Merge branch 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6

* 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6: (72 commits)
  [S390] 3215/3270 console: remove wrong comment
  [S390] dasd: remove BKL from extended error reporting code
  [S390] vmlogrdr: remove BKL
  [S390] vmur: remove BKL
  [S390] zcrypt: remove BKL
  [S390] 3270: remove BKL
  [S390] vmwatchdog: remove lock_kernel() from open() function
  [S390] monwriter: remove lock_kernel() from open() function
  [S390] monreader: remove lock_kernel() from open() function
  [S390] s390: remove unused nfsd #includes
  [S390] ftrace: build ftrace.o when CONFIG_FTRACE_SYSCALLS is set for s390
  [S390] etr/stp: put correct per cpu variable
  [S390] tty3270: move keyboard compat ioctls
  [S390] sclp: improve servicability setting
  [S390] s390: use change recording override for kernel mapping
  [S390] MAINTAINERS: Add s390 drivers block
  [S390] use generic sockios.h header file
  [S390] use generic termbits.h header file
  [S390] smp: remove unused typedef and defines
  [S390] cmm: free pages on hibernate.
  ...

MAINTAINERS

@@ -3139,6 +3139,7 @@ S:	Supported
 F:	Documentation/s390/kvm.txt
 F:	arch/s390/include/asm/kvm*
 F:	arch/s390/kvm/
+F:	drivers/s390/kvm/
 
 KEXEC
 M:	Eric Biederman <ebiederm@xmission.com>
@@ -4553,6 +4554,7 @@ L:	linux-s390@vger.kernel.org
 W:	http://www.ibm.com/developerworks/linux/linux390/
 S:	Supported
 F:	arch/s390/
+F:	drivers/s390/
 
 S390 NETWORK DRIVERS
 M:	Ursula Braun <ursula.braun@de.ibm.com>
@@ -4568,6 +4570,7 @@ M:	Felix Beck <felix.beck@de.ibm.com>
 M:	Ralph Wuerthner <ralph.wuerthner@de.ibm.com>
 M:	linux390@de.ibm.com
 L:	linux-s390@vger.kernel.org
+W:	http://www.ibm.com/developerworks/linux/linux390/
 S:	Supported
 F:	drivers/s390/crypto/
 

arch/s390/Kconfig

@@ -220,23 +220,8 @@ config AUDIT_ARCH
 	bool
 	default y
 
-config S390_SWITCH_AMODE
-	bool "Switch kernel/user addressing modes"
-	help
-	  This option allows to switch the addressing modes of kernel and user
-	  space. The kernel parameter switch_amode=on will enable this feature,
-	  default is disabled. Enabling this (via kernel parameter) on machines
-	  earlier than IBM System z9-109 EC/BC will reduce system performance.
-
-	  Note that this option will also be selected by selecting the execute
-	  protection option below. Enabling the execute protection via the
-	  noexec kernel parameter will also switch the addressing modes,
-	  independent of the switch_amode kernel parameter.
-
-
 config S390_EXEC_PROTECT
 	bool "Data execute protection"
-	select S390_SWITCH_AMODE
 	help
 	  This option allows to enable a buffer overflow protection for user
 	  space programs and it also selects the addressing mode option above.

arch/s390/defconfig

@@ -185,7 +185,6 @@ CONFIG_HOTPLUG_CPU=y
 CONFIG_COMPAT=y
 CONFIG_SYSVIPC_COMPAT=y
 CONFIG_AUDIT_ARCH=y
-CONFIG_S390_SWITCH_AMODE=y
 CONFIG_S390_EXEC_PROTECT=y
 
 #

arch/s390/include/asm/atomic.h

@@ -21,7 +21,7 @@
 #if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
 
 #define __CS_LOOP(ptr, op_val, op_string) ({				\
-	typeof(ptr->counter) old_val, new_val;				\
+	int old_val, new_val;						\
 	asm volatile(							\
 		"	l	%0,%2\n"				\
 		"0:	lr	%1,%0\n"				\
@@ -38,7 +38,7 @@
 #else /* __GNUC__ */
 
 #define __CS_LOOP(ptr, op_val, op_string) ({				\
-	typeof(ptr->counter) old_val, new_val;				\
+	int old_val, new_val;						\
 	asm volatile(							\
 		"	l	%0,0(%3)\n"				\
 		"0:	lr	%1,%0\n"				\
@@ -143,7 +143,7 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u)
 #if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
 
 #define __CSG_LOOP(ptr, op_val, op_string) ({				\
-	typeof(ptr->counter) old_val, new_val;				\
+	long long old_val, new_val;					\
 	asm volatile(							\
 		"	lg	%0,%2\n"				\
 		"0:	lgr	%1,%0\n"				\
@@ -160,7 +160,7 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u)
 #else /* __GNUC__ */
 
 #define __CSG_LOOP(ptr, op_val, op_string) ({				\
-	typeof(ptr->counter) old_val, new_val;				\
+	long long old_val, new_val;					\
 	asm volatile(							\
 		"	lg	%0,0(%3)\n"				\
 		"0:	lgr	%1,%0\n"				\

arch/s390/include/asm/ccwdev.h

@@ -142,6 +142,8 @@ struct ccw1;
 extern int ccw_device_set_options_mask(struct ccw_device *, unsigned long);
 extern int ccw_device_set_options(struct ccw_device *, unsigned long);
 extern void ccw_device_clear_options(struct ccw_device *, unsigned long);
+int ccw_device_is_pathgroup(struct ccw_device *cdev);
+int ccw_device_is_multipath(struct ccw_device *cdev);
 
 /* Allow for i/o completion notification after primary interrupt status. */
 #define CCWDEV_EARLY_NOTIFICATION	0x0001
@@ -151,6 +153,8 @@ extern void ccw_device_clear_options(struct ccw_device *, unsigned long);
 #define CCWDEV_DO_PATHGROUP             0x0004
 /* Allow forced onlining of boxed devices. */
 #define CCWDEV_ALLOW_FORCE              0x0008
+/* Try to use multipath mode. */
+#define CCWDEV_DO_MULTIPATH		0x0010
 
 extern int ccw_device_start(struct ccw_device *, struct ccw1 *,
 			    unsigned long, __u8, unsigned long);

arch/s390/include/asm/mmu_context.h

@@ -36,7 +36,7 @@ static inline int init_new_context(struct task_struct *tsk,
 		mm->context.has_pgste = 1;
 		mm->context.alloc_pgste = 1;
 	} else {
-		mm->context.noexec = s390_noexec;
+		mm->context.noexec = (user_mode == SECONDARY_SPACE_MODE);
 		mm->context.has_pgste = 0;
 		mm->context.alloc_pgste = 0;
 	}
@@ -58,7 +58,7 @@ static inline void update_mm(struct mm_struct *mm, struct task_struct *tsk)
 	pgd_t *pgd = mm->pgd;
 
 	S390_lowcore.user_asce = mm->context.asce_bits | __pa(pgd);
-	if (switch_amode) {
+	if (user_mode != HOME_SPACE_MODE) {
 		/* Load primary space page table origin. */
 		pgd = mm->context.noexec ? get_shadow_table(pgd) : pgd;
 		S390_lowcore.user_exec_asce = mm->context.asce_bits | __pa(pgd);

arch/s390/include/asm/pgalloc.h

@@ -143,7 +143,8 @@ static inline pgd_t *pgd_alloc(struct mm_struct *mm)
 	spin_lock_init(&mm->context.list_lock);
 	INIT_LIST_HEAD(&mm->context.crst_list);
 	INIT_LIST_HEAD(&mm->context.pgtable_list);
-	return (pgd_t *) crst_table_alloc(mm, s390_noexec);
+	return (pgd_t *)
+		crst_table_alloc(mm, user_mode == SECONDARY_SPACE_MODE);
 }
 #define pgd_free(mm, pgd) crst_table_free(mm, (unsigned long *) pgd)
 

arch/s390/include/asm/pgtable.h

@@ -169,12 +169,13 @@ extern unsigned long VMALLOC_START;
  * STL Segment-Table-Length:  Segment-table length (STL+1*16 entries -> up to 2048)
  *
  * A 64 bit pagetable entry of S390 has following format:
- * |                     PFRA                         |0IP0|  OS  |
+ * |			 PFRA			      |0IPC|  OS  |
  * 0000000000111111111122222222223333333333444444444455555555556666
  * 0123456789012345678901234567890123456789012345678901234567890123
  *
  * I Page-Invalid Bit:    Page is not available for address-translation
  * P Page-Protection Bit: Store access not possible for page
+ * C Change-bit override: HW is not required to set change bit
  *
  * A 64 bit segmenttable entry of S390 has following format:
  * |        P-table origin                              |      TT
@@ -218,6 +219,7 @@ extern unsigned long VMALLOC_START;
  */
 
 /* Hardware bits in the page table entry */
+#define _PAGE_CO	0x100		/* HW Change-bit override */
 #define _PAGE_RO	0x200		/* HW read-only bit  */
 #define _PAGE_INVALID	0x400		/* HW invalid bit    */
 

arch/s390/include/asm/setup.h

@@ -49,17 +49,12 @@ extern unsigned long memory_end;
 
 void detect_memory_layout(struct mem_chunk chunk[]);
 
-#ifdef CONFIG_S390_SWITCH_AMODE
-extern unsigned int switch_amode;
-#else
-#define switch_amode	(0)
-#endif
-
-#ifdef CONFIG_S390_EXEC_PROTECT
-extern unsigned int s390_noexec;
-#else
-#define s390_noexec	(0)
-#endif
+#define PRIMARY_SPACE_MODE	0
+#define ACCESS_REGISTER_MODE	1
+#define SECONDARY_SPACE_MODE	2
+#define HOME_SPACE_MODE		3
+
+extern unsigned int user_mode;
 
 /*
  * Machine features detected in head.S

arch/s390/include/asm/smp.h

 /*
- *  include/asm-s390/smp.h
- *
- *  S390 version
- *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
- *    Author(s): Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
- *               Martin Schwidefsky (schwidefsky@de.ibm.com)
- *               Heiko Carstens (heiko.carstens@de.ibm.com)
+ *    Copyright IBM Corp. 1999,2009
+ *    Author(s): Denis Joseph Barrow,
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>,
+ *		 Heiko Carstens <heiko.carstens@de.ibm.com>,
  */
 #ifndef __ASM_SMP_H
 #define __ASM_SMP_H
 
-#include <linux/threads.h>
-#include <linux/cpumask.h>
-#include <linux/bitops.h>
+#ifdef CONFIG_SMP
 
-#if defined(__KERNEL__) && defined(CONFIG_SMP) && !defined(__ASSEMBLY__)
-
-#include <asm/lowcore.h>
-#include <asm/sigp.h>
-#include <asm/ptrace.h>
 #include <asm/system.h>
-
-/*
-  s390 specific smp.c headers
- */
-typedef struct
-{
-	int        intresting;
-	sigp_ccode ccode; 
-	__u32      status;
-	__u16      cpu;
-} sigp_info;
+#include <asm/sigp.h>
 
 extern void machine_restart_smp(char *);
 extern void machine_halt_smp(void);
 extern void machine_power_off_smp(void);
 
-#define NO_PROC_ID		0xFF		/* No processor magic marker */
-
-/*
- *	This magic constant controls our willingness to transfer
- *	a process across CPUs. Such a transfer incurs misses on the L1
- *	cache, and on a P6 or P5 with multiple L2 caches L2 hits. My
- *	gut feeling is this will vary by board in value. For a board
- *	with separate L2 cache it probably depends also on the RSS, and
- *	for a board with shared L2 cache it ought to decay fast as other
- *	processes are run.
- */
- 
-#define PROC_CHANGE_PENALTY	20		/* Schedule penalty */
-
 #define raw_smp_processor_id()	(S390_lowcore.cpu_nr)
-#define cpu_logical_map(cpu)	(cpu)
 
 extern int __cpu_disable (void);
 extern void __cpu_die (unsigned int cpu);
@@ -64,7 +29,9 @@ extern int smp_cpu_polarization[];
 extern void arch_send_call_function_single_ipi(int cpu);
 extern void arch_send_call_function_ipi_mask(const struct cpumask *mask);
 
-#endif
+extern union save_area *zfcpdump_save_areas[NR_CPUS + 1];
+
+#endif /* CONFIG_SMP */
 
 #ifdef CONFIG_HOTPLUG_CPU
 extern int smp_rescan_cpus(void);
@@ -72,5 +39,4 @@ extern int smp_rescan_cpus(void);
 static inline int smp_rescan_cpus(void) { return 0; }
 #endif
 
-extern union save_area *zfcpdump_save_areas[NR_CPUS + 1];
-#endif
+#endif /* __ASM_SMP_H */

arch/s390/include/asm/sockios.h

-/*
- *  include/asm-s390/sockios.h
- *
- *  S390 version
- *
- *  Derived from "include/asm-i386/sockios.h"
- */
+#ifndef _ASM_S390_SOCKIOS_H
+#define _ASM_S390_SOCKIOS_H
 
-#ifndef __ARCH_S390_SOCKIOS__
-#define __ARCH_S390_SOCKIOS__
-
-/* Socket-level I/O control calls. */
-#define FIOSETOWN 	0x8901
-#define SIOCSPGRP	0x8902
-#define FIOGETOWN	0x8903
-#define SIOCGPGRP	0x8904
-#define SIOCATMARK	0x8905
-#define SIOCGSTAMP	0x8906		/* Get stamp (timeval) */
-#define SIOCGSTAMPNS	0x8907		/* Get stamp (timespec) */
+#include <asm-generic/sockios.h>
 
 #endif

arch/s390/include/asm/termbits.h

-/*
- *  include/asm-s390/termbits.h
- *
- *  S390 version
- *
- *  Derived from "include/asm-i386/termbits.h"
- */
+#ifndef _ASM_S390_TERMBITS_H
+#define _ASM_S390_TERMBITS_H
 
