Merge ARM fixes

34 files changed, 1990 insertions, 797 deletions

diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile
index a81881c9b297..5492d25d7d69 100644
--- a/arch/s390/kernel/Makefile
+++ b/arch/s390/kernel/Makefile
@@ -4,9 +4,9 @@
 
 EXTRA_AFLAGS	:= -traditional
 
-obj-y	:=  bitmap.o traps.o time.o process.o reset.o \
+obj-y	:=  bitmap.o traps.o time.o process.o base.o early.o \
             setup.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o \
-	    semaphore.o s390_ext.o debug.o profile.o irq.o ipl.o
+	    semaphore.o s390_ext.o debug.o irq.o ipl.o
 
 obj-y	+= $(if $(CONFIG_64BIT),entry64.o,entry.o)
 obj-y	+= $(if $(CONFIG_64BIT),reipl64.o,reipl.o)
diff --git a/arch/s390/kernel/base.S b/arch/s390/kernel/base.S
new file mode 100644
index 000000000000..dc7e5259770f
--- /dev/null
+++ b/arch/s390/kernel/base.S
@@ -0,0 +1,150 @@
+/*
+ *  arch/s390/kernel/base.S
+ *
+ *    Copyright IBM Corp. 2006,2007
+ *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ *		 Michael Holzheu <holzheu@de.ibm.com>
+ */
+
+#include <asm/ptrace.h>
+#include <asm/lowcore.h>
+
+#ifdef CONFIG_64BIT
+
+	.globl	s390_base_mcck_handler
+s390_base_mcck_handler:
+	basr	%r13,0
+0:	lg	%r15,__LC_PANIC_STACK	# load panic stack
+	aghi	%r15,-STACK_FRAME_OVERHEAD
+	larl	%r1,s390_base_mcck_handler_fn
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	la	%r1,4095
+	lmg	%r0,%r15,__LC_GPREGS_SAVE_AREA-4095(%r1)
+	lpswe	__LC_MCK_OLD_PSW
+
+	.section .bss
+	.globl	s390_base_mcck_handler_fn
+s390_base_mcck_handler_fn:
+	.quad	0
+	.previous
+
+	.globl	s390_base_ext_handler
+s390_base_ext_handler:
+	stmg	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	aghi	%r15,-STACK_FRAME_OVERHEAD
+	larl	%r1,s390_base_ext_handler_fn
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	lmg	%r0,%r15,__LC_SAVE_AREA
+	ni	__LC_EXT_OLD_PSW+1,0xfd	# clear wait state bit
+	lpswe	__LC_EXT_OLD_PSW
+
+	.section .bss
+	.globl s390_base_ext_handler_fn
+s390_base_ext_handler_fn:
+	.quad	0
+	.previous
+
+	.globl	s390_base_pgm_handler
+s390_base_pgm_handler:
+	stmg	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	aghi	%r15,-STACK_FRAME_OVERHEAD
+	larl	%r1,s390_base_pgm_handler_fn
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+	lmg	%r0,%r15,__LC_SAVE_AREA
+	lpswe	__LC_PGM_OLD_PSW
+1:	lpswe	disabled_wait_psw-0b(%r13)
+
+	.align	8
+disabled_wait_psw:
+	.quad	0x0002000180000000,0x0000000000000000 + s390_base_pgm_handler
+
+	.section .bss
+	.globl s390_base_pgm_handler_fn
+s390_base_pgm_handler_fn:
+	.quad	0
+	.previous
+
+#else /* CONFIG_64BIT */
+
+	.globl	s390_base_mcck_handler
+s390_base_mcck_handler:
+	basr	%r13,0
+0:	l	%r15,__LC_PANIC_STACK	# load panic stack
+	ahi	%r15,-STACK_FRAME_OVERHEAD
+	l	%r1,2f-0b(%r13)
+	l	%r1,0(%r1)
+	ltr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	lm	%r0,%r15,__LC_GPREGS_SAVE_AREA
+	lpsw	__LC_MCK_OLD_PSW
+
+2:	.long	s390_base_mcck_handler_fn
+
+	.section .bss
+	.globl	s390_base_mcck_handler_fn
+s390_base_mcck_handler_fn:
+	.long	0
+	.previous
+
+	.globl	s390_base_ext_handler
+s390_base_ext_handler:
+	stm	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	ahi	%r15,-STACK_FRAME_OVERHEAD
+	l	%r1,2f-0b(%r13)
+	l	%r1,0(%r1)
+	ltr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	lm	%r0,%r15,__LC_SAVE_AREA
+	ni	__LC_EXT_OLD_PSW+1,0xfd	# clear wait state bit
+	lpsw	__LC_EXT_OLD_PSW
+
+2:	.long	s390_base_ext_handler_fn
+
+	.section .bss
+	.globl	s390_base_ext_handler_fn
+s390_base_ext_handler_fn:
+	.long	0
+	.previous
+
+	.globl	s390_base_pgm_handler
+s390_base_pgm_handler:
+	stm	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	ahi	%r15,-STACK_FRAME_OVERHEAD
+	l	%r1,2f-0b(%r13)
+	l	%r1,0(%r1)
+	ltr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+	lm	%r0,%r15,__LC_SAVE_AREA
+	lpsw	__LC_PGM_OLD_PSW
+
+1:	lpsw	disabled_wait_psw-0b(%r13)
+
+2:	.long	s390_base_pgm_handler_fn
+
+disabled_wait_psw:
+	.align	8
+	.long	0x000a0000,0x00000000 + s390_base_pgm_handler
+
+	.section .bss
+	.globl	s390_base_pgm_handler_fn
+s390_base_pgm_handler_fn:
+	.long	0
+	.previous
+
+#endif /* CONFIG_64BIT */
diff --git a/arch/s390/kernel/binfmt_elf32.c b/arch/s390/kernel/binfmt_elf32.c
index 5c46054195cb..f1e40ca00d8d 100644
--- a/arch/s390/kernel/binfmt_elf32.c
+++ b/arch/s390/kernel/binfmt_elf32.c
@@ -192,7 +192,7 @@ MODULE_AUTHOR("Gerhard Tonn <ton@de.ibm.com>");
 
 #undef cputime_to_timeval
 #define cputime_to_timeval cputime_to_compat_timeval
-static __inline__ void
+static inline void
 cputime_to_compat_timeval(const cputime_t cputime, struct compat_timeval *value)
 {
 	value->tv_usec = cputime % 1000000;
diff --git a/arch/s390/kernel/compat_exec_domain.c b/arch/s390/kernel/compat_exec_domain.c
index 71d27c493568..914d49444f92 100644
--- a/arch/s390/kernel/compat_exec_domain.c
+++ b/arch/s390/kernel/compat_exec_domain.c
@@ -12,10 +12,9 @@
 #include <linux/personality.h>
 #include <linux/sched.h>
 
-struct exec_domain s390_exec_domain;
+static struct exec_domain s390_exec_domain;
 
-static int __init
-s390_init (void)
+static int __init s390_init (void)
 {
 	s390_exec_domain.name = "Linux/s390";
 	s390_exec_domain.handler = NULL;
diff --git a/arch/s390/kernel/compat_linux.c b/arch/s390/kernel/compat_linux.c
index 5b33f823863a..664c669b1856 100644
--- a/arch/s390/kernel/compat_linux.c
+++ b/arch/s390/kernel/compat_linux.c
@@ -69,6 +69,12 @@
 
 #include "compat_linux.h"
 
+long psw_user32_bits	= (PSW_BASE32_BITS | PSW_MASK_DAT | PSW_ASC_HOME |
+			   PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			   PSW_MASK_PSTATE | PSW_DEFAULT_KEY);
+long psw32_user_bits	= (PSW32_BASE_BITS | PSW32_MASK_DAT | PSW32_ASC_HOME |
+			   PSW32_MASK_IO | PSW32_MASK_EXT | PSW32_MASK_MCHECK |
+			   PSW32_MASK_PSTATE);
  
 /* For this source file, we want overflow handling. */
 
@@ -392,51 +398,6 @@ int cp_compat_stat(struct kstat *stat, struct compat_stat __user *statbuf)
 	return err;
 }
 
-struct sysinfo32 {
-        s32 uptime;
-        u32 loads[3];
-        u32 totalram;
-        u32 freeram;
-        u32 sharedram;
-        u32 bufferram;
-        u32 totalswap;
-        u32 freeswap;
-        unsigned short procs;
-	unsigned short pads;
-	u32 totalhigh;
-	u32 freehigh;
-	unsigned int mem_unit;
-        char _f[8];
-};
-
-asmlinkage long sys32_sysinfo(struct sysinfo32 __user *info)
-{
-	struct sysinfo s;
-	int ret, err;
-	mm_segment_t old_fs = get_fs ();
-	
-	set_fs (KERNEL_DS);
-	ret = sys_sysinfo((struct sysinfo __user *) &s);
-	set_fs (old_fs);
-	err = put_user (s.uptime, &info->uptime);
-	err |= __put_user (s.loads[0], &info->loads[0]);
-	err |= __put_user (s.loads[1], &info->loads[1]);
-	err |= __put_user (s.loads[2], &info->loads[2]);
-	err |= __put_user (s.totalram, &info->totalram);
-	err |= __put_user (s.freeram, &info->freeram);
-	err |= __put_user (s.sharedram, &info->sharedram);
-	err |= __put_user (s.bufferram, &info->bufferram);
-	err |= __put_user (s.totalswap, &info->totalswap);
-	err |= __put_user (s.freeswap, &info->freeswap);
-	err |= __put_user (s.procs, &info->procs);
-	err |= __put_user (s.totalhigh, &info->totalhigh);
-	err |= __put_user (s.freehigh, &info->freehigh);
-	err |= __put_user (s.mem_unit, &info->mem_unit);
-	if (err)
-		return -EFAULT;
-	return ret;
-}
-
 asmlinkage long sys32_sched_rr_get_interval(compat_pid_t pid,
 				struct compat_timespec __user *interval)
 {
@@ -445,7 +406,8 @@ asmlinkage long sys32_sched_rr_get_interval(compat_pid_t pid,
 	mm_segment_t old_fs = get_fs ();
 	
 	set_fs (KERNEL_DS);
-	ret = sys_sched_rr_get_interval(pid, (struct timespec __user *) &t);
+	ret = sys_sched_rr_get_interval(pid,
+					(struct timespec __force __user *) &t);
 	set_fs (old_fs);
 	if (put_compat_timespec(&t, interval))
 		return -EFAULT;
@@ -472,8 +434,8 @@ asmlinkage long sys32_rt_sigprocmask(int how, compat_sigset_t __user *set,
 	}
 	set_fs (KERNEL_DS);
 	ret = sys_rt_sigprocmask(how,
-				 set ? (sigset_t __user *) &s : NULL,
-				 oset ? (sigset_t __user *) &s : NULL,
+				 set ? (sigset_t __force __user *) &s : NULL,
+				 oset ? (sigset_t __force __user *) &s : NULL,
 				 sigsetsize);
 	set_fs (old_fs);
 	if (ret) return ret;
@@ -499,7 +461,7 @@ asmlinkage long sys32_rt_sigpending(compat_sigset_t __user *set,
 	mm_segment_t old_fs = get_fs();
 		
 	set_fs (KERNEL_DS);
-	ret = sys_rt_sigpending((sigset_t __user *) &s, sigsetsize);
+	ret = sys_rt_sigpending((sigset_t __force __user *) &s, sigsetsize);
 	set_fs (old_fs);
 	if (!ret) {
 		switch (_NSIG_WORDS) {
@@ -524,7 +486,7 @@ sys32_rt_sigqueueinfo(int pid, int sig, compat_siginfo_t __user *uinfo)
 	if (copy_siginfo_from_user32(&info, uinfo))
 		return -EFAULT;
 	set_fs (KERNEL_DS);
-	ret = sys_rt_sigqueueinfo(pid, sig, (siginfo_t __user *) &info);
+	ret = sys_rt_sigqueueinfo(pid, sig, (siginfo_t __force __user *) &info);
 	set_fs (old_fs);
 	return ret;
 }
@@ -682,7 +644,7 @@ asmlinkage long sys32_sendfile(int out_fd, int in_fd, compat_off_t __user *offse
 		
 	set_fs(KERNEL_DS);
 	ret = sys_sendfile(out_fd, in_fd,
-			   offset ? (off_t __user *) &of : NULL, count);
+			   offset ? (off_t __force __user *) &of : NULL, count);
 	set_fs(old_fs);
 	
 	if (offset && put_user(of, offset))
@@ -703,7 +665,8 @@ asmlinkage long sys32_sendfile64(int out_fd, int in_fd,
 		
 	set_fs(KERNEL_DS);
 	ret = sys_sendfile64(out_fd, in_fd,
-			     offset ? (loff_t __user *) &lof : NULL, count);
+			     offset ? (loff_t __force __user *) &lof : NULL,
+			     count);
 	set_fs(old_fs);
 	
 	if (offset && put_user(lof, offset))
diff --git a/arch/s390/kernel/compat_linux.h b/arch/s390/kernel/compat_linux.h
index 1a18e29668ef..e89f8c0c42a0 100644
--- a/arch/s390/kernel/compat_linux.h
+++ b/arch/s390/kernel/compat_linux.h
@@ -115,37 +115,6 @@ typedef struct
         __u32	addr;
 } _psw_t32 __attribute__ ((aligned(8)));
 
-#define PSW32_MASK_PER		0x40000000UL
-#define PSW32_MASK_DAT		0x04000000UL
-#define PSW32_MASK_IO		0x02000000UL
-#define PSW32_MASK_EXT		0x01000000UL
-#define PSW32_MASK_KEY		0x00F00000UL
-#define PSW32_MASK_MCHECK	0x00040000UL
-#define PSW32_MASK_WAIT		0x00020000UL
-#define PSW32_MASK_PSTATE	0x00010000UL
-#define PSW32_MASK_ASC		0x0000C000UL
-#define PSW32_MASK_CC		0x00003000UL
-#define PSW32_MASK_PM		0x00000f00UL
-
-#define PSW32_ADDR_AMODE31	0x80000000UL
-#define PSW32_ADDR_INSN		0x7FFFFFFFUL
-
-#define PSW32_BASE_BITS		0x00080000UL
-
-#define PSW32_ASC_PRIMARY	0x00000000UL
-#define PSW32_ASC_ACCREG	0x00004000UL
-#define PSW32_ASC_SECONDARY	0x00008000UL
-#define PSW32_ASC_HOME		0x0000C000UL
-
-#define PSW32_USER_BITS	(PSW32_BASE_BITS | PSW32_MASK_DAT | PSW32_ASC_HOME | \
-			 PSW32_MASK_IO | PSW32_MASK_EXT | PSW32_MASK_MCHECK | \
-			 PSW32_MASK_PSTATE)
-
-#define PSW32_MASK_MERGE(CURRENT,NEW) \
-        (((CURRENT) & ~(PSW32_MASK_CC|PSW32_MASK_PM)) | \
-         ((NEW) & (PSW32_MASK_CC|PSW32_MASK_PM)))
-
-
 typedef struct
 {
 	_psw_t32	psw;
diff --git a/arch/s390/kernel/compat_signal.c b/arch/s390/kernel/compat_signal.c
index 861888ab8c13..887a9881d0d0 100644
--- a/arch/s390/kernel/compat_signal.c
+++ b/arch/s390/kernel/compat_signal.c
@@ -275,8 +275,8 @@ sys32_sigaltstack(const stack_t32 __user *uss, stack_t32 __user *uoss,
 	}
 
 	set_fs (KERNEL_DS);
-	ret = do_sigaltstack((stack_t __user *) (uss ? &kss : NULL),
-			     (stack_t __user *) (uoss ? &koss : NULL),
+	ret = do_sigaltstack((stack_t __force __user *) (uss ? &kss : NULL),
+			     (stack_t __force __user *) (uoss ? &koss : NULL),
 			     regs->gprs[15]);
 	set_fs (old_fs);
 
@@ -298,7 +298,7 @@ static int save_sigregs32(struct pt_regs *regs, _sigregs32 __user *sregs)
 	_s390_regs_common32 regs32;
 	int err, i;
 
-	regs32.psw.mask = PSW32_MASK_MERGE(PSW32_USER_BITS,
+	regs32.psw.mask = PSW32_MASK_MERGE(psw32_user_bits,
 					   (__u32)(regs->psw.mask >> 32));
 	regs32.psw.addr = PSW32_ADDR_AMODE31 | (__u32) regs->psw.addr;
 	for (i = 0; i < NUM_GPRS; i++)
@@ -401,7 +401,7 @@ asmlinkage long sys32_rt_sigreturn(struct pt_regs *regs)
 		goto badframe; 
 
 	set_fs (KERNEL_DS);
-	do_sigaltstack((stack_t __user *)&st, NULL, regs->gprs[15]);
+	do_sigaltstack((stack_t __force __user *)&st, NULL, regs->gprs[15]);
 	set_fs (old_fs);
 
 	return regs->gprs[2];
diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S
index 71e54ef0931e..97901296894e 100644
--- a/arch/s390/kernel/compat_wrapper.S
+++ b/arch/s390/kernel/compat_wrapper.S
@@ -517,10 +517,10 @@ sys32_swapoff_wrapper:
 	llgtr	%r2,%r2			# const char *
 	jg	sys_swapoff		# branch to system call
 
-	.globl	sys32_sysinfo_wrapper
-sys32_sysinfo_wrapper:
+	.globl	compat_sys_sysinfo_wrapper
+compat_sys_sysinfo_wrapper:
 	llgtr	%r2,%r2			# struct sysinfo_emu31 *
-	jg	sys32_sysinfo		# branch to system call
+	jg	compat_sys_sysinfo	# branch to system call
 
 	.globl	sys32_ipc_wrapper
 sys32_ipc_wrapper:
diff --git a/arch/s390/kernel/cpcmd.c b/arch/s390/kernel/cpcmd.c
index a5972f1541fe..6c89f30c8e31 100644
--- a/arch/s390/kernel/cpcmd.c
+++ b/arch/s390/kernel/cpcmd.c
@@ -16,6 +16,7 @@
 #include <asm/ebcdic.h>
 #include <asm/cpcmd.h>
 #include <asm/system.h>
+#include <asm/io.h>
 
 static DEFINE_SPINLOCK(cpcmd_lock);
 static char cpcmd_buf[241];
@@ -88,13 +89,8 @@ int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 	int len;
 	unsigned long flags;
 
-	if ((rlen == 0) || (response == NULL)
-	    || !((unsigned long)response >> 31)) {
-		spin_lock_irqsave(&cpcmd_lock, flags);
-		len = __cpcmd(cmd, response, rlen, response_code);
-		spin_unlock_irqrestore(&cpcmd_lock, flags);
-	}
-	else {
+	if ((virt_to_phys(response) != (unsigned long) response) ||
+			(((unsigned long)response + rlen) >> 31)) {
 		lowbuf = kmalloc(rlen, GFP_KERNEL | GFP_DMA);
 		if (!lowbuf) {
 			printk(KERN_WARNING
@@ -106,6 +102,10 @@ int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 		spin_unlock_irqrestore(&cpcmd_lock, flags);
 		memcpy(response, lowbuf, rlen);
 		kfree(lowbuf);
+	} else {
+		spin_lock_irqsave(&cpcmd_lock, flags);
+		len = __cpcmd(cmd, response, rlen, response_code);
+		spin_unlock_irqrestore(&cpcmd_lock, flags);
 	}
 	return len;
 }
diff --git a/arch/s390/kernel/crash.c b/arch/s390/kernel/crash.c
index 926cceeae0fa..8cc7c9fa64f5 100644
--- a/arch/s390/kernel/crash.c
+++ b/arch/s390/kernel/crash.c
@@ -9,6 +9,7 @@
 
