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

* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6: (37 commits)
  [S390] Avoid excessive inlining.
  [S390] Mark kernel text section read-only.
  [S390] Convert memory detection into C code.
  [S390] Calibrate delay and bogomips.
  [S390] Hypervisor filesystem (s390_hypfs) for z/VM
  [S390] Add crypto support for 3592 tape devices
  [S390] boot from NSS support
  [S390] Support for s390 Pseudo Random Number Generator
  [S390] ETR support.
  [S390] noexec protection
  [S390] move crypto options and some cleanup.
  [S390] cio: Don't spam debug feature.
  [S390] Cleanup of CHSC event handling.
  [S390] cio: declare hardware structures packed.
  [S390] Add set_fs(USER_DS) to start_thread().
  [S390] cio: Catch operand exceptions on stsch.
  [S390] Fix register usage description.
  [S390] kretprobe_trampoline_holder() in wrong section.
  [S390] Fix kprobes breakpoint handling.
  [S390] Update maintainers file.
  ...

Documentation/s390/Debugging390.txt

@@ -480,7 +480,7 @@ r2       argument 0 / return value 0                call-clobbered
 r3       argument 1 / return value 1 (if long long) call-clobbered
 r4       argument 2                                 call-clobbered
 r5       argument 3                                 call-clobbered
-r6	 argument 5                                 saved
+r6	 argument 4				    saved
 r7       pointer-to arguments 5 to ...              saved      
 r8       this & that                                saved
 r9       this & that                                saved

MAINTAINERS

@@ -2791,7 +2791,7 @@ M:	schwidefsky@de.ibm.com
 P:	Heiko Carstens
 M:	heiko.carstens@de.ibm.com
 M:	linux390@de.ibm.com
-L:	linux-390@vm.marist.edu
+L:	linux-s390@vger.kernel.org
 W:	http://www.ibm.com/developerworks/linux/linux390/
 S:	Supported
 
@@ -2799,7 +2799,7 @@ S390 NETWORK DRIVERS
 P:	Frank Pavlic
 M:	fpavlic@de.ibm.com
 M:	linux390@de.ibm.com
-L:	linux-390@vm.marist.edu
+L:	linux-s390@vger.kernel.org
 W:	http://www.ibm.com/developerworks/linux/linux390/
 S:	Supported
 
@@ -2807,7 +2807,7 @@ S390 ZFCP DRIVER
 P:	Swen Schillig
 M:	swen@vnet.ibm.com
 M:	linux390@de.ibm.com
-L:	linux-390@vm.marist.edu
+L:	linux-s390@vger.kernel.org
 W:	http://www.ibm.com/developerworks/linux/linux390/
 S:	Supported
 

arch/s390/Kconfig

@@ -34,10 +34,6 @@ config GENERIC_HWEIGHT
 	bool
 	default y
 
-config GENERIC_CALIBRATE_DELAY
-	bool
-	default y
-
 config GENERIC_TIME
 	def_bool y
 
@@ -134,6 +130,31 @@ 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.
+	  The kernel parameter noexec=on will enable this feature and also
+	  switch the addressing modes, default is disabled. Enabling this (via
+	  kernel parameter) on machines earlier than IBM System z9-109 EC/BC
+	  will reduce system performance.
+
 comment "Code generation options"
 
 choice

arch/s390/appldata/appldata_base.c

@@ -81,7 +81,7 @@ static struct ctl_table appldata_dir_table[] = {
 /*
  * Timer
  */
-DEFINE_PER_CPU(struct vtimer_list, appldata_timer);
+static DEFINE_PER_CPU(struct vtimer_list, appldata_timer);
 static atomic_t appldata_expire_count = ATOMIC_INIT(0);
 
 static DEFINE_SPINLOCK(appldata_timer_lock);

arch/s390/appldata/appldata_mem.c

@@ -36,7 +36,7 @@
  * book:
  * http://oss.software.ibm.com/developerworks/opensource/linux390/index.shtml
  */
-struct appldata_mem_data {
+static struct appldata_mem_data {
 	u64 timestamp;
 	u32 sync_count_1;       /* after VM collected the record data, */
 	u32 sync_count_2;	/* sync_count_1 and sync_count_2 should be the

arch/s390/appldata/appldata_net_sum.c

@@ -34,7 +34,7 @@
  * book:
  * http://oss.software.ibm.com/developerworks/opensource/linux390/index.shtml
  */
-struct appldata_net_sum_data {
+static struct appldata_net_sum_data {
 	u64 timestamp;
 	u32 sync_count_1;	/* after VM collected the record data, */
 	u32 sync_count_2;	/* sync_count_1 and sync_count_2 should be the

arch/s390/crypto/Kconfig

@@ -0,0 +1,60 @@
+config CRYPTO_SHA1_S390
+	tristate "SHA1 digest algorithm"
+	depends on S390
+	select CRYPTO_ALGAPI
+	help
+	  This is the s390 hardware accelerated implementation of the
+	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2).
+
+config CRYPTO_SHA256_S390
+	tristate "SHA256 digest algorithm"
+	depends on S390
+	select CRYPTO_ALGAPI
+	help
+	  This is the s390 hardware accelerated implementation of the
+	  SHA256 secure hash standard (DFIPS 180-2).
+
+	  This version of SHA implements a 256 bit hash with 128 bits of
+	  security against collision attacks.
+
+config CRYPTO_DES_S390
+	tristate "DES and Triple DES cipher algorithms"
+	depends on S390
+	select CRYPTO_ALGAPI
+	select CRYPTO_BLKCIPHER
+	help
+	  This us the s390 hardware accelerated implementation of the
+	  DES cipher algorithm (FIPS 46-2), and Triple DES EDE (FIPS 46-3).
+
+config CRYPTO_AES_S390
+	tristate "AES cipher algorithms"
+	depends on S390
+	select CRYPTO_ALGAPI
+	select CRYPTO_BLKCIPHER
+	help
+	  This is the s390 hardware accelerated implementation of the
+	  AES cipher algorithms (FIPS-197). AES uses the Rijndael
+	  algorithm.
+
+	  Rijndael appears to be consistently a very good performer in
+	  both hardware and software across a wide range of computing
+	  environments regardless of its use in feedback or non-feedback
+	  modes. Its key setup time is excellent, and its key agility is
+	  good. Rijndael's very low memory requirements make it very well
+	  suited for restricted-space environments, in which it also
+	  demonstrates excellent performance. Rijndael's operations are
+	  among the easiest to defend against power and timing attacks.
+
+	  On s390 the System z9-109 currently only supports the key size
+	  of 128 bit.
+
+config S390_PRNG
+	tristate "Pseudo random number generator device driver"
+	depends on S390
+	default "m"
+	help
+	  Select this option if you want to use the s390 pseudo random number
+	  generator. The PRNG is part of the cryptograhic processor functions
+	  and uses triple-DES to generate secure random numbers like the
+	  ANSI X9.17 standard. The PRNG is usable via the char device
+	  /dev/prandom.

arch/s390/crypto/Makefile

@@ -6,5 +6,4 @@ obj-$(CONFIG_CRYPTO_SHA1_S390) += sha1_s390.o
 obj-$(CONFIG_CRYPTO_SHA256_S390) += sha256_s390.o
 obj-$(CONFIG_CRYPTO_DES_S390) += des_s390.o des_check_key.o
 obj-$(CONFIG_CRYPTO_AES_S390) += aes_s390.o
-
-obj-$(CONFIG_CRYPTO_TEST) += crypt_s390_query.o
+obj-$(CONFIG_S390_PRNG) += prng.o

arch/s390/crypto/aes_s390.c

@@ -4,7 +4,7 @@
  * s390 implementation of the AES Cipher Algorithm.
  *
  * s390 Version:
- *   Copyright (C) 2005 IBM Deutschland GmbH, IBM Corporation
+ *   Copyright IBM Corp. 2005,2007
  *   Author(s): Jan Glauber (jang@de.ibm.com)
  *
  * Derived from "crypto/aes.c"
@@ -27,9 +27,11 @@
 /* data block size for all key lengths */
 #define AES_BLOCK_SIZE		16
 
-int has_aes_128 = 0;
-int has_aes_192 = 0;
-int has_aes_256 = 0;
+#define AES_KEYLEN_128		1
+#define AES_KEYLEN_192		2
+#define AES_KEYLEN_256		4
+
+static char keylen_flag = 0;
 
 struct s390_aes_ctx {
 	u8 iv[AES_BLOCK_SIZE];
@@ -47,20 +49,19 @@ static int aes_set_key(struct crypto_tfm *tfm, const u8 *in_key,
 
 	switch (key_len) {
 	case 16:
-		if (!has_aes_128)
+		if (!(keylen_flag & AES_KEYLEN_128))
 			goto fail;
 		break;
 	case 24:
-		if (!has_aes_192)
+		if (!(keylen_flag & AES_KEYLEN_192))
 			goto fail;
 
 		break;
 	case 32:
-		if (!has_aes_256)
+		if (!(keylen_flag & AES_KEYLEN_256))
 			goto fail;
 		break;
 	default:
-		/* invalid key length */
 		goto fail;
 		break;
 	}
@@ -322,34 +323,32 @@ static int __init aes_init(void)
 	int ret;
 
 	if (crypt_s390_func_available(KM_AES_128_ENCRYPT))
-		has_aes_128 = 1;
+		keylen_flag |= AES_KEYLEN_128;
 	if (crypt_s390_func_available(KM_AES_192_ENCRYPT))
-		has_aes_192 = 1;
+		keylen_flag |= AES_KEYLEN_192;
 	if (crypt_s390_func_available(KM_AES_256_ENCRYPT))
-		has_aes_256 = 1;
+		keylen_flag |= AES_KEYLEN_256;
+
+	if (!keylen_flag)
+		return -EOPNOTSUPP;
 
-	if (!has_aes_128 && !has_aes_192 && !has_aes_256)
-		return -ENOSYS;
+	/* z9 109 and z9 BC/EC only support 128 bit key length */
+	if (keylen_flag == AES_KEYLEN_128)
+		printk(KERN_INFO
+		       "aes_s390: hardware acceleration only available for"
+		       "128 bit keys\n");
 
 	ret = crypto_register_alg(&aes_alg);
-	if (ret != 0) {
-		printk(KERN_INFO "crypt_s390: aes-s390 couldn't be loaded.\n");
+	if (ret)
 		goto aes_err;
-	}
 
 	ret = crypto_register_alg(&ecb_aes_alg);
-	if (ret != 0) {
-		printk(KERN_INFO
-		       "crypt_s390: ecb-aes-s390 couldn't be loaded.\n");
+	if (ret)
 		goto ecb_aes_err;
-	}
 
 	ret = crypto_register_alg(&cbc_aes_alg);
-	if (ret != 0) {
-		printk(KERN_INFO
-		       "crypt_s390: cbc-aes-s390 couldn't be loaded.\n");
+	if (ret)
 		goto cbc_aes_err;
-	}
 
 out:
 	return ret;

arch/s390/crypto/crypt_s390.h

@@ -3,8 +3,9 @@
  *
  * Support for s390 cryptographic instructions.
  *
- *   Copyright (C) 2003 IBM Deutschland GmbH, IBM Corporation
- *   Author(s): Thomas Spatzier (tspat@de.ibm.com)
+ *   Copyright IBM Corp. 2003,2007
+ *   Author(s): Thomas Spatzier
+ *		Jan Glauber (jan.glauber@de.ibm.com)
  *
  * This program is free software; you can redistribute it and/or modify it
  * under the terms of the GNU General Public License as published by the Free
@@ -32,7 +33,8 @@ enum crypt_s390_operations {
 	CRYPT_S390_KMAC = 0x0500
 };
 
-/* function codes for KM (CIPHER MESSAGE) instruction
+/*
+ * function codes for KM (CIPHER MESSAGE) instruction
  * 0x80 is the decipher modifier bit
  */
 enum crypt_s390_km_func {
@@ -51,7 +53,8 @@ enum crypt_s390_km_func {
 	KM_AES_256_DECRYPT  = CRYPT_S390_KM | 0x14 | 0x80,
 };
 