-#ifndef __ARCH_S390_TERMBITS_H__
-#define __ARCH_S390_TERMBITS_H__
-
-#include <linux/posix_types.h>
-
-typedef unsigned char	cc_t;
-typedef unsigned int	speed_t;
-typedef unsigned int	tcflag_t;
-
-#define NCCS 19
-struct termios {
-	tcflag_t c_iflag;		/* input mode flags */
-	tcflag_t c_oflag;		/* output mode flags */
-	tcflag_t c_cflag;		/* control mode flags */
-	tcflag_t c_lflag;		/* local mode flags */
-	cc_t c_line;			/* line discipline */
-	cc_t c_cc[NCCS];		/* control characters */
-};
-
-struct termios2 {
-	tcflag_t c_iflag;		/* input mode flags */
-	tcflag_t c_oflag;		/* output mode flags */
-	tcflag_t c_cflag;		/* control mode flags */
-	tcflag_t c_lflag;		/* local mode flags */
-	cc_t c_line;			/* line discipline */
-	cc_t c_cc[NCCS];		/* control characters */
-	speed_t c_ispeed;		/* input speed */
-	speed_t c_ospeed;		/* output speed */
-};
-
-struct ktermios {
-	tcflag_t c_iflag;		/* input mode flags */
-	tcflag_t c_oflag;		/* output mode flags */
-	tcflag_t c_cflag;		/* control mode flags */
-	tcflag_t c_lflag;		/* local mode flags */
-	cc_t c_line;			/* line discipline */
-	cc_t c_cc[NCCS];		/* control characters */
-	speed_t c_ispeed;		/* input speed */
-	speed_t c_ospeed;		/* output speed */
-};
-
-/* c_cc characters */
-#define VINTR 0
-#define VQUIT 1
-#define VERASE 2
-#define VKILL 3
-#define VEOF 4
-#define VTIME 5
-#define VMIN 6
-#define VSWTC 7
-#define VSTART 8
-#define VSTOP 9
-#define VSUSP 10
-#define VEOL 11
-#define VREPRINT 12
-#define VDISCARD 13
-#define VWERASE 14
-#define VLNEXT 15
-#define VEOL2 16
-
-/* c_iflag bits */
-#define IGNBRK	0000001
-#define BRKINT	0000002
-#define IGNPAR	0000004
-#define PARMRK	0000010
-#define INPCK	0000020
-#define ISTRIP	0000040
-#define INLCR	0000100
-#define IGNCR	0000200
-#define ICRNL	0000400
-#define IUCLC	0001000
-#define IXON	0002000
-#define IXANY	0004000
-#define IXOFF	0010000
-#define IMAXBEL	0020000
-#define IUTF8	0040000
-
-/* c_oflag bits */
-#define OPOST	0000001
-#define OLCUC	0000002
-#define ONLCR	0000004
-#define OCRNL	0000010
-#define ONOCR	0000020
-#define ONLRET	0000040
-#define OFILL	0000100
-#define OFDEL	0000200
-#define NLDLY	0000400
-#define   NL0	0000000
-#define   NL1	0000400
-#define CRDLY	0003000
-#define   CR0	0000000
-#define   CR1	0001000
-#define   CR2	0002000
-#define   CR3	0003000
-#define TABDLY	0014000
-#define   TAB0	0000000
-#define   TAB1	0004000
-#define   TAB2	0010000
-#define   TAB3	0014000
-#define   XTABS	0014000
-#define BSDLY	0020000
-#define   BS0	0000000
-#define   BS1	0020000
-#define VTDLY	0040000
-#define   VT0	0000000
-#define   VT1	0040000
-#define FFDLY	0100000
-#define   FF0	0000000
-#define   FF1	0100000
-
-/* c_cflag bit meaning */
-#define CBAUD	0010017
-#define  B0	0000000		/* hang up */
-#define  B50	0000001
-#define  B75	0000002
-#define  B110	0000003
-#define  B134	0000004
-#define  B150	0000005
-#define  B200	0000006
-#define  B300	0000007
-#define  B600	0000010
-#define  B1200	0000011
-#define  B1800	0000012
-#define  B2400	0000013
-#define  B4800	0000014
-#define  B9600	0000015
-#define  B19200	0000016
-#define  B38400	0000017
-#define EXTA B19200
-#define EXTB B38400
-#define CSIZE	0000060
-#define   CS5	0000000
-#define   CS6	0000020
-#define   CS7	0000040
-#define   CS8	0000060
-#define CSTOPB	0000100
-#define CREAD	0000200
-#define PARENB	0000400
-#define PARODD	0001000
-#define HUPCL	0002000
-#define CLOCAL	0004000
-#define CBAUDEX 0010000
-#define  BOTHER  0010000
-#define  B57600  0010001
-#define  B115200 0010002
-#define  B230400 0010003
-#define  B460800 0010004
-#define   B500000 0010005
-#define   B576000 0010006
-#define   B921600 0010007
-#define  B1000000 0010010
-#define  B1152000 0010011
-#define  B1500000 0010012
-#define  B2000000 0010013
-#define  B2500000 0010014
-#define  B3000000 0010015
-#define  B3500000 0010016
-#define  B4000000 0010017
-#define CIBAUD	  002003600000	/* input baud rate */
-#define CMSPAR	  010000000000		/* mark or space (stick) parity */
-#define CRTSCTS	  020000000000		/* flow control */
-
-#define IBSHIFT	  16		/* Shift from CBAUD to CIBAUD */
-
-/* c_lflag bits */
-#define ISIG	0000001
-#define ICANON	0000002
-#define XCASE	0000004
-#define ECHO	0000010
-#define ECHOE	0000020
-#define ECHOK	0000040
-#define ECHONL	0000100
-#define NOFLSH	0000200
-#define TOSTOP	0000400
-#define ECHOCTL	0001000
-#define ECHOPRT	0002000
-#define ECHOKE	0004000
-#define FLUSHO	0010000
-#define PENDIN	0040000
-#define IEXTEN	0100000
-
-/* tcflow() and TCXONC use these */
-#define	TCOOFF		0
-#define	TCOON		1
-#define	TCIOFF		2
-#define	TCION		3
-
-/* tcflush() and TCFLSH use these */
-#define	TCIFLUSH	0
-#define	TCOFLUSH	1
-#define	TCIOFLUSH	2
-
-/* tcsetattr uses these */
-#define	TCSANOW		0
-#define	TCSADRAIN	1
-#define	TCSAFLUSH	2
+#include <asm-generic/termbits.h>
 
 #endif

arch/s390/include/asm/todclk.h

-/*
- * File...........: linux/include/asm/todclk.h
- * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
- * Bugreports.to..: <Linux390@de.ibm.com>
- * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
- *
- * History of changes (starts July 2000)
- */
-
-#ifndef __ASM_TODCLK_H
-#define __ASM_TODCLK_H
-
-#ifdef __KERNEL__
-
-#define TOD_uSEC (0x1000ULL)
-#define TOD_mSEC (1000 * TOD_uSEC)
-#define TOD_SEC (1000 * TOD_mSEC)
-#define TOD_MIN (60 * TOD_SEC)
-#define TOD_HOUR (60 * TOD_MIN)
-
-#endif
-
-#endif

arch/s390/include/asm/uaccess.h

@@ -93,6 +93,8 @@ extern struct uaccess_ops uaccess_mvcos;
 extern struct uaccess_ops uaccess_mvcos_switch;
 extern struct uaccess_ops uaccess_pt;
 
+extern int __handle_fault(unsigned long, unsigned long, int);
+
 static inline int __put_user_fn(size_t size, void __user *ptr, void *x)
 {
 	size = uaccess.copy_to_user_small(size, ptr, x);

arch/s390/kernel/Makefile

@@ -44,6 +44,7 @@ obj-$(CONFIG_KPROBES)		+= kprobes.o
 obj-$(CONFIG_FUNCTION_TRACER)	+= $(if $(CONFIG_64BIT),mcount64.o,mcount.o)
 obj-$(CONFIG_DYNAMIC_FTRACE)	+= ftrace.o
 obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o
+obj-$(CONFIG_FTRACE_SYSCALLS)  += ftrace.o
 
 # Kexec part
 S390_KEXEC_OBJS := machine_kexec.o crash.o

arch/s390/kernel/compat_linux.c

@@ -31,14 +31,8 @@
 #include <linux/shm.h>
 #include <linux/slab.h>
 #include <linux/uio.h>
-#include <linux/nfs_fs.h>
 #include <linux/quota.h>
 #include <linux/module.h>
-#include <linux/sunrpc/svc.h>
-#include <linux/nfsd/nfsd.h>
-#include <linux/nfsd/cache.h>
-#include <linux/nfsd/xdr.h>
-#include <linux/nfsd/syscall.h>
 #include <linux/poll.h>
 #include <linux/personality.h>
 #include <linux/stat.h>

arch/s390/kernel/compat_linux.h

@@ -4,10 +4,6 @@
 #include <linux/compat.h>
 #include <linux/socket.h>
 #include <linux/syscalls.h>
-#include <linux/nfs_fs.h>
-#include <linux/sunrpc/svc.h>
-#include <linux/nfsd/nfsd.h>
-#include <linux/nfsd/export.h>
 
 /* Macro that masks the high order bit of an 32 bit pointer and converts it*/
 /*       to a 64 bit pointer */

arch/s390/kernel/head64.S

@@ -83,6 +83,8 @@ startup_continue:
 	slr	%r0,%r0 		# set cpuid to zero
 	sigp	%r1,%r0,0x12		# switch to esame mode
 	sam64				# switch to 64 bit mode
+	llgfr	%r13,%r13		# clear high-order half of base reg
+	lmh	%r0,%r15,.Lzero64-.LPG1(%r13)	# clear high-order half
 	lctlg	%c0,%c15,.Lctl-.LPG1(%r13)	# load control registers
 	lg	%r12,.Lparmaddr-.LPG1(%r13)	# pointer to parameter area
 					# move IPL device to lowcore
@@ -127,6 +129,7 @@ startup_continue:
 .L4malign:.quad 0xffffffffffc00000
 .Lscan2g:.quad	0x80000000 + 0x20000 - 8	# 2GB + 128K - 8
 .Lnop:	.long	0x07000700
+.Lzero64:.fill	16,4,0x0
 #ifdef CONFIG_ZFCPDUMP
 .Lcurrent_cpu:
 	.long 0x0

arch/s390/kernel/setup.c

@@ -305,9 +305,8 @@ static int __init early_parse_mem(char *p)
 }
 early_param("mem", early_parse_mem);
 
-#ifdef CONFIG_S390_SWITCH_AMODE
-unsigned int switch_amode = 0;
-EXPORT_SYMBOL_GPL(switch_amode);
+unsigned int user_mode = HOME_SPACE_MODE;
+EXPORT_SYMBOL_GPL(user_mode);
 
 static int set_amode_and_uaccess(unsigned long user_amode,
 				 unsigned long user32_amode)
@@ -340,23 +339,29 @@ static int set_amode_and_uaccess(unsigned long user_amode,
  */
 static int __init early_parse_switch_amode(char *p)
 {
-	switch_amode = 1;
+	if (user_mode != SECONDARY_SPACE_MODE)
+		user_mode = PRIMARY_SPACE_MODE;
 	return 0;
 }
 early_param("switch_amode", early_parse_switch_amode);
 
-#else /* CONFIG_S390_SWITCH_AMODE */
-static inline int set_amode_and_uaccess(unsigned long user_amode,
-					unsigned long user32_amode)
+static int __init early_parse_user_mode(char *p)
 {
+	if (p && strcmp(p, "primary") == 0)
+		user_mode = PRIMARY_SPACE_MODE;
+#ifdef CONFIG_S390_EXEC_PROTECT
+	else if (p && strcmp(p, "secondary") == 0)
+		user_mode = SECONDARY_SPACE_MODE;
+#endif
+	else if (!p || strcmp(p, "home") == 0)
+		user_mode = HOME_SPACE_MODE;
+	else
+		return 1;
 	return 0;
 }
-#endif /* CONFIG_S390_SWITCH_AMODE */
+early_param("user_mode", early_parse_user_mode);
 
 #ifdef CONFIG_S390_EXEC_PROTECT
-unsigned int s390_noexec = 0;
-EXPORT_SYMBOL_GPL(s390_noexec);
-
 /*
  * Enable execute protection?
  */
@@ -364,8 +369,7 @@ static int __init early_parse_noexec(char *p)
 {
 	if (!strncmp(p, "off", 3))
 		return 0;
-	switch_amode = 1;
-	s390_noexec = 1;
+	user_mode = SECONDARY_SPACE_MODE;
 	return 0;
 }
 early_param("noexec", early_parse_noexec);
@@ -373,7 +377,7 @@ early_param("noexec", early_parse_noexec);
 
 static void setup_addressing_mode(void)
 {
-	if (s390_noexec) {
+	if (user_mode == SECONDARY_SPACE_MODE) {
 		if (set_amode_and_uaccess(PSW_ASC_SECONDARY,
 					  PSW32_ASC_SECONDARY))
 			pr_info("Execute protection active, "
@@ -381,7 +385,7 @@ static void setup_addressing_mode(void)
 		else
 			pr_info("Execute protection active, "
 				"mvcos not available\n");
-	} else if (switch_amode) {
+	} else if (user_mode == PRIMARY_SPACE_MODE) {
 		if (set_amode_and_uaccess(PSW_ASC_PRIMARY, PSW32_ASC_PRIMARY))
 			pr_info("Address spaces switched, "
 				"mvcos available\n");
@@ -411,7 +415,7 @@ 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;
-	if (switch_amode)
+	if (user_mode != HOME_SPACE_MODE)
 		lc->restart_psw.mask |= PSW_ASC_HOME;
 	lc->external_new_psw.mask = psw_kernel_bits;
 	lc->external_new_psw.addr =

arch/s390/kernel/time.c

@@ -335,7 +335,7 @@ int get_sync_clock(unsigned long long *clock)
 	sw0 = atomic_read(sw_ptr);
 	*clock = get_clock();
 	sw1 = atomic_read(sw_ptr);
-	put_cpu_var(clock_sync_sync);
+	put_cpu_var(clock_sync_word);
 	if (sw0 == sw1 && (sw0 & 0x80000000U))
 		/* Success: time is in sync. */
 		return 0;
@@ -385,7 +385,7 @@ static inline int check_sync_clock(void)
 
 	sw_ptr = &get_cpu_var(clock_sync_word);
 	rc = (atomic_read(sw_ptr) & 0x80000000U) != 0;
-	put_cpu_var(clock_sync_sync);
+	put_cpu_var(clock_sync_word);
 	return rc;
 }
 

arch/s390/kernel/vdso.c

@@ -86,7 +86,8 @@ static void vdso_init_data(struct vdso_data *vd)
 	unsigned int facility_list;
 
 	facility_list = stfl();
-	vd->ectg_available = switch_amode && (facility_list & 1);
+	vd->ectg_available =
+		user_mode != HOME_SPACE_MODE && (facility_list & 1);
 }
 
 #ifdef CONFIG_64BIT
@@ -114,7 +115,7 @@ int vdso_alloc_per_cpu(int cpu, struct _lowcore *lowcore)
 
 	lowcore->vdso_per_cpu_data = __LC_PASTE;
 
-	if (!switch_amode || !vdso_enabled)
+	if (user_mode == HOME_SPACE_MODE || !vdso_enabled)
 		return 0;
 
 	segment_table = __get_free_pages(GFP_KERNEL, SEGMENT_ORDER);
@@ -160,7 +161,7 @@ void vdso_free_per_cpu(int cpu, struct _lowcore *lowcore)
 	unsigned long segment_table, page_table, page_frame;
 	u32 *psal, *aste;
 
-	if (!switch_amode || !vdso_enabled)
+	if (user_mode == HOME_SPACE_MODE || !vdso_enabled)
 		return;
 
 	psal = (u32 *)(addr_t) lowcore->paste[4];
@@ -184,7 +185,7 @@ static void __vdso_init_cr5(void *dummy)
 
 static void vdso_init_cr5(void)
 {
-	if (switch_amode && vdso_enabled)
+	if (user_mode != HOME_SPACE_MODE && vdso_enabled)
 		on_each_cpu(__vdso_init_cr5, NULL, 1);
 }
 #endif /* CONFIG_64BIT */

arch/s390/kvm/Kconfig

@@ -20,7 +20,6 @@ config KVM
 	depends on HAVE_KVM && EXPERIMENTAL
 	select PREEMPT_NOTIFIERS
 	select ANON_INODES
-	select S390_SWITCH_AMODE
 	---help---
 	  Support hosting paravirtualized guest machines using the SIE
 	  virtualization capability on the mainframe. This should work

arch/s390/lib/uaccess_mvcos.c

@@ -162,7 +162,6 @@ static size_t clear_user_mvcos(size_t size, void __user *to)
 	return size;
 }
 
-#ifdef CONFIG_S390_SWITCH_AMODE
 static size_t strnlen_user_mvcos(size_t count, const char __user *src)
 {
 	char buf[256];
@@ -200,7 +199,6 @@ static size_t strncpy_from_user_mvcos(size_t count, const char __user *src,
 	} while ((len_str == len) && (done < count));
 	return done;
 }
-#endif /* CONFIG_S390_SWITCH_AMODE */
 
 struct uaccess_ops uaccess_mvcos = {
 	.copy_from_user = copy_from_user_mvcos_check,
@@ -215,7 +213,6 @@ struct uaccess_ops uaccess_mvcos = {
 	.futex_atomic_cmpxchg = futex_atomic_cmpxchg_std,
 };
 
