Merge with rsync://fileserver/linux

Update to Linus latest

30 files changed, 1280 insertions, 459 deletions

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index ab79af84699..6600ee87f89 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -226,6 +226,8 @@ config WARN_STACK_SIZE
 	  This allows you to specify the maximum frame size a function may
 	  have without the compiler complaining about it.
 
+source "mm/Kconfig"
+
 comment "I/O subsystem configuration"
 
 config MACHCHK_WARNING
@@ -453,6 +455,14 @@ config NO_IDLE_HZ_INIT
 	  The HZ timer is switched off in idle by default. That means the
 	  HZ timer is already disabled at boot time.
 
+config KEXEC
+	bool "kexec system call (EXPERIMENTAL)"
+	depends on EXPERIMENTAL
+	help
+	  kexec is a system call that implements the ability to shutdown your
+	  current kernel, and to start another kernel.  It is like a reboot
+	  but is independent of hardware/microcode support.
+
 endmenu
 
 config PCMCIA
diff --git a/arch/s390/appldata/appldata_base.c b/arch/s390/appldata/appldata_base.c
index 01ae1964c93..c067435bae4 100644
--- a/arch/s390/appldata/appldata_base.c
+++ b/arch/s390/appldata/appldata_base.c
@@ -28,6 +28,7 @@
 //#include <linux/kernel_stat.h>
 #include <linux/notifier.h>
 #include <linux/cpu.h>
+#include <linux/workqueue.h>
 
 #include "appldata.h"
 
@@ -133,9 +134,12 @@ static int appldata_interval = APPLDATA_CPU_INTERVAL;
 static int appldata_timer_active;
 
 /*
- * Tasklet
+ * Work queue
  */
-static struct tasklet_struct appldata_tasklet_struct;
+static struct workqueue_struct *appldata_wq;
+static void appldata_work_fn(void *data);
+static DECLARE_WORK(appldata_work, appldata_work_fn, NULL);
+
 
 /*
  * Ops list
@@ -144,11 +148,11 @@ static DEFINE_SPINLOCK(appldata_ops_lock);
 static LIST_HEAD(appldata_ops_list);
 
 
-/************************* timer, tasklet, DIAG ******************************/
+/*************************** timer, work, DIAG *******************************/
 /*
  * appldata_timer_function()
  *
- * schedule tasklet and reschedule timer
+ * schedule work and reschedule timer
  */
 static void appldata_timer_function(unsigned long data, struct pt_regs *regs)
 {
@@ -157,22 +161,22 @@ static void appldata_timer_function(unsigned long data, struct pt_regs *regs)
 		atomic_read(&appldata_expire_count));
 	if (atomic_dec_and_test(&appldata_expire_count)) {
 		atomic_set(&appldata_expire_count, num_online_cpus());
-		tasklet_schedule((struct tasklet_struct *) data);
+		queue_work(appldata_wq, (struct work_struct *) data);
 	}
 }
 
 /*
- * appldata_tasklet_function()
+ * appldata_work_fn()
  *
  * call data gathering function for each (active) module
  */
-static void appldata_tasklet_function(unsigned long data)
+static void appldata_work_fn(void *data)
 {
 	struct list_head *lh;
 	struct appldata_ops *ops;
 	int i;
 
-	P_DEBUG("  -= Tasklet =-\n");
+	P_DEBUG("  -= Work Queue =-\n");
 	i = 0;
 	spin_lock(&appldata_ops_lock);
 	list_for_each(lh, &appldata_ops_list) {
@@ -231,7 +235,7 @@ static int appldata_diag(char record_nr, u16 function, unsigned long buffer,
 			: "=d" (ry) : "d" (&(appldata_parameter_list)) : "cc");
 	return (int) ry;
 }
-/********************** timer, tasklet, DIAG <END> ***************************/
+/************************ timer, work, DIAG <END> ****************************/
 
 
 /****************************** /proc stuff **********************************/
@@ -411,7 +415,7 @@ appldata_generic_handler(ctl_table *ctl, int write, struct file *filp,
 	struct list_head *lh;
 
 	found = 0;
-	spin_lock_bh(&appldata_ops_lock);
+	spin_lock(&appldata_ops_lock);
 	list_for_each(lh, &appldata_ops_list) {
 		tmp_ops = list_entry(lh, struct appldata_ops, list);
 		if (&tmp_ops->ctl_table[2] == ctl) {
@@ -419,15 +423,15 @@ appldata_generic_handler(ctl_table *ctl, int write, struct file *filp,
 		}
 	}
 	if (!found) {
-		spin_unlock_bh(&appldata_ops_lock);
+		spin_unlock(&appldata_ops_lock);
 		return -ENODEV;
 	}
 	ops = ctl->data;
 	if (!try_module_get(ops->owner)) {	// protect this function
-		spin_unlock_bh(&appldata_ops_lock);
+		spin_unlock(&appldata_ops_lock);
 		return -ENODEV;
 	}
-	spin_unlock_bh(&appldata_ops_lock);
+	spin_unlock(&appldata_ops_lock);
 
 	if (!*lenp || *ppos) {
 		*lenp = 0;
@@ -451,10 +455,11 @@ appldata_generic_handler(ctl_table *ctl, int write, struct file *filp,
 		return -EFAULT;
 	}
 
-	spin_lock_bh(&appldata_ops_lock);
+	spin_lock(&appldata_ops_lock);
 	if ((buf[0] == '1') && (ops->active == 0)) {
-		if (!try_module_get(ops->owner)) {	// protect tasklet
-			spin_unlock_bh(&appldata_ops_lock);
+		// protect work queue callback
+		if (!try_module_get(ops->owner)) {
+			spin_unlock(&appldata_ops_lock);
 			module_put(ops->owner);
 			return -ENODEV;
 		}
@@ -485,7 +490,7 @@ appldata_generic_handler(ctl_table *ctl, int write, struct file *filp,
 		}
 		module_put(ops->owner);
 	}
-	spin_unlock_bh(&appldata_ops_lock);
+	spin_unlock(&appldata_ops_lock);
 out:
 	*lenp = len;
 	*ppos += len;
@@ -529,7 +534,7 @@ int appldata_register_ops(struct appldata_ops *ops)
 	}
 	memset(ops->ctl_table, 0, 4*sizeof(struct ctl_table));
 
-	spin_lock_bh(&appldata_ops_lock);
+	spin_lock(&appldata_ops_lock);
 	list_for_each(lh, &appldata_ops_list) {
 		tmp_ops = list_entry(lh, struct appldata_ops, list);
 		P_DEBUG("register_ops loop: %i) name = %s, ctl = %i\n",
@@ -541,18 +546,18 @@ int appldata_register_ops(struct appldata_ops *ops)
 				APPLDATA_PROC_NAME_LENGTH) == 0) {
 			P_ERROR("Name \"%s\" already registered!\n", ops->name);
 			kfree(ops->ctl_table);
-			spin_unlock_bh(&appldata_ops_lock);
+			spin_unlock(&appldata_ops_lock);
 			return -EBUSY;
 		}
 		if (tmp_ops->ctl_nr == ops->ctl_nr) {
 			P_ERROR("ctl_nr %i already registered!\n", ops->ctl_nr);
 			kfree(ops->ctl_table);
-			spin_unlock_bh(&appldata_ops_lock);
+			spin_unlock(&appldata_ops_lock);
 			return -EBUSY;
 		}
 	}
 	list_add(&ops->list, &appldata_ops_list);
-	spin_unlock_bh(&appldata_ops_lock);
+	spin_unlock(&appldata_ops_lock);
 
 	ops->ctl_table[0].ctl_name = CTL_APPLDATA;
 	ops->ctl_table[0].procname = appldata_proc_name;
@@ -583,12 +588,12 @@ int appldata_register_ops(struct appldata_ops *ops)
  */
 void appldata_unregister_ops(struct appldata_ops *ops)
 {
-	spin_lock_bh(&appldata_ops_lock);
+	spin_lock(&appldata_ops_lock);
 	unregister_sysctl_table(ops->sysctl_header);
 	list_del(&ops->list);
 	kfree(ops->ctl_table);
 	ops->ctl_table = NULL;
-	spin_unlock_bh(&appldata_ops_lock);
+	spin_unlock(&appldata_ops_lock);
 	P_INFO("%s-ops unregistered!\n", ops->name);
 }
 /********************** module-ops management <END> **************************/
@@ -602,7 +607,7 @@ appldata_online_cpu(int cpu)
 	init_virt_timer(&per_cpu(appldata_timer, cpu));
 	per_cpu(appldata_timer, cpu).function = appldata_timer_function;
 	per_cpu(appldata_timer, cpu).data = (unsigned long)
-		&appldata_tasklet_struct;
+		&appldata_work;
 	atomic_inc(&appldata_expire_count);
 	spin_lock(&appldata_timer_lock);
 	__appldata_vtimer_setup(APPLDATA_MOD_TIMER);
@@ -615,7 +620,7 @@ appldata_offline_cpu(int cpu)
 	del_virt_timer(&per_cpu(appldata_timer, cpu));
 	if (atomic_dec_and_test(&appldata_expire_count)) {
 		atomic_set(&appldata_expire_count, num_online_cpus());
-		tasklet_schedule(&appldata_tasklet_struct);
+		queue_work(appldata_wq, &appldata_work);
 	}
 	spin_lock(&appldata_timer_lock);
 	__appldata_vtimer_setup(APPLDATA_MOD_TIMER);
@@ -648,7 +653,7 @@ static struct notifier_block __devinitdata appldata_nb = {
 /*
  * appldata_init()
  *
- * init timer and tasklet, register /proc entries
+ * init timer, register /proc entries
  */
 static int __init appldata_init(void)
 {
@@ -657,6 +662,12 @@ static int __init appldata_init(void)
 	P_DEBUG("sizeof(parameter_list) = %lu\n",
 		sizeof(struct appldata_parameter_list));
 
+	appldata_wq = create_singlethread_workqueue("appldata");
+	if (!appldata_wq) {
+		P_ERROR("Could not create work queue\n");
+		return -ENOMEM;
+	}
+
 	for_each_online_cpu(i)
 		appldata_online_cpu(i);
 
@@ -670,7 +681,6 @@ static int __init appldata_init(void)
 	appldata_table[1].de->owner = THIS_MODULE;
 #endif
 
-	tasklet_init(&appldata_tasklet_struct, appldata_tasklet_function, 0);
 	P_DEBUG("Base interface initialized.\n");
 	return 0;
 }
@@ -678,7 +688,7 @@ static int __init appldata_init(void)
 /*
  * appldata_exit()
  *
- * stop timer and tasklet, unregister /proc entries
+ * stop timer, unregister /proc entries
  */
 static void __exit appldata_exit(void)
 {
@@ -690,7 +700,7 @@ static void __exit appldata_exit(void)
 	/*
 	 * ops list should be empty, but just in case something went wrong...
 	 */
-	spin_lock_bh(&appldata_ops_lock);
+	spin_lock(&appldata_ops_lock);
 	list_for_each(lh, &appldata_ops_list) {
 		ops = list_entry(lh, struct appldata_ops, list);
 		rc = appldata_diag(ops->record_nr, APPLDATA_STOP_REC,
@@ -700,7 +710,7 @@ static void __exit appldata_exit(void)
 				"return code: %d\n", ops->name, rc);
 		}
 	}
-	spin_unlock_bh(&appldata_ops_lock);
+	spin_unlock(&appldata_ops_lock);
 
 	for_each_online_cpu(i)
 		appldata_offline_cpu(i);
@@ -709,7 +719,7 @@ static void __exit appldata_exit(void)
 
 	unregister_sysctl_table(appldata_sysctl_header);
 
-	tasklet_kill(&appldata_tasklet_struct);
+	destroy_workqueue(appldata_wq);
 	P_DEBUG("... module unloaded!\n");
 }
 /**************************** init / exit <END> ******************************/
diff --git a/arch/s390/appldata/appldata_mem.c b/arch/s390/appldata/appldata_mem.c
index 462ee9a84e7..f0e2fbed3d4 100644
--- a/arch/s390/appldata/appldata_mem.c
+++ b/arch/s390/appldata/appldata_mem.c
@@ -68,7 +68,7 @@ struct appldata_mem_data {
 	u64 pgmajfault;		/* page faults (major only) */
 // <-- New in 2.6
 
-} appldata_mem_data;
+} __attribute__((packed)) appldata_mem_data;
 
 
 static inline void appldata_debug_print(struct appldata_mem_data *mem_data)
diff --git a/arch/s390/appldata/appldata_net_sum.c b/arch/s390/appldata/appldata_net_sum.c
index dd61638d302..2a4c7432db4 100644
--- a/arch/s390/appldata/appldata_net_sum.c
+++ b/arch/s390/appldata/appldata_net_sum.c
@@ -57,7 +57,7 @@ struct appldata_net_sum_data {
 	u64 rx_dropped;		/* no space in linux buffers     */
 	u64 tx_dropped;		/* no space available in linux   */
 	u64 collisions;		/* collisions while transmitting */
-} appldata_net_sum_data;
+} __attribute__((packed)) appldata_net_sum_data;
 
 
 static inline void appldata_print_debug(struct appldata_net_sum_data *net_data)
diff --git a/arch/s390/appldata/appldata_os.c b/arch/s390/appldata/appldata_os.c
index b83f0748455..e0a476bf4fd 100644
--- a/arch/s390/appldata/appldata_os.c
+++ b/arch/s390/appldata/appldata_os.c
@@ -49,7 +49,7 @@ struct appldata_os_per_cpu {
 	u32 per_cpu_softirq;	/* ... spent in softirqs            */
 	u32 per_cpu_iowait;	/* ... spent while waiting for I/O  */
 // <-- New in 2.6
-};
+} __attribute__((packed));
 
 struct appldata_os_data {
 	u64 timestamp;
@@ -75,7 +75,7 @@ struct appldata_os_data {
 
 	/* per cpu data */
 	struct appldata_os_per_cpu os_cpu[0];
-};
+} __attribute__((packed));
 
 static struct appldata_os_data *appldata_os_data;
 
diff --git a/arch/s390/boot/install.sh b/arch/s390/boot/install.sh
index 278a8139cb1..d4026f62cb0 100644
--- a/arch/s390/boot/install.sh
+++ b/arch/s390/boot/install.sh
@@ -21,8 +21,8 @@
 
 # User may have a custom install script
 
-if [ -x ~/bin/installkernel ]; then exec ~/bin/installkernel "$@"; fi
-if [ -x /sbin/installkernel ]; then exec /sbin/installkernel "$@"; fi
+if [ -x ~/bin/${CROSS_COMPILE}installkernel ]; then exec ~/bin/${CROSS_COMPILE}installkernel "$@"; fi
+if [ -x /sbin/${CROSS_COMPILE}installkernel ]; then exec /sbin/${CROSS_COMPILE}installkernel "$@"; fi
 
 # Default install - same as make zlilo
 
diff --git a/arch/s390/defconfig b/arch/s390/defconfig
index 07fd0414a4b..89850b2c27e 100644
--- a/arch/s390/defconfig
+++ b/arch/s390/defconfig
@@ -245,6 +245,7 @@ CONFIG_S390_TAPE_BLOCK=y
 #
 CONFIG_S390_TAPE_34XX=m
 # CONFIG_VMLOGRDR is not set
+# CONFIG_VMCP is not set
 # CONFIG_MONREADER is not set
 # CONFIG_DCSS_SHM is not set
 
diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile
index b41e0e199a7..ab1e49d2e51 100644
--- a/arch/s390/kernel/Makefile
+++ b/arch/s390/kernel/Makefile
@@ -25,6 +25,16 @@ obj-$(CONFIG_ARCH_S390X)	+= entry64.o reipl64.o
 
 obj-$(CONFIG_VIRT_TIMER)	+= vtime.o
 
+# Kexec part
+S390_KEXEC_OBJS := machine_kexec.o crash.o
+ifeq ($(CONFIG_ARCH_S390X),y)
+S390_KEXEC_OBJS += relocate_kernel64.o
+else
+S390_KEXEC_OBJS += relocate_kernel.o
+endif
+obj-$(CONFIG_KEXEC) += $(S390_KEXEC_OBJS)
+
+
 #
 # This is just to get the dependencies...
 #
diff --git a/arch/s390/kernel/compat_ioctl.c b/arch/s390/kernel/compat_ioctl.c
index 03d03c6d3cb..24a1e9f069a 100644
--- a/arch/s390/kernel/compat_ioctl.c
+++ b/arch/s390/kernel/compat_ioctl.c
@@ -42,7 +42,6 @@ struct ioctl_trans ioctl_start[] = {
 #include "../../../fs/compat_ioctl.c"
 