-/* function codes for KMC (CIPHER MESSAGE WITH CHAINING)
+/*
+ * function codes for KMC (CIPHER MESSAGE WITH CHAINING)
  * instruction
  */
 enum crypt_s390_kmc_func {
@@ -68,9 +71,11 @@ enum crypt_s390_kmc_func {
 	KMC_AES_192_DECRYPT  = CRYPT_S390_KMC | 0x13 | 0x80,
 	KMC_AES_256_ENCRYPT  = CRYPT_S390_KMC | 0x14,
 	KMC_AES_256_DECRYPT  = CRYPT_S390_KMC | 0x14 | 0x80,
+	KMC_PRNG	     = CRYPT_S390_KMC | 0x43,
 };
 
-/* function codes for KIMD (COMPUTE INTERMEDIATE MESSAGE DIGEST)
+/*
+ * function codes for KIMD (COMPUTE INTERMEDIATE MESSAGE DIGEST)
  * instruction
  */
 enum crypt_s390_kimd_func {
@@ -79,7 +84,8 @@ enum crypt_s390_kimd_func {
 	KIMD_SHA_256 = CRYPT_S390_KIMD | 2,
 };
 
-/* function codes for KLMD (COMPUTE LAST MESSAGE DIGEST)
+/*
+ * function codes for KLMD (COMPUTE LAST MESSAGE DIGEST)
  * instruction
  */
 enum crypt_s390_klmd_func {
@@ -88,7 +94,8 @@ enum crypt_s390_klmd_func {
 	KLMD_SHA_256 = CRYPT_S390_KLMD | 2,
 };
 
-/* function codes for KMAC (COMPUTE MESSAGE AUTHENTICATION CODE)
+/*
+ * function codes for KMAC (COMPUTE MESSAGE AUTHENTICATION CODE)
  * instruction
  */
 enum crypt_s390_kmac_func {
@@ -98,229 +105,219 @@ enum crypt_s390_kmac_func {
 	KMAC_TDEA_192 = CRYPT_S390_KMAC | 3
 };
 
-/* status word for s390 crypto instructions' QUERY functions */
-struct crypt_s390_query_status {
-	u64 high;
-	u64 low;
-};
-
-/*
+/**
+ * crypt_s390_km:
+ * @func: the function code passed to KM; see crypt_s390_km_func
+ * @param: address of parameter block; see POP for details on each func
+ * @dest: address of destination memory area
+ * @src: address of source memory area
+ * @src_len: length of src operand in bytes
+ *
  * Executes the KM (CIPHER MESSAGE) operation of the CPU.
- * @param func: the function code passed to KM; see crypt_s390_km_func
- * @param param: address of parameter block; see POP for details on each func
- * @param dest: address of destination memory area
- * @param src: address of source memory area
- * @param src_len: length of src operand in bytes
- * @returns < zero for failure, 0 for the query func, number of processed bytes
- * 	for encryption/decryption funcs
+ *
+ * Returns -1 for failure, 0 for the query func, number of processed
+ * bytes for encryption/decryption funcs
  */
-static inline int
-crypt_s390_km(long func, void* param, u8* dest, const u8* src, long src_len)
+static inline int crypt_s390_km(long func, void *param,
+				u8 *dest, const u8 *src, long src_len)
 {
 	register long __func asm("0") = func & CRYPT_S390_FUNC_MASK;
-	register void* __param asm("1") = param;
-	register const u8* __src asm("2") = src;
+	register void *__param asm("1") = param;
+	register const u8 *__src asm("2") = src;
 	register long __src_len asm("3") = src_len;
-	register u8* __dest asm("4") = dest;
+	register u8 *__dest asm("4") = dest;
 	int ret;
 
 	asm volatile(
 		"0:	.insn	rre,0xb92e0000,%3,%1 \n" /* KM opcode */
 		"1:	brc	1,0b \n" /* handle partial completion */
-		"	ahi	%0,%h7\n"
-		"2:	ahi	%0,%h8\n"
-		"3:\n"
-		EX_TABLE(0b,3b) EX_TABLE(1b,2b)
+		"	la	%0,0\n"
+		"2:\n"
+		EX_TABLE(0b,2b) EX_TABLE(1b,2b)
 		: "=d" (ret), "+a" (__src), "+d" (__src_len), "+a" (__dest)
-		: "d" (__func), "a" (__param), "0" (-EFAULT),
-		  "K" (ENOSYS), "K" (-ENOSYS + EFAULT) : "cc", "memory");
+		: "d" (__func), "a" (__param), "0" (-1) : "cc", "memory");
 	if (ret < 0)
 		return ret;
 	return (func & CRYPT_S390_FUNC_MASK) ? src_len - __src_len : __src_len;
 }
 
-/*
+/**
+ * crypt_s390_kmc:
+ * @func: the function code passed to KM; see crypt_s390_kmc_func
+ * @param: address of parameter block; see POP for details on each func
+ * @dest: address of destination memory area
+ * @src: address of source memory area
+ * @src_len: length of src operand in bytes
+ *
  * Executes the KMC (CIPHER MESSAGE WITH CHAINING) operation of the CPU.
- * @param func: the function code passed to KM; see crypt_s390_kmc_func
- * @param param: address of parameter block; see POP for details on each func
- * @param dest: address of destination memory area
- * @param src: address of source memory area
- * @param src_len: length of src operand in bytes
- * @returns < zero for failure, 0 for the query func, number of processed bytes
- * 	for encryption/decryption funcs
+ *
+ * Returns -1 for failure, 0 for the query func, number of processed
+ * bytes for encryption/decryption funcs
  */
-static inline int
-crypt_s390_kmc(long func, void* param, u8* dest, const u8* src, long src_len)
+static inline int crypt_s390_kmc(long func, void *param,
+				 u8 *dest, const u8 *src, long src_len)
 {
 	register long __func asm("0") = func & CRYPT_S390_FUNC_MASK;
-	register void* __param asm("1") = param;
-	register const u8* __src asm("2") = src;
+	register void *__param asm("1") = param;
+	register const u8 *__src asm("2") = src;
 	register long __src_len asm("3") = src_len;
-	register u8* __dest asm("4") = dest;
+	register u8 *__dest asm("4") = dest;
 	int ret;
 
 	asm volatile(
 		"0:	.insn	rre,0xb92f0000,%3,%1 \n" /* KMC opcode */
 		"1:	brc	1,0b \n" /* handle partial completion */
-		"	ahi	%0,%h7\n"
-		"2:	ahi	%0,%h8\n"
-		"3:\n"
-		EX_TABLE(0b,3b) EX_TABLE(1b,2b)
+		"	la	%0,0\n"
+		"2:\n"
+		EX_TABLE(0b,2b) EX_TABLE(1b,2b)
 		: "=d" (ret), "+a" (__src), "+d" (__src_len), "+a" (__dest)
-		: "d" (__func), "a" (__param), "0" (-EFAULT),
-		  "K" (ENOSYS), "K" (-ENOSYS + EFAULT) : "cc", "memory");
+		: "d" (__func), "a" (__param), "0" (-1) : "cc", "memory");
 	if (ret < 0)
 		return ret;
 	return (func & CRYPT_S390_FUNC_MASK) ? src_len - __src_len : __src_len;
 }
 
-/*
+/**
+ * crypt_s390_kimd:
+ * @func: the function code passed to KM; see crypt_s390_kimd_func
+ * @param: address of parameter block; see POP for details on each func
+ * @src: address of source memory area
+ * @src_len: length of src operand in bytes
+ *
  * Executes the KIMD (COMPUTE INTERMEDIATE MESSAGE DIGEST) operation
  * of the CPU.
- * @param func: the function code passed to KM; see crypt_s390_kimd_func
- * @param param: address of parameter block; see POP for details on each func
- * @param src: address of source memory area
- * @param src_len: length of src operand in bytes
- * @returns < zero for failure, 0 for the query func, number of processed bytes
- * 	for digest funcs
+ *
+ * Returns -1 for failure, 0 for the query func, number of processed
+ * bytes for digest funcs
  */
-static inline int
-crypt_s390_kimd(long func, void* param, const u8* src, long src_len)
+static inline int crypt_s390_kimd(long func, void *param,
+				  const u8 *src, long src_len)
 {
 	register long __func asm("0") = func & CRYPT_S390_FUNC_MASK;
-	register void* __param asm("1") = param;
-	register const u8* __src asm("2") = src;
+	register void *__param asm("1") = param;
+	register const u8 *__src asm("2") = src;
 	register long __src_len asm("3") = src_len;
 	int ret;
 
 	asm volatile(
 		"0:	.insn	rre,0xb93e0000,%1,%1 \n" /* KIMD opcode */
 		"1:	brc	1,0b \n" /* handle partial completion */
-		"	ahi	%0,%h6\n"
-		"2:	ahi	%0,%h7\n"
-		"3:\n"
-		EX_TABLE(0b,3b) EX_TABLE(1b,2b)
+		"	la	%0,0\n"
+		"2:\n"
+		EX_TABLE(0b,2b) EX_TABLE(1b,2b)
 		: "=d" (ret), "+a" (__src), "+d" (__src_len)
-		: "d" (__func), "a" (__param), "0" (-EFAULT),
-		  "K" (ENOSYS), "K" (-ENOSYS + EFAULT) : "cc", "memory");
+		: "d" (__func), "a" (__param), "0" (-1) : "cc", "memory");
 	if (ret < 0)
 		return ret;
 	return (func & CRYPT_S390_FUNC_MASK) ? src_len - __src_len : __src_len;
 }
 
-/*
+/**
+ * crypt_s390_klmd:
+ * @func: the function code passed to KM; see crypt_s390_klmd_func
+ * @param: address of parameter block; see POP for details on each func
+ * @src: address of source memory area
+ * @src_len: length of src operand in bytes
+ *
  * Executes the KLMD (COMPUTE LAST MESSAGE DIGEST) operation of the CPU.
- * @param func: the function code passed to KM; see crypt_s390_klmd_func
- * @param param: address of parameter block; see POP for details on each func
- * @param src: address of source memory area
- * @param src_len: length of src operand in bytes
- * @returns < zero for failure, 0 for the query func, number of processed bytes
- * 	for digest funcs
+ *
+ * Returns -1 for failure, 0 for the query func, number of processed
+ * bytes for digest funcs
  */
-static inline int
-crypt_s390_klmd(long func, void* param, const u8* src, long src_len)
+static inline int crypt_s390_klmd(long func, void *param,
+				  const u8 *src, long src_len)
 {
 	register long __func asm("0") = func & CRYPT_S390_FUNC_MASK;
-	register void* __param asm("1") = param;
-	register const u8* __src asm("2") = src;
+	register void *__param asm("1") = param;
+	register const u8 *__src asm("2") = src;
 	register long __src_len asm("3") = src_len;
 	int ret;
 
 	asm volatile(
 		"0:	.insn	rre,0xb93f0000,%1,%1 \n" /* KLMD opcode */
 		"1:	brc	1,0b \n" /* handle partial completion */
-		"	ahi	%0,%h6\n"
-		"2:	ahi	%0,%h7\n"
-		"3:\n"
-		EX_TABLE(0b,3b) EX_TABLE(1b,2b)
+		"	la	%0,0\n"
+		"2:\n"
+		EX_TABLE(0b,2b) EX_TABLE(1b,2b)
 		: "=d" (ret), "+a" (__src), "+d" (__src_len)
-		: "d" (__func), "a" (__param), "0" (-EFAULT),
-		  "K" (ENOSYS), "K" (-ENOSYS + EFAULT) : "cc", "memory");
+		: "d" (__func), "a" (__param), "0" (-1) : "cc", "memory");
 	if (ret < 0)
 		return ret;
 	return (func & CRYPT_S390_FUNC_MASK) ? src_len - __src_len : __src_len;
 }
 