-#ifdef CONFIG_S390_SWITCH_AMODE
 struct uaccess_ops uaccess_mvcos_switch = {
 	.copy_from_user = copy_from_user_mvcos,
 	.copy_from_user_small = copy_from_user_mvcos,
@@ -228,4 +225,3 @@ struct uaccess_ops uaccess_mvcos_switch = {
 	.futex_atomic_op = futex_atomic_op_pt,
 	.futex_atomic_cmpxchg = futex_atomic_cmpxchg_pt,
 };
-#endif

arch/s390/lib/uaccess_pt.c

@@ -23,85 +23,20 @@ static inline pte_t *follow_table(struct mm_struct *mm, unsigned long addr)
 
 	pgd = pgd_offset(mm, addr);
 	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
-		return NULL;
+		return (pte_t *) 0x3a;
 
 	pud = pud_offset(pgd, addr);
 	if (pud_none(*pud) || unlikely(pud_bad(*pud)))
-		return NULL;
+		return (pte_t *) 0x3b;
 
 	pmd = pmd_offset(pud, addr);
 	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
-		return NULL;
+		return (pte_t *) 0x10;
 
 	return pte_offset_map(pmd, addr);
 }
 
-static int __handle_fault(struct mm_struct *mm, unsigned long address,
-			  int write_access)
-{
-	struct vm_area_struct *vma;
-	int ret = -EFAULT;
-	int fault;
-
-	if (in_atomic())
-		return ret;
-	down_read(&mm->mmap_sem);
-	vma = find_vma(mm, address);
-	if (unlikely(!vma))
-		goto out;
-	if (unlikely(vma->vm_start > address)) {
-		if (!(vma->vm_flags & VM_GROWSDOWN))
-			goto out;
-		if (expand_stack(vma, address))
-			goto out;
-	}
-
-	if (!write_access) {
-		/* page not present, check vm flags */
-		if (!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)))
-			goto out;
-	} else {
-		if (!(vma->vm_flags & VM_WRITE))
-			goto out;
-	}
-
-survive:
-	fault = handle_mm_fault(mm, vma, address, write_access ? FAULT_FLAG_WRITE : 0);
-	if (unlikely(fault & VM_FAULT_ERROR)) {
-		if (fault & VM_FAULT_OOM)
-			goto out_of_memory;
-		else if (fault & VM_FAULT_SIGBUS)
-			goto out_sigbus;
-		BUG();
-	}
-	if (fault & VM_FAULT_MAJOR)
-		current->maj_flt++;
-	else
-		current->min_flt++;
-	ret = 0;
-out:
-	up_read(&mm->mmap_sem);
-	return ret;
-
-out_of_memory:
-	up_read(&mm->mmap_sem);
-	if (is_global_init(current)) {
-		yield();
-		down_read(&mm->mmap_sem);
-		goto survive;
-	}
-	printk("VM: killing process %s\n", current->comm);
-	return ret;
-
-out_sigbus:
-	up_read(&mm->mmap_sem);
-	current->thread.prot_addr = address;
-	current->thread.trap_no = 0x11;
-	force_sig(SIGBUS, current);
-	return ret;
-}
-
-static size_t __user_copy_pt(unsigned long uaddr, void *kptr,
+static __always_inline size_t __user_copy_pt(unsigned long uaddr, void *kptr,
 					     size_t n, int write_user)
 {
 	struct mm_struct *mm = current->mm;
@@ -114,12 +49,17 @@ static size_t __user_copy_pt(unsigned long uaddr, void *kptr,
 	spin_lock(&mm->page_table_lock);
 	do {
 		pte = follow_table(mm, uaddr);
-		if (!pte || !pte_present(*pte) ||
-		    (write_user && !pte_write(*pte)))
+		if ((unsigned long) pte < 0x1000)
+			goto fault;
+		if (!pte_present(*pte)) {
+			pte = (pte_t *) 0x11;
+			goto fault;
+		} else if (write_user && !pte_write(*pte)) {
+			pte = (pte_t *) 0x04;
 			goto fault;
+		}
 
 		pfn = pte_pfn(*pte);
-
 		offset = uaddr & (PAGE_SIZE - 1);
 		size = min(n - done, PAGE_SIZE - offset);
 		if (write_user) {
@@ -137,7 +77,7 @@ static size_t __user_copy_pt(unsigned long uaddr, void *kptr,
 	return n - done;
 fault:
 	spin_unlock(&mm->page_table_lock);
-	if (__handle_fault(mm, uaddr, write_user))
+	if (__handle_fault(uaddr, (unsigned long) pte, write_user))
 		return n - done;
 	goto retry;
 }
@@ -146,30 +86,31 @@ static size_t __user_copy_pt(unsigned long uaddr, void *kptr,
  * Do DAT for user address by page table walk, return kernel address.
  * This function needs to be called with current->mm->page_table_lock held.
  */
-static unsigned long __dat_user_addr(unsigned long uaddr)
+static __always_inline unsigned long __dat_user_addr(unsigned long uaddr)
 {
 	struct mm_struct *mm = current->mm;
-	unsigned long pfn, ret;
+	unsigned long pfn;
 	pte_t *pte;
 	int rc;
 
-	ret = 0;
 retry:
 	pte = follow_table(mm, uaddr);
-	if (!pte || !pte_present(*pte))
+	if ((unsigned long) pte < 0x1000)
+		goto fault;
+	if (!pte_present(*pte)) {
+		pte = (pte_t *) 0x11;
 		goto fault;
+	}
 
 	pfn = pte_pfn(*pte);
-	ret = (pfn << PAGE_SHIFT) + (uaddr & (PAGE_SIZE - 1));
-out:
-	return ret;
+	return (pfn << PAGE_SHIFT) + (uaddr & (PAGE_SIZE - 1));
 fault:
 	spin_unlock(&mm->page_table_lock);
-	rc = __handle_fault(mm, uaddr, 0);
+	rc = __handle_fault(uaddr, (unsigned long) pte, 0);
 	spin_lock(&mm->page_table_lock);
-	if (rc)
-		goto out;
+	if (!rc)
 		goto retry;
+	return 0;
 }
 
 size_t copy_from_user_pt(size_t n, const void __user *from, void *to)
@@ -234,8 +175,12 @@ static size_t strnlen_user_pt(size_t count, const char __user *src)
 	spin_lock(&mm->page_table_lock);
 	do {
 		pte = follow_table(mm, uaddr);
-		if (!pte || !pte_present(*pte))
+		if ((unsigned long) pte < 0x1000)
+			goto fault;
+		if (!pte_present(*pte)) {
+			pte = (pte_t *) 0x11;
 			goto fault;
+		}
 
 		pfn = pte_pfn(*pte);
 		offset = uaddr & (PAGE_SIZE-1);
@@ -249,9 +194,8 @@ static size_t strnlen_user_pt(size_t count, const char __user *src)
 	return done + 1;
 fault:
 	spin_unlock(&mm->page_table_lock);
-	if (__handle_fault(mm, uaddr, 0)) {
+	if (__handle_fault(uaddr, (unsigned long) pte, 0))
 		return 0;
-	}
 	goto retry;
 }
 
@@ -284,7 +228,7 @@ static size_t copy_in_user_pt(size_t n, void __user *to,
 {
 	struct mm_struct *mm = current->mm;
 	unsigned long offset_from, offset_to, offset_max, pfn_from, pfn_to,
-		      uaddr, done, size;
+		      uaddr, done, size, error_code;
 	unsigned long uaddr_from = (unsigned long) from;
 	unsigned long uaddr_to = (unsigned long) to;
 	pte_t *pte_from, *pte_to;
@@ -298,17 +242,28 @@ static size_t copy_in_user_pt(size_t n, void __user *to,
 retry:
 	spin_lock(&mm->page_table_lock);
 	do {
-		pte_from = follow_table(mm, uaddr_from);
-		if (!pte_from || !pte_present(*pte_from)) {
-			uaddr = uaddr_from;
 		write_user = 0;
+		uaddr = uaddr_from;
+		pte_from = follow_table(mm, uaddr_from);
+		error_code = (unsigned long) pte_from;
+		if (error_code < 0x1000)
+			goto fault;
+		if (!pte_present(*pte_from)) {
+			error_code = 0x11;
 			goto fault;
 		}
 
-		pte_to = follow_table(mm, uaddr_to);
-		if (!pte_to || !pte_present(*pte_to) || !pte_write(*pte_to)) {
-			uaddr = uaddr_to;
 		write_user = 1;
+		uaddr = uaddr_to;
+		pte_to = follow_table(mm, uaddr_to);
+		error_code = (unsigned long) pte_to;
+		if (error_code < 0x1000)
+			goto fault;
+		if (!pte_present(*pte_to)) {
+			error_code = 0x11;
+			goto fault;
+		} else if (!pte_write(*pte_to)) {
+			error_code = 0x04;
 			goto fault;
 		}
 
@@ -329,7 +284,7 @@ static size_t copy_in_user_pt(size_t n, void __user *to,
 	return n - done;
 fault:
 	spin_unlock(&mm->page_table_lock);
-	if (__handle_fault(mm, uaddr, write_user))
+	if (__handle_fault(uaddr, error_code, write_user))
 		return n - done;
 	goto retry;
 }

arch/s390/mm/cmm.c

@@ -18,6 +18,7 @@
 #include <linux/swap.h>
 #include <linux/kthread.h>
 #include <linux/oom.h>
+#include <linux/suspend.h>
 
 #include <asm/pgalloc.h>
 #include <asm/uaccess.h>
@@ -44,6 +45,7 @@ static volatile long cmm_pages_target;
 static volatile long cmm_timed_pages_target;
 static long cmm_timeout_pages;
 static long cmm_timeout_seconds;
+static int cmm_suspended;
 
 static struct cmm_page_array *cmm_page_list;
 static struct cmm_page_array *cmm_timed_page_list;
@@ -147,9 +149,9 @@ cmm_thread(void *dummy)
 
 	while (1) {
 		rc = wait_event_interruptible(cmm_thread_wait,
-			(cmm_pages != cmm_pages_target ||
-			 cmm_timed_pages != cmm_timed_pages_target ||
-			 kthread_should_stop()));
+			(!cmm_suspended && (cmm_pages != cmm_pages_target ||
+			 cmm_timed_pages != cmm_timed_pages_target)) ||
+			 kthread_should_stop());
 		if (kthread_should_stop() || rc == -ERESTARTSYS) {
 			cmm_pages_target = cmm_pages;
 			cmm_timed_pages_target = cmm_timed_pages;
@@ -410,6 +412,38 @@ cmm_smsg_target(char *from, char *msg)
 
 static struct ctl_table_header *cmm_sysctl_header;
 
+static int cmm_suspend(void)
+{
+	cmm_suspended = 1;
+	cmm_free_pages(cmm_pages, &cmm_pages, &cmm_page_list);
+	cmm_free_pages(cmm_timed_pages, &cmm_timed_pages, &cmm_timed_page_list);
+	return 0;
+}
+
+static int cmm_resume(void)
+{
+	cmm_suspended = 0;
+	cmm_kick_thread();
+	return 0;
+}
+
+static int cmm_power_event(struct notifier_block *this,
+			   unsigned long event, void *ptr)
+{
+	switch (event) {
+	case PM_POST_HIBERNATION:
+		return cmm_resume();
+	case PM_HIBERNATION_PREPARE:
+		return cmm_suspend();
+	default:
+		return NOTIFY_DONE;
+	}
+}
+
+static struct notifier_block cmm_power_notifier = {
+	.notifier_call = cmm_power_event,
+};
+
 static int
 cmm_init (void)
 {
@@ -418,7 +452,7 @@ cmm_init (void)
 #ifdef CONFIG_CMM_PROC
 	cmm_sysctl_header = register_sysctl_table(cmm_dir_table);
 	if (!cmm_sysctl_header)
-		goto out;
+		goto out_sysctl;
 #endif
 #ifdef CONFIG_CMM_IUCV
 	rc = smsg_register_callback(SMSG_PREFIX, cmm_smsg_target);
@@ -428,17 +462,21 @@ cmm_init (void)
 	rc = register_oom_notifier(&cmm_oom_nb);
 	if (rc < 0)
 		goto out_oom_notify;
+	rc = register_pm_notifier(&cmm_power_notifier);
+	if (rc)
+		goto out_pm;
 	init_waitqueue_head(&cmm_thread_wait);
 	init_timer(&cmm_timer);
 	cmm_thread_ptr = kthread_run(cmm_thread, NULL, "cmmthread");
 	rc = IS_ERR(cmm_thread_ptr) ? PTR_ERR(cmm_thread_ptr) : 0;
-	if (!rc)
-		goto out;
-	/*
-	 * kthread_create failed. undo all the stuff from above again.
-	 */
-	unregister_oom_notifier(&cmm_oom_nb);
+	if (rc)
+		goto out_kthread;
+	return 0;
 
+out_kthread:
+	unregister_pm_notifier(&cmm_power_notifier);
+out_pm:
+	unregister_oom_notifier(&cmm_oom_nb);
 out_oom_notify:
 #ifdef CONFIG_CMM_IUCV
 	smsg_unregister_callback(SMSG_PREFIX, cmm_smsg_target);
@@ -446,8 +484,8 @@ cmm_init (void)
 #endif
 #ifdef CONFIG_CMM_PROC
 	unregister_sysctl_table(cmm_sysctl_header);
+out_sysctl:
 #endif
-out:
 	return rc;
 }
 
@@ -455,6 +493,7 @@ static void
 cmm_exit(void)
 {
 	kthread_stop(cmm_thread_ptr);
+	unregister_pm_notifier(&cmm_power_notifier);
 	unregister_oom_notifier(&cmm_oom_nb);
 	cmm_free_pages(cmm_pages, &cmm_pages, &cmm_page_list);
 	cmm_free_pages(cmm_timed_pages, &cmm_timed_pages, &cmm_timed_page_list);

arch/s390/mm/pgtable.c

@@ -269,7 +269,7 @@ int s390_enable_sie(void)
 	struct mm_struct *mm, *old_mm;
 
 	/* Do we have switched amode? If no, we cannot do sie */
-	if (!switch_amode)
+	if (user_mode == HOME_SPACE_MODE)
 		return -EINVAL;
 
 	/* Do we have pgstes? if yes, we are done */

arch/s390/mm/vmem.c

@@ -70,6 +70,10 @@ static pte_t __ref *vmem_pte_alloc(void)
 		pte = alloc_bootmem(PTRS_PER_PTE * sizeof(pte_t));
 	if (!pte)
 		return NULL;
+	if (MACHINE_HAS_HPAGE)
+		clear_table((unsigned long *) pte, _PAGE_TYPE_EMPTY | _PAGE_CO,
+			    PTRS_PER_PTE * sizeof(pte_t));
+	else
 		clear_table((unsigned long *) pte, _PAGE_TYPE_EMPTY,
 			    PTRS_PER_PTE * sizeof(pte_t));
 	return pte;
@@ -112,7 +116,8 @@ static int vmem_add_mem(unsigned long start, unsigned long size, int ro)
 		if (MACHINE_HAS_HPAGE && !(address & ~HPAGE_MASK) &&
 		    (address + HPAGE_SIZE <= start + size) &&
 		    (address >= HPAGE_SIZE)) {
-			pte_val(pte) |= _SEGMENT_ENTRY_LARGE;
+			pte_val(pte) |= _SEGMENT_ENTRY_LARGE |
+					_SEGMENT_ENTRY_CO;
 			pmd_val(*pm_dir) = pte_val(pte);
 			address += HPAGE_SIZE - PAGE_SIZE;
 			continue;

drivers/s390/block/dasd_3990_erp.c

@@ -12,7 +12,6 @@
 #include <linux/timer.h>
 #include <linux/slab.h>
 #include <asm/idals.h>
-#include <asm/todclk.h>
 
 #define PRINTK_HEADER "dasd_erp(3990): "
 
@@ -70,8 +69,7 @@ dasd_3990_erp_cleanup(struct dasd_ccw_req * erp, char final_status)
  *   processing until the started timer has expired or an related
  *   interrupt was received.
  */
-static void
-dasd_3990_erp_block_queue(struct dasd_ccw_req * erp, int expires)
+static void dasd_3990_erp_block_queue(struct dasd_ccw_req *erp, int expires)
 {
 
 	struct dasd_device *device = erp->startdev;
@@ -81,10 +79,13 @@ dasd_3990_erp_block_queue(struct dasd_ccw_req * erp, int expires)
 		    "blocking request queue for %is", expires/HZ);
 
 	spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
-	device->stopped |= DASD_STOPPED_PENDING;
+	dasd_device_set_stop_bits(device, DASD_STOPPED_PENDING);
 	spin_unlock_irqrestore(get_ccwdev_lock(device->cdev), flags);
 	erp->status = DASD_CQR_FILLED;
-	dasd_block_set_timer(device->block, expires);
+	if (erp->block)
+		dasd_block_set_timer(erp->block, expires);
+	else
+		dasd_device_set_timer(device, expires);
 }
 