 /* s390 only ioctls */
-#if defined(CONFIG_DASD) || defined(CONFIG_DASD_MODULE)
 COMPATIBLE_IOCTL(DASDAPIVER)
 COMPATIBLE_IOCTL(BIODASDDISABLE)
 COMPATIBLE_IOCTL(BIODASDENABLE)
@@ -59,16 +58,11 @@ COMPATIBLE_IOCTL(BIODASDPRRD)
 COMPATIBLE_IOCTL(BIODASDPSRD)
 COMPATIBLE_IOCTL(BIODASDGATTR)
 COMPATIBLE_IOCTL(BIODASDSATTR)
-#if defined(CONFIG_DASD_CMB) || defined(CONFIG_DASD_CMB_MODULE)
 COMPATIBLE_IOCTL(BIODASDCMFENABLE)
 COMPATIBLE_IOCTL(BIODASDCMFDISABLE)
 COMPATIBLE_IOCTL(BIODASDREADALLCMB)
-#endif
-#endif
 
-#if defined(CONFIG_S390_TAPE) || defined(CONFIG_S390_TAPE_MODULE)
 COMPATIBLE_IOCTL(TAPE390_DISPLAY)
-#endif
 
 /* s390 doesn't need handlers here */
 COMPATIBLE_IOCTL(TIOCGSERIAL)
diff --git a/arch/s390/kernel/compat_linux.h b/arch/s390/kernel/compat_linux.h
index bf33dcfec7d..3898f66d0b2 100644
--- a/arch/s390/kernel/compat_linux.h
+++ b/arch/s390/kernel/compat_linux.h
@@ -45,7 +45,7 @@ typedef struct compat_siginfo {
 
 		/* POSIX.1b timers */
 		struct {
-			timer_t _tid;		/* timer id */
+			compat_timer_t _tid;		/* timer id */
 			int _overrun;		/* overrun count */
 			compat_sigval_t _sigval;	/* same as below */
 			int _sys_private;       /* not to be passed to user */
diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S
index 7a607b1d038..bf529739c8a 100644
--- a/arch/s390/kernel/compat_wrapper.S
+++ b/arch/s390/kernel/compat_wrapper.S
@@ -1441,3 +1441,11 @@ compat_sys_waitid_wrapper:
 	lgfr	%r5,%r5			# int
 	llgtr	%r6,%r6			# struct rusage_emu31 *
 	jg	compat_sys_waitid
+
+	.globl	compat_sys_kexec_load_wrapper
+compat_sys_kexec_load_wrapper:
+	llgfr	%r2,%r2			# unsigned long
+	llgfr	%r3,%r3			# unsigned long
+	llgtr	%r4,%r4			# struct kexec_segment *
+	llgfr	%r5,%r5			# unsigned long
+	jg	compat_sys_kexec_load
diff --git a/arch/s390/kernel/cpcmd.c b/arch/s390/kernel/cpcmd.c
index 44df8dc07c5..20062145e84 100644
--- a/arch/s390/kernel/cpcmd.c
+++ b/arch/s390/kernel/cpcmd.c
@@ -2,7 +2,7 @@
  *  arch/s390/kernel/cpcmd.c
  *
  *  S390 version
- *    Copyright (C) 1999,2000 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Copyright (C) 1999,2005 IBM Deutschland Entwicklung GmbH, IBM Corporation
  *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
  *               Christian Borntraeger (cborntra@de.ibm.com),
  */
@@ -18,93 +18,114 @@
 #include <asm/system.h>
 
 static DEFINE_SPINLOCK(cpcmd_lock);
-static char cpcmd_buf[240];
+static char cpcmd_buf[241];
 
 /*
  * the caller of __cpcmd has to ensure that the response buffer is below 2 GB
  */
-void __cpcmd(char *cmd, char *response, int rlen)
+int  __cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 {
 	const int mask = 0x40000000L;
 	unsigned long flags;
+	int return_code;
+	int return_len;
 	int cmdlen;
 
 	spin_lock_irqsave(&cpcmd_lock, flags);
 	cmdlen = strlen(cmd);
 	BUG_ON(cmdlen > 240);
-	strcpy(cpcmd_buf, cmd);
+	memcpy(cpcmd_buf, cmd, cmdlen);
 	ASCEBC(cpcmd_buf, cmdlen);
 
 	if (response != NULL && rlen > 0) {
 		memset(response, 0, rlen);
 #ifndef CONFIG_ARCH_S390X
-		asm volatile ("LRA   2,0(%0)\n\t"
-                              "LR    4,%1\n\t"
-                              "O     4,%4\n\t"
-                              "LRA   3,0(%2)\n\t"
-                              "LR    5,%3\n\t"
-                              ".long 0x83240008 # Diagnose X'08'\n\t"
-                              : /* no output */
-                              : "a" (cpcmd_buf), "d" (cmdlen),
-                                "a" (response), "d" (rlen), "m" (mask)
-                              : "cc", "2", "3", "4", "5" );
+		asm volatile (	"lra	2,0(%2)\n"
+				"lr	4,%3\n"
+				"o	4,%6\n"
+				"lra	3,0(%4)\n"
+				"lr	5,%5\n"
+				"diag	2,4,0x8\n"
+				"brc	8, .Litfits\n"
+				"ar	5, %5\n"
+				".Litfits: \n"
+				"lr	%0,4\n"
+				"lr	%1,5\n"
+				: "=d" (return_code), "=d" (return_len)
+				: "a" (cpcmd_buf), "d" (cmdlen),
+				"a" (response), "d" (rlen), "m" (mask)
+				: "cc", "2", "3", "4", "5" );
 #else /* CONFIG_ARCH_S390X */
-                asm volatile ("   lrag  2,0(%0)\n"
-                              "   lgr   4,%1\n"
-                              "   o     4,%4\n"
-                              "   lrag  3,0(%2)\n"
-                              "   lgr   5,%3\n"
-                              "   sam31\n"
-                              "   .long 0x83240008 # Diagnose X'08'\n"
-                              "   sam64"
-                              : /* no output */
-                              : "a" (cpcmd_buf), "d" (cmdlen),
-                                "a" (response), "d" (rlen), "m" (mask)
-                              : "cc", "2", "3", "4", "5" );
+                asm volatile (	"lrag	2,0(%2)\n"
+				"lgr	4,%3\n"
+				"o	4,%6\n"
+				"lrag	3,0(%4)\n"
+				"lgr	5,%5\n"
+				"sam31\n"
+				"diag	2,4,0x8\n"
+				"sam64\n"
+				"brc	8, .Litfits\n"
+				"agr	5, %5\n"
+				".Litfits: \n"
+				"lgr	%0,4\n"
+				"lgr	%1,5\n"
+				: "=d" (return_code), "=d" (return_len)
+				: "a" (cpcmd_buf), "d" (cmdlen),
+				"a" (response), "d" (rlen), "m" (mask)
+				: "cc", "2", "3", "4", "5" );
 #endif /* CONFIG_ARCH_S390X */
                 EBCASC(response, rlen);
         } else {
+		return_len = 0;
 #ifndef CONFIG_ARCH_S390X
-                asm volatile ("LRA   2,0(%0)\n\t"
-                              "LR    3,%1\n\t"
-                              ".long 0x83230008 # Diagnose X'08'\n\t"
-                              : /* no output */
-                              : "a" (cpcmd_buf), "d" (cmdlen)
-                              : "2", "3"  );
+                asm volatile (	"lra	2,0(%1)\n"
+				"lr	3,%2\n"
+				"diag	2,3,0x8\n"
+				"lr	%0,3\n"
+				: "=d" (return_code)
+				: "a" (cpcmd_buf), "d" (cmdlen)
+				: "2", "3"  );
 #else /* CONFIG_ARCH_S390X */
-                asm volatile ("   lrag  2,0(%0)\n"
-                              "   lgr   3,%1\n"
-                              "   sam31\n"
-                              "   .long 0x83230008 # Diagnose X'08'\n"
-                              "   sam64"
-                              : /* no output */
-                              : "a" (cpcmd_buf), "d" (cmdlen)
-                              : "2", "3"  );
+                asm volatile (	"lrag	2,0(%1)\n"
+				"lgr	3,%2\n"
+				"sam31\n"
+				"diag	2,3,0x8\n"
+				"sam64\n"
+				"lgr	%0,3\n"
+				: "=d" (return_code)
+				: "a" (cpcmd_buf), "d" (cmdlen)
+				: "2", "3" );
 #endif /* CONFIG_ARCH_S390X */
         }
 	spin_unlock_irqrestore(&cpcmd_lock, flags);
+	if (response_code != NULL)
+		*response_code = return_code;
+	return return_len;
 }
 
 EXPORT_SYMBOL(__cpcmd);
 
 #ifdef CONFIG_ARCH_S390X
-void cpcmd(char *cmd, char *response, int rlen)
+int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 {
 	char *lowbuf;
+	int len;
+
 	if ((rlen == 0) || (response == NULL)
 	    || !((unsigned long)response >> 31))
-		__cpcmd(cmd, response, rlen);
+		len = __cpcmd(cmd, response, rlen, response_code);
 	else {
 		lowbuf = kmalloc(rlen, GFP_KERNEL | GFP_DMA);
 		if (!lowbuf) {
 			printk(KERN_WARNING
 				"cpcmd: could not allocate response buffer\n");
-			return;
+			return -ENOMEM;
 		}
-		__cpcmd(cmd, lowbuf, rlen);
+		len = __cpcmd(cmd, lowbuf, rlen, response_code);
 		memcpy(response, lowbuf, rlen);
 		kfree(lowbuf);
 	}
+	return len;
 }
 
 EXPORT_SYMBOL(cpcmd);
diff --git a/arch/s390/kernel/crash.c b/arch/s390/kernel/crash.c
new file mode 100644
index 00000000000..7bd169c58b0
--- /dev/null
+++ b/arch/s390/kernel/crash.c
@@ -0,0 +1,17 @@
+/*
+ * arch/s390/kernel/crash.c
+ *
+ * (C) Copyright IBM Corp. 2005
+ *
+ * Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ *
+ */
+
+#include <linux/threads.h>
+#include <linux/kexec.h>
+
+note_buf_t crash_notes[NR_CPUS];
+
+void machine_crash_shutdown(struct pt_regs *regs)
+{
+}
diff --git a/arch/s390/kernel/debug.c b/arch/s390/kernel/debug.c
index 91f8ce5543d..960ba6029c3 100644
--- a/arch/s390/kernel/debug.c
+++ b/arch/s390/kernel/debug.c
@@ -19,22 +19,27 @@
 #include <linux/sysctl.h>
 #include <asm/uaccess.h>
 #include <asm/semaphore.h>
-
 #include <linux/module.h>
 #include <linux/init.h>
+#include <linux/fs.h>
+#include <linux/debugfs.h>
 
 #include <asm/debug.h>
 
 #define DEBUG_PROLOG_ENTRY -1
 
+#define ALL_AREAS 0 /* copy all debug areas */
+#define NO_AREAS  1 /* copy no debug areas */
+
 /* typedefs */
 
 typedef struct file_private_info {
 	loff_t offset;			/* offset of last read in file */
 	int    act_area;                /* number of last formated area */
+	int    act_page;                /* act page in given area */
 	int    act_entry;               /* last formated entry (offset */
                                         /* relative to beginning of last */
-                                        /* formated area) */ 
+                                        /* formated page) */
 	size_t act_entry_offset;        /* up to this offset we copied */
 					/* in last read the last formated */
 					/* entry to userland */
@@ -51,8 +56,8 @@ typedef struct
 	 * This assumes that all args are converted into longs 
 	 * on L/390 this is the case for all types of parameter 
 	 * except of floats, and long long (32 bit) 
-         *
-         */
+	 *
+	 */
 	long args[0];
 } debug_sprintf_entry_t;
 
@@ -63,32 +68,38 @@ extern void tod_to_timeval(uint64_t todval, struct timeval *xtime);
 
 static int debug_init(void);
 static ssize_t debug_output(struct file *file, char __user *user_buf,
-			    size_t user_len, loff_t * offset);
+			size_t user_len, loff_t * offset);
 static ssize_t debug_input(struct file *file, const char __user *user_buf,
-			   size_t user_len, loff_t * offset);
+			size_t user_len, loff_t * offset);
 static int debug_open(struct inode *inode, struct file *file);
 static int debug_close(struct inode *inode, struct file *file);
-static debug_info_t*  debug_info_create(char *name, int page_order, int nr_areas, int buf_size);
+static debug_info_t*  debug_info_create(char *name, int pages_per_area,
+			int nr_areas, int buf_size);
 static void debug_info_get(debug_info_t *);
 static void debug_info_put(debug_info_t *);
 static int debug_prolog_level_fn(debug_info_t * id,
-				 struct debug_view *view, char *out_buf);
+			struct debug_view *view, char *out_buf);
 static int debug_input_level_fn(debug_info_t * id, struct debug_view *view,
-				struct file *file, const char __user *user_buf,
-				size_t user_buf_size, loff_t * offset);
+			struct file *file, const char __user *user_buf,
+			size_t user_buf_size, loff_t * offset);
+static int debug_prolog_pages_fn(debug_info_t * id,
+			struct debug_view *view, char *out_buf);
+static int debug_input_pages_fn(debug_info_t * id, struct debug_view *view,
+			struct file *file, const char __user *user_buf,
+			size_t user_buf_size, loff_t * offset);
 static int debug_input_flush_fn(debug_info_t * id, struct debug_view *view,
-                                struct file *file, const char __user *user_buf,
-                                size_t user_buf_size, loff_t * offset);
+			struct file *file, const char __user *user_buf,
+			size_t user_buf_size, loff_t * offset);
 static int debug_hex_ascii_format_fn(debug_info_t * id, struct debug_view *view,
-                                char *out_buf, const char *in_buf);
+			char *out_buf, const char *in_buf);
 static int debug_raw_format_fn(debug_info_t * id,
-				 struct debug_view *view, char *out_buf,
-				 const char *in_buf);
+			struct debug_view *view, char *out_buf,
+			const char *in_buf);
 static int debug_raw_header_fn(debug_info_t * id, struct debug_view *view,
-                         int area, debug_entry_t * entry, char *out_buf);
+			int area, debug_entry_t * entry, char *out_buf);
 
 static int debug_sprintf_format_fn(debug_info_t * id, struct debug_view *view,
-				   char *out_buf, debug_sprintf_entry_t *curr_event);
+			char *out_buf, debug_sprintf_entry_t *curr_event);
 
 /* globals */
 
@@ -119,6 +130,15 @@ struct debug_view debug_level_view = {
 	NULL
 };
 
+struct debug_view debug_pages_view = {
+	"pages",
+	&debug_prolog_pages_fn,
+	NULL,
+	NULL,
+	&debug_input_pages_fn,
+	NULL
+};
+
 struct debug_view debug_flush_view = {
         "flush",
         NULL,
@@ -149,98 +169,161 @@ DECLARE_MUTEX(debug_lock);
 static int initialized;
 
 static struct file_operations debug_file_ops = {
-	.owner	 = THIS_MODULE,
+	.owner   = THIS_MODULE,
 	.read    = debug_output,
-	.write   = debug_input,	
+	.write   = debug_input,
 	.open    = debug_open,
 	.release = debug_close,
 };
 
-static struct proc_dir_entry *debug_proc_root_entry;
+static struct dentry *debug_debugfs_root_entry;
 
 /* functions */
 
 /*
+ * debug_areas_alloc
+ * - Debug areas are implemented as a threedimensonal array:
+ *   areas[areanumber][pagenumber][pageoffset]
+ */
+
+static debug_entry_t***
+debug_areas_alloc(int pages_per_area, int nr_areas)
+{
+	debug_entry_t*** areas;
+	int i,j;
+
+	areas = (debug_entry_t ***) kmalloc(nr_areas *
+					sizeof(debug_entry_t**),
+					GFP_KERNEL);
+	if (!areas)
+		goto fail_malloc_areas;
+	for (i = 0; i < nr_areas; i++) {
+		areas[i] = (debug_entry_t**) kmalloc(pages_per_area *
+				sizeof(debug_entry_t*),GFP_KERNEL);
+		if (!areas[i]) {
+			goto fail_malloc_areas2;
+		}
+		for(j = 0; j < pages_per_area; j++) {
+			areas[i][j] = (debug_entry_t*)kmalloc(PAGE_SIZE,
+						GFP_KERNEL);
+			if(!areas[i][j]) {
+				for(j--; j >=0 ; j--) {
+					kfree(areas[i][j]);
+				}
+				kfree(areas[i]);
+				goto fail_malloc_areas2;
+			} else {
+				memset(areas[i][j],0,PAGE_SIZE);
+			}
+		}
+	}
+	return areas;
+
+fail_malloc_areas2:
+	for(i--; i >= 0; i--){
+		for(j=0; j < pages_per_area;j++){
+			kfree(areas[i][j]);
+		}
+		kfree(areas[i]);
+	}
+	kfree(areas);
+fail_malloc_areas:
+	return NULL;
+
+}
+
+
+/*
  * debug_info_alloc
  * - alloc new debug-info
  */
 