-/*
+/**
+ * crypt_s390_kmac:
+ * @func: the function code passed to KM; see crypt_s390_klmd_func
+ * @param: address of parameter block; see POP for details on each func
+ * @src: address of source memory area
+ * @src_len: length of src operand in bytes
+ *
  * Executes the KMAC (COMPUTE MESSAGE AUTHENTICATION CODE) operation
  * of the CPU.
- * @param func: the function code passed to KM; see crypt_s390_klmd_func
- * @param param: address of parameter block; see POP for details on each func
- * @param src: address of source memory area
- * @param src_len: length of src operand in bytes
- * @returns < zero for failure, 0 for the query func, number of processed bytes
- * 	for digest funcs
+ *
+ * Returns -1 for failure, 0 for the query func, number of processed
+ * bytes for digest funcs
  */
-static inline int
-crypt_s390_kmac(long func, void* param, const u8* src, long src_len)
+static inline int crypt_s390_kmac(long func, void *param,
+				  const u8 *src, long src_len)
 {
 	register long __func asm("0") = func & CRYPT_S390_FUNC_MASK;
-	register void* __param asm("1") = param;
-	register const u8* __src asm("2") = src;
+	register void *__param asm("1") = param;
+	register const u8 *__src asm("2") = src;
 	register long __src_len asm("3") = src_len;
 	int ret;
 
 	asm volatile(
 		"0:	.insn	rre,0xb91e0000,%1,%1 \n" /* KLAC opcode */
 		"1:	brc	1,0b \n" /* handle partial completion */
-		"	ahi	%0,%h6\n"
-		"2:	ahi	%0,%h7\n"
-		"3:\n"
-		EX_TABLE(0b,3b) EX_TABLE(1b,2b)
+		"	la	%0,0\n"
+		"2:\n"
+		EX_TABLE(0b,2b) EX_TABLE(1b,2b)
 		: "=d" (ret), "+a" (__src), "+d" (__src_len)
-		: "d" (__func), "a" (__param), "0" (-EFAULT),
-		  "K" (ENOSYS), "K" (-ENOSYS + EFAULT) : "cc", "memory");
+		: "d" (__func), "a" (__param), "0" (-1) : "cc", "memory");
 	if (ret < 0)
 		return ret;
 	return (func & CRYPT_S390_FUNC_MASK) ? src_len - __src_len : __src_len;
 }
 
 /**
+ * crypt_s390_func_available:
+ * @func: the function code of the specific function; 0 if op in general
+ *
  * Tests if a specific crypto function is implemented on the machine.
- * @param func:	the function code of the specific function; 0 if op in general
- * @return	1 if func available; 0 if func or op in general not available
+ *
+ * Returns 1 if func available; 0 if func or op in general not available
  */
-static inline int
-crypt_s390_func_available(int func)
+static inline int crypt_s390_func_available(int func)
 {
+	unsigned char status[16];
 	int ret;
 
-	struct crypt_s390_query_status status = {
-		.high = 0,
-		.low = 0
-	};
-	switch (func & CRYPT_S390_OP_MASK){
-		case CRYPT_S390_KM:
-			ret = crypt_s390_km(KM_QUERY, &status, NULL, NULL, 0);
-			break;
-		case CRYPT_S390_KMC:
-			ret = crypt_s390_kmc(KMC_QUERY, &status, NULL, NULL, 0);
-			break;
-		case CRYPT_S390_KIMD:
-			ret = crypt_s390_kimd(KIMD_QUERY, &status, NULL, 0);
-			break;
-		case CRYPT_S390_KLMD:
-			ret = crypt_s390_klmd(KLMD_QUERY, &status, NULL, 0);
-			break;
-		case CRYPT_S390_KMAC:
-			ret = crypt_s390_kmac(KMAC_QUERY, &status, NULL, 0);
-			break;
-		default:
-			ret = 0;
-			return ret;
-	}
-	if (ret >= 0){
-		func &= CRYPT_S390_FUNC_MASK;
-		func &= 0x7f; //mask modifier bit
-		if (func < 64){
-			ret = (status.high >> (64 - func - 1)) & 0x1;
-		} else {
-			ret = (status.low >> (128 - func - 1)) & 0x1;
-		}
-	} else {
-		ret = 0;
+	switch (func & CRYPT_S390_OP_MASK) {
+	case CRYPT_S390_KM:
+		ret = crypt_s390_km(KM_QUERY, &status, NULL, NULL, 0);
+		break;
+	case CRYPT_S390_KMC:
+		ret = crypt_s390_kmc(KMC_QUERY, &status, NULL, NULL, 0);
+		break;
+	case CRYPT_S390_KIMD:
+		ret = crypt_s390_kimd(KIMD_QUERY, &status, NULL, 0);
+		break;
+	case CRYPT_S390_KLMD:
+		ret = crypt_s390_klmd(KLMD_QUERY, &status, NULL, 0);
+		break;
+	case CRYPT_S390_KMAC:
+		ret = crypt_s390_kmac(KMAC_QUERY, &status, NULL, 0);
+		break;
+	default:
+		return 0;
 	}
-	return ret;
+	if (ret < 0)
+		return 0;
+	func &= CRYPT_S390_FUNC_MASK;
+	func &= 0x7f;		/* mask modifier bit */
+	return (status[func >> 3] & (0x80 >> (func & 7))) != 0;
 }
 
-#endif // _CRYPTO_ARCH_S390_CRYPT_S390_H
+#endif	/* _CRYPTO_ARCH_S390_CRYPT_S390_H */

arch/s390/crypto/crypt_s390_query.c

@@ -1,129 +0,0 @@
-/*
- * Cryptographic API.
- *
- * Support for s390 cryptographic instructions.
- * Testing module for querying processor crypto capabilities.
- *
- * Copyright (c) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
- * Author(s): Thomas Spatzier (tspat@de.ibm.com)
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License as published by the Free
- * Software Foundation; either version 2 of the License, or (at your option)
- * any later version.
- *
- */
-#include <linux/module.h>
-#include <linux/init.h>
-#include <linux/kernel.h>
-#include <asm/errno.h>
-#include "crypt_s390.h"
-
-static void query_available_functions(void)
-{
-	printk(KERN_INFO "#####################\n");
-
-	/* query available KM functions */
-	printk(KERN_INFO "KM_QUERY: %d\n",
-		crypt_s390_func_available(KM_QUERY));
-	printk(KERN_INFO "KM_DEA: %d\n",
-		crypt_s390_func_available(KM_DEA_ENCRYPT));
-	printk(KERN_INFO "KM_TDEA_128: %d\n",
-		crypt_s390_func_available(KM_TDEA_128_ENCRYPT));
-	printk(KERN_INFO "KM_TDEA_192: %d\n",
-		crypt_s390_func_available(KM_TDEA_192_ENCRYPT));
-	printk(KERN_INFO "KM_AES_128: %d\n",
-		crypt_s390_func_available(KM_AES_128_ENCRYPT));
-	printk(KERN_INFO "KM_AES_192: %d\n",
-		crypt_s390_func_available(KM_AES_192_ENCRYPT));
-	printk(KERN_INFO "KM_AES_256: %d\n",
-		crypt_s390_func_available(KM_AES_256_ENCRYPT));
-
-	/* query available KMC functions */
-	printk(KERN_INFO "KMC_QUERY: %d\n",
-		crypt_s390_func_available(KMC_QUERY));
-	printk(KERN_INFO "KMC_DEA: %d\n",
-		crypt_s390_func_available(KMC_DEA_ENCRYPT));
-	printk(KERN_INFO "KMC_TDEA_128: %d\n",
-		crypt_s390_func_available(KMC_TDEA_128_ENCRYPT));
-	printk(KERN_INFO "KMC_TDEA_192: %d\n",
-		crypt_s390_func_available(KMC_TDEA_192_ENCRYPT));
-	printk(KERN_INFO "KMC_AES_128: %d\n",
-		crypt_s390_func_available(KMC_AES_128_ENCRYPT));
-	printk(KERN_INFO "KMC_AES_192: %d\n",
-		crypt_s390_func_available(KMC_AES_192_ENCRYPT));
-	printk(KERN_INFO "KMC_AES_256: %d\n",
-		crypt_s390_func_available(KMC_AES_256_ENCRYPT));
-
-	/* query available KIMD functions */
-	printk(KERN_INFO "KIMD_QUERY: %d\n",
-		crypt_s390_func_available(KIMD_QUERY));
-	printk(KERN_INFO "KIMD_SHA_1: %d\n",
-		crypt_s390_func_available(KIMD_SHA_1));
-	printk(KERN_INFO "KIMD_SHA_256: %d\n",
-		crypt_s390_func_available(KIMD_SHA_256));
-
-	/* query available KLMD functions */
-	printk(KERN_INFO "KLMD_QUERY: %d\n",
-		crypt_s390_func_available(KLMD_QUERY));
-	printk(KERN_INFO "KLMD_SHA_1: %d\n",
-		crypt_s390_func_available(KLMD_SHA_1));
-	printk(KERN_INFO "KLMD_SHA_256: %d\n",
-		crypt_s390_func_available(KLMD_SHA_256));
-
-	/* query available KMAC functions */
-	printk(KERN_INFO "KMAC_QUERY: %d\n",
-		crypt_s390_func_available(KMAC_QUERY));
-	printk(KERN_INFO "KMAC_DEA: %d\n",
-		crypt_s390_func_available(KMAC_DEA));
-	printk(KERN_INFO "KMAC_TDEA_128: %d\n",
-		crypt_s390_func_available(KMAC_TDEA_128));
-	printk(KERN_INFO "KMAC_TDEA_192: %d\n",
-		crypt_s390_func_available(KMAC_TDEA_192));
-}
-
-static int init(void)
-{
-	struct crypt_s390_query_status status = {
-		.high = 0,
-		.low = 0
-	};
-
-	printk(KERN_INFO "crypt_s390: querying available crypto functions\n");
-	crypt_s390_km(KM_QUERY, &status, NULL, NULL, 0);
-	printk(KERN_INFO "KM:\t%016llx %016llx\n",
-			(unsigned long long) status.high,
-			(unsigned long long) status.low);
-	status.high = status.low = 0;
-	crypt_s390_kmc(KMC_QUERY, &status, NULL, NULL, 0);
-	printk(KERN_INFO "KMC:\t%016llx %016llx\n",
-			(unsigned long long) status.high,
-			(unsigned long long) status.low);
-	status.high = status.low = 0;
-	crypt_s390_kimd(KIMD_QUERY, &status, NULL, 0);
-	printk(KERN_INFO "KIMD:\t%016llx %016llx\n",
-			(unsigned long long) status.high,
-			(unsigned long long) status.low);
-	status.high = status.low = 0;
-	crypt_s390_klmd(KLMD_QUERY, &status, NULL, 0);
-	printk(KERN_INFO "KLMD:\t%016llx %016llx\n",
-			(unsigned long long) status.high,
-			(unsigned long long) status.low);
-	status.high = status.low = 0;
-	crypt_s390_kmac(KMAC_QUERY, &status, NULL, 0);
-	printk(KERN_INFO "KMAC:\t%016llx %016llx\n",
-			(unsigned long long) status.high,
-			(unsigned long long) status.low);
-
-	query_available_functions();
-	return -ECANCELED;
-}
-
-static void __exit cleanup(void)
-{
-}
-
-module_init(init);
-module_exit(cleanup);
-
-MODULE_LICENSE("GPL");

arch/s390/crypto/des_check_key.c

@@ -10,8 +10,9 @@
  * scatterlist interface.  Changed LGPL to GPL per section 3 of the LGPL.
  *
  * s390 Version:
- *   Copyright (C) 2003 IBM Deutschland GmbH, IBM Corporation
- *   Author(s): Thomas Spatzier (tspat@de.ibm.com)
+ *   Copyright IBM Corp. 2003
+ *   Author(s): Thomas Spatzier
+ *		Jan Glauber (jan.glauber@de.ibm.com)
  *
  * Derived from "crypto/des.c"
  *   Copyright (c) 1992 Dana L. How.
@@ -30,6 +31,7 @@
 #include <linux/module.h>
 #include <linux/errno.h>
 #include <linux/crypto.h>
+#include "crypto_des.h"
 
