rc) : "a" (addr) : "cc");
-	return rc;
-}
-
 #define ADDR2G (1ULL << 31)
 
 static void find_memory_chunks(struct mem_chunk chunk[])

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 06/38] pm: ccw bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ccw.patch --]
[-- Type: text/plain, Size: 18543 bytes --]

From: Sebastian Ott <sebott@linux.vnet.ibm.com>

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/include/asm/ccwdev.h |   18 ++-
 drivers/s390/cio/cmf.c         |    5 
 drivers/s390/cio/device.c      |  237 +++++++++++++++++++++++++++++++++++++++++
 drivers/s390/cio/device.h      |    3 
 drivers/s390/cio/device_fsm.c  |   96 +++++++---------
 drivers/s390/cio/device_ops.c  |   26 +++-
 drivers/s390/cio/io_sch.h      |    1 
 7 files changed, 321 insertions(+), 65 deletions(-)

Index: linux-2.6/drivers/s390/cio/device.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device.c
+++ linux-2.6/drivers/s390/cio/device.c
@@ -1895,6 +1895,242 @@ static void ccw_device_shutdown(struct d
 	disable_cmf(cdev);
 }
 
+static int ccw_device_pm_prepare(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+
+	if (work_pending(&cdev->private->kick_work))
+		return -EAGAIN;
+	/* Fail while device is being set online/offline. */
+	if (atomic_read(&cdev->private->onoff))
+		return -EAGAIN;
+
+	if (cdev->online && cdev->drv && cdev->drv->prepare)
+		return cdev->drv->prepare(cdev);
+
+	return 0;
+}
+
+static void ccw_device_pm_complete(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+
+	if (cdev->online && cdev->drv && cdev->drv->complete)
+		cdev->drv->complete(cdev);
+}
+
+static int ccw_device_pm_freeze(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret, cm_enabled;
+
+	/* Fail suspend while device is in transistional state. */
+	if (!dev_fsm_final_state(cdev))
+		return -EAGAIN;
+	if (!cdev->online)
+		return 0;
+	if (cdev->drv && cdev->drv->freeze) {
+		ret = cdev->drv->freeze(cdev);
+		if (ret)
+			return ret;
+	}
+
+	spin_lock_irq(sch->lock);
+	cm_enabled = cdev->private->cmb != NULL;
+	spin_unlock_irq(sch->lock);
+	if (cm_enabled) {
+		/* Don't have the css write on memory. */
+		ret = ccw_set_cmf(cdev, 0);
+		if (ret)
+			return ret;
+	}
+	/* From here on, disallow device driver I/O. */
+	spin_lock_irq(sch->lock);
+	ret = cio_disable_subchannel(sch);
+	spin_unlock_irq(sch->lock);
+
+	return ret;
+}
+
+static int ccw_device_pm_thaw(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret, cm_enabled;
+
+	if (!cdev->online)
+		return 0;
+
+	spin_lock_irq(sch->lock);
+	/* Allow device driver I/O again. */
+	ret = cio_enable_subchannel(sch, (u32)(addr_t)sch);
+	cm_enabled = cdev->private->cmb != NULL;
+	spin_unlock_irq(sch->lock);
+	if (ret)
+		return ret;
+
+	if (cm_enabled) {
+		ret = ccw_set_cmf(cdev, 1);
+		if (ret)
+			return ret;
+	}
+
+	if (cdev->drv && cdev->drv->thaw)
+		ret = cdev->drv->thaw(cdev);
+
+	return ret;
+}
+
+static void __ccw_device_pm_restore(struct ccw_device *cdev)
+{
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret;
+
+	if (cio_is_console(sch->schid))
+		goto out;
+	/*
+	 * While we were sleeping, devices may have gone or become
+	 * available again. Kick re-detection.
+	 */
+	spin_lock_irq(sch->lock);
+	cdev->private->flags.resuming = 1;
+	ret = ccw_device_recognition(cdev);
+	spin_unlock_irq(sch->lock);
+	if (ret) {
+		CIO_MSG_EVENT(0, "Couldn't start recognition for device "
+			      "%s (ret=%d)\n", dev_name(&cdev->dev), ret);
+		spin_lock_irq(sch->lock);
+		cdev->private->state = DEV_STATE_DISCONNECTED;
+		spin_unlock_irq(sch->lock);
+		/* notify driver after the resume cb */
+		goto out;
+	}
+	wait_event(cdev->private->wait_q, dev_fsm_final_state(cdev) ||
+		   cdev->private->state == DEV_STATE_DISCONNECTED);
+
+out:
+	cdev->private->flags.resuming = 0;
+}
+
+static int resume_handle_boxed(struct ccw_device *cdev)
+{
+	cdev->private->state = DEV_STATE_BOXED;
+	if (ccw_device_notify(cdev, CIO_BOXED))
+		return 0;
+	ccw_device_schedule_sch_unregister(cdev);
+	return -ENODEV;
+}
+
+static int resume_handle_disc(struct ccw_device *cdev)
+{
+	cdev->private->state = DEV_STATE_DISCONNECTED;
+	if (ccw_device_notify(cdev, CIO_GONE))
+		return 0;
+	ccw_device_schedule_sch_unregister(cdev);
+	return -ENODEV;
+}
+
+static int ccw_device_pm_restore(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret = 0, cm_enabled;
+
+	__ccw_device_pm_restore(cdev);
+	spin_lock_irq(sch->lock);
+	if (cio_is_console(sch->schid)) {
+		cio_enable_subchannel(sch, (u32)(addr_t)sch);
+		spin_unlock_irq(sch->lock);
+		goto out_restore;
+	}
+	cdev->private->flags.donotify = 0;
+	/* check recognition results */
+	switch (cdev->private->state) {
+	case DEV_STATE_OFFLINE:
+		break;
+	case DEV_STATE_BOXED:
+		ret = resume_handle_boxed(cdev);
+		spin_unlock_irq(sch->lock);
+		if (ret)
+			goto out;
+		goto out_restore;
+	case DEV_STATE_DISCONNECTED:
+		goto out_disc_unlock;
+	default:
+		goto out_unreg_unlock;
+	}
+	/* check if the device id has changed */
+	if (sch->schib.pmcw.dev != cdev->private->dev_id.devno) {
+		CIO_MSG_EVENT(0, "resume: sch %s: failed (devno changed from "
+			      "%04x to %04x)\n", dev_name(&sch->dev),
+			      cdev->private->dev_id.devno,
+			      sch->schib.pmcw.dev);
+		goto out_unreg_unlock;
+	}
+	/* check if the device type has changed */
+	if (!ccw_device_test_sense_data(cdev)) {
+		ccw_device_update_sense_data(cdev);
+		PREPARE_WORK(&cdev->private->kick_work,
+			     ccw_device_do_unbind_bind);
+		queue_work(ccw_device_work, &cdev->private->kick_work);
+		ret = -ENODEV;
+		goto out_unlock;
+	}
+	if (!cdev->online) {
+		ret = 0;
+		goto out_unlock;
+	}
+	ret = ccw_device_online(cdev);
+	if (ret)
+		goto out_disc_unlock;
+
+	cm_enabled = cdev->private->cmb != NULL;
+	spin_unlock_irq(sch->lock);
+
+	wait_event(cdev->private->wait_q, dev_fsm_final_state(cdev));
+	if (cdev->private->state != DEV_STATE_ONLINE) {
+		spin_lock_irq(sch->lock);
+		goto out_disc_unlock;
+	}
+	if (cm_enabled) {
+		ret = ccw_set_cmf(cdev, 1);
+		if (ret) {
+			CIO_MSG_EVENT(2, "resume: cdev %s: cmf failed "
+				      "(rc=%d)\n", dev_name(&cdev->dev), ret);
+			ret = 0;
+		}
+	}
+
+out_restore:
+	if (cdev->online && cdev->drv && cdev->drv->restore)
+		ret = cdev->drv->restore(cdev);
+out:
+	return ret;
+
+out_disc_unlock:
+	ret = resume_handle_disc(cdev);
+	spin_unlock_irq(sch->lock);
+	if (ret)
+		return ret;
+	goto out_restore;
+
+out_unreg_unlock:
+	ccw_device_schedule_sch_unregister(cdev);
+	ret = -ENODEV;
+out_unlock:
+	spin_unlock_irq(sch->lock);
+	return ret;
+}
+
+static struct dev_pm_ops ccw_pm_ops = {
+	.prepare = ccw_device_pm_prepare,
+	.complete = ccw_device_pm_complete,
+	.freeze = ccw_device_pm_freeze,
+	.thaw = ccw_device_pm_thaw,
+	.restore = ccw_device_pm_restore,
+};
+
 struct bus_type ccw_bus_type = {
 	.name   = "ccw",
 	.match  = ccw_bus_match,
@@ -1902,6 +2138,7 @@ struct bus_type ccw_bus_type = {
 	.probe  = ccw_device_probe,
 	.remove = ccw_device_remove,
 	.shutdown = ccw_device_shutdown,
+	.pm = &ccw_pm_ops,
 };
 
 /**
Index: linux-2.6/drivers/s390/cio/io_sch.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/io_sch.h
+++ linux-2.6/drivers/s390/cio/io_sch.h
@@ -107,6 +107,7 @@ struct ccw_device_private {
 		unsigned int recog_done:1;  /* dev. recog. complete */
 		unsigned int fake_irb:1;    /* deliver faked irb */
 		unsigned int intretry:1;    /* retry internal operation */
+		unsigned int resuming:1;    /* recognition while resume */
 	} __attribute__((packed)) flags;
 	unsigned long intparm;	/* user interruption parameter */
 	struct qdio_irq *qdio_data;
Index: linux-2.6/arch/s390/include/asm/ccwdev.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/ccwdev.h
+++ linux-2.6/arch/s390/include/asm/ccwdev.h
@@ -1,11 +1,9 @@
 /*
- *  include/asm-s390/ccwdev.h
- *  include/asm-s390x/ccwdev.h
+ * Copyright  IBM Corp. 2002, 2009
  *
- *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Arnd Bergmann <arndb@de.ibm.com>
+ * Author(s): Arnd Bergmann <arndb@de.ibm.com>
  *
- *  Interface for CCW device drivers
+ * Interface for CCW device drivers
  */
 #ifndef _S390_CCWDEV_H_
 #define _S390_CCWDEV_H_
@@ -104,6 +102,11 @@ struct ccw_device {
  * @set_offline: called when setting device offline
  * @notify: notify driver of device state changes
  * @shutdown: called at device shutdown
+ * @prepare: prepare for pm state transition
+ * @complete: undo work done in @prepare
+ * @freeze: callback for freezing during hibernation snapshotting
+ * @thaw: undo work done in @freeze
+ * @restore: callback for restoring after hibernation
  * @driver: embedded device driver structure
  * @name: device driver name
  */
@@ -116,6 +119,11 @@ struct ccw_driver {
 	int (*set_offline) (struct ccw_device *);
 	int (*notify) (struct ccw_device *, int);
 	void (*shutdown) (struct ccw_device *);
+	int (*prepare) (struct ccw_device *);
+	void (*complete) (struct ccw_device *);
+	int (*freeze)(struct ccw_device *);
+	int (*thaw) (struct ccw_device *);
+	int (*restore)(struct ccw_device *);
 	struct device_driver driver;
 	char *name;
 };
Index: linux-2.6/drivers/s390/cio/device_ops.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device_ops.c
+++ linux-2.6/drivers/s390/cio/device_ops.c
@@ -1,10 +1,8 @@
 /*
- *  drivers/s390/cio/device_ops.c
+ * Copyright IBM Corp. 2002, 2009
  *
- *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
- *			 IBM Corporation
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
- *               Cornelia Huck (cornelia.huck@de.ibm.com)
+ * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *            Cornelia Huck (cornelia.huck@de.ibm.com)
  */
 #include <linux/module.h>
 #include <linux/init.h>
@@ -116,12 +114,15 @@ int ccw_device_clear(struct ccw_device *
 
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state != DEV_STATE_ONLINE &&
 	    cdev->private->state != DEV_STATE_W4SENSE)
 		return -EINVAL;
-	sch = to_subchannel(cdev->dev.parent);
+
 	ret = cio_clear(sch);
 	if (ret == 0)
 		cdev->private->intparm = intparm;
@@ -162,6 +163,8 @@ int ccw_device_start_key(struct ccw_devi
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
 	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state == DEV_STATE_VERIFY ||
@@ -337,12 +340,15 @@ int ccw_device_halt(struct ccw_device *c
 
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state != DEV_STATE_ONLINE &&
 	    cdev->private->state != DEV_STATE_W4SENSE)
 		return -EINVAL;
-	sch = to_subchannel(cdev->dev.parent);
+
 	ret = cio_halt(sch);
 	if (ret == 0)
 		cdev->private->intparm = intparm;
@@ -369,6 +375,8 @@ int ccw_device_resume(struct ccw_device 
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
 	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state != DEV_STATE_ONLINE ||
@@ -580,6 +588,8 @@ int ccw_device_tm_start_key(struct ccw_d
 	int rc;
 
 	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state != DEV_STATE_ONLINE)
 		return -EIO;
 	/* Adjust requested path mask to excluded varied off paths. */
@@ -669,6 +679,8 @@ int ccw_device_tm_intrg(struct ccw_devic
 {
 	struct subchannel *sch = to_subchannel(cdev->dev.parent);
 
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state != DEV_STATE_ONLINE)
 		return -EIO;
 	if (!scsw_is_tm(&sch->schib.scsw) ||
Index: linux-2.6/drivers/s390/cio/device.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device.h
+++ linux-2.6/drivers/s390/cio/device.h
@@ -87,6 +87,8 @@ int ccw_device_is_orphan(struct ccw_devi
 int ccw_device_recognition(struct ccw_device *);
 int ccw_device_online(struct ccw_device *);
 int ccw_device_offline(struct ccw_device *);
+void ccw_device_update_sense_data(struct ccw_device *);
+int ccw_device_test_sense_data(struct ccw_device *);
 void ccw_device_schedule_sch_unregister(struct ccw_device *);
 int ccw_purge_blacklisted(void);
 
@@ -133,5 +135,6 @@ extern struct bus_type ccw_bus_type;
 void retry_set_schib(struct ccw_device *cdev);
 void cmf_retry_copy_block(struct ccw_device *);
 int cmf_reenable(struct ccw_device *);
+int ccw_set_cmf(struct ccw_device *cdev, int enable);
 extern struct device_attribute dev_attr_cmb_enable;
 #endif
Index: linux-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device_fsm.c
+++ linux-2.6/drivers/s390/cio/device_fsm.c
@@ -177,29 +177,21 @@ ccw_device_cancel_halt_clear(struct ccw_
 	panic("Can't stop i/o on subchannel.\n");
 }
 
-static int
-ccw_device_handle_oper(struct ccw_device *cdev)
+void ccw_device_update_sense_data(struct ccw_device *cdev)
 {
-	struct subchannel *sch;
+	memset(&cdev->id, 0, sizeof(cdev->id));
+	cdev->id.cu_type   = cdev->private->senseid.cu_type;
+	cdev->id.cu_model  = cdev->private->senseid.cu_model;
+	cdev->id.dev_type  = cdev->private->senseid.dev_type;
+	cdev->id.dev_model = cdev->private->senseid.dev_model;
+}
 
-	sch = to_subchannel(cdev->dev.parent);
-	cdev->private->flags.recog_done = 1;
-	/*
-	 * Check if cu type and device type still match. If
-	 * not, it is certainly another device and we have to
-	 * de- and re-register.
-	 */
-	if (cdev->id.cu_type != cdev->private->senseid.cu_type ||
-	    cdev->id.cu_model != cdev->private->senseid.cu_model ||
-	    cdev->id.dev_type != cdev->private->senseid.dev_type ||
-	    cdev->id.dev_model != cdev->private->senseid.dev_model) {
-		PREPARE_WORK(&cdev->private->kick_work,
-			     ccw_device_do_unbind_bind);
-		queue_work(ccw_device_work, &cdev->private->kick_work);
-		return 0;
-	}
-	cdev->private->flags.donotify = 1;
-	return 1;
+int ccw_device_test_sense_data(struct ccw_device *cdev)
+{
+	return cdev->id.cu_type == cdev->private->senseid.cu_type &&
+		cdev->id.cu_model == cdev->private->senseid.cu_model &&
+		cdev->id.dev_type == cdev->private->senseid.dev_type &&
+		cdev->id.dev_model == cdev->private->senseid.dev_model;
 }
 
 /*
@@ -233,7 +225,7 @@ static void
 ccw_device_recog_done(struct ccw_device *cdev, int state)
 {
 	struct subchannel *sch;
-	int notify, old_lpm, same_dev;
+	int old_lpm;
 
 	sch = to_subchannel(cdev->dev.parent);
 
@@ -263,8 +255,12 @@ ccw_device_recog_done(struct ccw_device 
 		wake_up(&cdev->private->wait_q);
 		return;
 	}
-	notify = 0;
-	same_dev = 0; /* Keep the compiler quiet... */
+	if (cdev->private->flags.resuming) {
+		cdev->private->state = state;
+		cdev->private->flags.recog_done = 1;
+		wake_up(&cdev->private->wait_q);
+		return;
+	}
 	switch (state) {
 	case DEV_STATE_NOT_OPER:
 		CIO_MSG_EVENT(2, "SenseID : unknown device %04x on "
@@ -273,34 +269,31 @@ ccw_device_recog_done(struct ccw_device 
 			      sch->schid.ssid, sch->schid.sch_no);
 		break;
 	case DEV_STATE_OFFLINE:
-		if (cdev->online) {
-			same_dev = ccw_device_handle_oper(cdev);
-			notify = 1;
+		if (!cdev->online) {
+			ccw_device_update_sense_data(cdev);
+			/* Issue device info message. */
+			CIO_MSG_EVENT(4, "SenseID : device 0.%x.%04x reports: "
+				      "CU  Type/Mod = %04X/%02X, Dev Type/Mod "
+				      "= %04X/%02X\n",
+				      cdev->private->dev_id.ssid,
+				      cdev->private->dev_id.devno,
+				      cdev->id.cu_type, cdev->id.cu_model,
+				      cdev->id.dev_type, cdev->id.dev_model);
+			break;
 		}
-		/* fill out sense information */
-		memset(&cdev->id, 0, sizeof(cdev->id));
-		cdev->id.cu_type   = cdev->private->senseid.cu_type;
-		cdev->id.cu_model  = cdev->private->senseid.cu_model;
-		cdev->id.dev_type  = cdev->private->senseid.dev_type;
-		cdev->id.dev_model = cdev->private->senseid.dev_model;
-		if (notify) {
-			cdev->private->state = DEV_STATE_OFFLINE;
-			if (same_dev) {
-				/* Get device online again. */
-				ccw_device_online(cdev);
-				wake_up(&cdev->private->wait_q);
-			}
-			return;
+		cdev->private->state = DEV_STATE_OFFLINE;
+		cdev->private->flags.recog_done = 1;
+		if (ccw_device_test_sense_data(cdev)) {
+			cdev->private->flags.donotify = 1;
+			ccw_device_online(cdev);
+			wake_up(&cdev->private->wait_q);
+		} else {
+			ccw_device_update_sense_data(cdev);
+			PREPARE_WORK(&cdev->private->kick_work,
+				     ccw_device_do_unbind_bind);
+			queue_work(ccw_device_work, &cdev->private->kick_work);
 		}
-		/* Issue device info message. */
-		CIO_MSG_EVENT(4, "SenseID : device 0.%x.%04x reports: "
-			      "CU  Type/Mod = %04X/%02X, Dev Type/Mod = "
-			      "%04X/%02X\n",
-			      cdev->private->dev_id.ssid,
-			      cdev->private->dev_id.devno,
-			      cdev->id.cu_type, cdev->id.cu_model,
-			      cdev->id.dev_type, cdev->id.dev_model);
-		break;
+		return;
 	case DEV_STATE_BOXED:
 		CIO_MSG_EVENT(0, "SenseID : boxed device %04x on "
 			      " subchannel 0.%x.%04x\n",
@@ -502,9 +495,6 @@ ccw_device_recognition(struct ccw_device
 	struct subchannel *sch;
 	int ret;
 
-	if ((cdev->private->state != DEV_STATE_NOT_OPER) &&
-	    (cdev->private->state != DEV_STATE_BOXED))
-		return -EINVAL;
 	sch = to_subchannel(cdev->dev.parent);
 	ret = cio_enable_subchannel(sch, (u32)(addr_t)sch);
 	if (ret != 0)
Index: linux-2.6/drivers/s390/cio/cmf.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/cmf.c
+++ linux-2.6/drivers/s390/cio/cmf.c
@@ -1204,6 +1204,11 @@ static ssize_t cmb_enable_store(struct d
 
 DEVICE_ATTR(cmb_enable, 0644, cmb_enable_show, cmb_enable_store);
 
+int ccw_set_cmf(struct ccw_device *cdev, int enable)
+{
+	return cmbops->set(cdev, enable ? 2 : 0);
+}
+
 /**
  * enable_cmf() - switch on the channel measurement for a specific device
  *  @cdev:	The ccw device to be enabled

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 06/38] pm: ccw bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ccw.patch --]
[-- Type: text/plain, Size: 18542 bytes --]

From: Sebastian Ott <sebott@linux.vnet.ibm.com>

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/include/asm/ccwdev.h |   18 ++-
 drivers/s390/cio/cmf.c         |    5 
 drivers/s390/cio/device.c      |  237 +++++++++++++++++++++++++++++++++++++++++
 drivers/s390/cio/device.h      |    3 
 drivers/s390/cio/device_fsm.c  |   96 +++++++---------
 drivers/s390/cio/device_ops.c  |   26 +++-
 drivers/s390/cio/io_sch.h      |    1 
 7 files changed, 321 insertions(+), 65 deletions(-)

Index: linux-2.6/drivers/s390/cio/device.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device.c
+++ linux-2.6/drivers/s390/cio/device.c
@@ -1895,6 +1895,242 @@ static void ccw_device_shutdown(struct d
 	disable_cmf(cdev);
 }
 
+static int ccw_device_pm_prepare(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+
+	if (work_pending(&cdev->private->kick_work))
+		return -EAGAIN;
+	/* Fail while device is being set online/offline. */
+	if (atomic_read(&cdev->private->onoff))
+		return -EAGAIN;
+
+	if (cdev->online && cdev->drv && cdev->drv->prepare)
+		return cdev->drv->prepare(cdev);
+
+	return 0;
+}
+
+static void ccw_device_pm_complete(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+
+	if (cdev->online && cdev->drv && cdev->drv->complete)
+		cdev->drv->complete(cdev);
+}
+
+static int ccw_device_pm_freeze(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret, cm_enabled;
+
+	/* Fail suspend while device is in transistional state. */
+	if (!dev_fsm_final_state(cdev))
+		return -EAGAIN;
+	if (!cdev->online)
+		return 0;
+	if (cdev->drv && cdev->drv->freeze) {
+		ret = cdev->drv->freeze(cdev);
+		if (ret)
+			return ret;
+	}
+
+	spin_lock_irq(sch->lock);
+	cm_enabled = cdev->private->cmb != NULL;
+	spin_unlock_irq(sch->lock);
+	if (cm_enabled) {
+		/* Don't have the css write on memory. */
+		ret = ccw_set_cmf(cdev, 0);
+		if (ret)
+			return ret;
+	}
+	/* From here on, disallow device driver I/O. */
+	spin_lock_irq(sch->lock);
+	ret = cio_disable_subchannel(sch);
+	spin_unlock_irq(sch->lock);
+
+	return ret;
+}
+
+static int ccw_device_pm_thaw(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret, cm_enabled;
+
+	if (!cdev->online)
+		return 0;
+
+	spin_lock_irq(sch->lock);
+	/* Allow device driver I/O again. */
+	ret = cio_enable_subchannel(sch, (u32)(addr_t)sch);
+	cm_enabled = cdev->private->cmb != NULL;
+	spin_unlock_irq(sch->lock);
+	if (ret)
+		return ret;
+
+	if (cm_enabled) {
+		ret = ccw_set_cmf(cdev, 1);
+		if (ret)
+			return ret;
+	}
+
+	if (cdev->drv && cdev->drv->thaw)
+		ret = cdev->drv->thaw(cdev);
+
+	return ret;
+}
+
+static void __ccw_device_pm_restore(struct ccw_device *cdev)
+{
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret;
+
+	if (cio_is_console(sch->schid))
+		goto out;
+	/*
+	 * While we were sleeping, devices may have gone or become
+	 * available again. Kick re-detection.
+	 */
+	spin_lock_irq(sch->lock);
+	cdev->private->flags.resuming = 1;
+	ret = ccw_device_recognition(cdev);
+	spin_unlock_irq(sch->lock);
+	if (ret) {
+		CIO_MSG_EVENT(0, "Couldn't start recognition for device "
+			      "%s (ret=%d)\n", dev_name(&cdev->dev), ret);
+		spin_lock_irq(sch->lock);
+		cdev->private->state = DEV_STATE_DISCONNECTED;
+		spin_unlock_irq(sch->lock);
+		/* notify driver after the resume cb */
+		goto out;
+	}
+	wait_event(cdev->private->wait_q, dev_fsm_final_state(cdev) ||
+		   cdev->private->state == DEV_STATE_DISCONNECTED);
+
+out:
+	cdev->private->flags.resuming = 0;
+}
+
+static int resume_handle_boxed(struct ccw_device *cdev)
+{
+	cdev->private->state = DEV_STATE_BOXED;
+	if (ccw_device_notify(cdev, CIO_BOXED))
+		return 0;
+	ccw_device_schedule_sch_unregister(cdev);
+	return -ENODEV;
+}
+
+static int resume_handle_disc(struct ccw_device *cdev)
+{
+	cdev->private->state = DEV_STATE_DISCONNECTED;
+	if (ccw_device_notify(cdev, CIO_GONE))
+		return 0;
+	ccw_device_schedule_sch_unregister(cdev);
+	return -ENODEV;
+}
+
+static int ccw_device_pm_restore(struct device *dev)
+{
+	struct ccw_device *cdev = to_ccwdev(dev);
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	int ret = 0, cm_enabled;
+
+	__ccw_device_pm_restore(cdev);
+	spin_lock_irq(sch->lock);
+	if (cio_is_console(sch->schid)) {
+		cio_enable_subchannel(sch, (u32)(addr_t)sch);
+		spin_unlock_irq(sch->lock);
+		goto out_restore;
+	}
+	cdev->private->flags.donotify = 0;
+	/* check recognition results */
+	switch (cdev->private->state) {
+	case DEV_STATE_OFFLINE:
+		break;
+	case DEV_STATE_BOXED:
+		ret = resume_handle_boxed(cdev);
+		spin_unlock_irq(sch->lock);
+		if (ret)
+			goto out;
+		goto out_restore;
+	case DEV_STATE_DISCONNECTED:
+		goto out_disc_unlock;
+	default:
+		goto out_unreg_unlock;
+	}
+	/* check if the device id has changed */
+	if (sch->schib.pmcw.dev != cdev->private->dev_id.devno) {
+		CIO_MSG_EVENT(0, "resume: sch %s: failed (devno changed from "
+			      "%04x to %04x)\n", dev_name(&sch->dev),
+			      cdev->private->dev_id.devno,
+			      sch->schib.pmcw.dev);
+		goto out_unreg_unlock;
+	}
+	/* check if the device type has changed */
+	if (!ccw_device_test_sense_data(cdev)) {
+		ccw_device_update_sense_data(cdev);
+		PREPARE_WORK(&cdev->private->kick_work,
+			     ccw_device_do_unbind_bind);
+		queue_work(ccw_device_work, &cdev->private->kick_work);
+		ret = -ENODEV;
+		goto out_unlock;
+	}
+	if (!cdev->online) {
+		ret = 0;
+		goto out_unlock;
+	}
+	ret = ccw_device_online(cdev);
+	if (ret)
+		goto out_disc_unlock;
+
+	cm_enabled = cdev->private->cmb != NULL;
+	spin_unlock_irq(sch->lock);
+
+	wait_event(cdev->private->wait_q, dev_fsm_final_state(cdev));
+	if (cdev->private->state != DEV_STATE_ONLINE) {
+		spin_lock_irq(sch->lock);
+		goto out_disc_unlock;
+	}
+	if (cm_enabled) {
+		ret = ccw_set_cmf(cdev, 1);
+		if (ret) {
+			CIO_MSG_EVENT(2, "resume: cdev %s: cmf failed "
+				      "(rc=%d)\n", dev_name(&cdev->dev), ret);
+			ret = 0;
+		}
+	}
+
+out_restore:
+	if (cdev->online && cdev->drv && cdev->drv->restore)
+		ret = cdev->drv->restore(cdev);
+out:
+	return ret;
+
+out_disc_unlock:
+	ret = resume_handle_disc(cdev);
+	spin_unlock_irq(sch->lock);
+	if (ret)
+		return ret;
+	goto out_restore;
+
+out_unreg_unlock:
+	ccw_device_schedule_sch_unregister(cdev);
+	ret = -ENODEV;
+out_unlock:
+	spin_unlock_irq(sch->lock);
+	return ret;
+}
+
+static struct dev_pm_ops ccw_pm_ops = {
+	.prepare = ccw_device_pm_prepare,
+	.complete = ccw_device_pm_complete,
+	.freeze = ccw_device_pm_freeze,
+	.thaw = ccw_device_pm_thaw,
+	.restore = ccw_device_pm_restore,
+};
+
 struct bus_type ccw_bus_type = {
 	.name   = "ccw",
 	.match  = ccw_bus_match,
@@ -1902,6 +2138,7 @@ struct bus_type ccw_bus_type = {
 	.probe  = ccw_device_probe,
 	.remove = ccw_device_remove,
 	.shutdown = ccw_device_shutdown,
+	.pm = &ccw_pm_ops,
 };
 
 /**
Index: linux-2.6/drivers/s390/cio/io_sch.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/io_sch.h
+++ linux-2.6/drivers/s390/cio/io_sch.h
@@ -107,6 +107,7 @@ struct ccw_device_private {
 		unsigned int recog_done:1;  /* dev. recog. complete */
 		unsigned int fake_irb:1;    /* deliver faked irb */
 		unsigned int intretry:1;    /* retry internal operation */
+		unsigned int resuming:1;    /* recognition while resume */
 	} __attribute__((packed)) flags;
 	unsigned long intparm;	/* user interruption parameter */
 	struct qdio_irq *qdio_data;
Index: linux-2.6/arch/s390/include/asm/ccwdev.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/ccwdev.h
+++ linux-2.6/arch/s390/include/asm/ccwdev.h
@@ -1,11 +1,9 @@
 /*
- *  include/asm-s390/ccwdev.h
- *  include/asm-s390x/ccwdev.h
+ * Copyright  IBM Corp. 2002, 2009
  *
- *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Arnd Bergmann <arndb@de.ibm.com>
+ * Author(s): Arnd Bergmann <arndb@de.ibm.com>
  *
- *  Interface for CCW device drivers
+ * Interface for CCW device drivers
  */
 #ifndef _S390_CCWDEV_H_
 #define _S390_CCWDEV_H_
@@ -104,6 +102,11 @@ struct ccw_device {
  * @set_offline: called when setting device offline
  * @notify: notify driver of device state changes
  * @shutdown: called at device shutdown
+ * @prepare: prepare for pm state transition
+ * @complete: undo work done in @prepare
+ * @freeze: callback for freezing during hibernation snapshotting
+ * @thaw: undo work done in @freeze
+ * @restore: callback for restoring after hibernation
  * @driver: embedded device driver structure
  * @name: device driver name
  */
@@ -116,6 +119,11 @@ struct ccw_driver {
 	int (*set_offline) (struct ccw_device *);
 	int (*notify) (struct ccw_device *, int);
 	void (*shutdown) (struct ccw_device *);
+	int (*prepare) (struct ccw_device *);
+	void (*complete) (struct ccw_device *);
+	int (*freeze)(struct ccw_device *);
+	int (*thaw) (struct ccw_device *);
+	int (*restore)(struct ccw_device *);
 	struct device_driver driver;
 	char *name;
 };
Index: linux-2.6/drivers/s390/cio/device_ops.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device_ops.c
+++ linux-2.6/drivers/s390/cio/device_ops.c
@@ -1,10 +1,8 @@
 /*
- *  drivers/s390/cio/device_ops.c
+ * Copyright IBM Corp. 2002, 2009
  *
- *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
- *			 IBM Corporation
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
- *               Cornelia Huck (cornelia.huck@de.ibm.com)
+ * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *            Cornelia Huck (cornelia.huck@de.ibm.com)
  */
 #include <linux/module.h>
 #include <linux/init.h>
@@ -116,12 +114,15 @@ int ccw_device_clear(struct ccw_device *
 
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state != DEV_STATE_ONLINE &&
 	    cdev->private->state != DEV_STATE_W4SENSE)
 		return -EINVAL;
-	sch = to_subchannel(cdev->dev.parent);
+
 	ret = cio_clear(sch);
 	if (ret == 0)
 		cdev->private->intparm = intparm;
@@ -162,6 +163,8 @@ int ccw_device_start_key(struct ccw_devi
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
 	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state == DEV_STATE_VERIFY ||
@@ -337,12 +340,15 @@ int ccw_device_halt(struct ccw_device *c
 
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state != DEV_STATE_ONLINE &&
 	    cdev->private->state != DEV_STATE_W4SENSE)
 		return -EINVAL;
-	sch = to_subchannel(cdev->dev.parent);
+
 	ret = cio_halt(sch);
 	if (ret == 0)
 		cdev->private->intparm = intparm;
@@ -369,6 +375,8 @@ int ccw_device_resume(struct ccw_device 
 	if (!cdev || !cdev->dev.parent)
 		return -ENODEV;
 	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
 	if (cdev->private->state != DEV_STATE_ONLINE ||
@@ -580,6 +588,8 @@ int ccw_device_tm_start_key(struct ccw_d
 	int rc;
 
 	sch = to_subchannel(cdev->dev.parent);
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state != DEV_STATE_ONLINE)
 		return -EIO;
 	/* Adjust requested path mask to excluded varied off paths. */
@@ -669,6 +679,8 @@ int ccw_device_tm_intrg(struct ccw_devic
 {
 	struct subchannel *sch = to_subchannel(cdev->dev.parent);
 
+	if (!sch->schib.pmcw.ena)
+		return -EINVAL;
 	if (cdev->private->state != DEV_STATE_ONLINE)
 		return -EIO;
 	if (!scsw_is_tm(&sch->schib.scsw) ||
Index: linux-2.6/drivers/s390/cio/device.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device.h
+++ linux-2.6/drivers/s390/cio/device.h
@@ -87,6 +87,8 @@ int ccw_device_is_orphan(struct ccw_devi
 int ccw_device_recognition(struct ccw_device *);
 int ccw_device_online(struct ccw_device *);
 int ccw_device_offline(struct ccw_device *);
+void ccw_device_update_sense_data(struct ccw_device *);
+int ccw_device_test_sense_data(struct ccw_device *);
 void ccw_device_schedule_sch_unregister(struct ccw_device *);
 int ccw_purge_blacklisted(void);
 
@@ -133,5 +135,6 @@ extern struct bus_type ccw_bus_type;
 void retry_set_schib(struct ccw_device *cdev);
 void cmf_retry_copy_block(struct ccw_device *);
 int cmf_reenable(struct ccw_device *);
+int ccw_set_cmf(struct ccw_device *cdev, int enable);
 extern struct device_attribute dev_attr_cmb_enable;
 #endif
Index: linux-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device_fsm.c
+++ linux-2.6/drivers/s390/cio/device_fsm.c
@@ -177,29 +177,21 @@ ccw_device_cancel_halt_clear(struct ccw_
 	panic("Can't stop i/o on subchannel.\n");
 }
 
-static int
-ccw_device_handle_oper(struct ccw_device *cdev)
+void ccw_device_update_sense_data(struct ccw_device *cdev)
 {
-	struct subchannel *sch;
+	memset(&cdev->id, 0, sizeof(cdev->id));
+	cdev->id.cu_type   = cdev->private->senseid.cu_type;
+	cdev->id.cu_model  = cdev->private->senseid.cu_model;
+	cdev->id.dev_type  = cdev->private->senseid.dev_type;
+	cdev->id.dev_model = cdev->private->senseid.dev_model;
+}
 
-	sch = to_subchannel(cdev->dev.parent);
-	cdev->private->flags.recog_done = 1;
-	/*
-	 * Check if cu type and device type still match. If
-	 * not, it is certainly another device and we have to
-	 * de- and re-register.
-	 */
-	if (cdev->id.cu_type != cdev->private->senseid.cu_type ||
-	    cdev->id.cu_model != cdev->private->senseid.cu_model ||
-	    cdev->id.dev_type != cdev->private->senseid.dev_type ||
-	    cdev->id.dev_model != cdev->private->senseid.dev_model) {
-		PREPARE_WORK(&cdev->private->kick_work,
-			     ccw_device_do_unbind_bind);
-		queue_work(ccw_device_work, &cdev->private->kick_work);
-		return 0;
-	}
-	cdev->private->flags.donotify = 1;
-	return 1;
+int ccw_device_test_sense_data(struct ccw_device *cdev)
+{
+	return cdev->id.cu_type == cdev->private->senseid.cu_type &&
+		cdev->id.cu_model == cdev->private->senseid.cu_model &&
+		cdev->id.dev_type == cdev->private->senseid.dev_type &&
+		cdev->id.dev_model == cdev->private->senseid.dev_model;
 }
 
 /*
@@ -233,7 +225,7 @@ static void
 ccw_device_recog_done(struct ccw_device *cdev, int state)
 {
 	struct subchannel *sch;
-	int notify, old_lpm, same_dev;
+	int old_lpm;
 
 	sch = to_subchannel(cdev->dev.parent);
 
@@ -263,8 +255,12 @@ ccw_device_recog_done(struct ccw_device 
 		wake_up(&cdev->private->wait_q);
 		return;
 	}
-	notify = 0;
-	same_dev = 0; /* Keep the compiler quiet... */
+	if (cdev->private->flags.resuming) {
+		cdev->private->state = state;
+		cdev->private->flags.recog_done = 1;
+		wake_up(&cdev->private->wait_q);
+		return;
+	}
 	switch (state) {
 	case DEV_STATE_NOT_OPER:
 		CIO_MSG_EVENT(2, "SenseID : unknown device %04x on "
@@ -273,34 +269,31 @@ ccw_device_recog_done(struct ccw_device 
 			      sch->schid.ssid, sch->schid.sch_no);
 		break;
 	case DEV_STATE_OFFLINE:
-		if (cdev->online) {
-			same_dev = ccw_device_handle_oper(cdev);
-			notify = 1;
+		if (!cdev->online) {
+			ccw_device_update_sense_data(cdev);
+			/* Issue device info message. */
+			CIO_MSG_EVENT(4, "SenseID : device 0.%x.%04x reports: "
+				      "CU  Type/Mod = %04X/%02X, Dev Type/Mod "
+				      "= %04X/%02X\n",
+				      cdev->private->dev_id.ssid,
+				      cdev->private->dev_id.devno,
+				      cdev->id.cu_type, cdev->id.cu_model,
+				      cdev->id.dev_type, cdev->id.dev_model);
+			break;
 		}
-		/* fill out sense information */
-		memset(&cdev->id, 0, sizeof(cdev->id));
-		cdev->id.cu_type   = cdev->private->senseid.cu_type;
-		cdev->id.cu_model  = cdev->private->senseid.cu_model;
-		cdev->id.dev_type  = cdev->private->senseid.dev_type;
-		cdev->id.dev_model = cdev->private->senseid.dev_model;
-		if (notify) {
-			cdev->private->state = DEV_STATE_OFFLINE;
-			if (same_dev) {
-				/* Get device online again. */
-				ccw_device_online(cdev);
-				wake_up(&cdev->private->wait_q);
-			}
-			return;
+		cdev->private->state = DEV_STATE_OFFLINE;
+		cdev->private->flags.recog_done = 1;
+		if (ccw_device_test_sense_data(cdev)) {
+			cdev->private->flags.donotify = 1;
+			ccw_device_online(cdev);
+			wake_up(&cdev->private->wait_q);
+		} else {
+			ccw_device_update_sense_data(cdev);
+			PREPARE_WORK(&cdev->private->kick_work,
+				     ccw_device_do_unbind_bind);
+			queue_work(ccw_device_work, &cdev->private->kick_work);
 		}
-		/* Issue device info message. */
-		CIO_MSG_EVENT(4, "SenseID : device 0.%x.%04x reports: "
-			      "CU  Type/Mod = %04X/%02X, Dev Type/Mod = "
-			      "%04X/%02X\n",
-			      cdev->private->dev_id.ssid,
-			      cdev->private->dev_id.devno,
-			      cdev->id.cu_type, cdev->id.cu_model,
-			      cdev->id.dev_type, cdev->id.dev_model);
-		break;
+		return;
 	case DEV_STATE_BOXED:
 		CIO_MSG_EVENT(0, "SenseID : boxed device %04x on "
 			      " subchannel 0.%x.%04x\n",
@@ -502,9 +495,6 @@ ccw_device_recognition(struct ccw_device
 	struct subchannel *sch;
 	int ret;
 
-	if ((cdev->private->state != DEV_STATE_NOT_OPER) &&
-	    (cdev->private->state != DEV_STATE_BOXED))
-		return -EINVAL;
 	sch = to_subchannel(cdev->dev.parent);
 	ret = cio_enable_subchannel(sch, (u32)(addr_t)sch);
 	if (ret != 0)
Index: linux-2.6/drivers/s390/cio/cmf.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/cmf.c
+++ linux-2.6/drivers/s390/cio/cmf.c
@@ -1204,6 +1204,11 @@ static ssize_t cmb_enable_store(struct d
 
 DEVICE_ATTR(cmb_enable, 0644, cmb_enable_show, cmb_enable_store);
 
+int ccw_set_cmf(struct ccw_device *cdev, int enable)
+{
+	return cmbops->set(cdev, enable ? 2 : 0);
+}
+
 /**
  * enable_cmf() - switch on the channel measurement for a specific device
  *  @cdev:	The ccw device to be enabled

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 07/38] pm: ccwgroup bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ccwgroup.patch --]
[-- Type: text/plain, Size: 4550 bytes --]

From: Sebastian Ott <sebott@linux.vnet.ibm.com>

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/include/asm/ccwgroup.h |   10 +++++
 drivers/s390/cio/ccwgroup.c      |   78 ++++++++++++++++++++++++++++++++++++---
 2 files changed, 83 insertions(+), 5 deletions(-)

Index: linux-2.6/drivers/s390/cio/ccwgroup.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/ccwgroup.c
+++ linux-2.6/drivers/s390/cio/ccwgroup.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/cio/ccwgroup.c
  *  bus driver for ccwgroup
  *
- *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
- *                       IBM Corporation
- *    Author(s): Arnd Bergmann (arndb@de.ibm.com)
- *               Cornelia Huck (cornelia.huck@de.ibm.com)
+ *  Copyright IBM Corp. 2002, 2009
+ *
+ *  Author(s): Arnd Bergmann (arndb@de.ibm.com)
+ *             Cornelia Huck (cornelia.huck@de.ibm.com)
  */
 #include <linux/module.h>
 #include <linux/errno.h>
@@ -501,6 +500,74 @@ static void ccwgroup_shutdown(struct dev
 		gdrv->shutdown(gdev);
 }
 
+static int ccwgroup_pm_prepare(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	/* Fail while device is being set online/offline. */
+	if (atomic_read(&gdev->onoff))
+		return -EAGAIN;
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->prepare ? gdrv->prepare(gdev) : 0;
+}
+
+static void ccwgroup_pm_complete(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(dev->driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return;
+
+	if (gdrv->complete)
+		gdrv->complete(gdev);
+}
+
+static int ccwgroup_pm_freeze(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->freeze ? gdrv->freeze(gdev) : 0;
+}
+
+static int ccwgroup_pm_thaw(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->thaw ? gdrv->thaw(gdev) : 0;
+}
+
+static int ccwgroup_pm_restore(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->restore ? gdrv->restore(gdev) : 0;
+}
+
+static struct dev_pm_ops ccwgroup_pm_ops = {
+	.prepare = ccwgroup_pm_prepare,
+	.complete = ccwgroup_pm_complete,
+	.freeze = ccwgroup_pm_freeze,
+	.thaw = ccwgroup_pm_thaw,
+	.restore = ccwgroup_pm_restore,
+};
+
 static struct bus_type ccwgroup_bus_type = {
 	.name   = "ccwgroup",
 	.match  = ccwgroup_bus_match,
@@ -508,6 +575,7 @@ static struct bus_type ccwgroup_bus_type
 	.probe  = ccwgroup_probe,
 	.remove = ccwgroup_remove,
 	.shutdown = ccwgroup_shutdown,
+	.pm = &ccwgroup_pm_ops,
 };
 
 
Index: linux-2.6/arch/s390/include/asm/ccwgroup.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/ccwgroup.h
+++ linux-2.6/arch/s390/include/asm/ccwgroup.h
@@ -38,6 +38,11 @@ struct ccwgroup_device {
  * @set_online: function called when device is set online
  * @set_offline: function called when device is set offline
  * @shutdown: function called when device is shut down
+ * @prepare: prepare for pm state transition
+ * @complete: undo work done in @prepare
+ * @freeze: callback for freezing during hibernation snapshotting
+ * @thaw: undo work done in @freeze
+ * @restore: callback for restoring after hibernation
  * @driver: embedded driver structure
  */
 struct ccwgroup_driver {
@@ -51,6 +56,11 @@ struct ccwgroup_driver {
 	int (*set_online) (struct ccwgroup_device *);
 	int (*set_offline) (struct ccwgroup_device *);
 	void (*shutdown)(struct ccwgroup_device *);
+	int (*prepare) (struct ccwgroup_device *);
+	void (*complete) (struct ccwgroup_device *);
+	int (*freeze)(struct ccwgroup_device *);
+	int (*thaw) (struct ccwgroup_device *);
+	int (*restore)(struct ccwgroup_device *);
 
 	struct device_driver driver;
 };

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 07/38] pm: ccwgroup bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ccwgroup.patch --]
[-- Type: text/plain, Size: 4549 bytes --]

From: Sebastian Ott <sebott@linux.vnet.ibm.com>

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/include/asm/ccwgroup.h |   10 +++++
 drivers/s390/cio/ccwgroup.c      |   78 ++++++++++++++++++++++++++++++++++++---
 2 files changed, 83 insertions(+), 5 deletions(-)

Index: linux-2.6/drivers/s390/cio/ccwgroup.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/ccwgroup.c
+++ linux-2.6/drivers/s390/cio/ccwgroup.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/cio/ccwgroup.c
  *  bus driver for ccwgroup
  *
- *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
- *                       IBM Corporation
- *    Author(s): Arnd Bergmann (arndb@de.ibm.com)
- *               Cornelia Huck (cornelia.huck@de.ibm.com)
+ *  Copyright IBM Corp. 2002, 2009
+ *
+ *  Author(s): Arnd Bergmann (arndb@de.ibm.com)
+ *             Cornelia Huck (cornelia.huck@de.ibm.com)
  */
 #include <linux/module.h>
 #include <linux/errno.h>
@@ -501,6 +500,74 @@ static void ccwgroup_shutdown(struct dev
 		gdrv->shutdown(gdev);
 }
 
+static int ccwgroup_pm_prepare(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	/* Fail while device is being set online/offline. */
+	if (atomic_read(&gdev->onoff))
+		return -EAGAIN;
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->prepare ? gdrv->prepare(gdev) : 0;
+}
+
+static void ccwgroup_pm_complete(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(dev->driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return;
+
+	if (gdrv->complete)
+		gdrv->complete(gdev);
+}
+
+static int ccwgroup_pm_freeze(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->freeze ? gdrv->freeze(gdev) : 0;
+}
+
+static int ccwgroup_pm_thaw(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->thaw ? gdrv->thaw(gdev) : 0;
+}
+
+static int ccwgroup_pm_restore(struct device *dev)
+{
+	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+	struct ccwgroup_driver *gdrv = to_ccwgroupdrv(gdev->dev.driver);
+
+	if (!gdev->dev.driver || gdev->state != CCWGROUP_ONLINE)
+		return 0;
+
+	return gdrv->restore ? gdrv->restore(gdev) : 0;
+}
+
+static struct dev_pm_ops ccwgroup_pm_ops = {
+	.prepare = ccwgroup_pm_prepare,
+	.complete = ccwgroup_pm_complete,
+	.freeze = ccwgroup_pm_freeze,
+	.thaw = ccwgroup_pm_thaw,
+	.restore = ccwgroup_pm_restore,
+};
+
 static struct bus_type ccwgroup_bus_type = {
 	.name   = "ccwgroup",
 	.match  = ccwgroup_bus_match,
@@ -508,6 +575,7 @@ static struct bus_type ccwgroup_bus_type
 	.probe  = ccwgroup_probe,
 	.remove = ccwgroup_remove,
 	.shutdown = ccwgroup_shutdown,
+	.pm = &ccwgroup_pm_ops,
 };
 
 
Index: linux-2.6/arch/s390/include/asm/ccwgroup.h
===================================================================
--- linux-2.6.orig/arch/s390/include/asm/ccwgroup.h
+++ linux-2.6/arch/s390/include/asm/ccwgroup.h
@@ -38,6 +38,11 @@ struct ccwgroup_device {
  * @set_online: function called when device is set online
  * @set_offline: function called when device is set offline
  * @shutdown: function called when device is shut down
+ * @prepare: prepare for pm state transition
+ * @complete: undo work done in @prepare
+ * @freeze: callback for freezing during hibernation snapshotting
+ * @thaw: undo work done in @freeze
+ * @restore: callback for restoring after hibernation
  * @driver: embedded driver structure
  */
 struct ccwgroup_driver {
@@ -51,6 +56,11 @@ struct ccwgroup_driver {
 	int (*set_online) (struct ccwgroup_device *);
 	int (*set_offline) (struct ccwgroup_device *);
 	void (*shutdown)(struct ccwgroup_device *);
+	int (*prepare) (struct ccwgroup_device *);
+	void (*complete) (struct ccwgroup_device *);
+	int (*freeze)(struct ccwgroup_device *);
+	int (*thaw) (struct ccwgroup_device *);
+	int (*restore)(struct ccwgroup_device *);
 
 	struct device_driver driver;
 };

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 08/38] pm: css bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_css.patch --]
[-- Type: text/plain, Size: 7610 bytes --]

From: Sebastian Ott <sebott@linux.vnet.ibm.com>

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/cio/chsc.c |    3 
 drivers/s390/cio/chsc.h |    1 
 drivers/s390/cio/css.c  |  157 ++++++++++++++++++++++++++++++++++++++++++++++--
 drivers/s390/cio/css.h  |   10 +++
 4 files changed, 164 insertions(+), 7 deletions(-)

Index: linux-2.6/drivers/s390/cio/css.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/css.c
+++ linux-2.6/drivers/s390/cio/css.c
@@ -1,10 +1,10 @@
 /*
- *  drivers/s390/cio/css.c
- *  driver for channel subsystem
+ * driver for channel subsystem
  *
- *    Copyright IBM Corp. 2002,2008
- *    Author(s): Arnd Bergmann (arndb@de.ibm.com)
- *		 Cornelia Huck (cornelia.huck@de.ibm.com)
+ * Copyright IBM Corp. 2002, 2009
+ *
+ * Author(s): Arnd Bergmann (arndb@de.ibm.com)
+ *	      Cornelia Huck (cornelia.huck@de.ibm.com)
  */
 
 #define KMSG_COMPONENT "cio"
@@ -17,6 +17,7 @@
 #include <linux/errno.h>
 #include <linux/list.h>
 #include <linux/reboot.h>
+#include <linux/suspend.h>
 #include <asm/isc.h>
 #include <asm/crw.h>
 
@@ -780,6 +781,79 @@ static struct notifier_block css_reboot_
 };
 
 /*
+ * Since the css devices are neither on a bus nor have a class
+ * nor have a special device type, we cannot stop/restart channel
+ * path measurements via the normal suspend/resume callbacks, but have
+ * to use notifiers.
+ */
+static int css_power_event(struct notifier_block *this, unsigned long event,
+			   void *ptr)
+{
+	void *secm_area;
+	int ret, i;
+
+	switch (event) {
+	case PM_HIBERNATION_PREPARE:
+	case PM_SUSPEND_PREPARE:
+		ret = NOTIFY_DONE;
+		for (i = 0; i <= __MAX_CSSID; i++) {
+			struct channel_subsystem *css;
+
+			css = channel_subsystems[i];
+			mutex_lock(&css->mutex);
+			if (!css->cm_enabled) {
+				mutex_unlock(&css->mutex);
+				continue;
+			}
+			secm_area = (void *)get_zeroed_page(GFP_KERNEL |
+							    GFP_DMA);
+			if (secm_area) {
+				if (__chsc_do_secm(css, 0, secm_area))
+					ret = NOTIFY_BAD;
+				free_page((unsigned long)secm_area);
+			} else
+				ret = NOTIFY_BAD;
+
+			mutex_unlock(&css->mutex);
+		}
+		break;
+	case PM_POST_HIBERNATION:
+	case PM_POST_SUSPEND:
+		ret = NOTIFY_DONE;
+		for (i = 0; i <= __MAX_CSSID; i++) {
+			struct channel_subsystem *css;
+
+			css = channel_subsystems[i];
+			mutex_lock(&css->mutex);
+			if (!css->cm_enabled) {
+				mutex_unlock(&css->mutex);
+				continue;
+			}
+			secm_area = (void *)get_zeroed_page(GFP_KERNEL |
+							    GFP_DMA);
+			if (secm_area) {
+				if (__chsc_do_secm(css, 1, secm_area))
+					ret = NOTIFY_BAD;
+				free_page((unsigned long)secm_area);
+			} else
+				ret = NOTIFY_BAD;
+
+			mutex_unlock(&css->mutex);
+		}
+		/* search for subchannels, which appeared during hibernation */
+		css_schedule_reprobe();
+		break;
+	default:
+		ret = NOTIFY_DONE;
+	}
+	return ret;
+
+}
+static struct notifier_block css_power_notifier = {
+	.notifier_call = css_power_event,
+};
+
+/*
  * Now that the driver core is running, we can setup our channel subsystem.
  * The struct subchannel's are created during probing (except for the
  * static console subchannel).
@@ -852,6 +926,11 @@ init_channel_subsystem (void)
 	ret = register_reboot_notifier(&css_reboot_notifier);
 	if (ret)
 		goto out_unregister;
+	ret = register_pm_notifier(&css_power_notifier);
+	if (ret) {
+		unregister_reboot_notifier(&css_reboot_notifier);
+		goto out_unregister;
+	}
 	css_init_done = 1;
 
 	/* Enable default isc for I/O subchannels. */
@@ -953,6 +1032,73 @@ static int css_uevent(struct device *dev
 	return ret;
 }
 
+static int css_pm_prepare(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (mutex_is_locked(&sch->reg_mutex))
+		return -EAGAIN;
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	/* Notify drivers that they may not register children. */
+	return drv->prepare ? drv->prepare(sch) : 0;
+}
+
+static void css_pm_complete(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return;
+	drv = to_cssdriver(sch->dev.driver);
+	if (drv->complete)
+		drv->complete(sch);
+}
+
+static int css_pm_freeze(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	return drv->freeze ? drv->freeze(sch) : 0;
+}
+
+static int css_pm_thaw(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	return drv->thaw ? drv->thaw(sch) : 0;
+}
+
+static int css_pm_restore(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	return drv->restore ? drv->restore(sch) : 0;
+}
+
+static struct dev_pm_ops css_pm_ops = {
+	.prepare = css_pm_prepare,
+	.complete = css_pm_complete,
+	.freeze = css_pm_freeze,
+	.thaw = css_pm_thaw,
+	.restore = css_pm_restore,
+};
+
 struct bus_type css_bus_type = {
 	.name     = "css",
 	.match    = css_bus_match,
@@ -960,6 +1106,7 @@ struct bus_type css_bus_type = {
 	.remove   = css_remove,
 	.shutdown = css_shutdown,
 	.uevent   = css_uevent,
+	.pm = &css_pm_ops,
 };
 
 /**
Index: linux-2.6/drivers/s390/cio/css.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/css.h
+++ linux-2.6/drivers/s390/cio/css.h
@@ -70,6 +70,11 @@ struct chp_link;
  * @probe: function called on probe
  * @remove: function called on remove
  * @shutdown: called at device shutdown
+ * @prepare: prepare for pm state transition
+ * @complete: undo work done in @prepare
+ * @freeze: callback for freezing during hibernation snapshotting
+ * @thaw: undo work done in @freeze
+ * @restore: callback for restoring after hibernation
  * @name: name of the device driver
  */
 struct css_driver {
@@ -82,6 +87,11 @@ struct css_driver {
 	int (*probe)(struct subchannel *);
 	int (*remove)(struct subchannel *);
 	void (*shutdown)(struct subchannel *);
+	int (*prepare) (struct subchannel *);
+	void (*complete) (struct subchannel *);
+	int (*freeze)(struct subchannel *);
+	int (*thaw) (struct subchannel *);
+	int (*restore)(struct subchannel *);
 	const char *name;
 };
 
Index: linux-2.6/drivers/s390/cio/chsc.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/chsc.c
+++ linux-2.6/drivers/s390/cio/chsc.c
@@ -549,8 +549,7 @@ cleanup:
 	return ret;
 }
 
-static int
-__chsc_do_secm(struct channel_subsystem *css, int enable, void *page)
+int __chsc_do_secm(struct channel_subsystem *css, int enable, void *page)
 {
 	struct {
 		struct chsc_header request;
Index: linux-2.6/drivers/s390/cio/chsc.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/chsc.h
+++ linux-2.6/drivers/s390/cio/chsc.h
@@ -90,6 +90,7 @@ extern void chsc_free_sei_area(void);
 extern int chsc_enable_facility(int);
 struct channel_subsystem;
 extern int chsc_secm(struct channel_subsystem *, int);
+int __chsc_do_secm(struct channel_subsystem *css, int enable, void *page);
 
 int chsc_chp_vary(struct chp_id chpid, int on);
 int chsc_determine_channel_path_desc(struct chp_id chpid, int fmt, int rfmt,

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 08/38] pm: css bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_css.patch --]
[-- Type: text/plain, Size: 7609 bytes --]

From: Sebastian Ott <sebott@linux.vnet.ibm.com>

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/cio/chsc.c |    3 
 drivers/s390/cio/chsc.h |    1 
 drivers/s390/cio/css.c  |  157 ++++++++++++++++++++++++++++++++++++++++++++++--
 drivers/s390/cio/css.h  |   10 +++
 4 files changed, 164 insertions(+), 7 deletions(-)

Index: linux-2.6/drivers/s390/cio/css.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/css.c
+++ linux-2.6/drivers/s390/cio/css.c
@@ -1,10 +1,10 @@
 /*
- *  drivers/s390/cio/css.c
- *  driver for channel subsystem
+ * driver for channel subsystem
  *
- *    Copyright IBM Corp. 2002,2008
- *    Author(s): Arnd Bergmann (arndb@de.ibm.com)
- *		 Cornelia Huck (cornelia.huck@de.ibm.com)
+ * Copyright IBM Corp. 2002, 2009
+ *
+ * Author(s): Arnd Bergmann (arndb@de.ibm.com)
+ *	      Cornelia Huck (cornelia.huck@de.ibm.com)
  */
 
 #define KMSG_COMPONENT "cio"
@@ -17,6 +17,7 @@
 #include <linux/errno.h>
 #include <linux/list.h>
 #include <linux/reboot.h>
+#include <linux/suspend.h>
 #include <asm/isc.h>
 #include <asm/crw.h>
 
@@ -780,6 +781,79 @@ static struct notifier_block css_reboot_
 };
 
 /*
+ * Since the css devices are neither on a bus nor have a class
+ * nor have a special device type, we cannot stop/restart channel
+ * path measurements via the normal suspend/resume callbacks, but have
+ * to use notifiers.
+ */
+static int css_power_event(struct notifier_block *this, unsigned long event,
+			   void *ptr)
+{
+	void *secm_area;
+	int ret, i;
+
+	switch (event) {
+	case PM_HIBERNATION_PREPARE:
+	case PM_SUSPEND_PREPARE:
+		ret = NOTIFY_DONE;
+		for (i = 0; i <= __MAX_CSSID; i++) {
+			struct channel_subsystem *css;
+
+			css = channel_subsystems[i];
+			mutex_lock(&css->mutex);
+			if (!css->cm_enabled) {
+				mutex_unlock(&css->mutex);
+				continue;
+			}
+			secm_area = (void *)get_zeroed_page(GFP_KERNEL |
+							    GFP_DMA);
+			if (secm_area) {
+				if (__chsc_do_secm(css, 0, secm_area))
+					ret = NOTIFY_BAD;
+				free_page((unsigned long)secm_area);
+			} else
+				ret = NOTIFY_BAD;
+
+			mutex_unlock(&css->mutex);
+		}
+		break;
+	case PM_POST_HIBERNATION:
+	case PM_POST_SUSPEND:
+		ret = NOTIFY_DONE;
+		for (i = 0; i <= __MAX_CSSID; i++) {
+			struct channel_subsystem *css;
+
+			css = channel_subsystems[i];
+			mutex_lock(&css->mutex);
+			if (!css->cm_enabled) {
+				mutex_unlock(&css->mutex);
+				continue;
+			}
+			secm_area = (void *)get_zeroed_page(GFP_KERNEL |
+							    GFP_DMA);
+			if (secm_area) {
+				if (__chsc_do_secm(css, 1, secm_area))
+					ret = NOTIFY_BAD;
+				free_page((unsigned long)secm_area);
+			} else
+				ret = NOTIFY_BAD;
+
+			mutex_unlock(&css->mutex);
+		}
+		/* search for subchannels, which appeared during hibernation */
+		css_schedule_reprobe();
+		break;
+	default:
+		ret = NOTIFY_DONE;
+	}
+	return ret;
+
+}
+static struct notifier_block css_power_notifier = {
+	.notifier_call = css_power_event,
+};
+
+/*
  * Now that the driver core is running, we can setup our channel subsystem.
  * The struct subchannel's are created during probing (except for the
  * static console subchannel).
@@ -852,6 +926,11 @@ init_channel_subsystem (void)
 	ret = register_reboot_notifier(&css_reboot_notifier);
 	if (ret)
 		goto out_unregister;
+	ret = register_pm_notifier(&css_power_notifier);
+	if (ret) {
+		unregister_reboot_notifier(&css_reboot_notifier);
+		goto out_unregister;
+	}
 	css_init_done = 1;
 
 	/* Enable default isc for I/O subchannels. */
@@ -953,6 +1032,73 @@ static int css_uevent(struct device *dev
 	return ret;
 }
 
+static int css_pm_prepare(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (mutex_is_locked(&sch->reg_mutex))
+		return -EAGAIN;
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	/* Notify drivers that they may not register children. */
+	return drv->prepare ? drv->prepare(sch) : 0;
+}
+
+static void css_pm_complete(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return;
+	drv = to_cssdriver(sch->dev.driver);
+	if (drv->complete)
+		drv->complete(sch);
+}
+
+static int css_pm_freeze(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	return drv->freeze ? drv->freeze(sch) : 0;
+}
+
+static int css_pm_thaw(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	return drv->thaw ? drv->thaw(sch) : 0;
+}
+
+static int css_pm_restore(struct device *dev)
+{
+	struct subchannel *sch = to_subchannel(dev);
+	struct css_driver *drv;
+
+	if (!sch->dev.driver)
+		return 0;
+	drv = to_cssdriver(sch->dev.driver);
+	return drv->restore ? drv->restore(sch) : 0;
+}
+
+static struct dev_pm_ops css_pm_ops = {
+	.prepare = css_pm_prepare,
+	.complete = css_pm_complete,
+	.freeze = css_pm_freeze,
+	.thaw = css_pm_thaw,
+	.restore = css_pm_restore,
+};
+
 struct bus_type css_bus_type = {
 	.name     = "css",
 	.match    = css_bus_match,
@@ -960,6 +1106,7 @@ struct bus_type css_bus_type = {
 	.remove   = css_remove,
 	.shutdown = css_shutdown,
 	.uevent   = css_uevent,
+	.pm = &css_pm_ops,
 };
 
 /**
Index: linux-2.6/drivers/s390/cio/css.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/css.h
+++ linux-2.6/drivers/s390/cio/css.h
@@ -70,6 +70,11 @@ struct chp_link;
  * @probe: function called on probe
  * @remove: function called on remove
  * @shutdown: called at device shutdown
+ * @prepare: prepare for pm state transition
+ * @complete: undo work done in @prepare
+ * @freeze: callback for freezing during hibernation snapshotting
+ * @thaw: undo work done in @freeze
+ * @restore: callback for restoring after hibernation
  * @name: name of the device driver
  */
 struct css_driver {
@@ -82,6 +87,11 @@ struct css_driver {
 	int (*probe)(struct subchannel *);
 	int (*remove)(struct subchannel *);
 	void (*shutdown)(struct subchannel *);
+	int (*prepare) (struct subchannel *);
+	void (*complete) (struct subchannel *);
+	int (*freeze)(struct subchannel *);
+	int (*thaw) (struct subchannel *);
+	int (*restore)(struct subchannel *);
 	const char *name;
 };
 
Index: linux-2.6/drivers/s390/cio/chsc.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/chsc.c
+++ linux-2.6/drivers/s390/cio/chsc.c
@@ -549,8 +549,7 @@ cleanup:
 	return ret;
 }
 
-static int
-__chsc_do_secm(struct channel_subsystem *css, int enable, void *page)
+int __chsc_do_secm(struct channel_subsystem *css, int enable, void *page)
 {
 	struct {
 		struct chsc_header request;
Index: linux-2.6/drivers/s390/cio/chsc.h
===================================================================
--- linux-2.6.orig/drivers/s390/cio/chsc.h
+++ linux-2.6/drivers/s390/cio/chsc.h
@@ -90,6 +90,7 @@ extern void chsc_free_sei_area(void);
 extern int chsc_enable_facility(int);
 struct channel_subsystem;
 extern int chsc_secm(struct channel_subsystem *, int);
+int __chsc_do_secm(struct channel_subsystem *css, int enable, void *page);
 
 int chsc_chp_vary(struct chp_id chpid, int on);
 int chsc_determine_channel_path_desc(struct chp_id chpid, int fmt, int rfmt,

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 09/38] pm: io subchannel driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_io_sch.patch --]
[-- Type: text/plain, Size: 1356 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/cio/device.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)

Index: linux-2.6/drivers/s390/cio/device.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device.c
+++ linux-2.6/drivers/s390/cio/device.c
@@ -138,6 +138,19 @@ static struct css_device_id io_subchanne
 };
 MODULE_DEVICE_TABLE(css, io_subchannel_ids);
 
+static int io_subchannel_prepare(struct subchannel *sch)
+{
+	struct ccw_device *cdev;
+	/*
+	 * Don't allow suspend while a ccw device registration
+	 * is still outstanding.
+	 */
+	cdev = sch_get_cdev(sch);
+	if (cdev && !device_is_registered(&cdev->dev))
+		return -EAGAIN;
+	return 0;
+}
+
 static struct css_driver io_subchannel_driver = {
 	.owner = THIS_MODULE,
 	.subchannel_type = io_subchannel_ids,
@@ -148,6 +161,7 @@ static struct css_driver io_subchannel_d
 	.probe = io_subchannel_probe,
 	.remove = io_subchannel_remove,
 	.shutdown = io_subchannel_shutdown,
+	.prepare = io_subchannel_prepare,
 };
 
 struct workqueue_struct *ccw_device_work;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 09/38] pm: io subchannel driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_io_sch.patch --]
[-- Type: text/plain, Size: 1355 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/cio/device.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)

Index: linux-2.6/drivers/s390/cio/device.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/device.c
+++ linux-2.6/drivers/s390/cio/device.c
@@ -138,6 +138,19 @@ static struct css_device_id io_subchanne
 };
 MODULE_DEVICE_TABLE(css, io_subchannel_ids);
 
+static int io_subchannel_prepare(struct subchannel *sch)
+{
+	struct ccw_device *cdev;
+	/*
+	 * Don't allow suspend while a ccw device registration
+	 * is still outstanding.
+	 */
+	cdev = sch_get_cdev(sch);
+	if (cdev && !device_is_registered(&cdev->dev))
+		return -EAGAIN;
+	return 0;
+}
+
 static struct css_driver io_subchannel_driver = {
 	.owner = THIS_MODULE,
 	.subchannel_type = io_subchannel_ids,
@@ -148,6 +161,7 @@ static struct css_driver io_subchannel_d
 	.probe = io_subchannel_probe,
 	.remove = io_subchannel_remove,
 	.shutdown = io_subchannel_shutdown,
+	.prepare = io_subchannel_prepare,
 };
 
 struct workqueue_struct *ccw_device_work;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 10/38] pm: chsc subchannel driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_chsc_sch.patch --]
[-- Type: text/plain, Size: 2017 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/cio/chsc_sch.c |   32 +++++++++++++++++++++++++++++++-
 1 file changed, 31 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/cio/chsc_sch.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/chsc_sch.c
+++ linux-2.6/drivers/s390/cio/chsc_sch.c
@@ -1,7 +1,8 @@
 /*
  * Driver for s390 chsc subchannels
  *
- * Copyright IBM Corp. 2008
+ * Copyright IBM Corp. 2008, 2009
+ *
  * Author(s): Cornelia Huck <cornelia.huck@de.ibm.com>
  *
  */
@@ -112,6 +113,31 @@ static void chsc_subchannel_shutdown(str
 	cio_disable_subchannel(sch);
 }
 
+static int chsc_subchannel_prepare(struct subchannel *sch)
+{
+	int cc;
+	struct schib schib;
+	/*
+	 * Don't allow suspend while the subchannel is not idle
+	 * since we don't have a way to clear the subchannel and
+	 * cannot disable it with a request running.
+	 */
+	cc = stsch(sch->schid, &schib);
+	if (!cc && scsw_stctl(&schib.scsw))
+		return -EAGAIN;
+	return 0;
+}
+
+static int chsc_subchannel_freeze(struct subchannel *sch)
+{
+	return cio_disable_subchannel(sch);
+}
+
+static int chsc_subchannel_restore(struct subchannel *sch)
+{
+	return cio_enable_subchannel(sch, (u32)(unsigned long)sch);
+}
+
 static struct css_device_id chsc_subchannel_ids[] = {
 	{ .match_flags = 0x1, .type =SUBCHANNEL_TYPE_CHSC, },
 	{ /* end of list */ },
@@ -125,6 +151,10 @@ static struct css_driver chsc_subchannel
 	.probe = chsc_subchannel_probe,
 	.remove = chsc_subchannel_remove,
 	.shutdown = chsc_subchannel_shutdown,
+	.prepare = chsc_subchannel_prepare,
+	.freeze = chsc_subchannel_freeze,
+	.thaw = chsc_subchannel_restore,
+	.restore = chsc_subchannel_restore,
 	.name = "chsc_subchannel",
 };
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 10/38] pm: chsc subchannel driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_chsc_sch.patch --]
[-- Type: text/plain, Size: 2016 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/cio/chsc_sch.c |   32 +++++++++++++++++++++++++++++++-
 1 file changed, 31 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/cio/chsc_sch.c
===================================================================
--- linux-2.6.orig/drivers/s390/cio/chsc_sch.c
+++ linux-2.6/drivers/s390/cio/chsc_sch.c
@@ -1,7 +1,8 @@
 /*
  * Driver for s390 chsc subchannels
  *
- * Copyright IBM Corp. 2008
+ * Copyright IBM Corp. 2008, 2009
+ *
  * Author(s): Cornelia Huck <cornelia.huck@de.ibm.com>
  *
  */
@@ -112,6 +113,31 @@ static void chsc_subchannel_shutdown(str
 	cio_disable_subchannel(sch);
 }
 
+static int chsc_subchannel_prepare(struct subchannel *sch)
+{
+	int cc;
+	struct schib schib;
+	/*
+	 * Don't allow suspend while the subchannel is not idle
+	 * since we don't have a way to clear the subchannel and
+	 * cannot disable it with a request running.
+	 */
+	cc = stsch(sch->schid, &schib);
+	if (!cc && scsw_stctl(&schib.scsw))
+		return -EAGAIN;
+	return 0;
+}
+
+static int chsc_subchannel_freeze(struct subchannel *sch)
+{
+	return cio_disable_subchannel(sch);
+}
+
+static int chsc_subchannel_restore(struct subchannel *sch)
+{
+	return cio_enable_subchannel(sch, (u32)(unsigned long)sch);
+}
+
 static struct css_device_id chsc_subchannel_ids[] = {
 	{ .match_flags = 0x1, .type =SUBCHANNEL_TYPE_CHSC, },
 	{ /* end of list */ },
@@ -125,6 +151,10 @@ static struct css_driver chsc_subchannel
 	.probe = chsc_subchannel_probe,
 	.remove = chsc_subchannel_remove,
 	.shutdown = chsc_subchannel_shutdown,
+	.prepare = chsc_subchannel_prepare,
+	.freeze = chsc_subchannel_freeze,
+	.thaw = chsc_subchannel_restore,
+	.restore = chsc_subchannel_restore,
 	.name = "chsc_subchannel",
 };
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 11/38] pm: dasd power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_dasd.patch --]
[-- Type: text/plain, Size: 13052 bytes --]

From: Stefan Haberland <stefan.haberland@de.ibm.com>

Introduce the power management callbacks to the dasd driver. On suspend
the dasd devices are stopped and removed from the focus of alias 
management.
On resume they are reinitialized by rereading the device characteristics
and adding the device to the alias management.
In case the device has gone away during suspend it will caught in the 
suspend state with stopped flag set to UNRESUMED. After it appears again
the restore function is called again.