-static debug_info_t*  debug_info_alloc(char *name, int page_order,
-                                        int nr_areas, int buf_size)
+static debug_info_t*
+debug_info_alloc(char *name, int pages_per_area, int nr_areas, int buf_size,
+		int level, int mode)
 {
 	debug_info_t* rc;
-	int i;
 
 	/* alloc everything */
 
-	rc = (debug_info_t*) kmalloc(sizeof(debug_info_t), GFP_ATOMIC);
+	rc = (debug_info_t*) kmalloc(sizeof(debug_info_t), GFP_KERNEL);
 	if(!rc)
 		goto fail_malloc_rc;
-	rc->active_entry = (int*)kmalloc(nr_areas * sizeof(int), GFP_ATOMIC);
-	if(!rc->active_entry)
-		goto fail_malloc_active_entry;
-	memset(rc->active_entry, 0, nr_areas * sizeof(int));
-	rc->areas = (debug_entry_t **) kmalloc(nr_areas *
-						sizeof(debug_entry_t *),
-						GFP_ATOMIC);
-	if (!rc->areas)
-		goto fail_malloc_areas;
-	for (i = 0; i < nr_areas; i++) {
-		rc->areas[i] = (debug_entry_t *) __get_free_pages(GFP_ATOMIC,
-								page_order);
-		if (!rc->areas[i]) {
-			for (i--; i >= 0; i--) {
-				free_pages((unsigned long) rc->areas[i],
-						page_order);
-			}
-			goto fail_malloc_areas2;
-		} else {
-			memset(rc->areas[i], 0, PAGE_SIZE << page_order);
-		}
+	rc->active_entries = (int*)kmalloc(nr_areas * sizeof(int), GFP_KERNEL);
+	if(!rc->active_entries)
+		goto fail_malloc_active_entries;
+	memset(rc->active_entries, 0, nr_areas * sizeof(int));
+	rc->active_pages = (int*)kmalloc(nr_areas * sizeof(int), GFP_KERNEL);
+	if(!rc->active_pages)
+		goto fail_malloc_active_pages;
+	memset(rc->active_pages, 0, nr_areas * sizeof(int));
+	if((mode == ALL_AREAS) && (pages_per_area != 0)){
+		rc->areas = debug_areas_alloc(pages_per_area, nr_areas);
+		if(!rc->areas)
+			goto fail_malloc_areas;
+	} else {
+		rc->areas = NULL;
 	}
 
 	/* initialize members */
 
 	spin_lock_init(&rc->lock);
-	rc->page_order  = page_order;
-	rc->nr_areas    = nr_areas;
-	rc->active_area = 0;
-	rc->level       = DEBUG_DEFAULT_LEVEL;
-	rc->buf_size    = buf_size;
-	rc->entry_size  = sizeof(debug_entry_t) + buf_size;
-	strlcpy(rc->name, name, sizeof(rc->name));
+	rc->pages_per_area = pages_per_area;
+	rc->nr_areas       = nr_areas;
+	rc->active_area    = 0;
+	rc->level          = level;
+	rc->buf_size       = buf_size;
+	rc->entry_size     = sizeof(debug_entry_t) + buf_size;
+	strlcpy(rc->name, name, sizeof(rc->name)-1);
 	memset(rc->views, 0, DEBUG_MAX_VIEWS * sizeof(struct debug_view *));
-#ifdef CONFIG_PROC_FS
-	memset(rc->proc_entries, 0 ,DEBUG_MAX_VIEWS *
-		sizeof(struct proc_dir_entry*));
-#endif /* CONFIG_PROC_FS */
+	memset(rc->debugfs_entries, 0 ,DEBUG_MAX_VIEWS *
+		sizeof(struct dentry*));
 	atomic_set(&(rc->ref_count), 0);
 
 	return rc;
 
-fail_malloc_areas2:
-	kfree(rc->areas);
 fail_malloc_areas:
-	kfree(rc->active_entry);
-fail_malloc_active_entry:
+	kfree(rc->active_pages);
+fail_malloc_active_pages:
+	kfree(rc->active_entries);
+fail_malloc_active_entries:
 	kfree(rc);
 fail_malloc_rc:
 	return NULL;
 }
 
 /*
- * debug_info_free
- * - free memory debug-info
+ * debug_areas_free
+ * - free all debug areas
  */
 
-static void debug_info_free(debug_info_t* db_info){
-	int i;
+static void
+debug_areas_free(debug_info_t* db_info)
+{
+	int i,j;
+
+	if(!db_info->areas)
+		return;
 	for (i = 0; i < db_info->nr_areas; i++) {
-		free_pages((unsigned long) db_info->areas[i],
-		db_info->page_order);
+		for(j = 0; j < db_info->pages_per_area; j++) {
+			kfree(db_info->areas[i][j]);
+		}
+		kfree(db_info->areas[i]);
 	}
 	kfree(db_info->areas);
-	kfree(db_info->active_entry);
+	db_info->areas = NULL;
+}
+
+/*
+ * debug_info_free
+ * - free memory debug-info
+ */
+
+static void
+debug_info_free(debug_info_t* db_info){
+	debug_areas_free(db_info);
+	kfree(db_info->active_entries);
+	kfree(db_info->active_pages);
 	kfree(db_info);
 }
 
@@ -249,21 +332,22 @@ static void debug_info_free(debug_info_t* db_info){
  * - create new debug-info
  */
 
-static debug_info_t*  debug_info_create(char *name, int page_order, 
-                                        int nr_areas, int buf_size)
+static debug_info_t*
+debug_info_create(char *name, int pages_per_area, int nr_areas, int buf_size)
 {
 	debug_info_t* rc;
 
-        rc = debug_info_alloc(name, page_order, nr_areas, buf_size);
+        rc = debug_info_alloc(name, pages_per_area, nr_areas, buf_size,
+				DEBUG_DEFAULT_LEVEL, ALL_AREAS);
         if(!rc) 
 		goto out;
 
-
-	/* create proc rood directory */
-        rc->proc_root_entry = proc_mkdir(rc->name, debug_proc_root_entry);
+	/* create root directory */
+        rc->debugfs_root_entry = debugfs_create_dir(rc->name,
+					debug_debugfs_root_entry);
 
 	/* append new element to linked list */
-        if (debug_area_first == NULL) {
+        if (!debug_area_first) {
                 /* first element in list */
                 debug_area_first = rc;
                 rc->prev = NULL;
@@ -285,17 +369,21 @@ out:
  * - copy debug-info
  */
 
-static debug_info_t* debug_info_copy(debug_info_t* in)
+static debug_info_t*
+debug_info_copy(debug_info_t* in, int mode)
 {
-        int i;
+        int i,j;
         debug_info_t* rc;
-        rc = debug_info_alloc(in->name, in->page_order, 
-                                in->nr_areas, in->buf_size);
-        if(!rc)
+
+        rc = debug_info_alloc(in->name, in->pages_per_area, in->nr_areas,
+				in->buf_size, in->level, mode);
+        if(!rc || (mode == NO_AREAS))
                 goto out;
 
         for(i = 0; i < in->nr_areas; i++){
-                memcpy(rc->areas[i],in->areas[i], PAGE_SIZE << in->page_order);
+		for(j = 0; j < in->pages_per_area; j++) {
+			memcpy(rc->areas[i][j], in->areas[i][j],PAGE_SIZE);
+		}
         }
 out:
         return rc;
@@ -306,7 +394,8 @@ out:
  * - increments reference count for debug-info
  */
 
-static void debug_info_get(debug_info_t * db_info)
+static void
+debug_info_get(debug_info_t * db_info)
 {
 	if (db_info)
 		atomic_inc(&db_info->ref_count);
@@ -317,29 +406,20 @@ static void debug_info_get(debug_info_t * db_info)
  * - decreases reference count for debug-info and frees it if necessary
  */
 
-static void debug_info_put(debug_info_t *db_info)
+static void
+debug_info_put(debug_info_t *db_info)
 {
 	int i;
 
 	if (!db_info)
 		return;
 	if (atomic_dec_and_test(&db_info->ref_count)) {
-#ifdef DEBUG
-		printk(KERN_INFO "debug: freeing debug area %p (%s)\n",
-		       db_info, db_info->name);
-#endif
 		for (i = 0; i < DEBUG_MAX_VIEWS; i++) {
-			if (db_info->views[i] == NULL)
+			if (!db_info->views[i])
 				continue;
-#ifdef CONFIG_PROC_FS
-			remove_proc_entry(db_info->proc_entries[i]->name,
-					  db_info->proc_root_entry);
-#endif
+			debugfs_remove(db_info->debugfs_entries[i]);
 		}
-#ifdef CONFIG_PROC_FS
-		remove_proc_entry(db_info->proc_root_entry->name,
-				  debug_proc_root_entry);
-#endif
+		debugfs_remove(db_info->debugfs_root_entry);
 		if(db_info == debug_area_first)
 			debug_area_first = db_info->next;
 		if(db_info == debug_area_last)
@@ -355,9 +435,9 @@ static void debug_info_put(debug_info_t *db_info)
  * - format one debug entry and return size of formated data
  */
 
-static int debug_format_entry(file_private_info_t *p_info)
+static int
+debug_format_entry(file_private_info_t *p_info)
 {
-	debug_info_t *id_org    = p_info->debug_info_org;
 	debug_info_t *id_snap   = p_info->debug_info_snap;
 	struct debug_view *view = p_info->view;
 	debug_entry_t *act_entry;
@@ -365,22 +445,23 @@ static int debug_format_entry(file_private_info_t *p_info)
 	if(p_info->act_entry == DEBUG_PROLOG_ENTRY){
 		/* print prolog */
         	if (view->prolog_proc)
-                	len += view->prolog_proc(id_org, view,p_info->temp_buf);
+                	len += view->prolog_proc(id_snap,view,p_info->temp_buf);
 		goto out;
 	}
-
-	act_entry = (debug_entry_t *) ((char*)id_snap->areas[p_info->act_area] +
-					p_info->act_entry);
+	if (!id_snap->areas) /* this is true, if we have a prolog only view */
+		goto out;    /* or if 'pages_per_area' is 0 */
+	act_entry = (debug_entry_t *) ((char*)id_snap->areas[p_info->act_area]
+				[p_info->act_page] + p_info->act_entry);
                         
 	if (act_entry->id.stck == 0LL)
 			goto out;  /* empty entry */
 	if (view->header_proc)
-		len += view->header_proc(id_org, view, p_info->act_area, 
+		len += view->header_proc(id_snap, view, p_info->act_area,
 					act_entry, p_info->temp_buf + len);
 	if (view->format_proc)
-		len += view->format_proc(id_org, view, p_info->temp_buf + len,
+		len += view->format_proc(id_snap, view, p_info->temp_buf + len,
 						DEBUG_DATA(act_entry));
-      out:
+out:
         return len;
 }
 
@@ -389,20 +470,30 @@ static int debug_format_entry(file_private_info_t *p_info)
  * - goto next entry in p_info
  */
 
-extern inline int debug_next_entry(file_private_info_t *p_info)
+extern inline int
+debug_next_entry(file_private_info_t *p_info)
 {
-	debug_info_t *id = p_info->debug_info_snap;
+	debug_info_t *id;
+
+	id = p_info->debug_info_snap;
 	if(p_info->act_entry == DEBUG_PROLOG_ENTRY){
 		p_info->act_entry = 0;
+		p_info->act_page  = 0;
 		goto out;
 	}
-	if ((p_info->act_entry += id->entry_size)
-		> ((PAGE_SIZE << (id->page_order)) 
-		- id->entry_size)){
-
-		/* next area */
+	if(!id->areas)
+		return 1;
+	p_info->act_entry += id->entry_size;
+	/* switch to next page, if we reached the end of the page  */
+	if (p_info->act_entry > (PAGE_SIZE - id->entry_size)){
+		/* next page */
 		p_info->act_entry = 0;
-        	p_info->act_area++;
+		p_info->act_page += 1;
+		if((p_info->act_page % id->pages_per_area) == 0) {
+			/* next area */
+        		p_info->act_area++;
+			p_info->act_page=0;
+		}
         	if(p_info->act_area >= id->nr_areas)
 			return 1;
 	}
@@ -416,13 +507,14 @@ out:
  * - copies formated debug entries to the user buffer
  */
 
-static ssize_t debug_output(struct file *file,		/* file descriptor */
-			    char __user *user_buf,	/* user buffer */
-			    size_t  len,		/* length of buffer */
-			    loff_t *offset)	      /* offset in the file */
+static ssize_t
+debug_output(struct file *file,		/* file descriptor */
+	    char __user *user_buf,	/* user buffer */
+	    size_t  len,		/* length of buffer */
+	    loff_t *offset)		/* offset in the file */
 {
 	size_t count = 0;
-	size_t entry_offset, size = 0;
+	size_t entry_offset;
 	file_private_info_t *p_info;
 
 	p_info = ((file_private_info_t *) file->private_data);
@@ -430,27 +522,33 @@ static ssize_t debug_output(struct file *file,		/* file descriptor */
 		return -EPIPE;
 	if(p_info->act_area >= p_info->debug_info_snap->nr_areas)
 		return 0;
-
 	entry_offset = p_info->act_entry_offset;
-
 	while(count < len){
-		size = debug_format_entry(p_info);
-		size = min((len - count), (size - entry_offset));
-
-		if(size){
-			if (copy_to_user(user_buf + count, 
-					p_info->temp_buf + entry_offset, size))
-			return -EFAULT;
+		int formatted_line_size;
+		int formatted_line_residue;
+		int user_buf_residue;
+		size_t copy_size;
+
+		formatted_line_size = debug_format_entry(p_info);
+		formatted_line_residue = formatted_line_size - entry_offset;
+		user_buf_residue = len-count;
+		copy_size = min(user_buf_residue, formatted_line_residue);
+		if(copy_size){
+			if (copy_to_user(user_buf + count, p_info->temp_buf
+					+ entry_offset, copy_size))
+				return -EFAULT;
+			count += copy_size;
+			entry_offset += copy_size;
 		}
-		count += size;
-		entry_offset = 0;
-		if(count != len)
-			if(debug_next_entry(p_info)) 
+		if(copy_size == formatted_line_residue){
+			entry_offset = 0;
+			if(debug_next_entry(p_info))
 				goto out;
+		}
 	}
 out:
 	p_info->offset           = *offset + count;
-	p_info->act_entry_offset = size;	
+	p_info->act_entry_offset = entry_offset;
 	*offset = p_info->offset;
 	return count;
 }
@@ -461,9 +559,9 @@ out:
  * - calls input function of view
  */
 
-static ssize_t debug_input(struct file *file,
-			   const char __user *user_buf, size_t length,
-			   loff_t *offset)
+static ssize_t
+debug_input(struct file *file, const char __user *user_buf, size_t length,
+		loff_t *offset)
 {
 	int rc = 0;
 	file_private_info_t *p_info;
@@ -487,26 +585,23 @@ static ssize_t debug_input(struct file *file,
  *   handle
  */
 
-static int debug_open(struct inode *inode, struct file *file)
+static int
+debug_open(struct inode *inode, struct file *file)
 {
 	int i = 0, rc = 0;
 	file_private_info_t *p_info;
 	debug_info_t *debug_info, *debug_info_snapshot;
 
-#ifdef DEBUG
-	printk("debug_open\n");
-#endif
 	down(&debug_lock);
 
 	/* find debug log and view */
-
 	debug_info = debug_area_first;
 	while(debug_info != NULL){
 		for (i = 0; i < DEBUG_MAX_VIEWS; i++) {
-			if (debug_info->views[i] == NULL)
+			if (!debug_info->views[i])
 				continue;
-			else if (debug_info->proc_entries[i] ==
-				 PDE(file->f_dentry->d_inode)) {
+			else if (debug_info->debugfs_entries[i] ==
+				 file->f_dentry) {
 				goto found;	/* found view ! */
 			}
 		}
@@ -516,41 +611,42 @@ static int debug_open(struct inode *inode, struct file *file)
 	rc = -EINVAL;
 	goto out;
 
-      found:
+found:
 
-	/* make snapshot of current debug areas to get it consistent */
+	/* Make snapshot of current debug areas to get it consistent.     */
+	/* To copy all the areas is only needed, if we have a view which  */
+	/* formats the debug areas. */
 