 #include <linux/threads.h>
 #include <linux/kexec.h>
+#include <linux/reboot.h>
 
 void machine_crash_shutdown(struct pt_regs *regs)
 {
diff --git a/arch/s390/kernel/debug.c b/arch/s390/kernel/debug.c
index bb57bc0e3fc8..eca3fe595ff4 100644
--- a/arch/s390/kernel/debug.c
+++ b/arch/s390/kernel/debug.c
@@ -120,7 +120,7 @@ struct debug_view debug_hex_ascii_view = {
 	NULL
 };
 
-struct debug_view debug_level_view = {
+static struct debug_view debug_level_view = {
 	"level",
 	&debug_prolog_level_fn,
 	NULL,
@@ -129,7 +129,7 @@ struct debug_view debug_level_view = {
 	NULL
 };
 
-struct debug_view debug_pages_view = {
+static struct debug_view debug_pages_view = {
 	"pages",
 	&debug_prolog_pages_fn,
 	NULL,
@@ -138,7 +138,7 @@ struct debug_view debug_pages_view = {
 	NULL
 };
 
-struct debug_view debug_flush_view = {
+static struct debug_view debug_flush_view = {
         "flush",
         NULL,
         NULL,
@@ -156,18 +156,18 @@ struct debug_view debug_sprintf_view = {
 	NULL
 };
 
-
+/* used by dump analysis tools to determine version of debug feature */
 unsigned int debug_feature_version = __DEBUG_FEATURE_VERSION;
 
 /* static globals */
 
 static debug_info_t *debug_area_first = NULL;
 static debug_info_t *debug_area_last = NULL;
-DECLARE_MUTEX(debug_lock);
+static DECLARE_MUTEX(debug_lock);
 
 static int initialized;
 
-static struct file_operations debug_file_ops = {
+static const struct file_operations debug_file_ops = {
 	.owner   = THIS_MODULE,
 	.read    = debug_output,
 	.write   = debug_input,
@@ -852,7 +852,6 @@ debug_finish_entry(debug_info_t * id, debug_entry_t* active, int level,
 static int debug_stoppable=1;
 static int debug_active=1;
 
-#define CTL_S390DBF 5677
 #define CTL_S390DBF_STOPPABLE 5678
 #define CTL_S390DBF_ACTIVE 5679
 
@@ -905,7 +904,7 @@ static struct ctl_table s390dbf_dir_table[] = {
 	{ .ctl_name = 0 }
 };
 
-struct ctl_table_header *s390dbf_sysctl_header;
+static struct ctl_table_header *s390dbf_sysctl_header;
 
 void
 debug_stop_all(void)
@@ -1054,7 +1053,7 @@ __init debug_init(void)
 {
 	int rc = 0;
 
-	s390dbf_sysctl_header = register_sysctl_table(s390dbf_dir_table, 1);
+	s390dbf_sysctl_header = register_sysctl_table(s390dbf_dir_table);
 	down(&debug_lock);
 	debug_debugfs_root_entry = debugfs_create_dir(DEBUG_DIR_ROOT,NULL);
 	printk(KERN_INFO "debug: Initialization complete\n");
@@ -1300,8 +1299,7 @@ out:
  * flushes debug areas
  */
  
-void
-debug_flush(debug_info_t* id, int area)
+static void debug_flush(debug_info_t* id, int area)
 {
         unsigned long flags;
         int i,j;
@@ -1511,8 +1509,7 @@ out:
 /*
  * clean up module
  */
-void
-__exit debug_exit(void)
+static void __exit debug_exit(void)
 {
 	debugfs_remove(debug_debugfs_root_entry);
 	unregister_sysctl_table(s390dbf_sysctl_header);
diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c
new file mode 100644
index 000000000000..e518dd53eff5
--- /dev/null
+++ b/arch/s390/kernel/early.c
@@ -0,0 +1,306 @@
+/*
+ *  arch/s390/kernel/early.c
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Hongjie Yang <hongjie@us.ibm.com>,
+ *		 Heiko Carstens <heiko.carstens@de.ibm.com>
+ */
+
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/ctype.h>
+#include <linux/lockdep.h>
+#include <linux/module.h>
+#include <linux/pfn.h>
+#include <linux/uaccess.h>
+#include <asm/lowcore.h>
+#include <asm/processor.h>
+#include <asm/sections.h>
+#include <asm/setup.h>
+#include <asm/cpcmd.h>
+#include <asm/sclp.h>
+
+/*
+ * Create a Kernel NSS if the SAVESYS= parameter is defined
+ */
+#define DEFSYS_CMD_SIZE		96
+#define SAVESYS_CMD_SIZE	32
+
+char kernel_nss_name[NSS_NAME_SIZE + 1];
+
+#ifdef CONFIG_SHARED_KERNEL
+static noinline __init void create_kernel_nss(void)
+{
+	unsigned int i, stext_pfn, eshared_pfn, end_pfn, min_size;
+#ifdef CONFIG_BLK_DEV_INITRD
+	unsigned int sinitrd_pfn, einitrd_pfn;
+#endif
+	int response;
+	char *savesys_ptr;
+	char upper_command_line[COMMAND_LINE_SIZE];
+	char defsys_cmd[DEFSYS_CMD_SIZE];
+	char savesys_cmd[SAVESYS_CMD_SIZE];
+
+	/* Do nothing if we are not running under VM */
+	if (!MACHINE_IS_VM)
+		return;
+
+	/* Convert COMMAND_LINE to upper case */
+	for (i = 0; i < strlen(COMMAND_LINE); i++)
+		upper_command_line[i] = toupper(COMMAND_LINE[i]);
+
+	savesys_ptr = strstr(upper_command_line, "SAVESYS=");
+
+	if (!savesys_ptr)
+		return;
+
+	savesys_ptr += 8;    /* Point to the beginning of the NSS name */
+	for (i = 0; i < NSS_NAME_SIZE; i++) {
+		if (savesys_ptr[i] == ' ' || savesys_ptr[i] == '\0')
+			break;
+		kernel_nss_name[i] = savesys_ptr[i];
+	}
+
+	stext_pfn = PFN_DOWN(__pa(&_stext));
+	eshared_pfn = PFN_DOWN(__pa(&_eshared));
+	end_pfn = PFN_UP(__pa(&_end));
+	min_size = end_pfn << 2;
+
+	sprintf(defsys_cmd, "DEFSYS %s 00000-%.5X EW %.5X-%.5X SR %.5X-%.5X",
+		kernel_nss_name, stext_pfn - 1, stext_pfn, eshared_pfn - 1,
+		eshared_pfn, end_pfn);
+
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (INITRD_START && INITRD_SIZE) {
+		sinitrd_pfn = PFN_DOWN(__pa(INITRD_START));
+		einitrd_pfn = PFN_UP(__pa(INITRD_START + INITRD_SIZE));
+		min_size = einitrd_pfn << 2;
+		sprintf(defsys_cmd, "%s EW %.5X-%.5X", defsys_cmd,
+		sinitrd_pfn, einitrd_pfn);
+	}
+#endif
+
+	sprintf(defsys_cmd, "%s EW MINSIZE=%.7iK", defsys_cmd, min_size);
+	sprintf(savesys_cmd, "SAVESYS %s \n IPL %s",
+		kernel_nss_name, kernel_nss_name);
+
+	__cpcmd(defsys_cmd, NULL, 0, &response);
+
+	if (response != 0)
+		return;
+
+	__cpcmd(savesys_cmd, NULL, 0, &response);
+
+	if (response != strlen(savesys_cmd))
+		return;
+
+	ipl_flags = IPL_NSS_VALID;
+}
+
+#else /* CONFIG_SHARED_KERNEL */
+
+static inline void create_kernel_nss(void) { }
+
+#endif /* CONFIG_SHARED_KERNEL */
+
+/*
+ * Clear bss memory
+ */
+static noinline __init void clear_bss_section(void)
+{
+	memset(__bss_start, 0, _end - __bss_start);
+}
+
+/*
+ * Initialize storage key for kernel pages
+ */
+static noinline __init void init_kernel_storage_key(void)
+{
+	unsigned long end_pfn, init_pfn;
+
+	end_pfn = PFN_UP(__pa(&_end));
+
+	for (init_pfn = 0 ; init_pfn < end_pfn; init_pfn++)
+		page_set_storage_key(init_pfn << PAGE_SHIFT, PAGE_DEFAULT_KEY);
+}
+
+static noinline __init void detect_machine_type(void)
+{
+	struct cpuinfo_S390 *cpuinfo = &S390_lowcore.cpu_data;
+
+	asm volatile("stidp %0" : "=m" (S390_lowcore.cpu_data.cpu_id));
+
+	/* Running under z/VM ? */
+	if (cpuinfo->cpu_id.version == 0xff)
+		machine_flags |= 1;
+
+	/* Running on a P/390 ? */
+	if (cpuinfo->cpu_id.machine == 0x7490)
+		machine_flags |= 4;
+}
+
+static noinline __init int memory_fast_detect(void)
+{
+
+	unsigned long val0 = 0;
+	unsigned long val1 = 0xc;
+	int ret = -ENOSYS;
+
+	if (ipl_flags & IPL_NSS_VALID)
+		return -ENOSYS;
+
+	asm volatile(
+		"	diag	%1,%2,0x260\n"
+		"0:	lhi	%0,0\n"
+		"1:\n"
+		EX_TABLE(0b,1b)
+		: "+d" (ret), "+d" (val0), "+d" (val1) : : "cc");
+
+	if (ret || val0 != val1)
+		return -ENOSYS;
+
+	memory_chunk[0].size = val0;
+	return 0;
+}
+
+#define ADDR2G	(1UL << 31)
+
+static noinline __init unsigned long sclp_memory_detect(void)
+{
+	struct sclp_readinfo_sccb *sccb;
+	unsigned long long memsize;
+
+	sccb = &s390_readinfo_sccb;
+
+	if (sccb->header.response_code != 0x10)
+		return 0;
+
+	if (sccb->rnsize)
+		memsize = sccb->rnsize << 20;
+	else
+		memsize = sccb->rnsize2 << 20;
+	if (sccb->rnmax)
+		memsize *= sccb->rnmax;
+	else
+		memsize *= sccb->rnmax2;
+#ifndef CONFIG_64BIT
+	/*
+	 * Can't deal with more than 2G in 31 bit addressing mode, so
+	 * limit the value in order to avoid strange side effects.
+	 */
+	if (memsize > ADDR2G)
+		memsize = ADDR2G;
+#endif
+	return (unsigned long) memsize;
+}
+
+static inline __init unsigned long __tprot(unsigned long addr)
+{
+	int cc = -1;
+
+	asm volatile(
+		"	tprot	0(%1),0\n"
+		"0:	ipm	%0\n"
+		"	srl	%0,28\n"
+		"1:\n"
+		EX_TABLE(0b,1b)
+		: "+d" (cc) : "a" (addr) : "cc");
+	return (unsigned long)cc;
+}
+
+/* Checking memory in 128KB increments. */
+#define CHUNK_INCR	(1UL << 17)
+
+static noinline __init void find_memory_chunks(unsigned long memsize)
+{
+	unsigned long addr = 0, old_addr = 0;
+	unsigned long old_cc = CHUNK_READ_WRITE;
+	unsigned long cc;
+	int chunk = 0;
+
+	while (chunk < MEMORY_CHUNKS) {
+		cc = __tprot(addr);
+		while (cc == old_cc) {
+			addr += CHUNK_INCR;
+			cc = __tprot(addr);
+#ifndef CONFIG_64BIT
+			if (addr == ADDR2G)
+				break;
+#endif
+		}
+
+		if (old_addr != addr &&
+		    (old_cc == CHUNK_READ_WRITE || old_cc == CHUNK_READ_ONLY)) {
+			memory_chunk[chunk].addr = old_addr;
+			memory_chunk[chunk].size = addr - old_addr;
+			memory_chunk[chunk].type = old_cc;
+			chunk++;
+		}
+
+		old_addr = addr;
+		old_cc = cc;
+
+#ifndef CONFIG_64BIT
+		if (addr == ADDR2G)
+			break;
+#endif
+		/*
+		 * Finish memory detection at the first hole, unless
+		 * - we reached the hsa -> skip it.
+		 * - we know there must be more.
+		 */
+		if (cc == -1UL && !memsize && old_addr != ADDR2G)
+			break;
+		if (memsize && addr >= memsize)
+			break;
+	}
+}
+
+static __init void early_pgm_check_handler(void)
+{
+	unsigned long addr;
+	const struct exception_table_entry *fixup;
+
+	addr = S390_lowcore.program_old_psw.addr;
+	fixup = search_exception_tables(addr & PSW_ADDR_INSN);
+	if (!fixup)
+		disabled_wait(0);
+	S390_lowcore.program_old_psw.addr = fixup->fixup | PSW_ADDR_AMODE;
+}
+
+static noinline __init void setup_lowcore_early(void)
+{
+	psw_t psw;
+
+	psw.mask = PSW_BASE_BITS | PSW_DEFAULT_KEY;
+	psw.addr = PSW_ADDR_AMODE | (unsigned long) s390_base_ext_handler;
+	S390_lowcore.external_new_psw = psw;
+	psw.addr = PSW_ADDR_AMODE | (unsigned long) s390_base_pgm_handler;
+	S390_lowcore.program_new_psw = psw;
+	s390_base_pgm_handler_fn = early_pgm_check_handler;
+}
+
+/*
+ * Save ipl parameters, clear bss memory, initialize storage keys
+ * and create a kernel NSS at startup if the SAVESYS= parm is defined
+ */
+void __init startup_init(void)
+{
+	unsigned long memsize;
+
+	ipl_save_parameters();
+	clear_bss_section();
+	init_kernel_storage_key();
+	lockdep_init();
+	lockdep_off();
+	detect_machine_type();
+	create_kernel_nss();
+	sort_main_extable();
+	setup_lowcore_early();
+	sclp_readinfo_early();
+	memsize = sclp_memory_detect();
+	if (memory_fast_detect() < 0)
+		find_memory_chunks(memsize);
+	lockdep_on();
+}
diff --git a/arch/s390/kernel/ebcdic.c b/arch/s390/kernel/ebcdic.c
index bb0f973137f0..cc0dc609d738 100644
--- a/arch/s390/kernel/ebcdic.c
+++ b/arch/s390/kernel/ebcdic.c
@@ -11,6 +11,7 @@
 
 #include <linux/module.h>
 #include <asm/types.h>
+#include <asm/ebcdic.h>
 
 /*
  * ASCII (IBM PC 437)  -> EBCDIC 037
diff --git a/arch/s390/kernel/head31.S b/arch/s390/kernel/head31.S
index eca507050e47..453fd3b4edea 100644
--- a/arch/s390/kernel/head31.S
+++ b/arch/s390/kernel/head31.S
@@ -51,176 +51,15 @@ startup_continue:
 	st	%r15,__LC_KERNEL_STACK	# set end of kernel stack
 	ahi	%r15,-96
 	xc	__SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15) # clear backchain
-
-	l	%r14,.Lipl_save_parameters-.LPG1(%r13)
-	basr	%r14,%r14
 #
-# clear bss memory
+# Save ipl parameters, clear bss memory, initialize storage key for kernel pages,
+# and create a kernel NSS if the SAVESYS= parm is defined
 #
-	l	%r2,.Lbss_bgn-.LPG1(%r13) # start of bss
-	l	%r3,.Lbss_end-.LPG1(%r13) # end of bss
-	sr	%r3,%r2			# length of bss
-	sr	%r4,%r4
-	sr	%r5,%r5			# set src,length and pad to zero
-	sr	%r0,%r0
-	mvcle	%r2,%r4,0		# clear mem
-	jo	.-4			# branch back, if not finish
-
-	l	%r2,.Lrcp-.LPG1(%r13)	# Read SCP forced command word
-.Lservicecall:
-	stosm	.Lpmask-.LPG1(%r13),0x01	# authorize ext interrupts
-
-	stctl	%r0, %r0,.Lcr-.LPG1(%r13)	# get cr0
-	la	%r1,0x200		# set bit 22
-	o	%r1,.Lcr-.LPG1(%r13)	# or old cr0 with r1
-	st	%r1,.Lcr-.LPG1(%r13)
-	lctl	%r0, %r0,.Lcr-.LPG1(%r13)	# load modified cr0
-
-	mvc	__LC_EXT_NEW_PSW(8),.Lpcext-.LPG1(%r13) # set postcall psw
-	la	%r1, .Lsclph-.LPG1(%r13)
-	a	%r1,__LC_EXT_NEW_PSW+4	# set handler
-	st	%r1,__LC_EXT_NEW_PSW+4
-
-	l	%r4,.Lsccbaddr-.LPG1(%r13) # %r4 is our index for sccb stuff
-	lr	%r1,%r4			# our sccb
-	.insn	rre,0xb2200000,%r2,%r1	# service call
-	ipm	%r1
-	srl	%r1,28			# get cc code
-	xr	%r3, %r3
-	chi	%r1,3
-	be	.Lfchunk-.LPG1(%r13)	# leave
-	chi	%r1,2
-	be	.Lservicecall-.LPG1(%r13)
-	lpsw	.Lwaitsclp-.LPG1(%r13)
-.Lsclph:
-	lh	%r1,.Lsccbr-.Lsccb(%r4)
-	chi	%r1,0x10		# 0x0010 is the sucess code
-	je	.Lprocsccb		# let's process the sccb
-	chi	%r1,0x1f0
-	bne	.Lfchunk-.LPG1(%r13)	# unhandled error code
-	c	%r2, .Lrcp-.LPG1(%r13)	# Did we try Read SCP forced
-	bne	.Lfchunk-.LPG1(%r13)	# if no, give up
-	l	%r2, .Lrcp2-.LPG1(%r13)	# try with Read SCP
-	b	.Lservicecall-.LPG1(%r13)
-.Lprocsccb:
-	lhi	%r1,0
-	icm	%r1,3,.Lscpincr1-.Lsccb(%r4) # use this one if != 0
-	jnz	.Lscnd
-	lhi	%r1,0x800		# otherwise report 2GB
-.Lscnd:
-	lhi	%r3,0x800		# limit reported memory size to 2GB
-	cr	%r1,%r3
-	jl	.Lno2gb
-	lr	%r1,%r3
-.Lno2gb:
-	xr	%r3,%r3			# same logic
-	ic	%r3,.Lscpa1-.Lsccb(%r4)
-	chi	%r3,0x00
-	jne	.Lcompmem
-	l	%r3,.Lscpa2-.Lsccb(%r4)
-.Lcompmem:
-	mr	%r2,%r1			# mem in MB on 128-bit
-	l	%r1,.Lonemb-.LPG1(%r13)
-	mr	%r2,%r1			# mem size in bytes in %r3
-	b	.Lfchunk-.LPG1(%r13)
-
-	.align 4
-.Lipl_save_parameters:
-	.long	ipl_save_parameters
-.Linittu:
-	.long	init_thread_union
-.Lpmask:
-	.byte	0
-	.align	8
-.Lpcext:.long	0x00080000,0x80000000
-.Lcr:
-	.long	0x00			# place holder for cr0
-	.align	8
-.Lwaitsclp:
-	.long 0x010a0000,0x80000000 + .Lsclph
-.Lrcp:
-	.int	0x00120001		# Read SCP forced code
-.Lrcp2:
-	.int	0x00020001		# Read SCP code
-.Lonemb:
-	.int	0x100000
-.Lfchunk:
+	l	%r14,.Lstartup_init-.LPG1(%r13)
+	basr	%r14,%r14
 