Signed-off-by: Stefan Haberland <stefan.haberland@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/dasd.c      |  102 ++++++++++++++++++++++++++++++++++-
 drivers/s390/block/dasd_eckd.c |  118 ++++++++++++++++++++++++++++++++++++-----
 drivers/s390/block/dasd_fba.c  |    6 +-
 drivers/s390/block/dasd_int.h  |    9 ++-
 4 files changed, 216 insertions(+), 19 deletions(-)

Index: linux-2.6/drivers/s390/block/dasd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd.c
+++ linux-2.6/drivers/s390/block/dasd.c
@@ -5,8 +5,7 @@
  *		    Carsten Otte <Cotte@de.ibm.com>
  *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999-2001
- *
+ * Copyright IBM Corp. 1999, 2009
  */
 
 #define KMSG_COMPONENT "dasd"
@@ -61,6 +60,7 @@ static int dasd_flush_block_queue(struct
 static void dasd_device_tasklet(struct dasd_device *);
 static void dasd_block_tasklet(struct dasd_block *);
 static void do_kick_device(struct work_struct *);
+static void do_restore_device(struct work_struct *);
 static void dasd_return_cqr_cb(struct dasd_ccw_req *, void *);
 static void dasd_device_timeout(unsigned long);
 static void dasd_block_timeout(unsigned long);
@@ -109,6 +109,7 @@ struct dasd_device *dasd_alloc_device(vo
 	device->timer.function = dasd_device_timeout;
 	device->timer.data = (unsigned long) device;
 	INIT_WORK(&device->kick_work, do_kick_device);
+	INIT_WORK(&device->restore_device, do_restore_device);
 	device->state = DASD_STATE_NEW;
 	device->target = DASD_STATE_NEW;
 
@@ -512,6 +513,25 @@ void dasd_kick_device(struct dasd_device
 }
 
 /*
+ * dasd_restore_device will schedule a call do do_restore_device to the kernel
+ * event daemon.
+ */
+static void do_restore_device(struct work_struct *work)
+{
+	struct dasd_device *device = container_of(work, struct dasd_device,
+						  restore_device);
+	device->cdev->drv->restore(device->cdev);
+	dasd_put_device(device);
+}
+
+void dasd_restore_device(struct dasd_device *device)
+{
+	dasd_get_device(device);
+	/* queue call to dasd_restore_device to the kernel event daemon. */
+	schedule_work(&device->restore_device);
+}
+
+/*
  * Set the target state for a device and starts the state change.
  */
 void dasd_set_target_state(struct dasd_device *device, int target)
@@ -908,6 +928,12 @@ int dasd_start_IO(struct dasd_ccw_req *c
 		DBF_DEV_EVENT(DBF_DEBUG, device, "%s",
 			      "start_IO: -EIO device gone, retry");
 		break;
+	case -EINVAL:
+		/* most likely caused in power management context */
+		DBF_DEV_EVENT(DBF_DEBUG, device, "%s",
+			      "start_IO: -EINVAL device currently "
+			      "not accessible");
+		break;
 	default:
 		/* internal error 11 - unknown rc */
 		snprintf(errorstring, ERRORLENGTH, "11 %d", rc);
@@ -2413,6 +2439,12 @@ int dasd_generic_notify(struct ccw_devic
 		ret = 1;
 		break;
 	case CIO_OPER:
+		if (device->stopped & DASD_UNRESUMED_PM) {
+			device->stopped &= ~DASD_UNRESUMED_PM;
+			dasd_restore_device(device);
+			ret = 1;
+			break;
+		}
 		/* FIXME: add a sanity check. */
 		device->stopped &= ~DASD_STOPPED_DC_WAIT;
 		dasd_schedule_device_bh(device);
@@ -2425,6 +2457,72 @@ int dasd_generic_notify(struct ccw_devic
 	return ret;
 }
 
+int dasd_generic_pm_freeze(struct ccw_device *cdev)
+{
+	struct dasd_ccw_req *cqr, *n;
+	int rc;
+	struct list_head freeze_queue;
+	struct dasd_device *device = dasd_device_from_cdev(cdev);
+
+	if (IS_ERR(device))
+		return PTR_ERR(device);
+	/* disallow new I/O  */
+	device->stopped |= DASD_STOPPED_PM;
+	/* clear active requests */
+	INIT_LIST_HEAD(&freeze_queue);
+	spin_lock_irq(get_ccwdev_lock(cdev));
+	rc = 0;
+	list_for_each_entry_safe(cqr, n, &device->ccw_queue, devlist) {
+		/* Check status and move request to flush_queue */
+		if (cqr->status == DASD_CQR_IN_IO) {
+			rc = device->discipline->term_IO(cqr);
+			if (rc) {
+				/* unable to terminate requeust */
+				dev_err(&device->cdev->dev,
+					"Unable to terminate request %p\n",
+					cqr);
+				spin_unlock_irq(get_ccwdev_lock(cdev));
+				dasd_put_device(device);
+				return rc;
+			}
+		}
+		list_move_tail(&cqr->devlist, &freeze_queue);
+	}
+
+	spin_unlock_irq(get_ccwdev_lock(cdev));
+
+	list_for_each_entry_safe(cqr, n, &freeze_queue, devlist) {
+		wait_event(dasd_flush_wq,
+			   (cqr->status != DASD_CQR_CLEAR_PENDING));
+		if (cqr->status == DASD_CQR_CLEARED)
+			cqr->status = DASD_CQR_QUEUED;
+	}
+	/* move freeze_queue to start of the ccw_queue */
+	spin_lock_irq(get_ccwdev_lock(cdev));
+	list_splice_tail(&freeze_queue, &device->ccw_queue);
+	spin_unlock_irq(get_ccwdev_lock(cdev));
+
+	dasd_put_device(device);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dasd_generic_pm_freeze);
+
+int dasd_generic_restore_device(struct ccw_device *cdev)
+{
+	struct dasd_device *device = dasd_device_from_cdev(cdev);
+
+	if (IS_ERR(device))
+		return PTR_ERR(device);
+
+	dasd_schedule_device_bh(device);
+	if (device->block)
+		dasd_schedule_block_bh(device->block);
+
+	dasd_put_device(device);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dasd_generic_restore_device);
+
 static struct dasd_ccw_req *dasd_generic_build_rdc(struct dasd_device *device,
 						   void *rdc_buffer,
 						   int rdc_buffer_size,
Index: linux-2.6/drivers/s390/block/dasd_eckd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_eckd.c
+++ linux-2.6/drivers/s390/block/dasd_eckd.c
@@ -5,10 +5,9 @@
  *		    Carsten Otte <Cotte@de.ibm.com>
  *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ * Copyright IBM Corp. 1999, 2009
  * EMC Symmetrix ioctl Copyright EMC Corporation, 2008
  * Author.........: Nigel Hislop <hislop_nigel@emc.com>
- *
  */
 
 #define KMSG_COMPONENT "dasd"
@@ -104,17 +103,6 @@ dasd_eckd_set_online(struct ccw_device *
 	return dasd_generic_set_online(cdev, &dasd_eckd_discipline);
 }
 
-static struct ccw_driver dasd_eckd_driver = {
-	.name        = "dasd-eckd",
-	.owner       = THIS_MODULE,
-	.ids         = dasd_eckd_ids,
-	.probe       = dasd_eckd_probe,
-	.remove      = dasd_generic_remove,
-	.set_offline = dasd_generic_set_offline,
-	.set_online  = dasd_eckd_set_online,
-	.notify      = dasd_generic_notify,
-};
-
 static const int sizes_trk0[] = { 28, 148, 84 };
 #define LABEL_SIZE 140
 
@@ -3236,6 +3224,110 @@ static void dasd_eckd_dump_sense(struct 
 		dasd_eckd_dump_sense_ccw(device, req, irb);
 }
 
+int dasd_eckd_pm_freeze(struct ccw_device *cdev)
+{
+	struct dasd_device *device = dasd_device_from_cdev(cdev);
+	/*
+	 * the device should be disconnected from our LCU structure
+	 * on restore we will reconnect it and reread LCU specific
+	 * information like PAV support that might have changed
+	 */
+	dasd_alias_remove_device(device);
+	dasd_alias_disconnect_device_from_lcu(device);
+	dasd_put_device(device);
+	return dasd_generic_pm_freeze(cdev);
+}
+
+int dasd_eckd_restore_device(struct ccw_device *cdev)
+{
+	struct dasd_eckd_private *private;
+	struct dasd_device *device;
+	int is_known, rc;
+
+	device = dasd_device_from_cdev(cdev);
+
+	/* allow new IO again */
+	device->stopped &= ~DASD_STOPPED_PM;
+
+	private = (struct dasd_eckd_private *) device->private;
+
+	/* Read Configuration Data */
+	rc = dasd_eckd_read_conf(device);
+	if (rc)
+		goto out_err;
+
+	/* Generate device unique id and register in devmap */
+	rc = dasd_eckd_generate_uid(device, &private->uid);
+	if (rc)
+		goto out_err;
+	dasd_set_uid(device->cdev, &private->uid);
+
+	/* register lcu with alias handling, enable PAV if this is a new lcu */
+	is_known = dasd_alias_make_device_known_to_lcu(device);
+	if (is_known < 0)
+		return is_known;
+	if (!is_known) {
+		/* new lcu found */
+		rc = dasd_eckd_validate_server(device); /* will switch pav on */
+		if (rc)
+			goto out_err;
+	}
+
+	/* Read Feature Codes */
+	rc = dasd_eckd_read_features(device);
+	if (rc)
+		goto out_err;
+
+	/* Read Device Characteristics */
+	memset(&private->rdc_data, 0, sizeof(private->rdc_data));
+	rc = dasd_generic_read_dev_chars(device, "ECKD",
+					 &private->rdc_data, 64);
+	if (rc) {
+		DBF_EVENT(DBF_WARNING,
+			  "Read device characteristics failed, rc=%d for "
+			  "device: %s", rc, dev_name(&device->cdev->dev));
+		goto out_err;
+	}
+
+	/* add device to alias management */
+	dasd_alias_add_device(device);
+
+	dasd_put_device(device);
+
+	return dasd_generic_restore_device(cdev);
+
+out_err:
+	/*
+	 * we expect -EINVAL if device is not accessible during resume
+	 * we will keep the device in UNRESUMED stop state
+	 * CIO will inform us later weather to restore or to throw away
+	 * the device
+	 */
+	if (rc == -EINVAL) {
+		/* just disallow new IO and save state unresumed */
+		device->stopped |= DASD_UNRESUMED_PM;
+		return 0;
+	} else {
+		/* in any other case we do nothing for now */
+		dasd_put_device(device);
+
+		return 0;
+	}
+}
+
+static struct ccw_driver dasd_eckd_driver = {
+	.name        = "dasd-eckd",
+	.owner       = THIS_MODULE,
+	.ids         = dasd_eckd_ids,
+	.probe       = dasd_eckd_probe,
+	.remove      = dasd_generic_remove,
+	.set_offline = dasd_generic_set_offline,
+	.set_online  = dasd_eckd_set_online,
+	.notify      = dasd_generic_notify,
+	.freeze      = dasd_eckd_pm_freeze,
+	.thaw        = dasd_eckd_restore_device,
+	.restore     = dasd_eckd_restore_device,
+};
 
 /*
  * max_blocks is dependent on the amount of storage that is available
Index: linux-2.6/drivers/s390/block/dasd_fba.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_fba.c
+++ linux-2.6/drivers/s390/block/dasd_fba.c
@@ -2,8 +2,7 @@
  * File...........: linux/drivers/s390/block/dasd_fba.c
  * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
- *
+ * Copyright IBM Corp. 1999, 2009
  */
 
 #define KMSG_COMPONENT "dasd"
@@ -75,6 +74,9 @@ static struct ccw_driver dasd_fba_driver
 	.set_offline = dasd_generic_set_offline,
 	.set_online  = dasd_fba_set_online,
 	.notify      = dasd_generic_notify,
+	.freeze      = dasd_generic_pm_freeze,
+	.thaw        = dasd_generic_restore_device,
+	.restore     = dasd_generic_restore_device,
 };
 
 static void
Index: linux-2.6/drivers/s390/block/dasd_int.h
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_int.h
+++ linux-2.6/drivers/s390/block/dasd_int.h
@@ -4,8 +4,7 @@
  *		    Horst Hummel <Horst.Hummel@de.ibm.com>
  *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
- *
+ * Copyright IBM Corp. 1999, 2009
  */
 
 #ifndef DASD_INT_H
@@ -367,6 +366,7 @@ struct dasd_device {
 	atomic_t tasklet_scheduled;
         struct tasklet_struct tasklet;
 	struct work_struct kick_work;
+	struct work_struct restore_device;
 	struct timer_list timer;
 
 	debug_info_t *debug_area;
@@ -410,6 +410,8 @@ struct dasd_block {
 #define DASD_STOPPED_PENDING 4         /* long busy */
 #define DASD_STOPPED_DC_WAIT 8         /* disconnected, wait */
 #define DASD_STOPPED_SU      16        /* summary unit check handling */
+#define DASD_STOPPED_PM      32        /* pm state transition */
+#define DASD_UNRESUMED_PM    64        /* pm resume failed state */
 
 /* per device flags */
 #define DASD_FLAG_OFFLINE	3	/* device is in offline processing */
@@ -556,6 +558,7 @@ void dasd_free_block(struct dasd_block *
 void dasd_enable_device(struct dasd_device *);
 void dasd_set_target_state(struct dasd_device *, int);
 void dasd_kick_device(struct dasd_device *);
+void dasd_restore_device(struct dasd_device *);
 
 void dasd_add_request_head(struct dasd_ccw_req *);
 void dasd_add_request_tail(struct dasd_ccw_req *);
@@ -578,6 +581,8 @@ int dasd_generic_set_online(struct ccw_d
 int dasd_generic_set_offline (struct ccw_device *cdev);
 int dasd_generic_notify(struct ccw_device *, int);
 void dasd_generic_handle_state_change(struct dasd_device *);
+int dasd_generic_pm_freeze(struct ccw_device *);
+int dasd_generic_restore_device(struct ccw_device *);
 
 int dasd_generic_read_dev_chars(struct dasd_device *, char *, void *, int);
 char *dasd_get_sense(struct irb *);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 11/38] pm: dasd power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_dasd.patch --]
[-- Type: text/plain, Size: 13051 bytes --]

From: Stefan Haberland <stefan.haberland@de.ibm.com>

Introduce the power management callbacks to the dasd driver. On suspend
the dasd devices are stopped and removed from the focus of alias 
management.
On resume they are reinitialized by rereading the device characteristics
and adding the device to the alias management.
In case the device has gone away during suspend it will caught in the 
suspend state with stopped flag set to UNRESUMED. After it appears again
the restore function is called again.

Signed-off-by: Stefan Haberland <stefan.haberland@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/dasd.c      |  102 ++++++++++++++++++++++++++++++++++-
 drivers/s390/block/dasd_eckd.c |  118 ++++++++++++++++++++++++++++++++++++-----
 drivers/s390/block/dasd_fba.c  |    6 +-
 drivers/s390/block/dasd_int.h  |    9 ++-
 4 files changed, 216 insertions(+), 19 deletions(-)

Index: linux-2.6/drivers/s390/block/dasd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd.c
+++ linux-2.6/drivers/s390/block/dasd.c
@@ -5,8 +5,7 @@
  *		    Carsten Otte <Cotte@de.ibm.com>
  *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999-2001
- *
+ * Copyright IBM Corp. 1999, 2009
  */
 
 #define KMSG_COMPONENT "dasd"
@@ -61,6 +60,7 @@ static int dasd_flush_block_queue(struct
 static void dasd_device_tasklet(struct dasd_device *);
 static void dasd_block_tasklet(struct dasd_block *);
 static void do_kick_device(struct work_struct *);
+static void do_restore_device(struct work_struct *);
 static void dasd_return_cqr_cb(struct dasd_ccw_req *, void *);
 static void dasd_device_timeout(unsigned long);
 static void dasd_block_timeout(unsigned long);
@@ -109,6 +109,7 @@ struct dasd_device *dasd_alloc_device(vo
 	device->timer.function = dasd_device_timeout;
 	device->timer.data = (unsigned long) device;
 	INIT_WORK(&device->kick_work, do_kick_device);
+	INIT_WORK(&device->restore_device, do_restore_device);
 	device->state = DASD_STATE_NEW;
 	device->target = DASD_STATE_NEW;
 
@@ -512,6 +513,25 @@ void dasd_kick_device(struct dasd_device
 }
 
 /*
+ * dasd_restore_device will schedule a call do do_restore_device to the kernel
+ * event daemon.
+ */
+static void do_restore_device(struct work_struct *work)
+{
+	struct dasd_device *device = container_of(work, struct dasd_device,
+						  restore_device);
+	device->cdev->drv->restore(device->cdev);
+	dasd_put_device(device);
+}
+
+void dasd_restore_device(struct dasd_device *device)
+{
+	dasd_get_device(device);
+	/* queue call to dasd_restore_device to the kernel event daemon. */
+	schedule_work(&device->restore_device);
+}
+
+/*
  * Set the target state for a device and starts the state change.
  */
 void dasd_set_target_state(struct dasd_device *device, int target)
@@ -908,6 +928,12 @@ int dasd_start_IO(struct dasd_ccw_req *c
 		DBF_DEV_EVENT(DBF_DEBUG, device, "%s",
 			      "start_IO: -EIO device gone, retry");
 		break;
+	case -EINVAL:
+		/* most likely caused in power management context */
+		DBF_DEV_EVENT(DBF_DEBUG, device, "%s",
+			      "start_IO: -EINVAL device currently "
+			      "not accessible");
+		break;
 	default:
 		/* internal error 11 - unknown rc */
 		snprintf(errorstring, ERRORLENGTH, "11 %d", rc);
@@ -2413,6 +2439,12 @@ int dasd_generic_notify(struct ccw_devic
 		ret = 1;
 		break;
 	case CIO_OPER:
+		if (device->stopped & DASD_UNRESUMED_PM) {
+			device->stopped &= ~DASD_UNRESUMED_PM;
+			dasd_restore_device(device);
+			ret = 1;
+			break;
+		}
 		/* FIXME: add a sanity check. */
 		device->stopped &= ~DASD_STOPPED_DC_WAIT;
 		dasd_schedule_device_bh(device);
@@ -2425,6 +2457,72 @@ int dasd_generic_notify(struct ccw_devic
 	return ret;
 }
 
+int dasd_generic_pm_freeze(struct ccw_device *cdev)
+{
+	struct dasd_ccw_req *cqr, *n;
+	int rc;
+	struct list_head freeze_queue;
+	struct dasd_device *device = dasd_device_from_cdev(cdev);
+
+	if (IS_ERR(device))
+		return PTR_ERR(device);
+	/* disallow new I/O  */
+	device->stopped |= DASD_STOPPED_PM;
+	/* clear active requests */
+	INIT_LIST_HEAD(&freeze_queue);
+	spin_lock_irq(get_ccwdev_lock(cdev));
+	rc = 0;
+	list_for_each_entry_safe(cqr, n, &device->ccw_queue, devlist) {
+		/* Check status and move request to flush_queue */
+		if (cqr->status == DASD_CQR_IN_IO) {
+			rc = device->discipline->term_IO(cqr);
+			if (rc) {
+				/* unable to terminate requeust */
+				dev_err(&device->cdev->dev,
+					"Unable to terminate request %p\n",
+					cqr);
+				spin_unlock_irq(get_ccwdev_lock(cdev));
+				dasd_put_device(device);
+				return rc;
+			}
+		}
+		list_move_tail(&cqr->devlist, &freeze_queue);
+	}
+
+	spin_unlock_irq(get_ccwdev_lock(cdev));
+
+	list_for_each_entry_safe(cqr, n, &freeze_queue, devlist) {
+		wait_event(dasd_flush_wq,
+			   (cqr->status != DASD_CQR_CLEAR_PENDING));
+		if (cqr->status == DASD_CQR_CLEARED)
+			cqr->status = DASD_CQR_QUEUED;
+	}
+	/* move freeze_queue to start of the ccw_queue */
+	spin_lock_irq(get_ccwdev_lock(cdev));
+	list_splice_tail(&freeze_queue, &device->ccw_queue);
+	spin_unlock_irq(get_ccwdev_lock(cdev));
+
+	dasd_put_device(device);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dasd_generic_pm_freeze);
+
+int dasd_generic_restore_device(struct ccw_device *cdev)
+{
+	struct dasd_device *device = dasd_device_from_cdev(cdev);
+
+	if (IS_ERR(device))
+		return PTR_ERR(device);
+
+	dasd_schedule_device_bh(device);
+	if (device->block)
+		dasd_schedule_block_bh(device->block);
+
+	dasd_put_device(device);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dasd_generic_restore_device);
+
 static struct dasd_ccw_req *dasd_generic_build_rdc(struct dasd_device *device,
 						   void *rdc_buffer,
 						   int rdc_buffer_size,
Index: linux-2.6/drivers/s390/block/dasd_eckd.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_eckd.c
+++ linux-2.6/drivers/s390/block/dasd_eckd.c
@@ -5,10 +5,9 @@
  *		    Carsten Otte <Cotte@de.ibm.com>
  *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ * Copyright IBM Corp. 1999, 2009
  * EMC Symmetrix ioctl Copyright EMC Corporation, 2008
  * Author.........: Nigel Hislop <hislop_nigel@emc.com>
- *
  */
 
 #define KMSG_COMPONENT "dasd"
@@ -104,17 +103,6 @@ dasd_eckd_set_online(struct ccw_device *
 	return dasd_generic_set_online(cdev, &dasd_eckd_discipline);
 }
 
-static struct ccw_driver dasd_eckd_driver = {
-	.name        = "dasd-eckd",
-	.owner       = THIS_MODULE,
-	.ids         = dasd_eckd_ids,
-	.probe       = dasd_eckd_probe,
-	.remove      = dasd_generic_remove,
-	.set_offline = dasd_generic_set_offline,
-	.set_online  = dasd_eckd_set_online,
-	.notify      = dasd_generic_notify,
-};
-
 static const int sizes_trk0[] = { 28, 148, 84 };
 #define LABEL_SIZE 140
 
@@ -3236,6 +3224,110 @@ static void dasd_eckd_dump_sense(struct 
 		dasd_eckd_dump_sense_ccw(device, req, irb);
 }
 
+int dasd_eckd_pm_freeze(struct ccw_device *cdev)
+{
+	struct dasd_device *device = dasd_device_from_cdev(cdev);
+	/*
+	 * the device should be disconnected from our LCU structure
+	 * on restore we will reconnect it and reread LCU specific
+	 * information like PAV support that might have changed
+	 */
+	dasd_alias_remove_device(device);
+	dasd_alias_disconnect_device_from_lcu(device);
+	dasd_put_device(device);
+	return dasd_generic_pm_freeze(cdev);
+}
+
+int dasd_eckd_restore_device(struct ccw_device *cdev)
+{
+	struct dasd_eckd_private *private;
+	struct dasd_device *device;
+	int is_known, rc;
+
+	device = dasd_device_from_cdev(cdev);
+
+	/* allow new IO again */
+	device->stopped &= ~DASD_STOPPED_PM;
+
+	private = (struct dasd_eckd_private *) device->private;
+
+	/* Read Configuration Data */
+	rc = dasd_eckd_read_conf(device);
+	if (rc)
+		goto out_err;
+
+	/* Generate device unique id and register in devmap */
+	rc = dasd_eckd_generate_uid(device, &private->uid);
+	if (rc)
+		goto out_err;
+	dasd_set_uid(device->cdev, &private->uid);
+
+	/* register lcu with alias handling, enable PAV if this is a new lcu */
+	is_known = dasd_alias_make_device_known_to_lcu(device);
+	if (is_known < 0)
+		return is_known;
+	if (!is_known) {
+		/* new lcu found */
+		rc = dasd_eckd_validate_server(device); /* will switch pav on */
+		if (rc)
+			goto out_err;
+	}
+
+	/* Read Feature Codes */
+	rc = dasd_eckd_read_features(device);
+	if (rc)
+		goto out_err;
+
+	/* Read Device Characteristics */
+	memset(&private->rdc_data, 0, sizeof(private->rdc_data));
+	rc = dasd_generic_read_dev_chars(device, "ECKD",
+					 &private->rdc_data, 64);
+	if (rc) {
+		DBF_EVENT(DBF_WARNING,
+			  "Read device characteristics failed, rc=%d for "
+			  "device: %s", rc, dev_name(&device->cdev->dev));
+		goto out_err;
+	}
+
+	/* add device to alias management */
+	dasd_alias_add_device(device);
+
+	dasd_put_device(device);
+
+	return dasd_generic_restore_device(cdev);
+
+out_err:
+	/*
+	 * we expect -EINVAL if device is not accessible during resume
+	 * we will keep the device in UNRESUMED stop state
+	 * CIO will inform us later weather to restore or to throw away
+	 * the device
+	 */
+	if (rc == -EINVAL) {
+		/* just disallow new IO and save state unresumed */
+		device->stopped |= DASD_UNRESUMED_PM;
+		return 0;
+	} else {
+		/* in any other case we do nothing for now */
+		dasd_put_device(device);
+
+		return 0;
+	}
+}
+
+static struct ccw_driver dasd_eckd_driver = {
+	.name        = "dasd-eckd",
+	.owner       = THIS_MODULE,
+	.ids         = dasd_eckd_ids,
+	.probe       = dasd_eckd_probe,
+	.remove      = dasd_generic_remove,
+	.set_offline = dasd_generic_set_offline,
+	.set_online  = dasd_eckd_set_online,
+	.notify      = dasd_generic_notify,
+	.freeze      = dasd_eckd_pm_freeze,
+	.thaw        = dasd_eckd_restore_device,
+	.restore     = dasd_eckd_restore_device,
+};
 
 /*
  * max_blocks is dependent on the amount of storage that is available
Index: linux-2.6/drivers/s390/block/dasd_fba.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_fba.c
+++ linux-2.6/drivers/s390/block/dasd_fba.c
@@ -2,8 +2,7 @@
  * File...........: linux/drivers/s390/block/dasd_fba.c
  * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
- *
+ * Copyright IBM Corp. 1999, 2009
  */
 
 #define KMSG_COMPONENT "dasd"
@@ -75,6 +74,9 @@ static struct ccw_driver dasd_fba_driver
 	.set_offline = dasd_generic_set_offline,
 	.set_online  = dasd_fba_set_online,
 	.notify      = dasd_generic_notify,
+	.freeze      = dasd_generic_pm_freeze,
+	.thaw        = dasd_generic_restore_device,
+	.restore     = dasd_generic_restore_device,
 };
 
 static void
Index: linux-2.6/drivers/s390/block/dasd_int.h
===================================================================
--- linux-2.6.orig/drivers/s390/block/dasd_int.h
+++ linux-2.6/drivers/s390/block/dasd_int.h
@@ -4,8 +4,7 @@
  *		    Horst Hummel <Horst.Hummel@de.ibm.com>
  *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
- *
+ * Copyright IBM Corp. 1999, 2009
  */
 
 #ifndef DASD_INT_H
@@ -367,6 +366,7 @@ struct dasd_device {
 	atomic_t tasklet_scheduled;
         struct tasklet_struct tasklet;
 	struct work_struct kick_work;
+	struct work_struct restore_device;
 	struct timer_list timer;
 
 	debug_info_t *debug_area;
@@ -410,6 +410,8 @@ struct dasd_block {
 #define DASD_STOPPED_PENDING 4         /* long busy */
 #define DASD_STOPPED_DC_WAIT 8         /* disconnected, wait */
 #define DASD_STOPPED_SU      16        /* summary unit check handling */
+#define DASD_STOPPED_PM      32        /* pm state transition */
+#define DASD_UNRESUMED_PM    64        /* pm resume failed state */
 
 /* per device flags */
 #define DASD_FLAG_OFFLINE	3	/* device is in offline processing */
@@ -556,6 +558,7 @@ void dasd_free_block(struct dasd_block *
 void dasd_enable_device(struct dasd_device *);
 void dasd_set_target_state(struct dasd_device *, int);
 void dasd_kick_device(struct dasd_device *);
+void dasd_restore_device(struct dasd_device *);
 
 void dasd_add_request_head(struct dasd_ccw_req *);
 void dasd_add_request_tail(struct dasd_ccw_req *);
@@ -578,6 +581,8 @@ int dasd_generic_set_online(struct ccw_d
 int dasd_generic_set_offline (struct ccw_device *cdev);
 int dasd_generic_notify(struct ccw_device *, int);
 void dasd_generic_handle_state_change(struct dasd_device *);
+int dasd_generic_pm_freeze(struct ccw_device *);
+int dasd_generic_restore_device(struct ccw_device *);
 
 int dasd_generic_read_dev_chars(struct dasd_device *, char *, void *, int);
 char *dasd_get_sense(struct irb *);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 12/38] pm: add kernel_page_present

[Martin Schwidefsky]

[-- Attachment #1: pm_kernel_page_present.patch --]
[-- Type: text/plain, Size: 1390 bytes --]

From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>

Fix the following build failure caused by make allyesconfig using
CONFIG_HIBERNATION and CONFIG_DEBUG_PAGEALLOC

kernel/built-in.o: In function `saveable_page':
kernel/power/snapshot.c:897: undefined reference to `kernel_page_present'
kernel/built-in.o: In function `safe_copy_page':
kernel/power/snapshot.c:948: undefined reference to `kernel_page_present'
make: *** [.tmp_vmlinux1] Error 1

Signed-off-by: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/mm/pgtable.c |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

Index: linux-2.6/arch/s390/mm/pgtable.c
===================================================================
--- linux-2.6.orig/arch/s390/mm/pgtable.c
+++ linux-2.6/arch/s390/mm/pgtable.c
@@ -313,3 +313,22 @@ int s390_enable_sie(void)
 	return 0;
 }
 EXPORT_SYMBOL_GPL(s390_enable_sie);
+
+#ifdef CONFIG_DEBUG_PAGEALLOC
+#ifdef CONFIG_HIBERNATION
+bool kernel_page_present(struct page *page)
+{
+        unsigned long addr;
+	int cc;
+
+	addr = page_to_phys(page);
+	asm("lra %1,0(%1)\n"
+	    "ipm %0\n"
+            "srl %0,28"
+	    :"=d"(cc),"+a"(addr)::"cc");
+	return cc == 0;
+}
+
+#endif /* CONFIG_HIBERNATION */
+#endif /* CONFIG_DEBUG_PAGEALLOC */
+

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 12/38] pm: add kernel_page_present

[Martin Schwidefsky]

[-- Attachment #1: pm_kernel_page_present.patch --]
[-- Type: text/plain, Size: 1389 bytes --]

From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>

Fix the following build failure caused by make allyesconfig using
CONFIG_HIBERNATION and CONFIG_DEBUG_PAGEALLOC

kernel/built-in.o: In function `saveable_page':
kernel/power/snapshot.c:897: undefined reference to `kernel_page_present'
kernel/built-in.o: In function `safe_copy_page':
kernel/power/snapshot.c:948: undefined reference to `kernel_page_present'
make: *** [.tmp_vmlinux1] Error 1

Signed-off-by: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/mm/pgtable.c |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

Index: linux-2.6/arch/s390/mm/pgtable.c
===================================================================
--- linux-2.6.orig/arch/s390/mm/pgtable.c
+++ linux-2.6/arch/s390/mm/pgtable.c
@@ -313,3 +313,22 @@ int s390_enable_sie(void)
 	return 0;
 }
 EXPORT_SYMBOL_GPL(s390_enable_sie);
+
+#ifdef CONFIG_DEBUG_PAGEALLOC
+#ifdef CONFIG_HIBERNATION
+bool kernel_page_present(struct page *page)
+{
+        unsigned long addr;
+	int cc;
+
+	addr = page_to_phys(page);
+	asm("lra %1,0(%1)\n"
+	    "ipm %0\n"
+            "srl %0,28"
+	    :"=d"(cc),"+a"(addr)::"cc");
+	return cc == 0;
+}
+
+#endif /* CONFIG_HIBERNATION */
+#endif /* CONFIG_DEBUG_PAGEALLOC */
+

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 13/38] pm: xpram driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_xpram.patch --]
[-- Type: text/plain, Size: 4881 bytes --]

From: Michael Holzheu <holzheu@de.ibm.com>

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/xpram.c |  129 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 126 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/s390/block/xpram.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/xpram.c
+++ linux-2.6/drivers/s390/block/xpram.c
@@ -39,7 +39,10 @@
 #include <linux/hdreg.h>  /* HDIO_GETGEO */
 #include <linux/sysdev.h>
 #include <linux/bio.h>
+#include <linux/suspend.h>
+#include <linux/platform_device.h>
 #include <asm/uaccess.h>
+#include <asm/checksum.h>
 
 #define XPRAM_NAME	"xpram"
 #define XPRAM_DEVS	1	/* one partition */
@@ -48,6 +51,7 @@
 typedef struct {
 	unsigned int	size;		/* size of xpram segment in pages */
 	unsigned int	offset;		/* start page of xpram segment */
+	unsigned int	csum;		/* partition checksum for suspend */
 } xpram_device_t;
 
 static xpram_device_t xpram_devices[XPRAM_MAX_DEVS];
@@ -138,7 +142,7 @@ static long xpram_page_out (unsigned lon
 /*
  * Check if xpram is available.
  */
-static int __init xpram_present(void)
+static int xpram_present(void)
 {
 	unsigned long mem_page;
 	int rc;
@@ -154,7 +158,7 @@ static int __init xpram_present(void)
 /*
  * Return index of the last available xpram page.
  */
-static unsigned long __init xpram_highest_page_index(void)
+static unsigned long xpram_highest_page_index(void)
 {
 	unsigned int page_index, add_bit;
 	unsigned long mem_page;
@@ -383,6 +387,106 @@ out:
 }
 
 /*
+ * Save checksums for all partitions.
+ */
+static int xpram_save_checksums(void)
+{
+	unsigned long mem_page;
+	int rc, i;
+
+	rc = 0;
+	mem_page = (unsigned long) __get_free_page(GFP_KERNEL);
+	if (!mem_page)
+		return -ENOMEM;
+	for (i = 0; i < xpram_devs; i++) {
+		rc = xpram_page_in(mem_page, xpram_devices[i].offset);
+		if (rc)
+			goto fail;
+		xpram_devices[i].csum = csum_partial((const void *) mem_page,
+						     PAGE_SIZE, 0);
+	}
+fail:
+	free_page(mem_page);
+	return rc ? -ENXIO : 0;
+}
+
+/*
+ * Verify checksums for all partitions.
+ */
+static int xpram_validate_checksums(void)
+{
+	unsigned long mem_page;
+	unsigned int csum;
+	int rc, i;
+
+	rc = 0;
+	mem_page = (unsigned long) __get_free_page(GFP_KERNEL);
+	if (!mem_page)
+		return -ENOMEM;
+	for (i = 0; i < xpram_devs; i++) {
+		rc = xpram_page_in(mem_page, xpram_devices[i].offset);
+		if (rc)
+			goto fail;
+		csum = csum_partial((const void *) mem_page, PAGE_SIZE, 0);
+		if (xpram_devices[i].csum != csum) {
+			rc = -EINVAL;
+			goto fail;
+		}
+	}
+fail:
+	free_page(mem_page);
+	return rc ? -ENXIO : 0;
+}
+
+/*
+ * Resume failed: Print error message and call panic.
+ */
+static void xpram_resume_error(const char *message)
+{
+	pr_err("Resume error: %s\n", message);
+	panic("xpram resume error\n");
+}
+
+/*
+ * Check if xpram setup changed between suspend and resume.
+ */
+static int xpram_restore(struct device *dev)
+{
+	if (!xpram_pages)
+		return 0;
+	if (xpram_present() != 0)
+		xpram_resume_error("xpram disappeared");
+	if (xpram_pages != xpram_highest_page_index() + 1)
+		xpram_resume_error("Size of xpram changed");
+	if (xpram_validate_checksums())
+		xpram_resume_error("Data of xpram changed");
+	return 0;
+}
+
+/*
+ * Save necessary state in suspend.
+ */
+static int xpram_freeze(struct device *dev)
+{
+	return xpram_save_checksums();
+}
+
+static struct dev_pm_ops xpram_pm_ops = {
+	.freeze		= xpram_freeze,
+	.restore	= xpram_restore,
+};
+
+static struct platform_driver xpram_pdrv = {
+	.driver = {
+		.name	= XPRAM_NAME,
+		.owner	= THIS_MODULE,
+		.pm	= &xpram_pm_ops,
+	},
+};
+
+static struct platform_device *xpram_pdev;
+
+/*
  * Finally, the init/exit functions.
  */
 static void __exit xpram_exit(void)
@@ -394,6 +498,8 @@ static void __exit xpram_exit(void)
 		put_disk(xpram_disks[i]);
 	}
 	unregister_blkdev(XPRAM_MAJOR, XPRAM_NAME);
+	platform_device_unregister(xpram_pdev);
+	platform_driver_unregister(&xpram_pdrv);
 }
 
 static int __init xpram_init(void)
@@ -411,7 +517,24 @@ static int __init xpram_init(void)
 	rc = xpram_setup_sizes(xpram_pages);
 	if (rc)
 		return rc;
-	return xpram_setup_blkdev();
+	rc = platform_driver_register(&xpram_pdrv);
+	if (rc)
+		return rc;
+	xpram_pdev = platform_device_register_simple(XPRAM_NAME, -1, NULL, 0);
+	if (IS_ERR(xpram_pdev)) {
+		rc = PTR_ERR(xpram_pdev);
+		goto fail_platform_driver_unregister;
+	}
+	rc = xpram_setup_blkdev();
+	if (rc)
+		goto fail_platform_device_unregister;
+	return 0;
+
+fail_platform_device_unregister:
+	platform_device_unregister(xpram_pdev);
+fail_platform_driver_unregister:
+	platform_driver_unregister(&xpram_pdrv);
+	return rc;
 }
 
 module_init(xpram_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 13/38] pm: xpram driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_xpram.patch --]
[-- Type: text/plain, Size: 4880 bytes --]

From: Michael Holzheu <holzheu@de.ibm.com>

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/xpram.c |  129 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 126 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/s390/block/xpram.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/xpram.c
+++ linux-2.6/drivers/s390/block/xpram.c
@@ -39,7 +39,10 @@
 #include <linux/hdreg.h>  /* HDIO_GETGEO */
 #include <linux/sysdev.h>
 #include <linux/bio.h>
+#include <linux/suspend.h>
+#include <linux/platform_device.h>
 #include <asm/uaccess.h>
+#include <asm/checksum.h>
 
 #define XPRAM_NAME	"xpram"
 #define XPRAM_DEVS	1	/* one partition */
@@ -48,6 +51,7 @@
 typedef struct {
 	unsigned int	size;		/* size of xpram segment in pages */
 	unsigned int	offset;		/* start page of xpram segment */
+	unsigned int	csum;		/* partition checksum for suspend */
 } xpram_device_t;
 
 static xpram_device_t xpram_devices[XPRAM_MAX_DEVS];
@@ -138,7 +142,7 @@ static long xpram_page_out (unsigned lon
 /*
  * Check if xpram is available.
  */
-static int __init xpram_present(void)
+static int xpram_present(void)
 {
 	unsigned long mem_page;
 	int rc;
@@ -154,7 +158,7 @@ static int __init xpram_present(void)
 /*
  * Return index of the last available xpram page.
  */
-static unsigned long __init xpram_highest_page_index(void)
+static unsigned long xpram_highest_page_index(void)
 {
 	unsigned int page_index, add_bit;
 	unsigned long mem_page;
@@ -383,6 +387,106 @@ out:
 }
 
 /*
+ * Save checksums for all partitions.
+ */
+static int xpram_save_checksums(void)
+{
+	unsigned long mem_page;
+	int rc, i;
+
+	rc = 0;
+	mem_page = (unsigned long) __get_free_page(GFP_KERNEL);
+	if (!mem_page)
+		return -ENOMEM;
+	for (i = 0; i < xpram_devs; i++) {
+		rc = xpram_page_in(mem_page, xpram_devices[i].offset);
+		if (rc)
+			goto fail;
+		xpram_devices[i].csum = csum_partial((const void *) mem_page,
+						     PAGE_SIZE, 0);
+	}
+fail:
+	free_page(mem_page);
+	return rc ? -ENXIO : 0;
+}
+
+/*
+ * Verify checksums for all partitions.
+ */
+static int xpram_validate_checksums(void)
+{
+	unsigned long mem_page;
+	unsigned int csum;
+	int rc, i;
+
+	rc = 0;
+	mem_page = (unsigned long) __get_free_page(GFP_KERNEL);
+	if (!mem_page)
+		return -ENOMEM;
+	for (i = 0; i < xpram_devs; i++) {
+		rc = xpram_page_in(mem_page, xpram_devices[i].offset);
+		if (rc)
+			goto fail;
+		csum = csum_partial((const void *) mem_page, PAGE_SIZE, 0);
+		if (xpram_devices[i].csum != csum) {
+			rc = -EINVAL;
+			goto fail;
+		}
+	}
+fail:
+	free_page(mem_page);
+	return rc ? -ENXIO : 0;
+}
+
+/*
+ * Resume failed: Print error message and call panic.
+ */
+static void xpram_resume_error(const char *message)
+{
+	pr_err("Resume error: %s\n", message);
+	panic("xpram resume error\n");
+}
+
+/*
+ * Check if xpram setup changed between suspend and resume.
+ */
+static int xpram_restore(struct device *dev)
+{
+	if (!xpram_pages)
+		return 0;
+	if (xpram_present() != 0)
+		xpram_resume_error("xpram disappeared");
+	if (xpram_pages != xpram_highest_page_index() + 1)
+		xpram_resume_error("Size of xpram changed");
+	if (xpram_validate_checksums())
+		xpram_resume_error("Data of xpram changed");
+	return 0;
+}
+
+/*
+ * Save necessary state in suspend.
+ */
+static int xpram_freeze(struct device *dev)
+{
+	return xpram_save_checksums();
+}
+
+static struct dev_pm_ops xpram_pm_ops = {
+	.freeze		= xpram_freeze,
+	.restore	= xpram_restore,
+};
+
+static struct platform_driver xpram_pdrv = {
+	.driver = {
+		.name	= XPRAM_NAME,
+		.owner	= THIS_MODULE,
+		.pm	= &xpram_pm_ops,
+	},
+};
+
+static struct platform_device *xpram_pdev;
+
+/*
  * Finally, the init/exit functions.
  */
 static void __exit xpram_exit(void)
@@ -394,6 +498,8 @@ static void __exit xpram_exit(void)
 		put_disk(xpram_disks[i]);
 	}
 	unregister_blkdev(XPRAM_MAJOR, XPRAM_NAME);
+	platform_device_unregister(xpram_pdev);
+	platform_driver_unregister(&xpram_pdrv);
 }
 
 static int __init xpram_init(void)
@@ -411,7 +517,24 @@ static int __init xpram_init(void)
 	rc = xpram_setup_sizes(xpram_pages);
 	if (rc)
 		return rc;
-	return xpram_setup_blkdev();
+	rc = platform_driver_register(&xpram_pdrv);
+	if (rc)
+		return rc;
+	xpram_pdev = platform_device_register_simple(XPRAM_NAME, -1, NULL, 0);
+	if (IS_ERR(xpram_pdev)) {
+		rc = PTR_ERR(xpram_pdev);
+		goto fail_platform_driver_unregister;
+	}
+	rc = xpram_setup_blkdev();
+	if (rc)
+		goto fail_platform_device_unregister;
+	return 0;
+
+fail_platform_device_unregister:
+	platform_device_unregister(xpram_pdev);
+fail_platform_driver_unregister:
+	platform_driver_unregister(&xpram_pdrv);
+	return rc;
 }
 
 module_init(xpram_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 14/38] cio: force console function

[Martin Schwidefsky]

[-- Attachment #1: pm_force_console.patch --]
[-- Type: text/plain, Size: 1838 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

If something goes wrong in a suspend / resume cycle a ccw based console
if very likely in the suspended state and cannot print anything.
Introduce ccw_device_force_console to force the wake up of the console
device to be able to print the oops message. The console device drivers
should use this function only if the system paniced.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/include/asm/ccwdev.h |    1 +
 drivers/s390/cio/device.c      |    8 ++++++++
 2 files changed, 9 insertions(+)

Index: linux-2.5/drivers/s390/cio/device.c
===================================================================
--- linux-2.5.orig/drivers/s390/cio/device.c
+++ linux-2.5/drivers/s390/cio/device.c
@@ -1785,6 +1785,15 @@ ccw_device_probe_console(void)
 	return &console_cdev;
 }
 
+static int ccw_device_pm_restore(struct device *dev);
+
+int ccw_device_force_console(void)
+{
+	if (!console_cdev_in_use)
+		return -ENODEV;
+	return ccw_device_pm_restore(&console_cdev.dev);
+}
+EXPORT_SYMBOL_GPL(ccw_device_force_console);
 
 const char *cio_get_console_cdev_name(struct subchannel *sch)
 {
Index: linux-2.5/arch/s390/include/asm/ccwdev.h
===================================================================
--- linux-2.5.orig/arch/s390/include/asm/ccwdev.h
+++ linux-2.5/arch/s390/include/asm/ccwdev.h
@@ -192,6 +192,7 @@ extern void ccw_device_get_id(struct ccw
 #define to_ccwdrv(n) container_of(n, struct ccw_driver, driver)
 
 extern struct ccw_device *ccw_device_probe_console(void);
+extern int ccw_device_force_console(void);
 
 // FIXME: these have to go
 extern int _ccw_device_get_subchannel_number(struct ccw_device *);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 14/38] cio: force console function

[Martin Schwidefsky]

[-- Attachment #1: pm_force_console.patch --]
[-- Type: text/plain, Size: 1837 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

If something goes wrong in a suspend / resume cycle a ccw based console
if very likely in the suspended state and cannot print anything.
Introduce ccw_device_force_console to force the wake up of the console
device to be able to print the oops message. The console device drivers
should use this function only if the system paniced.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/include/asm/ccwdev.h |    1 +
 drivers/s390/cio/device.c      |    8 ++++++++
 2 files changed, 9 insertions(+)

Index: linux-2.5/drivers/s390/cio/device.c
===================================================================
--- linux-2.5.orig/drivers/s390/cio/device.c
+++ linux-2.5/drivers/s390/cio/device.c
@@ -1785,6 +1785,15 @@ ccw_device_probe_console(void)
 	return &console_cdev;
 }
 
+static int ccw_device_pm_restore(struct device *dev);
+
+int ccw_device_force_console(void)
+{
+	if (!console_cdev_in_use)
+		return -ENODEV;
+	return ccw_device_pm_restore(&console_cdev.dev);
+}
+EXPORT_SYMBOL_GPL(ccw_device_force_console);
 
 const char *cio_get_console_cdev_name(struct subchannel *sch)
 {
Index: linux-2.5/arch/s390/include/asm/ccwdev.h
===================================================================
--- linux-2.5.orig/arch/s390/include/asm/ccwdev.h
+++ linux-2.5/arch/s390/include/asm/ccwdev.h
@@ -192,6 +192,7 @@ extern void ccw_device_get_id(struct ccw
 #define to_ccwdrv(n) container_of(n, struct ccw_driver, driver)
 
 extern struct ccw_device *ccw_device_probe_console(void);
+extern int ccw_device_force_console(void);
 
 // FIXME: these have to go
 extern int _ccw_device_get_subchannel_number(struct ccw_device *);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 15/38] pm: con3215 power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_3215.patch --]
[-- Type: text/plain, Size: 15645 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Introduce the power management callbacks to the 3215 console. On suspend
the console buffer is flushed to the 3215 device to have an empty console
buffer. Printks done while the 3215 device is suspended are buffered in
the 64K buffer of the 3215 device. If the buffer is full new messages will
push out the oldest messages to make room for the most recent message.
On resume the buffered messages are printed. If the system panics before
the 3215 device is resumed ccw_device_force_console is used to get the
console working again.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/con3215.c |  205 +++++++++++++++++++++++++-------------------
 1 file changed, 120 insertions(+), 85 deletions(-)

Index: linux-2.6/drivers/s390/char/con3215.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/con3215.c
+++ linux-2.6/drivers/s390/char/con3215.c
@@ -1,14 +1,12 @@
 /*
- *  drivers/s390/char/con3215.c
- *    3215 line mode terminal driver.
+ * 3215 line mode terminal driver.
  *
- *  S390 version
- *    Copyright (C) 1999,2000 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
+ * Copyright IBM Corp. 1999, 2009
+ * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
  *
- *  Updated:
- *   Aug-2000: Added tab support
- *	       Dan Morrison, IBM Corporation (dmorriso@cse.buffalo.edu)
+ * Updated:
+ *  Aug-2000: Added tab support
+ *	      Dan Morrison, IBM Corporation <dmorriso@cse.buffalo.edu>
  */
 
 #include <linux/module.h>
@@ -56,6 +54,7 @@
 #define RAW3215_CLOSING	    32	      /* set while in close process */
 #define RAW3215_TIMER_RUNS  64	      /* set if the output delay timer is on */
 #define RAW3215_FLUSHING    128	      /* set to flush buffer (no delay) */
+#define RAW3215_FROZEN      256	      /* set if 3215 is frozen for suspend */
 
 #define TAB_STOP_SIZE	    8	      /* tab stop size */
 
@@ -111,8 +110,8 @@ static struct tty_driver *tty3215_driver
 /*
  * Get a request structure from the free list
  */
-static inline struct raw3215_req *
-raw3215_alloc_req(void) {
+static inline struct raw3215_req *raw3215_alloc_req(void)
+{
 	struct raw3215_req *req;
 	unsigned long flags;
 
@@ -126,8 +125,8 @@ raw3215_alloc_req(void) {
 /*
  * Put a request structure back to the free list
  */
-static inline void
-raw3215_free_req(struct raw3215_req *req) {
+static inline void raw3215_free_req(struct raw3215_req *req)
+{
 	unsigned long flags;
 
 	if (req->type == RAW3215_FREE)
@@ -145,8 +144,7 @@ raw3215_free_req(struct raw3215_req *req
  * because a 3215 terminal won't accept a new read before the old one is
  * completed.
  */
-static void
-raw3215_mk_read_req(struct raw3215_info *raw)
+static void raw3215_mk_read_req(struct raw3215_info *raw)
 {
 	struct raw3215_req *req;
 	struct ccw1 *ccw;
@@ -174,8 +172,7 @@ raw3215_mk_read_req(struct raw3215_info 
  * buffer to the 3215 device. If a queued write exists it is replaced by
  * the new, probably lengthened request.
  */
-static void
-raw3215_mk_write_req(struct raw3215_info *raw)
+static void raw3215_mk_write_req(struct raw3215_info *raw)
 {
 	struct raw3215_req *req;
 	struct ccw1 *ccw;
@@ -251,8 +248,7 @@ raw3215_mk_write_req(struct raw3215_info
 /*
  * Start a read or a write request
  */
-static void
-raw3215_start_io(struct raw3215_info *raw)
+static void raw3215_start_io(struct raw3215_info *raw)
 {
 	struct raw3215_req *req;
 	int res;
@@ -290,8 +286,7 @@ raw3215_start_io(struct raw3215_info *ra
 /*
  * Function to start a delayed output after RAW3215_TIMEOUT seconds
  */
-static void
-raw3215_timeout(unsigned long __data)
+static void raw3215_timeout(unsigned long __data)
 {
 	struct raw3215_info *raw = (struct raw3215_info *) __data;
 	unsigned long flags;
@@ -300,8 +295,10 @@ raw3215_timeout(unsigned long __data)
 	if (raw->flags & RAW3215_TIMER_RUNS) {
 		del_timer(&raw->timer);
 		raw->flags &= ~RAW3215_TIMER_RUNS;
-		raw3215_mk_write_req(raw);
-		raw3215_start_io(raw);
+		if (!(raw->flags & RAW3215_FROZEN)) {
+			raw3215_mk_write_req(raw);
+			raw3215_start_io(raw);
+		}
 	}
 	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
 }
@@ -312,10 +309,9 @@ raw3215_timeout(unsigned long __data)
  * amount of data is bigger than RAW3215_MIN_WRITE. If a write is not
  * done immediately a timer is started with a delay of RAW3215_TIMEOUT.
  */
-static inline void
-raw3215_try_io(struct raw3215_info *raw)
+static inline void raw3215_try_io(struct raw3215_info *raw)
 {
-	if (!(raw->flags & RAW3215_ACTIVE))
+	if (!(raw->flags & RAW3215_ACTIVE) || (raw->flags & RAW3215_FROZEN))
 		return;
 	if (raw->queued_read != NULL)
 		raw3215_start_io(raw);
@@ -359,8 +355,8 @@ static void raw3215_next_io(struct raw32
 /*
  * Interrupt routine, called from common io layer
  */
-static void
-raw3215_irq(struct ccw_device *cdev, unsigned long intparm, struct irb *irb)
+static void raw3215_irq(struct ccw_device *cdev, unsigned long intparm,
+			struct irb *irb)
 {
 	struct raw3215_info *raw;
 	struct raw3215_req *req;
@@ -459,14 +455,40 @@ raw3215_irq(struct ccw_device *cdev, uns
 }
 
 /*
+ * Drop the oldest line from the output buffer.
+ */
+static void raw3215_drop_line(struct raw3215_info *raw)
+{
+	int ix;
+	char ch;
+
+	BUG_ON(raw->written != 0);
+	ix = (raw->head - raw->count) & (RAW3215_BUFFER_SIZE - 1);
+	while (raw->count > 0) {
+		ch = raw->buffer[ix];
+		ix = (ix + 1) & (RAW3215_BUFFER_SIZE - 1);
+		raw->count--;
+		if (ch == 0x15)
+			break;
+	}
+	raw->head = ix;
+}
+
+/*
  * Wait until length bytes are available int the output buffer.
  * Has to be called with the s390irq lock held. Can be called
  * disabled.
  */
-static void
-raw3215_make_room(struct raw3215_info *raw, unsigned int length)
+static void raw3215_make_room(struct raw3215_info *raw, unsigned int length)
 {
 	while (RAW3215_BUFFER_SIZE - raw->count < length) {
+		/* While console is frozen for suspend we have no other
+		 * choice but to drop message from the buffer to make
+		 * room for even more messages. */
+		if (raw->flags & RAW3215_FROZEN) {
+			raw3215_drop_line(raw);
+			continue;
+		}
 		/* there might be a request pending */
 		raw->flags |= RAW3215_FLUSHING;
 		raw3215_mk_write_req(raw);
@@ -488,8 +510,8 @@ raw3215_make_room(struct raw3215_info *r
 /*
  * String write routine for 3215 devices
  */
-static void
-raw3215_write(struct raw3215_info *raw, const char *str, unsigned int length)
+static void raw3215_write(struct raw3215_info *raw, const char *str,
+			  unsigned int length)
 {
 	unsigned long flags;
 	int c, count;
@@ -529,8 +551,7 @@ raw3215_write(struct raw3215_info *raw, 
 /*
  * Put character routine for 3215 devices
  */
-static void
-raw3215_putchar(struct raw3215_info *raw, unsigned char ch)
+static void raw3215_putchar(struct raw3215_info *raw, unsigned char ch)
 {
 	unsigned long flags;
 	unsigned int length, i;
@@ -566,8 +587,7 @@ raw3215_putchar(struct raw3215_info *raw
  * Flush routine, it simply sets the flush flag and tries to start
  * pending IO.
  */
-static void
-raw3215_flush_buffer(struct raw3215_info *raw)
+static void raw3215_flush_buffer(struct raw3215_info *raw)
 {
 	unsigned long flags;
 
@@ -583,8 +603,7 @@ raw3215_flush_buffer(struct raw3215_info
 /*
  * Fire up a 3215 device.
  */
-static int
-raw3215_startup(struct raw3215_info *raw)
+static int raw3215_startup(struct raw3215_info *raw)
 {
 	unsigned long flags;
 
@@ -602,8 +621,7 @@ raw3215_startup(struct raw3215_info *raw
 /*
  * Shutdown a 3215 device.
  */
-static void
-raw3215_shutdown(struct raw3215_info *raw)
+static void raw3215_shutdown(struct raw3215_info *raw)
 {
 	DECLARE_WAITQUEUE(wait, current);
 	unsigned long flags;
@@ -628,8 +646,7 @@ raw3215_shutdown(struct raw3215_info *ra
 	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
 }
 
-static int
-raw3215_probe (struct ccw_device *cdev)
+static int raw3215_probe (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 	int line;
@@ -675,8 +692,7 @@ raw3215_probe (struct ccw_device *cdev)
 	return 0;
 }
 
-static void
-raw3215_remove (struct ccw_device *cdev)
+static void raw3215_remove (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 
@@ -689,8 +705,7 @@ raw3215_remove (struct ccw_device *cdev)
 	}
 }
 
-static int
-raw3215_set_online (struct ccw_device *cdev)
+static int raw3215_set_online (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 
@@ -701,8 +716,7 @@ raw3215_set_online (struct ccw_device *c
 	return raw3215_startup(raw);
 }
 
-static int
-raw3215_set_offline (struct ccw_device *cdev)
+static int raw3215_set_offline (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 
@@ -715,6 +729,36 @@ raw3215_set_offline (struct ccw_device *
 	return 0;
 }
 
+static int raw3215_pm_stop(struct ccw_device *cdev)
+{
+	struct raw3215_info *raw;
+	unsigned long flags;
+
+	/* Empty the output buffer, then prevent new I/O. */
+	raw = cdev->dev.driver_data;
+	spin_lock_irqsave(get_ccwdev_lock(raw->cdev), flags);
+	raw3215_make_room(raw, RAW3215_BUFFER_SIZE);
+	raw->flags |= RAW3215_FROZEN;
+	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
+	return 0;
+}
+
+static int raw3215_pm_start(struct ccw_device *cdev)
+{
+	struct raw3215_info *raw;
+	unsigned long flags;
+
+	/* Allow I/O again and flush output buffer. */
+	raw = cdev->dev.driver_data;
+	spin_lock_irqsave(get_ccwdev_lock(raw->cdev), flags);
+	raw->flags &= ~RAW3215_FROZEN;
+	raw->flags |= RAW3215_FLUSHING;
+	raw3215_try_io(raw);
+	raw->flags &= ~RAW3215_FLUSHING;
+	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
+	return 0;
+}
+
 static struct ccw_device_id raw3215_id[] = {
 	{ CCW_DEVICE(0x3215, 0) },
 	{ /* end of list */ },
@@ -728,14 +772,17 @@ static struct ccw_driver raw3215_ccw_dri
 	.remove		= &raw3215_remove,
 	.set_online	= &raw3215_set_online,
 	.set_offline	= &raw3215_set_offline,
+	.freeze		= &raw3215_pm_stop,
+	.thaw		= &raw3215_pm_start,
+	.restore	= &raw3215_pm_start,
 };
 
 #ifdef CONFIG_TN3215_CONSOLE
 /*
  * Write a string to the 3215 console
  */
-static void
-con3215_write(struct console *co, const char *str, unsigned int count)
+static void con3215_write(struct console *co, const char *str,
+			  unsigned int count)
 {
 	struct raw3215_info *raw;
 	int i;
@@ -768,13 +815,17 @@ static struct tty_driver *con3215_device
  * panic() calls con3215_flush through a panic_notifier
  * before the system enters a disabled, endless loop.
  */
-static void
-con3215_flush(void)
+static void con3215_flush(void)
 {
 	struct raw3215_info *raw;
 	unsigned long flags;
 
 	raw = raw3215[0];  /* console 3215 is the first one */
+	if (raw->flags & RAW3215_FROZEN)
+		/* The console is still frozen for suspend. */
+		if (ccw_device_force_console())
+			/* Forcing didn't work, no panic message .. */
+			return;
 	spin_lock_irqsave(get_ccwdev_lock(raw->cdev), flags);
 	raw3215_make_room(raw, RAW3215_BUFFER_SIZE);
 	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
@@ -811,8 +862,7 @@ static struct console con3215 = {
  * 3215 console initialization code called from console_init().
  * NOTE: This is called before kmalloc is available.
  */
-static int __init
-con3215_init(void)
+static int __init con3215_init(void)
 {
 	struct ccw_device *cdev;
 	struct raw3215_info *raw;
@@ -875,8 +925,7 @@ console_initcall(con3215_init);
  *
  * This routine is called whenever a 3215 tty is opened.
  */
-static int
-tty3215_open(struct tty_struct *tty, struct file * filp)
+static int tty3215_open(struct tty_struct *tty, struct file * filp)
 {
 	struct raw3215_info *raw;
 	int retval, line;
@@ -909,8 +958,7 @@ tty3215_open(struct tty_struct *tty, str
  * This routine is called when the 3215 tty is closed. We wait
  * for the remaining request to be completed. Then we clean up.
  */
-static void
-tty3215_close(struct tty_struct *tty, struct file * filp)
+static void tty3215_close(struct tty_struct *tty, struct file * filp)
 {
 	struct raw3215_info *raw;
 
@@ -927,8 +975,7 @@ tty3215_close(struct tty_struct *tty, st
 /*
  * Returns the amount of free space in the output buffer.
  */
-static int
-tty3215_write_room(struct tty_struct *tty)
+static int tty3215_write_room(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -944,9 +991,8 @@ tty3215_write_room(struct tty_struct *tt
 /*
  * String write routine for 3215 ttys
  */
-static int
-tty3215_write(struct tty_struct * tty,
-	      const unsigned char *buf, int count)
+static int tty3215_write(struct tty_struct * tty,
+			 const unsigned char *buf, int count)
 {
 	struct raw3215_info *raw;
 
@@ -960,8 +1006,7 @@ tty3215_write(struct tty_struct * tty,
 /*
  * Put character routine for 3215 ttys
  */
-static int
-tty3215_put_char(struct tty_struct *tty, unsigned char ch)
+static int tty3215_put_char(struct tty_struct *tty, unsigned char ch)
 {
 	struct raw3215_info *raw;
 
@@ -972,16 +1017,14 @@ tty3215_put_char(struct tty_struct *tty,
 	return 1;
 }
 
-static void
-tty3215_flush_chars(struct tty_struct *tty)
+static void tty3215_flush_chars(struct tty_struct *tty)
 {
 }
 
 /*
  * Returns the number of characters in the output buffer
  */
-static int
-tty3215_chars_in_buffer(struct tty_struct *tty)
+static int tty3215_chars_in_buffer(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -989,8 +1032,7 @@ tty3215_chars_in_buffer(struct tty_struc
 	return raw->count;
 }
 
-static void
-tty3215_flush_buffer(struct tty_struct *tty)
+static void tty3215_flush_buffer(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -1002,9 +1044,8 @@ tty3215_flush_buffer(struct tty_struct *
 /*
  * Currently we don't have any io controls for 3215 ttys
  */
-static int
-tty3215_ioctl(struct tty_struct *tty, struct file * file,
-	      unsigned int cmd, unsigned long arg)
+static int tty3215_ioctl(struct tty_struct *tty, struct file * file,
+			 unsigned int cmd, unsigned long arg)
 {
 	if (tty->flags & (1 << TTY_IO_ERROR))
 		return -EIO;
@@ -1019,8 +1060,7 @@ tty3215_ioctl(struct tty_struct *tty, st
 /*
  * Disable reading from a 3215 tty
  */
-static void
-tty3215_throttle(struct tty_struct * tty)
+static void tty3215_throttle(struct tty_struct * tty)
 {
 	struct raw3215_info *raw;
 
@@ -1031,8 +1071,7 @@ tty3215_throttle(struct tty_struct * tty
 /*
  * Enable reading from a 3215 tty
  */
-static void
-tty3215_unthrottle(struct tty_struct * tty)
+static void tty3215_unthrottle(struct tty_struct * tty)
 {
 	struct raw3215_info *raw;
 	unsigned long flags;
@@ -1049,8 +1088,7 @@ tty3215_unthrottle(struct tty_struct * t
 /*
  * Disable writing to a 3215 tty
  */
-static void
-tty3215_stop(struct tty_struct *tty)
+static void tty3215_stop(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -1061,8 +1099,7 @@ tty3215_stop(struct tty_struct *tty)
 /*
  * Enable writing to a 3215 tty
  */
-static void
-tty3215_start(struct tty_struct *tty)
+static void tty3215_start(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 	unsigned long flags;
@@ -1096,8 +1133,7 @@ static const struct tty_operations tty32
  * 3215 tty registration code called from tty_init().
  * Most kernel services (incl. kmalloc) are available at this poimt.
  */
-static int __init
-tty3215_init(void)
+static int __init tty3215_init(void)
 {
 	struct tty_driver *driver;
 	int ret;
@@ -1142,8 +1178,7 @@ tty3215_init(void)
 	return 0;
 }
 
-static void __exit
-tty3215_exit(void)
+static void __exit tty3215_exit(void)
 {
 	tty_unregister_driver(tty3215_driver);
 	put_tty_driver(tty3215_driver);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 15/38] pm: con3215 power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_3215.patch --]
[-- Type: text/plain, Size: 15644 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Introduce the power management callbacks to the 3215 console. On suspend
the console buffer is flushed to the 3215 device to have an empty console
buffer. Printks done while the 3215 device is suspended are buffered in
the 64K buffer of the 3215 device. If the buffer is full new messages will
push out the oldest messages to make room for the most recent message.
On resume the buffered messages are printed. If the system panics before
the 3215 device is resumed ccw_device_force_console is used to get the
console working again.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/con3215.c |  205 +++++++++++++++++++++++++-------------------
 1 file changed, 120 insertions(+), 85 deletions(-)

Index: linux-2.6/drivers/s390/char/con3215.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/con3215.c
+++ linux-2.6/drivers/s390/char/con3215.c
@@ -1,14 +1,12 @@
 /*
- *  drivers/s390/char/con3215.c
- *    3215 line mode terminal driver.
+ * 3215 line mode terminal driver.
  *
- *  S390 version
- *    Copyright (C) 1999,2000 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
+ * Copyright IBM Corp. 1999, 2009
+ * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
  *
- *  Updated:
- *   Aug-2000: Added tab support
- *	       Dan Morrison, IBM Corporation (dmorriso@cse.buffalo.edu)
+ * Updated:
+ *  Aug-2000: Added tab support
+ *	      Dan Morrison, IBM Corporation <dmorriso@cse.buffalo.edu>
  */
 
 #include <linux/module.h>
@@ -56,6 +54,7 @@
 #define RAW3215_CLOSING	    32	      /* set while in close process */
 #define RAW3215_TIMER_RUNS  64	      /* set if the output delay timer is on */
 #define RAW3215_FLUSHING    128	      /* set to flush buffer (no delay) */
+#define RAW3215_FROZEN      256	      /* set if 3215 is frozen for suspend */
 