-	debug_info_snapshot = debug_info_copy(debug_info);
+	if(!debug_info->views[i]->format_proc &&
+		!debug_info->views[i]->header_proc){
+		debug_info_snapshot = debug_info_copy(debug_info, NO_AREAS);
+	} else {
+		debug_info_snapshot = debug_info_copy(debug_info, ALL_AREAS);
+	}
 
 	if(!debug_info_snapshot){
-#ifdef DEBUG
-		printk(KERN_ERR "debug_open: debug_info_copy failed (out of mem)\n");
-#endif
 		rc = -ENOMEM;
 		goto out;
 	}
-
-	if ((file->private_data =
-	     kmalloc(sizeof(file_private_info_t), GFP_ATOMIC)) == 0) {
-#ifdef DEBUG
-		printk(KERN_ERR "debug_open: kmalloc failed\n");
-#endif
-		debug_info_free(debug_info_snapshot);	
+	p_info = (file_private_info_t *) kmalloc(sizeof(file_private_info_t),
+						GFP_KERNEL);
+	if(!p_info){
+		if(debug_info_snapshot)
+			debug_info_free(debug_info_snapshot);
 		rc = -ENOMEM;
 		goto out;
 	}
-	p_info = (file_private_info_t *) file->private_data;
 	p_info->offset = 0;
 	p_info->debug_info_snap = debug_info_snapshot;
 	p_info->debug_info_org  = debug_info;
 	p_info->view = debug_info->views[i];
 	p_info->act_area = 0;
+	p_info->act_page = 0;
 	p_info->act_entry = DEBUG_PROLOG_ENTRY;
 	p_info->act_entry_offset = 0;
-
+	file->private_data = p_info;
 	debug_info_get(debug_info);
-
-      out:
+out:
 	up(&debug_lock);
 	return rc;
 }
@@ -561,14 +657,13 @@ static int debug_open(struct inode *inode, struct file *file)
  * - deletes  private_data area of the file handle
  */
 
-static int debug_close(struct inode *inode, struct file *file)
+static int
+debug_close(struct inode *inode, struct file *file)
 {
 	file_private_info_t *p_info;
-#ifdef DEBUG
-	printk("debug_close\n");
-#endif
 	p_info = (file_private_info_t *) file->private_data;
-	debug_info_free(p_info->debug_info_snap);
+	if(p_info->debug_info_snap)
+		debug_info_free(p_info->debug_info_snap);
 	debug_info_put(p_info->debug_info_org);
 	kfree(file->private_data);
 	return 0;		/* success */
@@ -580,8 +675,8 @@ static int debug_close(struct inode *inode, struct file *file)
  * - returns handle for debug area
  */
 
-debug_info_t *debug_register
-    (char *name, int page_order, int nr_areas, int buf_size) 
+debug_info_t*
+debug_register (char *name, int pages_per_area, int nr_areas, int buf_size)
 {
 	debug_info_t *rc = NULL;
 
@@ -591,18 +686,14 @@ debug_info_t *debug_register
 
         /* create new debug_info */
 
-	rc = debug_info_create(name, page_order, nr_areas, buf_size);
+	rc = debug_info_create(name, pages_per_area, nr_areas, buf_size);
 	if(!rc) 
 		goto out;
 	debug_register_view(rc, &debug_level_view);
         debug_register_view(rc, &debug_flush_view);
-#ifdef DEBUG
-	printk(KERN_INFO
-	       "debug: reserved %d areas of %d pages for debugging %s\n",
-	       nr_areas, 1 << page_order, rc->name);
-#endif
-      out:
-        if (rc == NULL){
+	debug_register_view(rc, &debug_pages_view);
+out:
+        if (!rc){
 		printk(KERN_ERR "debug: debug_register failed for %s\n",name);
         }
 	up(&debug_lock);
@@ -614,27 +705,65 @@ debug_info_t *debug_register
  * - give back debug area
  */
 
-void debug_unregister(debug_info_t * id)
+void
+debug_unregister(debug_info_t * id)
 {
 	if (!id)
 		goto out;
 	down(&debug_lock);
-#ifdef DEBUG
-	printk(KERN_INFO "debug: unregistering %s\n", id->name);
-#endif
 	debug_info_put(id);
 	up(&debug_lock);
 
-      out:
+out:
 	return;
 }
 
 /*
+ * debug_set_size:
+ * - set area size (number of pages) and number of areas
+ */
+static int
+debug_set_size(debug_info_t* id, int nr_areas, int pages_per_area)
+{
+	unsigned long flags;
+	debug_entry_t *** new_areas;
+	int rc=0;
+
+	if(!id || (nr_areas <= 0) || (pages_per_area < 0))
+		return -EINVAL;
+	if(pages_per_area > 0){
+		new_areas = debug_areas_alloc(pages_per_area, nr_areas);
+		if(!new_areas) {
+			printk(KERN_WARNING "debug: could not allocate memory "\
+					 "for pagenumber: %i\n",pages_per_area);
+			rc = -ENOMEM;
+			goto out;
+		}
+	} else {
+		new_areas = NULL;
+	}
+	spin_lock_irqsave(&id->lock,flags);
+	debug_areas_free(id);
+	id->areas = new_areas;
+	id->nr_areas = nr_areas;
+	id->pages_per_area = pages_per_area;
+	id->active_area = 0;
+	memset(id->active_entries,0,sizeof(int)*id->nr_areas);
+	memset(id->active_pages, 0, sizeof(int)*id->nr_areas);
+	spin_unlock_irqrestore(&id->lock,flags);
+	printk(KERN_INFO "debug: %s: set new size (%i pages)\n"\
+			 ,id->name, pages_per_area);
+out:
+	return rc;
+}
+
+/*
  * debug_set_level:
  * - set actual debug level
  */
 
-void debug_set_level(debug_info_t* id, int new_level)
+void
+debug_set_level(debug_info_t* id, int new_level)
 {
 	unsigned long flags;
 	if(!id)
@@ -649,10 +778,6 @@ void debug_set_level(debug_info_t* id, int new_level)
                         id->name, new_level, 0, DEBUG_MAX_LEVEL);
         } else {
                 id->level = new_level;
-#ifdef DEBUG
-                printk(KERN_INFO 
-			"debug: %s: new level %i\n",id->name,id->level);
-#endif
         }
 	spin_unlock_irqrestore(&id->lock,flags);
 }
@@ -663,11 +788,16 @@ void debug_set_level(debug_info_t* id, int new_level)
  * - set active entry to next in the ring buffer
  */
 
-extern inline void proceed_active_entry(debug_info_t * id)
+extern inline void
+proceed_active_entry(debug_info_t * id)
 {
-	if ((id->active_entry[id->active_area] += id->entry_size)
-	    > ((PAGE_SIZE << (id->page_order)) - id->entry_size))
-		id->active_entry[id->active_area] = 0;
+	if ((id->active_entries[id->active_area] += id->entry_size)
+	    > (PAGE_SIZE - id->entry_size)){
+		id->active_entries[id->active_area] = 0;
+		id->active_pages[id->active_area] =
+			(id->active_pages[id->active_area] + 1) %
+			id->pages_per_area;
+	}
 }
 
 /*
@@ -675,7 +805,8 @@ extern inline void proceed_active_entry(debug_info_t * id)
  * - set active area to next in the ring buffer
  */
 
-extern inline void proceed_active_area(debug_info_t * id)
+extern inline void
+proceed_active_area(debug_info_t * id)
 {
 	id->active_area++;
 	id->active_area = id->active_area % id->nr_areas;
@@ -685,10 +816,12 @@ extern inline void proceed_active_area(debug_info_t * id)
  * get_active_entry:
  */
 
-extern inline debug_entry_t *get_active_entry(debug_info_t * id)
+extern inline debug_entry_t*
+get_active_entry(debug_info_t * id)
 {
-	return (debug_entry_t *) ((char *) id->areas[id->active_area] +
-				  id->active_entry[id->active_area]);
+	return (debug_entry_t *) (((char *) id->areas[id->active_area]
+					[id->active_pages[id->active_area]]) +
+					id->active_entries[id->active_area]);
 }
 
 /*
@@ -696,8 +829,9 @@ extern inline debug_entry_t *get_active_entry(debug_info_t * id)
  * - set timestamp, caller address, cpu number etc.
  */
 
-extern inline void debug_finish_entry(debug_info_t * id, debug_entry_t* active,
-		int level, int exception)
+extern inline void
+debug_finish_entry(debug_info_t * id, debug_entry_t* active, int level,
+			int exception)
 {
 	STCK(active->id.stck);
 	active->id.fields.cpuid = smp_processor_id();
@@ -721,7 +855,8 @@ static int debug_active=1;
  * always allow read, allow write only if debug_stoppable is set or
  * if debug_active is already off
  */
-static int s390dbf_procactive(ctl_table *table, int write, struct file *filp,
+static int
+s390dbf_procactive(ctl_table *table, int write, struct file *filp,
                      void __user *buffer, size_t *lenp, loff_t *ppos)
 {
 	if (!write || debug_stoppable || !debug_active)
@@ -766,7 +901,8 @@ static struct ctl_table s390dbf_dir_table[] = {
 
 struct ctl_table_header *s390dbf_sysctl_header;
 
-void debug_stop_all(void)
+void
+debug_stop_all(void)
 {
 	if (debug_stoppable)
 		debug_active = 0;
@@ -778,13 +914,13 @@ void debug_stop_all(void)
  * - write debug entry with given size
  */
 
-debug_entry_t *debug_event_common(debug_info_t * id, int level, const void *buf,
-			          int len)
+debug_entry_t*
+debug_event_common(debug_info_t * id, int level, const void *buf, int len)
 {
 	unsigned long flags;
 	debug_entry_t *active;
 
-	if (!debug_active)
+	if (!debug_active || !id->areas)
 		return NULL;
 	spin_lock_irqsave(&id->lock, flags);
 	active = get_active_entry(id);
@@ -801,13 +937,13 @@ debug_entry_t *debug_event_common(debug_info_t * id, int level, const void *buf,
  * - write debug entry with given size and switch to next debug area
  */
 
-debug_entry_t *debug_exception_common(debug_info_t * id, int level, 
-                                      const void *buf, int len)
+debug_entry_t
+*debug_exception_common(debug_info_t * id, int level, const void *buf, int len)
 {
 	unsigned long flags;
 	debug_entry_t *active;
 
-	if (!debug_active)
+	if (!debug_active || !id->areas)
 		return NULL;
 	spin_lock_irqsave(&id->lock, flags);
 	active = get_active_entry(id);
@@ -823,7 +959,8 @@ debug_entry_t *debug_exception_common(debug_info_t * id, int level,
  * counts arguments in format string for sprintf view
  */
 
-extern inline int debug_count_numargs(char *string)
+extern inline int
+debug_count_numargs(char *string)
 {
 	int numargs=0;
 
@@ -838,8 +975,8 @@ extern inline int debug_count_numargs(char *string)
  * debug_sprintf_event:
  */
 
-debug_entry_t *debug_sprintf_event(debug_info_t* id,
-                                   int level,char *string,...)
+debug_entry_t*
+debug_sprintf_event(debug_info_t* id, int level,char *string,...)
 {
 	va_list   ap;
 	int numargs,idx;
@@ -849,7 +986,7 @@ debug_entry_t *debug_sprintf_event(debug_info_t* id,
 
 	if((!id) || (level > id->level))
 		return NULL;
-	if (!debug_active)
+	if (!debug_active || !id->areas)
 		return NULL;
 	numargs=debug_count_numargs(string);
 
@@ -871,8 +1008,8 @@ debug_entry_t *debug_sprintf_event(debug_info_t* id,
  * debug_sprintf_exception:
  */
 
-debug_entry_t *debug_sprintf_exception(debug_info_t* id,
-                                       int level,char *string,...)
+debug_entry_t*
+debug_sprintf_exception(debug_info_t* id, int level,char *string,...)
 {
 	va_list   ap;
 	int numargs,idx;
@@ -882,7 +1019,7 @@ debug_entry_t *debug_sprintf_exception(debug_info_t* id,
 
 	if((!id) || (level > id->level))
 		return NULL;
-	if (!debug_active)
+	if (!debug_active || !id->areas)
 		return NULL;
 
 	numargs=debug_count_numargs(string);
@@ -906,15 +1043,14 @@ debug_entry_t *debug_sprintf_exception(debug_info_t* id,
  * - is called exactly once to initialize the debug feature
  */
 
-static int __init debug_init(void)
+static int
+__init debug_init(void)
 {
 	int rc = 0;
 
 	s390dbf_sysctl_header = register_sysctl_table(s390dbf_dir_table, 1);
 	down(&debug_lock);
-#ifdef CONFIG_PROC_FS
-	debug_proc_root_entry = proc_mkdir(DEBUG_DIR_ROOT, NULL);
-#endif /* CONFIG_PROC_FS */
+	debug_debugfs_root_entry = debugfs_create_dir(DEBUG_DIR_ROOT,NULL);
 	printk(KERN_INFO "debug: Initialization complete\n");
 	initialized = 1;
 	up(&debug_lock);
@@ -926,13 +1062,14 @@ static int __init debug_init(void)
  * debug_register_view:
  */
 
-int debug_register_view(debug_info_t * id, struct debug_view *view)
+int
+debug_register_view(debug_info_t * id, struct debug_view *view)
 {
 	int rc = 0;
 	int i;
 	unsigned long flags;
 	mode_t mode = S_IFREG;
-	struct proc_dir_entry *pde;
+	struct dentry *pde;
 
 	if (!id)
 		goto out;
@@ -940,16 +1077,17 @@ int debug_register_view(debug_info_t * id, struct debug_view *view)
 		mode |= S_IRUSR;
 	if (view->input_proc)
 		mode |= S_IWUSR;
-	pde = create_proc_entry(view->name, mode, id->proc_root_entry);
+	pde = debugfs_create_file(view->name, mode, id->debugfs_root_entry,
+				NULL, &debug_file_ops);
 	if (!pde){
-		printk(KERN_WARNING "debug: create_proc_entry() failed! Cannot register view %s/%s\n", id->name,view->name);
+		printk(KERN_WARNING "debug: debugfs_create_file() failed!"\
+			" Cannot register view %s/%s\n", id->name,view->name);
 		rc = -1;
 		goto out;
 	}
-
 	spin_lock_irqsave(&id->lock, flags);
 	for (i = 0; i < DEBUG_MAX_VIEWS; i++) {
-		if (id->views[i] == NULL)
+		if (!id->views[i])
 			break;
 	}
 	if (i == DEBUG_MAX_VIEWS) {
@@ -957,16 +1095,14 @@ int debug_register_view(debug_info_t * id, struct debug_view *view)
 			id->name,view->name);
 		printk(KERN_WARNING 
 			"debug: maximum number of views reached (%i)!\n", i);
-		remove_proc_entry(pde->name, id->proc_root_entry);
+		debugfs_remove(pde);
 		rc = -1;
-	}
-	else {
+	} else {
 		id->views[i] = view;
-		pde->proc_fops = &debug_file_ops;
-		id->proc_entries[i] = pde;
+		id->debugfs_entries[i] = pde;
 	}
 	spin_unlock_irqrestore(&id->lock, flags);
-      out:
+out:
 	return rc;
 }
 
@@ -974,7 +1110,8 @@ int debug_register_view(debug_info_t * id, struct debug_view *view)
  * debug_unregister_view:
  */
 
-int debug_unregister_view(debug_info_t * id, struct debug_view *view)
+int
+debug_unregister_view(debug_info_t * id, struct debug_view *view)
 {
 	int rc = 0;
 	int i;
@@ -990,15 +1127,46 @@ int debug_unregister_view(debug_info_t * id, struct debug_view *view)
 	if (i == DEBUG_MAX_VIEWS)
 		rc = -1;
 	else {
-#ifdef CONFIG_PROC_FS
-		remove_proc_entry(id->proc_entries[i]->name,
-				  id->proc_root_entry);
-#endif
+		debugfs_remove(id->debugfs_entries[i]);
 		id->views[i] = NULL;
 		rc = 0;
 	}
 	spin_unlock_irqrestore(&id->lock, flags);
-      out:
+out:
+	return rc;
+}
+
+static inline char *
+debug_get_user_string(const char __user *user_buf, size_t user_len)
+{
+	char* buffer;
+
+	buffer = kmalloc(user_len + 1, GFP_KERNEL);
+	if (!buffer)
+		return ERR_PTR(-ENOMEM);
+	if (copy_from_user(buffer, user_buf, user_len) != 0) {
+		kfree(buffer);
+		return ERR_PTR(-EFAULT);
+	}
+	/* got the string, now strip linefeed. */
+	if (buffer[user_len - 1] == '\n')
+		buffer[user_len - 1] = 0;
+	else
+		buffer[user_len] = 0;
+        return buffer;
+}
+
+static inline int
+debug_get_uint(char *buf)
+{
+	int rc;
+
+	for(; isspace(*buf); buf++);
+	rc = simple_strtoul(buf, &buf, 10);
+	if(*buf){
+		printk("debug: no integer specified!\n");
+		rc = -EINVAL;
+	}
 	return rc;
 }
 