-#
-# find memory chunks.
-#
-	lr	%r9,%r3			# end of mem
-	mvc	__LC_PGM_NEW_PSW(8),.Lpcmem-.LPG1(%r13)
-	la	%r1,1			# test in increments of 128KB
-	sll	%r1,17
-	l	%r3,.Lmchunk-.LPG1(%r13) # get pointer to memory_chunk array
-	slr	%r4,%r4			# set start of chunk to zero
-	slr	%r5,%r5			# set end of chunk to zero
-	slr	%r6,%r6			# set access code to zero
-	la	%r10,MEMORY_CHUNKS	# number of chunks
-.Lloop:
-	tprot	0(%r5),0		# test protection of first byte
-	ipm	%r7
-	srl	%r7,28
-	clr	%r6,%r7			# compare cc with last access code
-	be	.Lsame-.LPG1(%r13)
-	lhi	%r8,0			# no program checks
-	b	.Lsavchk-.LPG1(%r13)
-.Lsame:
-	ar	%r5,%r1			# add 128KB to end of chunk
-	bno	.Lloop-.LPG1(%r13)	# r1 < 0x80000000 -> loop
-.Lchkmem:				# > 2GB or tprot got a program check
-	lhi	%r8,1			# set program check flag
-.Lsavchk:
-	clr	%r4,%r5			# chunk size > 0?
-	be	.Lchkloop-.LPG1(%r13)
-	st	%r4,0(%r3)		# store start address of chunk
-	lr	%r0,%r5
-	slr	%r0,%r4
-	st	%r0,4(%r3)		# store size of chunk
-	st	%r6,8(%r3)		# store type of chunk
-	la	%r3,12(%r3)
-	ahi	%r10,-1			# update chunk number
-.Lchkloop:
-	lr	%r6,%r7			# set access code to last cc
-	# we got an exception or we're starting a new
-	# chunk , we must check if we should
-	# still try to find valid memory (if we detected
-	# the amount of available storage), and if we
-	# have chunks left
-	xr	%r0,%r0
-	clr	%r0,%r9			# did we detect memory?
-	je	.Ldonemem		# if not, leave
-	chi	%r10,0			# do we have chunks left?
-	je	.Ldonemem
-	chi	%r8,1			# program check ?
-	je	.Lpgmchk
-	lr	%r4,%r5			# potential new chunk
-	alr	%r5,%r1			# add 128KB to end of chunk
-	j	.Llpcnt
-.Lpgmchk:
-	alr	%r5,%r1			# add 128KB to end of chunk
-	lr	%r4,%r5			# potential new chunk
-.Llpcnt:
-	clr	%r5,%r9			# should we go on?
-	jl	.Lloop
-.Ldonemem:
 	l	%r12,.Lmflags-.LPG1(%r13) # get address of machine_flags
 #
-# find out if we are running under VM
-#
-	stidp	__LC_CPUID		# store cpuid
-	tm	__LC_CPUID,0xff		# running under VM ?
-	bno	.Lnovm-.LPG1(%r13)
-	oi	3(%r12),1		# set VM flag
-.Lnovm:
-	lh	%r0,__LC_CPUID+4	# get cpu version
-	chi	%r0,0x7490		# running on a P/390 ?
-	bne	.Lnop390-.LPG1(%r13)
-	oi	3(%r12),4		# set P/390 flag
-.Lnop390:
-
-#
 # find out if we have an IEEE fpu
 #
 	mvc	__LC_PGM_NEW_PSW(8),.Lpcfpu-.LPG1(%r13)
@@ -295,7 +134,6 @@ startup_continue:
 	.long	0			# cr15: linkage stack operations
 .Lduct:	.long	0,0,0,0,0,0,0,0
 	.long	0,0,0,0,0,0,0,0
-.Lpcmem:.long	0x00080000,0x80000000 + .Lchkmem
 .Lpcfpu:.long	0x00080000,0x80000000 + .Lchkfpu
 .Lpccsp:.long	0x00080000,0x80000000 + .Lchkcsp
 .Lpcmvpg:.long	0x00080000,0x80000000 + .Lchkmvpg
@@ -306,7 +144,9 @@ startup_continue:
 .Lbss_bgn:  .long __bss_start
 .Lbss_end:  .long _end
 .Lparmaddr: .long PARMAREA
-.Lsccbaddr: .long .Lsccb
+.Linittu:   .long init_thread_union
+.Lstartup_init:
+	    .long startup_init
 
 	.globl ipl_schib
 ipl_schib:
@@ -322,26 +162,6 @@ ipl_devno:
 	.word 0
 
 	.org	0x12000
-.globl s390_readinfo_sccb
-s390_readinfo_sccb:
-.Lsccb:
-	.hword	0x1000			# length, one page
-	.byte	0x00,0x00,0x00
-	.byte	0x80			# variable response bit set
-.Lsccbr:
-	.hword	0x00			# response code
-.Lscpincr1:
-	.hword	0x00
-.Lscpa1:
-	.byte	0x00
-	.fill	89,1,0
-.Lscpa2:
-	.int	0x00
-.Lscpincr2:
-	.quad	0x00
-	.fill	3984,1,0
-	.org	0x13000
-
 #ifdef CONFIG_SHARED_KERNEL
 	.org	0x100000
 #endif
diff --git a/arch/s390/kernel/head64.S b/arch/s390/kernel/head64.S
index 6ba3f4512dd1..b8fec4e5c5d4 100644
--- a/arch/s390/kernel/head64.S
+++ b/arch/s390/kernel/head64.S
@@ -58,183 +58,15 @@ startup_continue:
 	stg	%r15,__LC_KERNEL_STACK	# set end of kernel stack
 	aghi	%r15,-160
 	xc	__SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15) # clear backchain
-
-	brasl	%r14,ipl_save_parameters
 #
-# clear bss memory
+# Save ipl parameters, clear bss memory, initialize storage key for kernel pages,
+# and create a kernel NSS if the SAVESYS= parm is defined
 #
-	larl	%r2,__bss_start 	# start of bss segment
-	larl	%r3,_end		# end of bss segment
-	sgr	%r3,%r2 		# length of bss
-	sgr	%r4,%r4 		#
-	sgr	%r5,%r5 		# set src,length and pad to zero
-	mvcle	%r2,%r4,0		# clear mem
-	jo	.-4			# branch back, if not finish
+	brasl	%r14,startup_init
 					# set program check new psw mask
 	mvc	__LC_PGM_NEW_PSW(8),.Lpcmsk-.LPG1(%r13)
-	larl	%r1,.Lslowmemdetect	# set program check address
-	stg	%r1,__LC_PGM_NEW_PSW+8
-	lghi	%r1,0xc
-	diag	%r0,%r1,0x260		# get memory size of virtual machine
-	cgr	%r0,%r1			# different? -> old detection routine
-	jne	.Lslowmemdetect
-	aghi	%r1,1			# size is one more than end
-	larl	%r2,memory_chunk
-	stg	%r1,8(%r2)		# store size of chunk
-	j	.Ldonemem
-
-.Lslowmemdetect:
-	l	%r2,.Lrcp-.LPG1(%r13)	# Read SCP forced command word
-.Lservicecall:
-	stosm	.Lpmask-.LPG1(%r13),0x01	# authorize ext interrupts
-
-	stctg	%r0,%r0,.Lcr-.LPG1(%r13)	# get cr0
-	la	%r1,0x200		# set bit 22
-	og	%r1,.Lcr-.LPG1(%r13)	# or old cr0 with r1
-	stg	%r1,.Lcr-.LPG1(%r13)
-	lctlg	%r0,%r0,.Lcr-.LPG1(%r13)	# load modified cr0
-
-	mvc	__LC_EXT_NEW_PSW(8),.Lpcmsk-.LPG1(%r13) # set postcall psw
-	larl	%r1,.Lsclph
-	stg	%r1,__LC_EXT_NEW_PSW+8	# set handler
-
-	larl	%r4,.Lsccb		# %r4 is our index for sccb stuff
-	lgr	%r1,%r4			# our sccb
-	.insn	rre,0xb2200000,%r2,%r1	# service call
-	ipm	%r1
-	srl	%r1,28			# get cc code
-	xr	%r3,%r3
-	chi	%r1,3
-	be	.Lfchunk-.LPG1(%r13)	# leave
-	chi	%r1,2
-	be	.Lservicecall-.LPG1(%r13)
-	lpswe	.Lwaitsclp-.LPG1(%r13)
-.Lsclph:
-	lh	%r1,.Lsccbr-.Lsccb(%r4)
-	chi	%r1,0x10		# 0x0010 is the sucess code
-	je	.Lprocsccb		# let's process the sccb
-	chi	%r1,0x1f0
-	bne	.Lfchunk-.LPG1(%r13)	# unhandled error code
-	c	%r2,.Lrcp-.LPG1(%r13)	# Did we try Read SCP forced
-	bne	.Lfchunk-.LPG1(%r13)	# if no, give up
-	l	%r2,.Lrcp2-.LPG1(%r13)	# try with Read SCP
-	b	.Lservicecall-.LPG1(%r13)
-.Lprocsccb:
-	lghi	%r1,0
-	icm	%r1,3,.Lscpincr1-.Lsccb(%r4)	# use this one if != 0
-	jnz	.Lscnd
-	lg	%r1,.Lscpincr2-.Lsccb(%r4)	# otherwise use this one
-.Lscnd:
-	xr	%r3,%r3			# same logic
-	ic	%r3,.Lscpa1-.Lsccb(%r4)
-	chi	%r3,0x00
-	jne	.Lcompmem
-	l	%r3,.Lscpa2-.Lsccb(%r4)
-.Lcompmem:
-	mlgr	%r2,%r1			# mem in MB on 128-bit
-	l	%r1,.Lonemb-.LPG1(%r13)
-	mlgr	%r2,%r1			# mem size in bytes in %r3
-	b	.Lfchunk-.LPG1(%r13)
-
-	.align	4
-.Lpmask:
-	.byte	0
-	.align	8
-.Lcr:
-	.quad	0x00  # place holder for cr0
-.Lwaitsclp:
-	.quad	0x0102000180000000,.Lsclph
-.Lrcp:
-	.int	0x00120001 # Read SCP forced code
-.Lrcp2:
-	.int	0x00020001 # Read SCP code
-.Lonemb:
-	.int	0x100000
-
-.Lfchunk:
-
-#
-# find memory chunks.
-#
-	lgr	%r9,%r3			# end of mem
-	larl	%r1,.Lchkmem		# set program check address
-	stg	%r1,__LC_PGM_NEW_PSW+8
-	la	%r1,1			# test in increments of 128KB
-	sllg	%r1,%r1,17
-	larl	%r3,memory_chunk
-	slgr	%r4,%r4 		# set start of chunk to zero
-	slgr	%r5,%r5 		# set end of chunk to zero
-	slr	%r6,%r6			# set access code to zero
-	la	%r10,MEMORY_CHUNKS	# number of chunks
-.Lloop:
-	tprot	0(%r5),0		# test protection of first byte
-	ipm	%r7
-	srl	%r7,28
-	clr	%r6,%r7			# compare cc with last access code
-	je	.Lsame
-	lghi	%r8,0			# no program checks
-	j	.Lsavchk
-.Lsame:
-	algr	%r5,%r1			# add 128KB to end of chunk
-					# no need to check here,
-	brc	12,.Lloop		# this is the same chunk
-.Lchkmem:				# > 16EB or tprot got a program check
-	lghi	%r8,1			# set program check flag
-.Lsavchk:
-	clgr	%r4,%r5			# chunk size > 0?
-	je	.Lchkloop
-	stg	%r4,0(%r3)		# store start address of chunk
-	lgr	%r0,%r5
-	slgr	%r0,%r4
-	stg	%r0,8(%r3)		# store size of chunk
-	st	%r6,20(%r3)		# store type of chunk
-	la	%r3,24(%r3)
-	ahi	%r10,-1			# update chunk number
-.Lchkloop:
-	lr	%r6,%r7			# set access code to last cc
-	# we got an exception or we're starting a new
-	# chunk , we must check if we should
-	# still try to find valid memory (if we detected
-	# the amount of available storage), and if we
-	# have chunks left
-	lghi	%r4,1
-	sllg	%r4,%r4,31
-	clgr	%r5,%r4
-	je	.Lhsaskip
-	xr	%r0, %r0
-	clgr	%r0, %r9		# did we detect memory?
-	je	.Ldonemem		# if not, leave
-	chi	%r10, 0			# do we have chunks left?
-	je	.Ldonemem
-.Lhsaskip:
-	chi	%r8,1			# program check ?
-	je	.Lpgmchk
-	lgr	%r4,%r5			# potential new chunk
-	algr	%r5,%r1			# add 128KB to end of chunk
-	j	.Llpcnt
-.Lpgmchk:
-	algr	%r5,%r1			# add 128KB to end of chunk
-	lgr	%r4,%r5			# potential new chunk
-.Llpcnt:
-	clgr	%r5,%r9			# should we go on?
-	jl	.Lloop
-.Ldonemem:
-
 	larl	%r12,machine_flags
 #
-# find out if we are running under VM
-#
-	stidp	__LC_CPUID		# store cpuid
-	tm	__LC_CPUID,0xff 	# running under VM ?
-	bno	0f-.LPG1(%r13)
-	oi	7(%r12),1		# set VM flag
-0:	lh	%r0,__LC_CPUID+4	# get cpu version
-	chi	%r0,0x7490		# running on a P/390 ?
-	bne	1f-.LPG1(%r13)
-	oi	7(%r12),4		# set P/390 flag
-1:
-
-#
 # find out if we have the MVPG instruction
 #
 	la	%r1,0f-.LPG1(%r13)	# set program check address
@@ -336,25 +168,6 @@ ipl_devno:
 	.word 0
 
 	.org	0x12000
-.globl s390_readinfo_sccb
-s390_readinfo_sccb:
-.Lsccb:
-	.hword	0x1000			# length, one page
-	.byte	0x00,0x00,0x00
-	.byte	0x80			# variable response bit set
-.Lsccbr:
-	.hword	0x00			# response code
-.Lscpincr1:
-	.hword	0x00
-.Lscpa1:
-	.byte	0x00
-	.fill	89,1,0
-.Lscpa2:
-	.int	0x00
-.Lscpincr2:
-	.quad	0x00
-	.fill	3984,1,0
-	.org	0x13000
 
 #ifdef CONFIG_SHARED_KERNEL
 	.org	0x100000
diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index 9e9972e8a52b..052259530651 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -20,26 +20,27 @@
 #include <asm/cio.h>
 #include <asm/ebcdic.h>
 #include <asm/reset.h>
+#include <asm/sclp.h>
 
 #define IPL_PARM_BLOCK_VERSION 0
-#define LOADPARM_LEN 8
 
-extern char s390_readinfo_sccb[];
-#define SCCB_VALID (*((__u16*)&s390_readinfo_sccb[6]) == 0x0010)
-#define SCCB_LOADPARM (&s390_readinfo_sccb[24])
-#define SCCB_FLAG (s390_readinfo_sccb[91])
+#define SCCB_VALID (s390_readinfo_sccb.header.response_code == 0x10)
+#define SCCB_LOADPARM (&s390_readinfo_sccb.loadparm)
+#define SCCB_FLAG (s390_readinfo_sccb.flags)
 
 enum ipl_type {
 	IPL_TYPE_NONE	 = 1,
 	IPL_TYPE_UNKNOWN = 2,
 	IPL_TYPE_CCW	 = 4,
 	IPL_TYPE_FCP	 = 8,
+	IPL_TYPE_NSS	 = 16,
 };
 
 #define IPL_NONE_STR	 "none"
 #define IPL_UNKNOWN_STR  "unknown"
 #define IPL_CCW_STR	 "ccw"
 #define IPL_FCP_STR	 "fcp"
+#define IPL_NSS_STR	 "nss"
 
 static char *ipl_type_str(enum ipl_type type)
 {
@@ -50,6 +51,8 @@ static char *ipl_type_str(enum ipl_type type)
 		return IPL_CCW_STR;
 	case IPL_TYPE_FCP:
 		return IPL_FCP_STR;
+	case IPL_TYPE_NSS:
+		return IPL_NSS_STR;
 	case IPL_TYPE_UNKNOWN:
 	default:
 		return IPL_UNKNOWN_STR;
@@ -64,6 +67,7 @@ enum ipl_method {
 	IPL_METHOD_FCP_RO_DIAG,
 	IPL_METHOD_FCP_RW_DIAG,
 	IPL_METHOD_FCP_RO_VM,
+	IPL_METHOD_NSS,
 };
 
 enum shutdown_action {
@@ -114,11 +118,14 @@ enum diag308_rc {
 static int diag308_set_works = 0;
 
 static int reipl_capabilities = IPL_TYPE_UNKNOWN;
+
 static enum ipl_type reipl_type = IPL_TYPE_UNKNOWN;
 static enum ipl_method reipl_method = IPL_METHOD_NONE;
 static struct ipl_parameter_block *reipl_block_fcp;
 static struct ipl_parameter_block *reipl_block_ccw;
 
+static char reipl_nss_name[NSS_NAME_SIZE + 1];
+
 static int dump_capabilities = IPL_TYPE_NONE;
 static enum ipl_type dump_type = IPL_TYPE_NONE;
 static enum ipl_method dump_method = IPL_METHOD_NONE;
@@ -173,6 +180,24 @@ static struct subsys_attribute sys_##_prefix##_##_name##_attr =		\
 			sys_##_prefix##_##_name##_show,			\
 			sys_##_prefix##_##_name##_store);
 
+#define DEFINE_IPL_ATTR_STR_RW(_prefix, _name, _fmt_out, _fmt_in, _value)\
+static ssize_t sys_##_prefix##_##_name##_show(struct subsystem *subsys,	\
+		char *page)						\
+{									\
+	return sprintf(page, _fmt_out, _value);				\
+}									\
+static ssize_t sys_##_prefix##_##_name##_store(struct subsystem *subsys,\
+		const char *buf, size_t len)				\
+{									\
+	if (sscanf(buf, _fmt_in, _value) != 1)				\
+		return -EINVAL;						\
+	return len;							\
+}									\
+static struct subsys_attribute sys_##_prefix##_##_name##_attr =		\
+	__ATTR(_name,(S_IRUGO | S_IWUSR),				\
+			sys_##_prefix##_##_name##_show,			\
+			sys_##_prefix##_##_name##_store);
+
 static void make_attrs_ro(struct attribute **attrs)
 {
 	while (*attrs) {
@@ -189,6 +214,8 @@ static enum ipl_type ipl_get_type(void)
 {
 	struct ipl_parameter_block *ipl = IPL_PARMBLOCK_START;
 