 #define TAB_STOP_SIZE	    8	      /* tab stop size */
 
@@ -111,8 +110,8 @@ static struct tty_driver *tty3215_driver
 /*
  * Get a request structure from the free list
  */
-static inline struct raw3215_req *
-raw3215_alloc_req(void) {
+static inline struct raw3215_req *raw3215_alloc_req(void)
+{
 	struct raw3215_req *req;
 	unsigned long flags;
 
@@ -126,8 +125,8 @@ raw3215_alloc_req(void) {
 /*
  * Put a request structure back to the free list
  */
-static inline void
-raw3215_free_req(struct raw3215_req *req) {
+static inline void raw3215_free_req(struct raw3215_req *req)
+{
 	unsigned long flags;
 
 	if (req->type == RAW3215_FREE)
@@ -145,8 +144,7 @@ raw3215_free_req(struct raw3215_req *req
  * because a 3215 terminal won't accept a new read before the old one is
  * completed.
  */
-static void
-raw3215_mk_read_req(struct raw3215_info *raw)
+static void raw3215_mk_read_req(struct raw3215_info *raw)
 {
 	struct raw3215_req *req;
 	struct ccw1 *ccw;
@@ -174,8 +172,7 @@ raw3215_mk_read_req(struct raw3215_info 
  * buffer to the 3215 device. If a queued write exists it is replaced by
  * the new, probably lengthened request.
  */
-static void
-raw3215_mk_write_req(struct raw3215_info *raw)
+static void raw3215_mk_write_req(struct raw3215_info *raw)
 {
 	struct raw3215_req *req;
 	struct ccw1 *ccw;
@@ -251,8 +248,7 @@ raw3215_mk_write_req(struct raw3215_info
 /*
  * Start a read or a write request
  */
-static void
-raw3215_start_io(struct raw3215_info *raw)
+static void raw3215_start_io(struct raw3215_info *raw)
 {
 	struct raw3215_req *req;
 	int res;
@@ -290,8 +286,7 @@ raw3215_start_io(struct raw3215_info *ra
 /*
  * Function to start a delayed output after RAW3215_TIMEOUT seconds
  */
-static void
-raw3215_timeout(unsigned long __data)
+static void raw3215_timeout(unsigned long __data)
 {
 	struct raw3215_info *raw = (struct raw3215_info *) __data;
 	unsigned long flags;
@@ -300,8 +295,10 @@ raw3215_timeout(unsigned long __data)
 	if (raw->flags & RAW3215_TIMER_RUNS) {
 		del_timer(&raw->timer);
 		raw->flags &= ~RAW3215_TIMER_RUNS;
-		raw3215_mk_write_req(raw);
-		raw3215_start_io(raw);
+		if (!(raw->flags & RAW3215_FROZEN)) {
+			raw3215_mk_write_req(raw);
+			raw3215_start_io(raw);
+		}
 	}
 	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
 }
@@ -312,10 +309,9 @@ raw3215_timeout(unsigned long __data)
  * amount of data is bigger than RAW3215_MIN_WRITE. If a write is not
  * done immediately a timer is started with a delay of RAW3215_TIMEOUT.
  */
-static inline void
-raw3215_try_io(struct raw3215_info *raw)
+static inline void raw3215_try_io(struct raw3215_info *raw)
 {
-	if (!(raw->flags & RAW3215_ACTIVE))
+	if (!(raw->flags & RAW3215_ACTIVE) || (raw->flags & RAW3215_FROZEN))
 		return;
 	if (raw->queued_read != NULL)
 		raw3215_start_io(raw);
@@ -359,8 +355,8 @@ static void raw3215_next_io(struct raw32
 /*
  * Interrupt routine, called from common io layer
  */
-static void
-raw3215_irq(struct ccw_device *cdev, unsigned long intparm, struct irb *irb)
+static void raw3215_irq(struct ccw_device *cdev, unsigned long intparm,
+			struct irb *irb)
 {
 	struct raw3215_info *raw;
 	struct raw3215_req *req;
@@ -459,14 +455,40 @@ raw3215_irq(struct ccw_device *cdev, uns
 }
 
 /*
+ * Drop the oldest line from the output buffer.
+ */
+static void raw3215_drop_line(struct raw3215_info *raw)
+{
+	int ix;
+	char ch;
+
+	BUG_ON(raw->written != 0);
+	ix = (raw->head - raw->count) & (RAW3215_BUFFER_SIZE - 1);
+	while (raw->count > 0) {
+		ch = raw->buffer[ix];
+		ix = (ix + 1) & (RAW3215_BUFFER_SIZE - 1);
+		raw->count--;
+		if (ch == 0x15)
+			break;
+	}
+	raw->head = ix;
+}
+
+/*
  * Wait until length bytes are available int the output buffer.
  * Has to be called with the s390irq lock held. Can be called
  * disabled.
  */
-static void
-raw3215_make_room(struct raw3215_info *raw, unsigned int length)
+static void raw3215_make_room(struct raw3215_info *raw, unsigned int length)
 {
 	while (RAW3215_BUFFER_SIZE - raw->count < length) {
+		/* While console is frozen for suspend we have no other
+		 * choice but to drop message from the buffer to make
+		 * room for even more messages. */
+		if (raw->flags & RAW3215_FROZEN) {
+			raw3215_drop_line(raw);
+			continue;
+		}
 		/* there might be a request pending */
 		raw->flags |= RAW3215_FLUSHING;
 		raw3215_mk_write_req(raw);
@@ -488,8 +510,8 @@ raw3215_make_room(struct raw3215_info *r
 /*
  * String write routine for 3215 devices
  */
-static void
-raw3215_write(struct raw3215_info *raw, const char *str, unsigned int length)
+static void raw3215_write(struct raw3215_info *raw, const char *str,
+			  unsigned int length)
 {
 	unsigned long flags;
 	int c, count;
@@ -529,8 +551,7 @@ raw3215_write(struct raw3215_info *raw, 
 /*
  * Put character routine for 3215 devices
  */
-static void
-raw3215_putchar(struct raw3215_info *raw, unsigned char ch)
+static void raw3215_putchar(struct raw3215_info *raw, unsigned char ch)
 {
 	unsigned long flags;
 	unsigned int length, i;
@@ -566,8 +587,7 @@ raw3215_putchar(struct raw3215_info *raw
  * Flush routine, it simply sets the flush flag and tries to start
  * pending IO.
  */
-static void
-raw3215_flush_buffer(struct raw3215_info *raw)
+static void raw3215_flush_buffer(struct raw3215_info *raw)
 {
 	unsigned long flags;
 
@@ -583,8 +603,7 @@ raw3215_flush_buffer(struct raw3215_info
 /*
  * Fire up a 3215 device.
  */
-static int
-raw3215_startup(struct raw3215_info *raw)
+static int raw3215_startup(struct raw3215_info *raw)
 {
 	unsigned long flags;
 
@@ -602,8 +621,7 @@ raw3215_startup(struct raw3215_info *raw
 /*
  * Shutdown a 3215 device.
  */
-static void
-raw3215_shutdown(struct raw3215_info *raw)
+static void raw3215_shutdown(struct raw3215_info *raw)
 {
 	DECLARE_WAITQUEUE(wait, current);
 	unsigned long flags;
@@ -628,8 +646,7 @@ raw3215_shutdown(struct raw3215_info *ra
 	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
 }
 
-static int
-raw3215_probe (struct ccw_device *cdev)
+static int raw3215_probe (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 	int line;
@@ -675,8 +692,7 @@ raw3215_probe (struct ccw_device *cdev)
 	return 0;
 }
 
-static void
-raw3215_remove (struct ccw_device *cdev)
+static void raw3215_remove (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 
@@ -689,8 +705,7 @@ raw3215_remove (struct ccw_device *cdev)
 	}
 }
 
-static int
-raw3215_set_online (struct ccw_device *cdev)
+static int raw3215_set_online (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 
@@ -701,8 +716,7 @@ raw3215_set_online (struct ccw_device *c
 	return raw3215_startup(raw);
 }
 
-static int
-raw3215_set_offline (struct ccw_device *cdev)
+static int raw3215_set_offline (struct ccw_device *cdev)
 {
 	struct raw3215_info *raw;
 
@@ -715,6 +729,36 @@ raw3215_set_offline (struct ccw_device *
 	return 0;
 }
 
+static int raw3215_pm_stop(struct ccw_device *cdev)
+{
+	struct raw3215_info *raw;
+	unsigned long flags;
+
+	/* Empty the output buffer, then prevent new I/O. */
+	raw = cdev->dev.driver_data;
+	spin_lock_irqsave(get_ccwdev_lock(raw->cdev), flags);
+	raw3215_make_room(raw, RAW3215_BUFFER_SIZE);
+	raw->flags |= RAW3215_FROZEN;
+	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
+	return 0;
+}
+
+static int raw3215_pm_start(struct ccw_device *cdev)
+{
+	struct raw3215_info *raw;
+	unsigned long flags;
+
+	/* Allow I/O again and flush output buffer. */
+	raw = cdev->dev.driver_data;
+	spin_lock_irqsave(get_ccwdev_lock(raw->cdev), flags);
+	raw->flags &= ~RAW3215_FROZEN;
+	raw->flags |= RAW3215_FLUSHING;
+	raw3215_try_io(raw);
+	raw->flags &= ~RAW3215_FLUSHING;
+	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
+	return 0;
+}
+
 static struct ccw_device_id raw3215_id[] = {
 	{ CCW_DEVICE(0x3215, 0) },
 	{ /* end of list */ },
@@ -728,14 +772,17 @@ static struct ccw_driver raw3215_ccw_dri
 	.remove		= &raw3215_remove,
 	.set_online	= &raw3215_set_online,
 	.set_offline	= &raw3215_set_offline,
+	.freeze		= &raw3215_pm_stop,
+	.thaw		= &raw3215_pm_start,
+	.restore	= &raw3215_pm_start,
 };
 
 #ifdef CONFIG_TN3215_CONSOLE
 /*
  * Write a string to the 3215 console
  */
-static void
-con3215_write(struct console *co, const char *str, unsigned int count)
+static void con3215_write(struct console *co, const char *str,
+			  unsigned int count)
 {
 	struct raw3215_info *raw;
 	int i;
@@ -768,13 +815,17 @@ static struct tty_driver *con3215_device
  * panic() calls con3215_flush through a panic_notifier
  * before the system enters a disabled, endless loop.
  */
-static void
-con3215_flush(void)
+static void con3215_flush(void)
 {
 	struct raw3215_info *raw;
 	unsigned long flags;
 
 	raw = raw3215[0];  /* console 3215 is the first one */
+	if (raw->flags & RAW3215_FROZEN)
+		/* The console is still frozen for suspend. */
+		if (ccw_device_force_console())
+			/* Forcing didn't work, no panic message .. */
+			return;
 	spin_lock_irqsave(get_ccwdev_lock(raw->cdev), flags);
 	raw3215_make_room(raw, RAW3215_BUFFER_SIZE);
 	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
@@ -811,8 +862,7 @@ static struct console con3215 = {
  * 3215 console initialization code called from console_init().
  * NOTE: This is called before kmalloc is available.
  */
-static int __init
-con3215_init(void)
+static int __init con3215_init(void)
 {
 	struct ccw_device *cdev;
 	struct raw3215_info *raw;
@@ -875,8 +925,7 @@ console_initcall(con3215_init);
  *
  * This routine is called whenever a 3215 tty is opened.
  */
-static int
-tty3215_open(struct tty_struct *tty, struct file * filp)
+static int tty3215_open(struct tty_struct *tty, struct file * filp)
 {
 	struct raw3215_info *raw;
 	int retval, line;
@@ -909,8 +958,7 @@ tty3215_open(struct tty_struct *tty, str
  * This routine is called when the 3215 tty is closed. We wait
  * for the remaining request to be completed. Then we clean up.
  */
-static void
-tty3215_close(struct tty_struct *tty, struct file * filp)
+static void tty3215_close(struct tty_struct *tty, struct file * filp)
 {
 	struct raw3215_info *raw;
 
@@ -927,8 +975,7 @@ tty3215_close(struct tty_struct *tty, st
 /*
  * Returns the amount of free space in the output buffer.
  */
-static int
-tty3215_write_room(struct tty_struct *tty)
+static int tty3215_write_room(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -944,9 +991,8 @@ tty3215_write_room(struct tty_struct *tt
 /*
  * String write routine for 3215 ttys
  */
-static int
-tty3215_write(struct tty_struct * tty,
-	      const unsigned char *buf, int count)
+static int tty3215_write(struct tty_struct * tty,
+			 const unsigned char *buf, int count)
 {
 	struct raw3215_info *raw;
 
@@ -960,8 +1006,7 @@ tty3215_write(struct tty_struct * tty,
 /*
  * Put character routine for 3215 ttys
  */
-static int
-tty3215_put_char(struct tty_struct *tty, unsigned char ch)
+static int tty3215_put_char(struct tty_struct *tty, unsigned char ch)
 {
 	struct raw3215_info *raw;
 
@@ -972,16 +1017,14 @@ tty3215_put_char(struct tty_struct *tty,
 	return 1;
 }
 
-static void
-tty3215_flush_chars(struct tty_struct *tty)
+static void tty3215_flush_chars(struct tty_struct *tty)
 {
 }
 
 /*
  * Returns the number of characters in the output buffer
  */
-static int
-tty3215_chars_in_buffer(struct tty_struct *tty)
+static int tty3215_chars_in_buffer(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -989,8 +1032,7 @@ tty3215_chars_in_buffer(struct tty_struc
 	return raw->count;
 }
 
-static void
-tty3215_flush_buffer(struct tty_struct *tty)
+static void tty3215_flush_buffer(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -1002,9 +1044,8 @@ tty3215_flush_buffer(struct tty_struct *
 /*
  * Currently we don't have any io controls for 3215 ttys
  */
-static int
-tty3215_ioctl(struct tty_struct *tty, struct file * file,
-	      unsigned int cmd, unsigned long arg)
+static int tty3215_ioctl(struct tty_struct *tty, struct file * file,
+			 unsigned int cmd, unsigned long arg)
 {
 	if (tty->flags & (1 << TTY_IO_ERROR))
 		return -EIO;
@@ -1019,8 +1060,7 @@ tty3215_ioctl(struct tty_struct *tty, st
 /*
  * Disable reading from a 3215 tty
  */
-static void
-tty3215_throttle(struct tty_struct * tty)
+static void tty3215_throttle(struct tty_struct * tty)
 {
 	struct raw3215_info *raw;
 
@@ -1031,8 +1071,7 @@ tty3215_throttle(struct tty_struct * tty
 /*
  * Enable reading from a 3215 tty
  */
-static void
-tty3215_unthrottle(struct tty_struct * tty)
+static void tty3215_unthrottle(struct tty_struct * tty)
 {
 	struct raw3215_info *raw;
 	unsigned long flags;
@@ -1049,8 +1088,7 @@ tty3215_unthrottle(struct tty_struct * t
 /*
  * Disable writing to a 3215 tty
  */
-static void
-tty3215_stop(struct tty_struct *tty)
+static void tty3215_stop(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 
@@ -1061,8 +1099,7 @@ tty3215_stop(struct tty_struct *tty)
 /*
  * Enable writing to a 3215 tty
  */
-static void
-tty3215_start(struct tty_struct *tty)
+static void tty3215_start(struct tty_struct *tty)
 {
 	struct raw3215_info *raw;
 	unsigned long flags;
@@ -1096,8 +1133,7 @@ static const struct tty_operations tty32
  * 3215 tty registration code called from tty_init().
  * Most kernel services (incl. kmalloc) are available at this poimt.
  */
-static int __init
-tty3215_init(void)
+static int __init tty3215_init(void)
 {
 	struct tty_driver *driver;
 	int ret;
@@ -1142,8 +1178,7 @@ tty3215_init(void)
 	return 0;
 }
 
-static void __exit
-tty3215_exit(void)
+static void __exit tty3215_exit(void)
 {
 	tty_unregister_driver(tty3215_driver);
 	put_tty_driver(tty3215_driver);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 16/38] pm: lcs driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_lcs.patch --]
[-- Type: text/plain, Size: 3061 bytes --]

From: Klaus-Dieter Wacket <kdwacker@de.ibm.com>

Signed-off-by: Klaus-Dieter Wacker <kdwacker@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/lcs.c |   74 +++++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 65 insertions(+), 9 deletions(-)

Index: linux-2.6/drivers/s390/net/lcs.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/lcs.c
+++ linux-2.6/drivers/s390/net/lcs.c
@@ -1,15 +1,12 @@
 /*
- *  linux/drivers/s390/net/lcs.c
- *
  *  Linux for S/390 Lan Channel Station Network Driver
  *
- *  Copyright (C)  1999-2001 IBM Deutschland Entwicklung GmbH,
- *			     IBM Corporation
- *    Author(s): Original Code written by
- *			  DJ Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com)
- *		 Rewritten by
- *			  Frank Pavlic (fpavlic@de.ibm.com) and
- *		 	  Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *  Copyright IBM Corp. 1999, 2009
+ *  Author(s): Original Code written by
+ *			DJ Barrow <djbarrow@de.ibm.com,barrow_dj@yahoo.com>
+ *	       Rewritten by
+ *			Frank Pavlic <fpavlic@de.ibm.com> and
+ *		 	Martin Schwidefsky <schwidefsky@de.ibm.com>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -2313,6 +2310,60 @@ lcs_remove_device(struct ccwgroup_device
 	put_device(&ccwgdev->dev);
 }
 
+static int lcs_pm_suspend(struct lcs_card *card)
+{
+	if (card->dev)
+		netif_device_detach(card->dev);
+	lcs_set_allowed_threads(card, 0);
+	lcs_wait_for_threads(card, 0xffffffff);
+	if (card->state != DEV_STATE_DOWN)
+		__lcs_shutdown_device(card->gdev, 1);
+	return 0;
+}
+
+static int lcs_pm_resume(struct lcs_card *card)
+{
+	int rc = 0;
+
+	if (card->state == DEV_STATE_RECOVER)
+		rc = lcs_new_device(card->gdev);
+	if (card->dev)
+		netif_device_attach(card->dev);
+	if (rc) {
+		dev_warn(&card->gdev->dev, "The lcs device driver "
+			"failed to recover the device\n");
+	}
+	return rc;
+}
+
+static int lcs_prepare(struct ccwgroup_device *gdev)
+{
+	return 0;
+}
+
+static void lcs_complete(struct ccwgroup_device *gdev)
+{
+	return;
+}
+
+static int lcs_freeze(struct ccwgroup_device *gdev)
+{
+	struct lcs_card *card = dev_get_drvdata(&gdev->dev);
+	return lcs_pm_suspend(card);
+}
+
+static int lcs_thaw(struct ccwgroup_device *gdev)
+{
+	struct lcs_card *card = dev_get_drvdata(&gdev->dev);
+	return lcs_pm_resume(card);
+}
+
+static int lcs_restore(struct ccwgroup_device *gdev)
+{
+	struct lcs_card *card = dev_get_drvdata(&gdev->dev);
+	return lcs_pm_resume(card);
+}
+
 /**
  * LCS ccwgroup driver registration
  */
@@ -2325,6 +2376,11 @@ static struct ccwgroup_driver lcs_group_
 	.remove      = lcs_remove_device,
 	.set_online  = lcs_new_device,
 	.set_offline = lcs_shutdown_device,
+	.prepare     = lcs_prepare,
+	.complete    = lcs_complete,
+	.freeze	     = lcs_freeze,
+	.thaw	     = lcs_thaw,
+	.restore     = lcs_restore,
 };
 
 /**

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 16/38] pm: lcs driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_lcs.patch --]
[-- Type: text/plain, Size: 3060 bytes --]

From: Klaus-Dieter Wacket <kdwacker@de.ibm.com>

Signed-off-by: Klaus-Dieter Wacker <kdwacker@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/lcs.c |   74 +++++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 65 insertions(+), 9 deletions(-)

Index: linux-2.6/drivers/s390/net/lcs.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/lcs.c
+++ linux-2.6/drivers/s390/net/lcs.c
@@ -1,15 +1,12 @@
 /*
- *  linux/drivers/s390/net/lcs.c
- *
  *  Linux for S/390 Lan Channel Station Network Driver
  *
- *  Copyright (C)  1999-2001 IBM Deutschland Entwicklung GmbH,
- *			     IBM Corporation
- *    Author(s): Original Code written by
- *			  DJ Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com)
- *		 Rewritten by
- *			  Frank Pavlic (fpavlic@de.ibm.com) and
- *		 	  Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *  Copyright IBM Corp. 1999, 2009
+ *  Author(s): Original Code written by
+ *			DJ Barrow <djbarrow@de.ibm.com,barrow_dj@yahoo.com>
+ *	       Rewritten by
+ *			Frank Pavlic <fpavlic@de.ibm.com> and
+ *		 	Martin Schwidefsky <schwidefsky@de.ibm.com>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -2313,6 +2310,60 @@ lcs_remove_device(struct ccwgroup_device
 	put_device(&ccwgdev->dev);
 }
 
+static int lcs_pm_suspend(struct lcs_card *card)
+{
+	if (card->dev)
+		netif_device_detach(card->dev);
+	lcs_set_allowed_threads(card, 0);
+	lcs_wait_for_threads(card, 0xffffffff);
+	if (card->state != DEV_STATE_DOWN)
+		__lcs_shutdown_device(card->gdev, 1);
+	return 0;
+}
+
+static int lcs_pm_resume(struct lcs_card *card)
+{
+	int rc = 0;
+
+	if (card->state == DEV_STATE_RECOVER)
+		rc = lcs_new_device(card->gdev);
+	if (card->dev)
+		netif_device_attach(card->dev);
+	if (rc) {
+		dev_warn(&card->gdev->dev, "The lcs device driver "
+			"failed to recover the device\n");
+	}
+	return rc;
+}
+
+static int lcs_prepare(struct ccwgroup_device *gdev)
+{
+	return 0;
+}
+
+static void lcs_complete(struct ccwgroup_device *gdev)
+{
+	return;
+}
+
+static int lcs_freeze(struct ccwgroup_device *gdev)
+{
+	struct lcs_card *card = dev_get_drvdata(&gdev->dev);
+	return lcs_pm_suspend(card);
+}
+
+static int lcs_thaw(struct ccwgroup_device *gdev)
+{
+	struct lcs_card *card = dev_get_drvdata(&gdev->dev);
+	return lcs_pm_resume(card);
+}
+
+static int lcs_restore(struct ccwgroup_device *gdev)
+{
+	struct lcs_card *card = dev_get_drvdata(&gdev->dev);
+	return lcs_pm_resume(card);
+}
+
 /**
  * LCS ccwgroup driver registration
  */
@@ -2325,6 +2376,11 @@ static struct ccwgroup_driver lcs_group_
 	.remove      = lcs_remove_device,
 	.set_online  = lcs_new_device,
 	.set_offline = lcs_shutdown_device,
+	.prepare     = lcs_prepare,
+	.complete    = lcs_complete,
+	.freeze	     = lcs_freeze,
+	.thaw	     = lcs_thaw,
+	.restore     = lcs_restore,
 };
 
 /**

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 17/38] pm: qeth driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_qeth.patch --]
[-- Type: text/plain, Size: 7235 bytes --]

From: Frank Blaschka <frank.blaschka@de.ibm.com>

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/qeth_core_main.c |   51 ++++++++++++++++++++++++++++++++++++-
 drivers/s390/net/qeth_l2_main.c   |   52 +++++++++++++++++++++++++++++++++++++-
 drivers/s390/net/qeth_l3_main.c   |   52 +++++++++++++++++++++++++++++++++++++-
 3 files changed, 152 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/s390/net/qeth_core_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/qeth_core_main.c
+++ linux-2.6/drivers/s390/net/qeth_core_main.c
@@ -1,7 +1,7 @@
 /*
  *  drivers/s390/net/qeth_core_main.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Utz Bacher <utz.bacher@de.ibm.com>,
  *		 Frank Pavlic <fpavlic@de.ibm.com>,
  *		 Thomas Spatzier <tspat@de.ibm.com>,
@@ -4192,6 +4192,50 @@ static void qeth_core_shutdown(struct cc
 		card->discipline.ccwgdriver->shutdown(gdev);
 }
 
+static int qeth_core_prepare(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->prepare)
+		return card->discipline.ccwgdriver->prepare(gdev);
+	return 0;
+}
+
+static void qeth_core_complete(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->complete)
+		card->discipline.ccwgdriver->complete(gdev);
+}
+
+static int qeth_core_freeze(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->freeze)
+		return card->discipline.ccwgdriver->freeze(gdev);
+	return 0;
+}
+
+static int qeth_core_thaw(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->thaw)
+		return card->discipline.ccwgdriver->thaw(gdev);
+	return 0;
+}
+
+static int qeth_core_restore(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->restore)
+		return card->discipline.ccwgdriver->restore(gdev);
+	return 0;
+}
+
 static struct ccwgroup_driver qeth_core_ccwgroup_driver = {
 	.owner = THIS_MODULE,
 	.name = "qeth",
@@ -4201,6 +4245,11 @@ static struct ccwgroup_driver qeth_core_
 	.set_online = qeth_core_set_online,
 	.set_offline = qeth_core_set_offline,
 	.shutdown = qeth_core_shutdown,
+	.prepare = qeth_core_prepare,
+	.complete = qeth_core_complete,
+	.freeze = qeth_core_freeze,
+	.thaw = qeth_core_thaw,
+	.restore = qeth_core_restore,
 };
 
 static ssize_t
Index: linux-2.6/drivers/s390/net/qeth_l3_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/qeth_l3_main.c
+++ linux-2.6/drivers/s390/net/qeth_l3_main.c
@@ -1,7 +1,7 @@
 /*
  *  drivers/s390/net/qeth_l3_main.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Utz Bacher <utz.bacher@de.ibm.com>,
  *		 Frank Pavlic <fpavlic@de.ibm.com>,
  *		 Thomas Spatzier <tspat@de.ibm.com>,
@@ -3274,12 +3274,62 @@ static void qeth_l3_shutdown(struct ccwg
 	qeth_clear_qdio_buffers(card);
 }
 
+static int qeth_l3_pm_suspend(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+
+	if (card->dev)
+		netif_device_detach(card->dev);
+	qeth_set_allowed_threads(card, 0, 1);
+	wait_event(card->wait_q, qeth_threads_running(card, 0xffffffff) == 0);
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	if (card->state == CARD_STATE_UP) {
+		card->use_hard_stop = 1;
+		__qeth_l3_set_offline(card->gdev, 1);
+	} else
+		__qeth_l3_set_offline(card->gdev, 0);
+	return 0;
+}
+
+static int qeth_l3_pm_resume(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	int rc = 0;
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		goto out;
+
+	if (card->state == CARD_STATE_RECOVER) {
+		rc = __qeth_l3_set_online(card->gdev, 1);
+		if (rc) {
+			if (card->dev) {
+				rtnl_lock();
+				dev_close(card->dev);
+				rtnl_unlock();
+			}
+		}
+	} else
+		rc = __qeth_l3_set_online(card->gdev, 0);
+out:
+	qeth_set_allowed_threads(card, 0xffffffff, 0);
+	if (card->dev)
+		netif_device_attach(card->dev);
+	if (rc)
+		dev_warn(&card->gdev->dev, "The qeth device driver "
+			"failed to recover an error on the device\n");
+	return rc;
+}
+
 struct ccwgroup_driver qeth_l3_ccwgroup_driver = {
 	.probe = qeth_l3_probe_device,
 	.remove = qeth_l3_remove_device,
 	.set_online = qeth_l3_set_online,
 	.set_offline = qeth_l3_set_offline,
 	.shutdown = qeth_l3_shutdown,
+	.freeze = qeth_l3_pm_suspend,
+	.thaw = qeth_l3_pm_resume,
+	.restore = qeth_l3_pm_resume,
 };
 EXPORT_SYMBOL_GPL(qeth_l3_ccwgroup_driver);
 
Index: linux-2.6/drivers/s390/net/qeth_l2_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/qeth_l2_main.c
+++ linux-2.6/drivers/s390/net/qeth_l2_main.c
@@ -1,7 +1,7 @@
 /*
  *  drivers/s390/net/qeth_l2_main.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Utz Bacher <utz.bacher@de.ibm.com>,
  *		 Frank Pavlic <fpavlic@de.ibm.com>,
  *		 Thomas Spatzier <tspat@de.ibm.com>,
@@ -1113,12 +1113,62 @@ static void qeth_l2_shutdown(struct ccwg
 	qeth_clear_qdio_buffers(card);
 }
 
+static int qeth_l2_pm_suspend(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+
+	if (card->dev)
+		netif_device_detach(card->dev);
+	qeth_set_allowed_threads(card, 0, 1);
+	wait_event(card->wait_q, qeth_threads_running(card, 0xffffffff) == 0);
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	if (card->state == CARD_STATE_UP) {
+		card->use_hard_stop = 1;
+		__qeth_l2_set_offline(card->gdev, 1);
+	} else
+		__qeth_l2_set_offline(card->gdev, 0);
+	return 0;
+}
+
+static int qeth_l2_pm_resume(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	int rc = 0;
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		goto out;
+
+	if (card->state == CARD_STATE_RECOVER) {
+		rc = __qeth_l2_set_online(card->gdev, 1);
+		if (rc) {
+			if (card->dev) {
+				rtnl_lock();
+				dev_close(card->dev);
+				rtnl_unlock();
+			}
+		}
+	} else
+		rc = __qeth_l2_set_online(card->gdev, 0);
+out:
+	qeth_set_allowed_threads(card, 0xffffffff, 0);
+	if (card->dev)
+		netif_device_attach(card->dev);
+	if (rc)
+		dev_warn(&card->gdev->dev, "The qeth device driver "
+			"failed to recover an error on the device\n");
+	return rc;
+}
+
 struct ccwgroup_driver qeth_l2_ccwgroup_driver = {
 	.probe = qeth_l2_probe_device,
 	.remove = qeth_l2_remove_device,
 	.set_online = qeth_l2_set_online,
 	.set_offline = qeth_l2_set_offline,
 	.shutdown = qeth_l2_shutdown,
+	.freeze = qeth_l2_pm_suspend,
+	.thaw = qeth_l2_pm_resume,
+	.restore = qeth_l2_pm_resume,
 };
 EXPORT_SYMBOL_GPL(qeth_l2_ccwgroup_driver);
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 17/38] pm: qeth driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_qeth.patch --]
[-- Type: text/plain, Size: 7234 bytes --]

From: Frank Blaschka <frank.blaschka@de.ibm.com>

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/qeth_core_main.c |   51 ++++++++++++++++++++++++++++++++++++-
 drivers/s390/net/qeth_l2_main.c   |   52 +++++++++++++++++++++++++++++++++++++-
 drivers/s390/net/qeth_l3_main.c   |   52 +++++++++++++++++++++++++++++++++++++-
 3 files changed, 152 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/s390/net/qeth_core_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/qeth_core_main.c
+++ linux-2.6/drivers/s390/net/qeth_core_main.c
@@ -1,7 +1,7 @@
 /*
  *  drivers/s390/net/qeth_core_main.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Utz Bacher <utz.bacher@de.ibm.com>,
  *		 Frank Pavlic <fpavlic@de.ibm.com>,
  *		 Thomas Spatzier <tspat@de.ibm.com>,
@@ -4192,6 +4192,50 @@ static void qeth_core_shutdown(struct cc
 		card->discipline.ccwgdriver->shutdown(gdev);
 }
 
+static int qeth_core_prepare(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->prepare)
+		return card->discipline.ccwgdriver->prepare(gdev);
+	return 0;
+}
+
+static void qeth_core_complete(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->complete)
+		card->discipline.ccwgdriver->complete(gdev);
+}
+
+static int qeth_core_freeze(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->freeze)
+		return card->discipline.ccwgdriver->freeze(gdev);
+	return 0;
+}
+
+static int qeth_core_thaw(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->thaw)
+		return card->discipline.ccwgdriver->thaw(gdev);
+	return 0;
+}
+
+static int qeth_core_restore(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	if (card->discipline.ccwgdriver &&
+	    card->discipline.ccwgdriver->restore)
+		return card->discipline.ccwgdriver->restore(gdev);
+	return 0;
+}
+
 static struct ccwgroup_driver qeth_core_ccwgroup_driver = {
 	.owner = THIS_MODULE,
 	.name = "qeth",
@@ -4201,6 +4245,11 @@ static struct ccwgroup_driver qeth_core_
 	.set_online = qeth_core_set_online,
 	.set_offline = qeth_core_set_offline,
 	.shutdown = qeth_core_shutdown,
+	.prepare = qeth_core_prepare,
+	.complete = qeth_core_complete,
+	.freeze = qeth_core_freeze,
+	.thaw = qeth_core_thaw,
+	.restore = qeth_core_restore,
 };
 
 static ssize_t
Index: linux-2.6/drivers/s390/net/qeth_l3_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/qeth_l3_main.c
+++ linux-2.6/drivers/s390/net/qeth_l3_main.c
@@ -1,7 +1,7 @@
 /*
  *  drivers/s390/net/qeth_l3_main.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Utz Bacher <utz.bacher@de.ibm.com>,
  *		 Frank Pavlic <fpavlic@de.ibm.com>,
  *		 Thomas Spatzier <tspat@de.ibm.com>,
@@ -3274,12 +3274,62 @@ static void qeth_l3_shutdown(struct ccwg
 	qeth_clear_qdio_buffers(card);
 }
 
+static int qeth_l3_pm_suspend(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+
+	if (card->dev)
+		netif_device_detach(card->dev);
+	qeth_set_allowed_threads(card, 0, 1);
+	wait_event(card->wait_q, qeth_threads_running(card, 0xffffffff) == 0);
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	if (card->state == CARD_STATE_UP) {
+		card->use_hard_stop = 1;
+		__qeth_l3_set_offline(card->gdev, 1);
+	} else
+		__qeth_l3_set_offline(card->gdev, 0);
+	return 0;
+}
+
+static int qeth_l3_pm_resume(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	int rc = 0;
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		goto out;
+
+	if (card->state == CARD_STATE_RECOVER) {
+		rc = __qeth_l3_set_online(card->gdev, 1);
+		if (rc) {
+			if (card->dev) {
+				rtnl_lock();
+				dev_close(card->dev);
+				rtnl_unlock();
+			}
+		}
+	} else
+		rc = __qeth_l3_set_online(card->gdev, 0);
+out:
+	qeth_set_allowed_threads(card, 0xffffffff, 0);
+	if (card->dev)
+		netif_device_attach(card->dev);
+	if (rc)
+		dev_warn(&card->gdev->dev, "The qeth device driver "
+			"failed to recover an error on the device\n");
+	return rc;
+}
+
 struct ccwgroup_driver qeth_l3_ccwgroup_driver = {
 	.probe = qeth_l3_probe_device,
 	.remove = qeth_l3_remove_device,
 	.set_online = qeth_l3_set_online,
 	.set_offline = qeth_l3_set_offline,
 	.shutdown = qeth_l3_shutdown,
+	.freeze = qeth_l3_pm_suspend,
+	.thaw = qeth_l3_pm_resume,
+	.restore = qeth_l3_pm_resume,
 };
 EXPORT_SYMBOL_GPL(qeth_l3_ccwgroup_driver);
 
Index: linux-2.6/drivers/s390/net/qeth_l2_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/qeth_l2_main.c
+++ linux-2.6/drivers/s390/net/qeth_l2_main.c
@@ -1,7 +1,7 @@
 /*
  *  drivers/s390/net/qeth_l2_main.c
  *
- *    Copyright IBM Corp. 2007
+ *    Copyright IBM Corp. 2007, 2009
  *    Author(s): Utz Bacher <utz.bacher@de.ibm.com>,
  *		 Frank Pavlic <fpavlic@de.ibm.com>,
  *		 Thomas Spatzier <tspat@de.ibm.com>,
@@ -1113,12 +1113,62 @@ static void qeth_l2_shutdown(struct ccwg
 	qeth_clear_qdio_buffers(card);
 }
 
+static int qeth_l2_pm_suspend(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+
+	if (card->dev)
+		netif_device_detach(card->dev);
+	qeth_set_allowed_threads(card, 0, 1);
+	wait_event(card->wait_q, qeth_threads_running(card, 0xffffffff) == 0);
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	if (card->state == CARD_STATE_UP) {
+		card->use_hard_stop = 1;
+		__qeth_l2_set_offline(card->gdev, 1);
+	} else
+		__qeth_l2_set_offline(card->gdev, 0);
+	return 0;
+}
+
+static int qeth_l2_pm_resume(struct ccwgroup_device *gdev)
+{
+	struct qeth_card *card = dev_get_drvdata(&gdev->dev);
+	int rc = 0;
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		goto out;
+
+	if (card->state == CARD_STATE_RECOVER) {
+		rc = __qeth_l2_set_online(card->gdev, 1);
+		if (rc) {
+			if (card->dev) {
+				rtnl_lock();
+				dev_close(card->dev);
+				rtnl_unlock();
+			}
+		}
+	} else
+		rc = __qeth_l2_set_online(card->gdev, 0);
+out:
+	qeth_set_allowed_threads(card, 0xffffffff, 0);
+	if (card->dev)
+		netif_device_attach(card->dev);
+	if (rc)
+		dev_warn(&card->gdev->dev, "The qeth device driver "
+			"failed to recover an error on the device\n");
+	return rc;
+}
+
 struct ccwgroup_driver qeth_l2_ccwgroup_driver = {
 	.probe = qeth_l2_probe_device,
 	.remove = qeth_l2_remove_device,
 	.set_online = qeth_l2_set_online,
 	.set_offline = qeth_l2_set_offline,
 	.shutdown = qeth_l2_shutdown,
+	.freeze = qeth_l2_pm_suspend,
+	.thaw = qeth_l2_pm_resume,
+	.restore = qeth_l2_pm_resume,
 };
 EXPORT_SYMBOL_GPL(qeth_l2_ccwgroup_driver);
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 18/38] pm: ctcm driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ctcm.patch --]
[-- Type: text/plain, Size: 2092 bytes --]

From: Frank Blaschka <frank.blaschka@de.ibm.com>

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/ctcm_main.c |   37 ++++++++++++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/net/ctcm_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/ctcm_main.c
+++ linux-2.6/drivers/s390/net/ctcm_main.c
@@ -1,7 +1,7 @@
 /*
  * drivers/s390/net/ctcm_main.c
  *
- * Copyright IBM Corp. 2001, 2007
+ * Copyright IBM Corp. 2001, 2009
  * Author(s):
  *	Original CTC driver(s):
  *		Fritz Elfert (felfert@millenux.com)
@@ -1690,6 +1690,38 @@ static void ctcm_remove_device(struct cc
 	put_device(&cgdev->dev);
 }
 
+static int ctcm_pm_suspend(struct ccwgroup_device *gdev)
+{
+	struct ctcm_priv *priv = dev_get_drvdata(&gdev->dev);
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	netif_device_detach(priv->channel[READ]->netdev);
+	ctcm_close(priv->channel[READ]->netdev);
+	ccw_device_set_offline(gdev->cdev[1]);
+	ccw_device_set_offline(gdev->cdev[0]);
+	return 0;
+}
+
+static int ctcm_pm_resume(struct ccwgroup_device *gdev)
+{
+	struct ctcm_priv *priv = dev_get_drvdata(&gdev->dev);
+	int rc;
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	rc = ccw_device_set_online(gdev->cdev[1]);
+	if (rc)
+		goto err_out;
+	rc = ccw_device_set_online(gdev->cdev[0]);
+	if (rc)
+		goto err_out;
+	ctcm_open(priv->channel[READ]->netdev);
+err_out:
+	netif_device_attach(priv->channel[READ]->netdev);
+	return rc;
+}
+
 static struct ccwgroup_driver ctcm_group_driver = {
 	.owner       = THIS_MODULE,
 	.name        = CTC_DRIVER_NAME,
@@ -1699,6 +1731,9 @@ static struct ccwgroup_driver ctcm_group
 	.remove      = ctcm_remove_device,
 	.set_online  = ctcm_new_device,
 	.set_offline = ctcm_shutdown_device,
+	.freeze	     = ctcm_pm_suspend,
+	.thaw	     = ctcm_pm_resume,
+	.restore     = ctcm_pm_resume,
 };
 
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 18/38] pm: ctcm driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ctcm.patch --]
[-- Type: text/plain, Size: 2091 bytes --]

From: Frank Blaschka <frank.blaschka@de.ibm.com>

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/ctcm_main.c |   37 ++++++++++++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/net/ctcm_main.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/ctcm_main.c
+++ linux-2.6/drivers/s390/net/ctcm_main.c
@@ -1,7 +1,7 @@
 /*
  * drivers/s390/net/ctcm_main.c
  *
- * Copyright IBM Corp. 2001, 2007
+ * Copyright IBM Corp. 2001, 2009
  * Author(s):
  *	Original CTC driver(s):
  *		Fritz Elfert (felfert@millenux.com)
@@ -1690,6 +1690,38 @@ static void ctcm_remove_device(struct cc
 	put_device(&cgdev->dev);
 }
 
+static int ctcm_pm_suspend(struct ccwgroup_device *gdev)
+{
+	struct ctcm_priv *priv = dev_get_drvdata(&gdev->dev);
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	netif_device_detach(priv->channel[READ]->netdev);
+	ctcm_close(priv->channel[READ]->netdev);
+	ccw_device_set_offline(gdev->cdev[1]);
+	ccw_device_set_offline(gdev->cdev[0]);
+	return 0;
+}
+
+static int ctcm_pm_resume(struct ccwgroup_device *gdev)
+{
+	struct ctcm_priv *priv = dev_get_drvdata(&gdev->dev);
+	int rc;
+
+	if (gdev->state == CCWGROUP_OFFLINE)
+		return 0;
+	rc = ccw_device_set_online(gdev->cdev[1]);
+	if (rc)
+		goto err_out;
+	rc = ccw_device_set_online(gdev->cdev[0]);
+	if (rc)
+		goto err_out;
+	ctcm_open(priv->channel[READ]->netdev);
+err_out:
+	netif_device_attach(priv->channel[READ]->netdev);
+	return rc;
+}
+
 static struct ccwgroup_driver ctcm_group_driver = {
 	.owner       = THIS_MODULE,
 	.name        = CTC_DRIVER_NAME,
@@ -1699,6 +1731,9 @@ static struct ccwgroup_driver ctcm_group
 	.remove      = ctcm_remove_device,
 	.set_online  = ctcm_new_device,
 	.set_offline = ctcm_shutdown_device,
+	.freeze	     = ctcm_pm_suspend,
+	.thaw	     = ctcm_pm_resume,
+	.restore     = ctcm_pm_resume,
 };
 
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 19/38] pm: claw driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_claw.patch --]
[-- Type: text/plain, Size: 1677 bytes --]

From: Frank Blaschka <frank.blaschka@de.ibm.com>

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/claw.c |   14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

Index: linux-2.6/drivers/s390/net/claw.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/claw.c
+++ linux-2.6/drivers/s390/net/claw.c
@@ -3,12 +3,12 @@
  *    ESCON CLAW network driver
  *
  *  Linux for zSeries version
- *    Copyright (C) 2002,2005 IBM Corporation
+ *    Copyright IBM Corp. 2002, 2009
  *  Author(s) Original code written by:
- *              Kazuo Iimura (iimura@jp.ibm.com)
+ *              Kazuo Iimura <iimura@jp.ibm.com>
  *   	      Rewritten by
- *              Andy Richter (richtera@us.ibm.com)
- *              Marc Price (mwprice@us.ibm.com)
+ *              Andy Richter <richtera@us.ibm.com>
+ *              Marc Price <mwprice@us.ibm.com>
  *
  *    sysfs parms:
  *   group x.x.rrrr,x.x.wwww
@@ -253,6 +253,11 @@ static void claw_free_wrt_buf(struct net
 /* Functions for unpack reads   */
 static void unpack_read(struct net_device *dev);
 
+static int claw_pm_prepare(struct ccwgroup_device *gdev)
+{
+	return -EPERM;
+}
+
 /* ccwgroup table  */
 
 static struct ccwgroup_driver claw_group_driver = {
@@ -264,6 +269,7 @@ static struct ccwgroup_driver claw_group
         .remove      = claw_remove_device,
         .set_online  = claw_new_device,
         .set_offline = claw_shutdown_device,
+	.prepare     = claw_pm_prepare,
 };
 
 /*

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 19/38] pm: claw driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_claw.patch --]
[-- Type: text/plain, Size: 1676 bytes --]

From: Frank Blaschka <frank.blaschka@de.ibm.com>

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/claw.c |   14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

Index: linux-2.6/drivers/s390/net/claw.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/claw.c
+++ linux-2.6/drivers/s390/net/claw.c
@@ -3,12 +3,12 @@
  *    ESCON CLAW network driver
  *
  *  Linux for zSeries version
- *    Copyright (C) 2002,2005 IBM Corporation
+ *    Copyright IBM Corp. 2002, 2009
  *  Author(s) Original code written by:
- *              Kazuo Iimura (iimura@jp.ibm.com)
+ *              Kazuo Iimura <iimura@jp.ibm.com>
  *   	      Rewritten by
- *              Andy Richter (richtera@us.ibm.com)
- *              Marc Price (mwprice@us.ibm.com)
+ *              Andy Richter <richtera@us.ibm.com>
+ *              Marc Price <mwprice@us.ibm.com>
  *
  *    sysfs parms:
  *   group x.x.rrrr,x.x.wwww
@@ -253,6 +253,11 @@ static void claw_free_wrt_buf(struct net
 /* Functions for unpack reads   */
 static void unpack_read(struct net_device *dev);
 
+static int claw_pm_prepare(struct ccwgroup_device *gdev)
+{
+	return -EPERM;
+}
+
 /* ccwgroup table  */
 
 static struct ccwgroup_driver claw_group_driver = {
@@ -264,6 +269,7 @@ static struct ccwgroup_driver claw_group
         .remove      = claw_remove_device,
         .set_online  = claw_new_device,
         .set_offline = claw_shutdown_device,
+	.prepare     = claw_pm_prepare,
 };
 
 /*

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 20/38] pm: zfcp driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_zfcp.patch --]
[-- Type: text/plain, Size: 1817 bytes --]

From: Martin Petermann <martin.petermann@de.ibm.com>

Signed-off-by: Martin Petermann <martin.petermann@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/scsi/zfcp_ccw.c |   33 +++++++++++++++++++++++++++++++++
 1 file changed, 33 insertions(+)

Index: linux-2.6/drivers/s390/scsi/zfcp_ccw.c
===================================================================
--- linux-2.6.orig/drivers/s390/scsi/zfcp_ccw.c
+++ linux-2.6/drivers/s390/scsi/zfcp_ccw.c
@@ -199,6 +199,36 @@ static void zfcp_ccw_shutdown(struct ccw
 	up(&zfcp_data.config_sema);
 }
 
+static int zfcp_ccw_suspend(struct ccw_device *cdev)
+
+{
+	struct zfcp_adapter *adapter = dev_get_drvdata(&cdev->dev);
+
+	down(&zfcp_data.config_sema);
+
+	zfcp_erp_adapter_shutdown(adapter, 0, "ccsusp1", NULL);
+	zfcp_erp_wait(adapter);
+
+	up(&zfcp_data.config_sema);
+
+	return 0;
+}
+
+static int zfcp_ccw_activate(struct ccw_device *cdev)
+
+{
+	struct zfcp_adapter *adapter = dev_get_drvdata(&cdev->dev);
+
+	zfcp_erp_modify_adapter_status(adapter, "ccresu1", NULL,
+				       ZFCP_STATUS_COMMON_RUNNING, ZFCP_SET);
+	zfcp_erp_adapter_reopen(adapter, ZFCP_STATUS_COMMON_ERP_FAILED,
+				"ccresu2", NULL);
+	zfcp_erp_wait(adapter);
+	flush_work(&adapter->scan_work);
+
+	return 0;
+}
+
 static struct ccw_device_id zfcp_ccw_device_id[] = {
 	{ CCW_DEVICE_DEVTYPE(0x1731, 0x3, 0x1732, 0x3) },
 	{ CCW_DEVICE_DEVTYPE(0x1731, 0x3, 0x1732, 0x4) }, /* priv. */
@@ -217,6 +247,9 @@ static struct ccw_driver zfcp_ccw_driver
 	.set_offline = zfcp_ccw_set_offline,
 	.notify      = zfcp_ccw_notify,
 	.shutdown    = zfcp_ccw_shutdown,
+	.freeze      = zfcp_ccw_suspend,
+	.thaw        = zfcp_ccw_activate,
+	.restore     = zfcp_ccw_activate,
 };
 
 /**

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 20/38] pm: zfcp driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_zfcp.patch --]
[-- Type: text/plain, Size: 1816 bytes --]

From: Martin Petermann <martin.petermann@de.ibm.com>

Signed-off-by: Martin Petermann <martin.petermann@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/scsi/zfcp_ccw.c |   33 +++++++++++++++++++++++++++++++++
 1 file changed, 33 insertions(+)

Index: linux-2.6/drivers/s390/scsi/zfcp_ccw.c
===================================================================
--- linux-2.6.orig/drivers/s390/scsi/zfcp_ccw.c
+++ linux-2.6/drivers/s390/scsi/zfcp_ccw.c
@@ -199,6 +199,36 @@ static void zfcp_ccw_shutdown(struct ccw
 	up(&zfcp_data.config_sema);
 }
 
+static int zfcp_ccw_suspend(struct ccw_device *cdev)
+
+{
+	struct zfcp_adapter *adapter = dev_get_drvdata(&cdev->dev);
+
+	down(&zfcp_data.config_sema);
+
+	zfcp_erp_adapter_shutdown(adapter, 0, "ccsusp1", NULL);
+	zfcp_erp_wait(adapter);
+
+	up(&zfcp_data.config_sema);
+
+	return 0;
+}
+
+static int zfcp_ccw_activate(struct ccw_device *cdev)
+
+{
+	struct zfcp_adapter *adapter = dev_get_drvdata(&cdev->dev);
+
+	zfcp_erp_modify_adapter_status(adapter, "ccresu1", NULL,
+				       ZFCP_STATUS_COMMON_RUNNING, ZFCP_SET);
+	zfcp_erp_adapter_reopen(adapter, ZFCP_STATUS_COMMON_ERP_FAILED,
+				"ccresu2", NULL);
+	zfcp_erp_wait(adapter);
+	flush_work(&adapter->scan_work);
+
+	return 0;
+}
+
 static struct ccw_device_id zfcp_ccw_device_id[] = {
 	{ CCW_DEVICE_DEVTYPE(0x1731, 0x3, 0x1732, 0x3) },
 	{ CCW_DEVICE_DEVTYPE(0x1731, 0x3, 0x1732, 0x4) }, /* priv. */
@@ -217,6 +247,9 @@ static struct ccw_driver zfcp_ccw_driver
 	.set_offline = zfcp_ccw_set_offline,
 	.notify      = zfcp_ccw_notify,
 	.shutdown    = zfcp_ccw_shutdown,
+	.freeze      = zfcp_ccw_suspend,
+	.thaw        = zfcp_ccw_activate,
+	.restore     = zfcp_ccw_activate,
 };
 
 /**

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 21/38] pm: vmwatchdog power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_vmwatchdog.patch --]
[-- Type: text/plain, Size: 5240 bytes --]

From: Christian Borntraeger <borntraeger@de.ibm.com>

This patch implements suspend/hibernation for the vmwatchdog driver. The 
pm_notifier_callchain is used to get control on PM events. Since watchdog
operation and suspend cannot work together in a reliable fashion, the open
flag is also used to prevent suspend and open from happening at the same
time.
The watchdog can also be active with no open file descriptor. This patch
adds another flag which is only changed in vmwdt_keep_alive and
vmwdt_disable.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/vmwatchdog.c |   81 ++++++++++++++++++++++++++++++++++++++---
 1 file changed, 76 insertions(+), 5 deletions(-)

Index: linux-2.6/drivers/s390/char/vmwatchdog.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/vmwatchdog.c
+++ linux-2.6/drivers/s390/char/vmwatchdog.c
@@ -1,17 +1,23 @@
 /*
  * Watchdog implementation based on z/VM Watchdog Timer API
  *
+ * Copyright IBM Corp. 2004,2009
+ *
  * The user space watchdog daemon can use this driver as
  * /dev/vmwatchdog to have z/VM execute the specified CP
  * command when the timeout expires. The default command is
  * "IPL", which which cause an immediate reboot.
  */
+#define KMSG_COMPONENT "vmwatchdog"
+#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
+
 #include <linux/init.h>
 #include <linux/fs.h>
 #include <linux/kernel.h>
 #include <linux/miscdevice.h>
 #include <linux/module.h>
 #include <linux/moduleparam.h>
+#include <linux/suspend.h>
 #include <linux/watchdog.h>
 #include <linux/smp_lock.h>
 
@@ -43,6 +49,9 @@ static unsigned int vmwdt_interval = 60;
 static unsigned long vmwdt_is_open;
 static int vmwdt_expect_close;
 
+#define VMWDT_OPEN	0	/* devnode is open or suspend in progress */
+#define VMWDT_RUNNING	1	/* The watchdog is armed */
+
 enum vmwdt_func {
 	/* function codes */
 	wdt_init   = 0,
@@ -92,6 +101,7 @@ static int vmwdt_keepalive(void)
 	EBC_TOUPPER(ebc_cmd, MAX_CMDLEN);
 
 	func = vmwdt_conceal ? (wdt_init | wdt_conceal) : wdt_init;
+	set_bit(VMWDT_RUNNING, &vmwdt_is_open);
 	ret = __diag288(func, vmwdt_interval, ebc_cmd, len);
 	WARN_ON(ret != 0);
 	kfree(ebc_cmd);
@@ -102,6 +112,7 @@ static int vmwdt_disable(void)
 {
 	int ret = __diag288(wdt_cancel, 0, "", 0);
 	WARN_ON(ret != 0);
+	clear_bit(VMWDT_RUNNING, &vmwdt_is_open);
 	return ret;
 }
 
@@ -123,13 +134,13 @@ static int vmwdt_open(struct inode *i, s
 {
 	int ret;
 	lock_kernel();
-	if (test_and_set_bit(0, &vmwdt_is_open)) {
+	if (test_and_set_bit(VMWDT_OPEN, &vmwdt_is_open)) {
 		unlock_kernel();
 		return -EBUSY;
 	}
 	ret = vmwdt_keepalive();
 	if (ret)
-		clear_bit(0, &vmwdt_is_open);
+		clear_bit(VMWDT_OPEN, &vmwdt_is_open);
 	unlock_kernel();
 	return ret ? ret : nonseekable_open(i, f);
 }
@@ -139,7 +150,7 @@ static int vmwdt_close(struct inode *i, 
 	if (vmwdt_expect_close == 42)
 		vmwdt_disable();
 	vmwdt_expect_close = 0;
-	clear_bit(0, &vmwdt_is_open);
+	clear_bit(VMWDT_OPEN, &vmwdt_is_open);
 	return 0;
 }
 
@@ -223,6 +234,57 @@ static ssize_t vmwdt_write(struct file *
 	return count;
 }
 
+static int vmwdt_resume(void)
+{
+	clear_bit(VMWDT_OPEN, &vmwdt_is_open);
+	return NOTIFY_DONE;
+}
+
+/*
+ * It makes no sense to go into suspend while the watchdog is running.
+ * Depending on the memory size, the watchdog might trigger, while we
+ * are still saving the memory.
+ * We reuse the open flag to ensure that suspend and watchdog open are
+ * exclusive operations
+ */
+static int vmwdt_suspend(void)
+{
+	if (test_and_set_bit(VMWDT_OPEN, &vmwdt_is_open)) {
+		pr_err("The watchdog is in use. "
+			"This prevents hibernation or suspend.\n");
+		return NOTIFY_BAD;
+	}
+	if (test_bit(VMWDT_RUNNING, &vmwdt_is_open)) {
+		clear_bit(VMWDT_OPEN, &vmwdt_is_open);
+		pr_err("The watchdog is running. "
+			"This prevents hibernation or suspend.\n");
+		return NOTIFY_BAD;
+	}
+	return NOTIFY_DONE;
+}
+
+/*
+ * This function is called for suspend and resume.
+ */
+static int vmwdt_power_event(struct notifier_block *this, unsigned long event,
+			     void *ptr)
+{
+	switch (event) {
+	case PM_POST_HIBERNATION:
+	case PM_POST_SUSPEND:
+		return vmwdt_resume();
+	case PM_HIBERNATION_PREPARE:
+	case PM_SUSPEND_PREPARE:
+		return vmwdt_suspend();
+	default:
+		return NOTIFY_DONE;
+	}
+}
+
+static struct notifier_block vmwdt_power_notifier = {
+	.notifier_call = vmwdt_power_event,
+};
+
 static const struct file_operations vmwdt_fops = {
 	.open    = &vmwdt_open,
 	.release = &vmwdt_close,
@@ -244,12 +306,21 @@ static int __init vmwdt_init(void)
 	ret = vmwdt_probe();
 	if (ret)
 		return ret;
-	return misc_register(&vmwdt_dev);
+	ret = register_pm_notifier(&vmwdt_power_notifier);
+	if (ret)
+		return ret;
+	ret = misc_register(&vmwdt_dev);
+	if (ret) {
+		unregister_pm_notifier(&vmwdt_power_notifier);
+		return ret;
+	}
+	return 0;
 }
 module_init(vmwdt_init);
 
 static void __exit vmwdt_exit(void)
 {
-	WARN_ON(misc_deregister(&vmwdt_dev) != 0);
+	unregister_pm_notifier(&vmwdt_power_notifier);
+	misc_deregister(&vmwdt_dev);
 }
 module_exit(vmwdt_exit);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 21/38] pm: vmwatchdog power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_vmwatchdog.patch --]
[-- Type: text/plain, Size: 5239 bytes --]

From: Christian Borntraeger <borntraeger@de.ibm.com>

This patch implements suspend/hibernation for the vmwatchdog driver. The 
pm_notifier_callchain is used to get control on PM events. Since watchdog
operation and suspend cannot work together in a reliable fashion, the open
flag is also used to prevent suspend and open from happening at the same
time.
The watchdog can also be active with no open file descriptor. This patch
adds another flag which is only changed in vmwdt_keep_alive and
vmwdt_disable.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/vmwatchdog.c |   81 ++++++++++++++++++++++++++++++++++++++---
 1 file changed, 76 insertions(+), 5 deletions(-)

Index: linux-2.6/drivers/s390/char/vmwatchdog.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/vmwatchdog.c
+++ linux-2.6/drivers/s390/char/vmwatchdog.c
@@ -1,17 +1,23 @@
 /*
  * Watchdog implementation based on z/VM Watchdog Timer API
  *
+ * Copyright IBM Corp. 2004,2009
+ *
  * The user space watchdog daemon can use this driver as
  * /dev/vmwatchdog to have z/VM execute the specified CP
  * command when the timeout expires. The default command is
  * "IPL", which which cause an immediate reboot.
  */
+#define KMSG_COMPONENT "vmwatchdog"
+#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
+
 #include <linux/init.h>
 #include <linux/fs.h>
 #include <linux/kernel.h>
 #include <linux/miscdevice.h>
 #include <linux/module.h>
 #include <linux/moduleparam.h>
+#include <linux/suspend.h>
 #include <linux/watchdog.h>
 #include <linux/smp_lock.h>
 
@@ -43,6 +49,9 @@ static unsigned int vmwdt_interval = 60;
 static unsigned long vmwdt_is_open;
 static int vmwdt_expect_close;
 
+#define VMWDT_OPEN	0	/* devnode is open or suspend in progress */
+#define VMWDT_RUNNING	1	/* The watchdog is armed */
+
 enum vmwdt_func {
 	/* function codes */
 	wdt_init   = 0,
@@ -92,6 +101,7 @@ static int vmwdt_keepalive(void)
 	EBC_TOUPPER(ebc_cmd, MAX_CMDLEN);
 
 	func = vmwdt_conceal ? (wdt_init | wdt_conceal) : wdt_init;
+	set_bit(VMWDT_RUNNING, &vmwdt_is_open);
 	ret = __diag288(func, vmwdt_interval, ebc_cmd, len);
 	WARN_ON(ret != 0);
 	kfree(ebc_cmd);
@@ -102,6 +112,7 @@ static int vmwdt_disable(void)
 {
 	int ret = __diag288(wdt_cancel, 0, "", 0);
 	WARN_ON(ret != 0);
+	clear_bit(VMWDT_RUNNING, &vmwdt_is_open);
 	return ret;
 }
 
@@ -123,13 +134,13 @@ static int vmwdt_open(struct inode *i, s
 {
 	int ret;
 	lock_kernel();
-	if (test_and_set_bit(0, &vmwdt_is_open)) {
+	if (test_and_set_bit(VMWDT_OPEN, &vmwdt_is_open)) {
 		unlock_kernel();
 		return -EBUSY;
 	}
 	ret = vmwdt_keepalive();
 	if (ret)
-		clear_bit(0, &vmwdt_is_open);
+		clear_bit(VMWDT_OPEN, &vmwdt_is_open);
 	unlock_kernel();
 	return ret ? ret : nonseekable_open(i, f);
 }
@@ -139,7 +150,7 @@ static int vmwdt_close(struct inode *i, 
 	if (vmwdt_expect_close == 42)
 		vmwdt_disable();
 	vmwdt_expect_close = 0;
-	clear_bit(0, &vmwdt_is_open);
+	clear_bit(VMWDT_OPEN, &vmwdt_is_open);
 	return 0;
 }
 
@@ -223,6 +234,57 @@ static ssize_t vmwdt_write(struct file *
 	return count;
 }
 
+static int vmwdt_resume(void)
+{
+	clear_bit(VMWDT_OPEN, &vmwdt_is_open);
+	return NOTIFY_DONE;
+}
+
+/*
+ * It makes no sense to go into suspend while the watchdog is running.
+ * Depending on the memory size, the watchdog might trigger, while we
+ * are still saving the memory.
+ * We reuse the open flag to ensure that suspend and watchdog open are
+ * exclusive operations
+ */
+static int vmwdt_suspend(void)
+{
+	if (test_and_set_bit(VMWDT_OPEN, &vmwdt_is_open)) {
+		pr_err("The watchdog is in use. "
+			"This prevents hibernation or suspend.\n");
+		return NOTIFY_BAD;
+	}
+	if (test_bit(VMWDT_RUNNING, &vmwdt_is_open)) {
+		clear_bit(VMWDT_OPEN, &vmwdt_is_open);
+		pr_err("The watchdog is running. "
+			"This prevents hibernation or suspend.\n");
+		return NOTIFY_BAD;
+	}
+	return NOTIFY_DONE;
+}
+
+/*
+ * This function is called for suspend and resume.
+ */
+static int vmwdt_power_event(struct notifier_block *this, unsigned long event,
+			     void *ptr)
+{
+	switch (event) {
+	case PM_POST_HIBERNATION:
+	case PM_POST_SUSPEND:
+		return vmwdt_resume();
+	case PM_HIBERNATION_PREPARE:
+	case PM_SUSPEND_PREPARE:
+		return vmwdt_suspend();
+	default:
+		return NOTIFY_DONE;
+	}
+}
+
+static struct notifier_block vmwdt_power_notifier = {
+	.notifier_call = vmwdt_power_event,
+};
+
 static const struct file_operations vmwdt_fops = {
 	.open    = &vmwdt_open,
 	.release = &vmwdt_close,
@@ -244,12 +306,21 @@ static int __init vmwdt_init(void)
 	ret = vmwdt_probe();
 	if (ret)
 		return ret;
-	return misc_register(&vmwdt_dev);
+	ret = register_pm_notifier(&vmwdt_power_notifier);
+	if (ret)
+		return ret;
+	ret = misc_register(&vmwdt_dev);
+	if (ret) {
+		unregister_pm_notifier(&vmwdt_power_notifier);
+		return ret;
+	}
+	return 0;
 }
 module_init(vmwdt_init);
 
 static void __exit vmwdt_exit(void)
 {
-	WARN_ON(misc_deregister(&vmwdt_dev) != 0);
+	unregister_pm_notifier(&vmwdt_power_notifier);
+	misc_deregister(&vmwdt_dev);
 }
 module_exit(vmwdt_exit);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 22/38] pm: appldata power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_appldata.patch --]
[-- Type: text/plain, Size: 4965 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/appldata/appldata_base.c |  119 +++++++++++++++++++++++++++++++++++--
 1 file changed, 115 insertions(+), 4 deletions(-)

Index: linux-2.6/arch/s390/appldata/appldata_base.c
===================================================================
--- linux-2.6.orig/arch/s390/appldata/appldata_base.c
+++ linux-2.6/arch/s390/appldata/appldata_base.c
@@ -5,7 +5,7 @@
  * Exports appldata_register_ops() and appldata_unregister_ops() for the
  * data gathering modules.
  *
- * Copyright IBM Corp. 2003, 2008
+ * Copyright IBM Corp. 2003, 2009
  *
  * Author: Gerald Schaefer <gerald.schaefer@de.ibm.com>
  */
@@ -26,6 +26,8 @@
 #include <linux/notifier.h>
 #include <linux/cpu.h>
 #include <linux/workqueue.h>
+#include <linux/suspend.h>
+#include <linux/platform_device.h>
 #include <asm/appldata.h>
 #include <asm/timer.h>
 #include <asm/uaccess.h>
@@ -41,6 +43,9 @@
 
 #define TOD_MICRO	0x01000			/* nr. of TOD clock units
 						   for 1 microsecond */
+
+static struct platform_device *appldata_pdev;
+
 /*
  * /proc entries (sysctl)
  */
@@ -86,6 +91,7 @@ static atomic_t appldata_expire_count = 
 static DEFINE_SPINLOCK(appldata_timer_lock);
 static int appldata_interval = APPLDATA_CPU_INTERVAL;
 static int appldata_timer_active;
+static int appldata_timer_suspended = 0;
 
 /*
  * Work queue
@@ -475,6 +481,93 @@ void appldata_unregister_ops(struct appl
 /********************** module-ops management <END> **************************/
 
 
+/**************************** suspend / resume *******************************/
+static int appldata_freeze(struct device *dev)
+{
+	struct appldata_ops *ops;
+	int rc;
+	struct list_head *lh;
+
+	get_online_cpus();
+	spin_lock(&appldata_timer_lock);
+	if (appldata_timer_active) {
+		__appldata_vtimer_setup(APPLDATA_DEL_TIMER);
+		appldata_timer_suspended = 1;
+	}
+	spin_unlock(&appldata_timer_lock);
+	put_online_cpus();
+
+	mutex_lock(&appldata_ops_mutex);
+	list_for_each(lh, &appldata_ops_list) {
+		ops = list_entry(lh, struct appldata_ops, list);
+		if (ops->active == 1) {
+			rc = appldata_diag(ops->record_nr, APPLDATA_STOP_REC,
+					(unsigned long) ops->data, ops->size,
+					ops->mod_lvl);
+			if (rc != 0)
+				pr_err("Stopping the data collection for %s "
+				       "failed with rc=%d\n", ops->name, rc);
+		}
+	}
+	mutex_unlock(&appldata_ops_mutex);
+	return 0;
+}
+
+static int appldata_restore(struct device *dev)
+{
+	struct appldata_ops *ops;
+	int rc;
+	struct list_head *lh;
+
+	get_online_cpus();
+	spin_lock(&appldata_timer_lock);
+	if (appldata_timer_suspended) {
+		__appldata_vtimer_setup(APPLDATA_ADD_TIMER);
+		appldata_timer_suspended = 0;
+	}
+	spin_unlock(&appldata_timer_lock);
+	put_online_cpus();
+
+	mutex_lock(&appldata_ops_mutex);
+	list_for_each(lh, &appldata_ops_list) {
+		ops = list_entry(lh, struct appldata_ops, list);
+		if (ops->active == 1) {
+			ops->callback(ops->data);	// init record
+			rc = appldata_diag(ops->record_nr,
+					APPLDATA_START_INTERVAL_REC,
+					(unsigned long) ops->data, ops->size,
+					ops->mod_lvl);
+			if (rc != 0) {
+				pr_err("Starting the data collection for %s "
+				       "failed with rc=%d\n", ops->name, rc);
+			}
+		}
+	}
+	mutex_unlock(&appldata_ops_mutex);
+	return 0;
+}
+
+static int appldata_thaw(struct device *dev)
+{
+	return appldata_restore(dev);
+}
+
+static struct dev_pm_ops appldata_pm_ops = {
+	.freeze		= appldata_freeze,
+	.thaw		= appldata_thaw,
+	.restore	= appldata_restore,
+};
+
+static struct platform_driver appldata_pdrv = {
+	.driver = {
+		.name	= "appldata",
+		.owner	= THIS_MODULE,
+		.pm	= &appldata_pm_ops,
+	},
+};
+/************************* suspend / resume <END> ****************************/
+
+
 /******************************* init / exit *********************************/
 
 static void __cpuinit appldata_online_cpu(int cpu)
@@ -531,11 +624,23 @@ static struct notifier_block __cpuinitda
  */
 static int __init appldata_init(void)
 {
-	int i;
+	int i, rc;
 
+	rc = platform_driver_register(&appldata_pdrv);
+	if (rc)
+		return rc;
+
+	appldata_pdev = platform_device_register_simple("appldata", -1, NULL,
+							0);
+	if (IS_ERR(appldata_pdev)) {
+		rc = PTR_ERR(appldata_pdev);
+		goto out_driver;
+	}
 	appldata_wq = create_singlethread_workqueue("appldata");
-	if (!appldata_wq)
-		return -ENOMEM;
+	if (!appldata_wq) {
+		rc = -ENOMEM;
+		goto out_device;
+	}
 
 	get_online_cpus();
 	for_each_online_cpu(i)
@@ -547,6 +652,12 @@ static int __init appldata_init(void)
 
 	appldata_sysctl_header = register_sysctl_table(appldata_dir_table);
 	return 0;
+
+out_device:
+	platform_device_unregister(appldata_pdev);
+out_driver:
+	platform_driver_unregister(&appldata_pdrv);
+	return rc;
 }
 
 __initcall(appldata_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 22/38] pm: appldata power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_appldata.patch --]
[-- Type: text/plain, Size: 4964 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/appldata/appldata_base.c |  119 +++++++++++++++++++++++++++++++++++--
 1 file changed, 115 insertions(+), 4 deletions(-)