 #define ROR(d,c,o)	((d) = (d) >> (c) | (d) << (o))
 

arch/s390/crypto/des_s390.c

@@ -3,9 +3,9 @@
  *
  * s390 implementation of the DES Cipher Algorithm.
  *
- * Copyright (c) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
- * Author(s): Thomas Spatzier (tspat@de.ibm.com)
- *
+ * Copyright IBM Corp. 2003,2007
+ * Author(s): Thomas Spatzier
+ *	      Jan Glauber (jan.glauber@de.ibm.com)
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -557,7 +557,7 @@ static int init(void)
 	if (!crypt_s390_func_available(KM_DEA_ENCRYPT) ||
 	    !crypt_s390_func_available(KM_TDEA_128_ENCRYPT) ||
 	    !crypt_s390_func_available(KM_TDEA_192_ENCRYPT))
-		return -ENOSYS;
+		return -EOPNOTSUPP;
 
 	ret = crypto_register_alg(&des_alg);
 	if (ret)

arch/s390/crypto/prng.c

@@ -0,0 +1,213 @@
+/*
+ * Copyright IBM Corp. 2006,2007
+ * Author(s): Jan Glauber <jan.glauber@de.ibm.com>
+ * Driver for the s390 pseudo random number generator
+ */
+#include <linux/fs.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/miscdevice.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/random.h>
+#include <asm/debug.h>
+#include <asm/uaccess.h>
+
+#include "crypt_s390.h"
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Jan Glauber <jan.glauber@de.ibm.com>");
+MODULE_DESCRIPTION("s390 PRNG interface");
+
+static int prng_chunk_size = 256;
+module_param(prng_chunk_size, int, S_IRUSR | S_IRGRP | S_IROTH);
+MODULE_PARM_DESC(prng_chunk_size, "PRNG read chunk size in bytes");
+
+static int prng_entropy_limit = 4096;
+module_param(prng_entropy_limit, int, S_IRUSR | S_IRGRP | S_IROTH | S_IWUSR);
+MODULE_PARM_DESC(prng_entropy_limit,
+	"PRNG add entropy after that much bytes were produced");
+
+/*
+ * Any one who considers arithmetical methods of producing random digits is,
+ * of course, in a state of sin. -- John von Neumann
+ */
+
+struct s390_prng_data {
+	unsigned long count; /* how many bytes were produced */
+	char *buf;
+};
+
+static struct s390_prng_data *p;
+
+/* copied from libica, use a non-zero initial parameter block */
+static unsigned char parm_block[32] = {
+0x0F,0x2B,0x8E,0x63,0x8C,0x8E,0xD2,0x52,0x64,0xB7,0xA0,0x7B,0x75,0x28,0xB8,0xF4,
+0x75,0x5F,0xD2,0xA6,0x8D,0x97,0x11,0xFF,0x49,0xD8,0x23,0xF3,0x7E,0x21,0xEC,0xA0,
+};
+
+static int prng_open(struct inode *inode, struct file *file)
+{
+	return nonseekable_open(inode, file);
+}
+
+static void prng_add_entropy(void)
+{
+	__u64 entropy[4];
+	unsigned int i;
+	int ret;
+
+	for (i = 0; i < 16; i++) {
+		ret = crypt_s390_kmc(KMC_PRNG, parm_block, (char *)entropy,
+				     (char *)entropy, sizeof(entropy));
+		BUG_ON(ret < 0 || ret != sizeof(entropy));
+		memcpy(parm_block, entropy, sizeof(entropy));
+	}
+}
+
+static void prng_seed(int nbytes)
+{
+	char buf[16];
+	int i = 0;
+
+	BUG_ON(nbytes > 16);
+	get_random_bytes(buf, nbytes);
+
+	/* Add the entropy */
+	while (nbytes >= 8) {
+		*((__u64 *)parm_block) ^= *((__u64 *)buf+i*8);
+		prng_add_entropy();
+		i += 8;
+		nbytes -= 8;
+	}
+	prng_add_entropy();
+}
+
+static ssize_t prng_read(struct file *file, char __user *ubuf, size_t nbytes,
+			 loff_t *ppos)
+{
+	int chunk, n;
+	int ret = 0;
+	int tmp;
+
+	/* nbytes can be arbitrary long, we spilt it into chunks */
+	while (nbytes) {
+		/* same as in extract_entropy_user in random.c */
+		if (need_resched()) {
+			if (signal_pending(current)) {
+				if (ret == 0)
+					ret = -ERESTARTSYS;
+				break;
+			}
+			schedule();
+		}
+
+		/*
+		 * we lose some random bytes if an attacker issues
+		 * reads < 8 bytes, but we don't care
+		 */
+		chunk = min_t(int, nbytes, prng_chunk_size);
+
+		/* PRNG only likes multiples of 8 bytes */
+		n = (chunk + 7) & -8;
+
+		if (p->count > prng_entropy_limit)
+			prng_seed(8);
+
+		/* if the CPU supports PRNG stckf is present too */
+		asm volatile(".insn     s,0xb27c0000,%0"
+			     : "=m" (*((unsigned long long *)p->buf)) : : "cc");
+
+		/*
+		 * Beside the STCKF the input for the TDES-EDE is the output
+		 * of the last operation. We differ here from X9.17 since we
+		 * only store one timestamp into the buffer. Padding the whole
+		 * buffer with timestamps does not improve security, since
+		 * successive stckf have nearly constant offsets.
+		 * If an attacker knows the first timestamp it would be
+		 * trivial to guess the additional values. One timestamp
+		 * is therefore enough and still guarantees unique input values.
+		 *
+		 * Note: you can still get strict X9.17 conformity by setting
+		 * prng_chunk_size to 8 bytes.
+		*/
+		tmp = crypt_s390_kmc(KMC_PRNG, parm_block, p->buf, p->buf, n);
+		BUG_ON((tmp < 0) || (tmp != n));
+
+		p->count += n;
+
+		if (copy_to_user(ubuf, p->buf, chunk))
+			return -EFAULT;
+
+		nbytes -= chunk;
+		ret += chunk;
+		ubuf += chunk;
+	}
+	return ret;
+}
+
+static struct file_operations prng_fops = {
+	.owner		= THIS_MODULE,
+	.open		= &prng_open,
+	.release	= NULL,
+	.read		= &prng_read,
+};
+
+static struct miscdevice prng_dev = {
+	.name	= "prandom",
+	.minor	= MISC_DYNAMIC_MINOR,
+	.fops	= &prng_fops,
+};
+
+static int __init prng_init(void)
+{
+	int ret;
+
+	/* check if the CPU has a PRNG */
+	if (!crypt_s390_func_available(KMC_PRNG))
+		return -EOPNOTSUPP;
+
+	if (prng_chunk_size < 8)
+		return -EINVAL;
+
+	p = kmalloc(sizeof(struct s390_prng_data), GFP_KERNEL);
+	if (!p)
+		return -ENOMEM;
+	p->count = 0;
+
+	p->buf = kmalloc(prng_chunk_size, GFP_KERNEL);
+	if (!p->buf) {
+		ret = -ENOMEM;
+		goto out_free;
+	}
+
+	/* initialize the PRNG, add 128 bits of entropy */
+	prng_seed(16);
+
+	ret = misc_register(&prng_dev);
+	if (ret) {
+		printk(KERN_WARNING
+		       "Could not register misc device for PRNG.\n");
+		goto out_buf;
+	}
+	return 0;
+
+out_buf:
+	kfree(p->buf);
+out_free:
+	kfree(p);
+	return ret;
+}
+
+static void __exit prng_exit(void)
+{
+	/* wipe me */
+	memset(p->buf, 0, prng_chunk_size);
+	kfree(p->buf);
+	kfree(p);
+
+	misc_deregister(&prng_dev);
+}
+
+module_init(prng_init);
+module_exit(prng_exit);

arch/s390/crypto/sha1_s390.c

@@ -8,8 +8,9 @@
  * implementation written by Steve Reid.
  *
  * s390 Version:
- *   Copyright (C) 2003 IBM Deutschland GmbH, IBM Corporation
- *   Author(s): Thomas Spatzier (tspat@de.ibm.com)
+ *   Copyright IBM Corp. 2003,2007
+ *   Author(s): Thomas Spatzier
+ *		Jan Glauber (jan.glauber@de.ibm.com)
  *
  * Derived from "crypto/sha1.c"
  *   Copyright (c) Alan Smithee.
@@ -43,16 +44,14 @@ struct crypt_s390_sha1_ctx {
 static void sha1_init(struct crypto_tfm *tfm)
 {
 	struct crypt_s390_sha1_ctx *ctx = crypto_tfm_ctx(tfm);
-	static const u32 initstate[5] = {
-		0x67452301,
-		0xEFCDAB89,
-		0x98BADCFE,
-		0x10325476,
-		0xC3D2E1F0
-	};
+
+	ctx->state[0] = 0x67452301;
+	ctx->state[1] =	0xEFCDAB89;
+	ctx->state[2] =	0x98BADCFE;
+	ctx->state[3] = 0x10325476;
+	ctx->state[4] =	0xC3D2E1F0;
 
 	ctx->count = 0;
-	memcpy(ctx->state, &initstate, sizeof(initstate));
 	ctx->buf_len = 0;
 }
 
@@ -63,13 +62,13 @@ static void sha1_update(struct crypto_tfm *tfm, const u8 *data,
 	long imd_len;
 
 	sctx = crypto_tfm_ctx(tfm);
-	sctx->count += len * 8; //message bit length
+	sctx->count += len * 8; /* message bit length */
 
-	//anything in buffer yet? -> must be completed
+	/* anything in buffer yet? -> must be completed */
 	if (sctx->buf_len && (sctx->buf_len + len) >= SHA1_BLOCK_SIZE) {
-		//complete full block and hash
+		/* complete full block and hash */
 		memcpy(sctx->buffer + sctx->buf_len, data,
-				SHA1_BLOCK_SIZE - sctx->buf_len);
+		       SHA1_BLOCK_SIZE - sctx->buf_len);
 		crypt_s390_kimd(KIMD_SHA_1, sctx->state, sctx->buffer,
 				SHA1_BLOCK_SIZE);
 		data += SHA1_BLOCK_SIZE - sctx->buf_len;
@@ -77,37 +76,36 @@ static void sha1_update(struct crypto_tfm *tfm, const u8 *data,
 		sctx->buf_len = 0;
 	}
 
-	//rest of data contains full blocks?
+	/* rest of data contains full blocks? */
 	imd_len = len & ~0x3ful;
-	if (imd_len){
+	if (imd_len) {
 		crypt_s390_kimd(KIMD_SHA_1, sctx->state, data, imd_len);
 		data += imd_len;
 		len -= imd_len;
 	}
-	//anything left? store in buffer
-	if (len){
+	/* anything left? store in buffer */
+	if (len) {
 		memcpy(sctx->buffer + sctx->buf_len , data, len);
 		sctx->buf_len += len;
 	}
 }
 
 
-static void
-pad_message(struct crypt_s390_sha1_ctx* sctx)
+static void pad_message(struct crypt_s390_sha1_ctx* sctx)
 {
 	int index;
 
 	index = sctx->buf_len;
-	sctx->buf_len = (sctx->buf_len < 56)?
-		SHA1_BLOCK_SIZE:2 * SHA1_BLOCK_SIZE;
-	//start pad with 1
+	sctx->buf_len = (sctx->buf_len < 56) ?
+			 SHA1_BLOCK_SIZE:2 * SHA1_BLOCK_SIZE;
+	/* start pad with 1 */
 	sctx->buffer[index] = 0x80;
-	//pad with zeros
+	/* pad with zeros */
 	index++;
 	memset(sctx->buffer + index, 0x00, sctx->buf_len - index);
-	//append length
+	/* append length */
 	memcpy(sctx->buffer + sctx->buf_len - 8, &sctx->count,
-			sizeof sctx->count);
+	       sizeof sctx->count);
 }
 
 /* Add padding and return the message digest. */
@@ -115,47 +113,40 @@ static void sha1_final(struct crypto_tfm *tfm, u8 *out)
 {
 	struct crypt_s390_sha1_ctx *sctx = crypto_tfm_ctx(tfm);
 