 /*
@@ -243,9 +244,13 @@ dasd_3990_erp_DCTL(struct dasd_ccw_req * erp, char modifier)
  * DESCRIPTION
  *   Setup ERP to do the ERP action 1 (see Reference manual).
  *   Repeat the operation on a different channel path.
- *   If all alternate paths have been tried, the request is posted with a
- *   permanent error.
- *   Note: duplex handling is not implemented (yet).
+ *   As deviation from the recommended recovery action, we reset the path mask
+ *   after we have tried each path and go through all paths a second time.
+ *   This will cover situations where only one path at a time is actually down,
+ *   but all paths fail and recover just with the same sequence and timing as
+ *   we try to use them (flapping links).
+ *   If all alternate paths have been tried twice, the request is posted with
+ *   a permanent error.
  *
  *  PARAMETER
  *   erp		pointer to the current ERP
@@ -254,17 +259,25 @@ dasd_3990_erp_DCTL(struct dasd_ccw_req * erp, char modifier)
  *   erp		pointer to the ERP
  *
  */
-static struct dasd_ccw_req *
-dasd_3990_erp_action_1(struct dasd_ccw_req * erp)
+static struct dasd_ccw_req *dasd_3990_erp_action_1_sec(struct dasd_ccw_req *erp)
 {
+	erp->function = dasd_3990_erp_action_1_sec;
+	dasd_3990_erp_alternate_path(erp);
+	return erp;
+}
 
+static struct dasd_ccw_req *dasd_3990_erp_action_1(struct dasd_ccw_req *erp)
+{
 	erp->function = dasd_3990_erp_action_1;
-
 	dasd_3990_erp_alternate_path(erp);
-
+	if (erp->status == DASD_CQR_FAILED) {
+		erp->status = DASD_CQR_FILLED;
+		erp->retries = 10;
+		erp->lpm = LPM_ANYPATH;
+		erp->function = dasd_3990_erp_action_1_sec;
+	}
 	return erp;
-
-}				/* end dasd_3990_erp_action_1 */
+}				/* end dasd_3990_erp_action_1(b) */
 
 /*
  * DASD_3990_ERP_ACTION_4
@@ -2295,6 +2308,7 @@ static struct dasd_ccw_req *dasd_3990_erp_add_erp(struct dasd_ccw_req *cqr)
 		return cqr;
 	}
 
+	ccw = cqr->cpaddr;
 	if (cqr->cpmode == 1) {
 		/* make a shallow copy of the original tcw but set new tsb */
 		erp->cpmode = 1;
@@ -2303,6 +2317,9 @@ static struct dasd_ccw_req *dasd_3990_erp_add_erp(struct dasd_ccw_req *cqr)
 		tsb = (struct tsb *) &tcw[1];
 		*tcw = *((struct tcw *)cqr->cpaddr);
 		tcw->tsb = (long)tsb;
+	} else if (ccw->cmd_code == DASD_ECKD_CCW_PSF) {
+		/* PSF cannot be chained from NOOP/TIC */
+		erp->cpaddr = cqr->cpaddr;
 	} else {
 		/* initialize request with default TIC to current ERP/CQR */
 		ccw = erp->cpaddr;
@@ -2487,6 +2504,8 @@ dasd_3990_erp_further_erp(struct dasd_ccw_req *erp)
 
 		erp = dasd_3990_erp_action_1(erp);
 
+	} else if (erp->function == dasd_3990_erp_action_1_sec) {
+		erp = dasd_3990_erp_action_1_sec(erp);
 	} else if (erp->function == dasd_3990_erp_action_5) {
 
 		/* retries have not been successful */

drivers/s390/block/dasd_alias.c

@@ -152,6 +152,7 @@ static struct alias_lcu *_allocate_lcu(struct dasd_uid *uid)
 	INIT_WORK(&lcu->suc_data.worker, summary_unit_check_handling_work);
 	INIT_DELAYED_WORK(&lcu->ruac_data.dwork, lcu_update_work);
 	spin_lock_init(&lcu->lock);
+	init_completion(&lcu->lcu_setup);
 	return lcu;
 
 out_err4:
@@ -239,6 +240,67 @@ int dasd_alias_make_device_known_to_lcu(struct dasd_device *device)
 	return is_lcu_known;
 }
 
+/*
+ * The first device to be registered on an LCU will have to do
+ * some additional setup steps to configure that LCU on the
+ * storage server. All further devices should wait with their
+ * initialization until the first device is done.
+ * To synchronize this work, the first device will call
+ * dasd_alias_lcu_setup_complete when it is done, and all
+ * other devices will wait for it with dasd_alias_wait_for_lcu_setup.
+ */
+void dasd_alias_lcu_setup_complete(struct dasd_device *device)
+{
+	struct dasd_eckd_private *private;
+	unsigned long flags;
+	struct alias_server *server;
+	struct alias_lcu *lcu;
+	struct dasd_uid *uid;
+
+	private = (struct dasd_eckd_private *) device->private;
+	uid = &private->uid;
+	lcu = NULL;
+	spin_lock_irqsave(&aliastree.lock, flags);
+	server = _find_server(uid);
+	if (server)
+		lcu = _find_lcu(server, uid);
+	spin_unlock_irqrestore(&aliastree.lock, flags);
+	if (!lcu) {
+		DBF_EVENT_DEVID(DBF_ERR, device->cdev,
+				"could not find lcu for %04x %02x",
+				uid->ssid, uid->real_unit_addr);
+		WARN_ON(1);
+		return;
+	}
+	complete_all(&lcu->lcu_setup);
+}
+
+void dasd_alias_wait_for_lcu_setup(struct dasd_device *device)
+{
+	struct dasd_eckd_private *private;
+	unsigned long flags;
+	struct alias_server *server;
+	struct alias_lcu *lcu;
+	struct dasd_uid *uid;
+
+	private = (struct dasd_eckd_private *) device->private;
+	uid = &private->uid;
+	lcu = NULL;
+	spin_lock_irqsave(&aliastree.lock, flags);
+	server = _find_server(uid);
+	if (server)
+		lcu = _find_lcu(server, uid);
+	spin_unlock_irqrestore(&aliastree.lock, flags);
+	if (!lcu) {
+		DBF_EVENT_DEVID(DBF_ERR, device->cdev,
+				"could not find lcu for %04x %02x",
+				uid->ssid, uid->real_unit_addr);
+		WARN_ON(1);
+		return;
+	}
+	wait_for_completion(&lcu->lcu_setup);
+}
+
 /*
  * This function removes a device from the scope of alias management.
  * The complicated part is to make sure that it is not in use by
@@ -755,11 +817,11 @@ static void __stop_device_on_lcu(struct dasd_device *device,
 {
 	/* If pos == device then device is already locked! */
 	if (pos == device) {
-		pos->stopped |= DASD_STOPPED_SU;
+		dasd_device_set_stop_bits(pos, DASD_STOPPED_SU);
 		return;
 	}
 	spin_lock(get_ccwdev_lock(pos->cdev));
-	pos->stopped |= DASD_STOPPED_SU;
+	dasd_device_set_stop_bits(pos, DASD_STOPPED_SU);
 	spin_unlock(get_ccwdev_lock(pos->cdev));
 }
 
@@ -793,26 +855,26 @@ static void _unstop_all_devices_on_lcu(struct alias_lcu *lcu)
 
 	list_for_each_entry(device, &lcu->active_devices, alias_list) {
 		spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
-		device->stopped &= ~DASD_STOPPED_SU;
+		dasd_device_remove_stop_bits(device, DASD_STOPPED_SU);
 		spin_unlock_irqrestore(get_ccwdev_lock(device->cdev), flags);
 	}
 
 	list_for_each_entry(device, &lcu->inactive_devices, alias_list) {
 		spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
-		device->stopped &= ~DASD_STOPPED_SU;
+		dasd_device_remove_stop_bits(device, DASD_STOPPED_SU);
 		spin_unlock_irqrestore(get_ccwdev_lock(device->cdev), flags);
 	}
 
 	list_for_each_entry(pavgroup, &lcu->grouplist, group) {
 		list_for_each_entry(device, &pavgroup->baselist, alias_list) {
 			spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
-			device->stopped &= ~DASD_STOPPED_SU;
+			dasd_device_remove_stop_bits(device, DASD_STOPPED_SU);
 			spin_unlock_irqrestore(get_ccwdev_lock(device->cdev),
 					       flags);
 		}
 		list_for_each_entry(device, &pavgroup->aliaslist, alias_list) {
 			spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
-			device->stopped &= ~DASD_STOPPED_SU;
+			dasd_device_remove_stop_bits(device, DASD_STOPPED_SU);
 			spin_unlock_irqrestore(get_ccwdev_lock(device->cdev),
 					       flags);
 		}
@@ -836,7 +898,8 @@ static void summary_unit_check_handling_work(struct work_struct *work)
 
 	/* 2. reset summary unit check */
 	spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
-	device->stopped &= ~(DASD_STOPPED_SU | DASD_STOPPED_PENDING);
+	dasd_device_remove_stop_bits(device,
+				     (DASD_STOPPED_SU | DASD_STOPPED_PENDING));
 	spin_unlock_irqrestore(get_ccwdev_lock(device->cdev), flags);
 	reset_summary_unit_check(lcu, device, suc_data->reason);
 

drivers/s390/block/dasd_diag.c

@@ -24,7 +24,6 @@
 #include <asm/ebcdic.h>
 #include <asm/io.h>
 #include <asm/s390_ext.h>
-#include <asm/todclk.h>
 #include <asm/vtoc.h>
 #include <asm/diag.h>
 
@@ -145,6 +144,15 @@ dasd_diag_erp(struct dasd_device *device)
 
 	mdsk_term_io(device);
 	rc = mdsk_init_io(device, device->block->bp_block, 0, NULL);
+	if (rc == 4) {
+		if (!(device->features & DASD_FEATURE_READONLY)) {
+			dev_warn(&device->cdev->dev,
+				 "The access mode of a DIAG device changed"
+				 " to read-only");
+			device->features |= DASD_FEATURE_READONLY;
+		}
+		rc = 0;
+	}
 	if (rc)
 		dev_warn(&device->cdev->dev, "DIAG ERP failed with "
 			    "rc=%d\n", rc);
@@ -433,16 +441,20 @@ dasd_diag_check_device(struct dasd_device *device)
 	for (sb = 512; sb < bsize; sb = sb << 1)
 		block->s2b_shift++;
 	rc = mdsk_init_io(device, block->bp_block, 0, NULL);
-	if (rc) {
+	if (rc && (rc != 4)) {
 		dev_warn(&device->cdev->dev, "DIAG initialization "
 			"failed with rc=%d\n", rc);
 		rc = -EIO;
 	} else {
+		if (rc == 4)
+			device->features |= DASD_FEATURE_READONLY;
 		dev_info(&device->cdev->dev,
-			 "New DASD with %ld byte/block, total size %ld KB\n",
+			 "New DASD with %ld byte/block, total size %ld KB%s\n",
 			 (unsigned long) block->bp_block,
 			 (unsigned long) (block->blocks <<
-					  block->s2b_shift) >> 1);
+					  block->s2b_shift) >> 1,
+			 (rc == 4) ? ", read-only device" : "");
+		rc = 0;
 	}
 out_label:
 	free_page((long) label);

drivers/s390/block/dasd_eckd.h

@@ -414,6 +414,7 @@ struct alias_lcu {
 	struct summary_unit_check_work_data suc_data;
 	struct read_uac_work_data ruac_data;
 	struct dasd_ccw_req *rsu_cqr;
+	struct completion lcu_setup;
 };
 
 struct alias_pav_group {
@@ -460,5 +461,6 @@ int dasd_alias_remove_device(struct dasd_device *);
 struct dasd_device *dasd_alias_get_start_dev(struct dasd_device *);
 void dasd_alias_handle_summary_unit_check(struct dasd_device *, struct irb *);
 void dasd_eckd_reset_ccw_to_base_io(struct dasd_ccw_req *);
-
+void dasd_alias_lcu_setup_complete(struct dasd_device *);
+void dasd_alias_wait_for_lcu_setup(struct dasd_device *);
 #endif				/* DASD_ECKD_H */

drivers/s390/block/dasd_eer.c

@@ -536,7 +536,6 @@ static int dasd_eer_open(struct inode *inp, struct file *filp)
 	eerb = kzalloc(sizeof(struct eerbuffer), GFP_KERNEL);
 	if (!eerb)
 		return -ENOMEM;
-	lock_kernel();
 	eerb->buffer_page_count = eer_pages;
 	if (eerb->buffer_page_count < 1 ||
 	    eerb->buffer_page_count > INT_MAX / PAGE_SIZE) {
@@ -544,7 +543,6 @@ static int dasd_eer_open(struct inode *inp, struct file *filp)
 		DBF_EVENT(DBF_WARNING, "can't open device since module "
 			"parameter eer_pages is smaller than 1 or"
 			" bigger than %d", (int)(INT_MAX / PAGE_SIZE));
-		unlock_kernel();
 		return -EINVAL;
 	}
 	eerb->buffersize = eerb->buffer_page_count * PAGE_SIZE;
@@ -552,14 +550,12 @@ static int dasd_eer_open(struct inode *inp, struct file *filp)
 			       GFP_KERNEL);
         if (!eerb->buffer) {
 		kfree(eerb);
-		unlock_kernel();
                 return -ENOMEM;
 	}
 	if (dasd_eer_allocate_buffer_pages(eerb->buffer,
 					   eerb->buffer_page_count)) {
 		kfree(eerb->buffer);
 		kfree(eerb);
-		unlock_kernel();
 		return -ENOMEM;
 	}
 	filp->private_data = eerb;
@@ -567,7 +563,6 @@ static int dasd_eer_open(struct inode *inp, struct file *filp)
 	list_add(&eerb->list, &bufferlist);
 	spin_unlock_irqrestore(&bufferlock, flags);
 
-	unlock_kernel();
 	return nonseekable_open(inp,filp);
 }
 

drivers/s390/block/dasd_fba.c

@@ -20,7 +20,6 @@
 #include <asm/idals.h>
 #include <asm/ebcdic.h>
 #include <asm/io.h>
-#include <asm/todclk.h>
 #include <asm/ccwdev.h>
 
 #include "dasd_int.h"
@@ -141,9 +140,8 @@ dasd_fba_check_characteristics(struct dasd_device *device)
 	}
 	block = dasd_alloc_block();
 	if (IS_ERR(block)) {
-		DBF_EVENT(DBF_WARNING, "could not allocate dasd block "
-			  "structure for device: %s",
-			  dev_name(&device->cdev->dev));
+		DBF_EVENT_DEVID(DBF_WARNING, cdev, "%s", "could not allocate "
+				"dasd block structure");
 		device->private = NULL;
 		kfree(private);
 		return PTR_ERR(block);
@@ -155,9 +153,8 @@ dasd_fba_check_characteristics(struct dasd_device *device)
 	rc = dasd_generic_read_dev_chars(device, DASD_FBA_MAGIC,
 					 &private->rdc_data, 32);
 	if (rc) {
-		DBF_EVENT(DBF_WARNING, "Read device characteristics returned "
-			  "error %d for device: %s",
-			  rc, dev_name(&device->cdev->dev));
+		DBF_EVENT_DEVID(DBF_WARNING, cdev, "Read device "
+				"characteristics returned error %d", rc);
 		device->block = NULL;
 		dasd_free_block(block);
 		device->private = NULL;

drivers/s390/block/dasd_int.h

@@ -108,6 +108,16 @@ do { \
 			    d_data); \
 } while(0)
 
+#define DBF_EVENT_DEVID(d_level, d_cdev, d_str, d_data...)	\
+do { \
+	struct ccw_dev_id __dev_id;			\
+	ccw_device_get_id(d_cdev, &__dev_id);		\
+	debug_sprintf_event(dasd_debug_area,		\
+			    d_level,					\
+			    "0.%x.%04x " d_str "\n",			\
+			    __dev_id.ssid, __dev_id.devno, d_data);	\
+} while (0)
+
 #define DBF_EXC(d_level, d_str, d_data...)\
 do { \
 	debug_sprintf_exception(dasd_debug_area, \
@@ -595,6 +605,9 @@ int dasd_generic_restore_device(struct ccw_device *);
 int dasd_generic_read_dev_chars(struct dasd_device *, int, void *, int);
 char *dasd_get_sense(struct irb *);
 