+	if (ipl_flags & IPL_NSS_VALID)
+		return IPL_TYPE_NSS;
 	if (!(ipl_flags & IPL_DEVNO_VALID))
 		return IPL_TYPE_UNKNOWN;
 	if (!(ipl_flags & IPL_PARMBLOCK_VALID))
@@ -324,6 +351,20 @@ static struct attribute_group ipl_ccw_attr_group = {
 	.attrs = ipl_ccw_attrs,
 };
 
+/* NSS ipl device attributes */
+
+DEFINE_IPL_ATTR_RO(ipl_nss, name, "%s\n", kernel_nss_name);
+
+static struct attribute *ipl_nss_attrs[] = {
+	&sys_ipl_type_attr.attr,
+	&sys_ipl_nss_name_attr.attr,
+	NULL,
+};
+
+static struct attribute_group ipl_nss_attr_group = {
+	.attrs = ipl_nss_attrs,
+};
+
 /* UNKNOWN ipl device attributes */
 
 static struct attribute *ipl_unknown_attrs[] = {
@@ -432,6 +473,21 @@ static struct attribute_group reipl_ccw_attr_group = {
 	.attrs = reipl_ccw_attrs,
 };
 
+
+/* NSS reipl device attributes */
+
+DEFINE_IPL_ATTR_STR_RW(reipl_nss, name, "%s\n", "%s\n", reipl_nss_name);
+
+static struct attribute *reipl_nss_attrs[] = {
+	&sys_reipl_nss_name_attr.attr,
+	NULL,
+};
+
+static struct attribute_group reipl_nss_attr_group = {
+	.name  = IPL_NSS_STR,
+	.attrs = reipl_nss_attrs,
+};
+
 /* reipl type */
 
 static int reipl_set_type(enum ipl_type type)
@@ -454,6 +510,9 @@ static int reipl_set_type(enum ipl_type type)
 		else
 			reipl_method = IPL_METHOD_FCP_RO_DIAG;
 		break;
+	case IPL_TYPE_NSS:
+		reipl_method = IPL_METHOD_NSS;
+		break;
 	default:
 		reipl_method = IPL_METHOD_NONE;
 	}
@@ -475,6 +534,8 @@ static ssize_t reipl_type_store(struct subsystem *subsys, const char *buf,
 		rc = reipl_set_type(IPL_TYPE_CCW);
 	else if (strncmp(buf, IPL_FCP_STR, strlen(IPL_FCP_STR)) == 0)
 		rc = reipl_set_type(IPL_TYPE_FCP);
+	else if (strncmp(buf, IPL_NSS_STR, strlen(IPL_NSS_STR)) == 0)
+		rc = reipl_set_type(IPL_TYPE_NSS);
 	return (rc != 0) ? rc : len;
 }
 
@@ -647,6 +708,10 @@ void do_reipl(void)
 	case IPL_METHOD_FCP_RO_VM:
 		__cpcmd("IPL", NULL, 0, NULL);
 		break;
+	case IPL_METHOD_NSS:
+		sprintf(buf, "IPL %s", reipl_nss_name);
+		__cpcmd(buf, NULL, 0, NULL);
+		break;
 	case IPL_METHOD_NONE:
 	default:
 		if (MACHINE_IS_VM)
@@ -733,6 +798,10 @@ static int __init ipl_init(void)
 	case IPL_TYPE_FCP:
 		rc = ipl_register_fcp_files();
 		break;
+	case IPL_TYPE_NSS:
+		rc = sysfs_create_group(&ipl_subsys.kset.kobj,
+					&ipl_nss_attr_group);
+		break;
 	default:
 		rc = sysfs_create_group(&ipl_subsys.kset.kobj,
 					&ipl_unknown_attr_group);
@@ -755,6 +824,20 @@ static void __init reipl_probe(void)
 	free_page((unsigned long)buffer);
 }
 
+static int __init reipl_nss_init(void)
+{
+	int rc;
+
+	if (!MACHINE_IS_VM)
+		return 0;
+	rc = sysfs_create_group(&reipl_subsys.kset.kobj, &reipl_nss_attr_group);
+	if (rc)
+		return rc;
+	strncpy(reipl_nss_name, kernel_nss_name, NSS_NAME_SIZE + 1);
+	reipl_capabilities |= IPL_TYPE_NSS;
+	return 0;
+}
+
 static int __init reipl_ccw_init(void)
 {
 	int rc;
@@ -837,6 +920,9 @@ static int __init reipl_init(void)
 	rc = reipl_fcp_init();
 	if (rc)
 		return rc;
+	rc = reipl_nss_init();
+	if (rc)
+		return rc;
 	rc = reipl_set_type(ipl_get_type());
 	if (rc)
 		return rc;
@@ -993,8 +1079,6 @@ static void do_reset_calls(void)
 		reset->fn();
 }
 
-extern void reset_mcck_handler(void);
-extern void reset_pgm_handler(void);
 extern __u32 dump_prefix_page;
 
 void s390_reset_system(void)
@@ -1016,14 +1100,14 @@ void s390_reset_system(void)
 	__ctl_clear_bit(0,28);
 
 	/* Set new machine check handler */
-	S390_lowcore.mcck_new_psw.mask = PSW_KERNEL_BITS & ~PSW_MASK_MCHECK;
+	S390_lowcore.mcck_new_psw.mask = psw_kernel_bits & ~PSW_MASK_MCHECK;
 	S390_lowcore.mcck_new_psw.addr =
-		PSW_ADDR_AMODE | (unsigned long) &reset_mcck_handler;
+		PSW_ADDR_AMODE | (unsigned long) s390_base_mcck_handler;
 
 	/* Set new program check handler */
-	S390_lowcore.program_new_psw.mask = PSW_KERNEL_BITS & ~PSW_MASK_MCHECK;
+	S390_lowcore.program_new_psw.mask = psw_kernel_bits & ~PSW_MASK_MCHECK;
 	S390_lowcore.program_new_psw.addr =
-		PSW_ADDR_AMODE | (unsigned long) &reset_pgm_handler;
+		PSW_ADDR_AMODE | (unsigned long) s390_base_pgm_handler;
 
 	do_reset_calls();
 }
diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c
index 1eef50918615..8f0cbca31203 100644
--- a/arch/s390/kernel/irq.c
+++ b/arch/s390/kernel/irq.c
@@ -1,9 +1,9 @@
 /*
  *  arch/s390/kernel/irq.c
  *
- *  S390 version
- *    Copyright (C) 2004 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Copyright IBM Corp. 2004,2007
  *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
+ *		 Thomas Spatzier (tspat@de.ibm.com)
  *
  * This file contains interrupt related functions.
  */
@@ -14,6 +14,8 @@
 #include <linux/interrupt.h>
 #include <linux/seq_file.h>
 #include <linux/cpu.h>
+#include <linux/proc_fs.h>
+#include <linux/profile.h>
 
 /*
  * show_interrupts is needed by /proc/interrupts.
@@ -93,5 +95,12 @@ asmlinkage void do_softirq(void)
 
 	local_irq_restore(flags);
 }
-
 EXPORT_SYMBOL(do_softirq);
+
+void init_irq_proc(void)
+{
+	struct proc_dir_entry *root_irq_dir;
+
+	root_irq_dir = proc_mkdir("irq", NULL);
+	create_prof_cpu_mask(root_irq_dir);
+}
diff --git a/arch/s390/kernel/kprobes.c b/arch/s390/kernel/kprobes.c
index 576368c4f605..a466bab6677e 100644
--- a/arch/s390/kernel/kprobes.c
+++ b/arch/s390/kernel/kprobes.c
@@ -155,15 +155,34 @@ void __kprobes get_instruction_type(struct arch_specific_insn *ainsn)
 static int __kprobes swap_instruction(void *aref)
 {
 	struct ins_replace_args *args = aref;
+	u32 *addr;
+	u32 instr;
 	int err = -EFAULT;
 
+	/*
+	 * Text segment is read-only, hence we use stura to bypass dynamic
+	 * address translation to exchange the instruction. Since stura
+	 * always operates on four bytes, but we only want to exchange two
+	 * bytes do some calculations to get things right. In addition we
+	 * shall not cross any page boundaries (vmalloc area!) when writing
+	 * the new instruction.
+	 */
+	addr = (u32 *)ALIGN((unsigned long)args->ptr, 4);
+	if ((unsigned long)args->ptr & 2)
+		instr = ((*addr) & 0xffff0000) | args->new;
+	else
+		instr = ((*addr) & 0x0000ffff) | args->new << 16;
+
 	asm volatile(
-		"0: mvc  0(2,%2),0(%3)\n"
-		"1: la   %0,0\n"
+		"	lra	%1,0(%1)\n"
+		"0:	stura	%2,%1\n"
+		"1:	la	%0,0\n"
 		"2:\n"
 		EX_TABLE(0b,2b)
-		: "+d" (err), "=m" (*args->ptr)
-		: "a" (args->ptr), "a" (&args->new), "m" (args->new));
+		: "+d" (err)
+		: "a" (addr), "d" (instr)
+		: "memory", "cc");
+
 	return err;
 }
 
@@ -356,7 +375,7 @@ no_kprobe:
  *	- When the probed function returns, this probe
  *		causes the handlers to fire
  */
-void __kprobes kretprobe_trampoline_holder(void)
+void kretprobe_trampoline_holder(void)
 {
 	asm volatile(".global kretprobe_trampoline\n"
 		     "kretprobe_trampoline: bcr 0,0\n");
@@ -365,7 +384,8 @@ void __kprobes kretprobe_trampoline_holder(void)
 /*
  * Called when the probe at kretprobe trampoline is hit
  */
-int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
+static int __kprobes trampoline_probe_handler(struct kprobe *p,
+					      struct pt_regs *regs)
 {
 	struct kretprobe_instance *ri = NULL;
 	struct hlist_head *head, empty_rp;
diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c
index f6d9bcc0f75b..52f57af252b4 100644
--- a/arch/s390/kernel/machine_kexec.c
+++ b/arch/s390/kernel/machine_kexec.c
@@ -11,6 +11,7 @@
 #include <linux/mm.h>
 #include <linux/kexec.h>
 #include <linux/delay.h>
+#include <linux/reboot.h>
 #include <asm/cio.h>
 #include <asm/setup.h>
 #include <asm/pgtable.h>
diff --git a/arch/s390/kernel/module.c b/arch/s390/kernel/module.c
index d989ed45a7aa..39d1dd752529 100644
--- a/arch/s390/kernel/module.c
+++ b/arch/s390/kernel/module.c
@@ -30,6 +30,7 @@
 #include <linux/fs.h>
 #include <linux/string.h>
 #include <linux/kernel.h>
+#include <linux/moduleloader.h>
 
 #if 0
 #define DEBUGP printk
@@ -58,7 +59,7 @@ void module_free(struct module *mod, void *module_region)
            table entries. */
 }
 
-static inline void
+static void
 check_rela(Elf_Rela *rela, struct module *me)
 {
 	struct mod_arch_syminfo *info;
@@ -181,7 +182,7 @@ apply_relocate(Elf_Shdr *sechdrs, const char *strtab, unsigned int symindex,
 	return -ENOEXEC;
 }
 
-static inline int
+static int
 apply_rela(Elf_Rela *rela, Elf_Addr base, Elf_Sym *symtab, 
 	   struct module *me)
 {
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index 6603fbb41d07..5acfac654f9d 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -144,7 +144,7 @@ static void default_idle(void)
 
 	trace_hardirqs_on();
 	/* Wait for external, I/O or machine check interrupt. */
-	__load_psw_mask(PSW_KERNEL_BITS | PSW_MASK_WAIT |
+	__load_psw_mask(psw_kernel_bits | PSW_MASK_WAIT |
 			PSW_MASK_IO | PSW_MASK_EXT);
 }
 
@@ -190,7 +190,7 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
 	struct pt_regs regs;
 
 	memset(&regs, 0, sizeof(regs));
-	regs.psw.mask = PSW_KERNEL_BITS | PSW_MASK_IO | PSW_MASK_EXT;
+	regs.psw.mask = psw_kernel_bits | PSW_MASK_IO | PSW_MASK_EXT;
 	regs.psw.addr = (unsigned long) kernel_thread_starter | PSW_ADDR_AMODE;
 	regs.gprs[9] = (unsigned long) fn;
 	regs.gprs[10] = (unsigned long) arg;
diff --git a/arch/s390/kernel/profile.c b/arch/s390/kernel/profile.c
deleted file mode 100644
index b81aa1f569ca..000000000000
--- a/arch/s390/kernel/profile.c
+++ /dev/null
@@ -1,20 +0,0 @@
-/*
- * arch/s390/kernel/profile.c
- *
- * Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
- * Author(s): Thomas Spatzier (tspat@de.ibm.com)
- *
- */
-#include <linux/proc_fs.h>
-#include <linux/profile.h>
-
-static struct proc_dir_entry * root_irq_dir;
-
-void init_irq_proc(void)
-{
-	/* create /proc/irq */
-	root_irq_dir = proc_mkdir("irq", NULL);
-
-	/* create /proc/irq/prof_cpu_mask */
-	create_prof_cpu_mask(root_irq_dir);
-}
diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c
index 8f36504075ed..2a8f0872ea8b 100644
--- a/arch/s390/kernel/ptrace.c
+++ b/arch/s390/kernel/ptrace.c
@@ -86,15 +86,13 @@ FixPerRegisters(struct task_struct *task)
 		per_info->control_regs.bits.storage_alt_space_ctl = 0;
 }
 
-void
-set_single_step(struct task_struct *task)
+static void set_single_step(struct task_struct *task)
 {
 	task->thread.per_info.single_step = 1;
 	FixPerRegisters(task);
 }
 
-void
-clear_single_step(struct task_struct *task)
+static void clear_single_step(struct task_struct *task)
 {
 	task->thread.per_info.single_step = 0;
 	FixPerRegisters(task);
@@ -232,9 +230,9 @@ poke_user(struct task_struct *child, addr_t addr, addr_t data)
 		 */
 		if (addr == (addr_t) &dummy->regs.psw.mask &&
 #ifdef CONFIG_COMPAT
-		    data != PSW_MASK_MERGE(PSW_USER32_BITS, data) &&
+		    data != PSW_MASK_MERGE(psw_user32_bits, data) &&
 #endif
-		    data != PSW_MASK_MERGE(PSW_USER_BITS, data))
+		    data != PSW_MASK_MERGE(psw_user_bits, data))
 			/* Invalid psw mask. */
 			return -EINVAL;
 #ifndef CONFIG_64BIT
@@ -309,7 +307,7 @@ do_ptrace_normal(struct task_struct *child, long request, long addr, long data)
 		copied = access_process_vm(child, addr, &tmp, sizeof(tmp), 0);
 		if (copied != sizeof(tmp))
 			return -EIO;
-		return put_user(tmp, (unsigned long __user *) data);
+		return put_user(tmp, (unsigned long __force __user *) data);
 
 	case PTRACE_PEEKUSR:
 		/* read the word at location addr in the USER area. */
@@ -331,7 +329,7 @@ do_ptrace_normal(struct task_struct *child, long request, long addr, long data)
 
 	case PTRACE_PEEKUSR_AREA:
 	case PTRACE_POKEUSR_AREA:
-		if (copy_from_user(&parea, (void __user *) addr,
+		if (copy_from_user(&parea, (void __force __user *) addr,
 							sizeof(parea)))
 			return -EFAULT;
 		addr = parea.kernel_addr;
@@ -341,10 +339,11 @@ do_ptrace_normal(struct task_struct *child, long request, long addr, long data)
 			if (request == PTRACE_PEEKUSR_AREA)
 				ret = peek_user(child, addr, data);
 			else {
-				addr_t tmp;
-				if (get_user (tmp, (addr_t __user *) data))
+				addr_t utmp;
+				if (get_user(utmp,
+					     (addr_t __force __user *) data))
 					return -EFAULT;
-				ret = poke_user(child, addr, tmp);
+				ret = poke_user(child, addr, utmp);
 			}
 			if (ret)
 				return ret;
@@ -394,7 +393,7 @@ peek_user_emu31(struct task_struct *child, addr_t addr, addr_t data)
 		if (addr == (addr_t) &dummy32->regs.psw.mask) {
 			/* Fake a 31 bit psw mask. */
 			tmp = (__u32)(task_pt_regs(child)->psw.mask >> 32);
-			tmp = PSW32_MASK_MERGE(PSW32_USER_BITS, tmp);
+			tmp = PSW32_MASK_MERGE(psw32_user_bits, tmp);
 		} else if (addr == (addr_t) &dummy32->regs.psw.addr) {
 			/* Fake a 31 bit psw address. */
 			tmp = (__u32) task_pt_regs(child)->psw.addr |
@@ -469,11 +468,11 @@ poke_user_emu31(struct task_struct *child, addr_t addr, addr_t data)
 		 */
 		if (addr == (addr_t) &dummy32->regs.psw.mask) {
 			/* Build a 64 bit psw mask from 31 bit mask. */
-			if (tmp != PSW32_MASK_MERGE(PSW32_USER_BITS, tmp))
+			if (tmp != PSW32_MASK_MERGE(psw32_user_bits, tmp))
 				/* Invalid psw mask. */
 				return -EINVAL;
 			task_pt_regs(child)->psw.mask =
-				PSW_MASK_MERGE(PSW_USER32_BITS, (__u64) tmp << 32);
+				PSW_MASK_MERGE(psw_user32_bits, (__u64) tmp << 32);
 		} else if (addr == (addr_t) &dummy32->regs.psw.addr) {
 			/* Build a 64 bit psw address from 31 bit address. */
 			task_pt_regs(child)->psw.addr =
@@ -550,7 +549,7 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 		copied = access_process_vm(child, addr, &tmp, sizeof(tmp), 0);
 		if (copied != sizeof(tmp))
 			return -EIO;
-		return put_user(tmp, (unsigned int __user *) data);
+		return put_user(tmp, (unsigned int __force __user *) data);
 