-	//must perform manual padding
+	/* must perform manual padding */
 	pad_message(sctx);
 	crypt_s390_kimd(KIMD_SHA_1, sctx->state, sctx->buffer, sctx->buf_len);
-	//copy digest to out
+	/* copy digest to out */
 	memcpy(out, sctx->state, SHA1_DIGEST_SIZE);
-	/* Wipe context */
+	/* wipe context */
 	memset(sctx, 0, sizeof *sctx);
 }
 
 static struct crypto_alg alg = {
 	.cra_name	=	"sha1",
-	.cra_driver_name =	"sha1-s390",
+	.cra_driver_name=	"sha1-s390",
 	.cra_priority	=	CRYPT_S390_PRIORITY,
 	.cra_flags	=	CRYPTO_ALG_TYPE_DIGEST,
 	.cra_blocksize	=	SHA1_BLOCK_SIZE,
 	.cra_ctxsize	=	sizeof(struct crypt_s390_sha1_ctx),
 	.cra_module	=	THIS_MODULE,
-	.cra_list       =       LIST_HEAD_INIT(alg.cra_list),
+	.cra_list	=	LIST_HEAD_INIT(alg.cra_list),
 	.cra_u		=	{ .digest = {
 	.dia_digestsize	=	SHA1_DIGEST_SIZE,
-	.dia_init   	= 	sha1_init,
-	.dia_update 	=	sha1_update,
-	.dia_final  	=	sha1_final } }
+	.dia_init	=	sha1_init,
+	.dia_update	=	sha1_update,
+	.dia_final	=	sha1_final } }
 };
 
-static int
-init(void)
+static int __init init(void)
 {
-	int ret = -ENOSYS;
+	if (!crypt_s390_func_available(KIMD_SHA_1))
+		return -EOPNOTSUPP;
 
-	if (crypt_s390_func_available(KIMD_SHA_1)){
-		ret = crypto_register_alg(&alg);
-		if (ret == 0){
-			printk(KERN_INFO "crypt_s390: sha1_s390 loaded.\n");
-		}
-	}
-	return ret;
+	return crypto_register_alg(&alg);
 }
 
-static void __exit
-fini(void)
+static void __exit fini(void)
 {
 	crypto_unregister_alg(&alg);
 }

arch/s390/crypto/sha256_s390.c

@@ -4,7 +4,7 @@
  * s390 implementation of the SHA256 Secure Hash Algorithm.
  *
  * s390 Version:
- *   Copyright (C) 2005 IBM Deutschland GmbH, IBM Corporation
+ *   Copyright IBM Corp. 2005,2007
  *   Author(s): Jan Glauber (jang@de.ibm.com)
  *
  * Derived from "crypto/sha256.c"
@@ -143,15 +143,10 @@ static struct crypto_alg alg = {
 
 static int init(void)
 {
-	int ret;
-
 	if (!crypt_s390_func_available(KIMD_SHA_256))
-		return -ENOSYS;
+		return -EOPNOTSUPP;
 
-	ret = crypto_register_alg(&alg);
-	if (ret != 0)
-		printk(KERN_INFO "crypt_s390: sha256_s390 couldn't be loaded.");
-	return ret;
+	return crypto_register_alg(&alg);
 }
 
 static void __exit fini(void)

arch/s390/defconfig

@@ -108,6 +108,8 @@ CONFIG_DEFAULT_MIGRATION_COST=1000000
 CONFIG_COMPAT=y
 CONFIG_SYSVIPC_COMPAT=y
 CONFIG_AUDIT_ARCH=y
+CONFIG_S390_SWITCH_AMODE=y
+CONFIG_S390_EXEC_PROTECT=y
 
 #
 # Code generation options
@@ -431,7 +433,6 @@ CONFIG_TN3270_CONSOLE=y
 CONFIG_TN3215=y
 CONFIG_TN3215_CONSOLE=y
 CONFIG_CCW_CONSOLE=y
-CONFIG_SCLP=y
 CONFIG_SCLP_TTY=y
 CONFIG_SCLP_CONSOLE=y
 CONFIG_SCLP_VT220_TTY=y
@@ -724,9 +725,7 @@ CONFIG_CRYPTO_MANAGER=y
 # CONFIG_CRYPTO_MD4 is not set
 # CONFIG_CRYPTO_MD5 is not set
 # CONFIG_CRYPTO_SHA1 is not set
-# CONFIG_CRYPTO_SHA1_S390 is not set
 # CONFIG_CRYPTO_SHA256 is not set
-# CONFIG_CRYPTO_SHA256_S390 is not set
 # CONFIG_CRYPTO_SHA512 is not set
 # CONFIG_CRYPTO_WP512 is not set
 # CONFIG_CRYPTO_TGR192 is not set
@@ -735,12 +734,10 @@ CONFIG_CRYPTO_ECB=m
 CONFIG_CRYPTO_CBC=y
 # CONFIG_CRYPTO_LRW is not set
 # CONFIG_CRYPTO_DES is not set
-# CONFIG_CRYPTO_DES_S390 is not set
 # CONFIG_CRYPTO_BLOWFISH is not set
 # CONFIG_CRYPTO_TWOFISH is not set
 # CONFIG_CRYPTO_SERPENT is not set
 # CONFIG_CRYPTO_AES is not set
-# CONFIG_CRYPTO_AES_S390 is not set
 # CONFIG_CRYPTO_CAST5 is not set
 # CONFIG_CRYPTO_CAST6 is not set
 # CONFIG_CRYPTO_TEA is not set
@@ -755,6 +752,11 @@ CONFIG_CRYPTO_CBC=y
 #
 # Hardware crypto devices
 #
+# CONFIG_CRYPTO_SHA1_S390 is not set
+# CONFIG_CRYPTO_SHA256_S390 is not set
+# CONFIG_CRYPTO_DES_S390 is not set
+# CONFIG_CRYPTO_AES_S390 is not set
+CONFIG_S390_PRNG=m
 
 #
 # Library routines

arch/s390/hypfs/Makefile

@@ -4,4 +4,4 @@
 
 obj-$(CONFIG_S390_HYPFS_FS) += s390_hypfs.o
 
-s390_hypfs-objs := inode.o hypfs_diag.o
+s390_hypfs-objs := inode.o hypfs_diag.o hypfs_vm.o

arch/s390/hypfs/hypfs.h

@@ -27,4 +27,13 @@ extern struct dentry *hypfs_create_str(struct super_block *sb,
 				       struct dentry *dir, const char *name,
 				       char *string);
 
+/* LPAR Hypervisor */
+extern int hypfs_diag_init(void);
+extern void hypfs_diag_exit(void);
+extern int hypfs_diag_create_files(struct super_block *sb, struct dentry *root);
+
+/* VM Hypervisor */
+extern int hypfs_vm_init(void);
+extern int hypfs_vm_create_files(struct super_block *sb, struct dentry *root);
+
 #endif /* _HYPFS_H_ */

arch/s390/hypfs/hypfs_diag.h

@@ -1,16 +0,0 @@
-/*
- *  arch/s390/hypfs_diag.h
- *    Hypervisor filesystem for Linux on s390.
- *
- *    Copyright (C) IBM Corp. 2006
- *    Author(s): Michael Holzheu <holzheu@de.ibm.com>
- */
-
-#ifndef _HYPFS_DIAG_H_
-#define _HYPFS_DIAG_H_
-
-extern int hypfs_diag_init(void);
-extern void hypfs_diag_exit(void);
-extern int hypfs_diag_create_files(struct super_block *sb, struct dentry *root);
-
-#endif /* _HYPFS_DIAG_H_ */

arch/s390/hypfs/hypfs_vm.c

@@ -0,0 +1,231 @@
+/*
+ *    Hypervisor filesystem for Linux on s390. z/VM implementation.
+ *
+ *    Copyright (C) IBM Corp. 2006
+ *    Author(s): Michael Holzheu <holzheu@de.ibm.com>
+ */
+
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/vmalloc.h>
+#include <asm/ebcdic.h>
+#include "hypfs.h"
+
+#define NAME_LEN 8
+
+static char local_guest[] = "        ";
+static char all_guests[] = "*       ";
+static char *guest_query;
+
+struct diag2fc_data {
+	__u32 version;
+	__u32 flags;
+	__u64 used_cpu;
+	__u64 el_time;
+	__u64 mem_min_kb;
+	__u64 mem_max_kb;
+	__u64 mem_share_kb;
+	__u64 mem_used_kb;
+	__u32 pcpus;
+	__u32 lcpus;
+	__u32 vcpus;
+	__u32 cpu_min;
+	__u32 cpu_max;
+	__u32 cpu_shares;
+	__u32 cpu_use_samp;
+	__u32 cpu_delay_samp;
+	__u32 page_wait_samp;
+	__u32 idle_samp;
+	__u32 other_samp;
+	__u32 total_samp;
+	char  guest_name[NAME_LEN];
+};
+
+struct diag2fc_parm_list {
+	char userid[NAME_LEN];
+	char aci_grp[NAME_LEN];
+	__u64 addr;
+	__u32 size;
+	__u32 fmt;
+};
+
+static int diag2fc(int size, char* query, void *addr)
+{
+	unsigned long residual_cnt;
+	unsigned long rc;
+	struct diag2fc_parm_list parm_list;
+
+	memcpy(parm_list.userid, query, NAME_LEN);
+	ASCEBC(parm_list.userid, NAME_LEN);
+	parm_list.addr = (unsigned long) addr ;
+	parm_list.size = size;
+	parm_list.fmt = 0x02;
+	memset(parm_list.aci_grp, 0x40, NAME_LEN);
+	rc = -1;
+
+	asm volatile(
+		"	diag    %0,%1,0x2fc\n"
+		"0:\n"
+		EX_TABLE(0b,0b)
+		: "=d" (residual_cnt), "+d" (rc) : "0" (&parm_list) : "memory");
+
+	if ((rc != 0 ) && (rc != -2))
+		return rc;
+	else
+		return -residual_cnt;
+}
+
+static struct diag2fc_data *diag2fc_store(char *query, int *count)
+{
+	int size;
+	struct diag2fc_data *data;
+
+	do {
+		size = diag2fc(0, query, NULL);
+		if (size < 0)
+			return ERR_PTR(-EACCES);
+		data = vmalloc(size);
+		if (!data)
+			return ERR_PTR(-ENOMEM);
+		if (diag2fc(size, query, data) == 0)
+			break;
+		vfree(data);
+	} while (1);
+	*count = (size / sizeof(*data));
+
+	return data;
+}
+
+static void diag2fc_free(void *data)
+{
+	vfree(data);
+}
+
+#define ATTRIBUTE(sb, dir, name, member) \
+do { \
+	void *rc; \
+	rc = hypfs_create_u64(sb, dir, name, member); \
+	if (IS_ERR(rc)) \
+		return PTR_ERR(rc); \
+} while(0)
+
+static int hpyfs_vm_create_guest(struct super_block *sb,
+				 struct dentry *systems_dir,
+				 struct diag2fc_data *data)
+{
+	char guest_name[NAME_LEN + 1] = {};
+	struct dentry *guest_dir, *cpus_dir, *samples_dir, *mem_dir;
+	int dedicated_flag, capped_value;
+
+	capped_value = (data->flags & 0x00000006) >> 1;
+	dedicated_flag = (data->flags & 0x00000008) >> 3;
+
+	/* guest dir */
+	memcpy(guest_name, data->guest_name, NAME_LEN);
+	EBCASC(guest_name, NAME_LEN);
+	strstrip(guest_name);
+	guest_dir = hypfs_mkdir(sb, systems_dir, guest_name);
+	if (IS_ERR(guest_dir))
+		return PTR_ERR(guest_dir);
+	ATTRIBUTE(sb, guest_dir, "onlinetime_us", data->el_time);
+
+	/* logical cpu information */
+	cpus_dir = hypfs_mkdir(sb, guest_dir, "cpus");
+	if (IS_ERR(cpus_dir))
+		return PTR_ERR(cpus_dir);
+	ATTRIBUTE(sb, cpus_dir, "cputime_us", data->used_cpu);
+	ATTRIBUTE(sb, cpus_dir, "capped", capped_value);
+	ATTRIBUTE(sb, cpus_dir, "dedicated", dedicated_flag);
+	ATTRIBUTE(sb, cpus_dir, "count", data->vcpus);
+	ATTRIBUTE(sb, cpus_dir, "weight_min", data->cpu_min);
+	ATTRIBUTE(sb, cpus_dir, "weight_max", data->cpu_max);
+	ATTRIBUTE(sb, cpus_dir, "weight_cur", data->cpu_shares);
+
+	/* memory information */
+	mem_dir = hypfs_mkdir(sb, guest_dir, "mem");
+	if (IS_ERR(mem_dir))
+		return PTR_ERR(mem_dir);
+	ATTRIBUTE(sb, mem_dir, "min_KiB", data->mem_min_kb);
+	ATTRIBUTE(sb, mem_dir, "max_KiB", data->mem_max_kb);
+	ATTRIBUTE(sb, mem_dir, "used_KiB", data->mem_used_kb);
+	ATTRIBUTE(sb, mem_dir, "share_KiB", data->mem_share_kb);
+
+	/* samples */
+	samples_dir = hypfs_mkdir(sb, guest_dir, "samples");
+	if (IS_ERR(samples_dir))
+		return PTR_ERR(samples_dir);
+	ATTRIBUTE(sb, samples_dir, "cpu_using", data->cpu_use_samp);
+	ATTRIBUTE(sb, samples_dir, "cpu_delay", data->cpu_delay_samp);
+	ATTRIBUTE(sb, samples_dir, "mem_delay", data->page_wait_samp);
+	ATTRIBUTE(sb, samples_dir, "idle", data->idle_samp);
+	ATTRIBUTE(sb, samples_dir, "other", data->other_samp);
+	ATTRIBUTE(sb, samples_dir, "total", data->total_samp);
+	return 0;
+}
+
+int hypfs_vm_create_files(struct super_block *sb, struct dentry *root)
+{
+	struct dentry *dir, *file;
+	struct diag2fc_data *data;
+	int rc, i, count = 0;
+
+	data = diag2fc_store(guest_query, &count);
+	if (IS_ERR(data))
+		return PTR_ERR(data);
+
+	/* Hpervisor Info */
+	dir = hypfs_mkdir(sb, root, "hyp");
+	if (IS_ERR(dir)) {
+		rc = PTR_ERR(dir);
+		goto failed;
+	}
+	file = hypfs_create_str(sb, dir, "type", "z/VM Hypervisor");
+	if (IS_ERR(file)) {
+		rc = PTR_ERR(file);
+		goto failed;
+	}
+
+	/* physical cpus */
+	dir = hypfs_mkdir(sb, root, "cpus");
+	if (IS_ERR(dir)) {
+		rc = PTR_ERR(dir);
+		goto failed;
+	}
+	file = hypfs_create_u64(sb, dir, "count", data->lcpus);
+	if (IS_ERR(file)) {
+		rc = PTR_ERR(file);
+		goto failed;
+	}
+
+	/* guests */
+	dir = hypfs_mkdir(sb, root, "systems");
+	if (IS_ERR(dir)) {
+		rc = PTR_ERR(dir);
+		goto failed;
+	}
+
+	for (i = 0; i < count; i++) {
+		rc = hpyfs_vm_create_guest(sb, dir, &(data[i]));
+		if (rc)
+			goto failed;
+	}
+	diag2fc_free(data);
+	return 0;
+
+failed:
+	diag2fc_free(data);
+	return rc;
+}
+
+int hypfs_vm_init(void)
+{
+	if (diag2fc(0, all_guests, NULL) > 0)
+		guest_query = all_guests;
+	else if (diag2fc(0, local_guest, NULL) > 0)
+		guest_query = local_guest;
+	else
+		return -EACCES;
+
+	return 0;
+}