+void dasd_device_set_stop_bits(struct dasd_device *, int);
+void dasd_device_remove_stop_bits(struct dasd_device *, int);
+
 /* externals in dasd_devmap.c */
 extern int dasd_max_devindex;
 extern int dasd_probeonly;

drivers/s390/block/dasd_ioctl.c

@@ -101,7 +101,7 @@ static int dasd_ioctl_quiesce(struct dasd_block *block)
 	pr_info("%s: The DASD has been put in the quiesce "
 		"state\n", dev_name(&base->cdev->dev));
 	spin_lock_irqsave(get_ccwdev_lock(base->cdev), flags);
-	base->stopped |= DASD_STOPPED_QUIESCE;
+	dasd_device_set_stop_bits(base, DASD_STOPPED_QUIESCE);
 	spin_unlock_irqrestore(get_ccwdev_lock(base->cdev), flags);
 	return 0;
 }
@@ -122,7 +122,7 @@ static int dasd_ioctl_resume(struct dasd_block *block)
 	pr_info("%s: I/O operations have been resumed "
 		"on the DASD\n", dev_name(&base->cdev->dev));
 	spin_lock_irqsave(get_ccwdev_lock(base->cdev), flags);
-	base->stopped &= ~DASD_STOPPED_QUIESCE;
+	dasd_device_remove_stop_bits(base, DASD_STOPPED_QUIESCE);
 	spin_unlock_irqrestore(get_ccwdev_lock(base->cdev), flags);
 
 	dasd_schedule_block_bh(block);

drivers/s390/char/con3215.c

@@ -857,7 +857,6 @@ static struct console con3215 = {
 
 /*
  * 3215 console initialization code called from console_init().
- * NOTE: This is called before kmalloc is available.
  */
 static int __init con3215_init(void)
 {

drivers/s390/char/con3270.c

@@ -572,7 +572,6 @@ static struct console con3270 = {
 
 /*
  * 3270 console initialization code called from console_init().
- * NOTE: This is called before kmalloc is available.
  */
 static int __init
 con3270_init(void)

drivers/s390/char/fs3270.c

@@ -38,6 +38,8 @@ struct fs3270 {
 	size_t rdbuf_size;		/* size of data returned by RDBUF */
 };
 
+static DEFINE_MUTEX(fs3270_mutex);
+
 static void
 fs3270_wake_up(struct raw3270_request *rq, void *data)
 {
@@ -328,7 +330,7 @@ fs3270_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 	if (!fp)
 		return -ENODEV;
 	rc = 0;
-	lock_kernel();
+	mutex_lock(&fs3270_mutex);
 	switch (cmd) {
 	case TUBICMD:
 		fp->read_command = arg;
@@ -354,7 +356,7 @@ fs3270_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			rc = -EFAULT;
 		break;
 	}
-	unlock_kernel();
+	mutex_unlock(&fs3270_mutex);
 	return rc;
 }
 
@@ -437,7 +439,7 @@ fs3270_open(struct inode *inode, struct file *filp)
 		minor = tty->index + RAW3270_FIRSTMINOR;
 		tty_kref_put(tty);
 	}
-	lock_kernel();
+	mutex_lock(&fs3270_mutex);
 	/* Check if some other program is already using fullscreen mode. */
 	fp = (struct fs3270 *) raw3270_find_view(&fs3270_fn, minor);
 	if (!IS_ERR(fp)) {
@@ -478,7 +480,7 @@ fs3270_open(struct inode *inode, struct file *filp)
 	}
 	filp->private_data = fp;
 out:
-	unlock_kernel();
+	mutex_unlock(&fs3270_mutex);
 	return rc;
 }
 

drivers/s390/char/monreader.c

@@ -12,7 +12,6 @@
 #include <linux/module.h>
 #include <linux/moduleparam.h>
 #include <linux/init.h>
-#include <linux/smp_lock.h>
 #include <linux/errno.h>
 #include <linux/types.h>
 #include <linux/kernel.h>
@@ -283,7 +282,6 @@ static int mon_open(struct inode *inode, struct file *filp)
 	/*
 	 * only one user allowed
 	 */
-	lock_kernel();
 	rc = -EBUSY;
 	if (test_and_set_bit(MON_IN_USE, &mon_in_use))
 		goto out;
@@ -321,7 +319,6 @@ static int mon_open(struct inode *inode, struct file *filp)
 	}
 	filp->private_data = monpriv;
 	dev_set_drvdata(monreader_device, monpriv);
-	unlock_kernel();
 	return nonseekable_open(inode, filp);
 
 out_path:
@@ -331,7 +328,6 @@ static int mon_open(struct inode *inode, struct file *filp)
 out_use:
 	clear_bit(MON_IN_USE, &mon_in_use);
 out:
-	unlock_kernel();
 	return rc;
 }
 
@@ -607,6 +603,10 @@ static int __init mon_init(void)
 	}
 	dcss_mkname(mon_dcss_name, &user_data_connect[8]);
 
+	/*
+	 * misc_register() has to be the last action in module_init(), because
+	 * file operations will be available right after this.
+	 */
 	rc = misc_register(&mon_dev);
 	if (rc < 0 )
 		goto out;

drivers/s390/char/monwriter.c

@@ -13,7 +13,6 @@
 #include <linux/moduleparam.h>
 #include <linux/init.h>
 #include <linux/errno.h>
-#include <linux/smp_lock.h>
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/miscdevice.h>
@@ -185,13 +184,11 @@ static int monwrite_open(struct inode *inode, struct file *filp)
 	monpriv = kzalloc(sizeof(struct mon_private), GFP_KERNEL);
 	if (!monpriv)
 		return -ENOMEM;
-	lock_kernel();
 	INIT_LIST_HEAD(&monpriv->list);
 	monpriv->hdr_to_read = sizeof(monpriv->hdr);
 	mutex_init(&monpriv->thread_mutex);
 	filp->private_data = monpriv;
 	list_add_tail(&monpriv->priv_list, &mon_priv_list);
-	unlock_kernel();
 	return nonseekable_open(inode, filp);
 }
 
@@ -364,6 +361,10 @@ static int __init mon_init(void)
 		goto out_driver;
 	}
 
+	/*
+	 * misc_register() has to be the last action in module_init(), because
+	 * file operations will be available right after this.
+	 */
 	rc = misc_register(&mon_dev);
 	if (rc)
 		goto out_device;

drivers/s390/char/sclp_cmd.c

@@ -84,6 +84,7 @@ static void __init sclp_read_info_early(void)
 		do {
 			memset(sccb, 0, sizeof(*sccb));
 			sccb->header.length = sizeof(*sccb);
+			sccb->header.function_code = 0x80;
 			sccb->header.control_mask[2] = 0x80;
 			rc = sclp_cmd_sync_early(commands[i], sccb);
 		} while (rc == -EBUSY);

drivers/s390/char/tape.h

@@ -212,6 +212,9 @@ struct tape_device {
 	struct tape_class_device *	nt;
 	struct tape_class_device *	rt;
 
+	/* Device mutex to serialize tape commands. */
+	struct mutex			mutex;
+
 	/* Device discipline information. */
 	struct tape_discipline *	discipline;
 	void *				discdata;
@@ -292,9 +295,9 @@ extern int tape_generic_pm_suspend(struct ccw_device *);
 extern int tape_generic_probe(struct ccw_device *);
 extern void tape_generic_remove(struct ccw_device *);
 
-extern struct tape_device *tape_get_device(int devindex);
-extern struct tape_device *tape_get_device_reference(struct tape_device *);
-extern struct tape_device *tape_put_device(struct tape_device *);
+extern struct tape_device *tape_find_device(int devindex);
+extern struct tape_device *tape_get_device(struct tape_device *);
+extern void tape_put_device(struct tape_device *);
 
 /* Externals from tape_char.c */
 extern int tapechar_init(void);

drivers/s390/char/tape_34xx.c

@@ -113,16 +113,16 @@ tape_34xx_work_handler(struct work_struct *work)
 {
 	struct tape_34xx_work *p =
 		container_of(work, struct tape_34xx_work, work);
+	struct tape_device *device = p->device;
 
 	switch(p->op) {
 		case TO_MSEN:
-			tape_34xx_medium_sense(p->device);
+			tape_34xx_medium_sense(device);
 			break;
 		default:
 			DBF_EVENT(3, "T34XX: internal error: unknown work\n");
 	}
-
-	p->device = tape_put_device(p->device);
+	tape_put_device(device);
 	kfree(p);
 }
 
@@ -136,7 +136,7 @@ tape_34xx_schedule_work(struct tape_device *device, enum tape_op op)
 
 	INIT_WORK(&p->work, tape_34xx_work_handler);
 
-	p->device = tape_get_device_reference(device);
+	p->device = tape_get_device(device);
 	p->op     = op;
 
 	schedule_work(&p->work);

drivers/s390/char/tape_3590.c

@@ -608,7 +608,7 @@ tape_3590_schedule_work(struct tape_device *device, enum tape_op op)
 
 	INIT_WORK(&p->work, tape_3590_work_handler);
 
-	p->device = tape_get_device_reference(device);
+	p->device = tape_get_device(device);
 	p->op = op;
 
 	schedule_work(&p->work);

drivers/s390/char/tape_block.c

@@ -54,7 +54,7 @@ static const struct block_device_operations tapeblock_fops = {
 	.owner		 = THIS_MODULE,
 	.open		 = tapeblock_open,
 	.release	 = tapeblock_release,
-	.locked_ioctl           = tapeblock_ioctl,
+	.ioctl		 = tapeblock_ioctl,
 	.media_changed   = tapeblock_medium_changed,
 	.revalidate_disk = tapeblock_revalidate_disk,
 };
@@ -239,7 +239,7 @@ tapeblock_setup_device(struct tape_device * device)
 	disk->major = tapeblock_major;
 	disk->first_minor = device->first_minor;
 	disk->fops = &tapeblock_fops;
-	disk->private_data = tape_get_device_reference(device);
+	disk->private_data = tape_get_device(device);
 	disk->queue = blkdat->request_queue;
 	set_capacity(disk, 0);
 	sprintf(disk->disk_name, "btibm%d",
@@ -247,11 +247,11 @@ tapeblock_setup_device(struct tape_device * device)
 
 	blkdat->disk = disk;
 	blkdat->medium_changed = 1;
-	blkdat->request_queue->queuedata = tape_get_device_reference(device);
+	blkdat->request_queue->queuedata = tape_get_device(device);
 
 	add_disk(disk);
 
-	tape_get_device_reference(device);
+	tape_get_device(device);
 	INIT_WORK(&blkdat->requeue_task, tapeblock_requeue);
 
 	return 0;
@@ -274,13 +274,14 @@ tapeblock_cleanup_device(struct tape_device *device)
 	}
 
 	del_gendisk(device->blk_data.disk);
-	device->blk_data.disk->private_data =
-		tape_put_device(device->blk_data.disk->private_data);
+	device->blk_data.disk->private_data = NULL;
+	tape_put_device(device);
 	put_disk(device->blk_data.disk);
 
 	device->blk_data.disk = NULL;
 cleanup_queue:
-	device->blk_data.request_queue->queuedata = tape_put_device(device);
+	device->blk_data.request_queue->queuedata = NULL;
+	tape_put_device(device);
 
 	blk_cleanup_queue(device->blk_data.request_queue);
 	device->blk_data.request_queue = NULL;
@@ -363,7 +364,7 @@ tapeblock_open(struct block_device *bdev, fmode_t mode)
 	struct tape_device *	device;
 	int			rc;
 
-	device = tape_get_device_reference(disk->private_data);
+	device = tape_get_device(disk->private_data);
 
 	if (device->required_tapemarks) {
 		DBF_EVENT(2, "TBLOCK: missing tapemarks\n");

drivers/s390/char/tape_char.c

@@ -33,8 +33,7 @@ static ssize_t tapechar_read(struct file *, char __user *, size_t, loff_t *);
 static ssize_t tapechar_write(struct file *, const char __user *, size_t, loff_t *);
 static int tapechar_open(struct inode *,struct file *);
 static int tapechar_release(struct inode *,struct file *);
-static int tapechar_ioctl(struct inode *, struct file *, unsigned int,
-			  unsigned long);
+static long tapechar_ioctl(struct file *, unsigned int, unsigned long);
 static long tapechar_compat_ioctl(struct file *, unsigned int,
 			  unsigned long);
 
@@ -43,7 +42,7 @@ static const struct file_operations tape_fops =
 	.owner = THIS_MODULE,
 	.read = tapechar_read,
 	.write = tapechar_write,
-	.ioctl = tapechar_ioctl,
+	.unlocked_ioctl = tapechar_ioctl,
 	.compat_ioctl = tapechar_compat_ioctl,
 	.open = tapechar_open,
 	.release = tapechar_release,
@@ -170,7 +169,6 @@ tapechar_read(struct file *filp, char __user *data, size_t count, loff_t *ppos)
 	if (rc == 0) {
 		rc = block_size - request->rescnt;
 		DBF_EVENT(6, "TCHAR:rbytes:  %x\n", rc);
-		filp->f_pos += rc;
 		/* Copy data from idal buffer to user space. */
 		if (idal_buffer_to_user(device->char_data.idal_buf,
 					data, rc) != 0)
@@ -238,7 +236,6 @@ tapechar_write(struct file *filp, const char __user *data, size_t count, loff_t
 			break;
 		DBF_EVENT(6, "TCHAR:wbytes: %lx\n",
 			  block_size - request->rescnt);
-		filp->f_pos += block_size - request->rescnt;
 		written += block_size - request->rescnt;
 		if (request->rescnt != 0)
 			break;
@@ -286,26 +283,20 @@ tapechar_open (struct inode *inode, struct file *filp)
 	if (imajor(filp->f_path.dentry->d_inode) != tapechar_major)
 		return -ENODEV;
 
-	lock_kernel();
 	minor = iminor(filp->f_path.dentry->d_inode);
-	device = tape_get_device(minor / TAPE_MINORS_PER_DEV);
+	device = tape_find_device(minor / TAPE_MINORS_PER_DEV);
 	if (IS_ERR(device)) {
-		DBF_EVENT(3, "TCHAR:open: tape_get_device() failed\n");
-		rc = PTR_ERR(device);
-		goto out;
+		DBF_EVENT(3, "TCHAR:open: tape_find_device() failed\n");
+		return PTR_ERR(device);
 	}
 
-
 	rc = tape_open(device);
 	if (rc == 0) {
 		filp->private_data = device;
-		rc = nonseekable_open(inode, filp);
-	}
-	else
+		nonseekable_open(inode, filp);
+	} else
 		tape_put_device(device);
 
-out:
-	unlock_kernel();
 	return rc;
 }
 
@@ -342,7 +333,8 @@ tapechar_release(struct inode *inode, struct file *filp)
 		device->char_data.idal_buf = NULL;
 	}
 	tape_release(device);
-	filp->private_data = tape_put_device(device);
+	filp->private_data = NULL;
+	tape_put_device(device);
 
 	return 0;
 }
@@ -351,16 +343,11 @@ tapechar_release(struct inode *inode, struct file *filp)
  * Tape device io controls.
  */
 static int
-tapechar_ioctl(struct inode *inp, struct file *filp,
+__tapechar_ioctl(struct tape_device *device,
 		 unsigned int no, unsigned long data)
 {
-	struct tape_device *device;
 	int rc;
 
-	DBF_EVENT(6, "TCHAR:ioct\n");
-
-	device = (struct tape_device *) filp->private_data;
-
 	if (no == MTIOCTOP) {
 		struct mtop op;
 
@@ -452,6 +439,21 @@ tapechar_ioctl(struct inode *inp, struct file *filp,
 	return device->discipline->ioctl_fn(device, no, data);
 }
 
+static long
+tapechar_ioctl(struct file *filp, unsigned int no, unsigned long data)
+{
+	struct tape_device *device;
+	long rc;
+
+	DBF_EVENT(6, "TCHAR:ioct\n");
+
+	device = (struct tape_device *) filp->private_data;
+	mutex_lock(&device->mutex);
+	rc = __tapechar_ioctl(device, no, data);
+	mutex_unlock(&device->mutex);
+	return rc;
+}
+
 static long
 tapechar_compat_ioctl(struct file *filp, unsigned int no, unsigned long data)
 {
@@ -459,9 +461,9 @@ tapechar_compat_ioctl(struct file *filp, unsigned int no, unsigned long data)
 	int rval = -ENOIOCTLCMD;
 
 	if (device->discipline->ioctl_fn) {
-		lock_kernel();
+		mutex_lock(&device->mutex);
 		rval = device->discipline->ioctl_fn(device, no, data);
-		unlock_kernel();
+		mutex_unlock(&device->mutex);
 		if (rval == -EINVAL)
 			rval = -ENOIOCTLCMD;
 	}

drivers/s390/char/tape_core.c

@@ -492,6 +492,7 @@ tape_alloc_device(void)
 		kfree(device);
 		return ERR_PTR(-ENOMEM);
 	}
+	mutex_init(&device->mutex);
 	INIT_LIST_HEAD(&device->req_queue);
 	INIT_LIST_HEAD(&device->node);
 	init_waitqueue_head(&device->state_change_wq);
@@ -511,11 +512,12 @@ tape_alloc_device(void)
  * increment the reference count.
  */
 struct tape_device *
-tape_get_device_reference(struct tape_device *device)
+tape_get_device(struct tape_device *device)
 {
-	DBF_EVENT(4, "tape_get_device_reference(%p) = %i\n", device,
-		atomic_inc_return(&device->ref_count));
+	int count;
 