 	case PTRACE_PEEKUSR:
 		/* read the word at location addr in the USER area. */
@@ -571,7 +570,7 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 
 	case PTRACE_PEEKUSR_AREA:
 	case PTRACE_POKEUSR_AREA:
-		if (copy_from_user(&parea, (void __user *) addr,
+		if (copy_from_user(&parea, (void __force __user *) addr,
 							sizeof(parea)))
 			return -EFAULT;
 		addr = parea.kernel_addr;
@@ -581,10 +580,11 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 			if (request == PTRACE_PEEKUSR_AREA)
 				ret = peek_user_emu31(child, addr, data);
 			else {
-				__u32 tmp;
-				if (get_user (tmp, (__u32 __user *) data))
+				__u32 utmp;
+				if (get_user(utmp,
+					     (__u32 __force __user *) data))
 					return -EFAULT;
-				ret = poke_user_emu31(child, addr, tmp);
+				ret = poke_user_emu31(child, addr, utmp);
 			}
 			if (ret)
 				return ret;
@@ -595,17 +595,19 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 		return 0;
 	case PTRACE_GETEVENTMSG:
 		return put_user((__u32) child->ptrace_message,
-				(unsigned int __user *) data);
+				(unsigned int __force __user *) data);
 	case PTRACE_GETSIGINFO:
 		if (child->last_siginfo == NULL)
 			return -EINVAL;
-		return copy_siginfo_to_user32((compat_siginfo_t __user *) data,
+		return copy_siginfo_to_user32((compat_siginfo_t
+					       __force __user *) data,
 					      child->last_siginfo);
 	case PTRACE_SETSIGINFO:
 		if (child->last_siginfo == NULL)
 			return -EINVAL;
 		return copy_siginfo_from_user32(child->last_siginfo,
-						(compat_siginfo_t __user *) data);
+						(compat_siginfo_t
+						 __force __user *) data);
 	}
 	return ptrace_request(child, request, addr, data);
 }
diff --git a/arch/s390/kernel/reset.S b/arch/s390/kernel/reset.S
deleted file mode 100644
index 8a87355161fa..000000000000
--- a/arch/s390/kernel/reset.S
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- *  arch/s390/kernel/reset.S
- *
- *    Copyright (C) IBM Corp. 2006
- *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
- *		 Michael Holzheu <holzheu@de.ibm.com>
- */
-
-#include <asm/ptrace.h>
-#include <asm/lowcore.h>
-
-#ifdef CONFIG_64BIT
-
-	.globl	reset_mcck_handler
-reset_mcck_handler:
-	basr	%r13,0
-0:	lg	%r15,__LC_PANIC_STACK	# load panic stack
-	aghi	%r15,-STACK_FRAME_OVERHEAD
-	lg	%r1,s390_reset_mcck_handler-0b(%r13)
-	ltgr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-1:	la	%r1,4095
-	lmg	%r0,%r15,__LC_GPREGS_SAVE_AREA-4095(%r1)
-	lpswe	__LC_MCK_OLD_PSW
-
-	.globl	s390_reset_mcck_handler
-s390_reset_mcck_handler:
-	.quad	0
-
-	.globl	reset_pgm_handler
-reset_pgm_handler:
-	stmg	%r0,%r15,__LC_SAVE_AREA
-	basr	%r13,0
-0:	lg	%r15,__LC_PANIC_STACK	# load panic stack
-	aghi	%r15,-STACK_FRAME_OVERHEAD
-	lg	%r1,s390_reset_pgm_handler-0b(%r13)
-	ltgr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-	lmg	%r0,%r15,__LC_SAVE_AREA
-	lpswe	__LC_PGM_OLD_PSW
-1:	lpswe	disabled_wait_psw-0b(%r13)
-	.globl s390_reset_pgm_handler
-s390_reset_pgm_handler:
-	.quad	0
-	.align	8
-disabled_wait_psw:
-	.quad	0x0002000180000000,0x0000000000000000 + reset_pgm_handler
-
-#else /* CONFIG_64BIT */
-
-	.globl	reset_mcck_handler
-reset_mcck_handler:
-	basr	%r13,0
-0:	l	%r15,__LC_PANIC_STACK	# load panic stack
-	ahi	%r15,-STACK_FRAME_OVERHEAD
-	l	%r1,s390_reset_mcck_handler-0b(%r13)
-	ltr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-1:	lm	%r0,%r15,__LC_GPREGS_SAVE_AREA
-	lpsw	__LC_MCK_OLD_PSW
-
-	.globl	s390_reset_mcck_handler
-s390_reset_mcck_handler:
-	.long	0
-
-	.globl	reset_pgm_handler
-reset_pgm_handler:
-	stm	%r0,%r15,__LC_SAVE_AREA
-	basr	%r13,0
-0:	l	%r15,__LC_PANIC_STACK	# load panic stack
-	ahi	%r15,-STACK_FRAME_OVERHEAD
-	l	%r1,s390_reset_pgm_handler-0b(%r13)
-	ltr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-	lm	%r0,%r15,__LC_SAVE_AREA
-	lpsw	__LC_PGM_OLD_PSW
-
-1:	lpsw	disabled_wait_psw-0b(%r13)
-	.globl	s390_reset_pgm_handler
-s390_reset_pgm_handler:
-	.long	0
-disabled_wait_psw:
-	.align 8
-	.long	0x000a0000,0x00000000 + reset_pgm_handler
-
-#endif /* CONFIG_64BIT */
diff --git a/arch/s390/kernel/s390_ext.c b/arch/s390/kernel/s390_ext.c
index bc5beaa8f98e..acf93dba7727 100644
--- a/arch/s390/kernel/s390_ext.c
+++ b/arch/s390/kernel/s390_ext.c
@@ -125,14 +125,12 @@ void do_extint(struct pt_regs *regs, unsigned short code)
 		 * Make sure that the i/o interrupt did not "overtake"
 		 * the last HZ timer interrupt.
 		 */
-		account_ticks();
+		account_ticks(S390_lowcore.int_clock);
 	kstat_cpu(smp_processor_id()).irqs[EXTERNAL_INTERRUPT]++;
         index = ext_hash(code);
 	for (p = ext_int_hash[index]; p; p = p->next) {
-		if (likely(p->code == code)) {
-			if (likely(p->handler))
-				p->handler(code);
-		}
+		if (likely(p->code == code))
+			p->handler(code);
 	}
 	irq_exit();
 	set_irq_regs(old_regs);
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index 5d8ee3baac14..50c5210fbc64 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -38,6 +38,8 @@
 #include <linux/device.h>
 #include <linux/notifier.h>
 #include <linux/pfn.h>
+#include <linux/ctype.h>
+#include <linux/reboot.h>
 
 #include <asm/uaccess.h>
 #include <asm/system.h>
@@ -49,6 +51,14 @@
 #include <asm/page.h>
 #include <asm/ptrace.h>
 #include <asm/sections.h>
+#include <asm/ebcdic.h>
+#include <asm/compat.h>
+
+long psw_kernel_bits	= (PSW_BASE_BITS | PSW_MASK_DAT | PSW_ASC_PRIMARY |
+			   PSW_MASK_MCHECK | PSW_DEFAULT_KEY);
+long psw_user_bits	= (PSW_BASE_BITS | PSW_MASK_DAT | PSW_ASC_HOME |
+			   PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			   PSW_MASK_PSTATE | PSW_DEFAULT_KEY);
 
 /*
  * User copy operations.
@@ -117,9 +127,9 @@ void __devinit cpu_init (void)
  */
 char vmhalt_cmd[128] = "";
 char vmpoff_cmd[128] = "";
-char vmpanic_cmd[128] = "";
+static char vmpanic_cmd[128] = "";
 
-static inline void strncpy_skip_quote(char *dst, char *src, int n)
+static void strncpy_skip_quote(char *dst, char *src, int n)
 {
         int sx, dx;
 
@@ -275,10 +285,6 @@ static void __init conmode_default(void)
 }
 
 #ifdef CONFIG_SMP
-extern void machine_restart_smp(char *);
-extern void machine_halt_smp(void);
-extern void machine_power_off_smp(void);
-
 void (*_machine_restart)(char *command) = machine_restart_smp;
 void (*_machine_halt)(void) = machine_halt_smp;
 void (*_machine_power_off)(void) = machine_power_off_smp;
@@ -386,6 +392,84 @@ static int __init early_parse_ipldelay(char *p)
 }
 early_param("ipldelay", early_parse_ipldelay);
 
+#ifdef CONFIG_S390_SWITCH_AMODE
+unsigned int switch_amode = 0;
+EXPORT_SYMBOL_GPL(switch_amode);
+
+static void set_amode_and_uaccess(unsigned long user_amode,
+				  unsigned long user32_amode)
+{
+	psw_user_bits = PSW_BASE_BITS | PSW_MASK_DAT | user_amode |
+			PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			PSW_MASK_PSTATE | PSW_DEFAULT_KEY;
+#ifdef CONFIG_COMPAT
+	psw_user32_bits = PSW_BASE32_BITS | PSW_MASK_DAT | user_amode |
+			  PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			  PSW_MASK_PSTATE | PSW_DEFAULT_KEY;
+	psw32_user_bits = PSW32_BASE_BITS | PSW32_MASK_DAT | user32_amode |
+			  PSW32_MASK_IO | PSW32_MASK_EXT | PSW32_MASK_MCHECK |
+			  PSW32_MASK_PSTATE;
+#endif
+	psw_kernel_bits = PSW_BASE_BITS | PSW_MASK_DAT | PSW_ASC_HOME |
+			  PSW_MASK_MCHECK | PSW_DEFAULT_KEY;
+
+	if (MACHINE_HAS_MVCOS) {
+		printk("mvcos available.\n");
+		memcpy(&uaccess, &uaccess_mvcos_switch, sizeof(uaccess));
+	} else {
+		printk("mvcos not available.\n");
+		memcpy(&uaccess, &uaccess_pt, sizeof(uaccess));
+	}
+}
+
+/*
+ * Switch kernel/user addressing modes?
+ */
+static int __init early_parse_switch_amode(char *p)
+{
+	switch_amode = 1;
+	return 0;
+}
+early_param("switch_amode", early_parse_switch_amode);
+
+#else /* CONFIG_S390_SWITCH_AMODE */
+static inline void set_amode_and_uaccess(unsigned long user_amode,
+					 unsigned long user32_amode)
+{
+}
+#endif /* CONFIG_S390_SWITCH_AMODE */
+
+#ifdef CONFIG_S390_EXEC_PROTECT
+unsigned int s390_noexec = 0;
+EXPORT_SYMBOL_GPL(s390_noexec);
+
+/*
+ * Enable execute protection?
+ */
+static int __init early_parse_noexec(char *p)
+{
+	if (!strncmp(p, "off", 3))
+		return 0;
+	switch_amode = 1;
+	s390_noexec = 1;
+	return 0;
+}
+early_param("noexec", early_parse_noexec);
+#endif /* CONFIG_S390_EXEC_PROTECT */
+
+static void setup_addressing_mode(void)
+{
+	if (s390_noexec) {
+		printk("S390 execute protection active, ");
+		set_amode_and_uaccess(PSW_ASC_SECONDARY, PSW32_ASC_SECONDARY);
+		return;
+	}
+	if (switch_amode) {
+		printk("S390 address spaces switched, ");
+		set_amode_and_uaccess(PSW_ASC_PRIMARY, PSW32_ASC_PRIMARY);
+	}
+}
+
 static void __init
 setup_lowcore(void)
 {
@@ -402,19 +486,21 @@ setup_lowcore(void)
 	lc->restart_psw.mask = PSW_BASE_BITS | PSW_DEFAULT_KEY;
 	lc->restart_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long) restart_int_handler;
-	lc->external_new_psw.mask = PSW_KERNEL_BITS;
+	if (switch_amode)
+		lc->restart_psw.mask |= PSW_ASC_HOME;
+	lc->external_new_psw.mask = psw_kernel_bits;
 	lc->external_new_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long) ext_int_handler;
-	lc->svc_new_psw.mask = PSW_KERNEL_BITS | PSW_MASK_IO | PSW_MASK_EXT;
+	lc->svc_new_psw.mask = psw_kernel_bits | PSW_MASK_IO | PSW_MASK_EXT;
 	lc->svc_new_psw.addr = PSW_ADDR_AMODE | (unsigned long) system_call;
-	lc->program_new_psw.mask = PSW_KERNEL_BITS;
+	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 & ~PSW_MASK_MCHECK & ~PSW_MASK_DAT;
+		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;
+	lc->io_new_psw.mask = psw_kernel_bits;
 	lc->io_new_psw.addr = PSW_ADDR_AMODE | (unsigned long) io_int_handler;
 	lc->ipl_device = S390_lowcore.ipl_device;
 	lc->jiffy_timer = -1LL;
@@ -439,7 +525,7 @@ setup_lowcore(void)
 static void __init
 setup_resources(void)
 {
-	struct resource *res;
+	struct resource *res, *sub_res;
 	int i;
 
 	code_resource.start = (unsigned long) &_text;
@@ -464,8 +550,38 @@ setup_resources(void)
 		res->start = memory_chunk[i].addr;
 		res->end = memory_chunk[i].addr +  memory_chunk[i].size - 1;
 		request_resource(&iomem_resource, res);
-		request_resource(res, &code_resource);
-		request_resource(res, &data_resource);
+
+		if (code_resource.start >= res->start  &&
+			code_resource.start <= res->end &&
+			code_resource.end > res->end) {
+			sub_res = alloc_bootmem_low(sizeof(struct resource));
+			memcpy(sub_res, &code_resource,
+				sizeof(struct resource));
+			sub_res->end = res->end;
+			code_resource.start = res->end + 1;
+			request_resource(res, sub_res);
+		}
+
+		if (code_resource.start >= res->start &&
+			code_resource.start <= res->end &&
+			code_resource.end <= res->end)
+			request_resource(res, &code_resource);
+
+		if (data_resource.start >= res->start &&
+			data_resource.start <= res->end &&
+			data_resource.end > res->end) {
+			sub_res = alloc_bootmem_low(sizeof(struct resource));
+			memcpy(sub_res, &data_resource,
+				sizeof(struct resource));
+			sub_res->end = res->end;
+			data_resource.start = res->end + 1;
+			request_resource(res, sub_res);
+		}
+
+		if (data_resource.start >= res->start &&
+			data_resource.start <= res->end &&
+			data_resource.end <= res->end)
+			request_resource(res, &data_resource);
 	}
 }
 
@@ -495,16 +611,13 @@ static void __init setup_memory_end(void)
 	}
 	if (!memory_end)
 		memory_end = memory_size;
-	if (real_size > memory_end)
-		printk("More memory detected than supported. Unused: %luk\n",
-		       (real_size - memory_end) >> 10);
 }
 
 static void __init
 setup_memory(void)
 {
         unsigned long bootmap_size;
-	unsigned long start_pfn, end_pfn, init_pfn;
+	unsigned long start_pfn, end_pfn;
 	int i;
 
 	/*
@@ -514,10 +627,6 @@ setup_memory(void)
 	start_pfn = PFN_UP(__pa(&_end));
 	end_pfn = max_pfn = PFN_DOWN(memory_end);
 
-	/* Initialize storage key for kernel pages */
-	for (init_pfn = 0 ; init_pfn < start_pfn; init_pfn++)
-		page_set_storage_key(init_pfn << PAGE_SHIFT, PAGE_DEFAULT_KEY);
-
 #ifdef CONFIG_BLK_DEV_INITRD
 	/*
 	 * Move the initrd in case the bitmap of the bootmem allocater
@@ -631,7 +740,7 @@ setup_arch(char **cmdline_p)
 #endif /* CONFIG_64BIT */
 
 	/* Save unparsed command line copy for /proc/cmdline */
-	strlcpy(saved_command_line, COMMAND_LINE, COMMAND_LINE_SIZE);
+	strlcpy(boot_command_line, COMMAND_LINE, COMMAND_LINE_SIZE);
 
 	*cmdline_p = COMMAND_LINE;
 	*(*cmdline_p + COMMAND_LINE_SIZE - 1) = '\0';
@@ -651,6 +760,7 @@ setup_arch(char **cmdline_p)
 	parse_early_param();
 
 	setup_memory_end();
+	setup_addressing_mode();
 	setup_memory();
 	setup_resources();
 	setup_lowcore();
@@ -694,6 +804,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
         struct cpuinfo_S390 *cpuinfo;
 	unsigned long n = (unsigned long) v - 1;
 
+	s390_adjust_jiffies();
 	preempt_disable();
 	if (!n) {
 		seq_printf(m, "vendor_id       : IBM/S390\n"
diff --git a/arch/s390/kernel/signal.c b/arch/s390/kernel/signal.c
index 4c8a7954ef48..554f9cf7499c 100644
--- a/arch/s390/kernel/signal.c
+++ b/arch/s390/kernel/signal.c
@@ -119,7 +119,7 @@ static int save_sigregs(struct pt_regs *regs, _sigregs __user *sregs)
 
 	/* Copy a 'clean' PSW mask to the user to avoid leaking
 	   information about whether PER is currently on.  */
-	user_sregs.regs.psw.mask = PSW_MASK_MERGE(PSW_USER_BITS, regs->psw.mask);
+	user_sregs.regs.psw.mask = PSW_MASK_MERGE(psw_user_bits, regs->psw.mask);
 	user_sregs.regs.psw.addr = regs->psw.addr;
 	memcpy(&user_sregs.regs.gprs, &regs->gprs, sizeof(sregs->regs.gprs));
 	memcpy(&user_sregs.regs.acrs, current->thread.acrs,
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index c0cd255fddbd..83a4ea6e3d60 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -22,23 +22,23 @@
 
 #include <linux/module.h>
 #include <linux/init.h>
-
 #include <linux/mm.h>
 #include <linux/spinlock.h>
 #include <linux/kernel_stat.h>
 #include <linux/smp_lock.h>
-
 #include <linux/delay.h>
 #include <linux/cache.h>
 #include <linux/interrupt.h>
 #include <linux/cpu.h>
-
+#include <linux/timex.h>
+#include <asm/setup.h>
 #include <asm/sigp.h>
 #include <asm/pgalloc.h>
 #include <asm/irq.h>
 #include <asm/s390_ext.h>
 #include <asm/cpcmd.h>
 #include <asm/tlbflush.h>
+#include <asm/timer.h>
 
 extern volatile int __cpu_logical_map[];
 
@@ -53,17 +53,11 @@ cpumask_t cpu_possible_map = CPU_MASK_NONE;
 
 static struct task_struct *current_set[NR_CPUS];
 
-/*
- * Reboot, halt and power_off routines for SMP.
- */
-extern char vmhalt_cmd[];
-extern char vmpoff_cmd[];
-
 static void smp_ext_bitcall(int, ec_bit_sig);
 static void smp_ext_bitcall_others(ec_bit_sig);
 
 /*
-5B * Structure and data for smp_call_function(). This is designed to minimise
+ * Structure and data for smp_call_function(). This is designed to minimise
  * static memory requirements. It also looks cleaner.
  */
 static DEFINE_SPINLOCK(call_lock);
@@ -110,7 +104,7 @@ int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
  * remote CPUs are nearly ready to execute <<func>> or are or have executed.
  *
  * You must not call this function with disabled interrupts or from a
- * hardware interrupt handler or from a bottom half handler.
+ * hardware interrupt handler.
  */
 {
 	struct call_data_struct data;
@@ -119,8 +113,8 @@ int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
 	if (cpus <= 0)
 		return 0;
 
-	/* Can deadlock when called with interrupts disabled */
-	WARN_ON(irqs_disabled());
+	/* Can deadlock when interrupts are disabled or if in wrong context */
+	WARN_ON(irqs_disabled() || in_irq());
 
 	data.func = func;
 	data.info = info;
@@ -129,7 +123,7 @@ int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
 	if (wait)
 		atomic_set(&data.finished, 0);
 
-	spin_lock(&call_lock);
+	spin_lock_bh(&call_lock);
 	call_data = &data;
 	/* Send a message to all other CPUs and wait for them to respond */
         smp_ext_bitcall_others(ec_call_function);
@@ -141,7 +135,7 @@ int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
 	if (wait)
 		while (atomic_read(&data.finished) != cpus)
 			cpu_relax();
-	spin_unlock(&call_lock);
+	spin_unlock_bh(&call_lock);
 
 	return 0;
 }
@@ -165,6 +159,9 @@ int smp_call_function_on(void (*func) (void *info), void *info,
 	if (!cpu_online(cpu))
 		return -EINVAL;
 