@@ -1011,13 +1179,69 @@ int debug_unregister_view(debug_info_t * id, struct debug_view *view)
  * prints out actual debug level
  */
 
-static int debug_prolog_level_fn(debug_info_t * id,
+static int
+debug_prolog_pages_fn(debug_info_t * id,
 				 struct debug_view *view, char *out_buf)
 {
+	return sprintf(out_buf, "%i\n", id->pages_per_area);
+}
+
+/*
+ * reads new size (number of pages per debug area)
+ */
+
+static int
+debug_input_pages_fn(debug_info_t * id, struct debug_view *view,
+			struct file *file, const char __user *user_buf,
+			size_t user_len, loff_t * offset)
+{
+	char *str;
+	int rc,new_pages;
+
+	if (user_len > 0x10000)
+                user_len = 0x10000;
+	if (*offset != 0){
+		rc = -EPIPE;
+		goto out;
+	}
+	str = debug_get_user_string(user_buf,user_len);
+	if(IS_ERR(str)){
+		rc = PTR_ERR(str);
+		goto out;
+	}
+	new_pages = debug_get_uint(str);
+	if(new_pages < 0){
+		rc = -EINVAL;
+		goto free_str;
+	}
+	rc = debug_set_size(id,id->nr_areas, new_pages);
+	if(rc != 0){
+		rc = -EINVAL;
+		goto free_str;
+	}
+	rc = user_len;
+free_str:
+	kfree(str);
+out:
+	*offset += user_len;
+	return rc;		/* number of input characters */
+}
+
+/*
+ * prints out actual debug level
+ */
+
+static int
+debug_prolog_level_fn(debug_info_t * id, struct debug_view *view, char *out_buf)
+{
 	int rc = 0;
 
-	if(id->level == -1) rc = sprintf(out_buf,"-\n");
-	else rc = sprintf(out_buf, "%i\n", id->level);
+	if(id->level == DEBUG_OFF_LEVEL) {
+		rc = sprintf(out_buf,"-\n");
+	}
+	else {
+		rc = sprintf(out_buf, "%i\n", id->level);
+	}
 	return rc;
 }
 
@@ -1025,30 +1249,43 @@ static int debug_prolog_level_fn(debug_info_t * id,
  * reads new debug level
  */
 
-static int debug_input_level_fn(debug_info_t * id, struct debug_view *view,
-				struct file *file, const char __user *user_buf,
-				size_t in_buf_size, loff_t * offset)
+static int
+debug_input_level_fn(debug_info_t * id, struct debug_view *view,
+			struct file *file, const char __user *user_buf,
+			size_t user_len, loff_t * offset)
 {
-	char input_buf[1];
-	int rc = in_buf_size;
+	char *str;
+	int rc,new_level;
 
-	if (*offset != 0)
+	if (user_len > 0x10000)
+                user_len = 0x10000;
+	if (*offset != 0){
+		rc = -EPIPE;
 		goto out;
-	if (copy_from_user(input_buf, user_buf, 1)){
-		rc = -EFAULT;
+	}
+	str = debug_get_user_string(user_buf,user_len);
+	if(IS_ERR(str)){
+		rc = PTR_ERR(str);
 		goto out;
 	}
-	if (isdigit(input_buf[0])) {
-		int new_level = ((int) input_buf[0] - (int) '0');
-		debug_set_level(id, new_level);
-	} else if(input_buf[0] == '-') {
+	if(str[0] == '-'){
 		debug_set_level(id, DEBUG_OFF_LEVEL);
+		rc = user_len;
+		goto free_str;
 	} else {
-		printk(KERN_INFO "debug: level `%c` is not valid\n",
-		       input_buf[0]);
+		new_level = debug_get_uint(str);
 	}
-      out:
-	*offset += in_buf_size;
+	if(new_level < 0) {
+		printk(KERN_INFO "debug: level `%s` is not valid\n", str);
+		rc = -EINVAL;
+	} else {
+		debug_set_level(id, new_level);
+		rc = user_len;
+	}
+free_str:
+	kfree(str);
+out:
+	*offset += user_len;
 	return rc;		/* number of input characters */
 }
 
@@ -1057,29 +1294,36 @@ static int debug_input_level_fn(debug_info_t * id, struct debug_view *view,
  * flushes debug areas
  */
  
-void debug_flush(debug_info_t* id, int area)
+void
+debug_flush(debug_info_t* id, int area)
 {
         unsigned long flags;
-        int i;
+        int i,j;
 
-        if(!id)
+        if(!id || !id->areas)
                 return;
         spin_lock_irqsave(&id->lock,flags);
         if(area == DEBUG_FLUSH_ALL){
                 id->active_area = 0;
-                memset(id->active_entry, 0, id->nr_areas * sizeof(int));
-                for (i = 0; i < id->nr_areas; i++) 
-                        memset(id->areas[i], 0, PAGE_SIZE << id->page_order);
+                memset(id->active_entries, 0, id->nr_areas * sizeof(int));
+                for (i = 0; i < id->nr_areas; i++) {
+			id->active_pages[i] = 0;
+			for(j = 0; j < id->pages_per_area; j++) {
+                        	memset(id->areas[i][j], 0, PAGE_SIZE);
+			}
+		}
                 printk(KERN_INFO "debug: %s: all areas flushed\n",id->name);
         } else if(area >= 0 && area < id->nr_areas) {
-                id->active_entry[area] = 0;
-                memset(id->areas[area], 0, PAGE_SIZE << id->page_order);
-                printk(KERN_INFO
-                        "debug: %s: area %i has been flushed\n",
+                id->active_entries[area] = 0;
+		id->active_pages[area] = 0;
+		for(i = 0; i < id->pages_per_area; i++) {
+                	memset(id->areas[area][i],0,PAGE_SIZE);
+		}
+                printk(KERN_INFO "debug: %s: area %i has been flushed\n",
                         id->name, area);
         } else {
                 printk(KERN_INFO
-                        "debug: %s: area %i cannot be flushed (range: %i - %i)\n",
+                      "debug: %s: area %i cannot be flushed (range: %i - %i)\n",
                         id->name, area, 0, id->nr_areas-1);
         }
         spin_unlock_irqrestore(&id->lock,flags);
@@ -1089,15 +1333,20 @@ void debug_flush(debug_info_t* id, int area)
  * view function: flushes debug areas 
  */
 
-static int debug_input_flush_fn(debug_info_t * id, struct debug_view *view,
-                                struct file *file, const char __user *user_buf,
-                                size_t in_buf_size, loff_t * offset)
+static int
+debug_input_flush_fn(debug_info_t * id, struct debug_view *view,
+			struct file *file, const char __user *user_buf,
+			size_t user_len, loff_t * offset)
 {
         char input_buf[1];
-        int rc = in_buf_size;
- 
-        if (*offset != 0)
+        int rc = user_len;
+
+	if (user_len > 0x10000)
+                user_len = 0x10000;
+        if (*offset != 0){
+		rc = -EPIPE;
                 goto out;
+	}
         if (copy_from_user(input_buf, user_buf, 1)){
                 rc = -EFAULT;
                 goto out;
@@ -1114,8 +1363,8 @@ static int debug_input_flush_fn(debug_info_t * id, struct debug_view *view,
 
         printk(KERN_INFO "debug: area `%c` is not valid\n", input_buf[0]);
 
-      out:
-        *offset += in_buf_size;
+out:
+        *offset += user_len;
         return rc;              /* number of input characters */
 }
 
@@ -1123,8 +1372,9 @@ static int debug_input_flush_fn(debug_info_t * id, struct debug_view *view,
  * prints debug header in raw format
  */
 
-int debug_raw_header_fn(debug_info_t * id, struct debug_view *view,
-                         int area, debug_entry_t * entry, char *out_buf)
+static int
+debug_raw_header_fn(debug_info_t * id, struct debug_view *view,
+			int area, debug_entry_t * entry, char *out_buf)
 {
         int rc;
 
@@ -1137,7 +1387,8 @@ int debug_raw_header_fn(debug_info_t * id, struct debug_view *view,
  * prints debug data in raw format
  */
 
-static int debug_raw_format_fn(debug_info_t * id, struct debug_view *view,
+static int
+debug_raw_format_fn(debug_info_t * id, struct debug_view *view,
 			       char *out_buf, const char *in_buf)
 {
 	int rc;
@@ -1151,8 +1402,9 @@ static int debug_raw_format_fn(debug_info_t * id, struct debug_view *view,
  * prints debug data in hex/ascii format
  */
 
-static int debug_hex_ascii_format_fn(debug_info_t * id, struct debug_view *view,
-		    		  char *out_buf, const char *in_buf)
+static int
+debug_hex_ascii_format_fn(debug_info_t * id, struct debug_view *view,
+	    		  char *out_buf, const char *in_buf)
 {
 	int i, rc = 0;
 
@@ -1176,7 +1428,8 @@ static int debug_hex_ascii_format_fn(debug_info_t * id, struct debug_view *view,
  * prints header for debug entry
  */
 
-int debug_dflt_header_fn(debug_info_t * id, struct debug_view *view,
+int
+debug_dflt_header_fn(debug_info_t * id, struct debug_view *view,
 			 int area, debug_entry_t * entry, char *out_buf)
 {
 	struct timeval time_val;
@@ -1210,8 +1463,9 @@ int debug_dflt_header_fn(debug_info_t * id, struct debug_view *view,
 
 #define DEBUG_SPRINTF_MAX_ARGS 10
 
-int debug_sprintf_format_fn(debug_info_t * id, struct debug_view *view,
-                            char *out_buf, debug_sprintf_entry_t *curr_event)
+static int
+debug_sprintf_format_fn(debug_info_t * id, struct debug_view *view,
+                        char *out_buf, debug_sprintf_entry_t *curr_event)
 {
 	int num_longs, num_used_args = 0,i, rc = 0;
 	int index[DEBUG_SPRINTF_MAX_ARGS];
@@ -1251,14 +1505,10 @@ out:
 /*
  * clean up module
  */
-void __exit debug_exit(void)
+void
+__exit debug_exit(void)
 {
-#ifdef DEBUG
-	printk("debug_cleanup_module: \n");
-#endif
-#ifdef CONFIG_PROC_FS
-	remove_proc_entry(debug_proc_root_entry->name, NULL);
-#endif /* CONFIG_PROC_FS */
+	debugfs_remove(debug_debugfs_root_entry);
 	unregister_sysctl_table(s390dbf_sysctl_header);
 	return;
 }
@@ -1266,7 +1516,7 @@ void __exit debug_exit(void)
 /*
  * module definitions
  */
-core_initcall(debug_init);
+postcore_initcall(debug_init);
 module_exit(debug_exit);
 MODULE_LICENSE("GPL");
 
diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S
index c0e09b33feb..5b262b5d869 100644
--- a/arch/s390/kernel/entry.S
+++ b/arch/s390/kernel/entry.S
@@ -7,6 +7,7 @@
  *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
  *               Hartmut Penner (hp@de.ibm.com),
  *               Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
+ *		 Heiko Carstens <heiko.carstens@de.ibm.com>
  */
 
 #include <linux/sys.h>
@@ -49,9 +50,9 @@ SP_ILC       =  STACK_FRAME_OVERHEAD + __PT_ILC
 SP_TRAP      =  STACK_FRAME_OVERHEAD + __PT_TRAP
 SP_SIZE      =  STACK_FRAME_OVERHEAD + __PT_SIZE
 
-_TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NEED_RESCHED | \
+_TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NEED_RESCHED | _TIF_MCCK_PENDING | \
 		 _TIF_RESTART_SVC | _TIF_SINGLE_STEP )
-_TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NEED_RESCHED)
+_TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NEED_RESCHED | _TIF_MCCK_PENDING)
 
 STACK_SHIFT = PAGE_SHIFT + THREAD_ORDER
 STACK_SIZE  = 1 << STACK_SHIFT
@@ -121,7 +122,11 @@ STACK_SIZE  = 1 << STACK_SHIFT
 	bz	BASED(stack_overflow)
 3:
 #endif
-2:	s	%r15,BASED(.Lc_spsize)	# make room for registers & psw
+2:
+	.endm
+
+	.macro  CREATE_STACK_FRAME psworg,savearea
+	s	%r15,BASED(.Lc_spsize)	# make room for registers & psw
 	mvc	SP_PSW(8,%r15),0(%r12)	# move user PSW to stack
 	la	%r12,\psworg
 	st	%r2,SP_ORIG_R2(%r15)	# store original content of gpr 2
@@ -161,6 +166,13 @@ __switch_to_base:
         be      __switch_to_noper-__switch_to_base(%r1)	# we got away w/o bashing TLB's
         lctl    %c9,%c11,__THREAD_per(%r3)	# Nope we didn't
 __switch_to_noper:
+	l	%r4,__THREAD_info(%r2)		# get thread_info of prev
+	tm	__TI_flags+3(%r4),_TIF_MCCK_PENDING # machine check pending?
+	bz	__switch_to_no_mcck-__switch_to_base(%r1)
+	ni	__TI_flags+3(%r4),255-_TIF_MCCK_PENDING # clear flag in prev
+	l	%r4,__THREAD_info(%r3)		# get thread_info of next
+	oi	__TI_flags+3(%r4),_TIF_MCCK_PENDING # set it in next
+__switch_to_no_mcck:
         stm     %r6,%r15,__SF_GPRS(%r15)# store __switch_to registers of prev task
 	st	%r15,__THREAD_ksp(%r2)	# store kernel stack to prev->tss.ksp
 	l	%r15,__THREAD_ksp(%r3)	# load kernel stack from next->tss.ksp
@@ -185,6 +197,7 @@ system_call:
 sysc_saveall:
 	SAVE_ALL_BASE __LC_SAVE_AREA
         SAVE_ALL __LC_SVC_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_SVC_OLD_PSW,__LC_SAVE_AREA
 	lh	%r7,0x8a	  # get svc number from lowcore
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 sysc_vtime:
@@ -234,6 +247,8 @@ sysc_work_loop:
 # One of the work bits is on. Find out which one.
 #
 sysc_work:
+	tm	__TI_flags+3(%r9),_TIF_MCCK_PENDING
+	bo	BASED(sysc_mcck_pending)
 	tm	__TI_flags+3(%r9),_TIF_NEED_RESCHED
 	bo	BASED(sysc_reschedule)
 	tm	__TI_flags+3(%r9),_TIF_SIGPENDING
@@ -253,6 +268,14 @@ sysc_reschedule:
 	br      %r1		       # call scheduler
 
 #
+# _TIF_MCCK_PENDING is set, call handler
+#
+sysc_mcck_pending:
+	l	%r1,BASED(.Ls390_handle_mcck)
+	la	%r14,BASED(sysc_work_loop)
+	br	%r1			# TIF bit will be cleared by handler
+
+#
 # _TIF_SIGPENDING is set, call do_signal
 #
 sysc_sigpending:     
@@ -430,6 +453,7 @@ pgm_check_handler:
         tm      __LC_PGM_INT_CODE+1,0x80 # check whether we got a per exception
         bnz     BASED(pgm_per)           # got per exception -> special case
 	SAVE_ALL __LC_PGM_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_PGM_OLD_PSW,__LC_SAVE_AREA
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	bz	BASED(pgm_no_vtime)
@@ -468,6 +492,7 @@ pgm_per:
 #
 pgm_per_std:
 	SAVE_ALL __LC_PGM_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_PGM_OLD_PSW,__LC_SAVE_AREA
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	bz	BASED(pgm_no_vtime2)
@@ -493,6 +518,7 @@ pgm_no_vtime2:
 #
 pgm_svcper:
 	SAVE_ALL __LC_SVC_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_SVC_OLD_PSW,__LC_SAVE_AREA
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	bz	BASED(pgm_no_vtime3)
@@ -521,6 +547,7 @@ io_int_handler:
 	stck	__LC_INT_CLOCK
 	SAVE_ALL_BASE __LC_SAVE_AREA+16
         SAVE_ALL __LC_IO_OLD_PSW,__LC_SAVE_AREA+16,0
+	CREATE_STACK_FRAME __LC_IO_OLD_PSW,__LC_SAVE_AREA+16
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	bz	BASED(io_no_vtime)
@@ -578,9 +605,11 @@ io_work:
 	lr	%r15,%r1
 #
 # One of the work bits is on. Find out which one.
-# Checked are: _TIF_SIGPENDING and _TIF_NEED_RESCHED
+# Checked are: _TIF_SIGPENDING, _TIF_NEED_RESCHED and _TIF_MCCK_PENDING
 #
 io_work_loop:
+	tm	__TI_flags+3(%r9),_TIF_MCCK_PENDING
+	bo      BASED(io_mcck_pending)
 	tm	__TI_flags+3(%r9),_TIF_NEED_RESCHED
 	bo	BASED(io_reschedule)
 	tm	__TI_flags+3(%r9),_TIF_SIGPENDING
@@ -588,6 +617,14 @@ io_work_loop:
 	b	BASED(io_leave)
 