Index: linux-2.6/arch/s390/appldata/appldata_base.c
===================================================================
--- linux-2.6.orig/arch/s390/appldata/appldata_base.c
+++ linux-2.6/arch/s390/appldata/appldata_base.c
@@ -5,7 +5,7 @@
  * Exports appldata_register_ops() and appldata_unregister_ops() for the
  * data gathering modules.
  *
- * Copyright IBM Corp. 2003, 2008
+ * Copyright IBM Corp. 2003, 2009
  *
  * Author: Gerald Schaefer <gerald.schaefer@de.ibm.com>
  */
@@ -26,6 +26,8 @@
 #include <linux/notifier.h>
 #include <linux/cpu.h>
 #include <linux/workqueue.h>
+#include <linux/suspend.h>
+#include <linux/platform_device.h>
 #include <asm/appldata.h>
 #include <asm/timer.h>
 #include <asm/uaccess.h>
@@ -41,6 +43,9 @@
 
 #define TOD_MICRO	0x01000			/* nr. of TOD clock units
 						   for 1 microsecond */
+
+static struct platform_device *appldata_pdev;
+
 /*
  * /proc entries (sysctl)
  */
@@ -86,6 +91,7 @@ static atomic_t appldata_expire_count = 
 static DEFINE_SPINLOCK(appldata_timer_lock);
 static int appldata_interval = APPLDATA_CPU_INTERVAL;
 static int appldata_timer_active;
+static int appldata_timer_suspended = 0;
 
 /*
  * Work queue
@@ -475,6 +481,93 @@ void appldata_unregister_ops(struct appl
 /********************** module-ops management <END> **************************/
 
 
+/**************************** suspend / resume *******************************/
+static int appldata_freeze(struct device *dev)
+{
+	struct appldata_ops *ops;
+	int rc;
+	struct list_head *lh;
+
+	get_online_cpus();
+	spin_lock(&appldata_timer_lock);
+	if (appldata_timer_active) {
+		__appldata_vtimer_setup(APPLDATA_DEL_TIMER);
+		appldata_timer_suspended = 1;
+	}
+	spin_unlock(&appldata_timer_lock);
+	put_online_cpus();
+
+	mutex_lock(&appldata_ops_mutex);
+	list_for_each(lh, &appldata_ops_list) {
+		ops = list_entry(lh, struct appldata_ops, list);
+		if (ops->active == 1) {
+			rc = appldata_diag(ops->record_nr, APPLDATA_STOP_REC,
+					(unsigned long) ops->data, ops->size,
+					ops->mod_lvl);
+			if (rc != 0)
+				pr_err("Stopping the data collection for %s "
+				       "failed with rc=%d\n", ops->name, rc);
+		}
+	}
+	mutex_unlock(&appldata_ops_mutex);
+	return 0;
+}
+
+static int appldata_restore(struct device *dev)
+{
+	struct appldata_ops *ops;
+	int rc;
+	struct list_head *lh;
+
+	get_online_cpus();
+	spin_lock(&appldata_timer_lock);
+	if (appldata_timer_suspended) {
+		__appldata_vtimer_setup(APPLDATA_ADD_TIMER);
+		appldata_timer_suspended = 0;
+	}
+	spin_unlock(&appldata_timer_lock);
+	put_online_cpus();
+
+	mutex_lock(&appldata_ops_mutex);
+	list_for_each(lh, &appldata_ops_list) {
+		ops = list_entry(lh, struct appldata_ops, list);
+		if (ops->active == 1) {
+			ops->callback(ops->data);	// init record
+			rc = appldata_diag(ops->record_nr,
+					APPLDATA_START_INTERVAL_REC,
+					(unsigned long) ops->data, ops->size,
+					ops->mod_lvl);
+			if (rc != 0) {
+				pr_err("Starting the data collection for %s "
+				       "failed with rc=%d\n", ops->name, rc);
+			}
+		}
+	}
+	mutex_unlock(&appldata_ops_mutex);
+	return 0;
+}
+
+static int appldata_thaw(struct device *dev)
+{
+	return appldata_restore(dev);
+}
+
+static struct dev_pm_ops appldata_pm_ops = {
+	.freeze		= appldata_freeze,
+	.thaw		= appldata_thaw,
+	.restore	= appldata_restore,
+};
+
+static struct platform_driver appldata_pdrv = {
+	.driver = {
+		.name	= "appldata",
+		.owner	= THIS_MODULE,
+		.pm	= &appldata_pm_ops,
+	},
+};
+/************************* suspend / resume <END> ****************************/
+
+
 /******************************* init / exit *********************************/
 
 static void __cpuinit appldata_online_cpu(int cpu)
@@ -531,11 +624,23 @@ static struct notifier_block __cpuinitda
  */
 static int __init appldata_init(void)
 {
-	int i;
+	int i, rc;
 
+	rc = platform_driver_register(&appldata_pdrv);
+	if (rc)
+		return rc;
+
+	appldata_pdev = platform_device_register_simple("appldata", -1, NULL,
+							0);
+	if (IS_ERR(appldata_pdev)) {
+		rc = PTR_ERR(appldata_pdev);
+		goto out_driver;
+	}
 	appldata_wq = create_singlethread_workqueue("appldata");
-	if (!appldata_wq)
-		return -ENOMEM;
+	if (!appldata_wq) {
+		rc = -ENOMEM;
+		goto out_device;
+	}
 
 	get_online_cpus();
 	for_each_online_cpu(i)
@@ -547,6 +652,12 @@ static int __init appldata_init(void)
 
 	appldata_sysctl_header = register_sysctl_table(appldata_dir_table);
 	return 0;
+
+out_device:
+	platform_device_unregister(appldata_pdev);
+out_driver:
+	platform_driver_unregister(&appldata_pdrv);
+	return rc;
 }
 
 __initcall(appldata_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 23/38] pm: vmur driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_vmur.patch --]
[-- Type: text/plain, Size: 2377 bytes --]

From: Frank Munzert <munzert@de.ibm.com>

Signed-off-by: Frank Munzert <munzert@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/vmur.c |   26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/char/vmur.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/vmur.c
+++ linux-2.6/drivers/s390/char/vmur.c
@@ -2,7 +2,7 @@
  * Linux driver for System z and s390 unit record devices
  * (z/VM virtual punch, reader, printer)
  *
- * Copyright IBM Corp. 2001, 2007
+ * Copyright IBM Corp. 2001, 2009
  * Authors: Malcolm Beattie <beattiem@uk.ibm.com>
  *	    Michael Holzheu <holzheu@de.ibm.com>
  *	    Frank Munzert <munzert@de.ibm.com>
@@ -60,6 +60,7 @@ static int ur_probe(struct ccw_device *c
 static void ur_remove(struct ccw_device *cdev);
 static int ur_set_online(struct ccw_device *cdev);
 static int ur_set_offline(struct ccw_device *cdev);
+static int ur_pm_suspend(struct ccw_device *cdev);
 
 static struct ccw_driver ur_driver = {
 	.name		= "vmur",
@@ -69,6 +70,7 @@ static struct ccw_driver ur_driver = {
 	.remove		= ur_remove,
 	.set_online	= ur_set_online,
 	.set_offline	= ur_set_offline,
+	.freeze		= ur_pm_suspend,
 };
 
 static DEFINE_MUTEX(vmur_mutex);
@@ -158,6 +160,28 @@ static void urdev_put(struct urdev *urd)
 }
 
 /*
+ * State and contents of ur devices can be changed by class D users issuing
+ * CP commands such as PURGE or TRANSFER, while the Linux guest is suspended.
+ * Also the Linux guest might be logged off, which causes all active spool
+ * files to be closed.
+ * So we cannot guarantee that spool files are still the same when the Linux
+ * guest is resumed. In order to avoid unpredictable results at resume time
+ * we simply refuse to suspend if a ur device node is open.
+ */
+static int ur_pm_suspend(struct ccw_device *cdev)
+{
+	struct urdev *urd = cdev->dev.driver_data;
+
+	TRACE("ur_pm_suspend: cdev=%p\n", cdev);
+	if (urd->open_flag) {
+		pr_err("Unit record device %s is busy, %s refusing to "
+		       "suspend.\n", dev_name(&cdev->dev), ur_banner);
+		return -EBUSY;
+	}
+	return 0;
+}
+
+/*
  * Low-level functions to do I/O to a ur device.
  *     alloc_chan_prog
  *     free_chan_prog

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 23/38] pm: vmur driver power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_vmur.patch --]
[-- Type: text/plain, Size: 2376 bytes --]

From: Frank Munzert <munzert@de.ibm.com>

Signed-off-by: Frank Munzert <munzert@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/vmur.c |   26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/char/vmur.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/vmur.c
+++ linux-2.6/drivers/s390/char/vmur.c
@@ -2,7 +2,7 @@
  * Linux driver for System z and s390 unit record devices
  * (z/VM virtual punch, reader, printer)
  *
- * Copyright IBM Corp. 2001, 2007
+ * Copyright IBM Corp. 2001, 2009
  * Authors: Malcolm Beattie <beattiem@uk.ibm.com>
  *	    Michael Holzheu <holzheu@de.ibm.com>
  *	    Frank Munzert <munzert@de.ibm.com>
@@ -60,6 +60,7 @@ static int ur_probe(struct ccw_device *c
 static void ur_remove(struct ccw_device *cdev);
 static int ur_set_online(struct ccw_device *cdev);
 static int ur_set_offline(struct ccw_device *cdev);
+static int ur_pm_suspend(struct ccw_device *cdev);
 
 static struct ccw_driver ur_driver = {
 	.name		= "vmur",
@@ -69,6 +70,7 @@ static struct ccw_driver ur_driver = {
 	.remove		= ur_remove,
 	.set_online	= ur_set_online,
 	.set_offline	= ur_set_offline,
+	.freeze		= ur_pm_suspend,
 };
 
 static DEFINE_MUTEX(vmur_mutex);
@@ -158,6 +160,28 @@ static void urdev_put(struct urdev *urd)
 }
 
 /*
+ * State and contents of ur devices can be changed by class D users issuing
+ * CP commands such as PURGE or TRANSFER, while the Linux guest is suspended.
+ * Also the Linux guest might be logged off, which causes all active spool
+ * files to be closed.
+ * So we cannot guarantee that spool files are still the same when the Linux
+ * guest is resumed. In order to avoid unpredictable results at resume time
+ * we simply refuse to suspend if a ur device node is open.
+ */
+static int ur_pm_suspend(struct ccw_device *cdev)
+{
+	struct urdev *urd = cdev->dev.driver_data;
+
+	TRACE("ur_pm_suspend: cdev=%p\n", cdev);
+	if (urd->open_flag) {
+		pr_err("Unit record device %s is busy, %s refusing to "
+		       "suspend.\n", dev_name(&cdev->dev), ur_banner);
+		return -EBUSY;
+	}
+	return 0;
+}
+
+/*
  * Low-level functions to do I/O to a ur device.
  *     alloc_chan_prog
  *     free_chan_prog

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 24/38] pm: vmlogrdr power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_vmlogrdr.patch --]
[-- Type: text/plain, Size: 1950 bytes --]

From: Stefan Weinhuber <wein@de.ibm.com>

Signed-off-by: Stefan Weinhuber <wein@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/vmlogrdr.c |   27 ++++++++++++++++++++++++++-
 1 file changed, 26 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/char/vmlogrdr.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/vmlogrdr.c
+++ linux-2.6/drivers/s390/char/vmlogrdr.c
@@ -3,7 +3,7 @@
  *	character device driver for reading z/VM system service records
  *
  *
- *	Copyright 2004 IBM Corporation
+ *	Copyright IBM Corp. 2004, 2009
  *	character device driver for reading z/VM system service records,
  *	Version 1.0
  *	Author(s): Xenia Tkatschow <xenia@us.ibm.com>
@@ -660,6 +660,29 @@ static struct attribute *vmlogrdr_attrs[
 	NULL,
 };
 
+static int vmlogrdr_pm_prepare(struct device *dev)
+{
+	int rc;
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+
+	rc = 0;
+	if (priv) {
+		spin_lock_bh(&priv->priv_lock);
+		if (priv->dev_in_use)
+			rc = -EBUSY;
+		spin_unlock_bh(&priv->priv_lock);
+	}
+	if (rc)
+		pr_err("vmlogrdr: device %s is busy. Refuse to suspend.\n",
+		       dev_name(dev));
+	return rc;
+}
+
+
+static struct dev_pm_ops vmlogrdr_pm_ops = {
+	.prepare = vmlogrdr_pm_prepare,
+};
+
 static struct attribute_group vmlogrdr_attr_group = {
 	.attrs = vmlogrdr_attrs,
 };
@@ -668,6 +691,7 @@ static struct class *vmlogrdr_class;
 static struct device_driver vmlogrdr_driver = {
 	.name = "vmlogrdr",
 	.bus  = &iucv_bus,
+	.pm = &vmlogrdr_pm_ops,
 };
 
 
@@ -729,6 +753,7 @@ static int vmlogrdr_register_device(stru
 		dev->bus = &iucv_bus;
 		dev->parent = iucv_root;
 		dev->driver = &vmlogrdr_driver;
+		dev->driver_data = priv;
 		/*
 		 * The release function could be called after the
 		 * module has been unloaded. It's _only_ task is to

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 24/38] pm: vmlogrdr power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_vmlogrdr.patch --]
[-- Type: text/plain, Size: 1949 bytes --]

From: Stefan Weinhuber <wein@de.ibm.com>

Signed-off-by: Stefan Weinhuber <wein@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/vmlogrdr.c |   27 ++++++++++++++++++++++++++-
 1 file changed, 26 insertions(+), 1 deletion(-)

Index: linux-2.6/drivers/s390/char/vmlogrdr.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/vmlogrdr.c
+++ linux-2.6/drivers/s390/char/vmlogrdr.c
@@ -3,7 +3,7 @@
  *	character device driver for reading z/VM system service records
  *
  *
- *	Copyright 2004 IBM Corporation
+ *	Copyright IBM Corp. 2004, 2009
  *	character device driver for reading z/VM system service records,
  *	Version 1.0
  *	Author(s): Xenia Tkatschow <xenia@us.ibm.com>
@@ -660,6 +660,29 @@ static struct attribute *vmlogrdr_attrs[
 	NULL,
 };
 
+static int vmlogrdr_pm_prepare(struct device *dev)
+{
+	int rc;
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+
+	rc = 0;
+	if (priv) {
+		spin_lock_bh(&priv->priv_lock);
+		if (priv->dev_in_use)
+			rc = -EBUSY;
+		spin_unlock_bh(&priv->priv_lock);
+	}
+	if (rc)
+		pr_err("vmlogrdr: device %s is busy. Refuse to suspend.\n",
+		       dev_name(dev));
+	return rc;
+}
+
+
+static struct dev_pm_ops vmlogrdr_pm_ops = {
+	.prepare = vmlogrdr_pm_prepare,
+};
+
 static struct attribute_group vmlogrdr_attr_group = {
 	.attrs = vmlogrdr_attrs,
 };
@@ -668,6 +691,7 @@ static struct class *vmlogrdr_class;
 static struct device_driver vmlogrdr_driver = {
 	.name = "vmlogrdr",
 	.bus  = &iucv_bus,
+	.pm = &vmlogrdr_pm_ops,
 };
 
 
@@ -729,6 +753,7 @@ static int vmlogrdr_register_device(stru
 		dev->bus = &iucv_bus;
 		dev->parent = iucv_root;
 		dev->driver = &vmlogrdr_driver;
+		dev->driver_data = priv;
 		/*
 		 * The release function could be called after the
 		 * module has been unloaded. It's _only_ task is to

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 25/38] pm: tape power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_tape.patch --]
[-- Type: text/plain, Size: 5404 bytes --]

From: Frank Munzert <munzert@de.ibm.com>

Signed-off-by: Frank Munzert <munzert@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/tape.h      |    3 +-
 drivers/s390/char/tape_34xx.c |    3 +-
 drivers/s390/char/tape_3590.c |    3 +-
 drivers/s390/char/tape_core.c |   52 +++++++++++++++++++++++++++++++++++++++++-
 4 files changed, 57 insertions(+), 4 deletions(-)

Index: linux-2.6/drivers/s390/char/tape.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape.h
+++ linux-2.6/drivers/s390/char/tape.h
@@ -3,7 +3,7 @@
  *    tape device driver for 3480/3490E/3590 tapes.
  *
  *  S390 and zSeries version
- *    Copyright IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Carsten Otte <cotte@de.ibm.com>
  *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
  *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
@@ -286,6 +286,7 @@ extern void tape_state_set(struct tape_d
 
 extern int tape_generic_online(struct tape_device *, struct tape_discipline *);
 extern int tape_generic_offline(struct ccw_device *);
+extern int tape_generic_pm_suspend(struct ccw_device *);
 
 /* Externals from tape_devmap.c */
 extern int tape_generic_probe(struct ccw_device *);
Index: linux-2.6/drivers/s390/char/tape_34xx.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape_34xx.c
+++ linux-2.6/drivers/s390/char/tape_34xx.c
@@ -2,7 +2,7 @@
  *  drivers/s390/char/tape_34xx.c
  *    tape device discipline for 3480/3490 tapes.
  *
- *    Copyright (C) IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Carsten Otte <cotte@de.ibm.com>
  *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
  *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
@@ -1302,6 +1302,7 @@ static struct ccw_driver tape_34xx_drive
 	.remove = tape_generic_remove,
 	.set_online = tape_34xx_online,
 	.set_offline = tape_generic_offline,
+	.freeze = tape_generic_pm_suspend,
 };
 
 static int
Index: linux-2.6/drivers/s390/char/tape_3590.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape_3590.c
+++ linux-2.6/drivers/s390/char/tape_3590.c
@@ -2,7 +2,7 @@
  *  drivers/s390/char/tape_3590.c
  *    tape device discipline for 3590 tapes.
  *
- *    Copyright IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Stefan Bader <shbader@de.ibm.com>
  *		 Michael Holzheu <holzheu@de.ibm.com>
  *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
@@ -1715,6 +1715,7 @@ static struct ccw_driver tape_3590_drive
 	.remove = tape_generic_remove,
 	.set_offline = tape_generic_offline,
 	.set_online = tape_3590_online,
+	.freeze = tape_generic_pm_suspend,
 };
 
 /*
Index: linux-2.6/drivers/s390/char/tape_core.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape_core.c
+++ linux-2.6/drivers/s390/char/tape_core.c
@@ -3,7 +3,7 @@
  *    basic function of the tape device driver
  *
  *  S390 and zSeries version
- *    Copyright IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Carsten Otte <cotte@de.ibm.com>
  *		 Michael Holzheu <holzheu@de.ibm.com>
  *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
@@ -380,6 +380,55 @@ tape_cleanup_device(struct tape_device *
 }
 
 /*
+ * Suspend device.
+ *
+ * Called by the common I/O layer if the drive should be suspended on user
+ * request. We refuse to suspend if the device is loaded or in use for the
+ * following reason:
+ * While the Linux guest is suspended, it might be logged off which causes
+ * devices to be detached. Tape devices are automatically rewound and unloaded
+ * during DETACH processing (unless the tape device was attached with the
+ * NOASSIGN or MULTIUSER option). After rewind/unload, there is no way to
+ * resume the original state of the tape device, since we would need to
+ * manually re-load the cartridge which was active at suspend time.
+ */
+int tape_generic_pm_suspend(struct ccw_device *cdev)
+{
+	struct tape_device *device;
+
+	device = cdev->dev.driver_data;
+	if (!device) {
+		return -ENODEV;
+	}
+
+	DBF_LH(3, "(%08x): tape_generic_pm_suspend(%p)\n",
+		device->cdev_id, device);
+
+	if (device->medium_state != MS_UNLOADED) {
+		pr_err("A cartridge is loaded in tape device %s, "
+		       "refusing to suspend\n", dev_name(&cdev->dev));
+		return -EBUSY;
+	}
+
+	spin_lock_irq(get_ccwdev_lock(device->cdev));
+	switch (device->tape_state) {
+		case TS_INIT:
+		case TS_NOT_OPER:
+		case TS_UNUSED:
+			spin_unlock_irq(get_ccwdev_lock(device->cdev));
+			break;
+		default:
+			pr_err("Tape device %s is busy, refusing to "
+			       "suspend\n", dev_name(&cdev->dev));
+			spin_unlock_irq(get_ccwdev_lock(device->cdev));
+			return -EBUSY;
+	}
+
+	DBF_LH(3, "(%08x): Drive suspended.\n", device->cdev_id);
+	return 0;
+}
+
+/*
  * Set device offline.
  *
  * Called by the common I/O layer if the drive should set offline on user
@@ -1273,6 +1322,7 @@ EXPORT_SYMBOL(tape_generic_remove);
 EXPORT_SYMBOL(tape_generic_probe);
 EXPORT_SYMBOL(tape_generic_online);
 EXPORT_SYMBOL(tape_generic_offline);
+EXPORT_SYMBOL(tape_generic_pm_suspend);
 EXPORT_SYMBOL(tape_put_device);
 EXPORT_SYMBOL(tape_get_device_reference);
 EXPORT_SYMBOL(tape_state_verbose);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 25/38] pm: tape power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_tape.patch --]
[-- Type: text/plain, Size: 5403 bytes --]

From: Frank Munzert <munzert@de.ibm.com>

Signed-off-by: Frank Munzert <munzert@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/tape.h      |    3 +-
 drivers/s390/char/tape_34xx.c |    3 +-
 drivers/s390/char/tape_3590.c |    3 +-
 drivers/s390/char/tape_core.c |   52 +++++++++++++++++++++++++++++++++++++++++-
 4 files changed, 57 insertions(+), 4 deletions(-)

Index: linux-2.6/drivers/s390/char/tape.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape.h
+++ linux-2.6/drivers/s390/char/tape.h
@@ -3,7 +3,7 @@
  *    tape device driver for 3480/3490E/3590 tapes.
  *
  *  S390 and zSeries version
- *    Copyright IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Carsten Otte <cotte@de.ibm.com>
  *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
  *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
@@ -286,6 +286,7 @@ extern void tape_state_set(struct tape_d
 
 extern int tape_generic_online(struct tape_device *, struct tape_discipline *);
 extern int tape_generic_offline(struct ccw_device *);
+extern int tape_generic_pm_suspend(struct ccw_device *);
 
 /* Externals from tape_devmap.c */
 extern int tape_generic_probe(struct ccw_device *);
Index: linux-2.6/drivers/s390/char/tape_34xx.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape_34xx.c
+++ linux-2.6/drivers/s390/char/tape_34xx.c
@@ -2,7 +2,7 @@
  *  drivers/s390/char/tape_34xx.c
  *    tape device discipline for 3480/3490 tapes.
  *
- *    Copyright (C) IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Carsten Otte <cotte@de.ibm.com>
  *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
  *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
@@ -1302,6 +1302,7 @@ static struct ccw_driver tape_34xx_drive
 	.remove = tape_generic_remove,
 	.set_online = tape_34xx_online,
 	.set_offline = tape_generic_offline,
+	.freeze = tape_generic_pm_suspend,
 };
 
 static int
Index: linux-2.6/drivers/s390/char/tape_3590.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape_3590.c
+++ linux-2.6/drivers/s390/char/tape_3590.c
@@ -2,7 +2,7 @@
  *  drivers/s390/char/tape_3590.c
  *    tape device discipline for 3590 tapes.
  *
- *    Copyright IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Stefan Bader <shbader@de.ibm.com>
  *		 Michael Holzheu <holzheu@de.ibm.com>
  *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
@@ -1715,6 +1715,7 @@ static struct ccw_driver tape_3590_drive
 	.remove = tape_generic_remove,
 	.set_offline = tape_generic_offline,
 	.set_online = tape_3590_online,
+	.freeze = tape_generic_pm_suspend,
 };
 
 /*
Index: linux-2.6/drivers/s390/char/tape_core.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/tape_core.c
+++ linux-2.6/drivers/s390/char/tape_core.c
@@ -3,7 +3,7 @@
  *    basic function of the tape device driver
  *
  *  S390 and zSeries version
- *    Copyright IBM Corp. 2001,2006
+ *    Copyright IBM Corp. 2001, 2009
  *    Author(s): Carsten Otte <cotte@de.ibm.com>
  *		 Michael Holzheu <holzheu@de.ibm.com>
  *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
@@ -380,6 +380,55 @@ tape_cleanup_device(struct tape_device *
 }
 
 /*
+ * Suspend device.
+ *
+ * Called by the common I/O layer if the drive should be suspended on user
+ * request. We refuse to suspend if the device is loaded or in use for the
+ * following reason:
+ * While the Linux guest is suspended, it might be logged off which causes
+ * devices to be detached. Tape devices are automatically rewound and unloaded
+ * during DETACH processing (unless the tape device was attached with the
+ * NOASSIGN or MULTIUSER option). After rewind/unload, there is no way to
+ * resume the original state of the tape device, since we would need to
+ * manually re-load the cartridge which was active at suspend time.
+ */
+int tape_generic_pm_suspend(struct ccw_device *cdev)
+{
+	struct tape_device *device;
+
+	device = cdev->dev.driver_data;
+	if (!device) {
+		return -ENODEV;
+	}
+
+	DBF_LH(3, "(%08x): tape_generic_pm_suspend(%p)\n",
+		device->cdev_id, device);
+
+	if (device->medium_state != MS_UNLOADED) {
+		pr_err("A cartridge is loaded in tape device %s, "
+		       "refusing to suspend\n", dev_name(&cdev->dev));
+		return -EBUSY;
+	}
+
+	spin_lock_irq(get_ccwdev_lock(device->cdev));
+	switch (device->tape_state) {
+		case TS_INIT:
+		case TS_NOT_OPER:
+		case TS_UNUSED:
+			spin_unlock_irq(get_ccwdev_lock(device->cdev));
+			break;
+		default:
+			pr_err("Tape device %s is busy, refusing to "
+			       "suspend\n", dev_name(&cdev->dev));
+			spin_unlock_irq(get_ccwdev_lock(device->cdev));
+			return -EBUSY;
+	}
+
+	DBF_LH(3, "(%08x): Drive suspended.\n", device->cdev_id);
+	return 0;
+}
+
+/*
  * Set device offline.
  *
  * Called by the common I/O layer if the drive should set offline on user
@@ -1273,6 +1322,7 @@ EXPORT_SYMBOL(tape_generic_remove);
 EXPORT_SYMBOL(tape_generic_probe);
 EXPORT_SYMBOL(tape_generic_online);
 EXPORT_SYMBOL(tape_generic_offline);
+EXPORT_SYMBOL(tape_generic_pm_suspend);
 EXPORT_SYMBOL(tape_put_device);
 EXPORT_SYMBOL(tape_get_device_reference);
 EXPORT_SYMBOL(tape_state_verbose);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 26/38] pm: power management support for SCLP drivers.

[Martin Schwidefsky]

[-- Attachment #1: pm_sclp.patch --]
[-- Type: text/plain, Size: 29357 bytes --]

From: Michael Holzheu <holzheu@de.ibm.com>

The SCLP base driver defines a new notifier call back for all upper level SCLP
drivers, like the SCLP console, etc. This guarantees that in suspend first the
upper level drivers are suspended and afterwards the SCLP base driver. For
resume it is the other way round. The SCLP base driver itself registers a
new platform device at the platform bus and gets PM notifications via
the dev_pm_ops.

In suspend, the SCLP base driver switches off the receiver and sender mask 
This is done in sclp_deactivate(). After suspend all new requests will be
rejected with -EIO and no more interrupts will be received, because the masks
are switched off. For resume the sender and receiver masks are reset in
the sclp_reactivate() function.

When the SCLP console is suspended, all new messages are cached in the
sclp console buffers. In resume, all the cached messages are written to the
console. In addition to that we have an early resume function that removes
the cached messages from the suspend image.

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/sclp.c       |  248 ++++++++++++++++++++++++++++++++++++-----
 drivers/s390/char/sclp.h       |   23 ++-
 drivers/s390/char/sclp_con.c   |  139 ++++++++++++++++------
 drivers/s390/char/sclp_rw.c    |   20 ++-
 drivers/s390/char/sclp_rw.h    |   12 -
 drivers/s390/char/sclp_vt220.c |  118 +++++++++++++------
 6 files changed, 432 insertions(+), 128 deletions(-)

Index: linux-2.6/drivers/s390/char/sclp.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp.c
+++ linux-2.6/drivers/s390/char/sclp.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/sclp.c
- *     core function to access sclp interface
+ * core function to access sclp interface
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corp. 1999, 2009
+ *
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #include <linux/module.h>
@@ -16,6 +15,9 @@
 #include <linux/reboot.h>
 #include <linux/jiffies.h>
 #include <linux/init.h>
+#include <linux/suspend.h>
+#include <linux/completion.h>
+#include <linux/platform_device.h>
 #include <asm/types.h>
 #include <asm/s390_ext.h>
 
@@ -47,6 +49,16 @@ static struct sclp_req sclp_init_req;
 static char sclp_read_sccb[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
 static char sclp_init_sccb[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
 
+/* Suspend request */
+static DECLARE_COMPLETION(sclp_request_queue_flushed);
+
+static void sclp_suspend_req_cb(struct sclp_req *req, void *data)
+{
+	complete(&sclp_request_queue_flushed);
+}
+
+static struct sclp_req sclp_suspend_req;
+
 /* Timer for request retries. */
 static struct timer_list sclp_request_timer;
 
@@ -84,6 +96,12 @@ static volatile enum sclp_mask_state_t {
 	sclp_mask_state_initializing
 } sclp_mask_state = sclp_mask_state_idle;
 
+/* Internal state: is the driver suspended? */
+static enum sclp_suspend_state_t {
+	sclp_suspend_state_running,
+	sclp_suspend_state_suspended,
+} sclp_suspend_state = sclp_suspend_state_running;
+
 /* Maximum retry counts */
 #define SCLP_INIT_RETRY		3
 #define SCLP_MASK_RETRY		3
@@ -211,6 +229,8 @@ sclp_process_queue(void)
 	del_timer(&sclp_request_timer);
 	while (!list_empty(&sclp_req_queue)) {
 		req = list_entry(sclp_req_queue.next, struct sclp_req, list);
+		if (!req->sccb)
+			goto do_post;
 		rc = __sclp_start_request(req);
 		if (rc == 0)
 			break;
@@ -222,6 +242,7 @@ sclp_process_queue(void)
 						 sclp_request_timeout, 0);
 			break;
 		}
+do_post:
 		/* Post-processing for aborted request */
 		list_del(&req->list);
 		if (req->callback) {
@@ -233,6 +254,19 @@ sclp_process_queue(void)
 	spin_unlock_irqrestore(&sclp_lock, flags);
 }
 
+static int __sclp_can_add_request(struct sclp_req *req)
+{
+	if (req == &sclp_suspend_req || req == &sclp_init_req)
+		return 1;
+	if (sclp_suspend_state != sclp_suspend_state_running)
+		return 0;
+	if (sclp_init_state != sclp_init_state_initialized)
+		return 0;
+	if (sclp_activation_state != sclp_activation_state_active)
+		return 0;
+	return 1;
+}
+
 /* Queue a new request. Return zero on success, non-zero otherwise. */
 int
 sclp_add_request(struct sclp_req *req)
@@ -241,9 +275,7 @@ sclp_add_request(struct sclp_req *req)
 	int rc;
 
 	spin_lock_irqsave(&sclp_lock, flags);
-	if ((sclp_init_state != sclp_init_state_initialized ||
-	     sclp_activation_state != sclp_activation_state_active) &&
-	    req != &sclp_init_req) {
+	if (!__sclp_can_add_request(req)) {
 		spin_unlock_irqrestore(&sclp_lock, flags);
 		return -EIO;
 	}
@@ -254,10 +286,16 @@ sclp_add_request(struct sclp_req *req)
 	/* Start if request is first in list */
 	if (sclp_running_state == sclp_running_state_idle &&
 	    req->list.prev == &sclp_req_queue) {
+		if (!req->sccb) {
+			list_del(&req->list);
+			rc = -ENODATA;
+			goto out;
+		}
 		rc = __sclp_start_request(req);
 		if (rc)
 			list_del(&req->list);
 	}
+out:
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	return rc;
 }
@@ -560,6 +598,7 @@ sclp_register(struct sclp_register *reg)
 	/* Trigger initial state change callback */
 	reg->sclp_receive_mask = 0;
 	reg->sclp_send_mask = 0;
+	reg->pm_event_posted = 0;
 	list_add(&reg->list, &sclp_reg_list);
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	rc = sclp_init_mask(1);
@@ -880,20 +919,134 @@ static struct notifier_block sclp_reboot
 	.notifier_call = sclp_reboot_event
 };
 
+/*
+ * Suspend/resume SCLP notifier implementation
+ */
+
+static void sclp_pm_event(enum sclp_pm_event sclp_pm_event, int rollback)
+{
+	struct sclp_register *reg;
+	unsigned long flags;
+
+	if (!rollback) {
+		spin_lock_irqsave(&sclp_lock, flags);
+		list_for_each_entry(reg, &sclp_reg_list, list)
+			reg->pm_event_posted = 0;
+		spin_unlock_irqrestore(&sclp_lock, flags);
+	}
+	do {
+		spin_lock_irqsave(&sclp_lock, flags);
+		list_for_each_entry(reg, &sclp_reg_list, list) {
+			if (rollback && reg->pm_event_posted)
+				goto found;
+			if (!rollback && !reg->pm_event_posted)
+				goto found;
+		}
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return;
+found:
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		if (reg->pm_event_fn)
+			reg->pm_event_fn(reg, sclp_pm_event);
+		reg->pm_event_posted = rollback ? 0 : 1;
+	} while (1);
+}
+
+/*
+ * Susend/resume callbacks for platform device
+ */
+
+static int sclp_freeze(struct device *dev)
+{
+	unsigned long flags;
+	int rc;
+
+	sclp_pm_event(SCLP_PM_EVENT_FREEZE, 0);
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	sclp_suspend_state = sclp_suspend_state_suspended;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+
+	/* Init supend data */
+	memset(&sclp_suspend_req, 0, sizeof(sclp_suspend_req));
+	sclp_suspend_req.callback = sclp_suspend_req_cb;
+	sclp_suspend_req.status = SCLP_REQ_FILLED;
+	init_completion(&sclp_request_queue_flushed);
+
+	rc = sclp_add_request(&sclp_suspend_req);
+	if (rc == 0)
+		wait_for_completion(&sclp_request_queue_flushed);
+	else if (rc != -ENODATA)
+		goto fail_thaw;
+
+	rc = sclp_deactivate();
+	if (rc)
+		goto fail_thaw;
+	return 0;
+
+fail_thaw:
+	spin_lock_irqsave(&sclp_lock, flags);
+	sclp_suspend_state = sclp_suspend_state_running;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	sclp_pm_event(SCLP_PM_EVENT_THAW, 1);
+	return rc;
+}
+
+static int sclp_undo_suspend(enum sclp_pm_event event)
+{
+	unsigned long flags;
+	int rc;
+
+	rc = sclp_reactivate();
+	if (rc)
+		return rc;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	sclp_suspend_state = sclp_suspend_state_running;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+
+	sclp_pm_event(event, 0);
+	return 0;
+}
+
+static int sclp_thaw(struct device *dev)
+{
+	return sclp_undo_suspend(SCLP_PM_EVENT_THAW);
+}
+
+static int sclp_restore(struct device *dev)
+{
+	return sclp_undo_suspend(SCLP_PM_EVENT_RESTORE);
+}
+
+static struct dev_pm_ops sclp_pm_ops = {
+	.freeze		= sclp_freeze,
+	.thaw		= sclp_thaw,
+	.restore	= sclp_restore,
+};
+
+static struct platform_driver sclp_pdrv = {
+	.driver = {
+		.name	= "sclp",
+		.owner	= THIS_MODULE,
+		.pm	= &sclp_pm_ops,
+	},
+};
+
+static struct platform_device *sclp_pdev;
+
 /* Initialize SCLP driver. Return zero if driver is operational, non-zero
  * otherwise. */
 static int
 sclp_init(void)
 {
 	unsigned long flags;
-	int rc;
+	int rc = 0;
 
 	spin_lock_irqsave(&sclp_lock, flags);
 	/* Check for previous or running initialization */
-	if (sclp_init_state != sclp_init_state_uninitialized) {
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return 0;
-	}
+	if (sclp_init_state != sclp_init_state_uninitialized)
+		goto fail_unlock;
 	sclp_init_state = sclp_init_state_initializing;
 	/* Set up variables */
 	INIT_LIST_HEAD(&sclp_req_queue);
@@ -904,27 +1057,17 @@ sclp_init(void)
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	rc = sclp_check_interface();
 	spin_lock_irqsave(&sclp_lock, flags);
-	if (rc) {
-		sclp_init_state = sclp_init_state_uninitialized;
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return rc;
-	}
+	if (rc)
+		goto fail_init_state_uninitialized;
 	/* Register reboot handler */
 	rc = register_reboot_notifier(&sclp_reboot_notifier);
-	if (rc) {
-		sclp_init_state = sclp_init_state_uninitialized;
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return rc;
-	}
+	if (rc)
+		goto fail_init_state_uninitialized;
 	/* Register interrupt handler */
 	rc = register_early_external_interrupt(0x2401, sclp_interrupt_handler,
 					       &ext_int_info_hwc);
-	if (rc) {
-		unregister_reboot_notifier(&sclp_reboot_notifier);
-		sclp_init_state = sclp_init_state_uninitialized;
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return rc;
-	}
+	if (rc)
+		goto fail_unregister_reboot_notifier;
 	sclp_init_state = sclp_init_state_initialized;
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	/* Enable service-signal external interruption - needs to happen with
@@ -932,11 +1075,56 @@ sclp_init(void)
 	ctl_set_bit(0, 9);
 	sclp_init_mask(1);
 	return 0;
+
+fail_unregister_reboot_notifier:
+	unregister_reboot_notifier(&sclp_reboot_notifier);
+fail_init_state_uninitialized:
+	sclp_init_state = sclp_init_state_uninitialized;
+fail_unlock:
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
 }
 
+/*
+ * SCLP panic notifier: If we are suspended, we thaw SCLP in order to be able
+ * to print the panic message.
+ */
+static int sclp_panic_notify(struct notifier_block *self,
+			     unsigned long event, void *data)
+{
+	if (sclp_suspend_state == sclp_suspend_state_suspended)
+		sclp_undo_suspend(SCLP_PM_EVENT_THAW);
+	return NOTIFY_OK;
+}
+
+static struct notifier_block sclp_on_panic_nb = {
+	.notifier_call = sclp_panic_notify,
+	.priority = SCLP_PANIC_PRIO,
+};
+
 static __init int sclp_initcall(void)
 {
+	int rc;
+
+	rc = platform_driver_register(&sclp_pdrv);
+	if (rc)
+		return rc;
+	sclp_pdev = platform_device_register_simple("sclp", -1, NULL, 0);
+	rc = IS_ERR(sclp_pdev) ? PTR_ERR(sclp_pdev) : 0;
+	if (rc)
+		goto fail_platform_driver_unregister;
+	rc = atomic_notifier_chain_register(&panic_notifier_list,
+					    &sclp_on_panic_nb);
+	if (rc)
+		goto fail_platform_device_unregister;
+
 	return sclp_init();
+
+fail_platform_device_unregister:
+	platform_device_unregister(sclp_pdev);
+fail_platform_driver_unregister:
+	platform_driver_unregister(&sclp_pdrv);
+	return rc;
 }
 
 arch_initcall(sclp_initcall);
Index: linux-2.6/drivers/s390/char/sclp.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp.h
+++ linux-2.6/drivers/s390/char/sclp.h
@@ -1,10 +1,8 @@
 /*
- *  drivers/s390/char/sclp.h
+ * Copyright IBM Corp. 1999, 2009
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #ifndef __SCLP_H__
@@ -17,7 +15,7 @@
 
 /* maximum number of pages concerning our own memory management */
 #define MAX_KMEM_PAGES (sizeof(unsigned long) << 3)
-#define MAX_CONSOLE_PAGES	4
+#define MAX_CONSOLE_PAGES	6
 
 #define EVTYP_OPCMD		0x01
 #define EVTYP_MSG		0x02
@@ -68,6 +66,15 @@ typedef unsigned int sclp_cmdw_t;
 
 #define GDS_KEY_SELFDEFTEXTMSG	0x31
 
+enum sclp_pm_event {
+	SCLP_PM_EVENT_FREEZE,
+	SCLP_PM_EVENT_THAW,
+	SCLP_PM_EVENT_RESTORE,
+};
+
+#define SCLP_PANIC_PRIO		1
+#define SCLP_PANIC_PRIO_CLIENT	0
+
 typedef u32 sccb_mask_t;	/* ATTENTION: assumes 32bit mask !!! */
 
 struct sccb_header {
@@ -134,6 +141,10 @@ struct sclp_register {
 	void (*state_change_fn)(struct sclp_register *);
 	/* called for events in cp_receive_mask/sclp_receive_mask */
 	void (*receiver_fn)(struct evbuf_header *);
+	/* called for power management events */
+	void (*pm_event_fn)(struct sclp_register *, enum sclp_pm_event);
+	/* pm event posted flag */
+	int pm_event_posted;
 };
 
 /* externals from sclp.c */
Index: linux-2.6/drivers/s390/char/sclp_con.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_con.c
+++ linux-2.6/drivers/s390/char/sclp_con.c
@@ -1,11 +1,9 @@
 /*
- *  drivers/s390/char/sclp_con.c
- *    SCLP line mode console driver
+ * SCLP line mode console driver
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corp. 1999, 2009
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #include <linux/kmod.h>
@@ -32,13 +30,14 @@ static spinlock_t sclp_con_lock;
 static struct list_head sclp_con_pages;
 /* List of full struct sclp_buffer structures ready for output */
 static struct list_head sclp_con_outqueue;
-/* Counter how many buffers are emitted (max 1) and how many */
-/* are on the output queue. */
-static int sclp_con_buffer_count;
 /* Pointer to current console buffer */
 static struct sclp_buffer *sclp_conbuf;
 /* Timer for delayed output of console messages */
 static struct timer_list sclp_con_timer;
+/* Suspend mode flag */
+static int sclp_con_suspended;
+/* Flag that output queue is currently running */
+static int sclp_con_queue_running;
 
 /* Output format for console messages */
 static unsigned short sclp_con_columns;
@@ -53,42 +52,71 @@ sclp_conbuf_callback(struct sclp_buffer 
 	do {
 		page = sclp_unmake_buffer(buffer);
 		spin_lock_irqsave(&sclp_con_lock, flags);
+
 		/* Remove buffer from outqueue */
 		list_del(&buffer->list);
-		sclp_con_buffer_count--;
 		list_add_tail((struct list_head *) page, &sclp_con_pages);
+
 		/* Check if there is a pending buffer on the out queue. */
 		buffer = NULL;
 		if (!list_empty(&sclp_con_outqueue))
-			buffer = list_entry(sclp_con_outqueue.next,
-					    struct sclp_buffer, list);
+			buffer = list_first_entry(&sclp_con_outqueue,
+						  struct sclp_buffer, list);
+		if (!buffer || sclp_con_suspended) {
+			sclp_con_queue_running = 0;
+			spin_unlock_irqrestore(&sclp_con_lock, flags);
+			break;
+		}
 		spin_unlock_irqrestore(&sclp_con_lock, flags);
-	} while (buffer && sclp_emit_buffer(buffer, sclp_conbuf_callback));
+	} while (sclp_emit_buffer(buffer, sclp_conbuf_callback));
 }
 
-static void
-sclp_conbuf_emit(void)
+/*
+ * Finalize and emit first pending buffer.
+ */
+static void sclp_conbuf_emit(void)
 {
 	struct sclp_buffer* buffer;
 	unsigned long flags;
-	int count;
 	int rc;
 
 	spin_lock_irqsave(&sclp_con_lock, flags);
-	buffer = sclp_conbuf;
+	if (sclp_conbuf)
+		list_add_tail(&sclp_conbuf->list, &sclp_con_outqueue);
 	sclp_conbuf = NULL;
-	if (buffer == NULL) {
-		spin_unlock_irqrestore(&sclp_con_lock, flags);
-		return;
-	}
-	list_add_tail(&buffer->list, &sclp_con_outqueue);
-	count = sclp_con_buffer_count++;
+	if (sclp_con_queue_running || sclp_con_suspended)
+		goto out_unlock;
+	if (list_empty(&sclp_con_outqueue))
+		goto out_unlock;
+	buffer = list_first_entry(&sclp_con_outqueue, struct sclp_buffer,
+				  list);
+	sclp_con_queue_running = 1;
 	spin_unlock_irqrestore(&sclp_con_lock, flags);
-	if (count)
-		return;
+
 	rc = sclp_emit_buffer(buffer, sclp_conbuf_callback);
 	if (rc)
 		sclp_conbuf_callback(buffer, rc);
+	return;
+out_unlock:
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
+}
+
+/*
+ * Wait until out queue is empty
+ */
+static void sclp_console_sync_queue(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_con_lock, flags);
+	if (timer_pending(&sclp_con_timer))
+		del_timer_sync(&sclp_con_timer);
+	while (sclp_con_queue_running) {
+		spin_unlock_irqrestore(&sclp_con_lock, flags);
+		sclp_sync_wait();
+		spin_lock_irqsave(&sclp_con_lock, flags);
+	}
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
 }
 
 /*
@@ -123,6 +151,8 @@ sclp_console_write(struct console *conso
 		/* make sure we have a console output buffer */
 		if (sclp_conbuf == NULL) {
 			while (list_empty(&sclp_con_pages)) {
+				if (sclp_con_suspended)
+					goto out;
 				spin_unlock_irqrestore(&sclp_con_lock, flags);
 				sclp_sync_wait();
 				spin_lock_irqsave(&sclp_con_lock, flags);
@@ -157,6 +187,7 @@ sclp_console_write(struct console *conso
 		sclp_con_timer.expires = jiffies + HZ/10;
 		add_timer(&sclp_con_timer);
 	}
+out:
 	spin_unlock_irqrestore(&sclp_con_lock, flags);
 }
 
@@ -168,30 +199,43 @@ sclp_console_device(struct console *c, i
 }
 
 /*
- * This routine is called from panic when the kernel
- * is going to give up. We have to make sure that all buffers
- * will be flushed to the SCLP.
+ * Make sure that all buffers will be flushed to the SCLP.
  */
 static void
 sclp_console_flush(void)
 {
+	sclp_conbuf_emit();
+	sclp_console_sync_queue();
+}
+
+/*
+ * Resume console: If there are cached messages, emit them.
+ */
+static void sclp_console_resume(void)
+{
 	unsigned long flags;
 
+	spin_lock_irqsave(&sclp_con_lock, flags);
+	sclp_con_suspended = 0;
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
 	sclp_conbuf_emit();
+}
+
+/*
+ * Suspend console: Set suspend flag and flush console
+ */
+static void sclp_console_suspend(void)
+{
+	unsigned long flags;
+
 	spin_lock_irqsave(&sclp_con_lock, flags);
-	if (timer_pending(&sclp_con_timer))
-		del_timer(&sclp_con_timer);
-	while (sclp_con_buffer_count > 0) {
-		spin_unlock_irqrestore(&sclp_con_lock, flags);
-		sclp_sync_wait();
-		spin_lock_irqsave(&sclp_con_lock, flags);
-	}
+	sclp_con_suspended = 1;
 	spin_unlock_irqrestore(&sclp_con_lock, flags);
+	sclp_console_flush();
 }
 
-static int
-sclp_console_notify(struct notifier_block *self,
-			  unsigned long event, void *data)
+static int sclp_console_notify(struct notifier_block *self,
+			       unsigned long event, void *data)
 {
 	sclp_console_flush();
 	return NOTIFY_OK;
@@ -199,7 +243,7 @@ sclp_console_notify(struct notifier_bloc
 
 static struct notifier_block on_panic_nb = {
 	.notifier_call = sclp_console_notify,
-	.priority = 1,
+	.priority = SCLP_PANIC_PRIO_CLIENT,
 };
 
 static struct notifier_block on_reboot_nb = {
@@ -221,6 +265,22 @@ static struct console sclp_console =
 };
 
 /*
+ * This function is called for SCLP suspend and resume events.
+ */
+void sclp_console_pm_event(enum sclp_pm_event sclp_pm_event)
+{
+	switch (sclp_pm_event) {
+	case SCLP_PM_EVENT_FREEZE:
+		sclp_console_suspend();
+		break;
+	case SCLP_PM_EVENT_RESTORE:
+	case SCLP_PM_EVENT_THAW:
+		sclp_console_resume();
+		break;
+	}
+}
+
+/*
  * called by console_init() in drivers/char/tty_io.c at boot-time.
  */
 static int __init
@@ -243,7 +303,6 @@ sclp_console_init(void)
 	}
 	INIT_LIST_HEAD(&sclp_con_outqueue);
 	spin_lock_init(&sclp_con_lock);
-	sclp_con_buffer_count = 0;
 	sclp_conbuf = NULL;
 	init_timer(&sclp_con_timer);
 
Index: linux-2.6/drivers/s390/char/sclp_rw.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_rw.h
+++ linux-2.6/drivers/s390/char/sclp_rw.h
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/sclp_rw.h
- *    interface to the SCLP-read/write driver
+ * interface to the SCLP-read/write driver
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corporation 1999, 2009
+ *
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #ifndef __SCLP_RW_H__
@@ -93,4 +92,5 @@ void sclp_set_columns(struct sclp_buffer
 void sclp_set_htab(struct sclp_buffer *, unsigned short);
 int sclp_chars_in_buffer(struct sclp_buffer *);
 
+void sclp_console_pm_event(enum sclp_pm_event sclp_pm_event);
 #endif	/* __SCLP_RW_H__ */
Index: linux-2.6/drivers/s390/char/sclp_rw.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_rw.c
+++ linux-2.6/drivers/s390/char/sclp_rw.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/sclp_rw.c
- *     driver: reading from and writing to system console on S/390 via SCLP
+ * driver: reading from and writing to system console on S/390 via SCLP
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corp. 1999, 2009
+ *
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #include <linux/kmod.h>
@@ -26,9 +25,16 @@
  */
 #define MAX_SCCB_ROOM (PAGE_SIZE - sizeof(struct sclp_buffer))
 
+static void sclp_rw_pm_event(struct sclp_register *reg,
+			     enum sclp_pm_event sclp_pm_event)
+{
+	sclp_console_pm_event(sclp_pm_event);
+}
+
 /* Event type structure for write message and write priority message */
 static struct sclp_register sclp_rw_event = {
-	.send_mask = EVTYP_MSG_MASK | EVTYP_PMSGCMD_MASK
+	.send_mask = EVTYP_MSG_MASK | EVTYP_PMSGCMD_MASK,
+	.pm_event_fn = sclp_rw_pm_event,
 };
 
 /*
Index: linux-2.6/drivers/s390/char/sclp_vt220.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_vt220.c
+++ linux-2.6/drivers/s390/char/sclp_vt220.c
@@ -1,10 +1,9 @@
 /*
- *  drivers/s390/char/sclp_vt220.c
- *    SCLP VT220 terminal driver.
+ * SCLP VT220 terminal driver.
  *
- *  S390 version
- *    Copyright IBM Corp. 2003,2008
- *    Author(s): Peter Oberparleiter <Peter.Oberparleiter@de.ibm.com>
+ * Copyright IBM Corp. 2003, 2009
+ *
+ * Author(s): Peter Oberparleiter <Peter.Oberparleiter@de.ibm.com>
  */
 
 #include <linux/module.h>
@@ -69,8 +68,11 @@ static struct list_head sclp_vt220_empty
 /* List of pending requests */
 static struct list_head sclp_vt220_outqueue;
 
-/* Number of requests in outqueue */
-static int sclp_vt220_outqueue_count;
+/* Suspend mode flag */
+static int sclp_vt220_suspended;
+
+/* Flag that output queue is currently running */
+static int sclp_vt220_queue_running;
 
 /* Timer used for delaying write requests to merge subsequent messages into
  * a single buffer */
@@ -92,6 +94,8 @@ static int __initdata sclp_vt220_init_co
 static int sclp_vt220_flush_later;
 
 static void sclp_vt220_receiver_fn(struct evbuf_header *evbuf);
+static void sclp_vt220_pm_event_fn(struct sclp_register *reg,
+				   enum sclp_pm_event sclp_pm_event);
 static int __sclp_vt220_emit(struct sclp_vt220_request *request);
 static void sclp_vt220_emit_current(void);
 
@@ -100,7 +104,8 @@ static struct sclp_register sclp_vt220_r
 	.send_mask		= EVTYP_VT220MSG_MASK,
 	.receive_mask		= EVTYP_VT220MSG_MASK,
 	.state_change_fn	= NULL,
-	.receiver_fn		= sclp_vt220_receiver_fn
+	.receiver_fn		= sclp_vt220_receiver_fn,
+	.pm_event_fn		= sclp_vt220_pm_event_fn,
 };
 
 
@@ -120,15 +125,19 @@ sclp_vt220_process_queue(struct sclp_vt2
 		spin_lock_irqsave(&sclp_vt220_lock, flags);
 		/* Move request from outqueue to empty queue */
 		list_del(&request->list);
-		sclp_vt220_outqueue_count--;
 		list_add_tail((struct list_head *) page, &sclp_vt220_empty);
 		/* Check if there is a pending buffer on the out queue. */
 		request = NULL;
 		if (!list_empty(&sclp_vt220_outqueue))
 			request = list_entry(sclp_vt220_outqueue.next,
 					     struct sclp_vt220_request, list);
+		if (!request || sclp_vt220_suspended) {
+			sclp_vt220_queue_running = 0;
+			spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+			break;
+		}
 		spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-	} while (request && __sclp_vt220_emit(request));
+	} while (__sclp_vt220_emit(request));
 	if (request == NULL && sclp_vt220_flush_later)
 		sclp_vt220_emit_current();
 	/* Check if the tty needs a wake up call */
@@ -212,26 +221,7 @@ __sclp_vt220_emit(struct sclp_vt220_requ
 }
 
 /*
- * Queue and emit given request.
- */
-static void
-sclp_vt220_emit(struct sclp_vt220_request *request)
-{
-	unsigned long flags;
-	int count;
-
-	spin_lock_irqsave(&sclp_vt220_lock, flags);
-	list_add_tail(&request->list, &sclp_vt220_outqueue);
-	count = sclp_vt220_outqueue_count++;
-	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-	/* Emit only the first buffer immediately - callback takes care of
-	 * the rest */
-	if (count == 0 && __sclp_vt220_emit(request))
-		sclp_vt220_process_queue(request);
-}
-
-/*
- * Queue and emit current request. Return zero on success, non-zero otherwise.
+ * Queue and emit current request.
  */
 static void
 sclp_vt220_emit_current(void)
@@ -241,22 +231,33 @@ sclp_vt220_emit_current(void)
 	struct sclp_vt220_sccb *sccb;
 
 	spin_lock_irqsave(&sclp_vt220_lock, flags);
-	request = NULL;
-	if (sclp_vt220_current_request != NULL) {
+	if (sclp_vt220_current_request) {
 		sccb = (struct sclp_vt220_sccb *) 
 				sclp_vt220_current_request->sclp_req.sccb;
 		/* Only emit buffers with content */
 		if (sccb->header.length != sizeof(struct sclp_vt220_sccb)) {
-			request = sclp_vt220_current_request;
+			list_add_tail(&sclp_vt220_current_request->list,
+				      &sclp_vt220_outqueue);
 			sclp_vt220_current_request = NULL;
 			if (timer_pending(&sclp_vt220_timer))
 				del_timer(&sclp_vt220_timer);
 		}
 		sclp_vt220_flush_later = 0;
 	}
+	if (sclp_vt220_queue_running || sclp_vt220_suspended)
+		goto out_unlock;
+	if (list_empty(&sclp_vt220_outqueue))
+		goto out_unlock;
+	request = list_first_entry(&sclp_vt220_outqueue,
+				   struct sclp_vt220_request, list);
+	sclp_vt220_queue_running = 1;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+
+	if (__sclp_vt220_emit(request))
+		sclp_vt220_process_queue(request);
+	return;
+out_unlock:
 	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-	if (request != NULL)
-		sclp_vt220_emit(request);
 }
 
 #define SCLP_NORMAL_WRITE	0x00
@@ -396,7 +397,7 @@ __sclp_vt220_write(const unsigned char *
 		if (sclp_vt220_current_request == NULL) {
 			while (list_empty(&sclp_vt220_empty)) {
 				spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-				if (may_fail)
+				if (may_fail || sclp_vt220_suspended)
 					goto out;
 				else
 					sclp_sync_wait();
@@ -531,7 +532,7 @@ sclp_vt220_put_char(struct tty_struct *t
 static void
 sclp_vt220_flush_chars(struct tty_struct *tty)
 {
-	if (sclp_vt220_outqueue_count == 0)
+	if (!sclp_vt220_queue_running)
 		sclp_vt220_emit_current();
 	else
 		sclp_vt220_flush_later = 1;
@@ -635,7 +636,6 @@ static int __init __sclp_vt220_init(int 
 	init_timer(&sclp_vt220_timer);
 	sclp_vt220_current_request = NULL;
 	sclp_vt220_buffered_chars = 0;
-	sclp_vt220_outqueue_count = 0;
 	sclp_vt220_tty = NULL;
 	sclp_vt220_flush_later = 0;
 
@@ -736,7 +736,7 @@ static void __sclp_vt220_flush_buffer(vo
 	spin_lock_irqsave(&sclp_vt220_lock, flags);
 	if (timer_pending(&sclp_vt220_timer))
 		del_timer(&sclp_vt220_timer);
-	while (sclp_vt220_outqueue_count > 0) {
+	while (sclp_vt220_queue_running) {
 		spin_unlock_irqrestore(&sclp_vt220_lock, flags);
 		sclp_sync_wait();
 		spin_lock_irqsave(&sclp_vt220_lock, flags);
@@ -744,6 +744,46 @@ static void __sclp_vt220_flush_buffer(vo
 	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
 }
 
+/*
+ * Resume console: If there are cached messages, emit them.
+ */
+static void sclp_vt220_resume(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	sclp_vt220_suspended = 0;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	sclp_vt220_emit_current();
+}
+
+/*
+ * Suspend console: Set suspend flag and flush console
+ */
+static void sclp_vt220_suspend(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	sclp_vt220_suspended = 1;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	__sclp_vt220_flush_buffer();
+}
+
+static void sclp_vt220_pm_event_fn(struct sclp_register *reg,
+				   enum sclp_pm_event sclp_pm_event)
+{
+	switch (sclp_pm_event) {
+	case SCLP_PM_EVENT_FREEZE:
+		sclp_vt220_suspend();
+		break;
+	case SCLP_PM_EVENT_RESTORE:
+	case SCLP_PM_EVENT_THAW:
+		sclp_vt220_resume();
+		break;
+	}
+}
+
 static int
 sclp_vt220_notify(struct notifier_block *self,
 			  unsigned long event, void *data)

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 26/38] pm: power management support for SCLP drivers.

[Martin Schwidefsky]

[-- Attachment #1: pm_sclp.patch --]
[-- Type: text/plain, Size: 29356 bytes --]

From: Michael Holzheu <holzheu@de.ibm.com>

The SCLP base driver defines a new notifier call back for all upper level SCLP
drivers, like the SCLP console, etc. This guarantees that in suspend first the
upper level drivers are suspended and afterwards the SCLP base driver. For
resume it is the other way round. The SCLP base driver itself registers a
new platform device at the platform bus and gets PM notifications via
the dev_pm_ops.

In suspend, the SCLP base driver switches off the receiver and sender mask 
This is done in sclp_deactivate(). After suspend all new requests will be
rejected with -EIO and no more interrupts will be received, because the masks
are switched off. For resume the sender and receiver masks are reset in
the sclp_reactivate() function.

When the SCLP console is suspended, all new messages are cached in the
sclp console buffers. In resume, all the cached messages are written to the
console. In addition to that we have an early resume function that removes
the cached messages from the suspend image.