+	count = atomic_inc_return(&device->ref_count);
+	DBF_EVENT(4, "tape_get_device(%p) = %i\n", device, count);
 	return device;
 }
 
@@ -525,32 +527,25 @@ tape_get_device_reference(struct tape_device *device)
  * The function returns a NULL pointer to be used by the caller
  * for clearing reference pointers.
  */
-struct tape_device *
+void
 tape_put_device(struct tape_device *device)
 {
-	int remain;
+	int count;
 
-	remain = atomic_dec_return(&device->ref_count);
-	if (remain > 0) {
-		DBF_EVENT(4, "tape_put_device(%p) -> %i\n", device, remain);
-	} else {
-		if (remain < 0) {
-			DBF_EVENT(4, "put device without reference\n");
-		} else {
-			DBF_EVENT(4, "tape_free_device(%p)\n", device);
+	count = atomic_dec_return(&device->ref_count);
+	DBF_EVENT(4, "tape_put_device(%p) -> %i\n", device, count);
+	BUG_ON(count < 0);
+	if (count == 0) {
 		kfree(device->modeset_byte);
 		kfree(device);
 	}
-	}
-
-	return NULL;			
 }
 
 /*
  * Find tape device by a device index.
  */
 struct tape_device *
-tape_get_device(int devindex)
+tape_find_device(int devindex)
 {
 	struct tape_device *device, *tmp;
 
@@ -558,7 +553,7 @@ tape_get_device(int devindex)
 	read_lock(&tape_device_lock);
 	list_for_each_entry(tmp, &tape_device_list, node) {
 		if (tmp->first_minor / TAPE_MINORS_PER_DEV == devindex) {
-			device = tape_get_device_reference(tmp);
+			device = tape_get_device(tmp);
 			break;
 		}
 	}
@@ -579,7 +574,8 @@ tape_generic_probe(struct ccw_device *cdev)
 	device = tape_alloc_device();
 	if (IS_ERR(device))
 		return -ENODEV;
-	ccw_device_set_options(cdev, CCWDEV_DO_PATHGROUP);
+	ccw_device_set_options(cdev, CCWDEV_DO_PATHGROUP |
+				     CCWDEV_DO_MULTIPATH);
 	ret = sysfs_create_group(&cdev->dev.kobj, &tape_attr_group);
 	if (ret) {
 		tape_put_device(device);
@@ -606,7 +602,8 @@ __tape_discard_requests(struct tape_device *device)
 		list_del(&request->list);
 
 		/* Decrease ref_count for removed request. */
-		request->device = tape_put_device(device);
+		request->device = NULL;
+		tape_put_device(device);
 		request->rc = -EIO;
 		if (request->callback != NULL)
 			request->callback(request, request->callback_data);
@@ -664,9 +661,11 @@ tape_generic_remove(struct ccw_device *cdev)
 			tape_cleanup_device(device);
 	}
 
-	if (!dev_get_drvdata(&cdev->dev)) {
+	device = dev_get_drvdata(&cdev->dev);
+	if (device) {
 		sysfs_remove_group(&cdev->dev.kobj, &tape_attr_group);
-		dev_set_drvdata(&cdev->dev, tape_put_device(dev_get_drvdata(&cdev->dev)));
+		dev_set_drvdata(&cdev->dev, NULL);
+		tape_put_device(device);
 	}
 }
 
@@ -721,9 +720,8 @@ tape_free_request (struct tape_request * request)
 {
 	DBF_LH(6, "Free request %p\n", request);
 
-	if (request->device != NULL) {
-		request->device = tape_put_device(request->device);
-	}
+	if (request->device)
+		tape_put_device(request->device);
 	kfree(request->cpdata);
 	kfree(request->cpaddr);
 	kfree(request);
@@ -838,7 +836,8 @@ static void tape_long_busy_timeout(unsigned long data)
 	BUG_ON(request->status != TAPE_REQUEST_LONG_BUSY);
 	DBF_LH(6, "%08x: Long busy timeout.\n", device->cdev_id);
 	__tape_start_next_request(device);
-	device->lb_timeout.data = (unsigned long) tape_put_device(device);
+	device->lb_timeout.data = 0UL;
+	tape_put_device(device);
 	spin_unlock_irq(get_ccwdev_lock(device->cdev));
 }
 
@@ -918,7 +917,7 @@ __tape_start_request(struct tape_device *device, struct tape_request *request)
 	}
 
 	/* Increase use count of device for the added request. */
-	request->device = tape_get_device_reference(device);
+	request->device = tape_get_device(device);
 
 	if (list_empty(&device->req_queue)) {
 		/* No other requests are on the queue. Start this one. */
@@ -1117,7 +1116,7 @@ __tape_do_irq (struct ccw_device *cdev, unsigned long intparm, struct irb *irb)
 		if (req->status == TAPE_REQUEST_LONG_BUSY) {
 			DBF_EVENT(3, "(%08x): del timer\n", device->cdev_id);
 			if (del_timer(&device->lb_timeout)) {
-				device->lb_timeout.data = (unsigned long)
+				device->lb_timeout.data = 0UL;
 				tape_put_device(device);
 				__tape_start_next_request(device);
 			}
@@ -1173,7 +1172,7 @@ __tape_do_irq (struct ccw_device *cdev, unsigned long intparm, struct irb *irb)
 			break;
 		case TAPE_IO_LONG_BUSY:
 			device->lb_timeout.data =
-				(unsigned long)tape_get_device_reference(device);
+				(unsigned long) tape_get_device(device);
 			device->lb_timeout.expires = jiffies +
 				LONG_BUSY_TIMEOUT * HZ;
 			DBF_EVENT(3, "(%08x): add timer\n", device->cdev_id);
@@ -1326,7 +1325,7 @@ EXPORT_SYMBOL(tape_generic_online);
 EXPORT_SYMBOL(tape_generic_offline);
 EXPORT_SYMBOL(tape_generic_pm_suspend);
 EXPORT_SYMBOL(tape_put_device);
-EXPORT_SYMBOL(tape_get_device_reference);
+EXPORT_SYMBOL(tape_get_device);
 EXPORT_SYMBOL(tape_state_verbose);
 EXPORT_SYMBOL(tape_op_verbose);
 EXPORT_SYMBOL(tape_state_set);

drivers/s390/char/tape_proc.c

@@ -45,7 +45,7 @@ static int tape_proc_show(struct seq_file *m, void *v)
 		seq_printf(m, "TapeNo\tBusID      CuType/Model\t"
 			"DevType/Model\tBlkSize\tState\tOp\tMedState\n");
 	}
-	device = tape_get_device(n);
+	device = tape_find_device(n);
 	if (IS_ERR(device))
 		return 0;
 	spin_lock_irq(get_ccwdev_lock(device->cdev));

drivers/s390/char/tty3270.c

@@ -19,6 +19,7 @@
 
 #include <linux/slab.h>
 #include <linux/bootmem.h>
+#include <linux/compat.h>
 
 #include <asm/ccwdev.h>
 #include <asm/cio.h>
@@ -1731,6 +1732,22 @@ tty3270_ioctl(struct tty_struct *tty, struct file *file,
 	return kbd_ioctl(tp->kbd, file, cmd, arg);
 }
 
+#ifdef CONFIG_COMPAT
+static long
+tty3270_compat_ioctl(struct tty_struct *tty, struct file *file,
+	      unsigned int cmd, unsigned long arg)
+{
+	struct tty3270 *tp;
+
+	tp = tty->driver_data;
+	if (!tp)
+		return -ENODEV;
+	if (tty->flags & (1 << TTY_IO_ERROR))
+		return -EIO;
+	return kbd_ioctl(tp->kbd, file, cmd, (unsigned long)compat_ptr(arg));
+}
+#endif
+
 static const struct tty_operations tty3270_ops = {
 	.open = tty3270_open,
 	.close = tty3270_close,
@@ -1745,6 +1762,9 @@ static const struct tty_operations tty3270_ops = {
 	.hangup = tty3270_hangup,
 	.wait_until_sent = tty3270_wait_until_sent,
 	.ioctl = tty3270_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl = tty3270_compat_ioctl,
+#endif
 	.set_termios = tty3270_set_termios
 };
 

drivers/s390/char/vmlogrdr.c

@@ -312,11 +312,9 @@ static int vmlogrdr_open (struct inode *inode, struct file *filp)
 		return -ENOSYS;
 
 	/* Besure this device hasn't already been opened */
-	lock_kernel();
 	spin_lock_bh(&logptr->priv_lock);
 	if (logptr->dev_in_use)	{
 		spin_unlock_bh(&logptr->priv_lock);
-		unlock_kernel();
 		return -EBUSY;
 	}
 	logptr->dev_in_use = 1;
@@ -360,9 +358,8 @@ static int vmlogrdr_open (struct inode *inode, struct file *filp)
 		   || (logptr->iucv_path_severed));
 	if (logptr->iucv_path_severed)
 		goto out_record;
- 	ret = nonseekable_open(inode, filp);
-	unlock_kernel();
-	return ret;
+	nonseekable_open(inode, filp);
+	return 0;
 
 out_record:
 	if (logptr->autorecording)
@@ -372,7 +369,6 @@ static int vmlogrdr_open (struct inode *inode, struct file *filp)
 	logptr->path = NULL;
 out_dev:
 	logptr->dev_in_use = 0;
-	unlock_kernel();
 	return -EIO;
 }
 

drivers/s390/char/vmur.c

@@ -695,7 +695,6 @@ static int ur_open(struct inode *inode, struct file *file)
 
 	if (accmode == O_RDWR)
 		return -EACCES;
-	lock_kernel();
 	/*
 	 * We treat the minor number as the devno of the ur device
 	 * to find in the driver tree.
@@ -749,7 +748,6 @@ static int ur_open(struct inode *inode, struct file *file)
 		goto fail_urfile_free;
 	urf->file_reclen = rc;
 	file->private_data = urf;
-	unlock_kernel();
 	return 0;
 
 fail_urfile_free:
@@ -761,7 +759,6 @@ static int ur_open(struct inode *inode, struct file *file)
 fail_put:
 	urdev_put(urd);
 out:
-	unlock_kernel();
 	return rc;
 }
 

drivers/s390/char/vmwatchdog.c

@@ -19,7 +19,6 @@
 #include <linux/moduleparam.h>
 #include <linux/suspend.h>
 #include <linux/watchdog.h>
-#include <linux/smp_lock.h>
 
 #include <asm/ebcdic.h>
 #include <asm/io.h>
@@ -49,6 +48,8 @@ static unsigned int vmwdt_interval = 60;
 static unsigned long vmwdt_is_open;
 static int vmwdt_expect_close;
 
+static DEFINE_MUTEX(vmwdt_mutex);
+
 #define VMWDT_OPEN	0	/* devnode is open or suspend in progress */
 #define VMWDT_RUNNING	1	/* The watchdog is armed */
 
@@ -133,15 +134,11 @@ static int __init vmwdt_probe(void)
 static int vmwdt_open(struct inode *i, struct file *f)
 {
 	int ret;
-	lock_kernel();
-	if (test_and_set_bit(VMWDT_OPEN, &vmwdt_is_open)) {
-		unlock_kernel();
+	if (test_and_set_bit(VMWDT_OPEN, &vmwdt_is_open))
 		return -EBUSY;
-	}
 	ret = vmwdt_keepalive();
 	if (ret)
 		clear_bit(VMWDT_OPEN, &vmwdt_is_open);
-	unlock_kernel();
 	return ret ? ret : nonseekable_open(i, f);
 }
 
@@ -160,8 +157,7 @@ static struct watchdog_info vmwdt_info = {
 	.identity = "z/VM Watchdog Timer",
 };
 
-static int vmwdt_ioctl(struct inode *i, struct file *f,
-			  unsigned int cmd, unsigned long arg)
+static int __vmwdt_ioctl(unsigned int cmd, unsigned long arg)
 {
 	switch (cmd) {
 	case WDIOC_GETSUPPORT:
@@ -205,10 +201,19 @@ static int vmwdt_ioctl(struct inode *i, struct file *f,
 	case WDIOC_KEEPALIVE:
 		return vmwdt_keepalive();
 	}
-
 	return -EINVAL;
 }
 
+static long vmwdt_ioctl(struct file *f, unsigned int cmd, unsigned long arg)
+{
+	int rc;
+
+	mutex_lock(&vmwdt_mutex);
+	rc = __vmwdt_ioctl(cmd, arg);
+	mutex_unlock(&vmwdt_mutex);
+	return (long) rc;
+}
+
 static ssize_t vmwdt_write(struct file *f, const char __user *buf,
 				size_t count, loff_t *ppos)
 {
@@ -288,7 +293,7 @@ static struct notifier_block vmwdt_power_notifier = {
 static const struct file_operations vmwdt_fops = {
 	.open    = &vmwdt_open,
 	.release = &vmwdt_close,
-	.ioctl   = &vmwdt_ioctl,
+	.unlocked_ioctl = &vmwdt_ioctl,
 	.write   = &vmwdt_write,
 	.owner   = THIS_MODULE,
 };
@@ -309,6 +314,10 @@ static int __init vmwdt_init(void)
 	ret = register_pm_notifier(&vmwdt_power_notifier);
 	if (ret)
 		return ret;
+	/*
+	 * misc_register() has to be the last action in module_init(), because
+	 * file operations will be available right after this.
+	 */
 	ret = misc_register(&vmwdt_dev);
 	if (ret) {
 		unregister_pm_notifier(&vmwdt_power_notifier);

drivers/s390/cio/Makefile

@@ -3,7 +3,7 @@
 #
 
 obj-y += airq.o blacklist.o chsc.o cio.o css.o chp.o idset.o isc.o \
-	fcx.o itcw.o crw.o
+	fcx.o itcw.o crw.o ccwreq.o
 ccw_device-objs += device.o device_fsm.o device_ops.o
 ccw_device-objs += device_id.o device_pgid.o device_status.o
 obj-y += ccw_device.o cmf.o