 #
+# _TIF_MCCK_PENDING is set, call handler
+#
+io_mcck_pending:
+	l	%r1,BASED(.Ls390_handle_mcck)
+	l	%r14,BASED(io_work_loop)
+	br	%r1		       # TIF bit will be cleared by handler
+
+#
 # _TIF_NEED_RESCHED is set, call schedule
 #	
 io_reschedule:        
@@ -621,6 +658,7 @@ ext_int_handler:
 	stck	__LC_INT_CLOCK
 	SAVE_ALL_BASE __LC_SAVE_AREA+16
         SAVE_ALL __LC_EXT_OLD_PSW,__LC_SAVE_AREA+16,0
+	CREATE_STACK_FRAME __LC_EXT_OLD_PSW,__LC_SAVE_AREA+16
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	bz	BASED(ext_no_vtime)
@@ -642,19 +680,62 @@ ext_no_vtime:
 
         .globl mcck_int_handler
 mcck_int_handler:
-	STORE_TIMER __LC_ASYNC_ENTER_TIMER
+	spt	__LC_CPU_TIMER_SAVE_AREA	# revalidate cpu timer
+	lm	%r0,%r15,__LC_GPREGS_SAVE_AREA	# revalidate gprs
 	SAVE_ALL_BASE __LC_SAVE_AREA+32
-        SAVE_ALL __LC_MCK_OLD_PSW,__LC_SAVE_AREA+32,0
+	la	%r12,__LC_MCK_OLD_PSW
+	tm	__LC_MCCK_CODE,0x80     # system damage?
+	bo	BASED(mcck_int_main)	# yes -> rest of mcck code invalid
+	tm	__LC_MCCK_CODE+5,0x02   # stored cpu timer value valid?
+	bo	BASED(0f)
+	spt	__LC_LAST_UPDATE_TIMER	# revalidate cpu timer
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
-	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
+	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_ASYNC_ENTER_TIMER
+	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_SYNC_ENTER_TIMER
+	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_EXIT_TIMER
+0:	tm	__LC_MCCK_CODE+2,0x08   # mwp of old psw valid?
+	bno	BASED(mcck_no_vtime)	# no -> skip cleanup critical
+	tm	__LC_MCK_OLD_PSW+1,0x01 # interrupting from user ?
 	bz	BASED(mcck_no_vtime)
 	UPDATE_VTIME __LC_EXIT_TIMER,__LC_ASYNC_ENTER_TIMER,__LC_USER_TIMER
 	UPDATE_VTIME __LC_LAST_UPDATE_TIMER,__LC_EXIT_TIMER,__LC_SYSTEM_TIMER
 	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_ASYNC_ENTER_TIMER
 mcck_no_vtime:
 #endif
+0:
+	tm	__LC_MCCK_CODE+2,0x09   # mwp + ia of old psw valid?
+	bno	BASED(mcck_int_main)	# no -> skip cleanup critical
+	tm	__LC_MCK_OLD_PSW+1,0x01	# test problem state bit
+	bnz	BASED(mcck_int_main)	# from user -> load async stack
+	clc	__LC_MCK_OLD_PSW+4(4),BASED(.Lcritical_end)
+	bhe	BASED(mcck_int_main)
+	clc	__LC_MCK_OLD_PSW+4(4),BASED(.Lcritical_start)
+	bl	BASED(mcck_int_main)
+	l	%r14,BASED(.Lcleanup_critical)
+	basr	%r14,%r14
+mcck_int_main:
+	l	%r14,__LC_PANIC_STACK	# are we already on the panic stack?
+	slr	%r14,%r15
+	sra	%r14,PAGE_SHIFT
+	be	BASED(0f)
+	l	%r15,__LC_PANIC_STACK	# load panic stack
+0:	CREATE_STACK_FRAME __LC_MCK_OLD_PSW,__LC_SAVE_AREA+32
+	l	%r9,__LC_THREAD_INFO	# load pointer to thread_info struct
+	la	%r2,SP_PTREGS(%r15)	# load pt_regs
 	l       %r1,BASED(.Ls390_mcck)
-	basr    %r14,%r1	  # call machine check handler
+	basr    %r14,%r1		# call machine check handler
+	tm      SP_PSW+1(%r15),0x01	# returning to user ?
+	bno	BASED(mcck_return)
+	l	%r1,__LC_KERNEL_STACK   # switch to kernel stack
+	s	%r1,BASED(.Lc_spsize)
+	mvc	SP_PTREGS(__PT_SIZE,%r1),SP_PTREGS(%r15)
+	xc      __SF_BACKCHAIN(4,%r1),__SF_BACKCHAIN(%r1) # clear back chain
+	lr	%r15,%r1
+	stosm	__SF_EMPTY(%r15),0x04	# turn dat on
+	tm	__TI_flags+3(%r9),_TIF_MCCK_PENDING
+	bno	BASED(mcck_return)
+	l	%r1,BASED(.Ls390_handle_mcck)
+	basr	%r14,%r1		# call machine check handler
 mcck_return:
         RESTORE_ALL 0
 
@@ -742,7 +823,7 @@ cleanup_critical:
 	clc	4(4,%r12),BASED(cleanup_table_sysc_work_loop)
 	bl	BASED(0f)
 	clc	4(4,%r12),BASED(cleanup_table_sysc_work_loop+4)
-	bl	BASED(cleanup_sysc_leave)
+	bl	BASED(cleanup_sysc_return)
 0:
 	br	%r14
 
@@ -760,6 +841,7 @@ cleanup_system_call:
 	mvc	__LC_SAVE_AREA(16),__LC_SAVE_AREA+16
 0:	st	%r13,__LC_SAVE_AREA+20
 	SAVE_ALL __LC_SVC_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_SVC_OLD_PSW,__LC_SAVE_AREA
 	st	%r15,__LC_SAVE_AREA+28
 	lh	%r7,0x8a
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
@@ -834,6 +916,8 @@ cleanup_sysc_leave_insn:
  * Symbol constants
  */
 .Ls390_mcck:   .long  s390_do_machine_check
+.Ls390_handle_mcck:
+	       .long  s390_handle_mcck
 .Ldo_IRQ:      .long  do_IRQ
 .Ldo_extint:   .long  do_extint
 .Ldo_signal:   .long  do_signal
diff --git a/arch/s390/kernel/entry64.S b/arch/s390/kernel/entry64.S
index 51527ab8c8f..57ca75d0ad7 100644
--- a/arch/s390/kernel/entry64.S
+++ b/arch/s390/kernel/entry64.S
@@ -7,6 +7,7 @@
  *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
  *               Hartmut Penner (hp@de.ibm.com),
  *               Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
+ *		 Heiko Carstens <heiko.carstens@de.ibm.com>
  */
 
 #include <linux/sys.h>
@@ -52,9 +53,9 @@ SP_SIZE      =  STACK_FRAME_OVERHEAD + __PT_SIZE
 STACK_SHIFT = PAGE_SHIFT + THREAD_ORDER
 STACK_SIZE  = 1 << STACK_SHIFT
 
-_TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NEED_RESCHED | \
+_TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NEED_RESCHED | _TIF_MCCK_PENDING | \
 		 _TIF_RESTART_SVC | _TIF_SINGLE_STEP )
-_TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NEED_RESCHED)
+_TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NEED_RESCHED | _TIF_MCCK_PENDING)
 
 #define BASED(name) name-system_call(%r13)
 
@@ -114,7 +115,11 @@ _TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NEED_RESCHED)
 	jz	stack_overflow
 3:
 #endif
-2:	aghi    %r15,-SP_SIZE		# make room for registers & psw
+2:
+	.endm
+
+	.macro	CREATE_STACK_FRAME psworg,savearea
+	aghi    %r15,-SP_SIZE		# make room for registers & psw
 	mvc     SP_PSW(16,%r15),0(%r12)	# move user PSW to stack
 	la	%r12,\psworg
 	stg	%r2,SP_ORIG_R2(%r15)	# store original content of gpr 2
@@ -152,6 +157,13 @@ __switch_to:
         je      __switch_to_noper            # we got away without bashing TLB's
         lctlg   %c9,%c11,__THREAD_per(%r3)	# Nope we didn't
 __switch_to_noper:
+	lg	%r4,__THREAD_info(%r2)              # get thread_info of prev
+	tm	__TI_flags+7(%r4),_TIF_MCCK_PENDING # machine check pending?
+	jz	__switch_to_no_mcck
+	ni	__TI_flags+7(%r4),255-_TIF_MCCK_PENDING # clear flag in prev
+	lg	%r4,__THREAD_info(%r3)		    # get thread_info of next
+	oi	__TI_flags+7(%r4),_TIF_MCCK_PENDING # set it in next
+__switch_to_no_mcck:
         stmg    %r6,%r15,__SF_GPRS(%r15)# store __switch_to registers of prev task
 	stg	%r15,__THREAD_ksp(%r2)	# store kernel stack to prev->tss.ksp
 	lg	%r15,__THREAD_ksp(%r3)	# load kernel stack from next->tss.ksp
@@ -176,6 +188,7 @@ system_call:
 sysc_saveall:
 	SAVE_ALL_BASE __LC_SAVE_AREA
         SAVE_ALL __LC_SVC_OLD_PSW,__LC_SAVE_AREA,1
+        CREATE_STACK_FRAME __LC_SVC_OLD_PSW,__LC_SAVE_AREA
 	llgh    %r7,__LC_SVC_INT_CODE # get svc number from lowcore
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 sysc_vtime:
@@ -232,6 +245,8 @@ sysc_work_loop:
 # One of the work bits is on. Find out which one.
 #
 sysc_work:
+	tm	__TI_flags+7(%r9),_TIF_MCCK_PENDING
+	jo	sysc_mcck_pending
 	tm	__TI_flags+7(%r9),_TIF_NEED_RESCHED
 	jo	sysc_reschedule
 	tm	__TI_flags+7(%r9),_TIF_SIGPENDING
@@ -250,6 +265,13 @@ sysc_reschedule:
         jg      schedule            # return point is sysc_return
 
 #
+# _TIF_MCCK_PENDING is set, call handler
+#
+sysc_mcck_pending:
+	larl	%r14,sysc_work_loop
+	jg	s390_handle_mcck    # TIF bit will be cleared by handler
+
+#
 # _TIF_SIGPENDING is set, call do_signal
 #
 sysc_sigpending:     
@@ -474,6 +496,7 @@ pgm_check_handler:
         tm      __LC_PGM_INT_CODE+1,0x80 # check whether we got a per exception
         jnz     pgm_per                  # got per exception -> special case
 	SAVE_ALL __LC_PGM_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_PGM_OLD_PSW,__LC_SAVE_AREA
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	jz	pgm_no_vtime
@@ -512,6 +535,7 @@ pgm_per:
 #
 pgm_per_std:
 	SAVE_ALL __LC_PGM_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_PGM_OLD_PSW,__LC_SAVE_AREA
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	jz	pgm_no_vtime2
@@ -537,6 +561,7 @@ pgm_no_vtime2:
 #
 pgm_svcper:
 	SAVE_ALL __LC_SVC_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_SVC_OLD_PSW,__LC_SAVE_AREA
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	jz	pgm_no_vtime3
@@ -564,6 +589,7 @@ io_int_handler:
 	stck	__LC_INT_CLOCK
 	SAVE_ALL_BASE __LC_SAVE_AREA+32
         SAVE_ALL __LC_IO_OLD_PSW,__LC_SAVE_AREA+32,0
+	CREATE_STACK_FRAME __LC_IO_OLD_PSW,__LC_SAVE_AREA+32
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	jz	io_no_vtime
@@ -621,9 +647,11 @@ io_work:
 	lgr	%r15,%r1
 #
 # One of the work bits is on. Find out which one.
-# Checked are: _TIF_SIGPENDING and _TIF_NEED_RESCHED
+# Checked are: _TIF_SIGPENDING, _TIF_NEED_RESCHED and _TIF_MCCK_PENDING
 #
 io_work_loop:
+	tm	__TI_flags+7(%r9),_TIF_MCCK_PENDING
+	jo	io_mcck_pending
 	tm	__TI_flags+7(%r9),_TIF_NEED_RESCHED
 	jo	io_reschedule
 	tm	__TI_flags+7(%r9),_TIF_SIGPENDING
@@ -631,6 +659,13 @@ io_work_loop:
 	j	io_leave
 
 #
+# _TIF_MCCK_PENDING is set, call handler
+#
+io_mcck_pending:
+	larl	%r14,io_work_loop
+	jg	s390_handle_mcck	# TIF bit will be cleared by handler
+
+#
 # _TIF_NEED_RESCHED is set, call schedule
 #	
 io_reschedule:        
@@ -661,6 +696,7 @@ ext_int_handler:
 	stck	__LC_INT_CLOCK
 	SAVE_ALL_BASE __LC_SAVE_AREA+32
         SAVE_ALL __LC_EXT_OLD_PSW,__LC_SAVE_AREA+32,0
+	CREATE_STACK_FRAME __LC_EXT_OLD_PSW,__LC_SAVE_AREA+32
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
 	jz	ext_no_vtime
@@ -680,18 +716,60 @@ ext_no_vtime:
  */
         .globl mcck_int_handler
 mcck_int_handler:
-	STORE_TIMER __LC_ASYNC_ENTER_TIMER
+	la	%r1,4095		# revalidate r1
+	spt	__LC_CPU_TIMER_SAVE_AREA-4095(%r1)	# revalidate cpu timer
+  	lmg     %r0,%r15,__LC_GPREGS_SAVE_AREA-4095(%r1)# revalidate gprs
 	SAVE_ALL_BASE __LC_SAVE_AREA+64
-        SAVE_ALL __LC_MCK_OLD_PSW,__LC_SAVE_AREA+64,0
+	la	%r12,__LC_MCK_OLD_PSW
+	tm	__LC_MCCK_CODE,0x80     # system damage?
+	jo	mcck_int_main		# yes -> rest of mcck code invalid
+	tm	__LC_MCCK_CODE+5,0x02   # stored cpu timer value valid?
+	jo	0f
+	spt	__LC_LAST_UPDATE_TIMER
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
-	tm	SP_PSW+1(%r15),0x01	# interrupting from user ?
+	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_ASYNC_ENTER_TIMER
+	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_SYNC_ENTER_TIMER
+	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_EXIT_TIMER
+0:	tm	__LC_MCCK_CODE+2,0x08	# mwp of old psw valid?
+	jno	mcck_no_vtime		# no -> no timer update
+	tm      __LC_MCK_OLD_PSW+1,0x01 # interrupting from user ?
 	jz	mcck_no_vtime
 	UPDATE_VTIME __LC_EXIT_TIMER,__LC_ASYNC_ENTER_TIMER,__LC_USER_TIMER
 	UPDATE_VTIME __LC_LAST_UPDATE_TIMER,__LC_EXIT_TIMER,__LC_SYSTEM_TIMER
 	mvc	__LC_LAST_UPDATE_TIMER(8),__LC_ASYNC_ENTER_TIMER
 mcck_no_vtime:
 #endif
-	brasl   %r14,s390_do_machine_check
+0:
+	tm	__LC_MCCK_CODE+2,0x09   # mwp + ia of old psw valid?
+	jno	mcck_int_main		# no -> skip cleanup critical
+	tm      __LC_MCK_OLD_PSW+1,0x01 # test problem state bit
+	jnz	mcck_int_main		# from user -> load kernel stack
+	clc	__LC_MCK_OLD_PSW+8(8),BASED(.Lcritical_end)
+	jhe	mcck_int_main
+	clc     __LC_MCK_OLD_PSW+8(8),BASED(.Lcritical_start)
+	jl	mcck_int_main
+	brasl   %r14,cleanup_critical
+mcck_int_main:
+	lg      %r14,__LC_PANIC_STACK   # are we already on the panic stack?
+	slgr	%r14,%r15
+	srag	%r14,%r14,PAGE_SHIFT
+	jz	0f
+	lg      %r15,__LC_PANIC_STACK   # load panic stack
+0:	CREATE_STACK_FRAME __LC_MCK_OLD_PSW,__LC_SAVE_AREA+64
+	lg	%r9,__LC_THREAD_INFO	# load pointer to thread_info struct
+	la	%r2,SP_PTREGS(%r15)	# load pt_regs
+	brasl	%r14,s390_do_machine_check
+	tm	SP_PSW+1(%r15),0x01     # returning to user ?
+	jno	mcck_return
+	lg	%r1,__LC_KERNEL_STACK	# switch to kernel stack
+	aghi	%r1,-SP_SIZE
+	mvc	SP_PTREGS(__PT_SIZE,%r1),SP_PTREGS(%r15)
+	xc	__SF_BACKCHAIN(8,%r1),__SF_BACKCHAIN(%r1) # clear back chain
+	lgr	%r15,%r1
+	stosm	__SF_EMPTY(%r15),0x04	# turn dat on
+	tm	__TI_flags+7(%r9),_TIF_MCCK_PENDING
+	jno	mcck_return
+	brasl	%r14,s390_handle_mcck
 mcck_return:
         RESTORE_ALL 0
 