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/sclp.c       |  248 ++++++++++++++++++++++++++++++++++++-----
 drivers/s390/char/sclp.h       |   23 ++-
 drivers/s390/char/sclp_con.c   |  139 ++++++++++++++++------
 drivers/s390/char/sclp_rw.c    |   20 ++-
 drivers/s390/char/sclp_rw.h    |   12 -
 drivers/s390/char/sclp_vt220.c |  118 +++++++++++++------
 6 files changed, 432 insertions(+), 128 deletions(-)

Index: linux-2.6/drivers/s390/char/sclp.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp.c
+++ linux-2.6/drivers/s390/char/sclp.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/sclp.c
- *     core function to access sclp interface
+ * core function to access sclp interface
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corp. 1999, 2009
+ *
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #include <linux/module.h>
@@ -16,6 +15,9 @@
 #include <linux/reboot.h>
 #include <linux/jiffies.h>
 #include <linux/init.h>
+#include <linux/suspend.h>
+#include <linux/completion.h>
+#include <linux/platform_device.h>
 #include <asm/types.h>
 #include <asm/s390_ext.h>
 
@@ -47,6 +49,16 @@ static struct sclp_req sclp_init_req;
 static char sclp_read_sccb[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
 static char sclp_init_sccb[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
 
+/* Suspend request */
+static DECLARE_COMPLETION(sclp_request_queue_flushed);
+
+static void sclp_suspend_req_cb(struct sclp_req *req, void *data)
+{
+	complete(&sclp_request_queue_flushed);
+}
+
+static struct sclp_req sclp_suspend_req;
+
 /* Timer for request retries. */
 static struct timer_list sclp_request_timer;
 
@@ -84,6 +96,12 @@ static volatile enum sclp_mask_state_t {
 	sclp_mask_state_initializing
 } sclp_mask_state = sclp_mask_state_idle;
 
+/* Internal state: is the driver suspended? */
+static enum sclp_suspend_state_t {
+	sclp_suspend_state_running,
+	sclp_suspend_state_suspended,
+} sclp_suspend_state = sclp_suspend_state_running;
+
 /* Maximum retry counts */
 #define SCLP_INIT_RETRY		3
 #define SCLP_MASK_RETRY		3
@@ -211,6 +229,8 @@ sclp_process_queue(void)
 	del_timer(&sclp_request_timer);
 	while (!list_empty(&sclp_req_queue)) {
 		req = list_entry(sclp_req_queue.next, struct sclp_req, list);
+		if (!req->sccb)
+			goto do_post;
 		rc = __sclp_start_request(req);
 		if (rc == 0)
 			break;
@@ -222,6 +242,7 @@ sclp_process_queue(void)
 						 sclp_request_timeout, 0);
 			break;
 		}
+do_post:
 		/* Post-processing for aborted request */
 		list_del(&req->list);
 		if (req->callback) {
@@ -233,6 +254,19 @@ sclp_process_queue(void)
 	spin_unlock_irqrestore(&sclp_lock, flags);
 }
 
+static int __sclp_can_add_request(struct sclp_req *req)
+{
+	if (req == &sclp_suspend_req || req == &sclp_init_req)
+		return 1;
+	if (sclp_suspend_state != sclp_suspend_state_running)
+		return 0;
+	if (sclp_init_state != sclp_init_state_initialized)
+		return 0;
+	if (sclp_activation_state != sclp_activation_state_active)
+		return 0;
+	return 1;
+}
+
 /* Queue a new request. Return zero on success, non-zero otherwise. */
 int
 sclp_add_request(struct sclp_req *req)
@@ -241,9 +275,7 @@ sclp_add_request(struct sclp_req *req)
 	int rc;
 
 	spin_lock_irqsave(&sclp_lock, flags);
-	if ((sclp_init_state != sclp_init_state_initialized ||
-	     sclp_activation_state != sclp_activation_state_active) &&
-	    req != &sclp_init_req) {
+	if (!__sclp_can_add_request(req)) {
 		spin_unlock_irqrestore(&sclp_lock, flags);
 		return -EIO;
 	}
@@ -254,10 +286,16 @@ sclp_add_request(struct sclp_req *req)
 	/* Start if request is first in list */
 	if (sclp_running_state == sclp_running_state_idle &&
 	    req->list.prev == &sclp_req_queue) {
+		if (!req->sccb) {
+			list_del(&req->list);
+			rc = -ENODATA;
+			goto out;
+		}
 		rc = __sclp_start_request(req);
 		if (rc)
 			list_del(&req->list);
 	}
+out:
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	return rc;
 }
@@ -560,6 +598,7 @@ sclp_register(struct sclp_register *reg)
 	/* Trigger initial state change callback */
 	reg->sclp_receive_mask = 0;
 	reg->sclp_send_mask = 0;
+	reg->pm_event_posted = 0;
 	list_add(&reg->list, &sclp_reg_list);
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	rc = sclp_init_mask(1);
@@ -880,20 +919,134 @@ static struct notifier_block sclp_reboot
 	.notifier_call = sclp_reboot_event
 };
 
+/*
+ * Suspend/resume SCLP notifier implementation
+ */
+
+static void sclp_pm_event(enum sclp_pm_event sclp_pm_event, int rollback)
+{
+	struct sclp_register *reg;
+	unsigned long flags;
+
+	if (!rollback) {
+		spin_lock_irqsave(&sclp_lock, flags);
+		list_for_each_entry(reg, &sclp_reg_list, list)
+			reg->pm_event_posted = 0;
+		spin_unlock_irqrestore(&sclp_lock, flags);
+	}
+	do {
+		spin_lock_irqsave(&sclp_lock, flags);
+		list_for_each_entry(reg, &sclp_reg_list, list) {
+			if (rollback && reg->pm_event_posted)
+				goto found;
+			if (!rollback && !reg->pm_event_posted)
+				goto found;
+		}
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return;
+found:
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		if (reg->pm_event_fn)
+			reg->pm_event_fn(reg, sclp_pm_event);
+		reg->pm_event_posted = rollback ? 0 : 1;
+	} while (1);
+}
+
+/*
+ * Susend/resume callbacks for platform device
+ */
+
+static int sclp_freeze(struct device *dev)
+{
+	unsigned long flags;
+	int rc;
+
+	sclp_pm_event(SCLP_PM_EVENT_FREEZE, 0);
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	sclp_suspend_state = sclp_suspend_state_suspended;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+
+	/* Init supend data */
+	memset(&sclp_suspend_req, 0, sizeof(sclp_suspend_req));
+	sclp_suspend_req.callback = sclp_suspend_req_cb;
+	sclp_suspend_req.status = SCLP_REQ_FILLED;
+	init_completion(&sclp_request_queue_flushed);
+
+	rc = sclp_add_request(&sclp_suspend_req);
+	if (rc == 0)
+		wait_for_completion(&sclp_request_queue_flushed);
+	else if (rc != -ENODATA)
+		goto fail_thaw;
+
+	rc = sclp_deactivate();
+	if (rc)
+		goto fail_thaw;
+	return 0;
+
+fail_thaw:
+	spin_lock_irqsave(&sclp_lock, flags);
+	sclp_suspend_state = sclp_suspend_state_running;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	sclp_pm_event(SCLP_PM_EVENT_THAW, 1);
+	return rc;
+}
+
+static int sclp_undo_suspend(enum sclp_pm_event event)
+{
+	unsigned long flags;
+	int rc;
+
+	rc = sclp_reactivate();
+	if (rc)
+		return rc;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	sclp_suspend_state = sclp_suspend_state_running;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+
+	sclp_pm_event(event, 0);
+	return 0;
+}
+
+static int sclp_thaw(struct device *dev)
+{
+	return sclp_undo_suspend(SCLP_PM_EVENT_THAW);
+}
+
+static int sclp_restore(struct device *dev)
+{
+	return sclp_undo_suspend(SCLP_PM_EVENT_RESTORE);
+}
+
+static struct dev_pm_ops sclp_pm_ops = {
+	.freeze		= sclp_freeze,
+	.thaw		= sclp_thaw,
+	.restore	= sclp_restore,
+};
+
+static struct platform_driver sclp_pdrv = {
+	.driver = {
+		.name	= "sclp",
+		.owner	= THIS_MODULE,
+		.pm	= &sclp_pm_ops,
+	},
+};
+
+static struct platform_device *sclp_pdev;
+
 /* Initialize SCLP driver. Return zero if driver is operational, non-zero
  * otherwise. */
 static int
 sclp_init(void)
 {
 	unsigned long flags;
-	int rc;
+	int rc = 0;
 
 	spin_lock_irqsave(&sclp_lock, flags);
 	/* Check for previous or running initialization */
-	if (sclp_init_state != sclp_init_state_uninitialized) {
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return 0;
-	}
+	if (sclp_init_state != sclp_init_state_uninitialized)
+		goto fail_unlock;
 	sclp_init_state = sclp_init_state_initializing;
 	/* Set up variables */
 	INIT_LIST_HEAD(&sclp_req_queue);
@@ -904,27 +1057,17 @@ sclp_init(void)
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	rc = sclp_check_interface();
 	spin_lock_irqsave(&sclp_lock, flags);
-	if (rc) {
-		sclp_init_state = sclp_init_state_uninitialized;
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return rc;
-	}
+	if (rc)
+		goto fail_init_state_uninitialized;
 	/* Register reboot handler */
 	rc = register_reboot_notifier(&sclp_reboot_notifier);
-	if (rc) {
-		sclp_init_state = sclp_init_state_uninitialized;
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return rc;
-	}
+	if (rc)
+		goto fail_init_state_uninitialized;
 	/* Register interrupt handler */
 	rc = register_early_external_interrupt(0x2401, sclp_interrupt_handler,
 					       &ext_int_info_hwc);
-	if (rc) {
-		unregister_reboot_notifier(&sclp_reboot_notifier);
-		sclp_init_state = sclp_init_state_uninitialized;
-		spin_unlock_irqrestore(&sclp_lock, flags);
-		return rc;
-	}
+	if (rc)
+		goto fail_unregister_reboot_notifier;
 	sclp_init_state = sclp_init_state_initialized;
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	/* Enable service-signal external interruption - needs to happen with
@@ -932,11 +1075,56 @@ sclp_init(void)
 	ctl_set_bit(0, 9);
 	sclp_init_mask(1);
 	return 0;
+
+fail_unregister_reboot_notifier:
+	unregister_reboot_notifier(&sclp_reboot_notifier);
+fail_init_state_uninitialized:
+	sclp_init_state = sclp_init_state_uninitialized;
+fail_unlock:
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
 }
 
+/*
+ * SCLP panic notifier: If we are suspended, we thaw SCLP in order to be able
+ * to print the panic message.
+ */
+static int sclp_panic_notify(struct notifier_block *self,
+			     unsigned long event, void *data)
+{
+	if (sclp_suspend_state == sclp_suspend_state_suspended)
+		sclp_undo_suspend(SCLP_PM_EVENT_THAW);
+	return NOTIFY_OK;
+}
+
+static struct notifier_block sclp_on_panic_nb = {
+	.notifier_call = sclp_panic_notify,
+	.priority = SCLP_PANIC_PRIO,
+};
+
 static __init int sclp_initcall(void)
 {
+	int rc;
+
+	rc = platform_driver_register(&sclp_pdrv);
+	if (rc)
+		return rc;
+	sclp_pdev = platform_device_register_simple("sclp", -1, NULL, 0);
+	rc = IS_ERR(sclp_pdev) ? PTR_ERR(sclp_pdev) : 0;
+	if (rc)
+		goto fail_platform_driver_unregister;
+	rc = atomic_notifier_chain_register(&panic_notifier_list,
+					    &sclp_on_panic_nb);
+	if (rc)
+		goto fail_platform_device_unregister;
+
 	return sclp_init();
+
+fail_platform_device_unregister:
+	platform_device_unregister(sclp_pdev);
+fail_platform_driver_unregister:
+	platform_driver_unregister(&sclp_pdrv);
+	return rc;
 }
 
 arch_initcall(sclp_initcall);
Index: linux-2.6/drivers/s390/char/sclp.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp.h
+++ linux-2.6/drivers/s390/char/sclp.h
@@ -1,10 +1,8 @@
 /*
- *  drivers/s390/char/sclp.h
+ * Copyright IBM Corp. 1999, 2009
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #ifndef __SCLP_H__
@@ -17,7 +15,7 @@
 
 /* maximum number of pages concerning our own memory management */
 #define MAX_KMEM_PAGES (sizeof(unsigned long) << 3)
-#define MAX_CONSOLE_PAGES	4
+#define MAX_CONSOLE_PAGES	6
 
 #define EVTYP_OPCMD		0x01
 #define EVTYP_MSG		0x02
@@ -68,6 +66,15 @@ typedef unsigned int sclp_cmdw_t;
 
 #define GDS_KEY_SELFDEFTEXTMSG	0x31
 
+enum sclp_pm_event {
+	SCLP_PM_EVENT_FREEZE,
+	SCLP_PM_EVENT_THAW,
+	SCLP_PM_EVENT_RESTORE,
+};
+
+#define SCLP_PANIC_PRIO		1
+#define SCLP_PANIC_PRIO_CLIENT	0
+
 typedef u32 sccb_mask_t;	/* ATTENTION: assumes 32bit mask !!! */
 
 struct sccb_header {
@@ -134,6 +141,10 @@ struct sclp_register {
 	void (*state_change_fn)(struct sclp_register *);
 	/* called for events in cp_receive_mask/sclp_receive_mask */
 	void (*receiver_fn)(struct evbuf_header *);
+	/* called for power management events */
+	void (*pm_event_fn)(struct sclp_register *, enum sclp_pm_event);
+	/* pm event posted flag */
+	int pm_event_posted;
 };
 
 /* externals from sclp.c */
Index: linux-2.6/drivers/s390/char/sclp_con.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_con.c
+++ linux-2.6/drivers/s390/char/sclp_con.c
@@ -1,11 +1,9 @@
 /*
- *  drivers/s390/char/sclp_con.c
- *    SCLP line mode console driver
+ * SCLP line mode console driver
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corp. 1999, 2009
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #include <linux/kmod.h>
@@ -32,13 +30,14 @@ static spinlock_t sclp_con_lock;
 static struct list_head sclp_con_pages;
 /* List of full struct sclp_buffer structures ready for output */
 static struct list_head sclp_con_outqueue;
-/* Counter how many buffers are emitted (max 1) and how many */
-/* are on the output queue. */
-static int sclp_con_buffer_count;
 /* Pointer to current console buffer */
 static struct sclp_buffer *sclp_conbuf;
 /* Timer for delayed output of console messages */
 static struct timer_list sclp_con_timer;
+/* Suspend mode flag */
+static int sclp_con_suspended;
+/* Flag that output queue is currently running */
+static int sclp_con_queue_running;
 
 /* Output format for console messages */
 static unsigned short sclp_con_columns;
@@ -53,42 +52,71 @@ sclp_conbuf_callback(struct sclp_buffer 
 	do {
 		page = sclp_unmake_buffer(buffer);
 		spin_lock_irqsave(&sclp_con_lock, flags);
+
 		/* Remove buffer from outqueue */
 		list_del(&buffer->list);
-		sclp_con_buffer_count--;
 		list_add_tail((struct list_head *) page, &sclp_con_pages);
+
 		/* Check if there is a pending buffer on the out queue. */
 		buffer = NULL;
 		if (!list_empty(&sclp_con_outqueue))
-			buffer = list_entry(sclp_con_outqueue.next,
-					    struct sclp_buffer, list);
+			buffer = list_first_entry(&sclp_con_outqueue,
+						  struct sclp_buffer, list);
+		if (!buffer || sclp_con_suspended) {
+			sclp_con_queue_running = 0;
+			spin_unlock_irqrestore(&sclp_con_lock, flags);
+			break;
+		}
 		spin_unlock_irqrestore(&sclp_con_lock, flags);
-	} while (buffer && sclp_emit_buffer(buffer, sclp_conbuf_callback));
+	} while (sclp_emit_buffer(buffer, sclp_conbuf_callback));
 }
 
-static void
-sclp_conbuf_emit(void)
+/*
+ * Finalize and emit first pending buffer.
+ */
+static void sclp_conbuf_emit(void)
 {
 	struct sclp_buffer* buffer;
 	unsigned long flags;
-	int count;
 	int rc;
 
 	spin_lock_irqsave(&sclp_con_lock, flags);
-	buffer = sclp_conbuf;
+	if (sclp_conbuf)
+		list_add_tail(&sclp_conbuf->list, &sclp_con_outqueue);
 	sclp_conbuf = NULL;
-	if (buffer == NULL) {
-		spin_unlock_irqrestore(&sclp_con_lock, flags);
-		return;
-	}
-	list_add_tail(&buffer->list, &sclp_con_outqueue);
-	count = sclp_con_buffer_count++;
+	if (sclp_con_queue_running || sclp_con_suspended)
+		goto out_unlock;
+	if (list_empty(&sclp_con_outqueue))
+		goto out_unlock;
+	buffer = list_first_entry(&sclp_con_outqueue, struct sclp_buffer,
+				  list);
+	sclp_con_queue_running = 1;
 	spin_unlock_irqrestore(&sclp_con_lock, flags);
-	if (count)
-		return;
+
 	rc = sclp_emit_buffer(buffer, sclp_conbuf_callback);
 	if (rc)
 		sclp_conbuf_callback(buffer, rc);
+	return;
+out_unlock:
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
+}
+
+/*
+ * Wait until out queue is empty
+ */
+static void sclp_console_sync_queue(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_con_lock, flags);
+	if (timer_pending(&sclp_con_timer))
+		del_timer_sync(&sclp_con_timer);
+	while (sclp_con_queue_running) {
+		spin_unlock_irqrestore(&sclp_con_lock, flags);
+		sclp_sync_wait();
+		spin_lock_irqsave(&sclp_con_lock, flags);
+	}
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
 }
 
 /*
@@ -123,6 +151,8 @@ sclp_console_write(struct console *conso
 		/* make sure we have a console output buffer */
 		if (sclp_conbuf == NULL) {
 			while (list_empty(&sclp_con_pages)) {
+				if (sclp_con_suspended)
+					goto out;
 				spin_unlock_irqrestore(&sclp_con_lock, flags);
 				sclp_sync_wait();
 				spin_lock_irqsave(&sclp_con_lock, flags);
@@ -157,6 +187,7 @@ sclp_console_write(struct console *conso
 		sclp_con_timer.expires = jiffies + HZ/10;
 		add_timer(&sclp_con_timer);
 	}
+out:
 	spin_unlock_irqrestore(&sclp_con_lock, flags);
 }
 
@@ -168,30 +199,43 @@ sclp_console_device(struct console *c, i
 }
 
 /*
- * This routine is called from panic when the kernel
- * is going to give up. We have to make sure that all buffers
- * will be flushed to the SCLP.
+ * Make sure that all buffers will be flushed to the SCLP.
  */
 static void
 sclp_console_flush(void)
 {
+	sclp_conbuf_emit();
+	sclp_console_sync_queue();
+}
+
+/*
+ * Resume console: If there are cached messages, emit them.
+ */
+static void sclp_console_resume(void)
+{
 	unsigned long flags;
 
+	spin_lock_irqsave(&sclp_con_lock, flags);
+	sclp_con_suspended = 0;
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
 	sclp_conbuf_emit();
+}
+
+/*
+ * Suspend console: Set suspend flag and flush console
+ */
+static void sclp_console_suspend(void)
+{
+	unsigned long flags;
+
 	spin_lock_irqsave(&sclp_con_lock, flags);
-	if (timer_pending(&sclp_con_timer))
-		del_timer(&sclp_con_timer);
-	while (sclp_con_buffer_count > 0) {
-		spin_unlock_irqrestore(&sclp_con_lock, flags);
-		sclp_sync_wait();
-		spin_lock_irqsave(&sclp_con_lock, flags);
-	}
+	sclp_con_suspended = 1;
 	spin_unlock_irqrestore(&sclp_con_lock, flags);
+	sclp_console_flush();
 }
 
-static int
-sclp_console_notify(struct notifier_block *self,
-			  unsigned long event, void *data)
+static int sclp_console_notify(struct notifier_block *self,
+			       unsigned long event, void *data)
 {
 	sclp_console_flush();
 	return NOTIFY_OK;
@@ -199,7 +243,7 @@ sclp_console_notify(struct notifier_bloc
 
 static struct notifier_block on_panic_nb = {
 	.notifier_call = sclp_console_notify,
-	.priority = 1,
+	.priority = SCLP_PANIC_PRIO_CLIENT,
 };
 
 static struct notifier_block on_reboot_nb = {
@@ -221,6 +265,22 @@ static struct console sclp_console =
 };
 
 /*
+ * This function is called for SCLP suspend and resume events.
+ */
+void sclp_console_pm_event(enum sclp_pm_event sclp_pm_event)
+{
+	switch (sclp_pm_event) {
+	case SCLP_PM_EVENT_FREEZE:
+		sclp_console_suspend();
+		break;
+	case SCLP_PM_EVENT_RESTORE:
+	case SCLP_PM_EVENT_THAW:
+		sclp_console_resume();
+		break;
+	}
+}
+
+/*
  * called by console_init() in drivers/char/tty_io.c at boot-time.
  */
 static int __init
@@ -243,7 +303,6 @@ sclp_console_init(void)
 	}
 	INIT_LIST_HEAD(&sclp_con_outqueue);
 	spin_lock_init(&sclp_con_lock);
-	sclp_con_buffer_count = 0;
 	sclp_conbuf = NULL;
 	init_timer(&sclp_con_timer);
 
Index: linux-2.6/drivers/s390/char/sclp_rw.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_rw.h
+++ linux-2.6/drivers/s390/char/sclp_rw.h
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/sclp_rw.h
- *    interface to the SCLP-read/write driver
+ * interface to the SCLP-read/write driver
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corporation 1999, 2009
+ *
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #ifndef __SCLP_RW_H__
@@ -93,4 +92,5 @@ void sclp_set_columns(struct sclp_buffer
 void sclp_set_htab(struct sclp_buffer *, unsigned short);
 int sclp_chars_in_buffer(struct sclp_buffer *);
 
+void sclp_console_pm_event(enum sclp_pm_event sclp_pm_event);
 #endif	/* __SCLP_RW_H__ */
Index: linux-2.6/drivers/s390/char/sclp_rw.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_rw.c
+++ linux-2.6/drivers/s390/char/sclp_rw.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/sclp_rw.c
- *     driver: reading from and writing to system console on S/390 via SCLP
+ * driver: reading from and writing to system console on S/390 via SCLP
  *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
- *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Copyright IBM Corp. 1999, 2009
+ *
+ * Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *	      Martin Schwidefsky <schwidefsky@de.ibm.com>
  */
 
 #include <linux/kmod.h>
@@ -26,9 +25,16 @@
  */
 #define MAX_SCCB_ROOM (PAGE_SIZE - sizeof(struct sclp_buffer))
 
+static void sclp_rw_pm_event(struct sclp_register *reg,
+			     enum sclp_pm_event sclp_pm_event)
+{
+	sclp_console_pm_event(sclp_pm_event);
+}
+
 /* Event type structure for write message and write priority message */
 static struct sclp_register sclp_rw_event = {
-	.send_mask = EVTYP_MSG_MASK | EVTYP_PMSGCMD_MASK
+	.send_mask = EVTYP_MSG_MASK | EVTYP_PMSGCMD_MASK,
+	.pm_event_fn = sclp_rw_pm_event,
 };
 
 /*
Index: linux-2.6/drivers/s390/char/sclp_vt220.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_vt220.c
+++ linux-2.6/drivers/s390/char/sclp_vt220.c
@@ -1,10 +1,9 @@
 /*
- *  drivers/s390/char/sclp_vt220.c
- *    SCLP VT220 terminal driver.
+ * SCLP VT220 terminal driver.
  *
- *  S390 version
- *    Copyright IBM Corp. 2003,2008
- *    Author(s): Peter Oberparleiter <Peter.Oberparleiter@de.ibm.com>
+ * Copyright IBM Corp. 2003, 2009
+ *
+ * Author(s): Peter Oberparleiter <Peter.Oberparleiter@de.ibm.com>
  */
 
 #include <linux/module.h>
@@ -69,8 +68,11 @@ static struct list_head sclp_vt220_empty
 /* List of pending requests */
 static struct list_head sclp_vt220_outqueue;
 
-/* Number of requests in outqueue */
-static int sclp_vt220_outqueue_count;
+/* Suspend mode flag */
+static int sclp_vt220_suspended;
+
+/* Flag that output queue is currently running */
+static int sclp_vt220_queue_running;
 
 /* Timer used for delaying write requests to merge subsequent messages into
  * a single buffer */
@@ -92,6 +94,8 @@ static int __initdata sclp_vt220_init_co
 static int sclp_vt220_flush_later;
 
 static void sclp_vt220_receiver_fn(struct evbuf_header *evbuf);
+static void sclp_vt220_pm_event_fn(struct sclp_register *reg,
+				   enum sclp_pm_event sclp_pm_event);
 static int __sclp_vt220_emit(struct sclp_vt220_request *request);
 static void sclp_vt220_emit_current(void);
 
@@ -100,7 +104,8 @@ static struct sclp_register sclp_vt220_r
 	.send_mask		= EVTYP_VT220MSG_MASK,
 	.receive_mask		= EVTYP_VT220MSG_MASK,
 	.state_change_fn	= NULL,
-	.receiver_fn		= sclp_vt220_receiver_fn
+	.receiver_fn		= sclp_vt220_receiver_fn,
+	.pm_event_fn		= sclp_vt220_pm_event_fn,
 };
 
 
@@ -120,15 +125,19 @@ sclp_vt220_process_queue(struct sclp_vt2
 		spin_lock_irqsave(&sclp_vt220_lock, flags);
 		/* Move request from outqueue to empty queue */
 		list_del(&request->list);
-		sclp_vt220_outqueue_count--;
 		list_add_tail((struct list_head *) page, &sclp_vt220_empty);
 		/* Check if there is a pending buffer on the out queue. */
 		request = NULL;
 		if (!list_empty(&sclp_vt220_outqueue))
 			request = list_entry(sclp_vt220_outqueue.next,
 					     struct sclp_vt220_request, list);
+		if (!request || sclp_vt220_suspended) {
+			sclp_vt220_queue_running = 0;
+			spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+			break;
+		}
 		spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-	} while (request && __sclp_vt220_emit(request));
+	} while (__sclp_vt220_emit(request));
 	if (request == NULL && sclp_vt220_flush_later)
 		sclp_vt220_emit_current();
 	/* Check if the tty needs a wake up call */
@@ -212,26 +221,7 @@ __sclp_vt220_emit(struct sclp_vt220_requ
 }
 
 /*
- * Queue and emit given request.
- */
-static void
-sclp_vt220_emit(struct sclp_vt220_request *request)
-{
-	unsigned long flags;
-	int count;
-
-	spin_lock_irqsave(&sclp_vt220_lock, flags);
-	list_add_tail(&request->list, &sclp_vt220_outqueue);
-	count = sclp_vt220_outqueue_count++;
-	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-	/* Emit only the first buffer immediately - callback takes care of
-	 * the rest */
-	if (count == 0 && __sclp_vt220_emit(request))
-		sclp_vt220_process_queue(request);
-}
-
-/*
- * Queue and emit current request. Return zero on success, non-zero otherwise.
+ * Queue and emit current request.
  */
 static void
 sclp_vt220_emit_current(void)
@@ -241,22 +231,33 @@ sclp_vt220_emit_current(void)
 	struct sclp_vt220_sccb *sccb;
 
 	spin_lock_irqsave(&sclp_vt220_lock, flags);
-	request = NULL;
-	if (sclp_vt220_current_request != NULL) {
+	if (sclp_vt220_current_request) {
 		sccb = (struct sclp_vt220_sccb *) 
 				sclp_vt220_current_request->sclp_req.sccb;
 		/* Only emit buffers with content */
 		if (sccb->header.length != sizeof(struct sclp_vt220_sccb)) {
-			request = sclp_vt220_current_request;
+			list_add_tail(&sclp_vt220_current_request->list,
+				      &sclp_vt220_outqueue);
 			sclp_vt220_current_request = NULL;
 			if (timer_pending(&sclp_vt220_timer))
 				del_timer(&sclp_vt220_timer);
 		}
 		sclp_vt220_flush_later = 0;
 	}
+	if (sclp_vt220_queue_running || sclp_vt220_suspended)
+		goto out_unlock;
+	if (list_empty(&sclp_vt220_outqueue))
+		goto out_unlock;
+	request = list_first_entry(&sclp_vt220_outqueue,
+				   struct sclp_vt220_request, list);
+	sclp_vt220_queue_running = 1;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+
+	if (__sclp_vt220_emit(request))
+		sclp_vt220_process_queue(request);
+	return;
+out_unlock:
 	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-	if (request != NULL)
-		sclp_vt220_emit(request);
 }
 
 #define SCLP_NORMAL_WRITE	0x00
@@ -396,7 +397,7 @@ __sclp_vt220_write(const unsigned char *
 		if (sclp_vt220_current_request == NULL) {
 			while (list_empty(&sclp_vt220_empty)) {
 				spin_unlock_irqrestore(&sclp_vt220_lock, flags);
-				if (may_fail)
+				if (may_fail || sclp_vt220_suspended)
 					goto out;
 				else
 					sclp_sync_wait();
@@ -531,7 +532,7 @@ sclp_vt220_put_char(struct tty_struct *t
 static void
 sclp_vt220_flush_chars(struct tty_struct *tty)
 {
-	if (sclp_vt220_outqueue_count == 0)
+	if (!sclp_vt220_queue_running)
 		sclp_vt220_emit_current();
 	else
 		sclp_vt220_flush_later = 1;
@@ -635,7 +636,6 @@ static int __init __sclp_vt220_init(int 
 	init_timer(&sclp_vt220_timer);
 	sclp_vt220_current_request = NULL;
 	sclp_vt220_buffered_chars = 0;
-	sclp_vt220_outqueue_count = 0;
 	sclp_vt220_tty = NULL;
 	sclp_vt220_flush_later = 0;
 
@@ -736,7 +736,7 @@ static void __sclp_vt220_flush_buffer(vo
 	spin_lock_irqsave(&sclp_vt220_lock, flags);
 	if (timer_pending(&sclp_vt220_timer))
 		del_timer(&sclp_vt220_timer);
-	while (sclp_vt220_outqueue_count > 0) {
+	while (sclp_vt220_queue_running) {
 		spin_unlock_irqrestore(&sclp_vt220_lock, flags);
 		sclp_sync_wait();
 		spin_lock_irqsave(&sclp_vt220_lock, flags);
@@ -744,6 +744,46 @@ static void __sclp_vt220_flush_buffer(vo
 	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
 }
 
+/*
+ * Resume console: If there are cached messages, emit them.
+ */
+static void sclp_vt220_resume(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	sclp_vt220_suspended = 0;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	sclp_vt220_emit_current();
+}
+
+/*
+ * Suspend console: Set suspend flag and flush console
+ */
+static void sclp_vt220_suspend(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	sclp_vt220_suspended = 1;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	__sclp_vt220_flush_buffer();
+}
+
+static void sclp_vt220_pm_event_fn(struct sclp_register *reg,
+				   enum sclp_pm_event sclp_pm_event)
+{
+	switch (sclp_pm_event) {
+	case SCLP_PM_EVENT_FREEZE:
+		sclp_vt220_suspend();
+		break;
+	case SCLP_PM_EVENT_RESTORE:
+	case SCLP_PM_EVENT_THAW:
+		sclp_vt220_resume();
+		break;
+	}
+}
+
 static int
 sclp_vt220_notify(struct notifier_block *self,
 			  unsigned long event, void *data)

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 27/38] iucv: establish reboot notifier

[Martin Schwidefsky]

[-- Attachment #1: pm_iucv_reboot.patch --]
[-- Type: text/plain, Size: 8522 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

To guarantee a proper cleanup, patch adds a reboot notifier to
the iucv base code, which disables iucv interrupts, shuts down
established iucv pathes, and removes iucv declarations for z/VM.

Checks have to be added to the iucv-API functions, whether 
iucv-buffers removed at reboot time are still declared.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 net/iucv/iucv.c |   89 ++++++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 87 insertions(+), 2 deletions(-)

Index: linux-2.6/net/iucv/iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/iucv.c
+++ linux-2.6/net/iucv/iucv.c
@@ -1,7 +1,8 @@
 /*
  * IUCV base infrastructure.
  *
- * Copyright 2001, 2006 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Copyright IBM Corp. 2001, 2009
+ *
  * Author(s):
  *    Original source:
  *	Alan Altmark (Alan_Altmark@us.ibm.com)	Sept. 2000
@@ -45,6 +46,7 @@
 #include <linux/err.h>
 #include <linux/device.h>
 #include <linux/cpu.h>
+#include <linux/reboot.h>
 #include <net/iucv/iucv.h>
 #include <asm/atomic.h>
 #include <asm/ebcdic.h>
@@ -758,6 +760,28 @@ void iucv_unregister(struct iucv_handler
 }
 EXPORT_SYMBOL(iucv_unregister);
 
+static int iucv_reboot_event(struct notifier_block *this,
+			     unsigned long event, void *ptr)
+{
+	int i, rc;
+
+	get_online_cpus();
+	on_each_cpu(iucv_block_cpu, NULL, 1);
+	preempt_disable();
+	for (i = 0; i < iucv_max_pathid; i++) {
+		if (iucv_path_table[i])
+			rc = iucv_sever_pathid(i, NULL);
+	}
+	preempt_enable();
+	put_online_cpus();
+	iucv_disable();
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block iucv_reboot_notifier = {
+	.notifier_call = iucv_reboot_event,
+};
+
 /**
  * iucv_path_accept
  * @path: address of iucv path structure
@@ -777,6 +801,10 @@ int iucv_path_accept(struct iucv_path *p
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	/* Prepare parameter block. */
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
@@ -792,6 +820,7 @@ int iucv_path_accept(struct iucv_path *p
 		path->msglim = parm->ctrl.ipmsglim;
 		path->flags = parm->ctrl.ipflags1;
 	}
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -821,6 +850,10 @@ int iucv_path_connect(struct iucv_path *
 
 	spin_lock_bh(&iucv_table_lock);
 	iucv_cleanup_queue();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->ctrl.ipmsglim = path->msglim;
@@ -855,6 +888,7 @@ int iucv_path_connect(struct iucv_path *
 			rc = -EIO;
 		}
 	}
+out:
 	spin_unlock_bh(&iucv_table_lock);
 	return rc;
 }
@@ -876,12 +910,17 @@ int iucv_path_quiesce(struct iucv_path *
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (userdata)
 		memcpy(parm->ctrl.ipuser, userdata, sizeof(parm->ctrl.ipuser));
 	parm->ctrl.ippathid = path->pathid;
 	rc = iucv_call_b2f0(IUCV_QUIESCE, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -903,12 +942,17 @@ int iucv_path_resume(struct iucv_path *p
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (userdata)
 		memcpy(parm->ctrl.ipuser, userdata, sizeof(parm->ctrl.ipuser));
 	parm->ctrl.ippathid = path->pathid;
 	rc = iucv_call_b2f0(IUCV_RESUME, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -927,6 +971,10 @@ int iucv_path_sever(struct iucv_path *pa
 	int rc;
 
 	preempt_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	if (iucv_active_cpu != smp_processor_id())
 		spin_lock_bh(&iucv_table_lock);
 	rc = iucv_sever_pathid(path->pathid, userdata);
@@ -934,6 +982,7 @@ int iucv_path_sever(struct iucv_path *pa
 	list_del_init(&path->list);
 	if (iucv_active_cpu != smp_processor_id())
 		spin_unlock_bh(&iucv_table_lock);
+out:
 	preempt_enable();
 	return rc;
 }
@@ -956,6 +1005,10 @@ int iucv_message_purge(struct iucv_path 
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->purge.ippathid = path->pathid;
@@ -967,6 +1020,7 @@ int iucv_message_purge(struct iucv_path 
 		msg->audit = (*(u32 *) &parm->purge.ipaudit) >> 8;
 		msg->tag = parm->purge.ipmsgtag;
 	}
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1043,6 +1097,10 @@ int __iucv_message_receive(struct iucv_p
 	if (msg->flags & IUCV_IPRMDATA)
 		return iucv_message_receive_iprmdata(path, msg, flags,
 						     buffer, size, residual);
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->db.ipbfadr1 = (u32)(addr_t) buffer;
@@ -1058,6 +1116,7 @@ int __iucv_message_receive(struct iucv_p
 		if (residual)
 			*residual = parm->db.ipbfln1f;
 	}
+out:
 	return rc;
 }
 EXPORT_SYMBOL(__iucv_message_receive);
@@ -1111,6 +1170,10 @@ int iucv_message_reject(struct iucv_path
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->db.ippathid = path->pathid;
@@ -1118,6 +1181,7 @@ int iucv_message_reject(struct iucv_path
 	parm->db.iptrgcls = msg->class;
 	parm->db.ipflags1 = (IUCV_IPTRGCLS | IUCV_IPFGMID | IUCV_IPFGPID);
 	rc = iucv_call_b2f0(IUCV_REJECT, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1145,6 +1209,10 @@ int iucv_message_reply(struct iucv_path 
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (flags & IUCV_IPRMDATA) {
@@ -1162,6 +1230,7 @@ int iucv_message_reply(struct iucv_path 
 		parm->db.iptrgcls = msg->class;
 	}
 	rc = iucv_call_b2f0(IUCV_REPLY, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1190,6 +1259,10 @@ int __iucv_message_send(struct iucv_path
 	union iucv_param *parm;
 	int rc;
 
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (flags & IUCV_IPRMDATA) {
@@ -1212,6 +1285,7 @@ int __iucv_message_send(struct iucv_path
 	rc = iucv_call_b2f0(IUCV_SEND, parm);
 	if (!rc)
 		msg->id = parm->db.ipmsgid;
+out:
 	return rc;
 }
 EXPORT_SYMBOL(__iucv_message_send);
@@ -1272,6 +1346,10 @@ int iucv_message_send2way(struct iucv_pa
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (flags & IUCV_IPRMDATA) {
@@ -1297,6 +1375,7 @@ int iucv_message_send2way(struct iucv_pa
 	rc = iucv_call_b2f0(IUCV_SEND, parm);
 	if (!rc)
 		msg->id = parm->db.ipmsgid;
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1738,15 +1817,20 @@ static int __init iucv_init(void)
 	rc = register_hotcpu_notifier(&iucv_cpu_notifier);
 	if (rc)
 		goto out_free;
+	rc = register_reboot_notifier(&iucv_reboot_notifier);
+	if (rc)
+		goto out_cpu;
 	ASCEBC(iucv_error_no_listener, 16);
 	ASCEBC(iucv_error_no_memory, 16);
 	ASCEBC(iucv_error_pathid, 16);
 	iucv_available = 1;
 	rc = bus_register(&iucv_bus);
 	if (rc)
-		goto out_cpu;
+		goto out_reboot;
 	return 0;
 
+out_reboot:
+	unregister_reboot_notifier(&iucv_reboot_notifier);
 out_cpu:
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 out_free:
@@ -1781,6 +1865,7 @@ static void __exit iucv_exit(void)
 	list_for_each_entry_safe(p, n, &iucv_work_queue, list)
 		kfree(p);
 	spin_unlock_irq(&iucv_queue_lock);
+	unregister_reboot_notifier(&iucv_reboot_notifier);
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 	for_each_possible_cpu(cpu) {
 		kfree(iucv_param_irq[cpu]);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 27/38] iucv: establish reboot notifier

[Martin Schwidefsky]

[-- Attachment #1: pm_iucv_reboot.patch --]
[-- Type: text/plain, Size: 8521 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

To guarantee a proper cleanup, patch adds a reboot notifier to
the iucv base code, which disables iucv interrupts, shuts down
established iucv pathes, and removes iucv declarations for z/VM.

Checks have to be added to the iucv-API functions, whether 
iucv-buffers removed at reboot time are still declared.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 net/iucv/iucv.c |   89 ++++++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 87 insertions(+), 2 deletions(-)

Index: linux-2.6/net/iucv/iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/iucv.c
+++ linux-2.6/net/iucv/iucv.c
@@ -1,7 +1,8 @@
 /*
  * IUCV base infrastructure.
  *
- * Copyright 2001, 2006 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Copyright IBM Corp. 2001, 2009
+ *
  * Author(s):
  *    Original source:
  *	Alan Altmark (Alan_Altmark@us.ibm.com)	Sept. 2000
@@ -45,6 +46,7 @@
 #include <linux/err.h>
 #include <linux/device.h>
 #include <linux/cpu.h>
+#include <linux/reboot.h>
 #include <net/iucv/iucv.h>
 #include <asm/atomic.h>
 #include <asm/ebcdic.h>
@@ -758,6 +760,28 @@ void iucv_unregister(struct iucv_handler
 }
 EXPORT_SYMBOL(iucv_unregister);
 
+static int iucv_reboot_event(struct notifier_block *this,
+			     unsigned long event, void *ptr)
+{
+	int i, rc;
+
+	get_online_cpus();
+	on_each_cpu(iucv_block_cpu, NULL, 1);
+	preempt_disable();
+	for (i = 0; i < iucv_max_pathid; i++) {
+		if (iucv_path_table[i])
+			rc = iucv_sever_pathid(i, NULL);
+	}
+	preempt_enable();
+	put_online_cpus();
+	iucv_disable();
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block iucv_reboot_notifier = {
+	.notifier_call = iucv_reboot_event,
+};
+
 /**
  * iucv_path_accept
  * @path: address of iucv path structure
@@ -777,6 +801,10 @@ int iucv_path_accept(struct iucv_path *p
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	/* Prepare parameter block. */
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
@@ -792,6 +820,7 @@ int iucv_path_accept(struct iucv_path *p
 		path->msglim = parm->ctrl.ipmsglim;
 		path->flags = parm->ctrl.ipflags1;
 	}
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -821,6 +850,10 @@ int iucv_path_connect(struct iucv_path *
 
 	spin_lock_bh(&iucv_table_lock);
 	iucv_cleanup_queue();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->ctrl.ipmsglim = path->msglim;
@@ -855,6 +888,7 @@ int iucv_path_connect(struct iucv_path *
 			rc = -EIO;
 		}
 	}
+out:
 	spin_unlock_bh(&iucv_table_lock);
 	return rc;
 }
@@ -876,12 +910,17 @@ int iucv_path_quiesce(struct iucv_path *
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (userdata)
 		memcpy(parm->ctrl.ipuser, userdata, sizeof(parm->ctrl.ipuser));
 	parm->ctrl.ippathid = path->pathid;
 	rc = iucv_call_b2f0(IUCV_QUIESCE, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -903,12 +942,17 @@ int iucv_path_resume(struct iucv_path *p
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (userdata)
 		memcpy(parm->ctrl.ipuser, userdata, sizeof(parm->ctrl.ipuser));
 	parm->ctrl.ippathid = path->pathid;
 	rc = iucv_call_b2f0(IUCV_RESUME, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -927,6 +971,10 @@ int iucv_path_sever(struct iucv_path *pa
 	int rc;
 
 	preempt_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	if (iucv_active_cpu != smp_processor_id())
 		spin_lock_bh(&iucv_table_lock);
 	rc = iucv_sever_pathid(path->pathid, userdata);
@@ -934,6 +982,7 @@ int iucv_path_sever(struct iucv_path *pa
 	list_del_init(&path->list);
 	if (iucv_active_cpu != smp_processor_id())
 		spin_unlock_bh(&iucv_table_lock);
+out:
 	preempt_enable();
 	return rc;
 }
@@ -956,6 +1005,10 @@ int iucv_message_purge(struct iucv_path 
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->purge.ippathid = path->pathid;
@@ -967,6 +1020,7 @@ int iucv_message_purge(struct iucv_path 
 		msg->audit = (*(u32 *) &parm->purge.ipaudit) >> 8;
 		msg->tag = parm->purge.ipmsgtag;
 	}
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1043,6 +1097,10 @@ int __iucv_message_receive(struct iucv_p
 	if (msg->flags & IUCV_IPRMDATA)
 		return iucv_message_receive_iprmdata(path, msg, flags,
 						     buffer, size, residual);
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->db.ipbfadr1 = (u32)(addr_t) buffer;
@@ -1058,6 +1116,7 @@ int __iucv_message_receive(struct iucv_p
 		if (residual)
 			*residual = parm->db.ipbfln1f;
 	}
+out:
 	return rc;
 }
 EXPORT_SYMBOL(__iucv_message_receive);
@@ -1111,6 +1170,10 @@ int iucv_message_reject(struct iucv_path
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	parm->db.ippathid = path->pathid;
@@ -1118,6 +1181,7 @@ int iucv_message_reject(struct iucv_path
 	parm->db.iptrgcls = msg->class;
 	parm->db.ipflags1 = (IUCV_IPTRGCLS | IUCV_IPFGMID | IUCV_IPFGPID);
 	rc = iucv_call_b2f0(IUCV_REJECT, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1145,6 +1209,10 @@ int iucv_message_reply(struct iucv_path 
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (flags & IUCV_IPRMDATA) {
@@ -1162,6 +1230,7 @@ int iucv_message_reply(struct iucv_path 
 		parm->db.iptrgcls = msg->class;
 	}
 	rc = iucv_call_b2f0(IUCV_REPLY, parm);
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1190,6 +1259,10 @@ int __iucv_message_send(struct iucv_path
 	union iucv_param *parm;
 	int rc;
 
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (flags & IUCV_IPRMDATA) {
@@ -1212,6 +1285,7 @@ int __iucv_message_send(struct iucv_path
 	rc = iucv_call_b2f0(IUCV_SEND, parm);
 	if (!rc)
 		msg->id = parm->db.ipmsgid;
+out:
 	return rc;
 }
 EXPORT_SYMBOL(__iucv_message_send);
@@ -1272,6 +1346,10 @@ int iucv_message_send2way(struct iucv_pa
 	int rc;
 
 	local_bh_disable();
+	if (!cpu_isset(smp_processor_id(), iucv_buffer_cpumask)) {
+		rc = -EIO;
+		goto out;
+	}
 	parm = iucv_param[smp_processor_id()];
 	memset(parm, 0, sizeof(union iucv_param));
 	if (flags & IUCV_IPRMDATA) {
@@ -1297,6 +1375,7 @@ int iucv_message_send2way(struct iucv_pa
 	rc = iucv_call_b2f0(IUCV_SEND, parm);
 	if (!rc)
 		msg->id = parm->db.ipmsgid;
+out:
 	local_bh_enable();
 	return rc;
 }
@@ -1738,15 +1817,20 @@ static int __init iucv_init(void)
 	rc = register_hotcpu_notifier(&iucv_cpu_notifier);
 	if (rc)
 		goto out_free;
+	rc = register_reboot_notifier(&iucv_reboot_notifier);
+	if (rc)
+		goto out_cpu;
 	ASCEBC(iucv_error_no_listener, 16);
 	ASCEBC(iucv_error_no_memory, 16);
 	ASCEBC(iucv_error_pathid, 16);
 	iucv_available = 1;
 	rc = bus_register(&iucv_bus);
 	if (rc)
-		goto out_cpu;
+		goto out_reboot;
 	return 0;
 
+out_reboot:
+	unregister_reboot_notifier(&iucv_reboot_notifier);
 out_cpu:
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 out_free:
@@ -1781,6 +1865,7 @@ static void __exit iucv_exit(void)
 	list_for_each_entry_safe(p, n, &iucv_work_queue, list)
 		kfree(p);
 	spin_unlock_irq(&iucv_queue_lock);
+	unregister_reboot_notifier(&iucv_reboot_notifier);
 	unregister_hotcpu_notifier(&iucv_cpu_notifier);
 	for_each_possible_cpu(cpu) {
 		kfree(iucv_param_irq[cpu]);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 28/38] pm: iucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_iucv.patch --]
[-- Type: text/plain, Size: 6823 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Patch calls the PM callback functions of iucv-bus devices, which are
responsible for removal of their established iucv pathes.

The PM freeze callback for the first iucv-bus device disables all iucv 
interrupts except the connection severed interrupt.
The PM freeze callback for the last iucv-bus device shuts down iucv.

The PM thaw callback for the first iucv-bus device re-enables iucv
if it has been shut down during freeze. If freezing has been interrupted,
it re-enables iucv interrupts according to the needs of iucv-exploiters.

The PM restore callback for the first iucv-bus device re-enables iucv.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 net/iucv/iucv.c |  179 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 179 insertions(+)

Index: linux-2.6/net/iucv/iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/iucv.c
+++ linux-2.6/net/iucv/iucv.c
@@ -11,6 +11,8 @@
  *	Fritz Elfert (elfert@de.ibm.com, felfert@millenux.com)
  *    Rewritten for af_iucv:
  *	Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *    PM functions:
+ *	Ursula Braun (ursula.braun@de.ibm.com)
  *
  * Documentation used:
  *    The original source
@@ -77,9 +79,24 @@ static int iucv_bus_match(struct device 
 	return 0;
 }
 
+static int iucv_pm_prepare(struct device *);
+static void iucv_pm_complete(struct device *);
+static int iucv_pm_freeze(struct device *);
+static int iucv_pm_thaw(struct device *);
+static int iucv_pm_restore(struct device *);
+
+static struct dev_pm_ops iucv_pm_ops = {
+	.prepare = iucv_pm_prepare,
+	.complete = iucv_pm_complete,
+	.freeze = iucv_pm_freeze,
+	.thaw = iucv_pm_thaw,
+	.restore = iucv_pm_restore,
+};
+
 struct bus_type iucv_bus = {
 	.name = "iucv",
 	.match = iucv_bus_match,
+	.pm = &iucv_pm_ops,
 };
 EXPORT_SYMBOL(iucv_bus);
 
@@ -149,6 +166,7 @@ enum iucv_command_codes {
 	IUCV_RESUME = 14,
 	IUCV_SEVER = 15,
 	IUCV_SETMASK = 16,
+	IUCV_SETCONTROLMASK = 17,
 };
 
 /*
@@ -366,6 +384,18 @@ static void iucv_allow_cpu(void *data)
 	parm->set_mask.ipmask = 0xf8;
 	iucv_call_b2f0(IUCV_SETMASK, parm);
 
+	/*
+	 * Enable all iucv control interrupts.
+	 * ipmask contains bits for the different interrupts
+	 *	0x80 - Flag to allow pending connections interrupts
+	 *	0x40 - Flag to allow connection complete interrupts
+	 *	0x20 - Flag to allow connection severed interrupts
+	 *	0x10 - Flag to allow connection quiesced interrupts
+	 *	0x08 - Flag to allow connection resumed interrupts
+	 */
+	memset(parm, 0, sizeof(union iucv_param));
+	parm->set_mask.ipmask = 0xf8;
+	iucv_call_b2f0(IUCV_SETCONTROLMASK, parm);
 	/* Set indication that iucv interrupts are allowed for this cpu. */
 	cpu_set(cpu, iucv_irq_cpumask);
 }
@@ -391,6 +421,31 @@ static void iucv_block_cpu(void *data)
 }
 
 /**
+ * iucv_block_cpu_almost
+ * @data: unused
+ *
+ * Allow connection-severed interrupts only on this cpu.
+ */
+static void iucv_block_cpu_almost(void *data)
+{
+	int cpu = smp_processor_id();
+	union iucv_param *parm;
+
+	/* Allow iucv control interrupts only */
+	parm = iucv_param_irq[cpu];
+	memset(parm, 0, sizeof(union iucv_param));
+	parm->set_mask.ipmask = 0x08;
+	iucv_call_b2f0(IUCV_SETMASK, parm);
+	/* Allow iucv-severed interrupt only */
+	memset(parm, 0, sizeof(union iucv_param));
+	parm->set_mask.ipmask = 0x20;
+	iucv_call_b2f0(IUCV_SETCONTROLMASK, parm);
+
+	/* Clear indication that iucv interrupts are allowed for this cpu. */
+	cpu_clear(cpu, iucv_irq_cpumask);
+}
+
+/**
  * iucv_declare_cpu
  * @data: unused
  *
@@ -1764,6 +1819,130 @@ static void iucv_external_interrupt(u16 
 	spin_unlock(&iucv_queue_lock);
 }
 
+static int iucv_pm_prepare(struct device *dev)
+{
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_INFO "iucv_pm_prepare\n");
+#endif
+	if (dev->driver && dev->driver->pm && dev->driver->pm->prepare)
+		rc = dev->driver->pm->prepare(dev);
+	return rc;
+}
+
+static void iucv_pm_complete(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_INFO "iucv_pm_complete\n");
+#endif
+	if (dev->driver && dev->driver->pm && dev->driver->pm->complete)
+		dev->driver->pm->complete(dev);
+}
+
+/**
+ * iucv_path_table_empty() - determine if iucv path table is empty
+ *
+ * Returns 0 if there are still iucv pathes defined
+ *         1 if there are no iucv pathes defined
+ */
+int iucv_path_table_empty(void)
+{
+	int i;
+
+	for (i = 0; i < iucv_max_pathid; i++) {
+		if (iucv_path_table[i])
+			return 0;
+	}
+	return 1;
+}
+
+/**
+ * iucv_pm_freeze() - Freeze PM callback
+ * @dev:	iucv-based device
+ *
+ * disable iucv interrupts
+ * invoke callback function of the iucv-based driver
+ * shut down iucv, if no iucv-pathes are established anymore
+ */
+static int iucv_pm_freeze(struct device *dev)
+{
+	int cpu;
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "iucv_pm_freeze\n");
+#endif
+	for_each_cpu_mask_nr(cpu, iucv_irq_cpumask)
+		smp_call_function_single(cpu, iucv_block_cpu_almost, NULL, 1);
+	if (dev->driver && dev->driver->pm && dev->driver->pm->freeze)
+		rc = dev->driver->pm->freeze(dev);
+	if (iucv_path_table_empty())
+		iucv_disable();
+	return rc;
+}
+
+/**
+ * iucv_pm_thaw() - Thaw PM callback
+ * @dev:	iucv-based device
+ *
+ * make iucv ready for use again: allocate path table, declare interrupt buffers
+ *				  and enable iucv interrupts
+ * invoke callback function of the iucv-based driver
+ */
+static int iucv_pm_thaw(struct device *dev)
+{
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "iucv_pm_thaw\n");
+#endif
+	if (!iucv_path_table) {
+		rc = iucv_enable();
+		if (rc)
+			goto out;
+	}
+	if (cpus_empty(iucv_irq_cpumask)) {
+		if (iucv_nonsmp_handler)
+			/* enable interrupts on one cpu */
+			iucv_allow_cpu(NULL);
+		else
+			/* enable interrupts on all cpus */
+			iucv_setmask_mp();
+	}
+	if (dev->driver && dev->driver->pm && dev->driver->pm->thaw)
+		rc = dev->driver->pm->thaw(dev);
+out:
+	return rc;
+}
+
+/**
+ * iucv_pm_restore() - Restore PM callback
+ * @dev:	iucv-based device
+ *
+ * make iucv ready for use again: allocate path table, declare interrupt buffers
+ *				  and enable iucv interrupts
+ * invoke callback function of the iucv-based driver
+ */
+static int iucv_pm_restore(struct device *dev)
+{
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "iucv_pm_restore %p\n", iucv_path_table);
+#endif
+	if (cpus_empty(iucv_irq_cpumask)) {
+		rc = iucv_query_maxconn();
+		rc = iucv_enable();
+		if (rc)
+			goto out;
+	}
+	if (dev->driver && dev->driver->pm && dev->driver->pm->restore)
+		rc = dev->driver->pm->restore(dev);
+out:
+	return rc;
+}
+
 /**
  * iucv_init
  *

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 28/38] pm: iucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_iucv.patch --]
[-- Type: text/plain, Size: 6822 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Patch calls the PM callback functions of iucv-bus devices, which are
responsible for removal of their established iucv pathes.

The PM freeze callback for the first iucv-bus device disables all iucv 
interrupts except the connection severed interrupt.
The PM freeze callback for the last iucv-bus device shuts down iucv.

The PM thaw callback for the first iucv-bus device re-enables iucv
if it has been shut down during freeze. If freezing has been interrupted,
it re-enables iucv interrupts according to the needs of iucv-exploiters.

The PM restore callback for the first iucv-bus device re-enables iucv.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 net/iucv/iucv.c |  179 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 179 insertions(+)

Index: linux-2.6/net/iucv/iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/iucv.c
+++ linux-2.6/net/iucv/iucv.c
@@ -11,6 +11,8 @@
  *	Fritz Elfert (elfert@de.ibm.com, felfert@millenux.com)
  *    Rewritten for af_iucv:
  *	Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *    PM functions:
+ *	Ursula Braun (ursula.braun@de.ibm.com)
  *
  * Documentation used:
  *    The original source
@@ -77,9 +79,24 @@ static int iucv_bus_match(struct device 
 	return 0;
 }
 
+static int iucv_pm_prepare(struct device *);
+static void iucv_pm_complete(struct device *);
+static int iucv_pm_freeze(struct device *);
+static int iucv_pm_thaw(struct device *);
+static int iucv_pm_restore(struct device *);
+
+static struct dev_pm_ops iucv_pm_ops = {
+	.prepare = iucv_pm_prepare,
+	.complete = iucv_pm_complete,
+	.freeze = iucv_pm_freeze,
+	.thaw = iucv_pm_thaw,
+	.restore = iucv_pm_restore,
+};
+
 struct bus_type iucv_bus = {
 	.name = "iucv",
 	.match = iucv_bus_match,
+	.pm = &iucv_pm_ops,
 };
 EXPORT_SYMBOL(iucv_bus);
 
@@ -149,6 +166,7 @@ enum iucv_command_codes {
 	IUCV_RESUME = 14,
 	IUCV_SEVER = 15,
 	IUCV_SETMASK = 16,
+	IUCV_SETCONTROLMASK = 17,
 };
 
 /*
@@ -366,6 +384,18 @@ static void iucv_allow_cpu(void *data)
 	parm->set_mask.ipmask = 0xf8;
 	iucv_call_b2f0(IUCV_SETMASK, parm);
 
+	/*
+	 * Enable all iucv control interrupts.
+	 * ipmask contains bits for the different interrupts
+	 *	0x80 - Flag to allow pending connections interrupts
+	 *	0x40 - Flag to allow connection complete interrupts
+	 *	0x20 - Flag to allow connection severed interrupts
+	 *	0x10 - Flag to allow connection quiesced interrupts
+	 *	0x08 - Flag to allow connection resumed interrupts
+	 */
+	memset(parm, 0, sizeof(union iucv_param));
+	parm->set_mask.ipmask = 0xf8;
+	iucv_call_b2f0(IUCV_SETCONTROLMASK, parm);
 	/* Set indication that iucv interrupts are allowed for this cpu. */
 	cpu_set(cpu, iucv_irq_cpumask);
 }
@@ -391,6 +421,31 @@ static void iucv_block_cpu(void *data)
 }
 
 /**
+ * iucv_block_cpu_almost
+ * @data: unused
+ *
+ * Allow connection-severed interrupts only on this cpu.
+ */
+static void iucv_block_cpu_almost(void *data)
+{
+	int cpu = smp_processor_id();
+	union iucv_param *parm;
+
+	/* Allow iucv control interrupts only */
+	parm = iucv_param_irq[cpu];
+	memset(parm, 0, sizeof(union iucv_param));
+	parm->set_mask.ipmask = 0x08;
+	iucv_call_b2f0(IUCV_SETMASK, parm);
+	/* Allow iucv-severed interrupt only */
+	memset(parm, 0, sizeof(union iucv_param));
+	parm->set_mask.ipmask = 0x20;
+	iucv_call_b2f0(IUCV_SETCONTROLMASK, parm);
+
+	/* Clear indication that iucv interrupts are allowed for this cpu. */
+	cpu_clear(cpu, iucv_irq_cpumask);
+}
+
+/**
  * iucv_declare_cpu
  * @data: unused
  *
@@ -1764,6 +1819,130 @@ static void iucv_external_interrupt(u16 
 	spin_unlock(&iucv_queue_lock);
 }
 
+static int iucv_pm_prepare(struct device *dev)
+{
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_INFO "iucv_pm_prepare\n");
+#endif
+	if (dev->driver && dev->driver->pm && dev->driver->pm->prepare)
+		rc = dev->driver->pm->prepare(dev);
+	return rc;
+}
+
+static void iucv_pm_complete(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_INFO "iucv_pm_complete\n");
+#endif
+	if (dev->driver && dev->driver->pm && dev->driver->pm->complete)
+		dev->driver->pm->complete(dev);
+}
+
+/**
+ * iucv_path_table_empty() - determine if iucv path table is empty
+ *
+ * Returns 0 if there are still iucv pathes defined
+ *         1 if there are no iucv pathes defined
+ */
+int iucv_path_table_empty(void)
+{
+	int i;
+
+	for (i = 0; i < iucv_max_pathid; i++) {
+		if (iucv_path_table[i])
+			return 0;
+	}
+	return 1;
+}
+
+/**
+ * iucv_pm_freeze() - Freeze PM callback
+ * @dev:	iucv-based device
+ *
+ * disable iucv interrupts
+ * invoke callback function of the iucv-based driver
+ * shut down iucv, if no iucv-pathes are established anymore
+ */
+static int iucv_pm_freeze(struct device *dev)
+{
+	int cpu;
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "iucv_pm_freeze\n");
+#endif
+	for_each_cpu_mask_nr(cpu, iucv_irq_cpumask)
+		smp_call_function_single(cpu, iucv_block_cpu_almost, NULL, 1);
+	if (dev->driver && dev->driver->pm && dev->driver->pm->freeze)
+		rc = dev->driver->pm->freeze(dev);
+	if (iucv_path_table_empty())
+		iucv_disable();
+	return rc;
+}
+
+/**
+ * iucv_pm_thaw() - Thaw PM callback
+ * @dev:	iucv-based device
+ *
+ * make iucv ready for use again: allocate path table, declare interrupt buffers
+ *				  and enable iucv interrupts
+ * invoke callback function of the iucv-based driver
+ */
+static int iucv_pm_thaw(struct device *dev)
+{
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "iucv_pm_thaw\n");
+#endif
+	if (!iucv_path_table) {
+		rc = iucv_enable();
+		if (rc)
+			goto out;
+	}
+	if (cpus_empty(iucv_irq_cpumask)) {
+		if (iucv_nonsmp_handler)
+			/* enable interrupts on one cpu */
+			iucv_allow_cpu(NULL);
+		else
+			/* enable interrupts on all cpus */
+			iucv_setmask_mp();
+	}
+	if (dev->driver && dev->driver->pm && dev->driver->pm->thaw)
+		rc = dev->driver->pm->thaw(dev);
+out:
+	return rc;
+}
+
+/**
+ * iucv_pm_restore() - Restore PM callback
+ * @dev:	iucv-based device
+ *
+ * make iucv ready for use again: allocate path table, declare interrupt buffers
+ *				  and enable iucv interrupts
+ * invoke callback function of the iucv-based driver
+ */
+static int iucv_pm_restore(struct device *dev)
+{
+	int rc = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "iucv_pm_restore %p\n", iucv_path_table);
+#endif
+	if (cpus_empty(iucv_irq_cpumask)) {
+		rc = iucv_query_maxconn();
+		rc = iucv_enable();
+		if (rc)
+			goto out;
+	}
+	if (dev->driver && dev->driver->pm && dev->driver->pm->restore)
+		rc = dev->driver->pm->restore(dev);
+out:
+	return rc;
+}
+
 /**
  * iucv_init
  *

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 29/38] pm: netiucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_netiucv.patch --]
[-- Type: text/plain, Size: 5317 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Patch establishes a dummy netiucv device to make sure iucv is notified
about suspend/resume even if netiucv is the only loaded iucv-exploting
module without any real net_device defined.

The PM freeze callback closes all open netiucv connections. Thus the
corresponding iucv path is removed.
The PM thaw/restore callback re-opens previously closed netiucv
connections.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/netiucv.c |  115 ++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 109 insertions(+), 6 deletions(-)

Index: linux-2.6/drivers/s390/net/netiucv.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/netiucv.c
+++ linux-2.6/drivers/s390/net/netiucv.c
@@ -1,11 +1,15 @@
 /*
  * IUCV network driver
  *
- * Copyright 2001 IBM Deutschland Entwicklung GmbH, IBM Corporation
- * Author(s): Fritz Elfert (elfert@de.ibm.com, felfert@millenux.com)
+ * Copyright IBM Corp. 2001, 2009
  *
- * Sysfs integration and all bugs therein by Cornelia Huck
- * (cornelia.huck@de.ibm.com)
+ * Author(s):
+ *	Original netiucv driver:
+ *		Fritz Elfert (elfert@de.ibm.com, felfert@millenux.com)
+ *	Sysfs integration and all bugs therein:
+ *		Cornelia Huck (cornelia.huck@de.ibm.com)
+ *	PM functions:
+ *		Ursula Braun (ursula.braun@de.ibm.com)
  *
  * Documentation used:
  *  the source of the original IUCV driver by:
@@ -149,10 +153,27 @@ PRINT_##importance(header "%02x %02x %02
 
 #define PRINTK_HEADER " iucv: "       /* for debugging */
 
+/* dummy device to make sure netiucv_pm functions are called */
+static struct device *netiucv_dev;
+
+static int netiucv_pm_prepare(struct device *);
+static void netiucv_pm_complete(struct device *);
+static int netiucv_pm_freeze(struct device *);
+static int netiucv_pm_restore_thaw(struct device *);
+
+static struct dev_pm_ops netiucv_pm_ops = {
+	.prepare = netiucv_pm_prepare,
+	.complete = netiucv_pm_complete,
+	.freeze = netiucv_pm_freeze,
+	.thaw = netiucv_pm_restore_thaw,
+	.restore = netiucv_pm_restore_thaw,
+};
+
 static struct device_driver netiucv_driver = {
 	.owner = THIS_MODULE,
 	.name = "netiucv",
 	.bus  = &iucv_bus,
+	.pm = &netiucv_pm_ops,
 };
 
 static int netiucv_callback_connreq(struct iucv_path *,
@@ -233,6 +254,7 @@ struct netiucv_priv {
 	fsm_instance            *fsm;
         struct iucv_connection  *conn;
 	struct device           *dev;
+	int			 pm_state;
 };
 
 /**
@@ -1265,6 +1287,72 @@ static int netiucv_close(struct net_devi
 	return 0;
 }
 
+static int netiucv_pm_prepare(struct device *dev)
+{
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	return 0;
+}
+
+static void netiucv_pm_complete(struct device *dev)
+{
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	return;
+}
+
+/**
+ * netiucv_pm_freeze() - Freeze PM callback
+ * @dev:	netiucv device
+ *
+ * close open netiucv interfaces
+ */
+static int netiucv_pm_freeze(struct device *dev)
+{
+	struct netiucv_priv *priv = dev->driver_data;
+	struct net_device *ndev = NULL;
+	int rc = 0;
+
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	if (priv && priv->conn)
+		ndev = priv->conn->netdev;
+	if (!ndev)
+		goto out;
+	netif_device_detach(ndev);
+	priv->pm_state = fsm_getstate(priv->fsm);
+	rc = netiucv_close(ndev);
+out:
+	return rc;
+}
+
+/**
+ * netiucv_pm_restore_thaw() - Thaw and restore PM callback
+ * @dev:	netiucv device
+ *
+ * re-open netiucv interfaces closed during freeze
+ */
+static int netiucv_pm_restore_thaw(struct device *dev)
+{
+	struct netiucv_priv *priv = dev->driver_data;
+	struct net_device *ndev = NULL;
+	int rc = 0;
+
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	if (priv && priv->conn)
+		ndev = priv->conn->netdev;
+	if (!ndev)
+		goto out;
+	switch (priv->pm_state) {
+	case DEV_STATE_RUNNING:
+	case DEV_STATE_STARTWAIT:
+		rc = netiucv_open(ndev);
+		break;
+	default:
+		break;
+	}
+	netif_device_attach(ndev);
+out:
+	return rc;
+}
+
 /**
  * Start transmission of a packet.
  * Called from generic network device layer.
@@ -1731,7 +1819,6 @@ static int netiucv_register_device(struc
 	struct device *dev = kzalloc(sizeof(struct device), GFP_KERNEL);
 	int ret;
 
-
 	IUCV_DBF_TEXT(trace, 3, __func__);
 
 	if (dev) {
@@ -2100,6 +2187,7 @@ static void __exit netiucv_exit(void)
 		netiucv_unregister_device(dev);
 	}
 
+	device_unregister(netiucv_dev);
 	driver_unregister(&netiucv_driver);
 	iucv_unregister(&netiucv_handler, 1);
 	iucv_unregister_dbf_views();
@@ -2125,10 +2213,25 @@ static int __init netiucv_init(void)
 		IUCV_DBF_TEXT_(setup, 2, "ret %d from driver_register\n", rc);
 		goto out_iucv;
 	}
-
+	/* establish dummy device */
+	netiucv_dev = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!netiucv_dev) {
+		rc = -ENOMEM;
+		goto out_driver;
+	}
+	dev_set_name(netiucv_dev, "netiucv");
+	netiucv_dev->bus = &iucv_bus;
+	netiucv_dev->parent = iucv_root;
+	netiucv_dev->release = (void (*)(struct device *))kfree;
+	netiucv_dev->driver = &netiucv_driver;
+	rc = device_register(netiucv_dev);
+	if (rc)
+		goto out_driver;
 	netiucv_banner();
 	return rc;
 
+out_driver:
+	driver_unregister(&netiucv_driver);
 out_iucv:
 	iucv_unregister(&netiucv_handler, 1);
 out_dbf:

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 29/38] pm: netiucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_netiucv.patch --]
[-- Type: text/plain, Size: 5316 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Patch establishes a dummy netiucv device to make sure iucv is notified
about suspend/resume even if netiucv is the only loaded iucv-exploting
module without any real net_device defined.

The PM freeze callback closes all open netiucv connections. Thus the
corresponding iucv path is removed.
The PM thaw/restore callback re-opens previously closed netiucv
connections.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/netiucv.c |  115 ++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 109 insertions(+), 6 deletions(-)

Index: linux-2.6/drivers/s390/net/netiucv.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/netiucv.c
+++ linux-2.6/drivers/s390/net/netiucv.c
@@ -1,11 +1,15 @@
 /*
  * IUCV network driver
  *
- * Copyright 2001 IBM Deutschland Entwicklung GmbH, IBM Corporation
- * Author(s): Fritz Elfert (elfert@de.ibm.com, felfert@millenux.com)
+ * Copyright IBM Corp. 2001, 2009
  *
- * Sysfs integration and all bugs therein by Cornelia Huck
- * (cornelia.huck@de.ibm.com)
+ * Author(s):
+ *	Original netiucv driver:
+ *		Fritz Elfert (elfert@de.ibm.com, felfert@millenux.com)
+ *	Sysfs integration and all bugs therein:
+ *		Cornelia Huck (cornelia.huck@de.ibm.com)
+ *	PM functions:
+ *		Ursula Braun (ursula.braun@de.ibm.com)
  *
  * Documentation used:
  *  the source of the original IUCV driver by:
@@ -149,10 +153,27 @@ PRINT_##importance(header "%02x %02x %02
 
 #define PRINTK_HEADER " iucv: "       /* for debugging */
 
+/* dummy device to make sure netiucv_pm functions are called */
+static struct device *netiucv_dev;
+
+static int netiucv_pm_prepare(struct device *);
+static void netiucv_pm_complete(struct device *);
+static int netiucv_pm_freeze(struct device *);
+static int netiucv_pm_restore_thaw(struct device *);
+
+static struct dev_pm_ops netiucv_pm_ops = {
+	.prepare = netiucv_pm_prepare,
+	.complete = netiucv_pm_complete,
+	.freeze = netiucv_pm_freeze,
+	.thaw = netiucv_pm_restore_thaw,
+	.restore = netiucv_pm_restore_thaw,
+};
+
 static struct device_driver netiucv_driver = {
 	.owner = THIS_MODULE,
 	.name = "netiucv",
 	.bus  = &iucv_bus,
+	.pm = &netiucv_pm_ops,
 };
 
 static int netiucv_callback_connreq(struct iucv_path *,
@@ -233,6 +254,7 @@ struct netiucv_priv {
 	fsm_instance            *fsm;
         struct iucv_connection  *conn;
 	struct device           *dev;
+	int			 pm_state;
 };
 
 /**
@@ -1265,6 +1287,72 @@ static int netiucv_close(struct net_devi
 	return 0;
 }
 
+static int netiucv_pm_prepare(struct device *dev)
+{
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	return 0;
+}
+
+static void netiucv_pm_complete(struct device *dev)
+{
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	return;
+}
+
+/**
+ * netiucv_pm_freeze() - Freeze PM callback
+ * @dev:	netiucv device
+ *
+ * close open netiucv interfaces
+ */
+static int netiucv_pm_freeze(struct device *dev)
+{
+	struct netiucv_priv *priv = dev->driver_data;
+	struct net_device *ndev = NULL;
+	int rc = 0;
+
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	if (priv && priv->conn)
+		ndev = priv->conn->netdev;
+	if (!ndev)
+		goto out;
+	netif_device_detach(ndev);
+	priv->pm_state = fsm_getstate(priv->fsm);
+	rc = netiucv_close(ndev);
+out:
+	return rc;
+}
+
+/**
+ * netiucv_pm_restore_thaw() - Thaw and restore PM callback
+ * @dev:	netiucv device
+ *
+ * re-open netiucv interfaces closed during freeze
+ */
+static int netiucv_pm_restore_thaw(struct device *dev)
+{
+	struct netiucv_priv *priv = dev->driver_data;
+	struct net_device *ndev = NULL;
+	int rc = 0;
+
+	IUCV_DBF_TEXT(trace, 3, __func__);
+	if (priv && priv->conn)
+		ndev = priv->conn->netdev;
+	if (!ndev)
+		goto out;
+	switch (priv->pm_state) {
+	case DEV_STATE_RUNNING:
+	case DEV_STATE_STARTWAIT:
+		rc = netiucv_open(ndev);
+		break;
+	default:
+		break;
+	}
+	netif_device_attach(ndev);
+out:
+	return rc;
+}
+
 /**
  * Start transmission of a packet.
  * Called from generic network device layer.
@@ -1731,7 +1819,6 @@ static int netiucv_register_device(struc
 	struct device *dev = kzalloc(sizeof(struct device), GFP_KERNEL);
 	int ret;
 
-
 	IUCV_DBF_TEXT(trace, 3, __func__);
 
 	if (dev) {
@@ -2100,6 +2187,7 @@ static void __exit netiucv_exit(void)
 		netiucv_unregister_device(dev);
 	}
 
+	device_unregister(netiucv_dev);
 	driver_unregister(&netiucv_driver);
 	iucv_unregister(&netiucv_handler, 1);
 	iucv_unregister_dbf_views();
@@ -2125,10 +2213,25 @@ static int __init netiucv_init(void)
 		IUCV_DBF_TEXT_(setup, 2, "ret %d from driver_register\n", rc);
 		goto out_iucv;
 	}
-
+	/* establish dummy device */
+	netiucv_dev = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!netiucv_dev) {
+		rc = -ENOMEM;
+		goto out_driver;
+	}
+	dev_set_name(netiucv_dev, "netiucv");
+	netiucv_dev->bus = &iucv_bus;
+	netiucv_dev->parent = iucv_root;
+	netiucv_dev->release = (void (*)(struct device *))kfree;
+	netiucv_dev->driver = &netiucv_driver;
+	rc = device_register(netiucv_dev);
+	if (rc)
+		goto out_driver;
 	netiucv_banner();
 	return rc;
 
+out_driver:
+	driver_unregister(&netiucv_driver);
 out_iucv:
 	iucv_unregister(&netiucv_handler, 1);
 out_dbf:

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 30/38] PM: af_iucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_afiucv.patch --]
[-- Type: text/plain, Size: 4969 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Patch establishes a dummy afiucv-device to make sure af_iucv is
notified as iucv-bus device about suspend/resume.

The PM freeze callback severs all iucv pathes of connected af_iucv sockets.
The PM thaw/restore callback switches the state of all previously connected
sockets to IUCV_DISCONN.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 net/iucv/af_iucv.c |  147 +++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 144 insertions(+), 3 deletions(-)

Index: linux-2.6/net/iucv/af_iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/af_iucv.c
+++ linux-2.6/net/iucv/af_iucv.c
@@ -1,11 +1,12 @@
 /*
- *  linux/net/iucv/af_iucv.c
- *
  *  IUCV protocol stack for Linux on zSeries
  *
- *  Copyright 2006 IBM Corporation
+ *  Copyright IBM Corp. 2006, 2009
  *
  *  Author(s):	Jennifer Hunt <jenhunt@us.ibm.com>
+ *		Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
+ *  PM functions:
+ *		Ursula Braun <ursula.braun@de.ibm.com>
  */
 
 #define KMSG_COMPONENT "af_iucv"
@@ -78,6 +79,122 @@ static inline void low_nmcpy(unsigned ch
        memcpy(&dst[8], src, 8);
 }
 
+static int afiucv_pm_prepare(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_prepare\n");
+#endif
+	return 0;
+}
+
+static void afiucv_pm_complete(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_complete\n");
+#endif
+	return;
+}
+
+/**
+ * afiucv_pm_freeze() - Freeze PM callback
+ * @dev:	AFIUCV dummy device
+ *
+ * Sever all established IUCV communication pathes
+ */
+static int afiucv_pm_freeze(struct device *dev)
+{
+	struct iucv_sock *iucv;
+	struct sock *sk;
+	struct hlist_node *node;
+	int err = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_freeze\n");
+#endif
+	read_lock(&iucv_sk_list.lock);
+	sk_for_each(sk, node, &iucv_sk_list.head) {
+		iucv = iucv_sk(sk);
+		skb_queue_purge(&iucv->send_skb_q);
+		skb_queue_purge(&iucv->backlog_skb_q);
+		switch (sk->sk_state) {
+		case IUCV_SEVERED:
+		case IUCV_DISCONN:
+		case IUCV_CLOSING:
+		case IUCV_CONNECTED:
+			if (iucv->path) {
+				err = iucv_path_sever(iucv->path, NULL);
+				iucv_path_free(iucv->path);
+				iucv->path = NULL;
+			}
+			break;
+		case IUCV_OPEN:
+		case IUCV_BOUND:
+		case IUCV_LISTEN:
+		case IUCV_CLOSED:
+		default:
+			break;
+		}
+	}
+	read_unlock(&iucv_sk_list.lock);
+	return err;
+}
+
+/**
+ * afiucv_pm_restore_thaw() - Thaw and restore PM callback
+ * @dev:	AFIUCV dummy device
+ *
+ * socket clean up after freeze
+ */
+static int afiucv_pm_restore_thaw(struct device *dev)
+{
+	struct iucv_sock *iucv;
+	struct sock *sk;
+	struct hlist_node *node;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_restore_thaw\n");
+#endif
+	read_lock(&iucv_sk_list.lock);
+	sk_for_each(sk, node, &iucv_sk_list.head) {
+		iucv = iucv_sk(sk);
+		switch (sk->sk_state) {
+		case IUCV_CONNECTED:
+			sk->sk_err = EPIPE;
+			sk->sk_state = IUCV_DISCONN;
+			sk->sk_state_change(sk);
+			break;
+		case IUCV_DISCONN:
+		case IUCV_SEVERED:
+		case IUCV_CLOSING:
+		case IUCV_LISTEN:
+		case IUCV_BOUND:
+		case IUCV_OPEN:
+		default:
+			break;
+		}
+	}
+	read_unlock(&iucv_sk_list.lock);
+	return 0;
+}
+
+static struct dev_pm_ops afiucv_pm_ops = {
+	.prepare = afiucv_pm_prepare,
+	.complete = afiucv_pm_complete,
+	.freeze = afiucv_pm_freeze,
+	.thaw = afiucv_pm_restore_thaw,
+	.restore = afiucv_pm_restore_thaw,
+};
+
+static struct device_driver af_iucv_driver = {
+	.owner = THIS_MODULE,
+	.name = "afiucv",
+	.bus  = &iucv_bus,
+	.pm   = &afiucv_pm_ops,
+};
+
+/* dummy device used as trigger for PM functions */
+static struct device *af_iucv_dev;
+
 /* Timers */
 static void iucv_sock_timeout(unsigned long arg)
 {
@@ -1258,8 +1375,30 @@ static int __init afiucv_init(void)
 	err = sock_register(&iucv_sock_family_ops);
 	if (err)
 		goto out_proto;
+	/* establish dummy device */
+	err = driver_register(&af_iucv_driver);
+	if (err)
+		goto out_sock;
+	af_iucv_dev = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!af_iucv_dev) {
+		err = -ENOMEM;
+		goto out_driver;
+	}
+	dev_set_name(af_iucv_dev, "af_iucv");
+	af_iucv_dev->bus = &iucv_bus;
+	af_iucv_dev->parent = iucv_root;
+	af_iucv_dev->release = (void (*)(struct device *))kfree;
+	af_iucv_dev->driver = &af_iucv_driver;
+	err = device_register(af_iucv_dev);
+	if (err)
+		goto out_driver;
+
 	return 0;
 
+out_driver:
+	driver_unregister(&af_iucv_driver);
+out_sock:
+	sock_unregister(PF_IUCV);
 out_proto:
 	proto_unregister(&iucv_proto);
 out_iucv:
@@ -1270,6 +1409,8 @@ out:
 
 static void __exit afiucv_exit(void)
 {
+	device_unregister(af_iucv_dev);
+	driver_unregister(&af_iucv_driver);
 	sock_unregister(PF_IUCV);
 	proto_unregister(&iucv_proto);
 	iucv_unregister(&af_iucv_handler, 0);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 30/38] PM: af_iucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_afiucv.patch --]
[-- Type: text/plain, Size: 4968 bytes --]

From: Ursula Braun <ursula.braun@de.ibm.com>

Patch establishes a dummy afiucv-device to make sure af_iucv is
notified as iucv-bus device about suspend/resume.