+	/* Can deadlock when interrupts are disabled or if in wrong context */
+	WARN_ON(irqs_disabled() || in_irq());
+
 	/* disable preemption for local function call */
 	curr_cpu = get_cpu();
 
@@ -200,7 +197,7 @@ int smp_call_function_on(void (*func) (void *info), void *info,
 }
 EXPORT_SYMBOL(smp_call_function_on);
 
-static inline void do_send_stop(void)
+static void do_send_stop(void)
 {
         int cpu, rc;
 
@@ -214,7 +211,7 @@ static inline void do_send_stop(void)
 	}
 }
 
-static inline void do_store_status(void)
+static void do_store_status(void)
 {
         int cpu, rc;
 
@@ -230,7 +227,7 @@ static inline void do_store_status(void)
         }
 }
 
-static inline void do_wait_for_stop(void)
+static void do_wait_for_stop(void)
 {
 	int cpu;
 
@@ -250,7 +247,7 @@ static inline void do_wait_for_stop(void)
 void smp_send_stop(void)
 {
 	/* Disable all interrupts/machine checks */
-	__load_psw_mask(PSW_KERNEL_BITS & ~PSW_MASK_MCHECK);
+	__load_psw_mask(psw_kernel_bits & ~PSW_MASK_MCHECK);
 
         /* write magic number to zero page (absolute 0) */
 	lowcore_ptr[smp_processor_id()]->panic_magic = __PANIC_MAGIC;
@@ -298,7 +295,7 @@ void machine_power_off_smp(void)
  * cpus are handled.
  */
 
-void do_ext_call_interrupt(__u16 code)
+static void do_ext_call_interrupt(__u16 code)
 {
         unsigned long bits;
 
@@ -385,7 +382,7 @@ struct ec_creg_mask_parms {
 /*
  * callback for setting/clearing control bits
  */
-void smp_ctl_bit_callback(void *info) {
+static void smp_ctl_bit_callback(void *info) {
 	struct ec_creg_mask_parms *pp = info;
 	unsigned long cregs[16];
 	int i;
@@ -458,17 +455,15 @@ __init smp_count_cpus(void)
 /*
  *      Activate a secondary processor.
  */
-extern void init_cpu_timer(void);
-extern void init_cpu_vtimer(void);
-
 int __devinit start_secondary(void *cpuvoid)
 {
         /* Setup the cpu */
         cpu_init();
 	preempt_disable();
-        /* init per CPU timer */
+	/* Enable TOD clock interrupts on the secondary cpu. */
         init_cpu_timer();
 #ifdef CONFIG_VIRT_TIMER
+	/* Enable cpu timer interrupts on the secondary cpu. */
         init_cpu_vtimer();
 #endif
 	/* Enable pfault pseudo page faults on this cpu. */
@@ -542,7 +537,7 @@ smp_put_cpu(int cpu)
 	spin_unlock_irqrestore(&smp_reserve_lock, flags);
 }
 
-static inline int
+static int
 cpu_stopped(int cpu)
 {
 	__u32 status;
diff --git a/arch/s390/kernel/stacktrace.c b/arch/s390/kernel/stacktrace.c
index 0d14a4789bf2..2e5c65a1863e 100644
--- a/arch/s390/kernel/stacktrace.c
+++ b/arch/s390/kernel/stacktrace.c
@@ -11,11 +11,11 @@
 #include <linux/stacktrace.h>
 #include <linux/kallsyms.h>
 
-static inline unsigned long save_context_stack(struct stack_trace *trace,
-					       unsigned int *skip,
-					       unsigned long sp,
-					       unsigned long low,
-					       unsigned long high)
+static unsigned long save_context_stack(struct stack_trace *trace,
+					unsigned int *skip,
+					unsigned long sp,
+					unsigned long low,
+					unsigned long high)
 {
 	struct stack_frame *sf;
 	struct pt_regs *regs;
diff --git a/arch/s390/kernel/syscalls.S b/arch/s390/kernel/syscalls.S
index a4ceae3dbcf1..a52c44455bf0 100644
--- a/arch/s390/kernel/syscalls.S
+++ b/arch/s390/kernel/syscalls.S
@@ -124,7 +124,7 @@ NI_SYSCALL							/* old "idle" system call */
 NI_SYSCALL							/* vm86old for i386 */
 SYSCALL(sys_wait4,sys_wait4,compat_sys_wait4_wrapper)
 SYSCALL(sys_swapoff,sys_swapoff,sys32_swapoff_wrapper)		/* 115 */
-SYSCALL(sys_sysinfo,sys_sysinfo,sys32_sysinfo_wrapper)
+SYSCALL(sys_sysinfo,sys_sysinfo,compat_sys_sysinfo_wrapper)
 SYSCALL(sys_ipc,sys_ipc,sys32_ipc_wrapper)
 SYSCALL(sys_fsync,sys_fsync,sys32_fsync_wrapper)
 SYSCALL(sys_sigreturn_glue,sys_sigreturn_glue,sys32_sigreturn_glue)
diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c
index 6cceed4df73e..ee9fd7b85928 100644
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -37,11 +37,15 @@
 #include <asm/irq.h>
 #include <asm/irq_regs.h>
 #include <asm/timer.h>
+#include <asm/etr.h>
 
 /* change this if you have some constant time drift */
 #define USECS_PER_JIFFY     ((unsigned long) 1000000/HZ)
 #define CLK_TICKS_PER_JIFFY ((unsigned long) USECS_PER_JIFFY << 12)
 
+/* The value of the TOD clock for 1.1.1970. */
+#define TOD_UNIX_EPOCH 0x7d91048bca000000ULL
+
 /*
  * Create a small time difference between the timer interrupts
  * on the different cpus to avoid lock contention.
@@ -51,6 +55,7 @@
 #define TICK_SIZE tick
 
 static ext_int_info_t ext_int_info_cc;
+static ext_int_info_t ext_int_etr_cc;
 static u64 init_timer_cc;
 static u64 jiffies_timer_cc;
 static u64 xtime_cc;
@@ -89,29 +94,21 @@ void tod_to_timeval(__u64 todval, struct timespec *xtime)
 #define s390_do_profile()	do { ; } while(0)
 #endif /* CONFIG_PROFILING */
 
-
 /*
- * timer_interrupt() needs to keep up the real-time clock,
- * as well as call the "do_timer()" routine every clocktick
+ * Advance the per cpu tick counter up to the time given with the
+ * "time" argument. The per cpu update consists of accounting
+ * the virtual cpu time, calling update_process_times and calling
+ * the profiling hook. If xtime is before time it is advanced as well.
  */
-void account_ticks(void)
+void account_ticks(u64 time)
 {
-	__u64 tmp;
 	__u32 ticks;
+	__u64 tmp;
 
 	/* Calculate how many ticks have passed. */
-	if (S390_lowcore.int_clock < S390_lowcore.jiffy_timer) {
-		/*
-		 * We have to program the clock comparator even if
-		 * no tick has passed. That happens if e.g. an i/o
-		 * interrupt wakes up an idle processor that has
-		 * switched off its hz timer.
-		 */
-		tmp = S390_lowcore.jiffy_timer + CPU_DEVIATION;
-		asm volatile ("SCKC %0" : : "m" (tmp));
+	if (time < S390_lowcore.jiffy_timer)
 		return;
-	}
-	tmp = S390_lowcore.int_clock - S390_lowcore.jiffy_timer;
+	tmp = time - S390_lowcore.jiffy_timer;
 	if (tmp >= 2*CLK_TICKS_PER_JIFFY) {  /* more than two ticks ? */
 		ticks = __div(tmp, CLK_TICKS_PER_JIFFY) + 1;
 		S390_lowcore.jiffy_timer +=
@@ -124,10 +121,6 @@ void account_ticks(void)
 		S390_lowcore.jiffy_timer += CLK_TICKS_PER_JIFFY;
 	}
 
-	/* set clock comparator for next tick */
-	tmp = S390_lowcore.jiffy_timer + CPU_DEVIATION;
-        asm volatile ("SCKC %0" : : "m" (tmp));
-
 #ifdef CONFIG_SMP
 	/*
 	 * Do not rely on the boot cpu to do the calls to do_timer.
@@ -173,7 +166,7 @@ int sysctl_hz_timer = 1;
  * Stop the HZ tick on the current CPU.
  * Only cpu_idle may call this function.
  */
-static inline void stop_hz_timer(void)
+static void stop_hz_timer(void)
 {
 	unsigned long flags;
 	unsigned long seq, next;
@@ -210,20 +203,21 @@ static inline void stop_hz_timer(void)
 		if (timer >= jiffies_timer_cc)
 			todval = timer;
 	}
-	asm volatile ("SCKC %0" : : "m" (todval));
+	set_clock_comparator(todval);
 }
 
 /*
  * Start the HZ tick on the current CPU.
  * Only cpu_idle may call this function.
  */
-static inline void start_hz_timer(void)
+static void start_hz_timer(void)
 {
 	BUG_ON(!in_interrupt());
 
 	if (!cpu_isset(smp_processor_id(), nohz_cpu_mask))
 		return;
-	account_ticks();
+	account_ticks(get_clock());
+	set_clock_comparator(S390_lowcore.jiffy_timer + CPU_DEVIATION);
 	cpu_clear(smp_processor_id(), nohz_cpu_mask);
 }
 
@@ -245,7 +239,7 @@ static struct notifier_block nohz_idle_nb = {
 	.notifier_call = nohz_idle_notify,
 };
 
-void __init nohz_init(void)
+static void __init nohz_init(void)
 {
 	if (register_idle_notifier(&nohz_idle_nb))
 		panic("Couldn't register idle notifier");
@@ -254,24 +248,57 @@ void __init nohz_init(void)
 #endif
 
 /*
- * Start the clock comparator on the current CPU.
+ * Set up per cpu jiffy timer and set the clock comparator.
+ */
+static void setup_jiffy_timer(void)
+{
+	/* Set up clock comparator to next jiffy. */
+	S390_lowcore.jiffy_timer =
+		jiffies_timer_cc + (jiffies_64 + 1) * CLK_TICKS_PER_JIFFY;
+	set_clock_comparator(S390_lowcore.jiffy_timer + CPU_DEVIATION);
+}
+
+/*
+ * Set up lowcore and control register of the current cpu to
+ * enable TOD clock and clock comparator interrupts.
  */
 void init_cpu_timer(void)
 {
-	unsigned long cr0;
-	__u64 timer;
+	setup_jiffy_timer();
 
-	timer = jiffies_timer_cc + jiffies_64 * CLK_TICKS_PER_JIFFY;
-	S390_lowcore.jiffy_timer = timer + CLK_TICKS_PER_JIFFY;
-	timer += CLK_TICKS_PER_JIFFY + CPU_DEVIATION;
-	asm volatile ("SCKC %0" : : "m" (timer));
-        /* allow clock comparator timer interrupt */
-	__ctl_store(cr0, 0, 0);
-        cr0 |= 0x800;
-	__ctl_load(cr0, 0, 0);
+	/* Enable clock comparator timer interrupt. */
+	__ctl_set_bit(0,11);
+
+	/* Always allow ETR external interrupts, even without an ETR. */
+	__ctl_set_bit(0, 4);
 }
 
-extern void vtime_init(void);
+static void clock_comparator_interrupt(__u16 code)
+{
+	/* set clock comparator for next tick */
+	set_clock_comparator(S390_lowcore.jiffy_timer + CPU_DEVIATION);
+}
+
+static void etr_reset(void);
+static void etr_init(void);
+static void etr_ext_handler(__u16);
+
+/*
+ * Get the TOD clock running.
+ */
+static u64 __init reset_tod_clock(void)
+{
+	u64 time;
+
+	etr_reset();
+	if (store_clock(&time) == 0)
+		return time;
+	/* TOD clock not running. Set the clock to Unix Epoch. */
+	if (set_clock(TOD_UNIX_EPOCH) != 0 || store_clock(&time) != 0)
+		panic("TOD clock not operational.");
+
+	return TOD_UNIX_EPOCH;
+}
 
 static cycle_t read_tod_clock(void)
 {
@@ -285,7 +312,7 @@ static struct clocksource clocksource_tod = {
 	.mask		= -1ULL,
 	.mult		= 1000,
 	.shift		= 12,
-	.is_continuous	= 1,
+	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
 };
 
 
@@ -295,48 +322,31 @@ static struct clocksource clocksource_tod = {
  */
 void __init time_init(void)
 {
-	__u64 set_time_cc;
-	int cc;
-
-        /* kick the TOD clock */
-	asm volatile(
-		"	stck	0(%2)\n"
-		"	ipm	%0\n"
-		"	srl	%0,28"
-		: "=d" (cc), "=m" (init_timer_cc)
-		: "a" (&init_timer_cc) : "cc");
-        switch (cc) {
-        case 0: /* clock in set state: all is fine */
-                break;
-        case 1: /* clock in non-set state: FIXME */
-                printk("time_init: TOD clock in non-set state\n");
-                break;
-        case 2: /* clock in error state: FIXME */
-                printk("time_init: TOD clock in error state\n");
-                break;
-        case 3: /* clock in stopped or not-operational state: FIXME */
-                printk("time_init: TOD clock stopped/non-operational\n");
-                break;
-        }
+	init_timer_cc = reset_tod_clock();
+	xtime_cc = init_timer_cc + CLK_TICKS_PER_JIFFY;
 	jiffies_timer_cc = init_timer_cc - jiffies_64 * CLK_TICKS_PER_JIFFY;
 
 	/* set xtime */
-	xtime_cc = init_timer_cc + CLK_TICKS_PER_JIFFY;
-	set_time_cc = init_timer_cc - 0x8126d60e46000000LL +
-		(0x3c26700LL*1000000*4096);
-        tod_to_timeval(set_time_cc, &xtime);
+	tod_to_timeval(init_timer_cc - TOD_UNIX_EPOCH, &xtime);
         set_normalized_timespec(&wall_to_monotonic,
                                 -xtime.tv_sec, -xtime.tv_nsec);
 
 	/* request the clock comparator external interrupt */
-	if (register_early_external_interrupt(0x1004, NULL,
+	if (register_early_external_interrupt(0x1004,
+					      clock_comparator_interrupt,
 					      &ext_int_info_cc) != 0)
                 panic("Couldn't request external interrupt 0x1004");
 
 	if (clocksource_register(&clocksource_tod) != 0)
 		panic("Could not register TOD clock source");
 
-        init_cpu_timer();
+	/* request the etr external interrupt */
+	if (register_early_external_interrupt(0x1406, etr_ext_handler,
+					      &ext_int_etr_cc) != 0)
+		panic("Couldn't request external interrupt 0x1406");
+
+	/* Enable TOD clock interrupts on the boot cpu. */
+	init_cpu_timer();
 