@@ -775,7 +853,7 @@ cleanup_critical:
 	clc	8(8,%r12),BASED(cleanup_table_sysc_work_loop)
 	jl	0f
 	clc	8(8,%r12),BASED(cleanup_table_sysc_work_loop+8)
-	jl	cleanup_sysc_leave
+	jl	cleanup_sysc_return
 0:
 	br	%r14
 
@@ -793,6 +871,7 @@ cleanup_system_call:
 	mvc	__LC_SAVE_AREA(32),__LC_SAVE_AREA+32
 0:	stg	%r13,__LC_SAVE_AREA+40
 	SAVE_ALL __LC_SVC_OLD_PSW,__LC_SAVE_AREA,1
+	CREATE_STACK_FRAME __LC_SVC_OLD_PSW,__LC_SAVE_AREA
 	stg	%r15,__LC_SAVE_AREA+56
 	llgh	%r7,__LC_SVC_INT_CODE
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
diff --git a/arch/s390/kernel/head.S b/arch/s390/kernel/head.S
index b804c55bd91..fc8bf5e285f 100644
--- a/arch/s390/kernel/head.S
+++ b/arch/s390/kernel/head.S
@@ -345,10 +345,25 @@ iplstart:
 	bno   .Lnoreset
         la    %r2,.Lreset              
         lhi   %r3,26
-        .long 0x83230008
+	diag  %r2,%r3,8
+	mvc   0x78(8),.Lrdrnewpsw              # set up IO interrupt psw
+.Lwaitrdrirq:
+	lpsw  .Lrdrwaitpsw
+.Lrdrint:
+	c     %r1,0xb8                         # compare subchannel number
+	bne   .Lwaitrdrirq
+	la    %r5,.Lirb
+	tsch  0(%r5)
 .Lnoreset:
+        b     .Lnoload
+
+	.align 8
+.Lrdrnewpsw:
+	.long  0x00080000,0x80000000+.Lrdrint
+.Lrdrwaitpsw:
+	.long  0x020a0000,0x80000000+.Lrdrint
 #endif
-	
+
 #
 # everything loaded, go for it
 #
@@ -517,10 +532,10 @@ startup:basr  %r13,0                     # get base
 	l     %r2, .Lrcp2-.LPG1(%r13)	# try with Read SCP
 	b     .Lservicecall-.LPG1(%r13)
 .Lprocsccb:
-	lh    %r1,.Lscpincr1-PARMAREA(%r4) # use this one if != 0
-	chi   %r1,0x00
-	jne   .Lscnd
-	l     %r1,.Lscpincr2-PARMAREA(%r4) # otherwise use this one
+	lhi   %r1,0
+	icm   %r1,3,.Lscpincr1-PARMAREA(%r4) # use this one if != 0
+	jnz   .Lscnd
+	l     %r1,.Lscpincr2-PARMAREA+4(%r4) # otherwise use this one
 .Lscnd:
 	xr    %r3,%r3			# same logic
 	ic    %r3,.Lscpa1-PARMAREA(%r4)
diff --git a/arch/s390/kernel/head64.S b/arch/s390/kernel/head64.S
index 8366793bc37..f525c0c2125 100644
--- a/arch/s390/kernel/head64.S
+++ b/arch/s390/kernel/head64.S
@@ -344,10 +344,25 @@ iplstart:
 	bno   .Lnoreset
         la    %r2,.Lreset              
         lhi   %r3,26
-        .long 0x83230008
+	diag  %r2,%r3,8
+	mvc   0x78(8),.Lrdrnewpsw	       # set up IO interrupt psw
+.Lwaitrdrirq:
+	lpsw  .Lrdrwaitpsw
+.Lrdrint:
+	c     %r1,0xb8			       # compare subchannel number
+	bne   .Lwaitrdrirq
+	la    %r5,.Lirb
+	tsch  0(%r5)
 .Lnoreset:
+	b     .Lnoload
+
+	.align 8
+.Lrdrnewpsw:
+	.long  0x00080000,0x80000000+.Lrdrint
+.Lrdrwaitpsw:
+	.long  0x020a0000,0x80000000+.Lrdrint
 #endif
-	
+
 #
 # everything loaded, go for it
 #
@@ -518,9 +533,9 @@ startup:basr  %r13,0                     # get base
 	l     %r2,.Lrcp2-.LPG1(%r13)	# try with Read SCP
 	b     .Lservicecall-.LPG1(%r13)
 .Lprocsccb:
-	lh    %r1,.Lscpincr1-PARMAREA(%r4) # use this one if != 0
-	chi   %r1,0x00
-	jne   .Lscnd
+	lghi  %r1,0
+	icm   %r1,3,.Lscpincr1-PARMAREA(%r4) # use this one if != 0
+	jnz   .Lscnd
 	lg    %r1,.Lscpincr2-PARMAREA(%r4) # otherwise use this one
 .Lscnd:
 	xr    %r3,%r3			# same logic
diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c
new file mode 100644
index 00000000000..2721c3a32b8
--- /dev/null
+++ b/arch/s390/kernel/machine_kexec.c
@@ -0,0 +1,98 @@
+/*
+ * arch/s390/kernel/machine_kexec.c
+ *
+ * (C) Copyright IBM Corp. 2005
+ *
+ * Author(s): Rolf Adelsberger <adelsberger@de.ibm.com>
+ *
+ */
+
+/*
+ * s390_machine_kexec.c - handle the transition of Linux booting another kernel
+ * on the S390 architecture.
+ */
+
+#include <asm/cio.h>
+#include <asm/setup.h>
+#include <linux/device.h>
+#include <linux/mm.h>
+#include <linux/kexec.h>
+#include <linux/delay.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+#include <asm/system.h>
+
+static void kexec_halt_all_cpus(void *);
+
+typedef void (*relocate_kernel_t) (kimage_entry_t *, unsigned long);
+
+const extern unsigned char relocate_kernel[];
+const extern unsigned long long relocate_kernel_len;
+
+int
+machine_kexec_prepare(struct kimage *image)
+{
+	unsigned long reboot_code_buffer;
+
+	/* We don't support anything but the default image type for now. */
+	if (image->type != KEXEC_TYPE_DEFAULT)
+		return -EINVAL;
+
+	/* Get the destination where the assembler code should be copied to.*/
+	reboot_code_buffer = page_to_pfn(image->control_code_page)<<PAGE_SHIFT;
+
+	/* Then copy it */
+	memcpy((void *) reboot_code_buffer, relocate_kernel,
+	       relocate_kernel_len);
+	return 0;
+}
+
+void
+machine_kexec_cleanup(struct kimage *image)
+{
+}
+
+void
+machine_shutdown(void)
+{
+	printk(KERN_INFO "kexec: machine_shutdown called\n");
+}
+
+NORET_TYPE void
+machine_kexec(struct kimage *image)
+{
+	clear_all_subchannels();
+
+	/* Disable lowcore protection */
+	ctl_clear_bit(0,28);
+
+	on_each_cpu(kexec_halt_all_cpus, image, 0, 0);
+	for (;;);
+}
+
+static void
+kexec_halt_all_cpus(void *kernel_image)
+{
+	static atomic_t cpuid = ATOMIC_INIT(-1);
+	int cpu;
+	struct kimage *image;
+	relocate_kernel_t data_mover;
+
+	if (atomic_compare_and_swap(-1, smp_processor_id(), &cpuid))
+		signal_processor(smp_processor_id(), sigp_stop);
+
+	/* Wait for all other cpus to enter stopped state */
+	for_each_online_cpu(cpu) {
+		if (cpu == smp_processor_id())
+			continue;
+		while (!smp_cpu_not_running(cpu))
+			cpu_relax();
+	}
+
+	image = (struct kimage *) kernel_image;
+	data_mover = (relocate_kernel_t)
+		(page_to_pfn(image->control_code_page) << PAGE_SHIFT);
+
+	/* Call the moving routine */
+	(*data_mover) (&image->head, image->start);
+}
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index 7aea25d6e30..9f3dff6c0b7 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -91,13 +91,12 @@ void do_monitor_call(struct pt_regs *regs, long interruption_code)
 			    (void *)(long) smp_processor_id());
 }
 
+extern void s390_handle_mcck(void);
 /*
  * The idle loop on a S390...
  */
 void default_idle(void)
 {
-	psw_t wait_psw;
-	unsigned long reg;
 	int cpu, rc;
 
 	local_irq_disable();
@@ -125,38 +124,17 @@ void default_idle(void)
 		cpu_die();
 #endif
 
-	/* 
-	 * Wait for external, I/O or machine check interrupt and
-	 * switch off machine check bit after the wait has ended.
-	 */
-	wait_psw.mask = PSW_KERNEL_BITS | PSW_MASK_MCHECK | PSW_MASK_WAIT |
-		PSW_MASK_IO | PSW_MASK_EXT;
-#ifndef CONFIG_ARCH_S390X
-	asm volatile (
-		"    basr %0,0\n"
-		"0:  la   %0,1f-0b(%0)\n"
-		"    st   %0,4(%1)\n"
-		"    oi   4(%1),0x80\n"
-		"    lpsw 0(%1)\n"
-		"1:  la   %0,2f-1b(%0)\n"
-		"    st   %0,4(%1)\n"
-		"    oi   4(%1),0x80\n"
-		"    ni   1(%1),0xf9\n"
-		"    lpsw 0(%1)\n"
-		"2:"
-		: "=&a" (reg) : "a" (&wait_psw) : "memory", "cc" );
-#else /* CONFIG_ARCH_S390X */
-	asm volatile (
-		"    larl  %0,0f\n"
-		"    stg   %0,8(%1)\n"
-		"    lpswe 0(%1)\n"
-		"0:  larl  %0,1f\n"
-		"    stg   %0,8(%1)\n"
-		"    ni    1(%1),0xf9\n"
-		"    lpswe 0(%1)\n"
-		"1:"
-		: "=&a" (reg) : "a" (&wait_psw) : "memory", "cc" );
-#endif /* CONFIG_ARCH_S390X */
+	local_mcck_disable();
+	if (test_thread_flag(TIF_MCCK_PENDING)) {
+		local_mcck_enable();
+		local_irq_enable();
+		s390_handle_mcck();
+		return;
+	}
+
+	/* Wait for external, I/O or machine check interrupt. */
+	__load_psw_mask(PSW_KERNEL_BITS | PSW_MASK_WAIT |
+			PSW_MASK_IO | PSW_MASK_EXT);
 }
 
 void cpu_idle(void)
diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c
index 26889366929..06afa3103ac 100644
--- a/arch/s390/kernel/ptrace.c
+++ b/arch/s390/kernel/ptrace.c
@@ -40,6 +40,7 @@
 #include <asm/pgalloc.h>
 #include <asm/system.h>
 #include <asm/uaccess.h>
+#include <asm/unistd.h>
 
 #ifdef CONFIG_S390_SUPPORT
 #include "compat_ptrace.h"
@@ -130,13 +131,19 @@ static int
 peek_user(struct task_struct *child, addr_t addr, addr_t data)
 {
 	struct user *dummy = NULL;
-	addr_t offset, tmp;
+	addr_t offset, tmp, mask;
 
 	/*
 	 * Stupid gdb peeks/pokes the access registers in 64 bit with
 	 * an alignment of 4. Programmers from hell...
 	 */
-	if ((addr & 3) || addr > sizeof(struct user) - __ADDR_MASK)
+	mask = __ADDR_MASK;
+#ifdef CONFIG_ARCH_S390X
+	if (addr >= (addr_t) &dummy->regs.acrs &&
+	    addr < (addr_t) &dummy->regs.orig_gpr2)
+		mask = 3;
+#endif
+	if ((addr & mask) || addr > sizeof(struct user) - __ADDR_MASK)
 		return -EIO;
 
 	if (addr < (addr_t) &dummy->regs.acrs) {
@@ -153,6 +160,16 @@ peek_user(struct task_struct *child, addr_t addr, addr_t data)
 		 * access registers are stored in the thread structure
 		 */
 		offset = addr - (addr_t) &dummy->regs.acrs;
+#ifdef CONFIG_ARCH_S390X
+		/*
+		 * Very special case: old & broken 64 bit gdb reading
+		 * from acrs[15]. Result is a 64 bit value. Read the
+		 * 32 bit acrs[15] value and shift it by 32. Sick...
+		 */
+		if (addr == (addr_t) &dummy->regs.acrs[15])
+			tmp = ((unsigned long) child->thread.acrs[15]) << 32;
+		else
+#endif
 		tmp = *(addr_t *)((addr_t) &child->thread.acrs + offset);
 
 	} else if (addr == (addr_t) &dummy->regs.orig_gpr2) {
@@ -167,6 +184,9 @@ peek_user(struct task_struct *child, addr_t addr, addr_t data)
 		 */
 		offset = addr - (addr_t) &dummy->regs.fp_regs;
 		tmp = *(addr_t *)((addr_t) &child->thread.fp_regs + offset);
+		if (addr == (addr_t) &dummy->regs.fp_regs.fpc)
+			tmp &= (unsigned long) FPC_VALID_MASK
+				<< (BITS_PER_LONG - 32);
 
 	} else if (addr < (addr_t) (&dummy->regs.per_info + 1)) {
 		/*
@@ -191,13 +211,19 @@ static int
 poke_user(struct task_struct *child, addr_t addr, addr_t data)
 {
 	struct user *dummy = NULL;
-	addr_t offset;
+	addr_t offset, mask;
 
 	/*
 	 * Stupid gdb peeks/pokes the access registers in 64 bit with
 	 * an alignment of 4. Programmers from hell indeed...
 	 */
-	if ((addr & 3) || addr > sizeof(struct user) - __ADDR_MASK)
+	mask = __ADDR_MASK;
+#ifdef CONFIG_ARCH_S390X
+	if (addr >= (addr_t) &dummy->regs.acrs &&
+	    addr < (addr_t) &dummy->regs.orig_gpr2)
+		mask = 3;
+#endif
+	if ((addr & mask) || addr > sizeof(struct user) - __ADDR_MASK)
 		return -EIO;
 
 	if (addr < (addr_t) &dummy->regs.acrs) {
@@ -224,6 +250,17 @@ poke_user(struct task_struct *child, addr_t addr, addr_t data)
 		 * access registers are stored in the thread structure
 		 */
 		offset = addr - (addr_t) &dummy->regs.acrs;
+#ifdef CONFIG_ARCH_S390X
+		/*
+		 * Very special case: old & broken 64 bit gdb writing
+		 * to acrs[15] with a 64 bit value. Ignore the lower
+		 * half of the value and write the upper 32 bit to
+		 * acrs[15]. Sick...
+		 */
+		if (addr == (addr_t) &dummy->regs.acrs[15])
+			child->thread.acrs[15] = (unsigned int) (data >> 32);
+		else
+#endif
 		*(addr_t *)((addr_t) &child->thread.acrs + offset) = data;
 
 	} else if (addr == (addr_t) &dummy->regs.orig_gpr2) {
@@ -237,7 +274,8 @@ poke_user(struct task_struct *child, addr_t addr, addr_t data)
 		 * floating point regs. are stored in the thread structure
 		 */
 		if (addr == (addr_t) &dummy->regs.fp_regs.fpc &&
-		    (data & ~FPC_VALID_MASK) != 0)
+		    (data & ~((unsigned long) FPC_VALID_MASK
+			      << (BITS_PER_LONG - 32))) != 0)
 			return -EINVAL;
 		offset = addr - (addr_t) &dummy->regs.fp_regs;
 		*(addr_t *)((addr_t) &child->thread.fp_regs + offset) = data;
@@ -723,6 +761,13 @@ syscall_trace(struct pt_regs *regs, int entryexit)
 				 ? 0x80 : 0));
 