The PM freeze callback severs all iucv pathes of connected af_iucv sockets.
The PM thaw/restore callback switches the state of all previously connected
sockets to IUCV_DISCONN.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 net/iucv/af_iucv.c |  147 +++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 144 insertions(+), 3 deletions(-)

Index: linux-2.6/net/iucv/af_iucv.c
===================================================================
--- linux-2.6.orig/net/iucv/af_iucv.c
+++ linux-2.6/net/iucv/af_iucv.c
@@ -1,11 +1,12 @@
 /*
- *  linux/net/iucv/af_iucv.c
- *
  *  IUCV protocol stack for Linux on zSeries
  *
- *  Copyright 2006 IBM Corporation
+ *  Copyright IBM Corp. 2006, 2009
  *
  *  Author(s):	Jennifer Hunt <jenhunt@us.ibm.com>
+ *		Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
+ *  PM functions:
+ *		Ursula Braun <ursula.braun@de.ibm.com>
  */
 
 #define KMSG_COMPONENT "af_iucv"
@@ -78,6 +79,122 @@ static inline void low_nmcpy(unsigned ch
        memcpy(&dst[8], src, 8);
 }
 
+static int afiucv_pm_prepare(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_prepare\n");
+#endif
+	return 0;
+}
+
+static void afiucv_pm_complete(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_complete\n");
+#endif
+	return;
+}
+
+/**
+ * afiucv_pm_freeze() - Freeze PM callback
+ * @dev:	AFIUCV dummy device
+ *
+ * Sever all established IUCV communication pathes
+ */
+static int afiucv_pm_freeze(struct device *dev)
+{
+	struct iucv_sock *iucv;
+	struct sock *sk;
+	struct hlist_node *node;
+	int err = 0;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_freeze\n");
+#endif
+	read_lock(&iucv_sk_list.lock);
+	sk_for_each(sk, node, &iucv_sk_list.head) {
+		iucv = iucv_sk(sk);
+		skb_queue_purge(&iucv->send_skb_q);
+		skb_queue_purge(&iucv->backlog_skb_q);
+		switch (sk->sk_state) {
+		case IUCV_SEVERED:
+		case IUCV_DISCONN:
+		case IUCV_CLOSING:
+		case IUCV_CONNECTED:
+			if (iucv->path) {
+				err = iucv_path_sever(iucv->path, NULL);
+				iucv_path_free(iucv->path);
+				iucv->path = NULL;
+			}
+			break;
+		case IUCV_OPEN:
+		case IUCV_BOUND:
+		case IUCV_LISTEN:
+		case IUCV_CLOSED:
+		default:
+			break;
+		}
+	}
+	read_unlock(&iucv_sk_list.lock);
+	return err;
+}
+
+/**
+ * afiucv_pm_restore_thaw() - Thaw and restore PM callback
+ * @dev:	AFIUCV dummy device
+ *
+ * socket clean up after freeze
+ */
+static int afiucv_pm_restore_thaw(struct device *dev)
+{
+	struct iucv_sock *iucv;
+	struct sock *sk;
+	struct hlist_node *node;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "afiucv_pm_restore_thaw\n");
+#endif
+	read_lock(&iucv_sk_list.lock);
+	sk_for_each(sk, node, &iucv_sk_list.head) {
+		iucv = iucv_sk(sk);
+		switch (sk->sk_state) {
+		case IUCV_CONNECTED:
+			sk->sk_err = EPIPE;
+			sk->sk_state = IUCV_DISCONN;
+			sk->sk_state_change(sk);
+			break;
+		case IUCV_DISCONN:
+		case IUCV_SEVERED:
+		case IUCV_CLOSING:
+		case IUCV_LISTEN:
+		case IUCV_BOUND:
+		case IUCV_OPEN:
+		default:
+			break;
+		}
+	}
+	read_unlock(&iucv_sk_list.lock);
+	return 0;
+}
+
+static struct dev_pm_ops afiucv_pm_ops = {
+	.prepare = afiucv_pm_prepare,
+	.complete = afiucv_pm_complete,
+	.freeze = afiucv_pm_freeze,
+	.thaw = afiucv_pm_restore_thaw,
+	.restore = afiucv_pm_restore_thaw,
+};
+
+static struct device_driver af_iucv_driver = {
+	.owner = THIS_MODULE,
+	.name = "afiucv",
+	.bus  = &iucv_bus,
+	.pm   = &afiucv_pm_ops,
+};
+
+/* dummy device used as trigger for PM functions */
+static struct device *af_iucv_dev;
+
 /* Timers */
 static void iucv_sock_timeout(unsigned long arg)
 {
@@ -1258,8 +1375,30 @@ static int __init afiucv_init(void)
 	err = sock_register(&iucv_sock_family_ops);
 	if (err)
 		goto out_proto;
+	/* establish dummy device */
+	err = driver_register(&af_iucv_driver);
+	if (err)
+		goto out_sock;
+	af_iucv_dev = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!af_iucv_dev) {
+		err = -ENOMEM;
+		goto out_driver;
+	}
+	dev_set_name(af_iucv_dev, "af_iucv");
+	af_iucv_dev->bus = &iucv_bus;
+	af_iucv_dev->parent = iucv_root;
+	af_iucv_dev->release = (void (*)(struct device *))kfree;
+	af_iucv_dev->driver = &af_iucv_driver;
+	err = device_register(af_iucv_dev);
+	if (err)
+		goto out_driver;
+
 	return 0;
 
+out_driver:
+	driver_unregister(&af_iucv_driver);
+out_sock:
+	sock_unregister(PF_IUCV);
 out_proto:
 	proto_unregister(&iucv_proto);
 out_iucv:
@@ -1270,6 +1409,8 @@ out:
 
 static void __exit afiucv_exit(void)
 {
+	device_unregister(af_iucv_dev);
+	driver_unregister(&af_iucv_driver);
 	sock_unregister(PF_IUCV);
 	proto_unregister(&iucv_proto);
 	iucv_unregister(&af_iucv_handler, 0);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 31/38] pm: hvc_iucv power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_hvc_iucv.patch --]
[-- Type: text/plain, Size: 10135 bytes --]

From: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>

The patch adds supporting for suspending and resuming IUCV HVC terminal
devices from disk. The obligatory Linux device driver interfaces has
been added by registering a device driver on the IUCV bus.
For each IUCV HVC terminal device the driver creates a respective device
on the IUCV bus.

To support suspend and resume, the PM freeze callback severs any established
IUCV communication path and triggers a HVC tty hang-up when the system image
is restored.
IUCV communication path are no longer valid when the z/VM guest is halted.

The device driver initialization has been updated to register devices and
the a new routine has been extracted to facilitate the hang-up of IUCV HVC
terminal devices.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/char/hvc_iucv.c |  204 +++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 161 insertions(+), 43 deletions(-)

Index: linux-2.6/drivers/char/hvc_iucv.c
===================================================================
--- linux-2.6.orig/drivers/char/hvc_iucv.c
+++ linux-2.6/drivers/char/hvc_iucv.c
@@ -4,7 +4,7 @@
  * This HVC device driver provides terminal access using
  * z/VM IUCV communication paths.
  *
- * Copyright IBM Corp. 2008
+ * Copyright IBM Corp. 2008, 2009
  *
  * Author(s):	Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
  */
@@ -15,6 +15,7 @@
 #include <asm/ebcdic.h>
 #include <linux/ctype.h>
 #include <linux/delay.h>
+#include <linux/device.h>
 #include <linux/init.h>
 #include <linux/mempool.h>
 #include <linux/moduleparam.h>
@@ -74,6 +75,7 @@ struct hvc_iucv_private {
 	wait_queue_head_t	sndbuf_waitq;	/* wait for send completion */
 	struct list_head	tty_outqueue;	/* outgoing IUCV messages */
 	struct list_head	tty_inqueue;	/* incoming IUCV messages */
+	struct device		*dev;		/* device structure */
 };
 
 struct iucv_tty_buffer {
@@ -542,7 +544,68 @@ static void flush_sndbuf_sync(struct hvc
 
 	if (sync_wait)
 		wait_event_timeout(priv->sndbuf_waitq,
-				   tty_outqueue_empty(priv), HZ);
+				   tty_outqueue_empty(priv), HZ/10);
+}
+
+/**
+ * hvc_iucv_hangup() - Sever IUCV path and schedule hvc tty hang up
+ * @priv:	Pointer to hvc_iucv_private structure
+ *
+ * This routine severs an existing IUCV communication path and hangs
+ * up the underlying HVC terminal device.
+ * The hang-up occurs only if an IUCV communication path is established;
+ * otherwise there is no need to hang up the terminal device.
+ *
+ * The IUCV HVC hang-up is separated into two steps:
+ * 1. After the IUCV path has been severed, the iucv_state is set to
+ *    IUCV_SEVERED.
+ * 2. Later, when the HVC thread calls hvc_iucv_get_chars(), the
+ *    IUCV_SEVERED state causes the tty hang-up in the HVC layer.
+ *
+ * If the tty has not yet been opened, clean up the hvc_iucv_private
+ * structure to allow re-connects.
+ * If the tty has been opened, let get_chars() return -EPIPE to signal
+ * the HVC layer to hang up the tty and, if so, wake up the HVC thread
+ * to call get_chars()...
+ *
+ * Special notes on hanging up a HVC terminal instantiated as console:
+ * Hang-up:	1. do_tty_hangup() replaces file ops (= hung_up_tty_fops)
+ *		2. do_tty_hangup() calls tty->ops->close() for console_filp
+ *			=> no hangup notifier is called by HVC (default)
+ *		2. hvc_close() returns because of tty_hung_up_p(filp)
+ *			=> no delete notifier is called!
+ * Finally, the back-end is not being notified, thus, the tty session is
+ * kept active (TTY_OPEN) to be ready for re-connects.
+ *
+ * Locking:	spin_lock(&priv->lock) w/o disabling bh
+ */
+static void hvc_iucv_hangup(struct hvc_iucv_private *priv)
+{
+	struct iucv_path *path;
+
+	path = NULL;
+	spin_lock(&priv->lock);
+	if (priv->iucv_state == IUCV_CONNECTED) {
+		path = priv->path;
+		priv->path = NULL;
+		priv->iucv_state = IUCV_SEVERED;
+		if (priv->tty_state == TTY_CLOSED)
+			hvc_iucv_cleanup(priv);
+		else
+			/* console is special (see above) */
+			if (priv->is_console) {
+				hvc_iucv_cleanup(priv);
+				priv->tty_state = TTY_OPENED;
+			} else
+				hvc_kick();
+	}
+	spin_unlock(&priv->lock);
+
+	/* finally sever path (outside of priv->lock due to lock ordering) */
+	if (path) {
+		iucv_path_sever(path, NULL);
+		iucv_path_free(path);
+	}
 }
 
 /**
@@ -735,11 +798,8 @@ out_path_handled:
  * @ipuser:	User specified data for this path
  *		(AF_IUCV: port/service name and originator port)
  *
- * The function also severs the path (as required by the IUCV protocol) and
- * sets the iucv state to IUCV_SEVERED for the associated struct
- * hvc_iucv_private instance. Later, the IUCV_SEVERED state triggers a tty
- * hangup (hvc_iucv_get_chars() / hvc_iucv_write()).
- * If tty portion of the HVC is closed, clean up the outqueue.
+ * This function calls the hvc_iucv_hangup() function for the
+ * respective IUCV HVC terminal.
  *
  * Locking:	struct hvc_iucv_private->lock
  */
@@ -747,33 +807,7 @@ static void hvc_iucv_path_severed(struct
 {
 	struct hvc_iucv_private *priv = path->private;
 
-	spin_lock(&priv->lock);
-	priv->iucv_state = IUCV_SEVERED;
-
-	/* If the tty has not yet been opened, clean up the hvc_iucv_private
-	 * structure to allow re-connects.
-	 * This is also done for our console device because console hangups
-	 * are handled specially and no notifier is called by HVC.
-	 * The tty session is active (TTY_OPEN) and ready for re-connects...
-	 *
-	 * If it has been opened, let get_chars() return -EPIPE to signal the
-	 * HVC layer to hang up the tty.
-	 * If so, we need to wake up the HVC thread to call get_chars()...
-	 */
-	priv->path = NULL;
-	if (priv->tty_state == TTY_CLOSED)
-		hvc_iucv_cleanup(priv);
-	else
-		if (priv->is_console) {
-			hvc_iucv_cleanup(priv);
-			priv->tty_state = TTY_OPENED;
-		} else
-			hvc_kick();
-	spin_unlock(&priv->lock);
-
-	/* finally sever path (outside of priv->lock due to lock ordering) */
-	iucv_path_sever(path, ipuser);
-	iucv_path_free(path);
+	hvc_iucv_hangup(priv);
 }
 
 /**
@@ -853,6 +887,37 @@ static void hvc_iucv_msg_complete(struct
 	destroy_tty_buffer_list(&list_remove);
 }
 
+/**
+ * hvc_iucv_pm_freeze() - Freeze PM callback
+ * @dev:	IUVC HVC terminal device
+ *
+ * Sever an established IUCV communication path and
+ * trigger a hang-up of the underlying HVC terminal.
+ */
+static int hvc_iucv_pm_freeze(struct device *dev)
+{
+	struct hvc_iucv_private *priv = dev_get_drvdata(dev);
+
+	local_bh_disable();
+	hvc_iucv_hangup(priv);
+	local_bh_enable();
+
+	return 0;
+}
+
+/**
+ * hvc_iucv_pm_restore_thaw() - Thaw and restore PM callback
+ * @dev:	IUVC HVC terminal device
+ *
+ * Wake up the HVC thread to trigger hang-up and respective
+ * HVC back-end notifier invocations.
+ */
+static int hvc_iucv_pm_restore_thaw(struct device *dev)
+{
+	hvc_kick();
+	return 0;
+}
+
 
 /* HVC operations */
 static struct hv_ops hvc_iucv_ops = {
@@ -863,6 +928,20 @@ static struct hv_ops hvc_iucv_ops = {
 	.notifier_hangup = hvc_iucv_notifier_hangup,
 };
 
+/* Suspend / resume device operations */
+static struct dev_pm_ops hvc_iucv_pm_ops = {
+	.freeze	  = hvc_iucv_pm_freeze,
+	.thaw	  = hvc_iucv_pm_restore_thaw,
+	.restore  = hvc_iucv_pm_restore_thaw,
+};
+
+/* IUCV HVC device driver */
+static struct device_driver hvc_iucv_driver = {
+	.name = KMSG_COMPONENT,
+	.bus  = &iucv_bus,
+	.pm   = &hvc_iucv_pm_ops,
+};
+
 /**
  * hvc_iucv_alloc() - Allocates a new struct hvc_iucv_private instance
  * @id:			hvc_iucv_table index
@@ -897,14 +976,12 @@ static int __init hvc_iucv_alloc(int id,
 	/* set console flag */
 	priv->is_console = is_console;
 
-	/* finally allocate hvc */
+	/* allocate hvc device */
 	priv->hvc = hvc_alloc(HVC_IUCV_MAGIC + id, /*		  PAGE_SIZE */
 			      HVC_IUCV_MAGIC + id, &hvc_iucv_ops, 256);
 	if (IS_ERR(priv->hvc)) {
 		rc = PTR_ERR(priv->hvc);
-		free_page((unsigned long) priv->sndbuf);
-		kfree(priv);
-		return rc;
+		goto out_error_hvc;
 	}
 
 	/* notify HVC thread instead of using polling */
@@ -915,8 +992,45 @@ static int __init hvc_iucv_alloc(int id,
 	memcpy(priv->srv_name, name, 8);
 	ASCEBC(priv->srv_name, 8);
 
+	/* create and setup device */
+	priv->dev = kzalloc(sizeof(*priv->dev), GFP_KERNEL);
+	if (!priv->dev) {
+		rc = -ENOMEM;
+		goto out_error_dev;
+	}
+	dev_set_name(priv->dev, "hvc_iucv%d", id);
+	dev_set_drvdata(priv->dev, priv);
+	priv->dev->bus = &iucv_bus;
+	priv->dev->parent = iucv_root;
+	priv->dev->driver = &hvc_iucv_driver;
+	priv->dev->release = (void (*)(struct device *)) kfree;
+	rc = device_register(priv->dev);
+	if (rc) {
+		kfree(priv->dev);
+		goto out_error_dev;
+	}
+
 	hvc_iucv_table[id] = priv;
 	return 0;
+
+out_error_dev:
+	hvc_remove(priv->hvc);
+out_error_hvc:
+	free_page((unsigned long) priv->sndbuf);
+	kfree(priv);
+
+	return rc;
+}
+
+/**
+ * hvc_iucv_destroy() - Destroy and free hvc_iucv_private instances
+ */
+static void __init hvc_iucv_destroy(struct hvc_iucv_private *priv)
+{
+	hvc_remove(priv->hvc);
+	device_unregister(priv->dev);
+	free_page((unsigned long) priv->sndbuf);
+	kfree(priv);
 }
 
 /**
@@ -1109,6 +1223,11 @@ static int __init hvc_iucv_init(void)
 		goto out_error;
 	}
 
+	/* register IUCV HVC device driver */
+	rc = driver_register(&hvc_iucv_driver);
+	if (rc)
+		goto out_error;
+
 	/* parse hvc_iucv_allow string and create z/VM user ID filter list */
 	if (hvc_iucv_filter_string) {
 		rc = hvc_iucv_setup_filter(hvc_iucv_filter_string);
@@ -1183,15 +1302,14 @@ out_error_iucv:
 	iucv_unregister(&hvc_iucv_handler, 0);
 out_error_hvc:
 	for (i = 0; i < hvc_iucv_devices; i++)
-		if (hvc_iucv_table[i]) {
-			if (hvc_iucv_table[i]->hvc)
-				hvc_remove(hvc_iucv_table[i]->hvc);
-			kfree(hvc_iucv_table[i]);
-		}
+		if (hvc_iucv_table[i])
+			hvc_iucv_destroy(hvc_iucv_table[i]);
 out_error_memory:
 	mempool_destroy(hvc_iucv_mempool);
 	kmem_cache_destroy(hvc_iucv_buffer_cache);
 out_error:
+	if (hvc_iucv_filter)
+		kfree(hvc_iucv_filter);
 	hvc_iucv_devices = 0; /* ensure that we do not provide any device */
 	return rc;
 }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 31/38] pm: hvc_iucv power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_hvc_iucv.patch --]
[-- Type: text/plain, Size: 10134 bytes --]

From: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>

The patch adds supporting for suspending and resuming IUCV HVC terminal
devices from disk. The obligatory Linux device driver interfaces has
been added by registering a device driver on the IUCV bus.
For each IUCV HVC terminal device the driver creates a respective device
on the IUCV bus.

To support suspend and resume, the PM freeze callback severs any established
IUCV communication path and triggers a HVC tty hang-up when the system image
is restored.
IUCV communication path are no longer valid when the z/VM guest is halted.

The device driver initialization has been updated to register devices and
the a new routine has been extracted to facilitate the hang-up of IUCV HVC
terminal devices.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/char/hvc_iucv.c |  204 +++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 161 insertions(+), 43 deletions(-)

Index: linux-2.6/drivers/char/hvc_iucv.c
===================================================================
--- linux-2.6.orig/drivers/char/hvc_iucv.c
+++ linux-2.6/drivers/char/hvc_iucv.c
@@ -4,7 +4,7 @@
  * This HVC device driver provides terminal access using
  * z/VM IUCV communication paths.
  *
- * Copyright IBM Corp. 2008
+ * Copyright IBM Corp. 2008, 2009
  *
  * Author(s):	Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
  */
@@ -15,6 +15,7 @@
 #include <asm/ebcdic.h>
 #include <linux/ctype.h>
 #include <linux/delay.h>
+#include <linux/device.h>
 #include <linux/init.h>
 #include <linux/mempool.h>
 #include <linux/moduleparam.h>
@@ -74,6 +75,7 @@ struct hvc_iucv_private {
 	wait_queue_head_t	sndbuf_waitq;	/* wait for send completion */
 	struct list_head	tty_outqueue;	/* outgoing IUCV messages */
 	struct list_head	tty_inqueue;	/* incoming IUCV messages */
+	struct device		*dev;		/* device structure */
 };
 
 struct iucv_tty_buffer {
@@ -542,7 +544,68 @@ static void flush_sndbuf_sync(struct hvc
 
 	if (sync_wait)
 		wait_event_timeout(priv->sndbuf_waitq,
-				   tty_outqueue_empty(priv), HZ);
+				   tty_outqueue_empty(priv), HZ/10);
+}
+
+/**
+ * hvc_iucv_hangup() - Sever IUCV path and schedule hvc tty hang up
+ * @priv:	Pointer to hvc_iucv_private structure
+ *
+ * This routine severs an existing IUCV communication path and hangs
+ * up the underlying HVC terminal device.
+ * The hang-up occurs only if an IUCV communication path is established;
+ * otherwise there is no need to hang up the terminal device.
+ *
+ * The IUCV HVC hang-up is separated into two steps:
+ * 1. After the IUCV path has been severed, the iucv_state is set to
+ *    IUCV_SEVERED.
+ * 2. Later, when the HVC thread calls hvc_iucv_get_chars(), the
+ *    IUCV_SEVERED state causes the tty hang-up in the HVC layer.
+ *
+ * If the tty has not yet been opened, clean up the hvc_iucv_private
+ * structure to allow re-connects.
+ * If the tty has been opened, let get_chars() return -EPIPE to signal
+ * the HVC layer to hang up the tty and, if so, wake up the HVC thread
+ * to call get_chars()...
+ *
+ * Special notes on hanging up a HVC terminal instantiated as console:
+ * Hang-up:	1. do_tty_hangup() replaces file ops (= hung_up_tty_fops)
+ *		2. do_tty_hangup() calls tty->ops->close() for console_filp
+ *			=> no hangup notifier is called by HVC (default)
+ *		2. hvc_close() returns because of tty_hung_up_p(filp)
+ *			=> no delete notifier is called!
+ * Finally, the back-end is not being notified, thus, the tty session is
+ * kept active (TTY_OPEN) to be ready for re-connects.
+ *
+ * Locking:	spin_lock(&priv->lock) w/o disabling bh
+ */
+static void hvc_iucv_hangup(struct hvc_iucv_private *priv)
+{
+	struct iucv_path *path;
+
+	path = NULL;
+	spin_lock(&priv->lock);
+	if (priv->iucv_state == IUCV_CONNECTED) {
+		path = priv->path;
+		priv->path = NULL;
+		priv->iucv_state = IUCV_SEVERED;
+		if (priv->tty_state == TTY_CLOSED)
+			hvc_iucv_cleanup(priv);
+		else
+			/* console is special (see above) */
+			if (priv->is_console) {
+				hvc_iucv_cleanup(priv);
+				priv->tty_state = TTY_OPENED;
+			} else
+				hvc_kick();
+	}
+	spin_unlock(&priv->lock);
+
+	/* finally sever path (outside of priv->lock due to lock ordering) */
+	if (path) {
+		iucv_path_sever(path, NULL);
+		iucv_path_free(path);
+	}
 }
 
 /**
@@ -735,11 +798,8 @@ out_path_handled:
  * @ipuser:	User specified data for this path
  *		(AF_IUCV: port/service name and originator port)
  *
- * The function also severs the path (as required by the IUCV protocol) and
- * sets the iucv state to IUCV_SEVERED for the associated struct
- * hvc_iucv_private instance. Later, the IUCV_SEVERED state triggers a tty
- * hangup (hvc_iucv_get_chars() / hvc_iucv_write()).
- * If tty portion of the HVC is closed, clean up the outqueue.
+ * This function calls the hvc_iucv_hangup() function for the
+ * respective IUCV HVC terminal.
  *
  * Locking:	struct hvc_iucv_private->lock
  */
@@ -747,33 +807,7 @@ static void hvc_iucv_path_severed(struct
 {
 	struct hvc_iucv_private *priv = path->private;
 
-	spin_lock(&priv->lock);
-	priv->iucv_state = IUCV_SEVERED;
-
-	/* If the tty has not yet been opened, clean up the hvc_iucv_private
-	 * structure to allow re-connects.
-	 * This is also done for our console device because console hangups
-	 * are handled specially and no notifier is called by HVC.
-	 * The tty session is active (TTY_OPEN) and ready for re-connects...
-	 *
-	 * If it has been opened, let get_chars() return -EPIPE to signal the
-	 * HVC layer to hang up the tty.
-	 * If so, we need to wake up the HVC thread to call get_chars()...
-	 */
-	priv->path = NULL;
-	if (priv->tty_state == TTY_CLOSED)
-		hvc_iucv_cleanup(priv);
-	else
-		if (priv->is_console) {
-			hvc_iucv_cleanup(priv);
-			priv->tty_state = TTY_OPENED;
-		} else
-			hvc_kick();
-	spin_unlock(&priv->lock);
-
-	/* finally sever path (outside of priv->lock due to lock ordering) */
-	iucv_path_sever(path, ipuser);
-	iucv_path_free(path);
+	hvc_iucv_hangup(priv);
 }
 
 /**
@@ -853,6 +887,37 @@ static void hvc_iucv_msg_complete(struct
 	destroy_tty_buffer_list(&list_remove);
 }
 
+/**
+ * hvc_iucv_pm_freeze() - Freeze PM callback
+ * @dev:	IUVC HVC terminal device
+ *
+ * Sever an established IUCV communication path and
+ * trigger a hang-up of the underlying HVC terminal.
+ */
+static int hvc_iucv_pm_freeze(struct device *dev)
+{
+	struct hvc_iucv_private *priv = dev_get_drvdata(dev);
+
+	local_bh_disable();
+	hvc_iucv_hangup(priv);
+	local_bh_enable();
+
+	return 0;
+}
+
+/**
+ * hvc_iucv_pm_restore_thaw() - Thaw and restore PM callback
+ * @dev:	IUVC HVC terminal device
+ *
+ * Wake up the HVC thread to trigger hang-up and respective
+ * HVC back-end notifier invocations.
+ */
+static int hvc_iucv_pm_restore_thaw(struct device *dev)
+{
+	hvc_kick();
+	return 0;
+}
+
 
 /* HVC operations */
 static struct hv_ops hvc_iucv_ops = {
@@ -863,6 +928,20 @@ static struct hv_ops hvc_iucv_ops = {
 	.notifier_hangup = hvc_iucv_notifier_hangup,
 };
 
+/* Suspend / resume device operations */
+static struct dev_pm_ops hvc_iucv_pm_ops = {
+	.freeze	  = hvc_iucv_pm_freeze,
+	.thaw	  = hvc_iucv_pm_restore_thaw,
+	.restore  = hvc_iucv_pm_restore_thaw,
+};
+
+/* IUCV HVC device driver */
+static struct device_driver hvc_iucv_driver = {
+	.name = KMSG_COMPONENT,
+	.bus  = &iucv_bus,
+	.pm   = &hvc_iucv_pm_ops,
+};
+
 /**
  * hvc_iucv_alloc() - Allocates a new struct hvc_iucv_private instance
  * @id:			hvc_iucv_table index
@@ -897,14 +976,12 @@ static int __init hvc_iucv_alloc(int id,
 	/* set console flag */
 	priv->is_console = is_console;
 
-	/* finally allocate hvc */
+	/* allocate hvc device */
 	priv->hvc = hvc_alloc(HVC_IUCV_MAGIC + id, /*		  PAGE_SIZE */
 			      HVC_IUCV_MAGIC + id, &hvc_iucv_ops, 256);
 	if (IS_ERR(priv->hvc)) {
 		rc = PTR_ERR(priv->hvc);
-		free_page((unsigned long) priv->sndbuf);
-		kfree(priv);
-		return rc;
+		goto out_error_hvc;
 	}
 
 	/* notify HVC thread instead of using polling */
@@ -915,8 +992,45 @@ static int __init hvc_iucv_alloc(int id,
 	memcpy(priv->srv_name, name, 8);
 	ASCEBC(priv->srv_name, 8);
 
+	/* create and setup device */
+	priv->dev = kzalloc(sizeof(*priv->dev), GFP_KERNEL);
+	if (!priv->dev) {
+		rc = -ENOMEM;
+		goto out_error_dev;
+	}
+	dev_set_name(priv->dev, "hvc_iucv%d", id);
+	dev_set_drvdata(priv->dev, priv);
+	priv->dev->bus = &iucv_bus;
+	priv->dev->parent = iucv_root;
+	priv->dev->driver = &hvc_iucv_driver;
+	priv->dev->release = (void (*)(struct device *)) kfree;
+	rc = device_register(priv->dev);
+	if (rc) {
+		kfree(priv->dev);
+		goto out_error_dev;
+	}
+
 	hvc_iucv_table[id] = priv;
 	return 0;
+
+out_error_dev:
+	hvc_remove(priv->hvc);
+out_error_hvc:
+	free_page((unsigned long) priv->sndbuf);
+	kfree(priv);
+
+	return rc;
+}
+
+/**
+ * hvc_iucv_destroy() - Destroy and free hvc_iucv_private instances
+ */
+static void __init hvc_iucv_destroy(struct hvc_iucv_private *priv)
+{
+	hvc_remove(priv->hvc);
+	device_unregister(priv->dev);
+	free_page((unsigned long) priv->sndbuf);
+	kfree(priv);
 }
 
 /**
@@ -1109,6 +1223,11 @@ static int __init hvc_iucv_init(void)
 		goto out_error;
 	}
 
+	/* register IUCV HVC device driver */
+	rc = driver_register(&hvc_iucv_driver);
+	if (rc)
+		goto out_error;
+
 	/* parse hvc_iucv_allow string and create z/VM user ID filter list */
 	if (hvc_iucv_filter_string) {
 		rc = hvc_iucv_setup_filter(hvc_iucv_filter_string);
@@ -1183,15 +1302,14 @@ out_error_iucv:
 	iucv_unregister(&hvc_iucv_handler, 0);
 out_error_hvc:
 	for (i = 0; i < hvc_iucv_devices; i++)
-		if (hvc_iucv_table[i]) {
-			if (hvc_iucv_table[i]->hvc)
-				hvc_remove(hvc_iucv_table[i]->hvc);
-			kfree(hvc_iucv_table[i]);
-		}
+		if (hvc_iucv_table[i])
+			hvc_iucv_destroy(hvc_iucv_table[i]);
 out_error_memory:
 	mempool_destroy(hvc_iucv_mempool);
 	kmem_cache_destroy(hvc_iucv_buffer_cache);
 out_error:
+	if (hvc_iucv_filter)
+		kfree(hvc_iucv_filter);
 	hvc_iucv_devices = 0; /* ensure that we do not provide any device */
 	return rc;
 }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 32/38] pm: smsgiucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_smsgiucv.patch --]
[-- Type: text/plain, Size: 3281 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Create dummy iucv-device to get control when the system is suspended
and resumed. Server the smsg iucv path on suspend, reestablish the
path on resume.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/smsgiucv.c |   63 +++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 60 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/s390/net/smsgiucv.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/smsgiucv.c
+++ linux-2.6/drivers/s390/net/smsgiucv.c
@@ -1,7 +1,8 @@
 /*
  * IUCV special message driver
  *
- * Copyright 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Copyright IBM Corp. 2003, 2009
+ *
  * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
  *
  * This program is free software; you can redistribute it and/or modify
@@ -40,6 +41,8 @@ MODULE_AUTHOR
 MODULE_DESCRIPTION ("Linux for S/390 IUCV special message driver");
 
 static struct iucv_path *smsg_path;
+/* dummy device used as trigger for PM functions */
+static struct device *smsg_dev;
 
 static DEFINE_SPINLOCK(smsg_list_lock);
 static LIST_HEAD(smsg_list);
@@ -132,14 +135,51 @@ void smsg_unregister_callback(char *pref
 	kfree(cb);
 }
 
+static int smsg_pm_freeze(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "smsg_pm_freeze\n");
+#endif
+	if (smsg_path)
+		iucv_path_sever(smsg_path, NULL);
+	return 0;
+}
+
+static int smsg_pm_restore_thaw(struct device *dev)
+{
+	int rc;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "smsg_pm_restore_thaw\n");
+#endif
+	if (smsg_path) {
+		memset(smsg_path, 0, sizeof(*smsg_path));
+		smsg_path->msglim = 255;
+		smsg_path->flags = 0;
+		rc = iucv_path_connect(smsg_path, &smsg_handler, "*MSG    ",
+				       NULL, NULL, NULL);
+		printk(KERN_ERR "iucv_path_connect returned with rc %i\n", rc);
+	}
+	return 0;
+}
+
+static struct dev_pm_ops smsg_pm_ops = {
+	.freeze = smsg_pm_freeze,
+	.thaw = smsg_pm_restore_thaw,
+	.restore = smsg_pm_restore_thaw,
+};
+
 static struct device_driver smsg_driver = {
+	.owner = THIS_MODULE,
 	.name = "SMSGIUCV",
 	.bus  = &iucv_bus,
+	.pm = &smsg_pm_ops,
 };
 
 static void __exit smsg_exit(void)
 {
 	cpcmd("SET SMSG IUCV", NULL, 0, NULL);
+	device_unregister(smsg_dev);
 	iucv_unregister(&smsg_handler, 1);
 	driver_unregister(&smsg_driver);
 }
@@ -166,12 +206,29 @@ static int __init smsg_init(void)
 	rc = iucv_path_connect(smsg_path, &smsg_handler, "*MSG    ",
 			       NULL, NULL, NULL);
 	if (rc)
-		goto out_free;
+		goto out_free_path;
+	smsg_dev = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!smsg_dev) {
+		rc = -ENOMEM;
+		goto out_free_path;
+	}
+	dev_set_name(smsg_dev, "smsg_iucv");
+	smsg_dev->bus = &iucv_bus;
+	smsg_dev->parent = iucv_root;
+	smsg_dev->release = (void (*)(struct device *))kfree;
+	smsg_dev->driver = &smsg_driver;
+	rc = device_register(smsg_dev);
+	if (rc)
+		goto out_free_dev;
+
 	cpcmd("SET SMSG IUCV", NULL, 0, NULL);
 	return 0;
 
-out_free:
+out_free_dev:
+	kfree(smsg_dev);
+out_free_path:
 	iucv_path_free(smsg_path);
+	smsg_path = NULL;
 out_register:
 	iucv_unregister(&smsg_handler, 1);
 out_driver:

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 32/38] pm: smsgiucv power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_smsgiucv.patch --]
[-- Type: text/plain, Size: 3280 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Create dummy iucv-device to get control when the system is suspended
and resumed. Server the smsg iucv path on suspend, reestablish the
path on resume.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/net/smsgiucv.c |   63 +++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 60 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/s390/net/smsgiucv.c
===================================================================
--- linux-2.6.orig/drivers/s390/net/smsgiucv.c
+++ linux-2.6/drivers/s390/net/smsgiucv.c
@@ -1,7 +1,8 @@
 /*
  * IUCV special message driver
  *
- * Copyright 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Copyright IBM Corp. 2003, 2009
+ *
  * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
  *
  * This program is free software; you can redistribute it and/or modify
@@ -40,6 +41,8 @@ MODULE_AUTHOR
 MODULE_DESCRIPTION ("Linux for S/390 IUCV special message driver");
 
 static struct iucv_path *smsg_path;
+/* dummy device used as trigger for PM functions */
+static struct device *smsg_dev;
 
 static DEFINE_SPINLOCK(smsg_list_lock);
 static LIST_HEAD(smsg_list);
@@ -132,14 +135,51 @@ void smsg_unregister_callback(char *pref
 	kfree(cb);
 }
 
+static int smsg_pm_freeze(struct device *dev)
+{
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "smsg_pm_freeze\n");
+#endif
+	if (smsg_path)
+		iucv_path_sever(smsg_path, NULL);
+	return 0;
+}
+
+static int smsg_pm_restore_thaw(struct device *dev)
+{
+	int rc;
+
+#ifdef CONFIG_PM_DEBUG
+	printk(KERN_WARNING "smsg_pm_restore_thaw\n");
+#endif
+	if (smsg_path) {
+		memset(smsg_path, 0, sizeof(*smsg_path));
+		smsg_path->msglim = 255;
+		smsg_path->flags = 0;
+		rc = iucv_path_connect(smsg_path, &smsg_handler, "*MSG    ",
+				       NULL, NULL, NULL);
+		printk(KERN_ERR "iucv_path_connect returned with rc %i\n", rc);
+	}
+	return 0;
+}
+
+static struct dev_pm_ops smsg_pm_ops = {
+	.freeze = smsg_pm_freeze,
+	.thaw = smsg_pm_restore_thaw,
+	.restore = smsg_pm_restore_thaw,
+};
+
 static struct device_driver smsg_driver = {
+	.owner = THIS_MODULE,
 	.name = "SMSGIUCV",
 	.bus  = &iucv_bus,
+	.pm = &smsg_pm_ops,
 };
 
 static void __exit smsg_exit(void)
 {
 	cpcmd("SET SMSG IUCV", NULL, 0, NULL);
+	device_unregister(smsg_dev);
 	iucv_unregister(&smsg_handler, 1);
 	driver_unregister(&smsg_driver);
 }
@@ -166,12 +206,29 @@ static int __init smsg_init(void)
 	rc = iucv_path_connect(smsg_path, &smsg_handler, "*MSG    ",
 			       NULL, NULL, NULL);
 	if (rc)
-		goto out_free;
+		goto out_free_path;
+	smsg_dev = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!smsg_dev) {
+		rc = -ENOMEM;
+		goto out_free_path;
+	}
+	dev_set_name(smsg_dev, "smsg_iucv");
+	smsg_dev->bus = &iucv_bus;
+	smsg_dev->parent = iucv_root;
+	smsg_dev->release = (void (*)(struct device *))kfree;
+	smsg_dev->driver = &smsg_driver;
+	rc = device_register(smsg_dev);
+	if (rc)
+		goto out_free_dev;
+
 	cpcmd("SET SMSG IUCV", NULL, 0, NULL);
 	return 0;
 
-out_free:
+out_free_dev:
+	kfree(smsg_dev);
+out_free_path:
 	iucv_path_free(smsg_path);
+	smsg_path = NULL;
 out_register:
 	iucv_unregister(&smsg_handler, 1);
 out_driver:

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 33/38] pm: con3270 power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_3270.patch --]
[-- Type: text/plain, Size: 9164 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Introduce the power management callbacks to the 3270 driver. On suspend
the current 3270 view is deactivated and for non-console 3270 device
the release callback is done. This disconnects the current tty /
fullscreen application from the 3270 device. On resume the current
view is reactivated, on the tty you get a fresh login.
If the system panics before the 3270 device has been resumed, the ccw
device for the 3270 console is reactivated with ccw_device_force_console.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/con3270.c |   12 +++---
 drivers/s390/char/fs3270.c  |   16 +++++---
 drivers/s390/char/raw3270.c |   84 ++++++++++++++++++++++++++++++++++++++------
 drivers/s390/char/raw3270.h |   12 +++---
 4 files changed, 95 insertions(+), 29 deletions(-)

Index: linux-2.6/drivers/s390/char/raw3270.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/raw3270.c
+++ linux-2.6/drivers/s390/char/raw3270.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/raw3270.c
- *    IBM/3270 Driver - core functions.
+ * IBM/3270 Driver - core functions.
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <linux/bootmem.h>
@@ -61,6 +60,7 @@ struct raw3270 {
 #define RAW3270_FLAGS_ATTN	2	/* Device sent an ATTN interrupt */
 #define RAW3270_FLAGS_READY	4	/* Device is useable by views */
 #define RAW3270_FLAGS_CONSOLE	8	/* Device is the console. */
+#define RAW3270_FLAGS_FROZEN	16      /* set if 3270 is frozen for suspend */
 
 /* Semaphore to protect global data of raw3270 (devices, views, etc). */
 static DEFINE_MUTEX(raw3270_mutex);
@@ -306,7 +306,8 @@ raw3270_start(struct raw3270_view *view,
 
 	spin_lock_irqsave(get_ccwdev_lock(view->dev->cdev), flags);
 	rp = view->dev;
-	if (!rp || rp->view != view)
+	if (!rp || rp->view != view ||
+	    test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
 		rc = -EACCES;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
@@ -323,7 +324,8 @@ raw3270_start_locked(struct raw3270_view
 	int rc;
 
 	rp = view->dev;
-	if (!rp || rp->view != view)
+	if (!rp || rp->view != view ||
+	    test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
 		rc = -EACCES;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
@@ -764,7 +766,8 @@ raw3270_reset(struct raw3270_view *view)
 	int rc;
 
 	rp = view->dev;
-	if (!rp || rp->view != view)
+	if (!rp || rp->view != view ||
+	    test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
 		rc = -EACCES;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
@@ -922,6 +925,8 @@ raw3270_activate_view(struct raw3270_vie
 		rc = 0;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
+	else if (test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
+		rc = -EACCES;
 	else {
 		oldview = NULL;
 		if (rp->view) {
@@ -969,7 +974,8 @@ raw3270_deactivate_view(struct raw3270_v
 		list_del_init(&view->list);
 		list_add_tail(&view->list, &rp->view_list);
 		/* Try to activate another view. */
-		if (test_bit(RAW3270_FLAGS_READY, &rp->flags)) {
+		if (test_bit(RAW3270_FLAGS_READY, &rp->flags) &&
+		    !test_bit(RAW3270_FLAGS_FROZEN, &rp->flags)) {
 			list_for_each_entry(view, &rp->view_list, list) {
 				rp->view = view;
 				if (view->fn->activate(view) == 0)
@@ -1068,7 +1074,8 @@ raw3270_del_view(struct raw3270_view *vi
 		rp->view = NULL;
 	}
 	list_del_init(&view->list);
-	if (!rp->view && test_bit(RAW3270_FLAGS_READY, &rp->flags)) {
+	if (!rp->view && test_bit(RAW3270_FLAGS_READY, &rp->flags) &&
+	    !test_bit(RAW3270_FLAGS_FROZEN, &rp->flags)) {
 		/* Try to activate another view. */
 		list_for_each_entry(nv, &rp->view_list, list) {
 			if (nv->fn->activate(nv) == 0) {
@@ -1337,6 +1344,58 @@ raw3270_set_offline (struct ccw_device *
 	return 0;
 }
 
+static int raw3270_pm_stop(struct ccw_device *cdev)
+{
+	struct raw3270 *rp;
+	struct raw3270_view *view;
+	unsigned long flags;
+
+	rp = cdev->dev.driver_data;
+	if (!rp)
+		return 0;
+	spin_lock_irqsave(get_ccwdev_lock(rp->cdev), flags);
+	if (rp->view)
+		rp->view->fn->deactivate(rp->view);
+	if (!test_bit(RAW3270_FLAGS_CONSOLE, &rp->flags)) {
+		/*
+		 * Release tty and fullscreen for all non-console
+		 * devices.
+		 */
+		list_for_each_entry(view, &rp->view_list, list) {
+			if (view->fn->release)
+				view->fn->release(view);
+		}
+	}
+	set_bit(RAW3270_FLAGS_FROZEN, &rp->flags);
+	spin_unlock_irqrestore(get_ccwdev_lock(rp->cdev), flags);
+	return 0;
+}
+
+static int raw3270_pm_start(struct ccw_device *cdev)
+{
+	struct raw3270 *rp;
+	unsigned long flags;
+
+	rp = cdev->dev.driver_data;
+	if (!rp)
+		return 0;
+	spin_lock_irqsave(get_ccwdev_lock(rp->cdev), flags);
+	clear_bit(RAW3270_FLAGS_FROZEN, &rp->flags);
+	if (rp->view)
+		rp->view->fn->activate(rp->view);
+	spin_unlock_irqrestore(get_ccwdev_lock(rp->cdev), flags);
+	return 0;
+}
+
+void raw3270_pm_unfreeze(struct raw3270_view *view)
+{
+	struct raw3270 *rp;
+
+	rp = view->dev;
+	if (rp && test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
+		ccw_device_force_console();
+}
+
 static struct ccw_device_id raw3270_id[] = {
 	{ CCW_DEVICE(0x3270, 0) },
 	{ CCW_DEVICE(0x3271, 0) },
@@ -1360,6 +1419,9 @@ static struct ccw_driver raw3270_ccw_dri
 	.remove		= &raw3270_remove,
 	.set_online	= &raw3270_set_online,
 	.set_offline	= &raw3270_set_offline,
+	.freeze		= &raw3270_pm_stop,
+	.thaw		= &raw3270_pm_start,
+	.restore	= &raw3270_pm_start,
 };
 
 static int
Index: linux-2.6/drivers/s390/char/con3270.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/con3270.c
+++ linux-2.6/drivers/s390/char/con3270.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/con3270.c
- *    IBM/3270 Driver - console view.
+ * IBM/3270 Driver - console view.
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <linux/bootmem.h>
@@ -530,6 +529,7 @@ con3270_flush(void)
 	cp = condev;
 	if (!cp->view.dev)
 		return;
+	raw3270_pm_unfreeze(&cp->view);
 	spin_lock_irqsave(&cp->view.lock, flags);
 	con3270_wait_write(cp);
 	cp->nr_up = 0;
Index: linux-2.6/drivers/s390/char/fs3270.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/fs3270.c
+++ linux-2.6/drivers/s390/char/fs3270.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/fs3270.c
- *    IBM/3270 Driver - fullscreen driver.
+ * IBM/3270 Driver - fullscreen driver.
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5/2.6 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5/2.6 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <linux/bootmem.h>
@@ -399,6 +398,11 @@ fs3270_free_view(struct raw3270_view *vi
 static void
 fs3270_release(struct raw3270_view *view)
 {
+	struct fs3270 *fp;
+
+	fp = (struct fs3270 *) view;
+	if (fp->fs_pid)
+		kill_pid(fp->fs_pid, SIGHUP, 1);
 }
 
 /* View to a 3270 device. Can be console, tty or fullscreen. */
Index: linux-2.6/drivers/s390/char/raw3270.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/raw3270.h
+++ linux-2.6/drivers/s390/char/raw3270.h
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/raw3270.h
- *    IBM/3270 Driver
+ * IBM/3270 Driver
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <asm/idals.h>
@@ -195,6 +194,7 @@ void raw3270_wait_cons_dev(struct raw327
 /* Notifier for device addition/removal */
 int raw3270_register_notifier(void (*notifier)(int, int));
 void raw3270_unregister_notifier(void (*notifier)(int, int));
+void raw3270_pm_unfreeze(struct raw3270_view *);
 
 /*
  * Little memory allocator for string objects. 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 33/38] pm: con3270 power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_3270.patch --]
[-- Type: text/plain, Size: 9163 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Introduce the power management callbacks to the 3270 driver. On suspend
the current 3270 view is deactivated and for non-console 3270 device
the release callback is done. This disconnects the current tty /
fullscreen application from the 3270 device. On resume the current
view is reactivated, on the tty you get a fresh login.
If the system panics before the 3270 device has been resumed, the ccw
device for the 3270 console is reactivated with ccw_device_force_console.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/con3270.c |   12 +++---
 drivers/s390/char/fs3270.c  |   16 +++++---
 drivers/s390/char/raw3270.c |   84 ++++++++++++++++++++++++++++++++++++++------
 drivers/s390/char/raw3270.h |   12 +++---
 4 files changed, 95 insertions(+), 29 deletions(-)

Index: linux-2.6/drivers/s390/char/raw3270.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/raw3270.c
+++ linux-2.6/drivers/s390/char/raw3270.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/raw3270.c
- *    IBM/3270 Driver - core functions.
+ * IBM/3270 Driver - core functions.
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <linux/bootmem.h>
@@ -61,6 +60,7 @@ struct raw3270 {
 #define RAW3270_FLAGS_ATTN	2	/* Device sent an ATTN interrupt */
 #define RAW3270_FLAGS_READY	4	/* Device is useable by views */
 #define RAW3270_FLAGS_CONSOLE	8	/* Device is the console. */
+#define RAW3270_FLAGS_FROZEN	16      /* set if 3270 is frozen for suspend */
 
 /* Semaphore to protect global data of raw3270 (devices, views, etc). */
 static DEFINE_MUTEX(raw3270_mutex);
@@ -306,7 +306,8 @@ raw3270_start(struct raw3270_view *view,
 
 	spin_lock_irqsave(get_ccwdev_lock(view->dev->cdev), flags);
 	rp = view->dev;
-	if (!rp || rp->view != view)
+	if (!rp || rp->view != view ||
+	    test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
 		rc = -EACCES;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
@@ -323,7 +324,8 @@ raw3270_start_locked(struct raw3270_view
 	int rc;
 
 	rp = view->dev;
-	if (!rp || rp->view != view)
+	if (!rp || rp->view != view ||
+	    test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
 		rc = -EACCES;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
@@ -764,7 +766,8 @@ raw3270_reset(struct raw3270_view *view)
 	int rc;
 
 	rp = view->dev;
-	if (!rp || rp->view != view)
+	if (!rp || rp->view != view ||
+	    test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
 		rc = -EACCES;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
@@ -922,6 +925,8 @@ raw3270_activate_view(struct raw3270_vie
 		rc = 0;
 	else if (!test_bit(RAW3270_FLAGS_READY, &rp->flags))
 		rc = -ENODEV;
+	else if (test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
+		rc = -EACCES;
 	else {
 		oldview = NULL;
 		if (rp->view) {
@@ -969,7 +974,8 @@ raw3270_deactivate_view(struct raw3270_v
 		list_del_init(&view->list);
 		list_add_tail(&view->list, &rp->view_list);
 		/* Try to activate another view. */
-		if (test_bit(RAW3270_FLAGS_READY, &rp->flags)) {
+		if (test_bit(RAW3270_FLAGS_READY, &rp->flags) &&
+		    !test_bit(RAW3270_FLAGS_FROZEN, &rp->flags)) {
 			list_for_each_entry(view, &rp->view_list, list) {
 				rp->view = view;
 				if (view->fn->activate(view) == 0)
@@ -1068,7 +1074,8 @@ raw3270_del_view(struct raw3270_view *vi
 		rp->view = NULL;
 	}
 	list_del_init(&view->list);
-	if (!rp->view && test_bit(RAW3270_FLAGS_READY, &rp->flags)) {
+	if (!rp->view && test_bit(RAW3270_FLAGS_READY, &rp->flags) &&
+	    !test_bit(RAW3270_FLAGS_FROZEN, &rp->flags)) {
 		/* Try to activate another view. */
 		list_for_each_entry(nv, &rp->view_list, list) {
 			if (nv->fn->activate(nv) == 0) {
@@ -1337,6 +1344,58 @@ raw3270_set_offline (struct ccw_device *
 	return 0;
 }
 
+static int raw3270_pm_stop(struct ccw_device *cdev)
+{
+	struct raw3270 *rp;
+	struct raw3270_view *view;
+	unsigned long flags;
+
+	rp = cdev->dev.driver_data;
+	if (!rp)
+		return 0;
+	spin_lock_irqsave(get_ccwdev_lock(rp->cdev), flags);
+	if (rp->view)
+		rp->view->fn->deactivate(rp->view);
+	if (!test_bit(RAW3270_FLAGS_CONSOLE, &rp->flags)) {
+		/*
+		 * Release tty and fullscreen for all non-console
+		 * devices.
+		 */
+		list_for_each_entry(view, &rp->view_list, list) {
+			if (view->fn->release)
+				view->fn->release(view);
+		}
+	}
+	set_bit(RAW3270_FLAGS_FROZEN, &rp->flags);
+	spin_unlock_irqrestore(get_ccwdev_lock(rp->cdev), flags);
+	return 0;
+}
+
+static int raw3270_pm_start(struct ccw_device *cdev)
+{
+	struct raw3270 *rp;
+	unsigned long flags;
+
+	rp = cdev->dev.driver_data;
+	if (!rp)
+		return 0;
+	spin_lock_irqsave(get_ccwdev_lock(rp->cdev), flags);
+	clear_bit(RAW3270_FLAGS_FROZEN, &rp->flags);
+	if (rp->view)
+		rp->view->fn->activate(rp->view);
+	spin_unlock_irqrestore(get_ccwdev_lock(rp->cdev), flags);
+	return 0;
+}
+
+void raw3270_pm_unfreeze(struct raw3270_view *view)
+{
+	struct raw3270 *rp;
+
+	rp = view->dev;
+	if (rp && test_bit(RAW3270_FLAGS_FROZEN, &rp->flags))
+		ccw_device_force_console();
+}
+
 static struct ccw_device_id raw3270_id[] = {
 	{ CCW_DEVICE(0x3270, 0) },
 	{ CCW_DEVICE(0x3271, 0) },
@@ -1360,6 +1419,9 @@ static struct ccw_driver raw3270_ccw_dri
 	.remove		= &raw3270_remove,
 	.set_online	= &raw3270_set_online,
 	.set_offline	= &raw3270_set_offline,
+	.freeze		= &raw3270_pm_stop,
+	.thaw		= &raw3270_pm_start,
+	.restore	= &raw3270_pm_start,
 };
 
 static int
Index: linux-2.6/drivers/s390/char/con3270.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/con3270.c
+++ linux-2.6/drivers/s390/char/con3270.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/con3270.c
- *    IBM/3270 Driver - console view.
+ * IBM/3270 Driver - console view.
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <linux/bootmem.h>
@@ -530,6 +529,7 @@ con3270_flush(void)
 	cp = condev;
 	if (!cp->view.dev)
 		return;
+	raw3270_pm_unfreeze(&cp->view);
 	spin_lock_irqsave(&cp->view.lock, flags);
 	con3270_wait_write(cp);
 	cp->nr_up = 0;
Index: linux-2.6/drivers/s390/char/fs3270.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/fs3270.c
+++ linux-2.6/drivers/s390/char/fs3270.c
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/fs3270.c
- *    IBM/3270 Driver - fullscreen driver.
+ * IBM/3270 Driver - fullscreen driver.
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5/2.6 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5/2.6 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <linux/bootmem.h>
@@ -399,6 +398,11 @@ fs3270_free_view(struct raw3270_view *vi
 static void
 fs3270_release(struct raw3270_view *view)
 {
+	struct fs3270 *fp;
+
+	fp = (struct fs3270 *) view;
+	if (fp->fs_pid)
+		kill_pid(fp->fs_pid, SIGHUP, 1);
 }
 
 /* View to a 3270 device. Can be console, tty or fullscreen. */
Index: linux-2.6/drivers/s390/char/raw3270.h
===================================================================
--- linux-2.6.orig/drivers/s390/char/raw3270.h
+++ linux-2.6/drivers/s390/char/raw3270.h
@@ -1,11 +1,10 @@
 /*
- *  drivers/s390/char/raw3270.h
- *    IBM/3270 Driver
+ * IBM/3270 Driver
  *
- *  Author(s):
- *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
- *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
- *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s):
+ *   Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *   Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *     Copyright IBM Corp. 2003, 2009
  */
 
 #include <asm/idals.h>
@@ -195,6 +194,7 @@ void raw3270_wait_cons_dev(struct raw327
 /* Notifier for device addition/removal */
 int raw3270_register_notifier(void (*notifier)(int, int));
 void raw3270_unregister_notifier(void (*notifier)(int, int));
+void raw3270_pm_unfreeze(struct raw3270_view *);
 
 /*
  * Little memory allocator for string objects. 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 34/38] pm: memory hotplug power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_sclp_mem.patch --]
[-- Type: text/plain, Size: 3915 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/sclp_cmd.c |   42 ++++++++++++++++++++++++++++++++++++++----
 mm/Kconfig                   |    4 ++--
 2 files changed, 40 insertions(+), 6 deletions(-)

Index: linux-2.6/mm/Kconfig
===================================================================
--- linux-2.6.orig/mm/Kconfig
+++ linux-2.6/mm/Kconfig
@@ -128,11 +128,11 @@ config SPARSEMEM_VMEMMAP
 config MEMORY_HOTPLUG
 	bool "Allow for memory hot-add"
 	depends on SPARSEMEM || X86_64_ACPI_NUMA
-	depends on HOTPLUG && !HIBERNATION && ARCH_ENABLE_MEMORY_HOTPLUG
+	depends on HOTPLUG && !(HIBERNATION && !S390) && ARCH_ENABLE_MEMORY_HOTPLUG
 	depends on (IA64 || X86 || PPC64 || SUPERH || S390)
 
 comment "Memory hotplug is currently incompatible with Software Suspend"
-	depends on SPARSEMEM && HOTPLUG && HIBERNATION
+	depends on SPARSEMEM && HOTPLUG && HIBERNATION && !S390
 
 config MEMORY_HOTPLUG_SPARSE
 	def_bool y
Index: linux-2.6/drivers/s390/char/sclp_cmd.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_cmd.c
+++ linux-2.6/drivers/s390/char/sclp_cmd.c
@@ -1,9 +1,8 @@
 /*
- *  drivers/s390/char/sclp_cmd.c
+ * Copyright IBM Corp. 2007, 2009
  *
- *    Copyright IBM Corp. 2007
- *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>,
- *		 Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ * Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>,
+ *	      Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
  */
 
 #define KMSG_COMPONENT "sclp_cmd"
@@ -12,11 +11,13 @@
 #include <linux/completion.h>
 #include <linux/init.h>
 #include <linux/errno.h>
+#include <linux/err.h>
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/mm.h>
 #include <linux/mmzone.h>
 #include <linux/memory.h>
+#include <linux/platform_device.h>
 #include <asm/chpid.h>
 #include <asm/sclp.h>
 #include <asm/setup.h>
@@ -292,6 +293,7 @@ static DEFINE_MUTEX(sclp_mem_mutex);
 static LIST_HEAD(sclp_mem_list);
 static u8 sclp_max_storage_id;
 static unsigned long sclp_storage_ids[256 / BITS_PER_LONG];
+static int sclp_mem_state_changed;
 
 struct memory_increment {
 	struct list_head list;
@@ -450,6 +452,8 @@ static int sclp_mem_notifier(struct noti
 		rc = -EINVAL;
 		break;
 	}
+	if (!rc)
+		sclp_mem_state_changed = 1;
 	mutex_unlock(&sclp_mem_mutex);
 	return rc ? NOTIFY_BAD : NOTIFY_OK;
 }
@@ -525,6 +529,14 @@ static void __init insert_increment(u16 
 	list_add(&new_incr->list, prev);
 }
 
+static int sclp_mem_freeze(struct device *dev)
+{
+	if (!sclp_mem_state_changed)
+		return 0;
+	pr_err("Memory hotplug state changed, suspend refused.\n");
+	return -EPERM;
+}
+
 struct read_storage_sccb {
 	struct sccb_header header;
 	u16 max_id;
@@ -534,8 +546,20 @@ struct read_storage_sccb {
 	u32 entries[0];
 } __packed;
 
+static struct dev_pm_ops sclp_mem_pm_ops = {
+	.freeze		= sclp_mem_freeze,
+};
+
+static struct platform_driver sclp_mem_pdrv = {
+	.driver = {
+		.name	= "sclp_mem",
+		.pm	= &sclp_mem_pm_ops,
+	},
+};
+
 static int __init sclp_detect_standby_memory(void)
 {
+	struct platform_device *sclp_pdev;
 	struct read_storage_sccb *sccb;
 	int i, id, assigned, rc;
 
@@ -588,7 +612,17 @@ static int __init sclp_detect_standby_me
 	rc = register_memory_notifier(&sclp_mem_nb);
 	if (rc)
 		goto out;
+	rc = platform_driver_register(&sclp_mem_pdrv);
+	if (rc)
+		goto out;
+	sclp_pdev = platform_device_register_simple("sclp_mem", -1, NULL, 0);
+	rc = IS_ERR(sclp_pdev) ? PTR_ERR(sclp_pdev) : 0;
+	if (rc)
+		goto out_driver;
 	sclp_add_standby_memory();
+	goto out;
+out_driver:
+	platform_driver_unregister(&sclp_mem_pdrv);
 out:
 	free_page((unsigned long) sccb);
 	return rc;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 34/38] pm: memory hotplug power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_sclp_mem.patch --]
[-- Type: text/plain, Size: 3914 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/sclp_cmd.c |   42 ++++++++++++++++++++++++++++++++++++++----
 mm/Kconfig                   |    4 ++--
 2 files changed, 40 insertions(+), 6 deletions(-)

Index: linux-2.6/mm/Kconfig
===================================================================
--- linux-2.6.orig/mm/Kconfig
+++ linux-2.6/mm/Kconfig
@@ -128,11 +128,11 @@ config SPARSEMEM_VMEMMAP
 config MEMORY_HOTPLUG
 	bool "Allow for memory hot-add"
 	depends on SPARSEMEM || X86_64_ACPI_NUMA
-	depends on HOTPLUG && !HIBERNATION && ARCH_ENABLE_MEMORY_HOTPLUG
+	depends on HOTPLUG && !(HIBERNATION && !S390) && ARCH_ENABLE_MEMORY_HOTPLUG
 	depends on (IA64 || X86 || PPC64 || SUPERH || S390)
 
 comment "Memory hotplug is currently incompatible with Software Suspend"
-	depends on SPARSEMEM && HOTPLUG && HIBERNATION
+	depends on SPARSEMEM && HOTPLUG && HIBERNATION && !S390
 
 config MEMORY_HOTPLUG_SPARSE
 	def_bool y
Index: linux-2.6/drivers/s390/char/sclp_cmd.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/sclp_cmd.c
+++ linux-2.6/drivers/s390/char/sclp_cmd.c
@@ -1,9 +1,8 @@
 /*
- *  drivers/s390/char/sclp_cmd.c
+ * Copyright IBM Corp. 2007, 2009
  *
- *    Copyright IBM Corp. 2007
- *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>,
- *		 Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ * Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>,
+ *	      Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
  */
 
 #define KMSG_COMPONENT "sclp_cmd"
@@ -12,11 +11,13 @@
 #include <linux/completion.h>
 #include <linux/init.h>
 #include <linux/errno.h>
+#include <linux/err.h>
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/mm.h>
 #include <linux/mmzone.h>
 #include <linux/memory.h>
+#include <linux/platform_device.h>
 #include <asm/chpid.h>
 #include <asm/sclp.h>
 #include <asm/setup.h>
@@ -292,6 +293,7 @@ static DEFINE_MUTEX(sclp_mem_mutex);
 static LIST_HEAD(sclp_mem_list);
 static u8 sclp_max_storage_id;
 static unsigned long sclp_storage_ids[256 / BITS_PER_LONG];
+static int sclp_mem_state_changed;
 
 struct memory_increment {
 	struct list_head list;
@@ -450,6 +452,8 @@ static int sclp_mem_notifier(struct noti
 		rc = -EINVAL;
 		break;
 	}
+	if (!rc)
+		sclp_mem_state_changed = 1;
 	mutex_unlock(&sclp_mem_mutex);
 	return rc ? NOTIFY_BAD : NOTIFY_OK;
 }
@@ -525,6 +529,14 @@ static void __init insert_increment(u16 
 	list_add(&new_incr->list, prev);
 }
 
+static int sclp_mem_freeze(struct device *dev)
+{
+	if (!sclp_mem_state_changed)
+		return 0;
+	pr_err("Memory hotplug state changed, suspend refused.\n");
+	return -EPERM;
+}
+
 struct read_storage_sccb {
 	struct sccb_header header;
 	u16 max_id;
@@ -534,8 +546,20 @@ struct read_storage_sccb {
 	u32 entries[0];
 } __packed;
 
+static struct dev_pm_ops sclp_mem_pm_ops = {
+	.freeze		= sclp_mem_freeze,
+};
+
+static struct platform_driver sclp_mem_pdrv = {
+	.driver = {
+		.name	= "sclp_mem",
+		.pm	= &sclp_mem_pm_ops,
+	},
+};
+
 static int __init sclp_detect_standby_memory(void)
 {
+	struct platform_device *sclp_pdev;
 	struct read_storage_sccb *sccb;
 	int i, id, assigned, rc;
 
@@ -588,7 +612,17 @@ static int __init sclp_detect_standby_me
 	rc = register_memory_notifier(&sclp_mem_nb);
 	if (rc)
 		goto out;
+	rc = platform_driver_register(&sclp_mem_pdrv);
+	if (rc)
+		goto out;
+	sclp_pdev = platform_device_register_simple("sclp_mem", -1, NULL, 0);
+	rc = IS_ERR(sclp_pdev) ? PTR_ERR(sclp_pdev) : 0;
+	if (rc)
+		goto out_driver;
 	sclp_add_standby_memory();
+	goto out;
+out_driver:
+	platform_driver_unregister(&sclp_mem_pdrv);
 out:
 	free_page((unsigned long) sccb);
 	return rc;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 35/38] pm: monwriter power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_monwriter.patch --]
[-- Type: text/plain, Size: 4069 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/monwriter.c |   98 +++++++++++++++++++++++++++++++++++++++---
 1 file changed, 92 insertions(+), 6 deletions(-)

Index: linux-2.6/drivers/s390/char/monwriter.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/monwriter.c
+++ linux-2.6/drivers/s390/char/monwriter.c
@@ -1,9 +1,7 @@
 /*
- * drivers/s390/char/monwriter.c
- *
  * Character device driver for writing z/VM *MONITOR service records.
  *
- * Copyright (C) IBM Corp. 2006
+ * Copyright IBM Corp. 2006, 2009
  *
  * Author(s): Melissa Howland <Melissa.Howland@us.ibm.com>
  */
@@ -22,6 +20,7 @@
 #include <linux/ctype.h>
 #include <linux/poll.h>
 #include <linux/mutex.h>
+#include <linux/platform_device.h>
 #include <asm/uaccess.h>
 #include <asm/ebcdic.h>
 #include <asm/io.h>
@@ -40,7 +39,10 @@ struct mon_buf {
 	char *data;
 };
 
+static LIST_HEAD(mon_priv_list);
+
 struct mon_private {
+	struct list_head priv_list;
 	struct list_head list;
 	struct monwrite_hdr hdr;
 	size_t hdr_to_read;
@@ -188,6 +190,7 @@ static int monwrite_open(struct inode *i
 	monpriv->hdr_to_read = sizeof(monpriv->hdr);
 	mutex_init(&monpriv->thread_mutex);
 	filp->private_data = monpriv;
+	list_add_tail(&monpriv->priv_list, &mon_priv_list);
 	unlock_kernel();
 	return nonseekable_open(inode, filp);
 }
@@ -206,6 +209,7 @@ static int monwrite_close(struct inode *
 		kfree(entry->data);
 		kfree(entry);
 	}
+	list_del(&monpriv->priv_list);
 	kfree(monpriv);
 	return 0;
 }
@@ -281,20 +285,102 @@ static struct miscdevice mon_dev = {
 };
 
 /*
+ * suspend/resume
+ */
+
+static int monwriter_freeze(struct device *dev)
+{
+	struct mon_private *monpriv;
+	struct mon_buf *monbuf;
+
+	list_for_each_entry(monpriv, &mon_priv_list, priv_list) {
+		list_for_each_entry(monbuf, &monpriv->list, list) {
+			if (monbuf->hdr.mon_function != MONWRITE_GEN_EVENT)
+				monwrite_diag(&monbuf->hdr, monbuf->data,
+					      APPLDATA_STOP_REC);
+		}
+	}
+	return 0;
+}
+
+static int monwriter_restore(struct device *dev)
+{
+	struct mon_private *monpriv;
+	struct mon_buf *monbuf;
+
+	list_for_each_entry(monpriv, &mon_priv_list, priv_list) {
+		list_for_each_entry(monbuf, &monpriv->list, list) {
+			if (monbuf->hdr.mon_function == MONWRITE_START_INTERVAL)
+				monwrite_diag(&monbuf->hdr, monbuf->data,
+					      APPLDATA_START_INTERVAL_REC);
+			if (monbuf->hdr.mon_function == MONWRITE_START_CONFIG)
+				monwrite_diag(&monbuf->hdr, monbuf->data,
+					      APPLDATA_START_CONFIG_REC);
+		}
+	}
+	return 0;
+}
+
+static int monwriter_thaw(struct device *dev)
+{
+	return monwriter_restore(dev);
+}
+
+static struct dev_pm_ops monwriter_pm_ops = {
+	.freeze		= monwriter_freeze,
+	.thaw		= monwriter_thaw,
+	.restore	= monwriter_restore,
+};
+
+static struct platform_driver monwriter_pdrv = {
+	.driver = {
+		.name	= "monwriter",
+		.owner	= THIS_MODULE,
+		.pm	= &monwriter_pm_ops,
+	},
+};
+
+static struct platform_device *monwriter_pdev;
+
+/*
  * module init/exit
  */
 
 static int __init mon_init(void)
 {
-	if (MACHINE_IS_VM)
-		return misc_register(&mon_dev);
-	else
+	int rc;
+
+	if (!MACHINE_IS_VM)
 		return -ENODEV;
+
+	rc = platform_driver_register(&monwriter_pdrv);
+	if (rc)
+		return rc;
+
+	monwriter_pdev = platform_device_register_simple("monwriter", -1, NULL,
+							0);
+	if (IS_ERR(monwriter_pdev)) {
+		rc = PTR_ERR(monwriter_pdev);
+		goto out_driver;
+	}
+
+	rc = misc_register(&mon_dev);
+	if (rc)
+		goto out_device;
+	return 0;
+
+out_device:
+	platform_device_unregister(monwriter_pdev);
+out_driver:
+	platform_driver_unregister(&monwriter_pdrv);
+	return rc;
 }
 
 static void __exit mon_exit(void)
 {
 	WARN_ON(misc_deregister(&mon_dev) != 0);
+	platform_device_unregister(monwriter_pdev);
+	platform_driver_unregister(&monwriter_pdrv);
 }
 
 module_init(mon_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 35/38] pm: monwriter power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_monwriter.patch --]
[-- Type: text/plain, Size: 4068 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/monwriter.c |   98 +++++++++++++++++++++++++++++++++++++++---
 1 file changed, 92 insertions(+), 6 deletions(-)