arch/s390/hypfs/inode.c

@@ -19,7 +19,6 @@
 #include <linux/module.h>
 #include <asm/ebcdic.h>
 #include "hypfs.h"
-#include "hypfs_diag.h"
 
 #define HYPFS_MAGIC 0x687970	/* ASCII 'hyp' */
 #define TMP_SIZE 64		/* size of temporary buffers */
@@ -192,7 +191,10 @@ static ssize_t hypfs_aio_write(struct kiocb *iocb, const struct iovec *iov,
 		goto out;
 	}
 	hypfs_delete_tree(sb->s_root);
-	rc = hypfs_diag_create_files(sb, sb->s_root);
+	if (MACHINE_IS_VM)
+		rc = hypfs_vm_create_files(sb, sb->s_root);
+	else
+		rc = hypfs_diag_create_files(sb, sb->s_root);
 	if (rc) {
 		printk(KERN_ERR "hypfs: Update failed\n");
 		hypfs_delete_tree(sb->s_root);
@@ -289,7 +291,10 @@ static int hypfs_fill_super(struct super_block *sb, void *data, int silent)
 		rc = -ENOMEM;
 		goto err_alloc;
 	}
-	rc = hypfs_diag_create_files(sb, root_dentry);
+	if (MACHINE_IS_VM)
+		rc = hypfs_vm_create_files(sb, root_dentry);
+	else
+		rc = hypfs_diag_create_files(sb, root_dentry);
 	if (rc)
 		goto err_tree;
 	sbi->update_file = hypfs_create_update_file(sb, root_dentry);
@@ -462,11 +467,15 @@ static int __init hypfs_init(void)
 {
 	int rc;
 
-	if (MACHINE_IS_VM)
-		return -ENODATA;
-	if (hypfs_diag_init()) {
-		rc = -ENODATA;
-		goto fail_diag;
+	if (MACHINE_IS_VM) {
+		if (hypfs_vm_init())
+			/* no diag 2fc, just exit */
+			return -ENODATA;
+	} else {
+		if (hypfs_diag_init()) {
+			rc = -ENODATA;
+			goto fail_diag;
+		}
 	}
 	kset_set_kset_s(&s390_subsys, hypervisor_subsys);
 	rc = subsystem_register(&s390_subsys);
@@ -480,7 +489,8 @@ static int __init hypfs_init(void)
 fail_filesystem:
 	subsystem_unregister(&s390_subsys);
 fail_sysfs:
-	hypfs_diag_exit();
+	if (!MACHINE_IS_VM)
+		hypfs_diag_exit();
 fail_diag:
 	printk(KERN_ERR "hypfs: Initialization failed with rc = %i.\n", rc);
 	return rc;
@@ -488,7 +498,8 @@ fail_diag:
 
 static void __exit hypfs_exit(void)
 {
-	hypfs_diag_exit();
+	if (!MACHINE_IS_VM)
+		hypfs_diag_exit();
 	unregister_filesystem(&hypfs_type);
 	subsystem_unregister(&s390_subsys);
 }

arch/s390/kernel/Makefile

@@ -4,9 +4,9 @@
 
 EXTRA_AFLAGS	:= -traditional
 
-obj-y	:=  bitmap.o traps.o time.o process.o reset.o \
+obj-y	:=  bitmap.o traps.o time.o process.o base.o early.o \
             setup.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o \
-	    semaphore.o s390_ext.o debug.o profile.o irq.o ipl.o
+	    semaphore.o s390_ext.o debug.o irq.o ipl.o
 
 obj-y	+= $(if $(CONFIG_64BIT),entry64.o,entry.o)
 obj-y	+= $(if $(CONFIG_64BIT),reipl64.o,reipl.o)

arch/s390/kernel/base.S

@@ -0,0 +1,150 @@
+/*
+ *  arch/s390/kernel/base.S
+ *
+ *    Copyright IBM Corp. 2006,2007
+ *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ *		 Michael Holzheu <holzheu@de.ibm.com>
+ */
+
+#include <asm/ptrace.h>
+#include <asm/lowcore.h>
+
+#ifdef CONFIG_64BIT
+
+	.globl	s390_base_mcck_handler
+s390_base_mcck_handler:
+	basr	%r13,0
+0:	lg	%r15,__LC_PANIC_STACK	# load panic stack
+	aghi	%r15,-STACK_FRAME_OVERHEAD
+	larl	%r1,s390_base_mcck_handler_fn
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	la	%r1,4095
+	lmg	%r0,%r15,__LC_GPREGS_SAVE_AREA-4095(%r1)
+	lpswe	__LC_MCK_OLD_PSW
+
+	.section .bss
+	.globl	s390_base_mcck_handler_fn
+s390_base_mcck_handler_fn:
+	.quad	0
+	.previous
+
+	.globl	s390_base_ext_handler
+s390_base_ext_handler:
+	stmg	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	aghi	%r15,-STACK_FRAME_OVERHEAD
+	larl	%r1,s390_base_ext_handler_fn
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	lmg	%r0,%r15,__LC_SAVE_AREA
+	ni	__LC_EXT_OLD_PSW+1,0xfd	# clear wait state bit
+	lpswe	__LC_EXT_OLD_PSW
+
+	.section .bss
+	.globl s390_base_ext_handler_fn
+s390_base_ext_handler_fn:
+	.quad	0
+	.previous
+
+	.globl	s390_base_pgm_handler
+s390_base_pgm_handler:
+	stmg	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	aghi	%r15,-STACK_FRAME_OVERHEAD
+	larl	%r1,s390_base_pgm_handler_fn
+	lg	%r1,0(%r1)
+	ltgr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+	lmg	%r0,%r15,__LC_SAVE_AREA
+	lpswe	__LC_PGM_OLD_PSW
+1:	lpswe	disabled_wait_psw-0b(%r13)
+
+	.align	8
+disabled_wait_psw:
+	.quad	0x0002000180000000,0x0000000000000000 + s390_base_pgm_handler
+
+	.section .bss
+	.globl s390_base_pgm_handler_fn
+s390_base_pgm_handler_fn:
+	.quad	0
+	.previous
+
+#else /* CONFIG_64BIT */
+
+	.globl	s390_base_mcck_handler
+s390_base_mcck_handler:
+	basr	%r13,0
+0:	l	%r15,__LC_PANIC_STACK	# load panic stack
+	ahi	%r15,-STACK_FRAME_OVERHEAD
+	l	%r1,2f-0b(%r13)
+	l	%r1,0(%r1)
+	ltr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	lm	%r0,%r15,__LC_GPREGS_SAVE_AREA
+	lpsw	__LC_MCK_OLD_PSW
+
+2:	.long	s390_base_mcck_handler_fn
+
+	.section .bss
+	.globl	s390_base_mcck_handler_fn
+s390_base_mcck_handler_fn:
+	.long	0
+	.previous
+
+	.globl	s390_base_ext_handler
+s390_base_ext_handler:
+	stm	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	ahi	%r15,-STACK_FRAME_OVERHEAD
+	l	%r1,2f-0b(%r13)
+	l	%r1,0(%r1)
+	ltr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+1:	lm	%r0,%r15,__LC_SAVE_AREA
+	ni	__LC_EXT_OLD_PSW+1,0xfd	# clear wait state bit
+	lpsw	__LC_EXT_OLD_PSW
+
+2:	.long	s390_base_ext_handler_fn
+
+	.section .bss
+	.globl	s390_base_ext_handler_fn
+s390_base_ext_handler_fn:
+	.long	0
+	.previous
+
+	.globl	s390_base_pgm_handler
+s390_base_pgm_handler:
+	stm	%r0,%r15,__LC_SAVE_AREA
+	basr	%r13,0
+0:	ahi	%r15,-STACK_FRAME_OVERHEAD
+	l	%r1,2f-0b(%r13)
+	l	%r1,0(%r1)
+	ltr	%r1,%r1
+	jz	1f
+	basr	%r14,%r1
+	lm	%r0,%r15,__LC_SAVE_AREA
+	lpsw	__LC_PGM_OLD_PSW
+
+1:	lpsw	disabled_wait_psw-0b(%r13)
+
+2:	.long	s390_base_pgm_handler_fn
+
+disabled_wait_psw:
+	.align	8
+	.long	0x000a0000,0x00000000 + s390_base_pgm_handler
+
+	.section .bss
+	.globl	s390_base_pgm_handler_fn
+s390_base_pgm_handler_fn:
+	.long	0
+	.previous
+
+#endif /* CONFIG_64BIT */

arch/s390/kernel/binfmt_elf32.c

@@ -192,7 +192,7 @@ MODULE_AUTHOR("Gerhard Tonn <ton@de.ibm.com>");
 