 	/*
+	 * If the debuffer has set an invalid system call number,
+	 * we prepare to skip the system call restart handling.
+	 */
+	if (!entryexit && regs->gprs[2] >= NR_syscalls)
+		regs->trap = -1;
+
+	/*
 	 * this isn't the same as continuing with a signal, but it will do
 	 * for normal use.  strace only continues with a signal if the
 	 * stopping signal is not SIGTRAP.  -brl
diff --git a/arch/s390/kernel/relocate_kernel.S b/arch/s390/kernel/relocate_kernel.S
new file mode 100644
index 00000000000..d5e4a62fbb7
--- /dev/null
+++ b/arch/s390/kernel/relocate_kernel.S
@@ -0,0 +1,81 @@
+/*
+ * arch/s390/kernel/relocate_kernel.S
+ *
+ * (C) Copyright IBM Corp. 2005
+ *
+ * Author(s): Rolf Adelsberger <adelsberger@de.ibm.com>
+ *
+ */
+
+/*
+ * moves the new kernel to its destination...
+ * %r2 = pointer to first kimage_entry_t
+ * %r3 = start address - where to jump to after the job is done...
+ *
+ * %r5 will be used as temp. storage
+ * %r6 holds the destination address
+ * %r7 = PAGE_SIZE
+ * %r8 holds the source address
+ * %r9 = PAGE_SIZE
+ * %r10 is a page mask
+ */
+
+	.text
+	.globl		relocate_kernel
+	relocate_kernel:
+		basr	%r13,0		#base address
+	.base:
+		spx	zero64-.base(%r13)	#absolute addressing mode
+		stnsm	sys_msk-.base(%r13),0xf8	#disable DAT and IRQ (external)
+		lhi	%r10,-1		#preparing the mask
+		sll	%r10,12		#shift it such that it becomes 0xf000
+	.top:
+		lhi	%r7,4096	#load PAGE_SIZE in r7
+		lhi	%r9,4096	#load PAGE_SIZE in r9
+		l	%r5,0(%r2)	#read another word for indirection page
+		ahi	%r2,4		#increment pointer
+		tml	%r5,0x1		#is it a destination page?
+		je	.indir_check	#NO, goto "indir_check"
+		lr	%r6,%r5		#r6 = r5
+		nr	%r6,%r10	#mask it out and...
+		j	.top		#...next iteration
+	.indir_check:
+		tml	%r5,0x2		#is it a indirection page?
+		je	.done_test	#NO, goto "done_test"
+		nr	%r5,%r10	#YES, mask out,
+		lr	%r2,%r5		#move it into the right register,
+		j	.top		#and read next...
+	.done_test:
+		tml	%r5,0x4		#is it the done indicator?
+		je	.source_test	#NO! Well, then it should be the source indicator...
+		j	.done		#ok, lets finish it here...
+	.source_test:
+		tml	%r5,0x8		#it should be a source indicator...
+		je	.top		#NO, ignore it...
+		lr	%r8,%r5		#r8 = r5
+		nr	%r8,%r10	#masking
+	0:	mvcle	%r6,%r8,0x0	#copy PAGE_SIZE bytes from r8 to r6 - pad with 0
+		jo	0b
+		j	.top
+	.done:
+		sr	%r0,%r0		#clear register r0
+		la	%r4,load_psw-.base(%r13)	#load psw-address into the register
+		o	%r3,4(%r4)	#or load address into psw
+		st	%r3,4(%r4)
+		mvc	0(8,%r0),0(%r4)	#copy psw to absolute address 0
+		sr	%r1,%r1		#clear %r1
+		sr	%r2,%r2		#clear %r2
+		sigp	%r1,%r2,0x12	#set cpuid to zero
+		lpsw	0		#hopefully start new kernel...
+
+		.align	8
+	zero64:
+		.quad	0
+	load_psw:
+		.long	0x00080000,0x80000000
+	sys_msk:
+		.quad	0
+	relocate_kernel_end:
+	.globl	relocate_kernel_len
+	relocate_kernel_len:
+		.quad	relocate_kernel_end - relocate_kernel
diff --git a/arch/s390/kernel/relocate_kernel64.S b/arch/s390/kernel/relocate_kernel64.S
new file mode 100644
index 00000000000..96290cc4eb3
--- /dev/null
+++ b/arch/s390/kernel/relocate_kernel64.S
@@ -0,0 +1,82 @@
+/*
+ * arch/s390/kernel/relocate_kernel64.S
+ *
+ * (C) Copyright IBM Corp. 2005
+ *
+ * Author(s): Rolf Adelsberger <adelsberger@de.ibm.com>
+ *
+ */
+
+/*
+ * moves the new kernel to its destination...
+ * %r2 = pointer to first kimage_entry_t
+ * %r3 = start address - where to jump to after the job is done...
+ *
+ * %r5 will be used as temp. storage
+ * %r6 holds the destination address
+ * %r7 = PAGE_SIZE
+ * %r8 holds the source address
+ * %r9 = PAGE_SIZE
+ *
+ * 0xf000 is a page_mask
+ */
+
+	.text
+	.globl		relocate_kernel
+	relocate_kernel:
+		basr	%r13,0		#base address
+	.base:
+		spx	zero64-.base(%r13)	#absolute addressing mode
+		stnsm	sys_msk-.base(%r13),0xf8	#disable DAT and IRQ (external)
+	.top:
+		lghi	%r7,4096	#load PAGE_SIZE in r7
+		lghi	%r9,4096	#load PAGE_SIZE in r9
+		lg	%r5,0(%r2)	#read another word for indirection page
+		aghi	%r2,8		#increment pointer
+		tml	%r5,0x1		#is it a destination page?
+		je	.indir_check	#NO, goto "indir_check"
+		lgr	%r6,%r5		#r6 = r5
+		nill	%r6,0xf000	#mask it out and...
+		j	.top		#...next iteration
+	.indir_check:
+		tml     %r5,0x2		#is it a indirection page?
+		je      .done_test	#NO, goto "done_test"
+		nill    %r5,0xf000	#YES, mask out,
+		lgr     %r2,%r5		#move it into the right register,
+		j       .top		#and read next...
+	.done_test:
+		tml     %r5,0x4		#is it the done indicator?
+		je      .source_test	#NO! Well, then it should be the source indicator...
+		j       .done		#ok, lets finish it here...
+	.source_test:
+		tml     %r5,0x8		#it should be a source indicator...
+		je      .top		#NO, ignore it...
+		lgr     %r8,%r5		#r8 = r5
+		nill    %r8,0xf000	#masking
+	0:	mvcle   %r6,%r8,0x0	#copy PAGE_SIZE bytes from r8 to r6 - pad with 0
+		jo	0b
+		j       .top
+	.done:
+		sgr     %r0,%r0		#clear register r0
+		la      %r4,load_psw-.base(%r13)	#load psw-address into the register
+		o	%r3,4(%r4)	#or load address into psw
+		st	%r3,4(%r4)
+		mvc     0(8,%r0),0(%r4)	#copy psw to absolute address 0
+		sam31			#31 bit mode
+		sr      %r1,%r1		#erase register r1
+		sr      %r2,%r2		#erase register r2
+		sigp    %r1,%r2,0x12	#set cpuid to zero
+		lpsw    0		#hopefully start new kernel...
+
+	        .align	8
+	zero64:
+		.quad	0
+	load_psw:
+		.long	0x00080000,0x80000000
+	sys_msk:
+		.quad	0
+	relocate_kernel_end:
+	.globl	relocate_kernel_len
+	relocate_kernel_len:
+		.quad	relocate_kernel_end - relocate_kernel
+
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index df83215beac..b6d740ac0e6 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -198,11 +198,11 @@ static void __init conmode_default(void)
 	char *ptr;
 
         if (MACHINE_IS_VM) {
-		__cpcmd("QUERY CONSOLE", query_buffer, 1024);
+		__cpcmd("QUERY CONSOLE", query_buffer, 1024, NULL);
 		console_devno = simple_strtoul(query_buffer + 5, NULL, 16);
 		ptr = strstr(query_buffer, "SUBCHANNEL =");
 		console_irq = simple_strtoul(ptr + 13, NULL, 16);
-		__cpcmd("QUERY TERM", query_buffer, 1024);
+		__cpcmd("QUERY TERM", query_buffer, 1024, NULL);
 		ptr = strstr(query_buffer, "CONMODE");
 		/*
 		 * Set the conmode to 3215 so that the device recognition 
@@ -211,7 +211,7 @@ static void __init conmode_default(void)
 		 * 3215 and the 3270 driver will try to access the console
 		 * device (3215 as console and 3270 as normal tty).
 		 */
-		__cpcmd("TERM CONMODE 3215", NULL, 0);
+		__cpcmd("TERM CONMODE 3215", NULL, 0, NULL);
 		if (ptr == NULL) {
 #if defined(CONFIG_SCLP_CONSOLE)
 			SET_CONSOLE_SCLP;
@@ -414,7 +414,8 @@ setup_lowcore(void)
 	lc->program_new_psw.mask = PSW_KERNEL_BITS;
 	lc->program_new_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long)pgm_check_handler;
-	lc->mcck_new_psw.mask = PSW_KERNEL_BITS;
+	lc->mcck_new_psw.mask =
+		PSW_KERNEL_BITS & ~PSW_MASK_MCHECK & ~PSW_MASK_DAT;
 	lc->mcck_new_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long) mcck_int_handler;
 	lc->io_new_psw.mask = PSW_KERNEL_BITS;
@@ -424,12 +425,18 @@ setup_lowcore(void)
 	lc->kernel_stack = ((unsigned long) &init_thread_union) + THREAD_SIZE;
 	lc->async_stack = (unsigned long)
 		__alloc_bootmem(ASYNC_SIZE, ASYNC_SIZE, 0) + ASYNC_SIZE;
-#ifdef CONFIG_CHECK_STACK
 	lc->panic_stack = (unsigned long)
 		__alloc_bootmem(PAGE_SIZE, PAGE_SIZE, 0) + PAGE_SIZE;
-#endif
 	lc->current_task = (unsigned long) init_thread_union.thread_info.task;
 	lc->thread_info = (unsigned long) &init_thread_union;
+#ifndef CONFIG_ARCH_S390X
+	if (MACHINE_HAS_IEEE) {
+		lc->extended_save_area_addr = (__u32)
+			__alloc_bootmem(PAGE_SIZE, PAGE_SIZE, 0);
+		/* enable extended save area */
+		ctl_set_bit(14, 29);
+	}
+#endif
 #ifdef CONFIG_ARCH_S390X
 	if (MACHINE_HAS_DIAG44)
 		lc->diag44_opcode = 0x83000044;
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index fdfcf0488b4..642572a8e33 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -284,7 +284,7 @@ static void do_machine_restart(void * __unused)
 	 * locks are always held disabled).
 	 */
 	if (MACHINE_IS_VM)
-		cpcmd ("IPL", NULL, 0);
+		cpcmd ("IPL", NULL, 0, NULL);
 	else
 		reipl (0x10000 | S390_lowcore.ipl_device);
 }
@@ -313,7 +313,7 @@ static void do_machine_halt(void * __unused)
 	if (atomic_compare_and_swap(-1, smp_processor_id(), &cpuid) == 0) {
 		smp_send_stop();
 		if (MACHINE_IS_VM && strlen(vmhalt_cmd) > 0)
-			cpcmd(vmhalt_cmd, NULL, 0);
+			cpcmd(vmhalt_cmd, NULL, 0, NULL);
 		signal_processor(smp_processor_id(),
 				 sigp_stop_and_store_status);
 	}
@@ -332,7 +332,7 @@ static void do_machine_power_off(void * __unused)
 	if (atomic_compare_and_swap(-1, smp_processor_id(), &cpuid) == 0) {
 		smp_send_stop();
 		if (MACHINE_IS_VM && strlen(vmpoff_cmd) > 0)
-			cpcmd(vmpoff_cmd, NULL, 0);
+			cpcmd(vmpoff_cmd, NULL, 0, NULL);
 		signal_processor(smp_processor_id(),
 				 sigp_stop_and_store_status);
 	}
@@ -679,12 +679,14 @@ __cpu_disable(void)
 {
 	unsigned long flags;
 	ec_creg_mask_parms cr_parms;
+	int cpu = smp_processor_id();
 
 	spin_lock_irqsave(&smp_reserve_lock, flags);
-	if (smp_cpu_reserved[smp_processor_id()] != 0) {
+	if (smp_cpu_reserved[cpu] != 0) {
 		spin_unlock_irqrestore(&smp_reserve_lock, flags);
 		return -EBUSY;
 	}
+	cpu_clear(cpu, cpu_online_map);
 
 #ifdef CONFIG_PFAULT
 	/* Disable pfault pseudo page faults on this cpu. */
@@ -771,13 +773,24 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
 
 		*(lowcore_ptr[i]) = S390_lowcore;
 		lowcore_ptr[i]->async_stack = stack + (ASYNC_SIZE);
-#ifdef CONFIG_CHECK_STACK
 		stack = __get_free_pages(GFP_KERNEL,0);
 		if (stack == 0ULL)
 			panic("smp_boot_cpus failed to allocate memory\n");
 		lowcore_ptr[i]->panic_stack = stack + (PAGE_SIZE);
+#ifndef __s390x__
+		if (MACHINE_HAS_IEEE) {
+			lowcore_ptr[i]->extended_save_area_addr =
+				(__u32) __get_free_pages(GFP_KERNEL,0);
+			if (lowcore_ptr[i]->extended_save_area_addr == 0)
+				panic("smp_boot_cpus failed to "
+				      "allocate memory\n");
+		}
 #endif
 	}
+#ifndef __s390x__
+	if (MACHINE_HAS_IEEE)
+		ctl_set_bit(14, 29); /* enable extended save area */
+#endif
 	set_prefix((u32)(unsigned long) lowcore_ptr[smp_processor_id()]);
 
 	for_each_cpu(cpu)
diff --git a/arch/s390/kernel/syscalls.S b/arch/s390/kernel/syscalls.S
index 515938628f8..a8668afb5f8 100644
--- a/arch/s390/kernel/syscalls.S
+++ b/arch/s390/kernel/syscalls.S
@@ -285,7 +285,7 @@ SYSCALL(sys_mq_timedsend,sys_mq_timedsend,compat_sys_mq_timedsend_wrapper)
 SYSCALL(sys_mq_timedreceive,sys_mq_timedreceive,compat_sys_mq_timedreceive_wrapper)
 SYSCALL(sys_mq_notify,sys_mq_notify,compat_sys_mq_notify_wrapper) /* 275 */
 SYSCALL(sys_mq_getsetattr,sys_mq_getsetattr,compat_sys_mq_getsetattr_wrapper)
-NI_SYSCALL							/* reserved for kexec */
+SYSCALL(sys_kexec_load,sys_kexec_load,compat_sys_kexec_load_wrapper)
 SYSCALL(sys_add_key,sys_add_key,compat_sys_add_key_wrapper)
 SYSCALL(sys_request_key,sys_request_key,compat_sys_request_key_wrapper)
 SYSCALL(sys_keyctl,sys_keyctl,compat_sys_keyctl)		/* 280 */
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index 8b90e9528b9..bc7b7be7acb 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -668,7 +668,10 @@ asmlinkage void space_switch_exception(struct pt_regs * regs, long int_code)
 
 asmlinkage void kernel_stack_overflow(struct pt_regs * regs)
 {
-	die("Kernel stack overflow", regs, 0);
+	bust_spinlocks(1);
+	printk("Kernel stack overflow.\n");
+	show_regs(regs);
+	bust_spinlocks(0);
 	panic("Corrupt kernel stack, can't continue.");
 }
 
@@ -732,7 +735,7 @@ void __init trap_init(void)
 						    &ext_int_pfault);
 #endif
 #ifndef CONFIG_ARCH_S390X
-		cpcmd("SET PAGEX ON", NULL, 0);
+		cpcmd("SET PAGEX ON", NULL, 0, NULL);
 #endif
 	}
 }
diff --git a/arch/s390/mm/cmm.c b/arch/s390/mm/cmm.c
index f5a5bc09b8f..2d5cb138575 100644
--- a/arch/s390/mm/cmm.c
+++ b/arch/s390/mm/cmm.c
@@ -21,7 +21,7 @@
 #include <asm/uaccess.h>
 
 static char *sender = "VMRMSVM";
-module_param(sender, charp, 0);
+module_param(sender, charp, 0400);
 MODULE_PARM_DESC(sender,
 		 "Guest name that may send SMSG messages (default VMRMSVM)");
 
diff --git a/arch/s390/mm/extmem.c b/arch/s390/mm/extmem.c
index 648deed17e2..c5348108ca3 100644
--- a/arch/s390/mm/extmem.c
+++ b/arch/s390/mm/extmem.c
@@ -576,8 +576,8 @@ segment_save(char *name)
 			segtype_string[seg->range[i].start & 0xff]);
 	}
 	sprintf(cmd2, "SAVESEG %s", name);
-	cpcmd(cmd1, NULL, 0);
-	cpcmd(cmd2, NULL, 0);
+	cpcmd(cmd1, NULL, 0, NULL);
+	cpcmd(cmd2, NULL, 0, NULL);
 	spin_unlock(&dcss_lock);
 }
 
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 80306bc8c79..75fde949d12 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -207,7 +207,7 @@ do_exception(struct pt_regs *regs, unsigned long error_code, int is_protection)
 	 * we are not in an interrupt and that there is a 
 	 * user context.
 	 */
-        if (user_address == 0 || in_interrupt() || !mm)
+        if (user_address == 0 || in_atomic() || !mm)
                 goto no_context;
 
 	/*