 #ifdef CONFIG_NO_IDLE_HZ
 	nohz_init();
@@ -345,5 +355,1048 @@ void __init time_init(void)
 #ifdef CONFIG_VIRT_TIMER
 	vtime_init();
 #endif
+	etr_init();
+}
+
+/*
+ * External Time Reference (ETR) code.
+ */
+static int etr_port0_online;
+static int etr_port1_online;
+
+static int __init early_parse_etr(char *p)
+{
+	if (strncmp(p, "off", 3) == 0)
+		etr_port0_online = etr_port1_online = 0;
+	else if (strncmp(p, "port0", 5) == 0)
+		etr_port0_online = 1;
+	else if (strncmp(p, "port1", 5) == 0)
+		etr_port1_online = 1;
+	else if (strncmp(p, "on", 2) == 0)
+		etr_port0_online = etr_port1_online = 1;
+	return 0;
+}
+early_param("etr", early_parse_etr);
+
+enum etr_event {
+	ETR_EVENT_PORT0_CHANGE,
+	ETR_EVENT_PORT1_CHANGE,
+	ETR_EVENT_PORT_ALERT,
+	ETR_EVENT_SYNC_CHECK,
+	ETR_EVENT_SWITCH_LOCAL,
+	ETR_EVENT_UPDATE,
+};
+
+enum etr_flags {
+	ETR_FLAG_ENOSYS,
+	ETR_FLAG_EACCES,
+	ETR_FLAG_STEAI,
+};
+
+/*
+ * Valid bit combinations of the eacr register are (x = don't care):
+ * e0 e1 dp p0 p1 ea es sl
+ *  0  0  x  0	0  0  0  0  initial, disabled state
+ *  0  0  x  0	1  1  0  0  port 1 online
+ *  0  0  x  1	0  1  0  0  port 0 online
+ *  0  0  x  1	1  1  0  0  both ports online
+ *  0  1  x  0	1  1  0  0  port 1 online and usable, ETR or PPS mode
+ *  0  1  x  0	1  1  0  1  port 1 online, usable and ETR mode
+ *  0  1  x  0	1  1  1  0  port 1 online, usable, PPS mode, in-sync
+ *  0  1  x  0	1  1  1  1  port 1 online, usable, ETR mode, in-sync
+ *  0  1  x  1	1  1  0  0  both ports online, port 1 usable
+ *  0  1  x  1	1  1  1  0  both ports online, port 1 usable, PPS mode, in-sync
+ *  0  1  x  1	1  1  1  1  both ports online, port 1 usable, ETR mode, in-sync
+ *  1  0  x  1	0  1  0  0  port 0 online and usable, ETR or PPS mode
+ *  1  0  x  1	0  1  0  1  port 0 online, usable and ETR mode
+ *  1  0  x  1	0  1  1  0  port 0 online, usable, PPS mode, in-sync
+ *  1  0  x  1	0  1  1  1  port 0 online, usable, ETR mode, in-sync
+ *  1  0  x  1	1  1  0  0  both ports online, port 0 usable
+ *  1  0  x  1	1  1  1  0  both ports online, port 0 usable, PPS mode, in-sync
+ *  1  0  x  1	1  1  1  1  both ports online, port 0 usable, ETR mode, in-sync
+ *  1  1  x  1	1  1  1  0  both ports online & usable, ETR, in-sync
+ *  1  1  x  1	1  1  1  1  both ports online & usable, ETR, in-sync
+ */
+static struct etr_eacr etr_eacr;
+static u64 etr_tolec;			/* time of last eacr update */
+static unsigned long etr_flags;
+static struct etr_aib etr_port0;
+static int etr_port0_uptodate;
+static struct etr_aib etr_port1;
+static int etr_port1_uptodate;
+static unsigned long etr_events;
+static struct timer_list etr_timer;
+static struct tasklet_struct etr_tasklet;
+static DEFINE_PER_CPU(atomic_t, etr_sync_word);
+
+static void etr_timeout(unsigned long dummy);
+static void etr_tasklet_fn(unsigned long dummy);
+
+/*
+ * The etr get_clock function. It will write the current clock value
+ * to the clock pointer and return 0 if the clock is in sync with the
+ * external time source. If the clock mode is local it will return
+ * -ENOSYS and -EAGAIN if the clock is not in sync with the external
+ * reference. This function is what ETR is all about..
+ */
+int get_sync_clock(unsigned long long *clock)
+{
+	atomic_t *sw_ptr;
+	unsigned int sw0, sw1;
+
+	sw_ptr = &get_cpu_var(etr_sync_word);
+	sw0 = atomic_read(sw_ptr);
+	*clock = get_clock();
+	sw1 = atomic_read(sw_ptr);
+	put_cpu_var(etr_sync_sync);
+	if (sw0 == sw1 && (sw0 & 0x80000000U))
+		/* Success: time is in sync. */
+		return 0;
+	if (test_bit(ETR_FLAG_ENOSYS, &etr_flags))
+		return -ENOSYS;
+	if (test_bit(ETR_FLAG_EACCES, &etr_flags))
+		return -EACCES;
+	return -EAGAIN;
+}
+EXPORT_SYMBOL(get_sync_clock);
+
+/*
+ * Make get_sync_clock return -EAGAIN.
+ */
+static void etr_disable_sync_clock(void *dummy)
+{
+	atomic_t *sw_ptr = &__get_cpu_var(etr_sync_word);
+	/*
+	 * Clear the in-sync bit 2^31. All get_sync_clock calls will
+	 * fail until the sync bit is turned back on. In addition
+	 * increase the "sequence" counter to avoid the race of an
+	 * etr event and the complete recovery against get_sync_clock.
+	 */
+	atomic_clear_mask(0x80000000, sw_ptr);
+	atomic_inc(sw_ptr);
+}
+
+/*
+ * Make get_sync_clock return 0 again.
+ * Needs to be called from a context disabled for preemption.
+ */
+static void etr_enable_sync_clock(void)
+{
+	atomic_t *sw_ptr = &__get_cpu_var(etr_sync_word);
+	atomic_set_mask(0x80000000, sw_ptr);
+}
+
+/*
+ * Reset ETR attachment.
+ */
+static void etr_reset(void)
+{
+	etr_eacr =  (struct etr_eacr) {
+		.e0 = 0, .e1 = 0, ._pad0 = 4, .dp = 0,
+		.p0 = 0, .p1 = 0, ._pad1 = 0, .ea = 0,
+		.es = 0, .sl = 0 };
+	if (etr_setr(&etr_eacr) == 0)
+		etr_tolec = get_clock();
+	else {
+		set_bit(ETR_FLAG_ENOSYS, &etr_flags);
+		if (etr_port0_online || etr_port1_online) {
+			printk(KERN_WARNING "Running on non ETR capable "
+			       "machine, only local mode available.\n");
+			etr_port0_online = etr_port1_online = 0;
+		}
+	}
+}
+
+static void etr_init(void)
+{
+	struct etr_aib aib;
+
+	if (test_bit(ETR_FLAG_ENOSYS, &etr_flags))
+		return;
+	/* Check if this machine has the steai instruction. */
+	if (etr_steai(&aib, ETR_STEAI_STEPPING_PORT) == 0)
+		set_bit(ETR_FLAG_STEAI, &etr_flags);
+	setup_timer(&etr_timer, etr_timeout, 0UL);
+	tasklet_init(&etr_tasklet, etr_tasklet_fn, 0);
+	if (!etr_port0_online && !etr_port1_online)
+		set_bit(ETR_FLAG_EACCES, &etr_flags);
+	if (etr_port0_online) {
+		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	}
+	if (etr_port1_online) {
+		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	}
+}
+
+/*
+ * Two sorts of ETR machine checks. The architecture reads:
+ * "When a machine-check niterruption occurs and if a switch-to-local or
+ *  ETR-sync-check interrupt request is pending but disabled, this pending
+ *  disabled interruption request is indicated and is cleared".
+ * Which means that we can get etr_switch_to_local events from the machine
+ * check handler although the interruption condition is disabled. Lovely..
+ */
+
+/*
+ * Switch to local machine check. This is called when the last usable
+ * ETR port goes inactive. After switch to local the clock is not in sync.
+ */
+void etr_switch_to_local(void)
+{
+	if (!etr_eacr.sl)
+		return;
+	etr_disable_sync_clock(NULL);
+	set_bit(ETR_EVENT_SWITCH_LOCAL, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+/*
+ * ETR sync check machine check. This is called when the ETR OTE and the
+ * local clock OTE are farther apart than the ETR sync check tolerance.
+ * After a ETR sync check the clock is not in sync. The machine check
+ * is broadcasted to all cpus at the same time.
+ */
+void etr_sync_check(void)
+{
+	if (!etr_eacr.es)
+		return;
+	etr_disable_sync_clock(NULL);
+	set_bit(ETR_EVENT_SYNC_CHECK, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+/*
+ * ETR external interrupt. There are two causes:
+ * 1) port state change, check the usability of the port
+ * 2) port alert, one of the ETR-data-validity bits (v1-v2 bits of the
+ *    sldr-status word) or ETR-data word 1 (edf1) or ETR-data word 3 (edf3)
+ *    or ETR-data word 4 (edf4) has changed.
+ */
+static void etr_ext_handler(__u16 code)
+{
+	struct etr_interruption_parameter *intparm =
+		(struct etr_interruption_parameter *) &S390_lowcore.ext_params;
+
+	if (intparm->pc0)
+		/* ETR port 0 state change. */
+		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
+	if (intparm->pc1)
+		/* ETR port 1 state change. */
+		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
+	if (intparm->eai)
+		/*
+		 * ETR port alert on either port 0, 1 or both.
+		 * Both ports are not up-to-date now.
+		 */
+		set_bit(ETR_EVENT_PORT_ALERT, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+static void etr_timeout(unsigned long dummy)
+{
+	set_bit(ETR_EVENT_UPDATE, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+/*
+ * Check if the etr mode is pss.
+ */
+static inline int etr_mode_is_pps(struct etr_eacr eacr)
+{
+	return eacr.es && !eacr.sl;
+}
+
+/*
+ * Check if the etr mode is etr.
+ */
+static inline int etr_mode_is_etr(struct etr_eacr eacr)
+{
+	return eacr.es && eacr.sl;
+}
+
+/*
+ * Check if the port can be used for TOD synchronization.
+ * For PPS mode the port has to receive OTEs. For ETR mode
+ * the port has to receive OTEs, the ETR stepping bit has to
+ * be zero and the validity bits for data frame 1, 2, and 3
+ * have to be 1.
+ */
+static int etr_port_valid(struct etr_aib *aib, int port)
+{
+	unsigned int psc;
+
+	/* Check that this port is receiving OTEs. */
+	if (aib->tsp == 0)
+		return 0;
+
+	psc = port ? aib->esw.psc1 : aib->esw.psc0;
+	if (psc == etr_lpsc_pps_mode)
+		return 1;
+	if (psc == etr_lpsc_operational_step)
+		return !aib->esw.y && aib->slsw.v1 &&
+			aib->slsw.v2 && aib->slsw.v3;
+	return 0;
+}
+
+/*
+ * Check if two ports are on the same network.
+ */
+static int etr_compare_network(struct etr_aib *aib1, struct etr_aib *aib2)
+{
+	// FIXME: any other fields we have to compare?
+	return aib1->edf1.net_id == aib2->edf1.net_id;
+}
+
+/*
+ * Wrapper for etr_stei that converts physical port states
+ * to logical port states to be consistent with the output
+ * of stetr (see etr_psc vs. etr_lpsc).
+ */
+static void etr_steai_cv(struct etr_aib *aib, unsigned int func)
+{
+	BUG_ON(etr_steai(aib, func) != 0);
+	/* Convert port state to logical port state. */
+	if (aib->esw.psc0 == 1)
+		aib->esw.psc0 = 2;
+	else if (aib->esw.psc0 == 0 && aib->esw.p == 0)
+		aib->esw.psc0 = 1;
+	if (aib->esw.psc1 == 1)
+		aib->esw.psc1 = 2;
+	else if (aib->esw.psc1 == 0 && aib->esw.p == 1)
+		aib->esw.psc1 = 1;
+}
+
+/*
+ * Check if the aib a2 is still connected to the same attachment as
+ * aib a1, the etv values differ by one and a2 is valid.
+ */
+static int etr_aib_follows(struct etr_aib *a1, struct etr_aib *a2, int p)
+{
+	int state_a1, state_a2;
+
+	/* Paranoia check: e0/e1 should better be the same. */
+	if (a1->esw.eacr.e0 != a2->esw.eacr.e0 ||
+	    a1->esw.eacr.e1 != a2->esw.eacr.e1)
+		return 0;
+
+	/* Still connected to the same etr ? */
+	state_a1 = p ? a1->esw.psc1 : a1->esw.psc0;
+	state_a2 = p ? a2->esw.psc1 : a2->esw.psc0;
+	if (state_a1 == etr_lpsc_operational_step) {
+		if (state_a2 != etr_lpsc_operational_step ||
+		    a1->edf1.net_id != a2->edf1.net_id ||
+		    a1->edf1.etr_id != a2->edf1.etr_id ||
+		    a1->edf1.etr_pn != a2->edf1.etr_pn)
+			return 0;
+	} else if (state_a2 != etr_lpsc_pps_mode)
+		return 0;
+
+	/* The ETV value of a2 needs to be ETV of a1 + 1. */
+	if (a1->edf2.etv + 1 != a2->edf2.etv)
+		return 0;
+
+	if (!etr_port_valid(a2, p))
+		return 0;
+
+	return 1;
+}
+
+/*
+ * The time is "clock". xtime is what we think the time is.
+ * Adjust the value by a multiple of jiffies and add the delta to ntp.
+ * "delay" is an approximation how long the synchronization took. If
+ * the time correction is positive, then "delay" is subtracted from
+ * the time difference and only the remaining part is passed to ntp.
+ */
+static void etr_adjust_time(unsigned long long clock, unsigned long long delay)
+{
+	unsigned long long delta, ticks;
+	struct timex adjust;
+
+	/*
+	 * We don't have to take the xtime lock because the cpu
+	 * executing etr_adjust_time is running disabled in
+	 * tasklet context and all other cpus are looping in
+	 * etr_sync_cpu_start.
+	 */
+	if (clock > xtime_cc) {
+		/* It is later than we thought. */
+		delta = ticks = clock - xtime_cc;
+		delta = ticks = (delta < delay) ? 0 : delta - delay;
+		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
+		init_timer_cc = init_timer_cc + delta;
+		jiffies_timer_cc = jiffies_timer_cc + delta;
+		xtime_cc = xtime_cc + delta;
+		adjust.offset = ticks * (1000000 / HZ);
+	} else {
+		/* It is earlier than we thought. */
+		delta = ticks = xtime_cc - clock;
+		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
+		init_timer_cc = init_timer_cc - delta;
+		jiffies_timer_cc = jiffies_timer_cc - delta;
+		xtime_cc = xtime_cc - delta;
+		adjust.offset = -ticks * (1000000 / HZ);
+	}
+	if (adjust.offset != 0) {
+		printk(KERN_NOTICE "etr: time adjusted by %li micro-seconds\n",
+		       adjust.offset);
+		adjust.modes = ADJ_OFFSET_SINGLESHOT;
+		do_adjtimex(&adjust);
+	}
+}
+
+static void etr_sync_cpu_start(void *dummy)
+{
+	int *in_sync = dummy;
+
+	etr_enable_sync_clock();
+	/*
+	 * This looks like a busy wait loop but it isn't. etr_sync_cpus
+	 * is called on all other cpus while the TOD clocks is stopped.
+	 * __udelay will stop the cpu on an enabled wait psw until the
+	 * TOD is running again.
+	 */
+	while (*in_sync == 0)
+		__udelay(1);
+	if (*in_sync != 1)
+		/* Didn't work. Clear per-cpu in sync bit again. */
+		etr_disable_sync_clock(NULL);
+	/*
+	 * This round of TOD syncing is done. Set the clock comparator
+	 * to the next tick and let the processor continue.
+	 */
+	setup_jiffy_timer();
+}
+
+static void etr_sync_cpu_end(void *dummy)
+{
+}
+
+/*
+ * Sync the TOD clock using the port refered to by aibp. This port
+ * has to be enabled and the other port has to be disabled. The
+ * last eacr update has to be more than 1.6 seconds in the past.
+ */
+static int etr_sync_clock(struct etr_aib *aib, int port)
+{
+	struct etr_aib *sync_port;
+	unsigned long long clock, delay;
+	int in_sync, follows;
+	int rc;
+
+	/* Check if the current aib is adjacent to the sync port aib. */
+	sync_port = (port == 0) ? &etr_port0 : &etr_port1;
+	follows = etr_aib_follows(sync_port, aib, port);
+	memcpy(sync_port, aib, sizeof(*aib));
+	if (!follows)
+		return -EAGAIN;
+
+	/*
+	 * Catch all other cpus and make them wait until we have
+	 * successfully synced the clock. smp_call_function will
+	 * return after all other cpus are in etr_sync_cpu_start.
+	 */
+	in_sync = 0;
+	preempt_disable();
+	smp_call_function(etr_sync_cpu_start,&in_sync,0,0);
+	local_irq_disable();
+	etr_enable_sync_clock();
+
+	/* Set clock to next OTE. */
+	__ctl_set_bit(14, 21);
+	__ctl_set_bit(0, 29);
+	clock = ((unsigned long long) (aib->edf2.etv + 1)) << 32;
+	if (set_clock(clock) == 0) {
+		__udelay(1);	/* Wait for the clock to start. */
+		__ctl_clear_bit(0, 29);
+		__ctl_clear_bit(14, 21);
+		etr_stetr(aib);
+		/* Adjust Linux timing variables. */
+		delay = (unsigned long long)
+			(aib->edf2.etv - sync_port->edf2.etv) << 32;
+		etr_adjust_time(clock, delay);
+		setup_jiffy_timer();
+		/* Verify that the clock is properly set. */
+		if (!etr_aib_follows(sync_port, aib, port)) {
+			/* Didn't work. */
+			etr_disable_sync_clock(NULL);
+			in_sync = -EAGAIN;
+			rc = -EAGAIN;
+		} else {
+			in_sync = 1;
+			rc = 0;
+		}
+	} else {
+		/* Could not set the clock ?!? */
+		__ctl_clear_bit(0, 29);
+		__ctl_clear_bit(14, 21);
+		etr_disable_sync_clock(NULL);
+		in_sync = -EAGAIN;
+		rc = -EAGAIN;
+	}
+	local_irq_enable();
+	smp_call_function(etr_sync_cpu_end,NULL,0,0);
+	preempt_enable();
+	return rc;
+}
+
+/*
+ * Handle the immediate effects of the different events.
+ * The port change event is used for online/offline changes.
+ */
+static struct etr_eacr etr_handle_events(struct etr_eacr eacr)
+{
+	if (test_and_clear_bit(ETR_EVENT_SYNC_CHECK, &etr_events))
+		eacr.es = 0;
+	if (test_and_clear_bit(ETR_EVENT_SWITCH_LOCAL, &etr_events))
+		eacr.es = eacr.sl = 0;
+	if (test_and_clear_bit(ETR_EVENT_PORT_ALERT, &etr_events))
+		etr_port0_uptodate = etr_port1_uptodate = 0;
+
+	if (test_and_clear_bit(ETR_EVENT_PORT0_CHANGE, &etr_events)) {
+		if (eacr.e0)
+			/*
+			 * Port change of an enabled port. We have to
+			 * assume that this can have caused an stepping
+			 * port switch.
+			 */
+			etr_tolec = get_clock();
+		eacr.p0 = etr_port0_online;
+		if (!eacr.p0)
+			eacr.e0 = 0;
+		etr_port0_uptodate = 0;
+	}
+	if (test_and_clear_bit(ETR_EVENT_PORT1_CHANGE, &etr_events)) {
+		if (eacr.e1)
+			/*
+			 * Port change of an enabled port. We have to
+			 * assume that this can have caused an stepping
+			 * port switch.
+			 */
+			etr_tolec = get_clock();
+		eacr.p1 = etr_port1_online;
+		if (!eacr.p1)
+			eacr.e1 = 0;
+		etr_port1_uptodate = 0;
+	}
+	clear_bit(ETR_EVENT_UPDATE, &etr_events);
+	return eacr;
+}
+
+/*
+ * Set up a timer that expires after the etr_tolec + 1.6 seconds if
+ * one of the ports needs an update.
+ */
+static void etr_set_tolec_timeout(unsigned long long now)
+{
+	unsigned long micros;
+
+	if ((!etr_eacr.p0 || etr_port0_uptodate) &&
+	    (!etr_eacr.p1 || etr_port1_uptodate))
+		return;
+	micros = (now > etr_tolec) ? ((now - etr_tolec) >> 12) : 0;
+	micros = (micros > 1600000) ? 0 : 1600000 - micros;
+	mod_timer(&etr_timer, jiffies + (micros * HZ) / 1000000 + 1);
+}
+
+/*
+ * Set up a time that expires after 1/2 second.
+ */
+static void etr_set_sync_timeout(void)
+{
+	mod_timer(&etr_timer, jiffies + HZ/2);
+}
+
+/*
+ * Update the aib information for one or both ports.
+ */
+static struct etr_eacr etr_handle_update(struct etr_aib *aib,
+					 struct etr_eacr eacr)
+{
+	/* With both ports disabled the aib information is useless. */
+	if (!eacr.e0 && !eacr.e1)
+		return eacr;
+
+	/* Update port0 or port1 with aib stored in etr_tasklet_fn. */
+	if (aib->esw.q == 0) {
+		/* Information for port 0 stored. */
+		if (eacr.p0 && !etr_port0_uptodate) {
+			etr_port0 = *aib;
+			if (etr_port0_online)
+				etr_port0_uptodate = 1;
+		}
+	} else {
+		/* Information for port 1 stored. */
+		if (eacr.p1 && !etr_port1_uptodate) {
+			etr_port1 = *aib;
+			if (etr_port0_online)
+				etr_port1_uptodate = 1;
+		}
+	}
+
+	/*
+	 * Do not try to get the alternate port aib if the clock
+	 * is not in sync yet.
+	 */
+	if (!eacr.es)
+		return eacr;
+
+	/*
+	 * If steai is available we can get the information about
+	 * the other port immediately. If only stetr is available the
+	 * data-port bit toggle has to be used.
+	 */
+	if (test_bit(ETR_FLAG_STEAI, &etr_flags)) {
+		if (eacr.p0 && !etr_port0_uptodate) {
+			etr_steai_cv(&etr_port0, ETR_STEAI_PORT_0);
+			etr_port0_uptodate = 1;
+		}
+		if (eacr.p1 && !etr_port1_uptodate) {
+			etr_steai_cv(&etr_port1, ETR_STEAI_PORT_1);
+			etr_port1_uptodate = 1;
+		}
+	} else {
+		/*
+		 * One port was updated above, if the other
+		 * port is not uptodate toggle dp bit.
+		 */
+		if ((eacr.p0 && !etr_port0_uptodate) ||
+		    (eacr.p1 && !etr_port1_uptodate))
+			eacr.dp ^= 1;
+		else
+			eacr.dp = 0;
+	}
+	return eacr;
+}
+
+/*
+ * Write new etr control register if it differs from the current one.
+ * Return 1 if etr_tolec has been updated as well.
+ */
+static void etr_update_eacr(struct etr_eacr eacr)
+{
+	int dp_changed;
+
+	if (memcmp(&etr_eacr, &eacr, sizeof(eacr)) == 0)
+		/* No change, return. */
+		return;
+	/*
+	 * The disable of an active port of the change of the data port
+	 * bit can/will cause a change in the data port.
+	 */
+	dp_changed = etr_eacr.e0 > eacr.e0 || etr_eacr.e1 > eacr.e1 ||
+		(etr_eacr.dp ^ eacr.dp) != 0;
+	etr_eacr = eacr;
+	etr_setr(&etr_eacr);
+	if (dp_changed)
+		etr_tolec = get_clock();
+}
+
+/*
+ * ETR tasklet. In this function you'll find the main logic. In
+ * particular this is the only function that calls etr_update_eacr(),
+ * it "controls" the etr control register.
+ */
+static void etr_tasklet_fn(unsigned long dummy)
+{
+	unsigned long long now;
+	struct etr_eacr eacr;
+	struct etr_aib aib;
+	int sync_port;
+
+	/* Create working copy of etr_eacr. */
+	eacr = etr_eacr;
+
+	/* Check for the different events and their immediate effects. */
+	eacr = etr_handle_events(eacr);
+
+	/* Check if ETR is supposed to be active. */
+	eacr.ea = eacr.p0 || eacr.p1;
+	if (!eacr.ea) {
+		/* Both ports offline. Reset everything. */
+		eacr.dp = eacr.es = eacr.sl = 0;
+		on_each_cpu(etr_disable_sync_clock, NULL, 0, 1);
+		del_timer_sync(&etr_timer);
+		etr_update_eacr(eacr);
+		set_bit(ETR_FLAG_EACCES, &etr_flags);
+		return;
+	}
+
+	/* Store aib to get the current ETR status word. */
+	BUG_ON(etr_stetr(&aib) != 0);
+	etr_port0.esw = etr_port1.esw = aib.esw;	/* Copy status word. */
+	now = get_clock();
+
+	/*
+	 * Update the port information if the last stepping port change
+	 * or data port change is older than 1.6 seconds.
+	 */
+	if (now >= etr_tolec + (1600000 << 12))
+		eacr = etr_handle_update(&aib, eacr);
+
+	/*
+	 * Select ports to enable. The prefered synchronization mode is PPS.
+	 * If a port can be enabled depends on a number of things:
+	 * 1) The port needs to be online and uptodate. A port is not
+	 *    disabled just because it is not uptodate, but it is only
+	 *    enabled if it is uptodate.
+	 * 2) The port needs to have the same mode (pps / etr).
+	 * 3) The port needs to be usable -> etr_port_valid() == 1
+	 * 4) To enable the second port the clock needs to be in sync.
+	 * 5) If both ports are useable and are ETR ports, the network id
+	 *    has to be the same.
+	 * The eacr.sl bit is used to indicate etr mode vs. pps mode.
+	 */
+	if (eacr.p0 && aib.esw.psc0 == etr_lpsc_pps_mode) {
+		eacr.sl = 0;
+		eacr.e0 = 1;
+		if (!etr_mode_is_pps(etr_eacr))
+			eacr.es = 0;
+		if (!eacr.es || !eacr.p1 || aib.esw.psc1 != etr_lpsc_pps_mode)
+			eacr.e1 = 0;
+		// FIXME: uptodate checks ?
+		else if (etr_port0_uptodate && etr_port1_uptodate)
+			eacr.e1 = 1;
+		sync_port = (etr_port0_uptodate &&
+			     etr_port_valid(&etr_port0, 0)) ? 0 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else if (eacr.p1 && aib.esw.psc1 == etr_lpsc_pps_mode) {
+		eacr.sl = 0;
+		eacr.e0 = 0;
+		eacr.e1 = 1;
+		if (!etr_mode_is_pps(etr_eacr))
+			eacr.es = 0;
+		sync_port = (etr_port1_uptodate &&
+			     etr_port_valid(&etr_port1, 1)) ? 1 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else if (eacr.p0 && aib.esw.psc0 == etr_lpsc_operational_step) {
+		eacr.sl = 1;
+		eacr.e0 = 1;
+		if (!etr_mode_is_etr(etr_eacr))
+			eacr.es = 0;
+		if (!eacr.es || !eacr.p1 ||
+		    aib.esw.psc1 != etr_lpsc_operational_alt)
+			eacr.e1 = 0;
+		else if (etr_port0_uptodate && etr_port1_uptodate &&
+			 etr_compare_network(&etr_port0, &etr_port1))
+			eacr.e1 = 1;
+		sync_port = (etr_port0_uptodate &&
+			     etr_port_valid(&etr_port0, 0)) ? 0 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else if (eacr.p1 && aib.esw.psc1 == etr_lpsc_operational_step) {
+		eacr.sl = 1;
+		eacr.e0 = 0;
+		eacr.e1 = 1;
+		if (!etr_mode_is_etr(etr_eacr))
+			eacr.es = 0;
+		sync_port = (etr_port1_uptodate &&
+			     etr_port_valid(&etr_port1, 1)) ? 1 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else {
+		/* Both ports not usable. */
+		eacr.es = eacr.sl = 0;
+		sync_port = -1;
+		set_bit(ETR_FLAG_EACCES, &etr_flags);
+	}
+
+	/*
+	 * If the clock is in sync just update the eacr and return.
+	 * If there is no valid sync port wait for a port update.
+	 */
+	if (eacr.es || sync_port < 0) {
+		etr_update_eacr(eacr);
+		etr_set_tolec_timeout(now);
+		return;
+	}
+
+	/*
+	 * Prepare control register for clock syncing
+	 * (reset data port bit, set sync check control.
+	 */
+	eacr.dp = 0;
+	eacr.es = 1;
+
+	/*
+	 * Update eacr and try to synchronize the clock. If the update
+	 * of eacr caused a stepping port switch (or if we have to
+	 * assume that a stepping port switch has occured) or the
+	 * clock syncing failed, reset the sync check control bit
+	 * and set up a timer to try again after 0.5 seconds
+	 */
+	etr_update_eacr(eacr);
+	if (now < etr_tolec + (1600000 << 12) ||
+	    etr_sync_clock(&aib, sync_port) != 0) {
+		/* Sync failed. Try again in 1/2 second. */
+		eacr.es = 0;
+		etr_update_eacr(eacr);
+		etr_set_sync_timeout();
+	} else
+		etr_set_tolec_timeout(now);
+}
+
+/*
+ * Sysfs interface functions
+ */
+static struct sysdev_class etr_sysclass = {
+	set_kset_name("etr")
+};
+
+static struct sys_device etr_port0_dev = {
+	.id	= 0,
+	.cls	= &etr_sysclass,
+};
+
+static struct sys_device etr_port1_dev = {
+	.id	= 1,
+	.cls	= &etr_sysclass,
+};
+
+/*
+ * ETR class attributes
+ */
+static ssize_t etr_stepping_port_show(struct sysdev_class *class, char *buf)
+{
+	return sprintf(buf, "%i\n", etr_port0.esw.p);
+}
+
+static SYSDEV_CLASS_ATTR(stepping_port, 0400, etr_stepping_port_show, NULL);
+
+static ssize_t etr_stepping_mode_show(struct sysdev_class *class, char *buf)
+{
+	char *mode_str;
+
+	if (etr_mode_is_pps(etr_eacr))
+		mode_str = "pps";
+	else if (etr_mode_is_etr(etr_eacr))
+		mode_str = "etr";
+	else
+		mode_str = "local";
+	return sprintf(buf, "%s\n", mode_str);
+}
+
+static SYSDEV_CLASS_ATTR(stepping_mode, 0400, etr_stepping_mode_show, NULL);
+
+/*
+ * ETR port attributes
+ */
+static inline struct etr_aib *etr_aib_from_dev(struct sys_device *dev)
+{
+	if (dev == &etr_port0_dev)
+		return etr_port0_online ? &etr_port0 : NULL;
+	else
+		return etr_port1_online ? &etr_port1 : NULL;
+}
+
+static ssize_t etr_online_show(struct sys_device *dev, char *buf)
+{
+	unsigned int online;
+
+	online = (dev == &etr_port0_dev) ? etr_port0_online : etr_port1_online;
+	return sprintf(buf, "%i\n", online);
+}
+
+static ssize_t etr_online_store(struct sys_device *dev,
+			      const char *buf, size_t count)
+{
+	unsigned int value;
+
+	value = simple_strtoul(buf, NULL, 0);
+	if (value != 0 && value != 1)
+		return -EINVAL;
+	if (test_bit(ETR_FLAG_ENOSYS, &etr_flags))
+		return -ENOSYS;
+	if (dev == &etr_port0_dev) {
+		if (etr_port0_online == value)
+			return count;	/* Nothing to do. */
+		etr_port0_online = value;
+		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	} else {
+		if (etr_port1_online == value)
+			return count;	/* Nothing to do. */
+		etr_port1_online = value;
+		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	}
+	return count;
+}
+
+static SYSDEV_ATTR(online, 0600, etr_online_show, etr_online_store);
+
+static ssize_t etr_stepping_control_show(struct sys_device *dev, char *buf)
+{
+	return sprintf(buf, "%i\n", (dev == &etr_port0_dev) ?
+		       etr_eacr.e0 : etr_eacr.e1);
+}
+
+static SYSDEV_ATTR(stepping_control, 0400, etr_stepping_control_show, NULL);
+
+static ssize_t etr_mode_code_show(struct sys_device *dev, char *buf)
+{
+	if (!etr_port0_online && !etr_port1_online)
+		/* Status word is not uptodate if both ports are offline. */
+		return -ENODATA;
+	return sprintf(buf, "%i\n", (dev == &etr_port0_dev) ?
+		       etr_port0.esw.psc0 : etr_port0.esw.psc1);
+}
+
+static SYSDEV_ATTR(state_code, 0400, etr_mode_code_show, NULL);
+
+static ssize_t etr_untuned_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.u);
+}
+
+static SYSDEV_ATTR(untuned, 0400, etr_untuned_show, NULL);
+
+static ssize_t etr_network_id_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.net_id);
+}
+
+static SYSDEV_ATTR(network, 0400, etr_network_id_show, NULL);
+
+static ssize_t etr_id_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.etr_id);
+}
+
+static SYSDEV_ATTR(id, 0400, etr_id_show, NULL);
+
+static ssize_t etr_port_number_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.etr_pn);
+}
+
+static SYSDEV_ATTR(port, 0400, etr_port_number_show, NULL);
+
+static ssize_t etr_coupled_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v3)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf3.c);
+}
+
+static SYSDEV_ATTR(coupled, 0400, etr_coupled_show, NULL);
+
+static ssize_t etr_local_time_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v3)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf3.blto);
+}
+
+static SYSDEV_ATTR(local_time, 0400, etr_local_time_show, NULL);
+
+static ssize_t etr_utc_offset_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v3)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf3.buo);
+}
+
+static SYSDEV_ATTR(utc_offset, 0400, etr_utc_offset_show, NULL);
+
+static struct sysdev_attribute *etr_port_attributes[] = {
+	&attr_online,
+	&attr_stepping_control,
+	&attr_state_code,
+	&attr_untuned,
+	&attr_network,
+	&attr_id,
+	&attr_port,
+	&attr_coupled,
+	&attr_local_time,
+	&attr_utc_offset,
+	NULL
+};
+
+static int __init etr_register_port(struct sys_device *dev)
+{
+	struct sysdev_attribute **attr;
+	int rc;
+
+	rc = sysdev_register(dev);
+	if (rc)
+		goto out;
+	for (attr = etr_port_attributes; *attr; attr++) {
+		rc = sysdev_create_file(dev, *attr);
+		if (rc)
+			goto out_unreg;
+	}
+	return 0;
+out_unreg:
+	for (; attr >= etr_port_attributes; attr--)
+		sysdev_remove_file(dev, *attr);
+	sysdev_unregister(dev);
+out:
+	return rc;
+}
+
+static void __init etr_unregister_port(struct sys_device *dev)
+{
+	struct sysdev_attribute **attr;
+
+	for (attr = etr_port_attributes; *attr; attr++)
+		sysdev_remove_file(dev, *attr);
+	sysdev_unregister(dev);
+}
+
+static int __init etr_init_sysfs(void)
+{
+	int rc;
+
+	rc = sysdev_class_register(&etr_sysclass);
+	if (rc)
+		goto out;
+	rc = sysdev_class_create_file(&etr_sysclass, &attr_stepping_port);
+	if (rc)
+		goto out_unreg_class;
+	rc = sysdev_class_create_file(&etr_sysclass, &attr_stepping_mode);
+	if (rc)
+		goto out_remove_stepping_port;
+	rc = etr_register_port(&etr_port0_dev);
+	if (rc)
+		goto out_remove_stepping_mode;
+	rc = etr_register_port(&etr_port1_dev);
+	if (rc)
+		goto out_remove_port0;
+	return 0;
+
+out_remove_port0:
+	etr_unregister_port(&etr_port0_dev);
+out_remove_stepping_mode:
+	sysdev_class_remove_file(&etr_sysclass, &attr_stepping_mode);
+out_remove_stepping_port:
+	sysdev_class_remove_file(&etr_sysclass, &attr_stepping_port);
+out_unreg_class:
+	sysdev_class_unregister(&etr_sysclass);
+out:
+	return rc;
 }
 