 #undef cputime_to_timeval
 #define cputime_to_timeval cputime_to_compat_timeval
-static __inline__ void
+static inline void
 cputime_to_compat_timeval(const cputime_t cputime, struct compat_timeval *value)
 {
 	value->tv_usec = cputime % 1000000;

arch/s390/kernel/compat_exec_domain.c

@@ -12,10 +12,9 @@
 #include <linux/personality.h>
 #include <linux/sched.h>
 
-struct exec_domain s390_exec_domain;
+static struct exec_domain s390_exec_domain;
 
-static int __init
-s390_init (void)
+static int __init s390_init (void)
 {
 	s390_exec_domain.name = "Linux/s390";
 	s390_exec_domain.handler = NULL;

arch/s390/kernel/compat_linux.c

@@ -69,6 +69,12 @@
 
 #include "compat_linux.h"
 
+long psw_user32_bits	= (PSW_BASE32_BITS | PSW_MASK_DAT | PSW_ASC_HOME |
+			   PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			   PSW_MASK_PSTATE | PSW_DEFAULT_KEY);
+long psw32_user_bits	= (PSW32_BASE_BITS | PSW32_MASK_DAT | PSW32_ASC_HOME |
+			   PSW32_MASK_IO | PSW32_MASK_EXT | PSW32_MASK_MCHECK |
+			   PSW32_MASK_PSTATE);
  
 /* For this source file, we want overflow handling. */
 
@@ -416,7 +422,7 @@ asmlinkage long sys32_sysinfo(struct sysinfo32 __user *info)
 	mm_segment_t old_fs = get_fs ();
 	
 	set_fs (KERNEL_DS);
-	ret = sys_sysinfo((struct sysinfo __user *) &s);
+	ret = sys_sysinfo((struct sysinfo __force __user *) &s);
 	set_fs (old_fs);
 	err = put_user (s.uptime, &info->uptime);
 	err |= __put_user (s.loads[0], &info->loads[0]);
@@ -445,7 +451,8 @@ asmlinkage long sys32_sched_rr_get_interval(compat_pid_t pid,
 	mm_segment_t old_fs = get_fs ();
 	
 	set_fs (KERNEL_DS);
-	ret = sys_sched_rr_get_interval(pid, (struct timespec __user *) &t);
+	ret = sys_sched_rr_get_interval(pid,
+					(struct timespec __force __user *) &t);
 	set_fs (old_fs);
 	if (put_compat_timespec(&t, interval))
 		return -EFAULT;
@@ -472,8 +479,8 @@ asmlinkage long sys32_rt_sigprocmask(int how, compat_sigset_t __user *set,
 	}
 	set_fs (KERNEL_DS);
 	ret = sys_rt_sigprocmask(how,
-				 set ? (sigset_t __user *) &s : NULL,
-				 oset ? (sigset_t __user *) &s : NULL,
+				 set ? (sigset_t __force __user *) &s : NULL,
+				 oset ? (sigset_t __force __user *) &s : NULL,
 				 sigsetsize);
 	set_fs (old_fs);
 	if (ret) return ret;
@@ -499,7 +506,7 @@ asmlinkage long sys32_rt_sigpending(compat_sigset_t __user *set,
 	mm_segment_t old_fs = get_fs();
 		
 	set_fs (KERNEL_DS);
-	ret = sys_rt_sigpending((sigset_t __user *) &s, sigsetsize);
+	ret = sys_rt_sigpending((sigset_t __force __user *) &s, sigsetsize);
 	set_fs (old_fs);
 	if (!ret) {
 		switch (_NSIG_WORDS) {
@@ -524,7 +531,7 @@ sys32_rt_sigqueueinfo(int pid, int sig, compat_siginfo_t __user *uinfo)
 	if (copy_siginfo_from_user32(&info, uinfo))
 		return -EFAULT;
 	set_fs (KERNEL_DS);
-	ret = sys_rt_sigqueueinfo(pid, sig, (siginfo_t __user *) &info);
+	ret = sys_rt_sigqueueinfo(pid, sig, (siginfo_t __force __user *) &info);
 	set_fs (old_fs);
 	return ret;
 }
@@ -682,7 +689,7 @@ asmlinkage long sys32_sendfile(int out_fd, int in_fd, compat_off_t __user *offse
 		
 	set_fs(KERNEL_DS);
 	ret = sys_sendfile(out_fd, in_fd,
-			   offset ? (off_t __user *) &of : NULL, count);
+			   offset ? (off_t __force __user *) &of : NULL, count);
 	set_fs(old_fs);
 	
 	if (offset && put_user(of, offset))
@@ -703,7 +710,8 @@ asmlinkage long sys32_sendfile64(int out_fd, int in_fd,
 		
 	set_fs(KERNEL_DS);
 	ret = sys_sendfile64(out_fd, in_fd,
-			     offset ? (loff_t __user *) &lof : NULL, count);
+			     offset ? (loff_t __force __user *) &lof : NULL,
+			     count);
 	set_fs(old_fs);
 	
 	if (offset && put_user(lof, offset))

arch/s390/kernel/compat_linux.h

@@ -115,37 +115,6 @@ typedef struct
         __u32	addr;
 } _psw_t32 __attribute__ ((aligned(8)));
 
-#define PSW32_MASK_PER		0x40000000UL
-#define PSW32_MASK_DAT		0x04000000UL
-#define PSW32_MASK_IO		0x02000000UL
-#define PSW32_MASK_EXT		0x01000000UL
-#define PSW32_MASK_KEY		0x00F00000UL
-#define PSW32_MASK_MCHECK	0x00040000UL
-#define PSW32_MASK_WAIT		0x00020000UL
-#define PSW32_MASK_PSTATE	0x00010000UL
-#define PSW32_MASK_ASC		0x0000C000UL
-#define PSW32_MASK_CC		0x00003000UL
-#define PSW32_MASK_PM		0x00000f00UL
-
-#define PSW32_ADDR_AMODE31	0x80000000UL
-#define PSW32_ADDR_INSN		0x7FFFFFFFUL
-
-#define PSW32_BASE_BITS		0x00080000UL
-
-#define PSW32_ASC_PRIMARY	0x00000000UL
-#define PSW32_ASC_ACCREG	0x00004000UL
-#define PSW32_ASC_SECONDARY	0x00008000UL
-#define PSW32_ASC_HOME		0x0000C000UL
-
-#define PSW32_USER_BITS	(PSW32_BASE_BITS | PSW32_MASK_DAT | PSW32_ASC_HOME | \
-			 PSW32_MASK_IO | PSW32_MASK_EXT | PSW32_MASK_MCHECK | \
-			 PSW32_MASK_PSTATE)
-
-#define PSW32_MASK_MERGE(CURRENT,NEW) \
-        (((CURRENT) & ~(PSW32_MASK_CC|PSW32_MASK_PM)) | \
-         ((NEW) & (PSW32_MASK_CC|PSW32_MASK_PM)))
-
-
 typedef struct
 {
 	_psw_t32	psw;

arch/s390/kernel/compat_signal.c

@@ -275,8 +275,8 @@ sys32_sigaltstack(const stack_t32 __user *uss, stack_t32 __user *uoss,
 	}
 
 	set_fs (KERNEL_DS);
-	ret = do_sigaltstack((stack_t __user *) (uss ? &kss : NULL),
-			     (stack_t __user *) (uoss ? &koss : NULL),
+	ret = do_sigaltstack((stack_t __force __user *) (uss ? &kss : NULL),
+			     (stack_t __force __user *) (uoss ? &koss : NULL),
 			     regs->gprs[15]);
 	set_fs (old_fs);
 
@@ -298,7 +298,7 @@ static int save_sigregs32(struct pt_regs *regs, _sigregs32 __user *sregs)
 	_s390_regs_common32 regs32;
 	int err, i;
 
-	regs32.psw.mask = PSW32_MASK_MERGE(PSW32_USER_BITS,
+	regs32.psw.mask = PSW32_MASK_MERGE(psw32_user_bits,
 					   (__u32)(regs->psw.mask >> 32));
 	regs32.psw.addr = PSW32_ADDR_AMODE31 | (__u32) regs->psw.addr;
 	for (i = 0; i < NUM_GPRS; i++)
@@ -401,7 +401,7 @@ asmlinkage long sys32_rt_sigreturn(struct pt_regs *regs)
 		goto badframe; 
 
 	set_fs (KERNEL_DS);
-	do_sigaltstack((stack_t __user *)&st, NULL, regs->gprs[15]);
+	do_sigaltstack((stack_t __force __user *)&st, NULL, regs->gprs[15]);
 	set_fs (old_fs);
 
 	return regs->gprs[2];

arch/s390/kernel/cpcmd.c

@@ -16,6 +16,7 @@
 #include <asm/ebcdic.h>
 #include <asm/cpcmd.h>
 #include <asm/system.h>
+#include <asm/io.h>
 
 static DEFINE_SPINLOCK(cpcmd_lock);
 static char cpcmd_buf[241];
@@ -88,13 +89,8 @@ int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 	int len;
 	unsigned long flags;
 
-	if ((rlen == 0) || (response == NULL)
-	    || !((unsigned long)response >> 31)) {
-		spin_lock_irqsave(&cpcmd_lock, flags);
-		len = __cpcmd(cmd, response, rlen, response_code);
-		spin_unlock_irqrestore(&cpcmd_lock, flags);
-	}
-	else {
+	if ((virt_to_phys(response) != (unsigned long) response) ||
+			(((unsigned long)response + rlen) >> 31)) {
 		lowbuf = kmalloc(rlen, GFP_KERNEL | GFP_DMA);
 		if (!lowbuf) {
 			printk(KERN_WARNING
@@ -106,6 +102,10 @@ int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 		spin_unlock_irqrestore(&cpcmd_lock, flags);
 		memcpy(response, lowbuf, rlen);
 		kfree(lowbuf);
+	} else {
+		spin_lock_irqsave(&cpcmd_lock, flags);
+		len = __cpcmd(cmd, response, rlen, response_code);
+		spin_unlock_irqrestore(&cpcmd_lock, flags);
 	}
 	return len;
 }

arch/s390/kernel/crash.c

@@ -9,6 +9,7 @@
 
 #include <linux/threads.h>
 #include <linux/kexec.h>
+#include <linux/reboot.h>
 
 void machine_crash_shutdown(struct pt_regs *regs)
 {

arch/s390/kernel/debug.c

@@ -120,7 +120,7 @@ struct debug_view debug_hex_ascii_view = {
 	NULL
 };
 
-struct debug_view debug_level_view = {
+static struct debug_view debug_level_view = {
 	"level",
 	&debug_prolog_level_fn,
 	NULL,
@@ -129,7 +129,7 @@ struct debug_view debug_level_view = {
 	NULL
 };
 
-struct debug_view debug_pages_view = {
+static struct debug_view debug_pages_view = {
 	"pages",
 	&debug_prolog_pages_fn,
 	NULL,
@@ -138,7 +138,7 @@ struct debug_view debug_pages_view = {
 	NULL
 };
 
-struct debug_view debug_flush_view = {
+static struct debug_view debug_flush_view = {
         "flush",
         NULL,
         NULL,
@@ -156,14 +156,14 @@ struct debug_view debug_sprintf_view = {
 	NULL
 };
 
-
+/* used by dump analysis tools to determine version of debug feature */
 unsigned int debug_feature_version = __DEBUG_FEATURE_VERSION;
 
 /* static globals */
 
 static debug_info_t *debug_area_first = NULL;
 static debug_info_t *debug_area_last = NULL;
-DECLARE_MUTEX(debug_lock);
+static DECLARE_MUTEX(debug_lock);
 
 static int initialized;
 
@@ -905,7 +905,7 @@ static struct ctl_table s390dbf_dir_table[] = {
 	{ .ctl_name = 0 }
 };
 