Index: linux-2.6/drivers/s390/char/monwriter.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/monwriter.c
+++ linux-2.6/drivers/s390/char/monwriter.c
@@ -1,9 +1,7 @@
 /*
- * drivers/s390/char/monwriter.c
- *
  * Character device driver for writing z/VM *MONITOR service records.
  *
- * Copyright (C) IBM Corp. 2006
+ * Copyright IBM Corp. 2006, 2009
  *
  * Author(s): Melissa Howland <Melissa.Howland@us.ibm.com>
  */
@@ -22,6 +20,7 @@
 #include <linux/ctype.h>
 #include <linux/poll.h>
 #include <linux/mutex.h>
+#include <linux/platform_device.h>
 #include <asm/uaccess.h>
 #include <asm/ebcdic.h>
 #include <asm/io.h>
@@ -40,7 +39,10 @@ struct mon_buf {
 	char *data;
 };
 
+static LIST_HEAD(mon_priv_list);
+
 struct mon_private {
+	struct list_head priv_list;
 	struct list_head list;
 	struct monwrite_hdr hdr;
 	size_t hdr_to_read;
@@ -188,6 +190,7 @@ static int monwrite_open(struct inode *i
 	monpriv->hdr_to_read = sizeof(monpriv->hdr);
 	mutex_init(&monpriv->thread_mutex);
 	filp->private_data = monpriv;
+	list_add_tail(&monpriv->priv_list, &mon_priv_list);
 	unlock_kernel();
 	return nonseekable_open(inode, filp);
 }
@@ -206,6 +209,7 @@ static int monwrite_close(struct inode *
 		kfree(entry->data);
 		kfree(entry);
 	}
+	list_del(&monpriv->priv_list);
 	kfree(monpriv);
 	return 0;
 }
@@ -281,20 +285,102 @@ static struct miscdevice mon_dev = {
 };
 
 /*
+ * suspend/resume
+ */
+
+static int monwriter_freeze(struct device *dev)
+{
+	struct mon_private *monpriv;
+	struct mon_buf *monbuf;
+
+	list_for_each_entry(monpriv, &mon_priv_list, priv_list) {
+		list_for_each_entry(monbuf, &monpriv->list, list) {
+			if (monbuf->hdr.mon_function != MONWRITE_GEN_EVENT)
+				monwrite_diag(&monbuf->hdr, monbuf->data,
+					      APPLDATA_STOP_REC);
+		}
+	}
+	return 0;
+}
+
+static int monwriter_restore(struct device *dev)
+{
+	struct mon_private *monpriv;
+	struct mon_buf *monbuf;
+
+	list_for_each_entry(monpriv, &mon_priv_list, priv_list) {
+		list_for_each_entry(monbuf, &monpriv->list, list) {
+			if (monbuf->hdr.mon_function == MONWRITE_START_INTERVAL)
+				monwrite_diag(&monbuf->hdr, monbuf->data,
+					      APPLDATA_START_INTERVAL_REC);
+			if (monbuf->hdr.mon_function == MONWRITE_START_CONFIG)
+				monwrite_diag(&monbuf->hdr, monbuf->data,
+					      APPLDATA_START_CONFIG_REC);
+		}
+	}
+	return 0;
+}
+
+static int monwriter_thaw(struct device *dev)
+{
+	return monwriter_restore(dev);
+}
+
+static struct dev_pm_ops monwriter_pm_ops = {
+	.freeze		= monwriter_freeze,
+	.thaw		= monwriter_thaw,
+	.restore	= monwriter_restore,
+};
+
+static struct platform_driver monwriter_pdrv = {
+	.driver = {
+		.name	= "monwriter",
+		.owner	= THIS_MODULE,
+		.pm	= &monwriter_pm_ops,
+	},
+};
+
+static struct platform_device *monwriter_pdev;
+
+/*
  * module init/exit
  */
 
 static int __init mon_init(void)
 {
-	if (MACHINE_IS_VM)
-		return misc_register(&mon_dev);
-	else
+	int rc;
+
+	if (!MACHINE_IS_VM)
 		return -ENODEV;
+
+	rc = platform_driver_register(&monwriter_pdrv);
+	if (rc)
+		return rc;
+
+	monwriter_pdev = platform_device_register_simple("monwriter", -1, NULL,
+							0);
+	if (IS_ERR(monwriter_pdev)) {
+		rc = PTR_ERR(monwriter_pdev);
+		goto out_driver;
+	}
+
+	rc = misc_register(&mon_dev);
+	if (rc)
+		goto out_device;
+	return 0;
+
+out_device:
+	platform_device_unregister(monwriter_pdev);
+out_driver:
+	platform_driver_unregister(&monwriter_pdrv);
+	return rc;
 }
 
 static void __exit mon_exit(void)
 {
 	WARN_ON(misc_deregister(&mon_dev) != 0);
+	platform_device_unregister(monwriter_pdev);
+	platform_driver_unregister(&monwriter_pdrv);
 }
 
 module_init(mon_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 36/38] pm: monreader power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_monreader.patch --]
[-- Type: text/plain, Size: 6876 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/monreader.c |  140 ++++++++++++++++++++++++++++++++++++++----
 1 file changed, 128 insertions(+), 12 deletions(-)

Index: linux-2.6/drivers/s390/char/monreader.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/monreader.c
+++ linux-2.6/drivers/s390/char/monreader.c
@@ -1,10 +1,9 @@
 /*
- * drivers/s390/char/monreader.c
- *
  * Character device driver for reading z/VM *MONITOR service records.
  *
- *   Copyright IBM Corp. 2004, 2008
- *   Author: Gerald Schaefer <gerald.schaefer@de.ibm.com>
+ * Copyright IBM Corp. 2004, 2009
+ *
+ * Author: Gerald Schaefer <gerald.schaefer@de.ibm.com>
  */
 
 #define KMSG_COMPONENT "monreader"
@@ -22,6 +21,7 @@
 #include <linux/spinlock.h>
 #include <linux/interrupt.h>
 #include <linux/poll.h>
+#include <linux/device.h>
 #include <net/iucv/iucv.h>
 #include <asm/uaccess.h>
 #include <asm/ebcdic.h>
@@ -78,6 +78,7 @@ static u8 user_data_sever[16] = {
 	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
 };
 
+static struct device *monreader_device;
 
 /******************************************************************************
  *                             helper functions                               *
@@ -319,11 +320,12 @@ static int mon_open(struct inode *inode,
 		goto out_path;
 	}
 	filp->private_data = monpriv;
+	monreader_device->driver_data = monpriv;
 	unlock_kernel();
 	return nonseekable_open(inode, filp);
 
 out_path:
-	kfree(monpriv->path);
+	iucv_path_free(monpriv->path);
 out_priv:
 	mon_free_mem(monpriv);
 out_use:
@@ -341,10 +343,13 @@ static int mon_close(struct inode *inode
 	/*
 	 * Close IUCV connection and unregister
 	 */
-	rc = iucv_path_sever(monpriv->path, user_data_sever);
-	if (rc)
-		pr_warning("Disconnecting the z/VM *MONITOR system service "
-			   "failed with rc=%i\n", rc);
+	if (monpriv->path) {
+		rc = iucv_path_sever(monpriv->path, user_data_sever);
+		if (rc)
+			pr_warning("Disconnecting the z/VM *MONITOR system "
+				   "service failed with rc=%i\n", rc);
+		iucv_path_free(monpriv->path);
+	}
 
 	atomic_set(&monpriv->iucv_severed, 0);
 	atomic_set(&monpriv->iucv_connected, 0);
@@ -452,6 +457,94 @@ static struct miscdevice mon_dev = {
 	.minor      = MISC_DYNAMIC_MINOR,
 };
 
+
+/******************************************************************************
+ *                              suspend / resume                              *
+ *****************************************************************************/
+static int monreader_freeze(struct device *dev)
+{
+	struct mon_private *monpriv = dev->driver_data;
+	int rc;
+
+	if (!monpriv)
+		return 0;
+	if (monpriv->path) {
+		rc = iucv_path_sever(monpriv->path, user_data_sever);
+		if (rc)
+			pr_warning("Disconnecting the z/VM *MONITOR system "
+				   "service failed with rc=%i\n", rc);
+		iucv_path_free(monpriv->path);
+	}
+	atomic_set(&monpriv->iucv_severed, 0);
+	atomic_set(&monpriv->iucv_connected, 0);
+	atomic_set(&monpriv->read_ready, 0);
+	atomic_set(&monpriv->msglim_count, 0);
+	monpriv->write_index  = 0;
+	monpriv->read_index   = 0;
+	monpriv->path = NULL;
+	return 0;
+}
+
+static int monreader_thaw(struct device *dev)
+{
+	struct mon_private *monpriv = dev->driver_data;
+	int rc;
+
+	if (!monpriv)
+		return 0;
+	rc = -ENOMEM;
+	monpriv->path = iucv_path_alloc(MON_MSGLIM, IUCV_IPRMDATA, GFP_KERNEL);
+	if (!monpriv->path)
+		goto out;
+	rc = iucv_path_connect(monpriv->path, &monreader_iucv_handler,
+			       MON_SERVICE, NULL, user_data_connect, monpriv);
+	if (rc) {
+		pr_err("Connecting to the z/VM *MONITOR system service "
+		       "failed with rc=%i\n", rc);
+		goto out_path;
+	}
+	wait_event(mon_conn_wait_queue,
+		   atomic_read(&monpriv->iucv_connected) ||
+		   atomic_read(&monpriv->iucv_severed));
+	if (atomic_read(&monpriv->iucv_severed))
+		goto out_path;
+	return 0;
+out_path:
+	rc = -EIO;
+	iucv_path_free(monpriv->path);
+	monpriv->path = NULL;
+out:
+	atomic_set(&monpriv->iucv_severed, 1);
+	return rc;
+}
+
+static int monreader_restore(struct device *dev)
+{
+	int rc;
+
+	segment_unload(mon_dcss_name);
+	rc = segment_load(mon_dcss_name, SEGMENT_SHARED,
+			  &mon_dcss_start, &mon_dcss_end);
+	if (rc < 0) {
+		segment_warning(rc, mon_dcss_name);
+		panic("fatal monreader resume error: no monitor dcss\n");
+	}
+	return monreader_thaw(dev);
+}
+
+static struct dev_pm_ops monreader_pm_ops = {
+	.freeze  = monreader_freeze,
+	.thaw    = monreader_thaw,
+	.restore = monreader_restore,
+};
+
+static struct device_driver monreader_driver = {
+	.name = "monreader",
+	.bus  = &iucv_bus,
+	.pm   = &monreader_pm_ops,
+};
+
+
 /******************************************************************************
  *                              module init/exit                              *
  *****************************************************************************/
@@ -475,16 +568,33 @@ static int __init mon_init(void)
 		return rc;
 	}
 
+	rc = driver_register(&monreader_driver);
+	if (rc)
+		goto out_iucv;
+	monreader_device = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!monreader_device)
+		goto out_driver;
+	dev_set_name(monreader_device, "monreader-dev");
+	monreader_device->bus = &iucv_bus;
+	monreader_device->parent = iucv_root;
+	monreader_device->driver = &monreader_driver;
+	monreader_device->release = (void (*)(struct device *))kfree;
+	rc = device_register(monreader_device);
+	if (rc) {
+		kfree(monreader_device);
+		goto out_driver;
+	}
+
 	rc = segment_type(mon_dcss_name);
 	if (rc < 0) {
 		segment_warning(rc, mon_dcss_name);
-		goto out_iucv;
+		goto out_device;
 	}
 	if (rc != SEG_TYPE_SC) {
 		pr_err("The specified *MONITOR DCSS %s does not have the "
 		       "required type SC\n", mon_dcss_name);
 		rc = -EINVAL;
-		goto out_iucv;
+		goto out_device;
 	}
 
 	rc = segment_load(mon_dcss_name, SEGMENT_SHARED,
@@ -492,7 +602,7 @@ static int __init mon_init(void)
 	if (rc < 0) {
 		segment_warning(rc, mon_dcss_name);
 		rc = -EINVAL;
-		goto out_iucv;
+		goto out_device;
 	}
 	dcss_mkname(mon_dcss_name, &user_data_connect[8]);
 
@@ -503,6 +613,10 @@ static int __init mon_init(void)
 
 out:
 	segment_unload(mon_dcss_name);
+out_device:
+	device_unregister(monreader_device);
+out_driver:
+	driver_unregister(&monreader_driver);
 out_iucv:
 	iucv_unregister(&monreader_iucv_handler, 1);
 	return rc;
@@ -512,6 +626,8 @@ static void __exit mon_exit(void)
 {
 	segment_unload(mon_dcss_name);
 	WARN_ON(misc_deregister(&mon_dev) != 0);
+	device_unregister(monreader_device);
+	driver_unregister(&monreader_driver);
 	iucv_unregister(&monreader_iucv_handler, 1);
 	return;
 }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 36/38] pm: monreader power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_monreader.patch --]
[-- Type: text/plain, Size: 6875 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/char/monreader.c |  140 ++++++++++++++++++++++++++++++++++++++----
 1 file changed, 128 insertions(+), 12 deletions(-)

Index: linux-2.6/drivers/s390/char/monreader.c
===================================================================
--- linux-2.6.orig/drivers/s390/char/monreader.c
+++ linux-2.6/drivers/s390/char/monreader.c
@@ -1,10 +1,9 @@
 /*
- * drivers/s390/char/monreader.c
- *
  * Character device driver for reading z/VM *MONITOR service records.
  *
- *   Copyright IBM Corp. 2004, 2008
- *   Author: Gerald Schaefer <gerald.schaefer@de.ibm.com>
+ * Copyright IBM Corp. 2004, 2009
+ *
+ * Author: Gerald Schaefer <gerald.schaefer@de.ibm.com>
  */
 
 #define KMSG_COMPONENT "monreader"
@@ -22,6 +21,7 @@
 #include <linux/spinlock.h>
 #include <linux/interrupt.h>
 #include <linux/poll.h>
+#include <linux/device.h>
 #include <net/iucv/iucv.h>
 #include <asm/uaccess.h>
 #include <asm/ebcdic.h>
@@ -78,6 +78,7 @@ static u8 user_data_sever[16] = {
 	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
 };
 
+static struct device *monreader_device;
 
 /******************************************************************************
  *                             helper functions                               *
@@ -319,11 +320,12 @@ static int mon_open(struct inode *inode,
 		goto out_path;
 	}
 	filp->private_data = monpriv;
+	monreader_device->driver_data = monpriv;
 	unlock_kernel();
 	return nonseekable_open(inode, filp);
 
 out_path:
-	kfree(monpriv->path);
+	iucv_path_free(monpriv->path);
 out_priv:
 	mon_free_mem(monpriv);
 out_use:
@@ -341,10 +343,13 @@ static int mon_close(struct inode *inode
 	/*
 	 * Close IUCV connection and unregister
 	 */
-	rc = iucv_path_sever(monpriv->path, user_data_sever);
-	if (rc)
-		pr_warning("Disconnecting the z/VM *MONITOR system service "
-			   "failed with rc=%i\n", rc);
+	if (monpriv->path) {
+		rc = iucv_path_sever(monpriv->path, user_data_sever);
+		if (rc)
+			pr_warning("Disconnecting the z/VM *MONITOR system "
+				   "service failed with rc=%i\n", rc);
+		iucv_path_free(monpriv->path);
+	}
 
 	atomic_set(&monpriv->iucv_severed, 0);
 	atomic_set(&monpriv->iucv_connected, 0);
@@ -452,6 +457,94 @@ static struct miscdevice mon_dev = {
 	.minor      = MISC_DYNAMIC_MINOR,
 };
 
+
+/******************************************************************************
+ *                              suspend / resume                              *
+ *****************************************************************************/
+static int monreader_freeze(struct device *dev)
+{
+	struct mon_private *monpriv = dev->driver_data;
+	int rc;
+
+	if (!monpriv)
+		return 0;
+	if (monpriv->path) {
+		rc = iucv_path_sever(monpriv->path, user_data_sever);
+		if (rc)
+			pr_warning("Disconnecting the z/VM *MONITOR system "
+				   "service failed with rc=%i\n", rc);
+		iucv_path_free(monpriv->path);
+	}
+	atomic_set(&monpriv->iucv_severed, 0);
+	atomic_set(&monpriv->iucv_connected, 0);
+	atomic_set(&monpriv->read_ready, 0);
+	atomic_set(&monpriv->msglim_count, 0);
+	monpriv->write_index  = 0;
+	monpriv->read_index   = 0;
+	monpriv->path = NULL;
+	return 0;
+}
+
+static int monreader_thaw(struct device *dev)
+{
+	struct mon_private *monpriv = dev->driver_data;
+	int rc;
+
+	if (!monpriv)
+		return 0;
+	rc = -ENOMEM;
+	monpriv->path = iucv_path_alloc(MON_MSGLIM, IUCV_IPRMDATA, GFP_KERNEL);
+	if (!monpriv->path)
+		goto out;
+	rc = iucv_path_connect(monpriv->path, &monreader_iucv_handler,
+			       MON_SERVICE, NULL, user_data_connect, monpriv);
+	if (rc) {
+		pr_err("Connecting to the z/VM *MONITOR system service "
+		       "failed with rc=%i\n", rc);
+		goto out_path;
+	}
+	wait_event(mon_conn_wait_queue,
+		   atomic_read(&monpriv->iucv_connected) ||
+		   atomic_read(&monpriv->iucv_severed));
+	if (atomic_read(&monpriv->iucv_severed))
+		goto out_path;
+	return 0;
+out_path:
+	rc = -EIO;
+	iucv_path_free(monpriv->path);
+	monpriv->path = NULL;
+out:
+	atomic_set(&monpriv->iucv_severed, 1);
+	return rc;
+}
+
+static int monreader_restore(struct device *dev)
+{
+	int rc;
+
+	segment_unload(mon_dcss_name);
+	rc = segment_load(mon_dcss_name, SEGMENT_SHARED,
+			  &mon_dcss_start, &mon_dcss_end);
+	if (rc < 0) {
+		segment_warning(rc, mon_dcss_name);
+		panic("fatal monreader resume error: no monitor dcss\n");
+	}
+	return monreader_thaw(dev);
+}
+
+static struct dev_pm_ops monreader_pm_ops = {
+	.freeze  = monreader_freeze,
+	.thaw    = monreader_thaw,
+	.restore = monreader_restore,
+};
+
+static struct device_driver monreader_driver = {
+	.name = "monreader",
+	.bus  = &iucv_bus,
+	.pm   = &monreader_pm_ops,
+};
+
+
 /******************************************************************************
  *                              module init/exit                              *
  *****************************************************************************/
@@ -475,16 +568,33 @@ static int __init mon_init(void)
 		return rc;
 	}
 
+	rc = driver_register(&monreader_driver);
+	if (rc)
+		goto out_iucv;
+	monreader_device = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!monreader_device)
+		goto out_driver;
+	dev_set_name(monreader_device, "monreader-dev");
+	monreader_device->bus = &iucv_bus;
+	monreader_device->parent = iucv_root;
+	monreader_device->driver = &monreader_driver;
+	monreader_device->release = (void (*)(struct device *))kfree;
+	rc = device_register(monreader_device);
+	if (rc) {
+		kfree(monreader_device);
+		goto out_driver;
+	}
+
 	rc = segment_type(mon_dcss_name);
 	if (rc < 0) {
 		segment_warning(rc, mon_dcss_name);
-		goto out_iucv;
+		goto out_device;
 	}
 	if (rc != SEG_TYPE_SC) {
 		pr_err("The specified *MONITOR DCSS %s does not have the "
 		       "required type SC\n", mon_dcss_name);
 		rc = -EINVAL;
-		goto out_iucv;
+		goto out_device;
 	}
 
 	rc = segment_load(mon_dcss_name, SEGMENT_SHARED,
@@ -492,7 +602,7 @@ static int __init mon_init(void)
 	if (rc < 0) {
 		segment_warning(rc, mon_dcss_name);
 		rc = -EINVAL;
-		goto out_iucv;
+		goto out_device;
 	}
 	dcss_mkname(mon_dcss_name, &user_data_connect[8]);
 
@@ -503,6 +613,10 @@ static int __init mon_init(void)
 
 out:
 	segment_unload(mon_dcss_name);
+out_device:
+	device_unregister(monreader_device);
+out_driver:
+	driver_unregister(&monreader_driver);
 out_iucv:
 	iucv_unregister(&monreader_iucv_handler, 1);
 	return rc;
@@ -512,6 +626,8 @@ static void __exit mon_exit(void)
 {
 	segment_unload(mon_dcss_name);
 	WARN_ON(misc_deregister(&mon_dev) != 0);
+	device_unregister(monreader_device);
+	driver_unregister(&monreader_driver);
 	iucv_unregister(&monreader_iucv_handler, 1);
 	return;
 }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 37/38] pm: dcssblk power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_dcssblk.patch --]
[-- Type: text/plain, Size: 4477 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/dcssblk.c |  132 +++++++++++++++++++++++++++++++++++++------
 1 file changed, 115 insertions(+), 17 deletions(-)

Index: linux-2.6/drivers/s390/block/dcssblk.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dcssblk.c
+++ linux-2.6/drivers/s390/block/dcssblk.c
@@ -14,10 +14,11 @@
 #include <linux/init.h>
 #include <linux/slab.h>
 #include <linux/blkdev.h>
-#include <asm/extmem.h>
-#include <asm/io.h>
 #include <linux/completion.h>
 #include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <asm/extmem.h>
+#include <asm/io.h>
 
 #define DCSSBLK_NAME "dcssblk"
 #define DCSSBLK_MINORS_PER_DISK 1
@@ -940,11 +941,94 @@ dcssblk_check_params(void)
 }
 
 /*
+ * Suspend / Resume
+ */
+static int dcssblk_freeze(struct device *dev)
+{
+	struct dcssblk_dev_info *dev_info;
+	int rc = 0;
+
+	list_for_each_entry(dev_info, &dcssblk_devices, lh) {
+		switch (dev_info->segment_type) {
+			case SEG_TYPE_SR:
+			case SEG_TYPE_ER:
+			case SEG_TYPE_SC:
+				if (!dev_info->is_shared)
+					rc = -EINVAL;
+				break;
+			default:
+				rc = -EINVAL;
+				break;
+		}
+		if (rc)
+			break;
+	}
+	if (rc)
+		pr_err("Suspend failed because device %s is writeable.\n",
+		       dev_info->segment_name);
+	return rc;
+}
+
+static int dcssblk_restore(struct device *dev)
+{
+	struct dcssblk_dev_info *dev_info;
+	struct segment_info *entry;
+	unsigned long start, end;
+	int rc = 0;
+
+	list_for_each_entry(dev_info, &dcssblk_devices, lh) {
+		list_for_each_entry(entry, &dev_info->seg_list, lh) {
+			segment_unload(entry->segment_name);
+			rc = segment_load(entry->segment_name, SEGMENT_SHARED,
+					  &start, &end);
+			if (rc < 0) {
+// TODO in_use check ?
+				segment_warning(rc, entry->segment_name);
+				goto out_panic;
+			}
+			if (start != entry->start || end != entry->end) {
+				pr_err("Mismatch of start / end address after "
+				       "resuming device %s\n",
+				       entry->segment_name);
+				goto out_panic;
+			}
+		}
+	}
+	return 0;
+out_panic:
+	panic("fatal dcssblk resume error\n");
+}
+
+static int dcssblk_thaw(struct device *dev)
+{
+	return 0;
+}
+
+static struct dev_pm_ops dcssblk_pm_ops = {
+	.freeze		= dcssblk_freeze,
+	.thaw		= dcssblk_thaw,
+	.restore	= dcssblk_restore,
+};
+
+static struct platform_driver dcssblk_pdrv = {
+	.driver = {
+		.name	= "dcssblk",
+		.owner	= THIS_MODULE,
+		.pm	= &dcssblk_pm_ops,
+	},
+};
+
+static struct platform_device *dcssblk_pdev;
+
+
+/*
  * The init/exit functions.
  */
 static void __exit
 dcssblk_exit(void)
 {
+	platform_device_unregister(dcssblk_pdev);
+	platform_driver_unregister(&dcssblk_pdrv);
 	root_device_unregister(dcssblk_root_dev);
 	unregister_blkdev(dcssblk_major, DCSSBLK_NAME);
 }
@@ -954,30 +1038,44 @@ dcssblk_init(void)
 {
 	int rc;
 
-	dcssblk_root_dev = root_device_register("dcssblk");
-	if (IS_ERR(dcssblk_root_dev))
-		return PTR_ERR(dcssblk_root_dev);
-	rc = device_create_file(dcssblk_root_dev, &dev_attr_add);
-	if (rc) {
-		root_device_unregister(dcssblk_root_dev);
+	rc = platform_driver_register(&dcssblk_pdrv);
+	if (rc)
 		return rc;
+
+	dcssblk_pdev = platform_device_register_simple("dcssblk", -1, NULL,
+							0);
+	if (IS_ERR(dcssblk_pdev)) {
+		rc = PTR_ERR(dcssblk_pdev);
+		goto out_pdrv;
 	}
-	rc = device_create_file(dcssblk_root_dev, &dev_attr_remove);
-	if (rc) {
-		root_device_unregister(dcssblk_root_dev);
-		return rc;
+
+	dcssblk_root_dev = root_device_register("dcssblk");
+	if (IS_ERR(dcssblk_root_dev)) {
+		rc = PTR_ERR(dcssblk_root_dev);
+		goto out_pdev;
 	}
+	rc = device_create_file(dcssblk_root_dev, &dev_attr_add);
+	if (rc)
+		goto out_root;
+	rc = device_create_file(dcssblk_root_dev, &dev_attr_remove);
+	if (rc)
+		goto out_root;
 	rc = register_blkdev(0, DCSSBLK_NAME);
-	if (rc < 0) {
-		root_device_unregister(dcssblk_root_dev);
-		return rc;
-	}
+	if (rc < 0)
+		goto out_root;
 	dcssblk_major = rc;
 	init_rwsem(&dcssblk_devices_sem);
 
 	dcssblk_check_params();
-
 	return 0;
+
+out_root:
+	root_device_unregister(dcssblk_root_dev);
+out_pdev:
+	platform_device_unregister(dcssblk_pdev);
+out_pdrv:
+	platform_driver_unregister(&dcssblk_pdrv);
+	return rc;
 }
 
 module_init(dcssblk_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 37/38] pm: dcssblk power management callbacks.

[Martin Schwidefsky]

[-- Attachment #1: pm_dcssblk.patch --]
[-- Type: text/plain, Size: 4476 bytes --]

From: Gerald Schaefer <gerald.schaefer@de.ibm.com>

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/block/dcssblk.c |  132 +++++++++++++++++++++++++++++++++++++------
 1 file changed, 115 insertions(+), 17 deletions(-)

Index: linux-2.6/drivers/s390/block/dcssblk.c
===================================================================
--- linux-2.6.orig/drivers/s390/block/dcssblk.c
+++ linux-2.6/drivers/s390/block/dcssblk.c
@@ -14,10 +14,11 @@
 #include <linux/init.h>
 #include <linux/slab.h>
 #include <linux/blkdev.h>
-#include <asm/extmem.h>
-#include <asm/io.h>
 #include <linux/completion.h>
 #include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <asm/extmem.h>
+#include <asm/io.h>
 
 #define DCSSBLK_NAME "dcssblk"
 #define DCSSBLK_MINORS_PER_DISK 1
@@ -940,11 +941,94 @@ dcssblk_check_params(void)
 }
 
 /*
+ * Suspend / Resume
+ */
+static int dcssblk_freeze(struct device *dev)
+{
+	struct dcssblk_dev_info *dev_info;
+	int rc = 0;
+
+	list_for_each_entry(dev_info, &dcssblk_devices, lh) {
+		switch (dev_info->segment_type) {
+			case SEG_TYPE_SR:
+			case SEG_TYPE_ER:
+			case SEG_TYPE_SC:
+				if (!dev_info->is_shared)
+					rc = -EINVAL;
+				break;
+			default:
+				rc = -EINVAL;
+				break;
+		}
+		if (rc)
+			break;
+	}
+	if (rc)
+		pr_err("Suspend failed because device %s is writeable.\n",
+		       dev_info->segment_name);
+	return rc;
+}
+
+static int dcssblk_restore(struct device *dev)
+{
+	struct dcssblk_dev_info *dev_info;
+	struct segment_info *entry;
+	unsigned long start, end;
+	int rc = 0;
+
+	list_for_each_entry(dev_info, &dcssblk_devices, lh) {
+		list_for_each_entry(entry, &dev_info->seg_list, lh) {
+			segment_unload(entry->segment_name);
+			rc = segment_load(entry->segment_name, SEGMENT_SHARED,
+					  &start, &end);
+			if (rc < 0) {
+// TODO in_use check ?
+				segment_warning(rc, entry->segment_name);
+				goto out_panic;
+			}
+			if (start != entry->start || end != entry->end) {
+				pr_err("Mismatch of start / end address after "
+				       "resuming device %s\n",
+				       entry->segment_name);
+				goto out_panic;
+			}
+		}
+	}
+	return 0;
+out_panic:
+	panic("fatal dcssblk resume error\n");
+}
+
+static int dcssblk_thaw(struct device *dev)
+{
+	return 0;
+}
+
+static struct dev_pm_ops dcssblk_pm_ops = {
+	.freeze		= dcssblk_freeze,
+	.thaw		= dcssblk_thaw,
+	.restore	= dcssblk_restore,
+};
+
+static struct platform_driver dcssblk_pdrv = {
+	.driver = {
+		.name	= "dcssblk",
+		.owner	= THIS_MODULE,
+		.pm	= &dcssblk_pm_ops,
+	},
+};
+
+static struct platform_device *dcssblk_pdev;
+
+
+/*
  * The init/exit functions.
  */
 static void __exit
 dcssblk_exit(void)
 {
+	platform_device_unregister(dcssblk_pdev);
+	platform_driver_unregister(&dcssblk_pdrv);
 	root_device_unregister(dcssblk_root_dev);
 	unregister_blkdev(dcssblk_major, DCSSBLK_NAME);
 }
@@ -954,30 +1038,44 @@ dcssblk_init(void)
 {
 	int rc;
 
-	dcssblk_root_dev = root_device_register("dcssblk");
-	if (IS_ERR(dcssblk_root_dev))
-		return PTR_ERR(dcssblk_root_dev);
-	rc = device_create_file(dcssblk_root_dev, &dev_attr_add);
-	if (rc) {
-		root_device_unregister(dcssblk_root_dev);
+	rc = platform_driver_register(&dcssblk_pdrv);
+	if (rc)
 		return rc;
+
+	dcssblk_pdev = platform_device_register_simple("dcssblk", -1, NULL,
+							0);
+	if (IS_ERR(dcssblk_pdev)) {
+		rc = PTR_ERR(dcssblk_pdev);
+		goto out_pdrv;
 	}
-	rc = device_create_file(dcssblk_root_dev, &dev_attr_remove);
-	if (rc) {
-		root_device_unregister(dcssblk_root_dev);
-		return rc;
+
+	dcssblk_root_dev = root_device_register("dcssblk");
+	if (IS_ERR(dcssblk_root_dev)) {
+		rc = PTR_ERR(dcssblk_root_dev);
+		goto out_pdev;
 	}
+	rc = device_create_file(dcssblk_root_dev, &dev_attr_add);
+	if (rc)
+		goto out_root;
+	rc = device_create_file(dcssblk_root_dev, &dev_attr_remove);
+	if (rc)
+		goto out_root;
 	rc = register_blkdev(0, DCSSBLK_NAME);
-	if (rc < 0) {
-		root_device_unregister(dcssblk_root_dev);
-		return rc;
-	}
+	if (rc < 0)
+		goto out_root;
 	dcssblk_major = rc;
 	init_rwsem(&dcssblk_devices_sem);
 
 	dcssblk_check_params();
-
 	return 0;
+
+out_root:
+	root_device_unregister(dcssblk_root_dev);
+out_pdev:
+	platform_device_unregister(dcssblk_pdev);
+out_pdrv:
+	platform_driver_unregister(&dcssblk_pdrv);
+	return rc;
 }
 
 module_init(dcssblk_init);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 38/38] pm: ap bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ap_bus.patch --]
[-- Type: text/plain, Size: 4198 bytes --]

From: Felix Beck <felix.beck@de.ibm.com>

Signed-off-by: Felix Beck <felix.beck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/crypto/ap_bus.c |   85 +++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 83 insertions(+), 2 deletions(-)

Index: linux-2.6/drivers/s390/crypto/ap_bus.c
===================================================================
--- linux-2.6.orig/drivers/s390/crypto/ap_bus.c
+++ linux-2.6/drivers/s390/crypto/ap_bus.c
@@ -54,6 +54,12 @@ static int ap_poll_thread_start(void);
 static void ap_poll_thread_stop(void);
 static void ap_request_timeout(unsigned long);
 static inline void ap_schedule_poll_timer(void);
+static int __ap_poll_device(struct ap_device *ap_dev, unsigned long *flags);
+static int ap_device_remove(struct device *dev);
+static int ap_device_probe(struct device *dev);
+static void ap_interrupt_handler(void *unused1, void *unused2);
+static void ap_reset(struct ap_device *ap_dev);
+static void ap_config_timeout(unsigned long ptr);
 
 /*
  * Module description.
@@ -101,6 +107,10 @@ static struct hrtimer ap_poll_timer;
  * If z/VM change to 1500000 nanoseconds to adjust to z/VM polling.*/
 static unsigned long long poll_timeout = 250000;
 
+/* Suspend flag */
+static int ap_suspend_flag;
+static struct bus_type ap_bus_type;
+
 /**
  * ap_using_interrupts() - Returns non-zero if interrupt support is
  * available.
@@ -617,10 +627,79 @@ static int ap_uevent (struct device *dev
 	return retval;
 }
 
+static int ap_bus_suspend(struct device *dev, pm_message_t state)
+{
+	struct ap_device *ap_dev = to_ap_dev(dev);
+	unsigned long flags;
+
+	if (!ap_suspend_flag) {
+		ap_suspend_flag = 1;
+
+		/* Disable scanning for devices, thus we do not want to scan
+		 * for them after removing.
+		 */
+		del_timer_sync(&ap_config_timer);
+		if (ap_work_queue != NULL) {
+			destroy_workqueue(ap_work_queue);
+			ap_work_queue = NULL;
+		}
+		tasklet_disable(&ap_tasklet);
+	}
+	/* Poll on the device until all requests are finished. */
+	do {
+		flags = 0;
+		__ap_poll_device(ap_dev, &flags);
+	} while ((flags & 1) || (flags & 2));
+
+	ap_device_remove(dev);
+	return 0;
+}
+
+static int ap_bus_resume(struct device *dev)
+{
+	int rc = 0;
+	struct ap_device *ap_dev = to_ap_dev(dev);
+
+	if (ap_suspend_flag) {
+		ap_suspend_flag = 0;
+		if (!ap_interrupts_available())
+			ap_interrupt_indicator = NULL;
+		ap_device_probe(dev);
+		ap_reset(ap_dev);
+		setup_timer(&ap_dev->timeout, ap_request_timeout,
+			    (unsigned long) ap_dev);
+		ap_scan_bus(NULL);
+		init_timer(&ap_config_timer);
+		ap_config_timer.function = ap_config_timeout;
+		ap_config_timer.data = 0;
+		ap_config_timer.expires = jiffies + ap_config_time * HZ;
+		add_timer(&ap_config_timer);
+		ap_work_queue = create_singlethread_workqueue("kapwork");
+		if (!ap_work_queue)
+			return -ENOMEM;
+		tasklet_enable(&ap_tasklet);
+		if (!ap_using_interrupts())
+			ap_schedule_poll_timer();
+		else
+			tasklet_schedule(&ap_tasklet);
+		if (ap_thread_flag)
+			rc = ap_poll_thread_start();
+	} else {
+		ap_device_probe(dev);
+		ap_reset(ap_dev);
+		setup_timer(&ap_dev->timeout, ap_request_timeout,
+			    (unsigned long) ap_dev);
+	}
+
+	return rc;
+}
+
 static struct bus_type ap_bus_type = {
 	.name = "ap",
 	.match = &ap_bus_match,
 	.uevent = &ap_uevent,
+	.suspend = ap_bus_suspend,
+	.resume = ap_bus_resume
 };
 
 static int ap_device_probe(struct device *dev)
@@ -1066,7 +1145,7 @@ ap_config_timeout(unsigned long ptr)
  */
 static inline void ap_schedule_poll_timer(void)
 {
-	if (ap_using_interrupts())
+	if (ap_using_interrupts() || ap_suspend_flag)
 		return;
 	if (hrtimer_is_queued(&ap_poll_timer))
 		return;
@@ -1384,6 +1463,8 @@ static int ap_poll_thread(void *data)
 
 	set_user_nice(current, 19);
 	while (1) {
+		if (ap_suspend_flag)
+			return 0;
 		if (need_resched()) {
 			schedule();
 			continue;
@@ -1414,7 +1495,7 @@ static int ap_poll_thread_start(void)
 {
 	int rc;
 
-	if (ap_using_interrupts())
+	if (ap_using_interrupts() || ap_suspend_flag)
 		return 0;
 	mutex_lock(&ap_poll_thread_mutex);
 	if (!ap_poll_kthread) {

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 38/38] pm: ap bus power management callbacks

[Martin Schwidefsky]

[-- Attachment #1: pm_ap_bus.patch --]
[-- Type: text/plain, Size: 4197 bytes --]

From: Felix Beck <felix.beck@de.ibm.com>

Signed-off-by: Felix Beck <felix.beck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 drivers/s390/crypto/ap_bus.c |   85 +++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 83 insertions(+), 2 deletions(-)

Index: linux-2.6/drivers/s390/crypto/ap_bus.c
===================================================================
--- linux-2.6.orig/drivers/s390/crypto/ap_bus.c
+++ linux-2.6/drivers/s390/crypto/ap_bus.c
@@ -54,6 +54,12 @@ static int ap_poll_thread_start(void);
 static void ap_poll_thread_stop(void);
 static void ap_request_timeout(unsigned long);
 static inline void ap_schedule_poll_timer(void);
+static int __ap_poll_device(struct ap_device *ap_dev, unsigned long *flags);
+static int ap_device_remove(struct device *dev);
+static int ap_device_probe(struct device *dev);
+static void ap_interrupt_handler(void *unused1, void *unused2);
+static void ap_reset(struct ap_device *ap_dev);
+static void ap_config_timeout(unsigned long ptr);
 
 /*
  * Module description.
@@ -101,6 +107,10 @@ static struct hrtimer ap_poll_timer;
  * If z/VM change to 1500000 nanoseconds to adjust to z/VM polling.*/
 static unsigned long long poll_timeout = 250000;
 
+/* Suspend flag */
+static int ap_suspend_flag;
+static struct bus_type ap_bus_type;
+
 /**
  * ap_using_interrupts() - Returns non-zero if interrupt support is
  * available.
@@ -617,10 +627,79 @@ static int ap_uevent (struct device *dev
 	return retval;
 }
 
+static int ap_bus_suspend(struct device *dev, pm_message_t state)
+{
+	struct ap_device *ap_dev = to_ap_dev(dev);
+	unsigned long flags;
+
+	if (!ap_suspend_flag) {
+		ap_suspend_flag = 1;
+
+		/* Disable scanning for devices, thus we do not want to scan
+		 * for them after removing.
+		 */
+		del_timer_sync(&ap_config_timer);
+		if (ap_work_queue != NULL) {
+			destroy_workqueue(ap_work_queue);
+			ap_work_queue = NULL;
+		}
+		tasklet_disable(&ap_tasklet);
+	}
+	/* Poll on the device until all requests are finished. */
+	do {
+		flags = 0;
+		__ap_poll_device(ap_dev, &flags);
+	} while ((flags & 1) || (flags & 2));
+
+	ap_device_remove(dev);
+	return 0;
+}
+
+static int ap_bus_resume(struct device *dev)
+{
+	int rc = 0;
+	struct ap_device *ap_dev = to_ap_dev(dev);
+
+	if (ap_suspend_flag) {
+		ap_suspend_flag = 0;
+		if (!ap_interrupts_available())
+			ap_interrupt_indicator = NULL;
+		ap_device_probe(dev);
+		ap_reset(ap_dev);
+		setup_timer(&ap_dev->timeout, ap_request_timeout,
+			    (unsigned long) ap_dev);
+		ap_scan_bus(NULL);
+		init_timer(&ap_config_timer);
+		ap_config_timer.function = ap_config_timeout;
+		ap_config_timer.data = 0;
+		ap_config_timer.expires = jiffies + ap_config_time * HZ;
+		add_timer(&ap_config_timer);
+		ap_work_queue = create_singlethread_workqueue("kapwork");
+		if (!ap_work_queue)
+			return -ENOMEM;
+		tasklet_enable(&ap_tasklet);
+		if (!ap_using_interrupts())
+			ap_schedule_poll_timer();
+		else
+			tasklet_schedule(&ap_tasklet);
+		if (ap_thread_flag)
+			rc = ap_poll_thread_start();
+	} else {
+		ap_device_probe(dev);
+		ap_reset(ap_dev);
+		setup_timer(&ap_dev->timeout, ap_request_timeout,
+			    (unsigned long) ap_dev);
+	}
+
+	return rc;
+}
+
 static struct bus_type ap_bus_type = {
 	.name = "ap",
 	.match = &ap_bus_match,
 	.uevent = &ap_uevent,
+	.suspend = ap_bus_suspend,
+	.resume = ap_bus_resume
 };
 
 static int ap_device_probe(struct device *dev)
@@ -1066,7 +1145,7 @@ ap_config_timeout(unsigned long ptr)
  */
 static inline void ap_schedule_poll_timer(void)
 {
-	if (ap_using_interrupts())
+	if (ap_using_interrupts() || ap_suspend_flag)
 		return;
 	if (hrtimer_is_queued(&ap_poll_timer))
 		return;
@@ -1384,6 +1463,8 @@ static int ap_poll_thread(void *data)
 
 	set_user_nice(current, 19);
 	while (1) {
+		if (ap_suspend_flag)
+			return 0;
 		if (need_resched()) {
 			schedule();
 			continue;
@@ -1414,7 +1495,7 @@ static int ap_poll_thread_start(void)
 {
 	int rc;
 
-	if (ap_using_interrupts())
+	if (ap_using_interrupts() || ap_suspend_flag)
 		return 0;
 	mutex_lock(&ap_poll_thread_mutex);
 	if (!ap_poll_kthread) {

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [patch 01/38] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Thu 2009-06-04 18:18:48, Martin Schwidefsky wrote:
> From: Cornelia Huck <cornelia.huck@de.ibm.com>
> 
> The *_nvs_* routines in swsusp.c make use of the io*map()
> functions, which are only provided for HAS_IOMEM, thus
> breaking compilation if HAS_IOMEM is not set. Fix this
> by moving the *_nvs_* routines into nvs.c, which is only
> compiled if HAS_IOMEM is set.
> 
> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
> ---
>  include/linux/suspend.h |   18 ++++--
>  kernel/power/Kconfig    |    4 +
>  kernel/power/Makefile   |    1 
>  kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
>  kernel/power/swsusp.c   |  122 --------------------------------------------
>  5 files changed, 147 insertions(+), 129 deletions(-)
> 
> Index: linux-2.6/kernel/power/nvs.c
> ===================================================================
> --- /dev/null
> +++ linux-2.6/kernel/power/nvs.c

Ideally, filename would be a bit more descriptive. 

> @@ -0,0 +1,131 @@
> +/*
> + * Routines for NVS memory handling
> + */

If you copy&pasted code, you need to copy&paste copyright notices, too.

> --- linux-2.6.orig/kernel/power/Makefile
> +++ linux-2.6/kernel/power/Makefile
> @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
>  obj-$(CONFIG_PM_SLEEP)		+= console.o
>  obj-$(CONFIG_FREEZER)		+= process.o
>  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> +obj-$(CONFIG_NVS)		+= nvs.o
>  
>  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o

CONFIG_NVS is definitely not descriptive enough. C_HIBERNATION_NVS?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [patch 01/38] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Thu 2009-06-04 18:18:48, Martin Schwidefsky wrote:
> From: Cornelia Huck <cornelia.huck@de.ibm.com>
> 
> The *_nvs_* routines in swsusp.c make use of the io*map()
> functions, which are only provided for HAS_IOMEM, thus
> breaking compilation if HAS_IOMEM is not set. Fix this
> by moving the *_nvs_* routines into nvs.c, which is only
> compiled if HAS_IOMEM is set.
> 
> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
> ---
>  include/linux/suspend.h |   18 ++++--
>  kernel/power/Kconfig    |    4 +
>  kernel/power/Makefile   |    1 
>  kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
>  kernel/power/swsusp.c   |  122 --------------------------------------------
>  5 files changed, 147 insertions(+), 129 deletions(-)
> 
> Index: linux-2.6/kernel/power/nvs.c
> ===================================================================
> --- /dev/null
> +++ linux-2.6/kernel/power/nvs.c

Ideally, filename would be a bit more descriptive. 

> @@ -0,0 +1,131 @@
> +/*
> + * Routines for NVS memory handling
> + */

If you copy&pasted code, you need to copy&paste copyright notices, too.

> --- linux-2.6.orig/kernel/power/Makefile
> +++ linux-2.6/kernel/power/Makefile
> @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
>  obj-$(CONFIG_PM_SLEEP)		+= console.o
>  obj-$(CONFIG_FREEZER)		+= process.o
>  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> +obj-$(CONFIG_NVS)		+= nvs.o
>  
>  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o

CONFIG_NVS is definitely not descriptive enough. C_HIBERNATION_NVS?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [patch 05/38] s390: hibernation support for s390

[Pavel Machek]

On Thu 2009-06-04 18:18:52, Martin Schwidefsky wrote:
> From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> 
> This patch introduces the hibernation backend support to the 
> s390 architecture. Now it is possible to suspend a mainframe Linux 
> guest using the following command:
> 
> echo disk > /sys/power/state

Should this one be last in series?

Otherwise no obvious problems...


> Index: linux-2.6/arch/s390/power/swsusp.c
> ===================================================================
> --- /dev/null
> +++ linux-2.6/arch/s390/power/swsusp.c
> @@ -0,0 +1,31 @@
> +/*
> + * Support for suspend and resume on s390
> + *
> + * Copyright IBM Corp. 2009
> + *
> + * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> + *
> + */
> +
> +#include <linux/suspend.h>
> +#include <linux/gfp.h>
> +#include <linux/cpu.h>
> +#include <linux/smp.h>
> +#include <linux/cpumask.h>
> +#include <asm/ipl.h>
> +#include <asm/processor.h>

Empty functions do not need 1000 includes...

> +/*
> + * save CPU registers before creating a hibernation image and before
> + * restoring the memory state from it
> + */
> +void save_processor_state(void)
> +{
> +}
> +
> +/*
> + * restore the contents of CPU registers
> + */
> +void restore_processor_state(void)
> +{
> +}

...and explaining why the functions can be emmpty would be nice.

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [patch 05/38] s390: hibernation support for s390

[Pavel Machek]

On Thu 2009-06-04 18:18:52, Martin Schwidefsky wrote:
> From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> 
> This patch introduces the hibernation backend support to the 
> s390 architecture. Now it is possible to suspend a mainframe Linux 
> guest using the following command:
> 
> echo disk > /sys/power/state

Should this one be last in series?

Otherwise no obvious problems...


> Index: linux-2.6/arch/s390/power/swsusp.c
> ===================================================================
> --- /dev/null
> +++ linux-2.6/arch/s390/power/swsusp.c
> @@ -0,0 +1,31 @@
> +/*
> + * Support for suspend and resume on s390
> + *
> + * Copyright IBM Corp. 2009
> + *
> + * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> + *
> + */
> +
> +#include <linux/suspend.h>
> +#include <linux/gfp.h>
> +#include <linux/cpu.h>
> +#include <linux/smp.h>
> +#include <linux/cpumask.h>
> +#include <asm/ipl.h>
> +#include <asm/processor.h>

Empty functions do not need 1000 includes...

> +/*
> + * save CPU registers before creating a hibernation image and before
> + * restoring the memory state from it
> + */
> +void save_processor_state(void)
> +{
> +}
> +
> +/*
> + * restore the contents of CPU registers
> + */
> +void restore_processor_state(void)
> +{
> +}

...and explaining why the functions can be emmpty would be nice.

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [patch 01/38] pm: Move nvs routines into a seperate file.

[Cornelia Huck]

On Mon, 8 Jun 2009 08:35:01 +0200,
Pavel Machek <pavel@ucw.cz> wrote:

> On Thu 2009-06-04 18:18:48, Martin Schwidefsky wrote:
> > From: Cornelia Huck <cornelia.huck@de.ibm.com>
> > 
> > The *_nvs_* routines in swsusp.c make use of the io*map()
> > functions, which are only provided for HAS_IOMEM, thus
> > breaking compilation if HAS_IOMEM is not set. Fix this
> > by moving the *_nvs_* routines into nvs.c, which is only
> > compiled if HAS_IOMEM is set.
> > 
> > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
> > ---
> >  include/linux/suspend.h |   18 ++++--
> >  kernel/power/Kconfig    |    4 +
> >  kernel/power/Makefile   |    1 
> >  kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
> >  kernel/power/swsusp.c   |  122 --------------------------------------------
> >  5 files changed, 147 insertions(+), 129 deletions(-)
> > 
> > Index: linux-2.6/kernel/power/nvs.c
> > ===================================================================
> > --- /dev/null
> > +++ linux-2.6/kernel/power/nvs.c
> 
> Ideally, filename would be a bit more descriptive. 

hibernate_nvs.c?

> 
> > @@ -0,0 +1,131 @@
> > +/*
> > + * Routines for NVS memory handling
> > + */
> 
> If you copy&pasted code, you need to copy&paste copyright notices, too.

Which ones? Rafael, it seems you wrote the NVS code - which copyright
should I use?

> 
> > --- linux-2.6.orig/kernel/power/Makefile
> > +++ linux-2.6/kernel/power/Makefile
> > @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
> >  obj-$(CONFIG_PM_SLEEP)		+= console.o
> >  obj-$(CONFIG_FREEZER)		+= process.o
> >  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> > +obj-$(CONFIG_NVS)		+= nvs.o
> >  
> >  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
> 
> CONFIG_NVS is definitely not descriptive enough. C_HIBERNATION_NVS?

It's a hidden config variable - but I can change it to
CONFIG_HIBERNATION_NVS.

Re: [patch 01/38] pm: Move nvs routines into a seperate file.

[Cornelia Huck]

On Mon, 8 Jun 2009 08:35:01 +0200,
Pavel Machek <pavel@ucw.cz> wrote:

> On Thu 2009-06-04 18:18:48, Martin Schwidefsky wrote:
> > From: Cornelia Huck <cornelia.huck@de.ibm.com>
> > 
> > The *_nvs_* routines in swsusp.c make use of the io*map()
> > functions, which are only provided for HAS_IOMEM, thus
> > breaking compilation if HAS_IOMEM is not set. Fix this
> > by moving the *_nvs_* routines into nvs.c, which is only
> > compiled if HAS_IOMEM is set.
> > 
> > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
> > ---
> >  include/linux/suspend.h |   18 ++++--
> >  kernel/power/Kconfig    |    4 +
> >  kernel/power/Makefile   |    1 
> >  kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
> >  kernel/power/swsusp.c   |  122 --------------------------------------------
> >  5 files changed, 147 insertions(+), 129 deletions(-)
> > 
> > Index: linux-2.6/kernel/power/nvs.c
> > ===================================================================
> > --- /dev/null
> > +++ linux-2.6/kernel/power/nvs.c
> 
> Ideally, filename would be a bit more descriptive. 

hibernate_nvs.c?

> 
> > @@ -0,0 +1,131 @@
> > +/*
> > + * Routines for NVS memory handling
> > + */
> 
> If you copy&pasted code, you need to copy&paste copyright notices, too.

Which ones? Rafael, it seems you wrote the NVS code - which copyright
should I use?

> 
> > --- linux-2.6.orig/kernel/power/Makefile
> > +++ linux-2.6/kernel/power/Makefile
> > @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
> >  obj-$(CONFIG_PM_SLEEP)		+= console.o
> >  obj-$(CONFIG_FREEZER)		+= process.o
> >  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> > +obj-$(CONFIG_NVS)		+= nvs.o
> >  
> >  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
> 
> CONFIG_NVS is definitely not descriptive enough. C_HIBERNATION_NVS?

It's a hidden config variable - but I can change it to
CONFIG_HIBERNATION_NVS.

Re: [patch 01/38] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Monday 08 June 2009, Cornelia Huck wrote:
> On Mon, 8 Jun 2009 08:35:01 +0200,
> Pavel Machek <pavel@ucw.cz> wrote:
> 
> > On Thu 2009-06-04 18:18:48, Martin Schwidefsky wrote:
> > > From: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > 
> > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > functions, which are only provided for HAS_IOMEM, thus
> > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > by moving the *_nvs_* routines into nvs.c, which is only
> > > compiled if HAS_IOMEM is set.
> > > 
> > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
> > > ---
> > >  include/linux/suspend.h |   18 ++++--
> > >  kernel/power/Kconfig    |    4 +
> > >  kernel/power/Makefile   |    1 
> > >  kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
> > >  kernel/power/swsusp.c   |  122 --------------------------------------------
> > >  5 files changed, 147 insertions(+), 129 deletions(-)
> > > 
> > > Index: linux-2.6/kernel/power/nvs.c
> > > ===================================================================
> > > --- /dev/null
> > > +++ linux-2.6/kernel/power/nvs.c
> > 
> > Ideally, filename would be a bit more descriptive. 
> 
> hibernate_nvs.c?

Yes, please.

> > > @@ -0,0 +1,131 @@
> > > +/*
> > > + * Routines for NVS memory handling
> > > + */
> > 
> > If you copy&pasted code, you need to copy&paste copyright notices, too.
> 
> Which ones? Rafael, it seems you wrote the NVS code - which copyright
> should I use?

Same as in kernel/irq/pm.c , please.

> > > --- linux-2.6.orig/kernel/power/Makefile
> > > +++ linux-2.6/kernel/power/Makefile
> > > @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
> > >  obj-$(CONFIG_PM_SLEEP)		+= console.o
> > >  obj-$(CONFIG_FREEZER)		+= process.o
> > >  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> > > +obj-$(CONFIG_NVS)		+= nvs.o
> > >  
> > >  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
> > 
> > CONFIG_NVS is definitely not descriptive enough. C_HIBERNATION_NVS?
> 
> It's a hidden config variable - but I can change it to
> CONFIG_HIBERNATION_NVS.

Please do that too.

When you're done, please send the patch to me and I'll put it into the suspend
tree for 2.6.31.

Best,
Rafael

Re: [patch 01/38] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Monday 08 June 2009, Cornelia Huck wrote:
> On Mon, 8 Jun 2009 08:35:01 +0200,
> Pavel Machek <pavel@ucw.cz> wrote:
> 
> > On Thu 2009-06-04 18:18:48, Martin Schwidefsky wrote:
> > > From: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > 
> > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > functions, which are only provided for HAS_IOMEM, thus
> > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > by moving the *_nvs_* routines into nvs.c, which is only
> > > compiled if HAS_IOMEM is set.
> > > 
> > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
> > > ---
> > >  include/linux/suspend.h |   18 ++++--
> > >  kernel/power/Kconfig    |    4 +
> > >  kernel/power/Makefile   |    1 
> > >  kernel/power/nvs.c      |  131 ++++++++++++++++++++++++++++++++++++++++++++++++
> > >  kernel/power/swsusp.c   |  122 --------------------------------------------
> > >  5 files changed, 147 insertions(+), 129 deletions(-)
> > > 
> > > Index: linux-2.6/kernel/power/nvs.c
> > > ===================================================================
> > > --- /dev/null
> > > +++ linux-2.6/kernel/power/nvs.c
> > 
> > Ideally, filename would be a bit more descriptive. 
> 
> hibernate_nvs.c?

Yes, please.

> > > @@ -0,0 +1,131 @@
> > > +/*
> > > + * Routines for NVS memory handling
> > > + */
> > 
> > If you copy&pasted code, you need to copy&paste copyright notices, too.
> 
> Which ones? Rafael, it seems you wrote the NVS code - which copyright
> should I use?

Same as in kernel/irq/pm.c , please.

> > > --- linux-2.6.orig/kernel/power/Makefile
> > > +++ linux-2.6/kernel/power/Makefile
> > > @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
> > >  obj-$(CONFIG_PM_SLEEP)		+= console.o
> > >  obj-$(CONFIG_FREEZER)		+= process.o
> > >  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> > > +obj-$(CONFIG_NVS)		+= nvs.o
> > >  
> > >  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
> > 
> > CONFIG_NVS is definitely not descriptive enough. C_HIBERNATION_NVS?
> 
> It's a hidden config variable - but I can change it to
> CONFIG_HIBERNATION_NVS.

Please do that too.

When you're done, please send the patch to me and I'll put it into the suspend
tree for 2.6.31.

Best,
Rafael

[PATCH v2] pm: Move nvs routines into a seperate file.

[Cornelia Huck]

The *_nvs_* routines in swsusp.c make use of the io*map()
functions, which are only provided for HAS_IOMEM, thus
breaking compilation if HAS_IOMEM is not set. Fix this
by moving the *_nvs_* routines into hibernation_nvs.c, which
is only compiled if HAS_IOMEM is set.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

---
 include/linux/suspend.h        |   18 +++--
 kernel/power/Kconfig           |    4 +
 kernel/power/Makefile          |    1 
 kernel/power/hibernation_nvs.c |  135 +++++++++++++++++++++++++++++++++++++++++
 kernel/power/swsusp.c          |  122 -------------------------------------
 5 files changed, 151 insertions(+), 129 deletions(-)

--- /dev/null
+++ linux-suspend/kernel/power/hibernation_nvs.c
@@ -0,0 +1,135 @@
+/*
+ * linux/kernel/power/hibernation_nvs.c
+ *
+ * Copyright (C) 2008,2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
+ *
+ * Routines for NVS memory handling
+ */
+
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/mm.h>
+#include <linux/suspend.h>
+
+/*
+ * Platforms, like ACPI, may want us to save some memory used by them during
+ * hibernation and to restore the contents of this memory during the subsequent
+ * resume.  The code below implements a mechanism allowing us to do that.
+ */
+
+struct nvs_page {
+	unsigned long phys_start;
+	unsigned int size;
+	void *kaddr;
+	void *data;
+	struct list_head node;
+};
+
+static LIST_HEAD(nvs_list);
+
+/**
+ *	hibernate_nvs_register - register platform NVS memory region to save
+ *	@start - physical address of the region
+ *	@size - size of the region
+ *
+ *	The NVS region need not be page-aligned (both ends) and we arrange
+ *	things so that the data from page-aligned addresses in this region will
+ *	be copied into separate RAM pages.
+ */
+int hibernate_nvs_register(unsigned long start, unsigned long size)
+{
+	struct nvs_page *entry, *next;
+
+	while (size > 0) {
+		unsigned int nr_bytes;
+
+		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
+		if (!entry)
+			goto Error;
+
+		list_add_tail(&entry->node, &nvs_list);
+		entry->phys_start = start;
+		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
+		entry->size = (size < nr_bytes) ? size : nr_bytes;
+
+		start += entry->size;
+		size -= entry->size;
+	}
+	return 0;
+
+ Error:
+	list_for_each_entry_safe(entry, next, &nvs_list, node) {
+		list_del(&entry->node);
+		kfree(entry);
+	}
+	return -ENOMEM;
+}
+
+/**
+ *	hibernate_nvs_free - free data pages allocated for saving NVS regions
+ */
+void hibernate_nvs_free(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			free_page((unsigned long)entry->data);
+			entry->data = NULL;
+			if (entry->kaddr) {
+				iounmap(entry->kaddr);
+				entry->kaddr = NULL;
+			}
+		}
+}
+
+/**
+ *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
+ */
+int hibernate_nvs_alloc(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node) {
+		entry->data = (void *)__get_free_page(GFP_KERNEL);
+		if (!entry->data) {
+			hibernate_nvs_free();
+			return -ENOMEM;
+		}
+	}
+	return 0;
+}
+
+/**
+ *	hibernate_nvs_save - save NVS memory regions
+ */
+void hibernate_nvs_save(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Saving platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			entry->kaddr = ioremap(entry->phys_start, entry->size);
+			memcpy(entry->data, entry->kaddr, entry->size);
+		}
+}
+
+/**
+ *	hibernate_nvs_restore - restore NVS memory regions
+ *
+ *	This function is going to be called with interrupts disabled, so it
+ *	cannot iounmap the virtual addresses used to access the NVS region.
+ */
+void hibernate_nvs_restore(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data)
+			memcpy(entry->kaddr, entry->data, entry->size);
+}
--- linux-suspend.orig/kernel/power/swsusp.c
+++ linux-suspend/kernel/power/swsusp.c
@@ -186,125 +186,3 @@ void swsusp_show_speed(struct timeval *s
 			centisecs / 100, centisecs % 100,
 			kps / 1000, (kps % 1000) / 10);
 }
-
-/*
- * Platforms, like ACPI, may want us to save some memory used by them during
- * hibernation and to restore the contents of this memory during the subsequent
- * resume.  The code below implements a mechanism allowing us to do that.
- */
-
-struct nvs_page {
-	unsigned long phys_start;
-	unsigned int size;
-	void *kaddr;
-	void *data;
-	struct list_head node;
-};
-
-static LIST_HEAD(nvs_list);
-
-/**
- *	hibernate_nvs_register - register platform NVS memory region to save
- *	@start - physical address of the region
- *	@size - size of the region
- *
- *	The NVS region need not be page-aligned (both ends) and we arrange
- *	things so that the data from page-aligned addresses in this region will
- *	be copied into separate RAM pages.
- */
-int hibernate_nvs_register(unsigned long start, unsigned long size)
-{
-	struct nvs_page *entry, *next;
-
-	while (size > 0) {
-		unsigned int nr_bytes;
-
-		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
-		if (!entry)
-			goto Error;
-
-		list_add_tail(&entry->node, &nvs_list);
-		entry->phys_start = start;
-		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
-		entry->size = (size < nr_bytes) ? size : nr_bytes;
-
-		start += entry->size;
-		size -= entry->size;
-	}
-	return 0;
-
- Error:
-	list_for_each_entry_safe(entry, next, &nvs_list, node) {
-		list_del(&entry->node);
-		kfree(entry);
-	}
-	return -ENOMEM;
-}
-
-/**
- *	hibernate_nvs_free - free data pages allocated for saving NVS regions
- */
-void hibernate_nvs_free(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			free_page((unsigned long)entry->data);
-			entry->data = NULL;
-			if (entry->kaddr) {
-				iounmap(entry->kaddr);
-				entry->kaddr = NULL;
-			}
-		}
-}
-
-/**
- *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
- */
-int hibernate_nvs_alloc(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node) {
-		entry->data = (void *)__get_free_page(GFP_KERNEL);
-		if (!entry->data) {
-			hibernate_nvs_free();
-			return -ENOMEM;
-		}
-	}
-	return 0;
-}
-
-/**
- *	hibernate_nvs_save - save NVS memory regions
- */
-void hibernate_nvs_save(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Saving platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			entry->kaddr = ioremap(entry->phys_start, entry->size);
-			memcpy(entry->data, entry->kaddr, entry->size);
-		}
-}
-
-/**
- *	hibernate_nvs_restore - restore NVS memory regions
- *
- *	This function is going to be called with interrupts disabled, so it
- *	cannot iounmap the virtual addresses used to access the NVS region.
- */
-void hibernate_nvs_restore(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data)
-			memcpy(entry->kaddr, entry->data, entry->size);
-}
--- linux-suspend.orig/kernel/power/Kconfig
+++ linux-suspend/kernel/power/Kconfig
@@ -116,9 +116,13 @@ config SUSPEND_FREEZER
 
 	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
 
+config HIBERNATION_NVS
+	bool
+
 config HIBERNATION
 	bool "Hibernation (aka 'suspend to disk')"
 	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
+	select HIBERNATION_NVS if HAS_IOMEM
 	---help---
 	  Enable the suspend to disk (STD) functionality, which is usually
 	  called "hibernation" in user interfaces.  STD checkpoints the
--- linux-suspend.orig/kernel/power/Makefile
+++ linux-suspend/kernel/power/Makefile
@@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
 obj-$(CONFIG_PM_SLEEP)		+= console.o
 obj-$(CONFIG_FREEZER)		+= process.o
 obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
+obj-$(CONFIG_HIBERNATION_NVS)	+= hibernation_nvs.o
 
 obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
--- linux-suspend.orig/include/linux/suspend.h
+++ linux-suspend/include/linux/suspend.h
@@ -245,11 +245,6 @@ extern unsigned long get_safe_page(gfp_t
 
 extern void hibernation_set_ops(struct platform_hibernation_ops *ops);
 extern int hibernate(void);
-extern int hibernate_nvs_register(unsigned long start, unsigned long size);
-extern int hibernate_nvs_alloc(void);
-extern void hibernate_nvs_free(void);
-extern void hibernate_nvs_save(void);
-extern void hibernate_nvs_restore(void);
 extern bool system_entering_hibernation(void);
 #else /* CONFIG_HIBERNATION */
 static inline int swsusp_page_is_forbidden(struct page *p) { return 0; }
@@ -258,6 +253,16 @@ static inline void swsusp_unset_page_fre
 
 static inline void hibernation_set_ops(struct platform_hibernation_ops *ops) {}
 static inline int hibernate(void) { return -ENOSYS; }
+static inline bool system_entering_hibernation(void) { return false; }
+#endif /* CONFIG_HIBERNATION */
+
+#ifdef CONFIG_HIBERNATION_NVS
+extern int hibernate_nvs_register(unsigned long start, unsigned long size);
+extern int hibernate_nvs_alloc(void);
+extern void hibernate_nvs_free(void);
+extern void hibernate_nvs_save(void);
+extern void hibernate_nvs_restore(void);
+#else /* CONFIG_HIBERNATION_NVS */
 static inline int hibernate_nvs_register(unsigned long a, unsigned long b)
 {
 	return 0;
@@ -266,8 +271,7 @@ static inline int hibernate_nvs_alloc(vo
 static inline void hibernate_nvs_free(void) {}
 static inline void hibernate_nvs_save(void) {}
 static inline void hibernate_nvs_restore(void) {}
-static inline bool system_entering_hibernation(void) { return false; }
-#endif /* CONFIG_HIBERNATION */
+#endif /* CONFIG_HIBERNATION_NVS */
 
 #ifdef CONFIG_PM_SLEEP
 void save_processor_state(void);

[PATCH v2] pm: Move nvs routines into a seperate file.