+device_initcall(etr_init_sysfs);
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index 3cbb0dcf1f1d..f0e5a320e2ec 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -283,7 +283,7 @@ char *task_show_regs(struct task_struct *task, char *buffer)
 	return buffer;
 }
 
-DEFINE_SPINLOCK(die_lock);
+static DEFINE_SPINLOCK(die_lock);
 
 void die(const char * str, struct pt_regs * regs, long err)
 {
@@ -364,8 +364,7 @@ void __kprobes do_single_step(struct pt_regs *regs)
 		force_sig(SIGTRAP, current);
 }
 
-asmlinkage void
-default_trap_handler(struct pt_regs * regs, long interruption_code)
+static void default_trap_handler(struct pt_regs * regs, long interruption_code)
 {
         if (regs->psw.mask & PSW_MASK_PSTATE) {
 		local_irq_enable();
@@ -376,7 +375,7 @@ default_trap_handler(struct pt_regs * regs, long interruption_code)
 }
 
 #define DO_ERROR_INFO(signr, str, name, sicode, siaddr) \
-asmlinkage void name(struct pt_regs * regs, long interruption_code) \
+static void name(struct pt_regs * regs, long interruption_code) \
 { \
         siginfo_t info; \
         info.si_signo = signr; \
@@ -442,7 +441,7 @@ do_fp_trap(struct pt_regs *regs, void __user *location,
 		"floating point exception", regs, &si);
 }
 
-asmlinkage void illegal_op(struct pt_regs * regs, long interruption_code)
+static void illegal_op(struct pt_regs * regs, long interruption_code)
 {
 	siginfo_t info;
         __u8 opcode[6];
@@ -491,8 +490,15 @@ asmlinkage void illegal_op(struct pt_regs * regs, long interruption_code)
 #endif
 		} else
 			signal = SIGILL;
-	} else
-		signal = SIGILL;
+	} else {
+		/*
+		 * If we get an illegal op in kernel mode, send it through the
+		 * kprobes notifier. If kprobes doesn't pick it up, SIGILL
+		 */
+		if (notify_die(DIE_BPT, "bpt", regs, interruption_code,
+			       3, SIGTRAP) != NOTIFY_STOP)
+			signal = SIGILL;
+	}
 
 #ifdef CONFIG_MATHEMU
         if (signal == SIGFPE)
@@ -585,7 +591,7 @@ DO_ERROR_INFO(SIGILL, "specification exception", specification_exception,
 	      ILL_ILLOPN, get_check_address(regs));
 #endif
 
-asmlinkage void data_exception(struct pt_regs * regs, long interruption_code)
+static void data_exception(struct pt_regs * regs, long interruption_code)
 {
 	__u16 __user *location;
 	int signal = 0;
@@ -675,7 +681,7 @@ asmlinkage void data_exception(struct pt_regs * regs, long interruption_code)
 	}
 }
 
-asmlinkage void space_switch_exception(struct pt_regs * regs, long int_code)
+static void space_switch_exception(struct pt_regs * regs, long int_code)
 {
         siginfo_t info;
 
diff --git a/arch/s390/kernel/vmlinux.lds.S b/arch/s390/kernel/vmlinux.lds.S
index fe0f2e97ba7b..c30716ae130c 100644
--- a/arch/s390/kernel/vmlinux.lds.S
+++ b/arch/s390/kernel/vmlinux.lds.S
@@ -31,18 +31,19 @@ SECTIONS
 
   _etext = .;			/* End of text section */
 
-  . = ALIGN(16);		/* Exception table */
-  __start___ex_table = .;
-  __ex_table : { *(__ex_table) }
-  __stop___ex_table = .;
-
   RODATA
 
 #ifdef CONFIG_SHARED_KERNEL
   . = ALIGN(1048576);		/* VM shared segments are 1MB aligned */
+#endif
 
+  . = ALIGN(4096);
   _eshared = .;			/* End of shareable data */
-#endif
+
+  . = ALIGN(16);		/* Exception table */
+  __start___ex_table = .;
+  __ex_table : { *(__ex_table) }
+  __stop___ex_table = .;
 
   .data : {			/* Data */
 	*(.data)
@@ -90,11 +91,14 @@ SECTIONS
   .con_initcall.init : { *(.con_initcall.init) }
   __con_initcall_end = .;
   SECURITY_INIT
+
+#ifdef CONFIG_BLK_DEV_INITRD
   . = ALIGN(256);
   __initramfs_start = .;
   .init.ramfs : { *(.init.initramfs) }
   . = ALIGN(2);
   __initramfs_end = .;
+#endif
   . = ALIGN(256);
   __per_cpu_start = .;
   .data.percpu  : { *(.data.percpu) }
diff --git a/arch/s390/kernel/vtime.c b/arch/s390/kernel/vtime.c
index 21baaf5496d6..9d5b02801b46 100644
--- a/arch/s390/kernel/vtime.c
+++ b/arch/s390/kernel/vtime.c
@@ -25,7 +25,7 @@
 #include <asm/irq_regs.h>
 
 static ext_int_info_t ext_int_info_timer;
-DEFINE_PER_CPU(struct vtimer_queue, virt_cpu_timer);
+static DEFINE_PER_CPU(struct vtimer_queue, virt_cpu_timer);
 
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 /*
@@ -524,16 +524,15 @@ EXPORT_SYMBOL(del_virt_timer);
 void init_cpu_vtimer(void)
 {
 	struct vtimer_queue *vt_list;
-	unsigned long cr0;
 
 	/* kick the virtual timer */
 	S390_lowcore.exit_timer = VTIMER_MAX_SLICE;
 	S390_lowcore.last_update_timer = VTIMER_MAX_SLICE;
 	asm volatile ("SPT %0" : : "m" (S390_lowcore.last_update_timer));
 	asm volatile ("STCK %0" : "=m" (S390_lowcore.last_update_clock));
-	__ctl_store(cr0, 0, 0);
-	cr0 |= 0x400;
-	__ctl_load(cr0, 0, 0);
+
+	/* enable cpu timer interrupts */
+	__ctl_set_bit(0,10);
 
 	vt_list = &per_cpu(virt_cpu_timer, smp_processor_id());
 	INIT_LIST_HEAD(&vt_list->list);
@@ -572,6 +571,7 @@ void __init vtime_init(void)
 	if (register_idle_notifier(&vtimer_idle_nb))
 		panic("Couldn't register idle notifier");
 
+	/* Enable cpu timer interrupts on the boot cpu. */
 	init_cpu_vtimer();
 }