drivers/s390/cio/ccwreq.c

+/*
+ *  Handling of internal CCW device requests.
+ *
+ *    Copyright IBM Corp. 2009
+ *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#include <linux/types.h>
+#include <linux/err.h>
+#include <asm/ccwdev.h>
+#include <asm/cio.h>
+
+#include "io_sch.h"
+#include "cio.h"
+#include "device.h"
+#include "cio_debug.h"
+
+/**
+ * lpm_adjust - adjust path mask
+ * @lpm: path mask to adjust
+ * @mask: mask of available paths
+ *
+ * Shift @lpm right until @lpm and @mask have at least one bit in common or
+ * until @lpm is zero. Return the resulting lpm.
+ */
+int lpm_adjust(int lpm, int mask)
+{
+	while (lpm && ((lpm & mask) == 0))
+		lpm >>= 1;
+	return lpm;
+}
+
+/*
+ * Adjust path mask to use next path and reset retry count. Return resulting
+ * path mask.
+ */
+static u16 ccwreq_next_path(struct ccw_device *cdev)
+{
+	struct ccw_request *req = &cdev->private->req;
+
+	req->retries	= req->maxretries;
+	req->mask	= lpm_adjust(req->mask >>= 1, req->lpm);
+
+	return req->mask;
+}
+
+/*
+ * Clean up device state and report to callback.
+ */
+static void ccwreq_stop(struct ccw_device *cdev, int rc)
+{
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	struct ccw_request *req = &cdev->private->req;
+
+	if (req->done)
+		return;
+	req->done = 1;
+	ccw_device_set_timeout(cdev, 0);
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+	sch->lpm = sch->schib.pmcw.pam;
+	if (rc && rc != -ENODEV && req->drc)
+		rc = req->drc;
+	req->callback(cdev, req->data, rc);
+}
+
+/*
+ * (Re-)Start the operation until retries and paths are exhausted.
+ */
+static void ccwreq_do(struct ccw_device *cdev)
+{
+	struct ccw_request *req = &cdev->private->req;
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	struct ccw1 *cp = req->cp;
+	int rc = -EACCES;
+
+	while (req->mask) {
+		if (req->retries-- == 0) {
+			/* Retries exhausted, try next path. */
+			ccwreq_next_path(cdev);
+			continue;
+		}
+		/* Perform start function. */
+		sch->lpm = 0xff;
+		memset(&cdev->private->irb, 0, sizeof(struct irb));
+		rc = cio_start(sch, cp, (u8) req->mask);
+		if (rc == 0) {
+			/* I/O started successfully. */
+			ccw_device_set_timeout(cdev, req->timeout);
+			return;
+		}
+		if (rc == -ENODEV) {
+			/* Permanent device error. */
+			break;
+		}
+		if (rc == -EACCES) {
+			/* Permant path error. */
+			ccwreq_next_path(cdev);
+			continue;
+		}
+		/* Temporary improper status. */
+		rc = cio_clear(sch);
+		if (rc)
+			break;
+		return;
+	}
+	ccwreq_stop(cdev, rc);
+}
+
+/**
+ * ccw_request_start - perform I/O request
+ * @cdev: ccw device
+ *
+ * Perform the I/O request specified by cdev->req.
+ */
+void ccw_request_start(struct ccw_device *cdev)
+{
+	struct ccw_request *req = &cdev->private->req;
+
+	/* Try all paths twice to counter link flapping. */
+	req->mask	= 0x8080;
+	req->retries	= req->maxretries;
+	req->mask	= lpm_adjust(req->mask, req->lpm);
+	req->drc	= 0;
+	req->done	= 0;
+	req->cancel	= 0;
+	if (!req->mask)
+		goto out_nopath;
+	ccwreq_do(cdev);
+	return;
+
+out_nopath:
+	ccwreq_stop(cdev, -EACCES);
+}
+
+/**
+ * ccw_request_cancel - cancel running I/O request
+ * @cdev: ccw device
+ *
+ * Cancel the I/O request specified by cdev->req. Return non-zero if request
+ * has already finished, zero otherwise.
+ */
+int ccw_request_cancel(struct ccw_device *cdev)
+{
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	struct ccw_request *req = &cdev->private->req;
+	int rc;
+
+	if (req->done)
+		return 1;
+	req->cancel = 1;
+	rc = cio_clear(sch);
+	if (rc)
+		ccwreq_stop(cdev, rc);
+	return 0;
+}
+
+/*
+ * Return the status of the internal I/O started on the specified ccw device.
+ * Perform BASIC SENSE if required.
+ */
+static enum io_status ccwreq_status(struct ccw_device *cdev, struct irb *lcirb)
+{
+	struct irb *irb = &cdev->private->irb;
+	struct cmd_scsw *scsw = &irb->scsw.cmd;
+
+	/* Perform BASIC SENSE if needed. */
+	if (ccw_device_accumulate_and_sense(cdev, lcirb))
+		return IO_RUNNING;
+	/* Check for halt/clear interrupt. */
+	if (scsw->fctl & (SCSW_FCTL_HALT_FUNC | SCSW_FCTL_CLEAR_FUNC))
+		return IO_KILLED;
+	/* Check for path error. */
+	if (scsw->cc == 3 || scsw->pno)
+		return IO_PATH_ERROR;
+	/* Handle BASIC SENSE data. */
+	if (irb->esw.esw0.erw.cons) {
+		CIO_TRACE_EVENT(2, "sensedata");
+		CIO_HEX_EVENT(2, &cdev->private->dev_id,
+			      sizeof(struct ccw_dev_id));
+		CIO_HEX_EVENT(2, &cdev->private->irb.ecw, SENSE_MAX_COUNT);
+		/* Check for command reject. */
+		if (irb->ecw[0] & SNS0_CMD_REJECT)
+			return IO_REJECTED;
+		/* Assume that unexpected SENSE data implies an error. */
+		return IO_STATUS_ERROR;
+	}
+	/* Check for channel errors. */
+	if (scsw->cstat != 0)
+		return IO_STATUS_ERROR;
+	/* Check for device errors. */
+	if (scsw->dstat & ~(DEV_STAT_CHN_END | DEV_STAT_DEV_END))
+		return IO_STATUS_ERROR;
+	/* Check for final state. */
+	if (!(scsw->dstat & DEV_STAT_DEV_END))
+		return IO_RUNNING;
+	/* Check for other improper status. */
+	if (scsw->cc == 1 && (scsw->stctl & SCSW_STCTL_ALERT_STATUS))
+		return IO_STATUS_ERROR;
+	return IO_DONE;
+}
+
+/*
+ * Log ccw request status.
+ */
+static void ccwreq_log_status(struct ccw_device *cdev, enum io_status status)
+{
+	struct ccw_request *req = &cdev->private->req;
+	struct {
+		struct ccw_dev_id dev_id;
+		u16 retries;
+		u8 lpm;
+		u8 status;
+	}  __attribute__ ((packed)) data;
+	data.dev_id	= cdev->private->dev_id;
+	data.retries	= req->retries;
+	data.lpm	= (u8) req->mask;
+	data.status	= (u8) status;
+	CIO_TRACE_EVENT(2, "reqstat");
+	CIO_HEX_EVENT(2, &data, sizeof(data));
+}
+
+/**
+ * ccw_request_handler - interrupt handler for I/O request procedure.
+ * @cdev: ccw device
+ *
+ * Handle interrupt during I/O request procedure.
+ */
+void ccw_request_handler(struct ccw_device *cdev)
+{
+	struct ccw_request *req = &cdev->private->req;
+	struct irb *irb = (struct irb *) __LC_IRB;
+	enum io_status status;
+	int rc = -EOPNOTSUPP;
+
+	/* Check status of I/O request. */
+	status = ccwreq_status(cdev, irb);
+	if (req->filter)
+		status = req->filter(cdev, req->data, irb, status);
+	if (status != IO_RUNNING)
+		ccw_device_set_timeout(cdev, 0);
+	if (status != IO_DONE && status != IO_RUNNING)
+		ccwreq_log_status(cdev, status);
+	switch (status) {
+	case IO_DONE:
+		break;
+	case IO_RUNNING:
+		return;
+	case IO_REJECTED:
+		goto err;
+	case IO_PATH_ERROR:
+		goto out_next_path;
+	case IO_STATUS_ERROR:
+		goto out_restart;
+	case IO_KILLED:
+		/* Check if request was cancelled on purpose. */
+		if (req->cancel) {
+			rc = -EIO;
+			goto err;
+		}
+		goto out_restart;
+	}
+	/* Check back with request initiator. */
+	if (!req->check)
+		goto out;
+	switch (req->check(cdev, req->data)) {
+	case 0:
+		break;
+	case -EAGAIN:
+		goto out_restart;
+	case -EACCES:
+		goto out_next_path;
+	default:
+		goto err;
+	}
+out:
+	ccwreq_stop(cdev, 0);
+	return;
+
+out_next_path:
+	/* Try next path and restart I/O. */
+	if (!ccwreq_next_path(cdev)) {
+		rc = -EACCES;
+		goto err;
+	}
+out_restart:
+	/* Restart. */
+	ccwreq_do(cdev);
+	return;
+err:
+	ccwreq_stop(cdev, rc);
+}
+
+
+/**
+ * ccw_request_timeout - timeout handler for I/O request procedure
+ * @cdev: ccw device
+ *
+ * Handle timeout during I/O request procedure.
+ */
+void ccw_request_timeout(struct ccw_device *cdev)
+{
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	struct ccw_request *req = &cdev->private->req;
+	int rc;
+
+	if (!ccwreq_next_path(cdev)) {
+		/* set the final return code for this request */
+		req->drc = -ETIME;
+	}
+	rc = cio_clear(sch);
+	if (rc)
+		goto err;
+	return;
+
+err:
+	ccwreq_stop(cdev, rc);
+}
+
+/**
+ * ccw_request_notoper - notoper handler for I/O request procedure
+ * @cdev: ccw device
+ *
+ * Handle timeout during I/O request procedure.
+ */
+void ccw_request_notoper(struct ccw_device *cdev)
+{
+	ccwreq_stop(cdev, -ENODEV);
+}

drivers/s390/cio/cio.h

@@ -68,6 +68,11 @@ struct schib {
 	__u8 mda[4];		 /* model dependent area */
 } __attribute__ ((packed,aligned(4)));
 
+enum sch_todo {
+	SCH_TODO_NOTHING,
+	SCH_TODO_UNREG,
+};
+
 /* subchannel data structure used by I/O subroutines */
 struct subchannel {
 	struct subchannel_id schid;
@@ -95,7 +100,8 @@ struct subchannel {
 	struct device dev;	/* entry in device tree */
 	struct css_driver *driver;
 	void *private; /* private per subchannel type data */
-	struct work_struct work;
+	enum sch_todo todo;
+	struct work_struct todo_work;
 	struct schib_config config;
 } __attribute__ ((aligned(8)));
 

drivers/s390/cio/css.c

@@ -133,6 +133,8 @@ int for_each_subchannel_staged(int (*fn_known)(struct subchannel *, void *),
 	return rc;
 }
 
+static void css_sch_todo(struct work_struct *work);
+
 static struct subchannel *
 css_alloc_subchannel(struct subchannel_id schid)
 {
@@ -147,6 +149,7 @@ css_alloc_subchannel(struct subchannel_id schid)
 		kfree(sch);
 		return ERR_PTR(ret);
 	}
+	INIT_WORK(&sch->todo_work, css_sch_todo);
 	return sch;
 }
 
@@ -190,6 +193,51 @@ void css_sch_device_unregister(struct subchannel *sch)
 }
 EXPORT_SYMBOL_GPL(css_sch_device_unregister);
 
+static void css_sch_todo(struct work_struct *work)
+{
+	struct subchannel *sch;
+	enum sch_todo todo;
+
+	sch = container_of(work, struct subchannel, todo_work);
+	/* Find out todo. */
+	spin_lock_irq(sch->lock);
+	todo = sch->todo;
+	CIO_MSG_EVENT(4, "sch_todo: sch=0.%x.%04x, todo=%d\n", sch->schid.ssid,
+		      sch->schid.sch_no, todo);
+	sch->todo = SCH_TODO_NOTHING;
+	spin_unlock_irq(sch->lock);
+	/* Perform todo. */
+	if (todo == SCH_TODO_UNREG)
+		css_sch_device_unregister(sch);
+	/* Release workqueue ref. */
+	put_device(&sch->dev);
+}
+
+/**
+ * css_sched_sch_todo - schedule a subchannel operation
+ * @sch: subchannel
+ * @todo: todo
+ *
+ * Schedule the operation identified by @todo to be performed on the slow path
+ * workqueue. Do nothing if another operation with higher priority is already
+ * scheduled. Needs to be called with subchannel lock held.
+ */
+void css_sched_sch_todo(struct subchannel *sch, enum sch_todo todo)
+{
+	CIO_MSG_EVENT(4, "sch_todo: sched sch=0.%x.%04x todo=%d\n",
+		      sch->schid.ssid, sch->schid.sch_no, todo);
+	if (sch->todo >= todo)
+		return;
+	/* Get workqueue ref. */
+	if (!get_device(&sch->dev))
+		return;
+	sch->todo = todo;
+	if (!queue_work(slow_path_wq, &sch->todo_work)) {
+		/* Already queued, release workqueue ref. */
+		put_device(&sch->dev);
+	}
+}
+
 static void ssd_from_pmcw(struct chsc_ssd_info *ssd, struct pmcw *pmcw)
 {
 	int i;
@@ -376,8 +424,8 @@ static int css_evaluate_new_subchannel(struct subchannel_id schid, int slow)
 		/* Unusable - ignore. */
 		return 0;
 	}
-	CIO_MSG_EVENT(4, "Evaluating schid 0.%x.%04x, event %d, unknown, "
-			 "slow path.\n", schid.ssid, schid.sch_no, CIO_OPER);
+	CIO_MSG_EVENT(4, "event: sch 0.%x.%04x, new\n", schid.ssid,
+		      schid.sch_no);
 
 	return css_probe_device(schid);
 }
@@ -394,6 +442,10 @@ static int css_evaluate_known_subchannel(struct subchannel *sch, int slow)
 				"Got subchannel machine check but "
 				"no sch_event handler provided.\n");
 	}
+	if (ret != 0 && ret != -EAGAIN) {
+		CIO_MSG_EVENT(2, "eval: sch 0.%x.%04x, rc=%d\n",
+			      sch->schid.ssid, sch->schid.sch_no, ret);
+	}
 	return ret;
 }
 
@@ -684,6 +736,7 @@ static int __init setup_css(int nr)
 	css->pseudo_subchannel->dev.parent = &css->device;
 	css->pseudo_subchannel->dev.release = css_subchannel_release;
 	dev_set_name(&css->pseudo_subchannel->dev, "defunct");
+	mutex_init(&css->pseudo_subchannel->reg_mutex);
 	ret = cio_create_sch_lock(css->pseudo_subchannel);
 	if (ret) {
 		kfree(css->pseudo_subchannel);

drivers/s390/cio/css.h

@@ -11,6 +11,8 @@
 #include <asm/chpid.h>
 #include <asm/schid.h>
 
+#include "cio.h"
+
 /*
  * path grouping stuff
  */
@@ -151,4 +153,5 @@ int css_sch_is_valid(struct schib *);
 
 extern struct workqueue_struct *slow_path_wq;
 void css_wait_for_slow_path(void);
+void css_sched_sch_todo(struct subchannel *sch, enum sch_todo todo);
 #endif

drivers/s390/cio/device.h

@@ -21,7 +21,6 @@ enum dev_state {
 	DEV_STATE_DISBAND_PGID,
 	DEV_STATE_BOXED,
 	/* states to wait for i/o completion before doing something */
-	DEV_STATE_CLEAR_VERIFY,
 	DEV_STATE_TIMEOUT_KILL,
 	DEV_STATE_QUIESCE,
 	/* special states for devices gone not operational */
@@ -29,6 +28,7 @@ enum dev_state {
 	DEV_STATE_DISCONNECTED_SENSE_ID,
 	DEV_STATE_CMFCHANGE,
 	DEV_STATE_CMFUPDATE,
+	DEV_STATE_STEAL_LOCK,
 	/* last element! */
 	NR_DEV_STATES
 };
@@ -81,17 +81,16 @@ void io_subchannel_init_config(struct subchannel *sch);
 
 int ccw_device_cancel_halt_clear(struct ccw_device *);
 
-void ccw_device_do_unbind_bind(struct work_struct *);
-void ccw_device_move_to_orphanage(struct work_struct *);
 int ccw_device_is_orphan(struct ccw_device *);
 
-int ccw_device_recognition(struct ccw_device *);
+void ccw_device_recognition(struct ccw_device *);
 int ccw_device_online(struct ccw_device *);
 int ccw_device_offline(struct ccw_device *);
 void ccw_device_update_sense_data(struct ccw_device *);
 int ccw_device_test_sense_data(struct ccw_device *);
 void ccw_device_schedule_sch_unregister(struct ccw_device *);
 int ccw_purge_blacklisted(void);
+void ccw_device_sched_todo(struct ccw_device *cdev, enum cdev_todo todo);
 
 /* Function prototypes for device status and basic sense stuff. */
 void ccw_device_accumulate_irb(struct ccw_device *, struct irb *);
@@ -99,24 +98,28 @@ void ccw_device_accumulate_basic_sense(struct ccw_device *, struct irb *);
 int ccw_device_accumulate_and_sense(struct ccw_device *, struct irb *);
 int ccw_device_do_sense(struct ccw_device *, struct irb *);
 