[Cornelia Huck]

The *_nvs_* routines in swsusp.c make use of the io*map()
functions, which are only provided for HAS_IOMEM, thus
breaking compilation if HAS_IOMEM is not set. Fix this
by moving the *_nvs_* routines into hibernation_nvs.c, which
is only compiled if HAS_IOMEM is set.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

---
 include/linux/suspend.h        |   18 +++--
 kernel/power/Kconfig           |    4 +
 kernel/power/Makefile          |    1 
 kernel/power/hibernation_nvs.c |  135 +++++++++++++++++++++++++++++++++++++++++
 kernel/power/swsusp.c          |  122 -------------------------------------
 5 files changed, 151 insertions(+), 129 deletions(-)

--- /dev/null
+++ linux-suspend/kernel/power/hibernation_nvs.c
@@ -0,0 +1,135 @@
+/*
+ * linux/kernel/power/hibernation_nvs.c
+ *
+ * Copyright (C) 2008,2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
+ *
+ * Routines for NVS memory handling
+ */
+
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/mm.h>
+#include <linux/suspend.h>
+
+/*
+ * Platforms, like ACPI, may want us to save some memory used by them during
+ * hibernation and to restore the contents of this memory during the subsequent
+ * resume.  The code below implements a mechanism allowing us to do that.
+ */
+
+struct nvs_page {
+	unsigned long phys_start;
+	unsigned int size;
+	void *kaddr;
+	void *data;
+	struct list_head node;
+};
+
+static LIST_HEAD(nvs_list);
+
+/**
+ *	hibernate_nvs_register - register platform NVS memory region to save
+ *	@start - physical address of the region
+ *	@size - size of the region
+ *
+ *	The NVS region need not be page-aligned (both ends) and we arrange
+ *	things so that the data from page-aligned addresses in this region will
+ *	be copied into separate RAM pages.
+ */
+int hibernate_nvs_register(unsigned long start, unsigned long size)
+{
+	struct nvs_page *entry, *next;
+
+	while (size > 0) {
+		unsigned int nr_bytes;
+
+		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
+		if (!entry)
+			goto Error;
+
+		list_add_tail(&entry->node, &nvs_list);
+		entry->phys_start = start;
+		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
+		entry->size = (size < nr_bytes) ? size : nr_bytes;
+
+		start += entry->size;
+		size -= entry->size;
+	}
+	return 0;
+
+ Error:
+	list_for_each_entry_safe(entry, next, &nvs_list, node) {
+		list_del(&entry->node);
+		kfree(entry);
+	}
+	return -ENOMEM;
+}
+
+/**
+ *	hibernate_nvs_free - free data pages allocated for saving NVS regions
+ */
+void hibernate_nvs_free(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			free_page((unsigned long)entry->data);
+			entry->data = NULL;
+			if (entry->kaddr) {
+				iounmap(entry->kaddr);
+				entry->kaddr = NULL;
+			}
+		}
+}
+
+/**
+ *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
+ */
+int hibernate_nvs_alloc(void)
+{
+	struct nvs_page *entry;
+
+	list_for_each_entry(entry, &nvs_list, node) {
+		entry->data = (void *)__get_free_page(GFP_KERNEL);
+		if (!entry->data) {
+			hibernate_nvs_free();
+			return -ENOMEM;
+		}
+	}
+	return 0;
+}
+
+/**
+ *	hibernate_nvs_save - save NVS memory regions
+ */
+void hibernate_nvs_save(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Saving platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data) {
+			entry->kaddr = ioremap(entry->phys_start, entry->size);
+			memcpy(entry->data, entry->kaddr, entry->size);
+		}
+}
+
+/**
+ *	hibernate_nvs_restore - restore NVS memory regions
+ *
+ *	This function is going to be called with interrupts disabled, so it
+ *	cannot iounmap the virtual addresses used to access the NVS region.
+ */
+void hibernate_nvs_restore(void)
+{
+	struct nvs_page *entry;
+
+	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
+
+	list_for_each_entry(entry, &nvs_list, node)
+		if (entry->data)
+			memcpy(entry->kaddr, entry->data, entry->size);
+}
--- linux-suspend.orig/kernel/power/swsusp.c
+++ linux-suspend/kernel/power/swsusp.c
@@ -186,125 +186,3 @@ void swsusp_show_speed(struct timeval *s
 			centisecs / 100, centisecs % 100,
 			kps / 1000, (kps % 1000) / 10);
 }
-
-/*
- * Platforms, like ACPI, may want us to save some memory used by them during
- * hibernation and to restore the contents of this memory during the subsequent
- * resume.  The code below implements a mechanism allowing us to do that.
- */
-
-struct nvs_page {
-	unsigned long phys_start;
-	unsigned int size;
-	void *kaddr;
-	void *data;
-	struct list_head node;
-};
-
-static LIST_HEAD(nvs_list);
-
-/**
- *	hibernate_nvs_register - register platform NVS memory region to save
- *	@start - physical address of the region
- *	@size - size of the region
- *
- *	The NVS region need not be page-aligned (both ends) and we arrange
- *	things so that the data from page-aligned addresses in this region will
- *	be copied into separate RAM pages.
- */
-int hibernate_nvs_register(unsigned long start, unsigned long size)
-{
-	struct nvs_page *entry, *next;
-
-	while (size > 0) {
-		unsigned int nr_bytes;
-
-		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
-		if (!entry)
-			goto Error;
-
-		list_add_tail(&entry->node, &nvs_list);
-		entry->phys_start = start;
-		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
-		entry->size = (size < nr_bytes) ? size : nr_bytes;
-
-		start += entry->size;
-		size -= entry->size;
-	}
-	return 0;
-
- Error:
-	list_for_each_entry_safe(entry, next, &nvs_list, node) {
-		list_del(&entry->node);
-		kfree(entry);
-	}
-	return -ENOMEM;
-}
-
-/**
- *	hibernate_nvs_free - free data pages allocated for saving NVS regions
- */
-void hibernate_nvs_free(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			free_page((unsigned long)entry->data);
-			entry->data = NULL;
-			if (entry->kaddr) {
-				iounmap(entry->kaddr);
-				entry->kaddr = NULL;
-			}
-		}
-}
-
-/**
- *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
- */
-int hibernate_nvs_alloc(void)
-{
-	struct nvs_page *entry;
-
-	list_for_each_entry(entry, &nvs_list, node) {
-		entry->data = (void *)__get_free_page(GFP_KERNEL);
-		if (!entry->data) {
-			hibernate_nvs_free();
-			return -ENOMEM;
-		}
-	}
-	return 0;
-}
-
-/**
- *	hibernate_nvs_save - save NVS memory regions
- */
-void hibernate_nvs_save(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Saving platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data) {
-			entry->kaddr = ioremap(entry->phys_start, entry->size);
-			memcpy(entry->data, entry->kaddr, entry->size);
-		}
-}
-
-/**
- *	hibernate_nvs_restore - restore NVS memory regions
- *
- *	This function is going to be called with interrupts disabled, so it
- *	cannot iounmap the virtual addresses used to access the NVS region.
- */
-void hibernate_nvs_restore(void)
-{
-	struct nvs_page *entry;
-
-	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
-
-	list_for_each_entry(entry, &nvs_list, node)
-		if (entry->data)
-			memcpy(entry->kaddr, entry->data, entry->size);
-}
--- linux-suspend.orig/kernel/power/Kconfig
+++ linux-suspend/kernel/power/Kconfig
@@ -116,9 +116,13 @@ config SUSPEND_FREEZER
 
 	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
 
+config HIBERNATION_NVS
+	bool
+
 config HIBERNATION
 	bool "Hibernation (aka 'suspend to disk')"
 	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
+	select HIBERNATION_NVS if HAS_IOMEM
 	---help---
 	  Enable the suspend to disk (STD) functionality, which is usually
 	  called "hibernation" in user interfaces.  STD checkpoints the
--- linux-suspend.orig/kernel/power/Makefile
+++ linux-suspend/kernel/power/Makefile
@@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
 obj-$(CONFIG_PM_SLEEP)		+= console.o
 obj-$(CONFIG_FREEZER)		+= process.o
 obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
+obj-$(CONFIG_HIBERNATION_NVS)	+= hibernation_nvs.o
 
 obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
--- linux-suspend.orig/include/linux/suspend.h
+++ linux-suspend/include/linux/suspend.h
@@ -245,11 +245,6 @@ extern unsigned long get_safe_page(gfp_t
 
 extern void hibernation_set_ops(struct platform_hibernation_ops *ops);
 extern int hibernate(void);
-extern int hibernate_nvs_register(unsigned long start, unsigned long size);
-extern int hibernate_nvs_alloc(void);
-extern void hibernate_nvs_free(void);
-extern void hibernate_nvs_save(void);
-extern void hibernate_nvs_restore(void);
 extern bool system_entering_hibernation(void);
 #else /* CONFIG_HIBERNATION */
 static inline int swsusp_page_is_forbidden(struct page *p) { return 0; }
@@ -258,6 +253,16 @@ static inline void swsusp_unset_page_fre
 
 static inline void hibernation_set_ops(struct platform_hibernation_ops *ops) {}
 static inline int hibernate(void) { return -ENOSYS; }
+static inline bool system_entering_hibernation(void) { return false; }
+#endif /* CONFIG_HIBERNATION */
+
+#ifdef CONFIG_HIBERNATION_NVS
+extern int hibernate_nvs_register(unsigned long start, unsigned long size);
+extern int hibernate_nvs_alloc(void);
+extern void hibernate_nvs_free(void);
+extern void hibernate_nvs_save(void);
+extern void hibernate_nvs_restore(void);
+#else /* CONFIG_HIBERNATION_NVS */
 static inline int hibernate_nvs_register(unsigned long a, unsigned long b)
 {
 	return 0;
@@ -266,8 +271,7 @@ static inline int hibernate_nvs_alloc(vo
 static inline void hibernate_nvs_free(void) {}
 static inline void hibernate_nvs_save(void) {}
 static inline void hibernate_nvs_restore(void) {}
-static inline bool system_entering_hibernation(void) { return false; }
-#endif /* CONFIG_HIBERNATION */
+#endif /* CONFIG_HIBERNATION_NVS */
 
 #ifdef CONFIG_PM_SLEEP
 void save_processor_state(void);

Re: [patch 05/38] s390: hibernation support for s390

[Hans-Joachim Picht]

On Mon, 8 Jun 2009 08:44:51 +0200
Pavel Machek <pavel@ucw.cz> wrote:


> > Index: linux-2.6/arch/s390/power/swsusp.c
> > ===================================================================
> > --- /dev/null
> > +++ linux-2.6/arch/s390/power/swsusp.c
> > @@ -0,0 +1,31 @@
> > +/*
> > + * Support for suspend and resume on s390
> > + *
> > + * Copyright IBM Corp. 2009
> > + *
> > + * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> > + *
> > + */
> > +
> > +#include <linux/suspend.h>
> > +#include <linux/gfp.h>
> > +#include <linux/cpu.h>
> > +#include <linux/smp.h>
> > +#include <linux/cpumask.h>
> > +#include <asm/ipl.h>
> > +#include <asm/processor.h>
> 
> Empty functions do not need 1000 includes...

Thanks for spotting this. I will remove these.
 
> > +/*
> > + * save CPU registers before creating a hibernation image and
> > before
> > + * restoring the memory state from it
> > + */
> > +void save_processor_state(void)
> > +{
> > +}
> > +
> > +/*
> > + * restore the contents of CPU registers
> > + */
> > +void restore_processor_state(void)
> > +{
> > +}
> 
> ...and explaining why the functions can be emmpty would be nice.

I integrated the code for this in the assembly backend in the
swsusp_arch_suspend & swsusp_arch_resume functions.

With best regards,

    --Hans

Re: [patch 05/38] s390: hibernation support for s390

[Hans-Joachim Picht]

On Mon, 8 Jun 2009 08:44:51 +0200
Pavel Machek <pavel@ucw.cz> wrote:


> > Index: linux-2.6/arch/s390/power/swsusp.c
> > ===================================================================
> > --- /dev/null
> > +++ linux-2.6/arch/s390/power/swsusp.c
> > @@ -0,0 +1,31 @@
> > +/*
> > + * Support for suspend and resume on s390
> > + *
> > + * Copyright IBM Corp. 2009
> > + *
> > + * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> > + *
> > + */
> > +
> > +#include <linux/suspend.h>
> > +#include <linux/gfp.h>
> > +#include <linux/cpu.h>
> > +#include <linux/smp.h>
> > +#include <linux/cpumask.h>
> > +#include <asm/ipl.h>
> > +#include <asm/processor.h>
> 
> Empty functions do not need 1000 includes...

Thanks for spotting this. I will remove these.
 
> > +/*
> > + * save CPU registers before creating a hibernation image and
> > before
> > + * restoring the memory state from it
> > + */
> > +void save_processor_state(void)
> > +{
> > +}
> > +
> > +/*
> > + * restore the contents of CPU registers
> > + */
> > +void restore_processor_state(void)
> > +{
> > +}
> 
> ...and explaining why the functions can be emmpty would be nice.

I integrated the code for this in the assembly backend in the
swsusp_arch_suspend & swsusp_arch_resume functions.

With best regards,

    --Hans

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Tue 2009-06-09 10:40:03, Cornelia Huck wrote:
> The *_nvs_* routines in swsusp.c make use of the io*map()
> functions, which are only provided for HAS_IOMEM, thus
> breaking compilation if HAS_IOMEM is not set. Fix this
> by moving the *_nvs_* routines into hibernation_nvs.c, which
> is only compiled if HAS_IOMEM is set.
> 
> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

Acked-by: Pavel Machek <pavel@ucw.cz>

> @@ -0,0 +1,135 @@
> +/*
> + * linux/kernel/power/hibernation_nvs.c
> + *
> + * Copyright (C) 2008,2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
> + *
> + * Routines for NVS memory handling
> + */

Specifying  GPLv2 here would be nice. If it contains copyright, it
should contain license, too.
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Tue 2009-06-09 10:40:03, Cornelia Huck wrote:
> The *_nvs_* routines in swsusp.c make use of the io*map()
> functions, which are only provided for HAS_IOMEM, thus
> breaking compilation if HAS_IOMEM is not set. Fix this
> by moving the *_nvs_* routines into hibernation_nvs.c, which
> is only compiled if HAS_IOMEM is set.
> 
> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

Acked-by: Pavel Machek <pavel@ucw.cz>

> @@ -0,0 +1,135 @@
> +/*
> + * linux/kernel/power/hibernation_nvs.c
> + *
> + * Copyright (C) 2008,2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
> + *
> + * Routines for NVS memory handling
> + */

Specifying  GPLv2 here would be nice. If it contains copyright, it
should contain license, too.
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [patch 05/38] s390: hibernation support for s390

[Pavel Machek]

On Tue 2009-06-09 15:34:02, Hans-Joachim Picht wrote:
> On Mon, 8 Jun 2009 08:44:51 +0200
> Pavel Machek <pavel@ucw.cz> wrote:
> 
> 
> > > Index: linux-2.6/arch/s390/power/swsusp.c
> > > ===================================================================
> > > --- /dev/null
> > > +++ linux-2.6/arch/s390/power/swsusp.c
> > > @@ -0,0 +1,31 @@
> > > +/*
> > > + * Support for suspend and resume on s390
> > > + *
> > > + * Copyright IBM Corp. 2009
> > > + *
> > > + * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> > > + *
> > > + */
> > > +
> > > +#include <linux/suspend.h>
> > > +#include <linux/gfp.h>
> > > +#include <linux/cpu.h>
> > > +#include <linux/smp.h>
> > > +#include <linux/cpumask.h>
> > > +#include <asm/ipl.h>
> > > +#include <asm/processor.h>
> > 
> > Empty functions do not need 1000 includes...
> 
> Thanks for spotting this. I will remove these.
>  
> > > +/*
> > > + * save CPU registers before creating a hibernation image and
> > > before
> > > + * restoring the memory state from it
> > > + */
> > > +void save_processor_state(void)
> > > +{
> > > +}
> > > +
> > > +/*
> > > + * restore the contents of CPU registers
> > > + */
> > > +void restore_processor_state(void)
> > > +{
> > > +}
> > 
> > ...and explaining why the functions can be emmpty would be nice.
> 
> I integrated the code for this in the assembly backend in the
> swsusp_arch_suspend & swsusp_arch_resume functions.

I guessed so. Maybe it is worth a comment?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [patch 05/38] s390: hibernation support for s390

[Pavel Machek]

On Tue 2009-06-09 15:34:02, Hans-Joachim Picht wrote:
> On Mon, 8 Jun 2009 08:44:51 +0200
> Pavel Machek <pavel@ucw.cz> wrote:
> 
> 
> > > Index: linux-2.6/arch/s390/power/swsusp.c
> > > ===================================================================
> > > --- /dev/null
> > > +++ linux-2.6/arch/s390/power/swsusp.c
> > > @@ -0,0 +1,31 @@
> > > +/*
> > > + * Support for suspend and resume on s390
> > > + *
> > > + * Copyright IBM Corp. 2009
> > > + *
> > > + * Author(s): Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> > > + *
> > > + */
> > > +
> > > +#include <linux/suspend.h>
> > > +#include <linux/gfp.h>
> > > +#include <linux/cpu.h>
> > > +#include <linux/smp.h>
> > > +#include <linux/cpumask.h>
> > > +#include <asm/ipl.h>
> > > +#include <asm/processor.h>
> > 
> > Empty functions do not need 1000 includes...
> 
> Thanks for spotting this. I will remove these.
>  
> > > +/*
> > > + * save CPU registers before creating a hibernation image and
> > > before
> > > + * restoring the memory state from it
> > > + */
> > > +void save_processor_state(void)
> > > +{
> > > +}
> > > +
> > > +/*
> > > + * restore the contents of CPU registers
> > > + */
> > > +void restore_processor_state(void)
> > > +{
> > > +}
> > 
> > ...and explaining why the functions can be emmpty would be nice.
> 
> I integrated the code for this in the assembly backend in the
> swsusp_arch_suspend & swsusp_arch_resume functions.

I guessed so. Maybe it is worth a comment?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Tuesday 09 June 2009, Cornelia Huck wrote:
> The *_nvs_* routines in swsusp.c make use of the io*map()
> functions, which are only provided for HAS_IOMEM, thus
> breaking compilation if HAS_IOMEM is not set. Fix this
> by moving the *_nvs_* routines into hibernation_nvs.c, which
> is only compiled if HAS_IOMEM is set.
> 
> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

Thanks, I added the GPLv2 line to the header comment and changed the name
of the file to hibernate_nvs.c (to match the other changes in the works).

I'll carry out some compilation testing on it and put it into the tree shortly.

Best,
Rafael

> ---
>  include/linux/suspend.h        |   18 +++--
>  kernel/power/Kconfig           |    4 +
>  kernel/power/Makefile          |    1 
>  kernel/power/hibernation_nvs.c |  135 +++++++++++++++++++++++++++++++++++++++++
>  kernel/power/swsusp.c          |  122 -------------------------------------
>  5 files changed, 151 insertions(+), 129 deletions(-)
> 
> --- /dev/null
> +++ linux-suspend/kernel/power/hibernation_nvs.c
> @@ -0,0 +1,135 @@
> +/*
> + * linux/kernel/power/hibernation_nvs.c
> + *
> + * Copyright (C) 2008,2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
> + *
> + * Routines for NVS memory handling
> + */
> +
> +#include <linux/io.h>
> +#include <linux/kernel.h>
> +#include <linux/list.h>
> +#include <linux/mm.h>
> +#include <linux/suspend.h>
> +
> +/*
> + * Platforms, like ACPI, may want us to save some memory used by them during
> + * hibernation and to restore the contents of this memory during the subsequent
> + * resume.  The code below implements a mechanism allowing us to do that.
> + */
> +
> +struct nvs_page {
> +	unsigned long phys_start;
> +	unsigned int size;
> +	void *kaddr;
> +	void *data;
> +	struct list_head node;
> +};
> +
> +static LIST_HEAD(nvs_list);
> +
> +/**
> + *	hibernate_nvs_register - register platform NVS memory region to save
> + *	@start - physical address of the region
> + *	@size - size of the region
> + *
> + *	The NVS region need not be page-aligned (both ends) and we arrange
> + *	things so that the data from page-aligned addresses in this region will
> + *	be copied into separate RAM pages.
> + */
> +int hibernate_nvs_register(unsigned long start, unsigned long size)
> +{
> +	struct nvs_page *entry, *next;
> +
> +	while (size > 0) {
> +		unsigned int nr_bytes;
> +
> +		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
> +		if (!entry)
> +			goto Error;
> +
> +		list_add_tail(&entry->node, &nvs_list);
> +		entry->phys_start = start;
> +		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
> +		entry->size = (size < nr_bytes) ? size : nr_bytes;
> +
> +		start += entry->size;
> +		size -= entry->size;
> +	}
> +	return 0;
> +
> + Error:
> +	list_for_each_entry_safe(entry, next, &nvs_list, node) {
> +		list_del(&entry->node);
> +		kfree(entry);
> +	}
> +	return -ENOMEM;
> +}
> +
> +/**
> + *	hibernate_nvs_free - free data pages allocated for saving NVS regions
> + */
> +void hibernate_nvs_free(void)
> +{
> +	struct nvs_page *entry;
> +
> +	list_for_each_entry(entry, &nvs_list, node)
> +		if (entry->data) {
> +			free_page((unsigned long)entry->data);
> +			entry->data = NULL;
> +			if (entry->kaddr) {
> +				iounmap(entry->kaddr);
> +				entry->kaddr = NULL;
> +			}
> +		}
> +}
> +
> +/**
> + *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
> + */
> +int hibernate_nvs_alloc(void)
> +{
> +	struct nvs_page *entry;
> +
> +	list_for_each_entry(entry, &nvs_list, node) {
> +		entry->data = (void *)__get_free_page(GFP_KERNEL);
> +		if (!entry->data) {
> +			hibernate_nvs_free();
> +			return -ENOMEM;
> +		}
> +	}
> +	return 0;
> +}
> +
> +/**
> + *	hibernate_nvs_save - save NVS memory regions
> + */
> +void hibernate_nvs_save(void)
> +{
> +	struct nvs_page *entry;
> +
> +	printk(KERN_INFO "PM: Saving platform NVS memory\n");
> +
> +	list_for_each_entry(entry, &nvs_list, node)
> +		if (entry->data) {
> +			entry->kaddr = ioremap(entry->phys_start, entry->size);
> +			memcpy(entry->data, entry->kaddr, entry->size);
> +		}
> +}
> +
> +/**
> + *	hibernate_nvs_restore - restore NVS memory regions
> + *
> + *	This function is going to be called with interrupts disabled, so it
> + *	cannot iounmap the virtual addresses used to access the NVS region.
> + */
> +void hibernate_nvs_restore(void)
> +{
> +	struct nvs_page *entry;
> +
> +	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
> +
> +	list_for_each_entry(entry, &nvs_list, node)
> +		if (entry->data)
> +			memcpy(entry->kaddr, entry->data, entry->size);
> +}
> --- linux-suspend.orig/kernel/power/swsusp.c
> +++ linux-suspend/kernel/power/swsusp.c
> @@ -186,125 +186,3 @@ void swsusp_show_speed(struct timeval *s
>  			centisecs / 100, centisecs % 100,
>  			kps / 1000, (kps % 1000) / 10);
>  }
> -
> -/*
> - * Platforms, like ACPI, may want us to save some memory used by them during
> - * hibernation and to restore the contents of this memory during the subsequent
> - * resume.  The code below implements a mechanism allowing us to do that.
> - */
> -
> -struct nvs_page {
> -	unsigned long phys_start;
> -	unsigned int size;
> -	void *kaddr;
> -	void *data;
> -	struct list_head node;
> -};
> -
> -static LIST_HEAD(nvs_list);
> -
> -/**
> - *	hibernate_nvs_register - register platform NVS memory region to save
> - *	@start - physical address of the region
> - *	@size - size of the region
> - *
> - *	The NVS region need not be page-aligned (both ends) and we arrange
> - *	things so that the data from page-aligned addresses in this region will
> - *	be copied into separate RAM pages.
> - */
> -int hibernate_nvs_register(unsigned long start, unsigned long size)
> -{
> -	struct nvs_page *entry, *next;
> -
> -	while (size > 0) {
> -		unsigned int nr_bytes;
> -
> -		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
> -		if (!entry)
> -			goto Error;
> -
> -		list_add_tail(&entry->node, &nvs_list);
> -		entry->phys_start = start;
> -		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
> -		entry->size = (size < nr_bytes) ? size : nr_bytes;
> -
> -		start += entry->size;
> -		size -= entry->size;
> -	}
> -	return 0;
> -
> - Error:
> -	list_for_each_entry_safe(entry, next, &nvs_list, node) {
> -		list_del(&entry->node);
> -		kfree(entry);
> -	}
> -	return -ENOMEM;
> -}
> -
> -/**
> - *	hibernate_nvs_free - free data pages allocated for saving NVS regions
> - */
> -void hibernate_nvs_free(void)
> -{
> -	struct nvs_page *entry;
> -
> -	list_for_each_entry(entry, &nvs_list, node)
> -		if (entry->data) {
> -			free_page((unsigned long)entry->data);
> -			entry->data = NULL;
> -			if (entry->kaddr) {
> -				iounmap(entry->kaddr);
> -				entry->kaddr = NULL;
> -			}
> -		}
> -}
> -
> -/**
> - *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
> - */
> -int hibernate_nvs_alloc(void)
> -{
> -	struct nvs_page *entry;
> -
> -	list_for_each_entry(entry, &nvs_list, node) {
> -		entry->data = (void *)__get_free_page(GFP_KERNEL);
> -		if (!entry->data) {
> -			hibernate_nvs_free();
> -			return -ENOMEM;
> -		}
> -	}
> -	return 0;
> -}
> -
> -/**
> - *	hibernate_nvs_save - save NVS memory regions
> - */
> -void hibernate_nvs_save(void)
> -{
> -	struct nvs_page *entry;
> -
> -	printk(KERN_INFO "PM: Saving platform NVS memory\n");
> -
> -	list_for_each_entry(entry, &nvs_list, node)
> -		if (entry->data) {
> -			entry->kaddr = ioremap(entry->phys_start, entry->size);
> -			memcpy(entry->data, entry->kaddr, entry->size);
> -		}
> -}
> -
> -/**
> - *	hibernate_nvs_restore - restore NVS memory regions
> - *
> - *	This function is going to be called with interrupts disabled, so it
> - *	cannot iounmap the virtual addresses used to access the NVS region.
> - */
> -void hibernate_nvs_restore(void)
> -{
> -	struct nvs_page *entry;
> -
> -	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
> -
> -	list_for_each_entry(entry, &nvs_list, node)
> -		if (entry->data)
> -			memcpy(entry->kaddr, entry->data, entry->size);
> -}
> --- linux-suspend.orig/kernel/power/Kconfig
> +++ linux-suspend/kernel/power/Kconfig
> @@ -116,9 +116,13 @@ config SUSPEND_FREEZER
>  
>  	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
>  
> +config HIBERNATION_NVS
> +	bool
> +
>  config HIBERNATION
>  	bool "Hibernation (aka 'suspend to disk')"
>  	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
> +	select HIBERNATION_NVS if HAS_IOMEM
>  	---help---
>  	  Enable the suspend to disk (STD) functionality, which is usually
>  	  called "hibernation" in user interfaces.  STD checkpoints the
> --- linux-suspend.orig/kernel/power/Makefile
> +++ linux-suspend/kernel/power/Makefile
> @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
>  obj-$(CONFIG_PM_SLEEP)		+= console.o
>  obj-$(CONFIG_FREEZER)		+= process.o
>  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> +obj-$(CONFIG_HIBERNATION_NVS)	+= hibernation_nvs.o
>  
>  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
> --- linux-suspend.orig/include/linux/suspend.h
> +++ linux-suspend/include/linux/suspend.h
> @@ -245,11 +245,6 @@ extern unsigned long get_safe_page(gfp_t
>  
>  extern void hibernation_set_ops(struct platform_hibernation_ops *ops);
>  extern int hibernate(void);
> -extern int hibernate_nvs_register(unsigned long start, unsigned long size);
> -extern int hibernate_nvs_alloc(void);
> -extern void hibernate_nvs_free(void);
> -extern void hibernate_nvs_save(void);
> -extern void hibernate_nvs_restore(void);
>  extern bool system_entering_hibernation(void);
>  #else /* CONFIG_HIBERNATION */
>  static inline int swsusp_page_is_forbidden(struct page *p) { return 0; }
> @@ -258,6 +253,16 @@ static inline void swsusp_unset_page_fre
>  
>  static inline void hibernation_set_ops(struct platform_hibernation_ops *ops) {}
>  static inline int hibernate(void) { return -ENOSYS; }
> +static inline bool system_entering_hibernation(void) { return false; }
> +#endif /* CONFIG_HIBERNATION */
> +
> +#ifdef CONFIG_HIBERNATION_NVS
> +extern int hibernate_nvs_register(unsigned long start, unsigned long size);
> +extern int hibernate_nvs_alloc(void);
> +extern void hibernate_nvs_free(void);
> +extern void hibernate_nvs_save(void);
> +extern void hibernate_nvs_restore(void);
> +#else /* CONFIG_HIBERNATION_NVS */
>  static inline int hibernate_nvs_register(unsigned long a, unsigned long b)
>  {
>  	return 0;
> @@ -266,8 +271,7 @@ static inline int hibernate_nvs_alloc(vo
>  static inline void hibernate_nvs_free(void) {}
>  static inline void hibernate_nvs_save(void) {}
>  static inline void hibernate_nvs_restore(void) {}
> -static inline bool system_entering_hibernation(void) { return false; }
> -#endif /* CONFIG_HIBERNATION */
> +#endif /* CONFIG_HIBERNATION_NVS */
>  
>  #ifdef CONFIG_PM_SLEEP
>  void save_processor_state(void);
> 
> 


-- 
Everyone knows that debugging is twice as hard as writing a program
in the first place.  So if you're as clever as you can be when you write it,
how will you ever debug it? --- Brian Kernighan

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Tuesday 09 June 2009, Cornelia Huck wrote:
> The *_nvs_* routines in swsusp.c make use of the io*map()
> functions, which are only provided for HAS_IOMEM, thus
> breaking compilation if HAS_IOMEM is not set. Fix this
> by moving the *_nvs_* routines into hibernation_nvs.c, which
> is only compiled if HAS_IOMEM is set.
> 
> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

Thanks, I added the GPLv2 line to the header comment and changed the name
of the file to hibernate_nvs.c (to match the other changes in the works).

I'll carry out some compilation testing on it and put it into the tree shortly.

Best,
Rafael

> ---
>  include/linux/suspend.h        |   18 +++--
>  kernel/power/Kconfig           |    4 +
>  kernel/power/Makefile          |    1 
>  kernel/power/hibernation_nvs.c |  135 +++++++++++++++++++++++++++++++++++++++++
>  kernel/power/swsusp.c          |  122 -------------------------------------
>  5 files changed, 151 insertions(+), 129 deletions(-)
> 
> --- /dev/null
> +++ linux-suspend/kernel/power/hibernation_nvs.c
> @@ -0,0 +1,135 @@
> +/*
> + * linux/kernel/power/hibernation_nvs.c
> + *
> + * Copyright (C) 2008,2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
> + *
> + * Routines for NVS memory handling
> + */
> +
> +#include <linux/io.h>
> +#include <linux/kernel.h>
> +#include <linux/list.h>
> +#include <linux/mm.h>
> +#include <linux/suspend.h>
> +
> +/*
> + * Platforms, like ACPI, may want us to save some memory used by them during
> + * hibernation and to restore the contents of this memory during the subsequent
> + * resume.  The code below implements a mechanism allowing us to do that.
> + */
> +
> +struct nvs_page {
> +	unsigned long phys_start;
> +	unsigned int size;
> +	void *kaddr;
> +	void *data;
> +	struct list_head node;
> +};
> +
> +static LIST_HEAD(nvs_list);
> +
> +/**
> + *	hibernate_nvs_register - register platform NVS memory region to save
> + *	@start - physical address of the region
> + *	@size - size of the region
> + *
> + *	The NVS region need not be page-aligned (both ends) and we arrange
> + *	things so that the data from page-aligned addresses in this region will
> + *	be copied into separate RAM pages.
> + */
> +int hibernate_nvs_register(unsigned long start, unsigned long size)
> +{
> +	struct nvs_page *entry, *next;
> +
> +	while (size > 0) {
> +		unsigned int nr_bytes;
> +
> +		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
> +		if (!entry)
> +			goto Error;
> +
> +		list_add_tail(&entry->node, &nvs_list);
> +		entry->phys_start = start;
> +		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
> +		entry->size = (size < nr_bytes) ? size : nr_bytes;
> +
> +		start += entry->size;
> +		size -= entry->size;
> +	}
> +	return 0;
> +
> + Error:
> +	list_for_each_entry_safe(entry, next, &nvs_list, node) {
> +		list_del(&entry->node);
> +		kfree(entry);
> +	}
> +	return -ENOMEM;
> +}
> +
> +/**
> + *	hibernate_nvs_free - free data pages allocated for saving NVS regions
> + */
> +void hibernate_nvs_free(void)
> +{
> +	struct nvs_page *entry;
> +
> +	list_for_each_entry(entry, &nvs_list, node)
> +		if (entry->data) {
> +			free_page((unsigned long)entry->data);
> +			entry->data = NULL;
> +			if (entry->kaddr) {
> +				iounmap(entry->kaddr);
> +				entry->kaddr = NULL;
> +			}
> +		}
> +}
> +
> +/**
> + *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
> + */
> +int hibernate_nvs_alloc(void)
> +{
> +	struct nvs_page *entry;
> +
> +	list_for_each_entry(entry, &nvs_list, node) {
> +		entry->data = (void *)__get_free_page(GFP_KERNEL);
> +		if (!entry->data) {
> +			hibernate_nvs_free();
> +			return -ENOMEM;
> +		}
> +	}
> +	return 0;
> +}
> +
> +/**
> + *	hibernate_nvs_save - save NVS memory regions
> + */
> +void hibernate_nvs_save(void)
> +{
> +	struct nvs_page *entry;
> +
> +	printk(KERN_INFO "PM: Saving platform NVS memory\n");
> +
> +	list_for_each_entry(entry, &nvs_list, node)
> +		if (entry->data) {
> +			entry->kaddr = ioremap(entry->phys_start, entry->size);
> +			memcpy(entry->data, entry->kaddr, entry->size);
> +		}
> +}
> +
> +/**
> + *	hibernate_nvs_restore - restore NVS memory regions
> + *
> + *	This function is going to be called with interrupts disabled, so it
> + *	cannot iounmap the virtual addresses used to access the NVS region.
> + */
> +void hibernate_nvs_restore(void)
> +{
> +	struct nvs_page *entry;
> +
> +	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
> +
> +	list_for_each_entry(entry, &nvs_list, node)
> +		if (entry->data)
> +			memcpy(entry->kaddr, entry->data, entry->size);
> +}
> --- linux-suspend.orig/kernel/power/swsusp.c
> +++ linux-suspend/kernel/power/swsusp.c
> @@ -186,125 +186,3 @@ void swsusp_show_speed(struct timeval *s
>  			centisecs / 100, centisecs % 100,
>  			kps / 1000, (kps % 1000) / 10);
>  }
> -
> -/*
> - * Platforms, like ACPI, may want us to save some memory used by them during
> - * hibernation and to restore the contents of this memory during the subsequent
> - * resume.  The code below implements a mechanism allowing us to do that.
> - */
> -
> -struct nvs_page {
> -	unsigned long phys_start;
> -	unsigned int size;
> -	void *kaddr;
> -	void *data;
> -	struct list_head node;
> -};
> -
> -static LIST_HEAD(nvs_list);
> -
> -/**
> - *	hibernate_nvs_register - register platform NVS memory region to save
> - *	@start - physical address of the region
> - *	@size - size of the region
> - *
> - *	The NVS region need not be page-aligned (both ends) and we arrange
> - *	things so that the data from page-aligned addresses in this region will
> - *	be copied into separate RAM pages.
> - */
> -int hibernate_nvs_register(unsigned long start, unsigned long size)
> -{
> -	struct nvs_page *entry, *next;
> -
> -	while (size > 0) {
> -		unsigned int nr_bytes;
> -
> -		entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL);
> -		if (!entry)
> -			goto Error;
> -
> -		list_add_tail(&entry->node, &nvs_list);
> -		entry->phys_start = start;
> -		nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK);
> -		entry->size = (size < nr_bytes) ? size : nr_bytes;
> -
> -		start += entry->size;
> -		size -= entry->size;
> -	}
> -	return 0;
> -
> - Error:
> -	list_for_each_entry_safe(entry, next, &nvs_list, node) {
> -		list_del(&entry->node);
> -		kfree(entry);
> -	}
> -	return -ENOMEM;
> -}
> -
> -/**
> - *	hibernate_nvs_free - free data pages allocated for saving NVS regions
> - */
> -void hibernate_nvs_free(void)
> -{
> -	struct nvs_page *entry;
> -
> -	list_for_each_entry(entry, &nvs_list, node)
> -		if (entry->data) {
> -			free_page((unsigned long)entry->data);
> -			entry->data = NULL;
> -			if (entry->kaddr) {
> -				iounmap(entry->kaddr);
> -				entry->kaddr = NULL;
> -			}
> -		}
> -}
> -
> -/**
> - *	hibernate_nvs_alloc - allocate memory necessary for saving NVS regions
> - */
> -int hibernate_nvs_alloc(void)
> -{
> -	struct nvs_page *entry;
> -
> -	list_for_each_entry(entry, &nvs_list, node) {
> -		entry->data = (void *)__get_free_page(GFP_KERNEL);
> -		if (!entry->data) {
> -			hibernate_nvs_free();
> -			return -ENOMEM;
> -		}
> -	}
> -	return 0;
> -}
> -
> -/**
> - *	hibernate_nvs_save - save NVS memory regions
> - */
> -void hibernate_nvs_save(void)
> -{
> -	struct nvs_page *entry;
> -
> -	printk(KERN_INFO "PM: Saving platform NVS memory\n");
> -
> -	list_for_each_entry(entry, &nvs_list, node)
> -		if (entry->data) {
> -			entry->kaddr = ioremap(entry->phys_start, entry->size);
> -			memcpy(entry->data, entry->kaddr, entry->size);
> -		}
> -}
> -
> -/**
> - *	hibernate_nvs_restore - restore NVS memory regions
> - *
> - *	This function is going to be called with interrupts disabled, so it
> - *	cannot iounmap the virtual addresses used to access the NVS region.
> - */
> -void hibernate_nvs_restore(void)
> -{
> -	struct nvs_page *entry;
> -
> -	printk(KERN_INFO "PM: Restoring platform NVS memory\n");
> -
> -	list_for_each_entry(entry, &nvs_list, node)
> -		if (entry->data)
> -			memcpy(entry->kaddr, entry->data, entry->size);
> -}
> --- linux-suspend.orig/kernel/power/Kconfig
> +++ linux-suspend/kernel/power/Kconfig
> @@ -116,9 +116,13 @@ config SUSPEND_FREEZER
>  
>  	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
>  
> +config HIBERNATION_NVS
> +	bool
> +
>  config HIBERNATION
>  	bool "Hibernation (aka 'suspend to disk')"
>  	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
> +	select HIBERNATION_NVS if HAS_IOMEM
>  	---help---
>  	  Enable the suspend to disk (STD) functionality, which is usually
>  	  called "hibernation" in user interfaces.  STD checkpoints the
> --- linux-suspend.orig/kernel/power/Makefile
> +++ linux-suspend/kernel/power/Makefile
> @@ -7,5 +7,6 @@ obj-$(CONFIG_PM)		+= main.o
>  obj-$(CONFIG_PM_SLEEP)		+= console.o
>  obj-$(CONFIG_FREEZER)		+= process.o
>  obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
> +obj-$(CONFIG_HIBERNATION_NVS)	+= hibernation_nvs.o
>  
>  obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o
> --- linux-suspend.orig/include/linux/suspend.h
> +++ linux-suspend/include/linux/suspend.h
> @@ -245,11 +245,6 @@ extern unsigned long get_safe_page(gfp_t
>  
>  extern void hibernation_set_ops(struct platform_hibernation_ops *ops);
>  extern int hibernate(void);
> -extern int hibernate_nvs_register(unsigned long start, unsigned long size);
> -extern int hibernate_nvs_alloc(void);
> -extern void hibernate_nvs_free(void);
> -extern void hibernate_nvs_save(void);
> -extern void hibernate_nvs_restore(void);
>  extern bool system_entering_hibernation(void);
>  #else /* CONFIG_HIBERNATION */
>  static inline int swsusp_page_is_forbidden(struct page *p) { return 0; }
> @@ -258,6 +253,16 @@ static inline void swsusp_unset_page_fre
>  
>  static inline void hibernation_set_ops(struct platform_hibernation_ops *ops) {}
>  static inline int hibernate(void) { return -ENOSYS; }
> +static inline bool system_entering_hibernation(void) { return false; }
> +#endif /* CONFIG_HIBERNATION */
> +
> +#ifdef CONFIG_HIBERNATION_NVS
> +extern int hibernate_nvs_register(unsigned long start, unsigned long size);
> +extern int hibernate_nvs_alloc(void);
> +extern void hibernate_nvs_free(void);
> +extern void hibernate_nvs_save(void);
> +extern void hibernate_nvs_restore(void);
> +#else /* CONFIG_HIBERNATION_NVS */
>  static inline int hibernate_nvs_register(unsigned long a, unsigned long b)
>  {
>  	return 0;
> @@ -266,8 +271,7 @@ static inline int hibernate_nvs_alloc(vo
>  static inline void hibernate_nvs_free(void) {}
>  static inline void hibernate_nvs_save(void) {}
>  static inline void hibernate_nvs_restore(void) {}
> -static inline bool system_entering_hibernation(void) { return false; }
> -#endif /* CONFIG_HIBERNATION */
> +#endif /* CONFIG_HIBERNATION_NVS */
>  
>  #ifdef CONFIG_PM_SLEEP
>  void save_processor_state(void);
> 
> 


-- 
Everyone knows that debugging is twice as hard as writing a program
in the first place.  So if you're as clever as you can be when you write it,
how will you ever debug it? --- Brian Kernighan

Re: [patch 05/38] s390: hibernation support for s390

[Hans-Joachim Picht]

On Tue, 9 Jun 2009 21:59:03 +0200
Pavel Machek <pavel@ucw.cz> wrote:

> > > > +/*
> > > > + * save CPU registers before creating a hibernation image and
> > > > before
> > > > + * restoring the memory state from it
> > > > + */
> > > > +void save_processor_state(void)
> > > > +{
> > > > +}
> > > > +
> > > > +/*
> > > > + * restore the contents of CPU registers
> > > > + */
> > > > +void restore_processor_state(void)
> > > > +{
> > > > +}
> > > 
> > > ...and explaining why the functions can be emmpty would be nice.
> > 
> > I integrated the code for this in the assembly backend in the
> > swsusp_arch_suspend & swsusp_arch_resume functions.
> 
> I guessed so. Maybe it is worth a comment?

Sure ;-)

Re: [patch 05/38] s390: hibernation support for s390

[Hans-Joachim Picht]

On Tue, 9 Jun 2009 21:59:03 +0200
Pavel Machek <pavel@ucw.cz> wrote:

> > > > +/*
> > > > + * save CPU registers before creating a hibernation image and
> > > > before
> > > > + * restoring the memory state from it
> > > > + */
> > > > +void save_processor_state(void)
> > > > +{
> > > > +}
> > > > +
> > > > +/*
> > > > + * restore the contents of CPU registers
> > > > + */
> > > > +void restore_processor_state(void)
> > > > +{
> > > > +}
> > > 
> > > ...and explaining why the functions can be emmpty would be nice.
> > 
> > I integrated the code for this in the assembly backend in the
> > swsusp_arch_suspend & swsusp_arch_resume functions.
> 
> I guessed so. Maybe it is worth a comment?

Sure ;-)

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Heiko Carstens]

On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> On Tuesday 09 June 2009, Cornelia Huck wrote:
> > The *_nvs_* routines in swsusp.c make use of the io*map()
> > functions, which are only provided for HAS_IOMEM, thus
> > breaking compilation if HAS_IOMEM is not set. Fix this
> > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > is only compiled if HAS_IOMEM is set.
> > 
> > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> 
> Thanks, I added the GPLv2 line to the header comment and changed the name
> of the file to hibernate_nvs.c (to match the other changes in the works).
> 
> I'll carry out some compilation testing on it and put it into the tree shortly.

Rafael, could you add the patch below as well?
Or should that go in via git390?

Subject: [PATCH] PM: add empty suspend/resume device irq functions

From: Heiko Carstens <heiko.carstens@de.ibm.com>

git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
device interrupts" introduced some helper functions. However these
functions are only available for architectures which support
GENERIC_HARDIRQS.

Other architectures will see this build error:

drivers/built-in.o: In function `sysdev_suspend':
(.text+0x15138): undefined reference to `check_wakeup_irqs'
drivers/built-in.o: In function `device_power_up':
(.text+0x1cb66): undefined reference to `resume_device_irqs'
drivers/built-in.o: In function `device_power_down':
(.text+0x1cb92): undefined reference to `suspend_device_irqs'

To fix this add some empty inline functions for !GENERIC_HARDIRQS.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 include/linux/interrupt.h |    6 ++++++
 1 file changed, 6 insertions(+)

diff -urpN linux-2.6/include/linux/interrupt.h linux-2.6-patched/include/linux/interrupt.h
--- linux-2.6/include/linux/interrupt.h	2009-06-11 13:07:46.000000000 +0200
+++ linux-2.6-patched/include/linux/interrupt.h	2009-06-11 13:08:31.000000000 +0200
@@ -183,6 +183,7 @@ extern void disable_irq(unsigned int irq
 extern void enable_irq(unsigned int irq);
 
 /* The following three functions are for the core kernel use only. */
+#ifdef CONFIG_GENERIC_HARDIRQS
 extern void suspend_device_irqs(void);
 extern void resume_device_irqs(void);
 #ifdef CONFIG_PM_SLEEP
@@ -190,6 +191,11 @@ extern int check_wakeup_irqs(void);
 #else
 static inline int check_wakeup_irqs(void) { return 0; }
 #endif
+#else
+static inline void suspend_device_irqs(void) { };
+static inline void resume_device_irqs(void) { };
+static inline int check_wakeup_irqs(void) { return 0; }
+#endif
 
 #if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS)
 

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Heiko Carstens]

On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> On Tuesday 09 June 2009, Cornelia Huck wrote:
> > The *_nvs_* routines in swsusp.c make use of the io*map()
> > functions, which are only provided for HAS_IOMEM, thus
> > breaking compilation if HAS_IOMEM is not set. Fix this
> > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > is only compiled if HAS_IOMEM is set.
> > 
> > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> 
> Thanks, I added the GPLv2 line to the header comment and changed the name
> of the file to hibernate_nvs.c (to match the other changes in the works).
> 
> I'll carry out some compilation testing on it and put it into the tree shortly.

Rafael, could you add the patch below as well?
Or should that go in via git390?

Subject: [PATCH] PM: add empty suspend/resume device irq functions

From: Heiko Carstens <heiko.carstens@de.ibm.com>

git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
device interrupts" introduced some helper functions. However these
functions are only available for architectures which support
GENERIC_HARDIRQS.

Other architectures will see this build error:

drivers/built-in.o: In function `sysdev_suspend':
(.text+0x15138): undefined reference to `check_wakeup_irqs'
drivers/built-in.o: In function `device_power_up':
(.text+0x1cb66): undefined reference to `resume_device_irqs'
drivers/built-in.o: In function `device_power_down':
(.text+0x1cb92): undefined reference to `suspend_device_irqs'

To fix this add some empty inline functions for !GENERIC_HARDIRQS.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 include/linux/interrupt.h |    6 ++++++
 1 file changed, 6 insertions(+)

diff -urpN linux-2.6/include/linux/interrupt.h linux-2.6-patched/include/linux/interrupt.h
--- linux-2.6/include/linux/interrupt.h	2009-06-11 13:07:46.000000000 +0200
+++ linux-2.6-patched/include/linux/interrupt.h	2009-06-11 13:08:31.000000000 +0200
@@ -183,6 +183,7 @@ extern void disable_irq(unsigned int irq
 extern void enable_irq(unsigned int irq);
 
 /* The following three functions are for the core kernel use only. */
+#ifdef CONFIG_GENERIC_HARDIRQS
 extern void suspend_device_irqs(void);
 extern void resume_device_irqs(void);
 #ifdef CONFIG_PM_SLEEP
@@ -190,6 +191,11 @@ extern int check_wakeup_irqs(void);
 #else
 static inline int check_wakeup_irqs(void) { return 0; }
 #endif
+#else
+static inline void suspend_device_irqs(void) { };
+static inline void resume_device_irqs(void) { };
+static inline int check_wakeup_irqs(void) { return 0; }
+#endif
 
 #if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS)
 

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Thursday 11 June 2009, Heiko Carstens wrote:
> On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > functions, which are only provided for HAS_IOMEM, thus
> > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > is only compiled if HAS_IOMEM is set.
> > > 
> > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > 
> > Thanks, I added the GPLv2 line to the header comment and changed the name
> > of the file to hibernate_nvs.c (to match the other changes in the works).
> > 
> > I'll carry out some compilation testing on it and put it into the tree shortly.
> 
> Rafael, could you add the patch below as well?
> Or should that go in via git390?
> 
> Subject: [PATCH] PM: add empty suspend/resume device irq functions
> 
> From: Heiko Carstens <heiko.carstens@de.ibm.com>
> 
> git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> device interrupts" introduced some helper functions. However these
> functions are only available for architectures which support
> GENERIC_HARDIRQS.
> 
> Other architectures will see this build error:
> 
> drivers/built-in.o: In function `sysdev_suspend':
> (.text+0x15138): undefined reference to `check_wakeup_irqs'
> drivers/built-in.o: In function `device_power_up':
> (.text+0x1cb66): undefined reference to `resume_device_irqs'
> drivers/built-in.o: In function `device_power_down':
> (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> 
> To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> 
> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
> ---
> 
>  include/linux/interrupt.h |    6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff -urpN linux-2.6/include/linux/interrupt.h linux-2.6-patched/include/linux/interrupt.h
> --- linux-2.6/include/linux/interrupt.h	2009-06-11 13:07:46.000000000 +0200
> +++ linux-2.6-patched/include/linux/interrupt.h	2009-06-11 13:08:31.000000000 +0200
> @@ -183,6 +183,7 @@ extern void disable_irq(unsigned int irq
>  extern void enable_irq(unsigned int irq);
>  
>  /* The following three functions are for the core kernel use only. */
> +#ifdef CONFIG_GENERIC_HARDIRQS
>  extern void suspend_device_irqs(void);
>  extern void resume_device_irqs(void);
>  #ifdef CONFIG_PM_SLEEP
> @@ -190,6 +191,11 @@ extern int check_wakeup_irqs(void);
>  #else
>  static inline int check_wakeup_irqs(void) { return 0; }
>  #endif
> +#else
> +static inline void suspend_device_irqs(void) { };
> +static inline void resume_device_irqs(void) { };
> +static inline int check_wakeup_irqs(void) { return 0; }
> +#endif
>  
>  #if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS)

Added, thanks a lot for the patch!

Best,
Rafael

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Thursday 11 June 2009, Heiko Carstens wrote:
> On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > functions, which are only provided for HAS_IOMEM, thus
> > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > is only compiled if HAS_IOMEM is set.
> > > 
> > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > 
> > Thanks, I added the GPLv2 line to the header comment and changed the name
> > of the file to hibernate_nvs.c (to match the other changes in the works).
> > 
> > I'll carry out some compilation testing on it and put it into the tree shortly.
> 
> Rafael, could you add the patch below as well?
> Or should that go in via git390?
> 
> Subject: [PATCH] PM: add empty suspend/resume device irq functions
> 
> From: Heiko Carstens <heiko.carstens@de.ibm.com>
> 
> git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> device interrupts" introduced some helper functions. However these
> functions are only available for architectures which support
> GENERIC_HARDIRQS.
> 
> Other architectures will see this build error:
> 
> drivers/built-in.o: In function `sysdev_suspend':
> (.text+0x15138): undefined reference to `check_wakeup_irqs'
> drivers/built-in.o: In function `device_power_up':
> (.text+0x1cb66): undefined reference to `resume_device_irqs'
> drivers/built-in.o: In function `device_power_down':
> (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> 
> To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> 
> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
> ---
> 
>  include/linux/interrupt.h |    6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff -urpN linux-2.6/include/linux/interrupt.h linux-2.6-patched/include/linux/interrupt.h
> --- linux-2.6/include/linux/interrupt.h	2009-06-11 13:07:46.000000000 +0200
> +++ linux-2.6-patched/include/linux/interrupt.h	2009-06-11 13:08:31.000000000 +0200
> @@ -183,6 +183,7 @@ extern void disable_irq(unsigned int irq
>  extern void enable_irq(unsigned int irq);
>  
>  /* The following three functions are for the core kernel use only. */
> +#ifdef CONFIG_GENERIC_HARDIRQS
>  extern void suspend_device_irqs(void);
>  extern void resume_device_irqs(void);
>  #ifdef CONFIG_PM_SLEEP
> @@ -190,6 +191,11 @@ extern int check_wakeup_irqs(void);
>  #else
>  static inline int check_wakeup_irqs(void) { return 0; }
>  #endif
> +#else
> +static inline void suspend_device_irqs(void) { };
> +static inline void resume_device_irqs(void) { };
> +static inline int check_wakeup_irqs(void) { return 0; }
> +#endif
>  
>  #if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS)

Added, thanks a lot for the patch!

Best,
Rafael

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > functions, which are only provided for HAS_IOMEM, thus
> > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > is only compiled if HAS_IOMEM is set.
> > > 
> > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > 
> > Thanks, I added the GPLv2 line to the header comment and changed the name
> > of the file to hibernate_nvs.c (to match the other changes in the works).
> > 
> > I'll carry out some compilation testing on it and put it into the tree shortly.
> 
> Rafael, could you add the patch below as well?
> Or should that go in via git390?
> 
> Subject: [PATCH] PM: add empty suspend/resume device irq functions
> 
> From: Heiko Carstens <heiko.carstens@de.ibm.com>
> 
> git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> device interrupts" introduced some helper functions. However these
> functions are only available for architectures which support
> GENERIC_HARDIRQS.
> 
> Other architectures will see this build error:
> 
> drivers/built-in.o: In function `sysdev_suspend':
> (.text+0x15138): undefined reference to `check_wakeup_irqs'
> drivers/built-in.o: In function `device_power_up':
> (.text+0x1cb66): undefined reference to `resume_device_irqs'
> drivers/built-in.o: In function `device_power_down':
> (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> 
> To fix this add some empty inline functions for !GENERIC_HARDIRQS.

I don't think that's right fix. If architecture does not use
GENERIC_HARDIRQS, it may want to implement *_device_irqs()
itself. Before your patch, it could, after your patch, it can not.

Better put those empty functions in arch/s390/include?
								Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > functions, which are only provided for HAS_IOMEM, thus
> > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > is only compiled if HAS_IOMEM is set.
> > > 
> > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > 
> > Thanks, I added the GPLv2 line to the header comment and changed the name
> > of the file to hibernate_nvs.c (to match the other changes in the works).
> > 
> > I'll carry out some compilation testing on it and put it into the tree shortly.
> 
> Rafael, could you add the patch below as well?
> Or should that go in via git390?
> 
> Subject: [PATCH] PM: add empty suspend/resume device irq functions
> 
> From: Heiko Carstens <heiko.carstens@de.ibm.com>
> 
> git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> device interrupts" introduced some helper functions. However these
> functions are only available for architectures which support
> GENERIC_HARDIRQS.
> 
> Other architectures will see this build error:
> 
> drivers/built-in.o: In function `sysdev_suspend':
> (.text+0x15138): undefined reference to `check_wakeup_irqs'
> drivers/built-in.o: In function `device_power_up':
> (.text+0x1cb66): undefined reference to `resume_device_irqs'
> drivers/built-in.o: In function `device_power_down':
> (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> 
> To fix this add some empty inline functions for !GENERIC_HARDIRQS.

I don't think that's right fix. If architecture does not use
GENERIC_HARDIRQS, it may want to implement *_device_irqs()
itself. Before your patch, it could, after your patch, it can not.

Better put those empty functions in arch/s390/include?
								Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Thursday 11 June 2009, Pavel Machek wrote:
> On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> > On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > > functions, which are only provided for HAS_IOMEM, thus
> > > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > > is only compiled if HAS_IOMEM is set.
> > > > 
> > > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > 
> > > Thanks, I added the GPLv2 line to the header comment and changed the name
> > > of the file to hibernate_nvs.c (to match the other changes in the works).
> > > 
> > > I'll carry out some compilation testing on it and put it into the tree shortly.
> > 
> > Rafael, could you add the patch below as well?
> > Or should that go in via git390?
> > 
> > Subject: [PATCH] PM: add empty suspend/resume device irq functions
> > 
> > From: Heiko Carstens <heiko.carstens@de.ibm.com>
> > 
> > git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> > device interrupts" introduced some helper functions. However these
> > functions are only available for architectures which support
> > GENERIC_HARDIRQS.
> > 
> > Other architectures will see this build error:
> > 
> > drivers/built-in.o: In function `sysdev_suspend':
> > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > drivers/built-in.o: In function `device_power_up':
> > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > drivers/built-in.o: In function `device_power_down':
> > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > 
> > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> 
> I don't think that's right fix. If architecture does not use
> GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> itself. Before your patch, it could, after your patch, it can not.
> 
> Better put those empty functions in arch/s390/include?

If any of the affected architectures wants to implement *_device_irqs()
itself, it can do the appropriate change in future.  For now, let's not break
compilation on them, shall we?

Thanks,
Rafael

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Thursday 11 June 2009, Pavel Machek wrote:
> On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> > On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > > functions, which are only provided for HAS_IOMEM, thus
> > > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > > is only compiled if HAS_IOMEM is set.
> > > > 
> > > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > 
> > > Thanks, I added the GPLv2 line to the header comment and changed the name
> > > of the file to hibernate_nvs.c (to match the other changes in the works).
> > > 
> > > I'll carry out some compilation testing on it and put it into the tree shortly.
> > 
> > Rafael, could you add the patch below as well?
> > Or should that go in via git390?
> > 
> > Subject: [PATCH] PM: add empty suspend/resume device irq functions
> > 
> > From: Heiko Carstens <heiko.carstens@de.ibm.com>
> > 
> > git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> > device interrupts" introduced some helper functions. However these
> > functions are only available for architectures which support
> > GENERIC_HARDIRQS.
> > 
> > Other architectures will see this build error:
> > 
> > drivers/built-in.o: In function `sysdev_suspend':
> > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > drivers/built-in.o: In function `device_power_up':
> > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > drivers/built-in.o: In function `device_power_down':
> > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > 
> > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> 
> I don't think that's right fix. If architecture does not use
> GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> itself. Before your patch, it could, after your patch, it can not.
> 
> Better put those empty functions in arch/s390/include?

If any of the affected architectures wants to implement *_device_irqs()
itself, it can do the appropriate change in future.  For now, let's not break
compilation on them, shall we?

Thanks,
Rafael

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Thu 2009-06-11 23:46:27, Rafael J. Wysocki wrote:
> On Thursday 11 June 2009, Pavel Machek wrote:
> > On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> > > On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > > > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > > > functions, which are only provided for HAS_IOMEM, thus
> > > > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > > > is only compiled if HAS_IOMEM is set.
> > > > > 
> > > > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > > 
> > > > Thanks, I added the GPLv2 line to the header comment and changed the name
> > > > of the file to hibernate_nvs.c (to match the other changes in the works).
> > > > 
> > > > I'll carry out some compilation testing on it and put it into the tree shortly.
> > > 
> > > Rafael, could you add the patch below as well?
> > > Or should that go in via git390?
> > > 
> > > Subject: [PATCH] PM: add empty suspend/resume device irq functions
> > > 
> > > From: Heiko Carstens <heiko.carstens@de.ibm.com>
> > > 
> > > git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> > > device interrupts" introduced some helper functions. However these
> > > functions are only available for architectures which support
> > > GENERIC_HARDIRQS.
> > > 
> > > Other architectures will see this build error:
> > > 
> > > drivers/built-in.o: In function `sysdev_suspend':
> > > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > > drivers/built-in.o: In function `device_power_up':
> > > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > > drivers/built-in.o: In function `device_power_down':
> > > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > > 
> > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > 
> > I don't think that's right fix. If architecture does not use
> > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > itself. Before your patch, it could, after your patch, it can not.
> > 
> > Better put those empty functions in arch/s390/include?
> 
> If any of the affected architectures wants to implement *_device_irqs()
> itself, it can do the appropriate change in future.  For now, let's not break
> compilation on them, shall we?

Well, if one of those architectures will want to implement
*_device_irqs(), it will have to either modify s390, and all other
!GENERIC_HARDIRQS architectures. Not exactly easy task; better do it
right from beggining?
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

On Thu 2009-06-11 23:46:27, Rafael J. Wysocki wrote:
> On Thursday 11 June 2009, Pavel Machek wrote:
> > On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> > > On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > > > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > > > functions, which are only provided for HAS_IOMEM, thus
> > > > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > > > is only compiled if HAS_IOMEM is set.
> > > > > 
> > > > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > > 
> > > > Thanks, I added the GPLv2 line to the header comment and changed the name
> > > > of the file to hibernate_nvs.c (to match the other changes in the works).
> > > > 
> > > > I'll carry out some compilation testing on it and put it into the tree shortly.
> > > 
> > > Rafael, could you add the patch below as well?
> > > Or should that go in via git390?
> > > 
> > > Subject: [PATCH] PM: add empty suspend/resume device irq functions
> > > 
> > > From: Heiko Carstens <heiko.carstens@de.ibm.com>
> > > 
> > > git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> > > device interrupts" introduced some helper functions. However these
> > > functions are only available for architectures which support
> > > GENERIC_HARDIRQS.
> > > 
> > > Other architectures will see this build error:
> > > 
> > > drivers/built-in.o: In function `sysdev_suspend':
> > > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > > drivers/built-in.o: In function `device_power_up':
> > > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > > drivers/built-in.o: In function `device_power_down':
> > > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > > 
> > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > 
> > I don't think that's right fix. If architecture does not use
> > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > itself. Before your patch, it could, after your patch, it can not.
> > 
> > Better put those empty functions in arch/s390/include?
> 
> If any of the affected architectures wants to implement *_device_irqs()
> itself, it can do the appropriate change in future.  For now, let's not break
> compilation on them, shall we?

Well, if one of those architectures will want to implement
*_device_irqs(), it will have to either modify s390, and all other
!GENERIC_HARDIRQS architectures. Not exactly easy task; better do it
right from beggining?
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Friday 12 June 2009, Pavel Machek wrote:
> On Thu 2009-06-11 23:46:27, Rafael J. Wysocki wrote:
> > On Thursday 11 June 2009, Pavel Machek wrote:
> > > On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> > > > On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > > > > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > > > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > > > > functions, which are only provided for HAS_IOMEM, thus
> > > > > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > > > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > > > > is only compiled if HAS_IOMEM is set.
> > > > > > 
> > > > > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > > > 
> > > > > Thanks, I added the GPLv2 line to the header comment and changed the name
> > > > > of the file to hibernate_nvs.c (to match the other changes in the works).
> > > > > 
> > > > > I'll carry out some compilation testing on it and put it into the tree shortly.
> > > > 
> > > > Rafael, could you add the patch below as well?
> > > > Or should that go in via git390?
> > > > 
> > > > Subject: [PATCH] PM: add empty suspend/resume device irq functions
> > > > 
> > > > From: Heiko Carstens <heiko.carstens@de.ibm.com>
> > > > 
> > > > git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> > > > device interrupts" introduced some helper functions. However these
> > > > functions are only available for architectures which support
> > > > GENERIC_HARDIRQS.
> > > > 
> > > > Other architectures will see this build error:
> > > > 
> > > > drivers/built-in.o: In function `sysdev_suspend':
> > > > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > > > drivers/built-in.o: In function `device_power_up':
> > > > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > > > drivers/built-in.o: In function `device_power_down':
> > > > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > > > 
> > > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > > 
> > > I don't think that's right fix. If architecture does not use
> > > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > > itself. Before your patch, it could, after your patch, it can not.
> > > 
> > > Better put those empty functions in arch/s390/include?
> > 
> > If any of the affected architectures wants to implement *_device_irqs()
> > itself, it can do the appropriate change in future.  For now, let's not break
> > compilation on them, shall we?
> 
> Well, if one of those architectures will want to implement
> *_device_irqs(), it will have to either modify s390, and all other
> !GENERIC_HARDIRQS architectures.

Why will it?  I think it will be sufficient to modify the header changed by
this patch and the architecture in question.

Best,
Rafael

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Rafael J. Wysocki]

On Friday 12 June 2009, Pavel Machek wrote:
> On Thu 2009-06-11 23:46:27, Rafael J. Wysocki wrote:
> > On Thursday 11 June 2009, Pavel Machek wrote:
> > > On Thu 2009-06-11 15:32:18, Heiko Carstens wrote:
> > > > On Wed, Jun 10, 2009 at 01:09:19AM +0200, Rafael J. Wysocki wrote:
> > > > > On Tuesday 09 June 2009, Cornelia Huck wrote:
> > > > > > The *_nvs_* routines in swsusp.c make use of the io*map()
> > > > > > functions, which are only provided for HAS_IOMEM, thus
> > > > > > breaking compilation if HAS_IOMEM is not set. Fix this
> > > > > > by moving the *_nvs_* routines into hibernation_nvs.c, which
> > > > > > is only compiled if HAS_IOMEM is set.
> > > > > > 
> > > > > > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> > > > > 
> > > > > Thanks, I added the GPLv2 line to the header comment and changed the name
> > > > > of the file to hibernate_nvs.c (to match the other changes in the works).
> > > > > 
> > > > > I'll carry out some compilation testing on it and put it into the tree shortly.
> > > > 
> > > > Rafael, could you add the patch below as well?
> > > > Or should that go in via git390?
> > > > 
> > > > Subject: [PATCH] PM: add empty suspend/resume device irq functions
> > > > 
> > > > From: Heiko Carstens <heiko.carstens@de.ibm.com>
> > > > 
> > > > git commit 0a0c5168 "PM: Introduce functions for suspending and resuming
> > > > device interrupts" introduced some helper functions. However these
> > > > functions are only available for architectures which support
> > > > GENERIC_HARDIRQS.
> > > > 
> > > > Other architectures will see this build error:
> > > > 
> > > > drivers/built-in.o: In function `sysdev_suspend':
> > > > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > > > drivers/built-in.o: In function `device_power_up':
> > > > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > > > drivers/built-in.o: In function `device_power_down':
> > > > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > > > 
> > > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > > 
> > > I don't think that's right fix. If architecture does not use
> > > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > > itself. Before your patch, it could, after your patch, it can not.
> > > 
> > > Better put those empty functions in arch/s390/include?
> > 
> > If any of the affected architectures wants to implement *_device_irqs()
> > itself, it can do the appropriate change in future.  For now, let's not break
> > compilation on them, shall we?
> 
> Well, if one of those architectures will want to implement
> *_device_irqs(), it will have to either modify s390, and all other
> !GENERIC_HARDIRQS architectures.

Why will it?  I think it will be sufficient to modify the header changed by
this patch and the architecture in question.

Best,
Rafael

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

> > > > > Other architectures will see this build error:
> > > > > 
> > > > > drivers/built-in.o: In function `sysdev_suspend':
> > > > > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > > > > drivers/built-in.o: In function `device_power_up':
> > > > > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > > > > drivers/built-in.o: In function `device_power_down':
> > > > > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > > > > 
> > > > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > > > 
> > > > I don't think that's right fix. If architecture does not use
> > > > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > > > itself. Before your patch, it could, after your patch, it can not.
> > > > 
> > > > Better put those empty functions in arch/s390/include?
> > > 
> > > If any of the affected architectures wants to implement *_device_irqs()
> > > itself, it can do the appropriate change in future.  For now, let's not break
> > > compilation on them, shall we?
> > 
> > Well, if one of those architectures will want to implement
> > *_device_irqs(), it will have to either modify s390, and all other
> > !GENERIC_HARDIRQS architectures.
> 
> Why will it?  I think it will be sufficient to modify the header changed by
> this patch and the architecture in question.

Hmm, how? Putting #ifndef MY_ARCH into generic header? Inventing
CONFIG_NON_GENERIC_HARDIRQS_BUT_I_NEED_DEVICE_IRQS?
									Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

> > > > > Other architectures will see this build error:
> > > > > 
> > > > > drivers/built-in.o: In function `sysdev_suspend':
> > > > > (.text+0x15138): undefined reference to `check_wakeup_irqs'
> > > > > drivers/built-in.o: In function `device_power_up':
> > > > > (.text+0x1cb66): undefined reference to `resume_device_irqs'
> > > > > drivers/built-in.o: In function `device_power_down':
> > > > > (.text+0x1cb92): undefined reference to `suspend_device_irqs'
> > > > > 
> > > > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > > > 
> > > > I don't think that's right fix. If architecture does not use
> > > > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > > > itself. Before your patch, it could, after your patch, it can not.
> > > > 
> > > > Better put those empty functions in arch/s390/include?
> > > 
> > > If any of the affected architectures wants to implement *_device_irqs()
> > > itself, it can do the appropriate change in future.  For now, let's not break
> > > compilation on them, shall we?
> > 
> > Well, if one of those architectures will want to implement
> > *_device_irqs(), it will have to either modify s390, and all other
> > !GENERIC_HARDIRQS architectures.
> 
> Why will it?  I think it will be sufficient to modify the header changed by
> this patch and the architecture in question.

Hmm, how? Putting #ifndef MY_ARCH into generic header? Inventing
CONFIG_NON_GENERIC_HARDIRQS_BUT_I_NEED_DEVICE_IRQS?
									Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

> > > > > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > > > > 
> > > > > I don't think that's right fix. If architecture does not use
> > > > > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > > > > itself. Before your patch, it could, after your patch, it can not.
> > > > > 
> > > > > Better put those empty functions in arch/s390/include?
> > > > 
> > > > If any of the affected architectures wants to implement *_device_irqs()
> > > > itself, it can do the appropriate change in future.  For now, let's not break
> > > > compilation on them, shall we?
> > > 
> > > Well, if one of those architectures will want to implement
> > > *_device_irqs(), it will have to either modify s390, and all other
> > > !GENERIC_HARDIRQS architectures.
> > 
> > Why will it?  I think it will be sufficient to modify the header changed by
> > this patch and the architecture in question.
> 
> Hmm, how? Putting #ifndef MY_ARCH into generic header? Inventing
> CONFIG_NON_GENERIC_HARDIRQS_BUT_I_NEED_DEVICE_IRQS?

Maybe playing with attribute((weak)) is the cleanest solution?
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH v2] pm: Move nvs routines into a seperate file.

[Pavel Machek]

> > > > > > To fix this add some empty inline functions for !GENERIC_HARDIRQS.
> > > > > 
> > > > > I don't think that's right fix. If architecture does not use
> > > > > GENERIC_HARDIRQS, it may want to implement *_device_irqs()
> > > > > itself. Before your patch, it could, after your patch, it can not.
> > > > > 
> > > > > Better put those empty functions in arch/s390/include?
> > > > 
> > > > If any of the affected architectures wants to implement *_device_irqs()
> > > > itself, it can do the appropriate change in future.  For now, let's not break
> > > > compilation on them, shall we?
> > > 
> > > Well, if one of those architectures will want to implement
> > > *_device_irqs(), it will have to either modify s390, and all other
> > > !GENERIC_HARDIRQS architectures.
> > 
> > Why will it?  I think it will be sufficient to modify the header changed by
> > this patch and the architecture in question.
> 
> Hmm, how? Putting #ifndef MY_ARCH into generic header? Inventing
> CONFIG_NON_GENERIC_HARDIRQS_BUT_I_NEED_DEVICE_IRQS?

Maybe playing with attribute((weak)) is the cleanest solution?
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [patch 05/38] s390: hibernation support for s390

[Martin Schwidefsky]

On Mon, 8 Jun 2009 08:44:51 +0200
Pavel Machek <pavel@ucw.cz> wrote:

> On Thu 2009-06-04 18:18:52, Martin Schwidefsky wrote:
> > From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> > 
> > This patch introduces the hibernation backend support to the 
> > s390 architecture. Now it is possible to suspend a mainframe Linux 
> > guest using the following command:
> > 
> > echo disk > /sys/power/state
> 
> Should this one be last in series?
> 
> Otherwise no obvious problems...

Ok, I will move the base patch to the end of the patch queue.
Thanks for the review.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [patch 05/38] s390: hibernation support for s390

[Martin Schwidefsky]

On Mon, 8 Jun 2009 08:44:51 +0200
Pavel Machek <pavel@ucw.cz> wrote:

> On Thu 2009-06-04 18:18:52, Martin Schwidefsky wrote:
> > From: Hans-Joachim Picht <hans@linux.vnet.ibm.com>
> > 
> > This patch introduces the hibernation backend support to the 
> > s390 architecture. Now it is possible to suspend a mainframe Linux 
> > guest using the following command:
> > 
> > echo disk > /sys/power/state
> 
> Should this one be last in series?
> 
> Otherwise no obvious problems...

Ok, I will move the base patch to the end of the patch queue.
Thanks for the review.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.