+/* Function prototype for internal request handling. */
+int lpm_adjust(int lpm, int mask);
+void ccw_request_start(struct ccw_device *);
+int ccw_request_cancel(struct ccw_device *cdev);
+void ccw_request_handler(struct ccw_device *cdev);
+void ccw_request_timeout(struct ccw_device *cdev);
+void ccw_request_notoper(struct ccw_device *cdev);
+
 /* Function prototypes for sense id stuff. */
 void ccw_device_sense_id_start(struct ccw_device *);
-void ccw_device_sense_id_irq(struct ccw_device *, enum dev_event);
 void ccw_device_sense_id_done(struct ccw_device *, int);
 
 /* Function prototypes for path grouping stuff. */
-void ccw_device_sense_pgid_start(struct ccw_device *);
-void ccw_device_sense_pgid_irq(struct ccw_device *, enum dev_event);
-void ccw_device_sense_pgid_done(struct ccw_device *, int);
-
 void ccw_device_verify_start(struct ccw_device *);
-void ccw_device_verify_irq(struct ccw_device *, enum dev_event);
 void ccw_device_verify_done(struct ccw_device *, int);
 
 void ccw_device_disband_start(struct ccw_device *);
-void ccw_device_disband_irq(struct ccw_device *, enum dev_event);
 void ccw_device_disband_done(struct ccw_device *, int);
 
+void ccw_device_stlck_start(struct ccw_device *, void *, void *, void *);
+void ccw_device_stlck_done(struct ccw_device *, void *, int);
+
 int ccw_device_call_handler(struct ccw_device *);
 
 int ccw_device_stlck(struct ccw_device *);

drivers/s390/cio/device_ops.c

@@ -11,6 +11,7 @@
 #include <linux/list.h>
 #include <linux/device.h>
 #include <linux/delay.h>
+#include <linux/completion.h>
 
 #include <asm/ccwdev.h>
 #include <asm/idals.h>
@@ -46,6 +47,7 @@ int ccw_device_set_options_mask(struct ccw_device *cdev, unsigned long flags)
 	cdev->private->options.repall = (flags & CCWDEV_REPORT_ALL) != 0;
 	cdev->private->options.pgroup = (flags & CCWDEV_DO_PATHGROUP) != 0;
 	cdev->private->options.force = (flags & CCWDEV_ALLOW_FORCE) != 0;
+	cdev->private->options.mpath = (flags & CCWDEV_DO_MULTIPATH) != 0;
 	return 0;
 }
 
@@ -74,6 +76,7 @@ int ccw_device_set_options(struct ccw_device *cdev, unsigned long flags)
 	cdev->private->options.repall |= (flags & CCWDEV_REPORT_ALL) != 0;
 	cdev->private->options.pgroup |= (flags & CCWDEV_DO_PATHGROUP) != 0;
 	cdev->private->options.force |= (flags & CCWDEV_ALLOW_FORCE) != 0;
+	cdev->private->options.mpath |= (flags & CCWDEV_DO_MULTIPATH) != 0;
 	return 0;
 }
 
@@ -90,8 +93,33 @@ void ccw_device_clear_options(struct ccw_device *cdev, unsigned long flags)
 	cdev->private->options.repall &= (flags & CCWDEV_REPORT_ALL) == 0;
 	cdev->private->options.pgroup &= (flags & CCWDEV_DO_PATHGROUP) == 0;
 	cdev->private->options.force &= (flags & CCWDEV_ALLOW_FORCE) == 0;
+	cdev->private->options.mpath &= (flags & CCWDEV_DO_MULTIPATH) == 0;
 }
 
+/**
+ * ccw_device_is_pathgroup - determine if paths to this device are grouped
+ * @cdev: ccw device
+ *
+ * Return non-zero if there is a path group, zero otherwise.
+ */
+int ccw_device_is_pathgroup(struct ccw_device *cdev)
+{
+	return cdev->private->flags.pgroup;
+}
+EXPORT_SYMBOL(ccw_device_is_pathgroup);
+
+/**
+ * ccw_device_is_multipath - determine if device is operating in multipath mode
+ * @cdev: ccw device
+ *
+ * Return non-zero if device is operating in multipath mode, zero otherwise.
+ */
+int ccw_device_is_multipath(struct ccw_device *cdev)
+{
+	return cdev->private->flags.mpath;
+}
+EXPORT_SYMBOL(ccw_device_is_multipath);
+
 /**
  * ccw_device_clear() - terminate I/O request processing
  * @cdev: target ccw device
@@ -167,8 +195,7 @@ int ccw_device_start_key(struct ccw_device *cdev, struct ccw1 *cpa,
 		return -EINVAL;
 	if (cdev->private->state == DEV_STATE_NOT_OPER)
 		return -ENODEV;
-	if (cdev->private->state == DEV_STATE_VERIFY ||
-	    cdev->private->state == DEV_STATE_CLEAR_VERIFY) {
+	if (cdev->private->state == DEV_STATE_VERIFY) {
 		/* Remember to fake irb when finished. */
 		if (!cdev->private->flags.fake_irb) {
 			cdev->private->flags.fake_irb = 1;
@@ -478,74 +505,65 @@ __u8 ccw_device_get_path_mask(struct ccw_device *cdev)
 	return sch->lpm;
 }
 
+struct stlck_data {
+	struct completion done;
+	int rc;
+};
+
+void ccw_device_stlck_done(struct ccw_device *cdev, void *data, int rc)
+{
+	struct stlck_data *sdata = data;
+
+	sdata->rc = rc;
+	complete(&sdata->done);
+}
+
 /*
- * Try to break the lock on a boxed device.
+ * Perform unconditional reserve + release.
  */
-int
-ccw_device_stlck(struct ccw_device *cdev)
+int ccw_device_stlck(struct ccw_device *cdev)
 {
-	void *buf, *buf2;
-	unsigned long flags;
-	struct subchannel *sch;
-	int ret;
-
-	if (!cdev)
-		return -ENODEV;
+	struct subchannel *sch = to_subchannel(cdev->dev.parent);
+	struct stlck_data data;
+	u8 *buffer;
+	int rc;
 
-	if (cdev->drv && !cdev->private->options.force)
+	/* Check if steal lock operation is valid for this device. */
+	if (cdev->drv) {
+		if (!cdev->private->options.force)
 			return -EINVAL;
-
-	sch = to_subchannel(cdev->dev.parent);
-	
-	CIO_TRACE_EVENT(2, "stl lock");
-	CIO_TRACE_EVENT(2, dev_name(&cdev->dev));
-
-	buf = kmalloc(32*sizeof(char), GFP_DMA|GFP_KERNEL);
-	if (!buf)
-		return -ENOMEM;
-	buf2 = kmalloc(32*sizeof(char), GFP_DMA|GFP_KERNEL);
-	if (!buf2) {
-		kfree(buf);
-		return -ENOMEM;
 	}
-	spin_lock_irqsave(sch->lock, flags);
-	ret = cio_enable_subchannel(sch, (u32)(addr_t)sch);
-	if (ret)
-		goto out_unlock;
-	/*
-	 * Setup ccw. We chain an unconditional reserve and a release so we
-	 * only break the lock.
-	 */
-	cdev->private->iccws[0].cmd_code = CCW_CMD_STLCK;
-	cdev->private->iccws[0].cda = (__u32) __pa(buf);
-	cdev->private->iccws[0].count = 32;
-	cdev->private->iccws[0].flags = CCW_FLAG_CC;
-	cdev->private->iccws[1].cmd_code = CCW_CMD_RELEASE;
-	cdev->private->iccws[1].cda = (__u32) __pa(buf2);
-	cdev->private->iccws[1].count = 32;
-	cdev->private->iccws[1].flags = 0;
-	ret = cio_start(sch, cdev->private->iccws, 0);
-	if (ret) {
-		cio_disable_subchannel(sch); //FIXME: return code?
+	buffer = kzalloc(64, GFP_DMA | GFP_KERNEL);
+	if (!buffer)
+		return -ENOMEM;
+	init_completion(&data.done);
+	data.rc = -EIO;
+	spin_lock_irq(sch->lock);
+	rc = cio_enable_subchannel(sch, (u32) (addr_t) sch);
+	if (rc)
 		goto out_unlock;
+	/* Perform operation. */
+	cdev->private->state = DEV_STATE_STEAL_LOCK,
+	ccw_device_stlck_start(cdev, &data, &buffer[0], &buffer[32]);
+	spin_unlock_irq(sch->lock);
+	/* Wait for operation to finish. */
+	if (wait_for_completion_interruptible(&data.done)) {
+		/* Got a signal. */
+		spin_lock_irq(sch->lock);
+		ccw_request_cancel(cdev);
+		spin_unlock_irq(sch->lock);
+		wait_for_completion(&data.done);
 	}
-	cdev->private->irb.scsw.cmd.actl |= SCSW_ACTL_START_PEND;
-	spin_unlock_irqrestore(sch->lock, flags);
-	wait_event(cdev->private->wait_q,
-		   cdev->private->irb.scsw.cmd.actl == 0);
-	spin_lock_irqsave(sch->lock, flags);
-	cio_disable_subchannel(sch); //FIXME: return code?
-	if ((cdev->private->irb.scsw.cmd.dstat !=
-	     (DEV_STAT_CHN_END|DEV_STAT_DEV_END)) ||
-	    (cdev->private->irb.scsw.cmd.cstat != 0))
-		ret = -EIO;
-	/* Clear irb. */
-	memset(&cdev->private->irb, 0, sizeof(struct irb));
+	rc = data.rc;
+	/* Check results. */
+	spin_lock_irq(sch->lock);
+	cio_disable_subchannel(sch);
+	cdev->private->state = DEV_STATE_BOXED;
 out_unlock:
-	kfree(buf);
-	kfree(buf2);
-	spin_unlock_irqrestore(sch->lock, flags);
-	return ret;
+	spin_unlock_irq(sch->lock);
+	kfree(buffer);
+
+	return rc;
 }
 
 void *ccw_device_get_chp_desc(struct ccw_device *cdev, int chp_no)

drivers/s390/cio/device_status.c

@@ -336,9 +336,6 @@ ccw_device_do_sense(struct ccw_device *cdev, struct irb *irb)
 	sense_ccw->count = SENSE_MAX_COUNT;
 	sense_ccw->flags = CCW_FLAG_SLI;
 
-	/* Reset internal retry indication. */
-	cdev->private->flags.intretry = 0;
-
 	rc = cio_start(sch, sense_ccw, 0xff);
 	if (rc == -ENODEV || rc == -EACCES)
 		dev_fsm_event(cdev, DEV_EVENT_VERIFY);

drivers/s390/cio/io_sch.h

 #ifndef S390_IO_SCH_H
 #define S390_IO_SCH_H
 
+#include <linux/types.h>
 #include <asm/schid.h>
+#include <asm/ccwdev.h>
+#include "css.h"
 
 /*
  * command-mode operation request block
@@ -67,6 +70,52 @@ struct io_subchannel_private {
 
 #define MAX_CIWS 8
 
+/*
+ * Possible status values for a CCW request's I/O.
+ */
+enum io_status {
+	IO_DONE,
+	IO_RUNNING,
+	IO_STATUS_ERROR,
+	IO_PATH_ERROR,
+	IO_REJECTED,
+	IO_KILLED
+};
+
+/**
+ * ccw_request - Internal CCW request.
+ * @cp: channel program to start
+ * @timeout: maximum allowable time in jiffies between start I/O and interrupt
+ * @maxretries: number of retries per I/O operation and path
+ * @lpm: mask of paths to use
+ * @check: optional callback that determines if results are final
+ * @filter: optional callback to adjust request status based on IRB data
+ * @callback: final callback
+ * @data: user-defined pointer passed to all callbacks
+ * @mask: current path mask
+ * @retries: current number of retries
+ * @drc: delayed return code
+ * @cancel: non-zero if request was cancelled
+ * @done: non-zero if request was finished
+ */
+struct ccw_request {
+	struct ccw1 *cp;
+	unsigned long timeout;
+	u16 maxretries;
+	u8 lpm;
+	int (*check)(struct ccw_device *, void *);
+	enum io_status (*filter)(struct ccw_device *, void *, struct irb *,
+				 enum io_status);
+	void (*callback)(struct ccw_device *, void *, int);
+	void *data;
+	/* These fields are used internally. */
+	u16 mask;
+	u16 retries;
+	int drc;
+	int cancel:1;
+	int done:1;
+} __attribute__((packed));
+
 /*
  * sense-id response buffer layout
  */
@@ -82,32 +131,43 @@ struct senseid {
 	struct ciw ciw[MAX_CIWS];	/* variable # of CIWs */
 }  __attribute__ ((packed, aligned(4)));
 
+enum cdev_todo {
+	CDEV_TODO_NOTHING,
+	CDEV_TODO_ENABLE_CMF,
+	CDEV_TODO_REBIND,
+	CDEV_TODO_REGISTER,
+	CDEV_TODO_UNREG,
+	CDEV_TODO_UNREG_EVAL,
+};
+
 struct ccw_device_private {
 	struct ccw_device *cdev;
 	struct subchannel *sch;
 	int state;		/* device state */
 	atomic_t onoff;
-	unsigned long registered;
 	struct ccw_dev_id dev_id;	/* device id */
 	struct subchannel_id schid;	/* subchannel number */
-	u8 imask;		/* lpm mask for SNID/SID/SPGID */
-	int iretry;		/* retry counter SNID/SID/SPGID */
+	struct ccw_request req;		/* internal I/O request */
+	int iretry;
+	u8 pgid_valid_mask;		/* mask of valid PGIDs */
 	struct {
 		unsigned int fast:1;	/* post with "channel end" */
 		unsigned int repall:1;	/* report every interrupt status */
 		unsigned int pgroup:1;	/* do path grouping */
 		unsigned int force:1;	/* allow forced online */
+		unsigned int mpath:1;	/* do multipathing */
 	} __attribute__ ((packed)) options;
 	struct {
-		unsigned int pgid_single:1; /* use single path for Set PGID */
 		unsigned int esid:1;	    /* Ext. SenseID supported by HW */
 		unsigned int dosense:1;	    /* delayed SENSE required */
 		unsigned int doverify:1;    /* delayed path verification */
 		unsigned int donotify:1;    /* call notify function */
 		unsigned int recog_done:1;  /* dev. recog. complete */
 		unsigned int fake_irb:1;    /* deliver faked irb */
-		unsigned int intretry:1;    /* retry internal operation */
 		unsigned int resuming:1;    /* recognition while resume */
+		unsigned int pgroup:1;	    /* pathgroup is set up */
+		unsigned int mpath:1;	    /* multipathing is set up */
+		unsigned int initialized:1; /* set if initial reference held */
 	} __attribute__((packed)) flags;
 	unsigned long intparm;	/* user interruption parameter */
 	struct qdio_irq *qdio_data;
@@ -115,7 +175,8 @@ struct ccw_device_private {
 	struct senseid senseid;	/* SenseID info */
 	struct pgid pgid[8];	/* path group IDs per chpid*/
 	struct ccw1 iccws[2];	/* ccws for SNID/SID/SPGID commands */
-	struct work_struct kick_work;
+	struct work_struct todo_work;
+	enum cdev_todo todo;
 	wait_queue_head_t wait_q;
 	struct timer_list timer;
 	void *cmb;			/* measurement information */

drivers/s390/crypto/zcrypt_api.h

@@ -71,6 +71,8 @@ struct ica_z90_status {
 #define ZCRYPT_PCIXCC_MCL3	4
 #define ZCRYPT_CEX2C		5
 #define ZCRYPT_CEX2A		6
+#define ZCRYPT_CEX3C		7
+#define ZCRYPT_CEX3A		8
 
 /**
  * Large random numbers are pulled in 4096 byte chunks from the crypto cards

drivers/s390/crypto/zcrypt_pcica.c

@@ -281,6 +281,7 @@ static long zcrypt_pcica_modexpo(struct zcrypt_device *zdev,
 	struct completion work;
 	int rc;
 
+	ap_init_message(&ap_msg);
 	ap_msg.message = kmalloc(PCICA_MAX_MESSAGE_SIZE, GFP_KERNEL);
 	if (!ap_msg.message)
 		return -ENOMEM;
@@ -318,6 +319,7 @@ static long zcrypt_pcica_modexpo_crt(struct zcrypt_device *zdev,
 	struct completion work;
 	int rc;
 
+	ap_init_message(&ap_msg);
 	ap_msg.message = kmalloc(PCICA_MAX_MESSAGE_SIZE, GFP_KERNEL);
 	if (!ap_msg.message)
 		return -ENOMEM;