-struct ctl_table_header *s390dbf_sysctl_header;
+static struct ctl_table_header *s390dbf_sysctl_header;
 
 void
 debug_stop_all(void)
@@ -1300,8 +1300,7 @@ out:
  * flushes debug areas
  */
  
-void
-debug_flush(debug_info_t* id, int area)
+static void debug_flush(debug_info_t* id, int area)
 {
         unsigned long flags;
         int i,j;
@@ -1511,8 +1510,7 @@ out:
 /*
  * clean up module
  */
-void
-__exit debug_exit(void)
+static void __exit debug_exit(void)
 {
 	debugfs_remove(debug_debugfs_root_entry);
 	unregister_sysctl_table(s390dbf_sysctl_header);

arch/s390/kernel/early.c

@@ -0,0 +1,306 @@
+/*
+ *  arch/s390/kernel/early.c
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Hongjie Yang <hongjie@us.ibm.com>,
+ *		 Heiko Carstens <heiko.carstens@de.ibm.com>
+ */
+
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/ctype.h>
+#include <linux/lockdep.h>
+#include <linux/module.h>
+#include <linux/pfn.h>
+#include <linux/uaccess.h>
+#include <asm/lowcore.h>
+#include <asm/processor.h>
+#include <asm/sections.h>
+#include <asm/setup.h>
+#include <asm/cpcmd.h>
+#include <asm/sclp.h>
+
+/*
+ * Create a Kernel NSS if the SAVESYS= parameter is defined
+ */
+#define DEFSYS_CMD_SIZE		96
+#define SAVESYS_CMD_SIZE	32
+
+char kernel_nss_name[NSS_NAME_SIZE + 1];
+
+#ifdef CONFIG_SHARED_KERNEL
+static noinline __init void create_kernel_nss(void)
+{
+	unsigned int i, stext_pfn, eshared_pfn, end_pfn, min_size;
+#ifdef CONFIG_BLK_DEV_INITRD
+	unsigned int sinitrd_pfn, einitrd_pfn;
+#endif
+	int response;
+	char *savesys_ptr;
+	char upper_command_line[COMMAND_LINE_SIZE];
+	char defsys_cmd[DEFSYS_CMD_SIZE];
+	char savesys_cmd[SAVESYS_CMD_SIZE];
+
+	/* Do nothing if we are not running under VM */
+	if (!MACHINE_IS_VM)
+		return;
+
+	/* Convert COMMAND_LINE to upper case */
+	for (i = 0; i < strlen(COMMAND_LINE); i++)
+		upper_command_line[i] = toupper(COMMAND_LINE[i]);
+
+	savesys_ptr = strstr(upper_command_line, "SAVESYS=");
+
+	if (!savesys_ptr)
+		return;
+
+	savesys_ptr += 8;    /* Point to the beginning of the NSS name */
+	for (i = 0; i < NSS_NAME_SIZE; i++) {
+		if (savesys_ptr[i] == ' ' || savesys_ptr[i] == '\0')
+			break;
+		kernel_nss_name[i] = savesys_ptr[i];
+	}
+
+	stext_pfn = PFN_DOWN(__pa(&_stext));
+	eshared_pfn = PFN_DOWN(__pa(&_eshared));
+	end_pfn = PFN_UP(__pa(&_end));
+	min_size = end_pfn << 2;
+
+	sprintf(defsys_cmd, "DEFSYS %s 00000-%.5X EW %.5X-%.5X SR %.5X-%.5X",
+		kernel_nss_name, stext_pfn - 1, stext_pfn, eshared_pfn - 1,
+		eshared_pfn, end_pfn);
+
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (INITRD_START && INITRD_SIZE) {
+		sinitrd_pfn = PFN_DOWN(__pa(INITRD_START));
+		einitrd_pfn = PFN_UP(__pa(INITRD_START + INITRD_SIZE));
+		min_size = einitrd_pfn << 2;
+		sprintf(defsys_cmd, "%s EW %.5X-%.5X", defsys_cmd,
+		sinitrd_pfn, einitrd_pfn);
+	}
+#endif
+
+	sprintf(defsys_cmd, "%s EW MINSIZE=%.7iK", defsys_cmd, min_size);
+	sprintf(savesys_cmd, "SAVESYS %s \n IPL %s",
+		kernel_nss_name, kernel_nss_name);
+
+	__cpcmd(defsys_cmd, NULL, 0, &response);
+
+	if (response != 0)
+		return;
+
+	__cpcmd(savesys_cmd, NULL, 0, &response);
+
+	if (response != strlen(savesys_cmd))
+		return;
+
+	ipl_flags = IPL_NSS_VALID;
+}
+
+#else /* CONFIG_SHARED_KERNEL */
+
+static inline void create_kernel_nss(void) { }
+
+#endif /* CONFIG_SHARED_KERNEL */
+
+/*
+ * Clear bss memory
+ */
+static noinline __init void clear_bss_section(void)
+{
+	memset(__bss_start, 0, _end - __bss_start);
+}
+
+/*
+ * Initialize storage key for kernel pages
+ */
+static noinline __init void init_kernel_storage_key(void)
+{
+	unsigned long end_pfn, init_pfn;
+
+	end_pfn = PFN_UP(__pa(&_end));
+
+	for (init_pfn = 0 ; init_pfn < end_pfn; init_pfn++)
+		page_set_storage_key(init_pfn << PAGE_SHIFT, PAGE_DEFAULT_KEY);
+}
+
+static noinline __init void detect_machine_type(void)
+{
+	struct cpuinfo_S390 *cpuinfo = &S390_lowcore.cpu_data;
+
+	asm volatile("stidp %0" : "=m" (S390_lowcore.cpu_data.cpu_id));
+
+	/* Running under z/VM ? */
+	if (cpuinfo->cpu_id.version == 0xff)
+		machine_flags |= 1;
+
+	/* Running on a P/390 ? */
+	if (cpuinfo->cpu_id.machine == 0x7490)
+		machine_flags |= 4;
+}
+
+static noinline __init int memory_fast_detect(void)
+{
+
+	unsigned long val0 = 0;
+	unsigned long val1 = 0xc;
+	int ret = -ENOSYS;
+
+	if (ipl_flags & IPL_NSS_VALID)
+		return -ENOSYS;
+
+	asm volatile(
+		"	diag	%1,%2,0x260\n"
+		"0:	lhi	%0,0\n"
+		"1:\n"
+		EX_TABLE(0b,1b)
+		: "+d" (ret), "+d" (val0), "+d" (val1) : : "cc");
+
+	if (ret || val0 != val1)
+		return -ENOSYS;
+
+	memory_chunk[0].size = val0;
+	return 0;
+}
+
+#define ADDR2G	(1UL << 31)
+
+static noinline __init unsigned long sclp_memory_detect(void)
+{
+	struct sclp_readinfo_sccb *sccb;
+	unsigned long long memsize;
+
+	sccb = &s390_readinfo_sccb;
+
+	if (sccb->header.response_code != 0x10)
+		return 0;
+
+	if (sccb->rnsize)
+		memsize = sccb->rnsize << 20;
+	else
+		memsize = sccb->rnsize2 << 20;
+	if (sccb->rnmax)
+		memsize *= sccb->rnmax;
+	else
+		memsize *= sccb->rnmax2;
+#ifndef CONFIG_64BIT
+	/*
+	 * Can't deal with more than 2G in 31 bit addressing mode, so
+	 * limit the value in order to avoid strange side effects.
+	 */
+	if (memsize > ADDR2G)
+		memsize = ADDR2G;
+#endif
+	return (unsigned long) memsize;
+}
+
+static inline __init unsigned long __tprot(unsigned long addr)
+{
+	int cc = -1;
+
+	asm volatile(
+		"	tprot	0(%1),0\n"
+		"0:	ipm	%0\n"
+		"	srl	%0,28\n"
+		"1:\n"
+		EX_TABLE(0b,1b)
+		: "+d" (cc) : "a" (addr) : "cc");
+	return (unsigned long)cc;
+}
+
+/* Checking memory in 128KB increments. */
+#define CHUNK_INCR	(1UL << 17)
+
+static noinline __init void find_memory_chunks(unsigned long memsize)
+{
+	unsigned long addr = 0, old_addr = 0;
+	unsigned long old_cc = CHUNK_READ_WRITE;
+	unsigned long cc;
+	int chunk = 0;
+
+	while (chunk < MEMORY_CHUNKS) {
+		cc = __tprot(addr);
+		while (cc == old_cc) {
+			addr += CHUNK_INCR;
+			cc = __tprot(addr);
+#ifndef CONFIG_64BIT
+			if (addr == ADDR2G)
+				break;
+#endif
+		}
+
+		if (old_addr != addr &&
+		    (old_cc == CHUNK_READ_WRITE || old_cc == CHUNK_READ_ONLY)) {
+			memory_chunk[chunk].addr = old_addr;
+			memory_chunk[chunk].size = addr - old_addr;
+			memory_chunk[chunk].type = old_cc;
+			chunk++;
+		}
+
+		old_addr = addr;
+		old_cc = cc;
+
+#ifndef CONFIG_64BIT
+		if (addr == ADDR2G)
+			break;
+#endif
+		/*
+		 * Finish memory detection at the first hole, unless
+		 * - we reached the hsa -> skip it.
+		 * - we know there must be more.
+		 */
+		if (cc == -1UL && !memsize && old_addr != ADDR2G)
+			break;
+		if (memsize && addr >= memsize)
+			break;
+	}
+}
+
+static __init void early_pgm_check_handler(void)
+{
+	unsigned long addr;
+	const struct exception_table_entry *fixup;
+
+	addr = S390_lowcore.program_old_psw.addr;
+	fixup = search_exception_tables(addr & PSW_ADDR_INSN);
+	if (!fixup)
+		disabled_wait(0);
+	S390_lowcore.program_old_psw.addr = fixup->fixup | PSW_ADDR_AMODE;
+}
+
+static noinline __init void setup_lowcore_early(void)
+{
+	psw_t psw;
+
+	psw.mask = PSW_BASE_BITS | PSW_DEFAULT_KEY;
+	psw.addr = PSW_ADDR_AMODE | (unsigned long) s390_base_ext_handler;
+	S390_lowcore.external_new_psw = psw;
+	psw.addr = PSW_ADDR_AMODE | (unsigned long) s390_base_pgm_handler;
+	S390_lowcore.program_new_psw = psw;
+	s390_base_pgm_handler_fn = early_pgm_check_handler;
+}
+
+/*
+ * Save ipl parameters, clear bss memory, initialize storage keys
+ * and create a kernel NSS at startup if the SAVESYS= parm is defined
+ */
+void __init startup_init(void)
+{
+	unsigned long memsize;
+
+	ipl_save_parameters();
+	clear_bss_section();
+	init_kernel_storage_key();
+	lockdep_init();
+	lockdep_off();
+	detect_machine_type();
+	create_kernel_nss();
+	sort_main_extable();
+	setup_lowcore_early();
+	sclp_readinfo_early();
+	memsize = sclp_memory_detect();
+	if (memory_fast_detect() < 0)
+		find_memory_chunks(memsize);
+	lockdep_on();
+}

arch/s390/kernel/ebcdic.c

@@ -11,6 +11,7 @@
 
 #include <linux/module.h>
 #include <asm/types.h>
+#include <asm/ebcdic.h>
 
 /*
  * ASCII (IBM PC 437)  -> EBCDIC 037

arch/s390/kernel/head31.S

@@ -51,175 +51,14 @@ startup_continue:
 	st	%r15,__LC_KERNEL_STACK	# set end of kernel stack
 	ahi	%r15,-96
 	xc	__SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15) # clear backchain
-
-	l	%r14,.Lipl_save_parameters-.LPG1(%r13)
-	basr	%r14,%r14
 #
-# clear bss memory
+# Save ipl parameters, clear bss memory, initialize storage key for kernel pages,
+# and create a kernel NSS if the SAVESYS= parm is defined
 #
-	l	%r2,.Lbss_bgn-.LPG1(%r13) # start of bss
-	l	%r3,.Lbss_end-.LPG1(%r13) # end of bss
-	sr	%r3,%r2			# length of bss
-	sr	%r4,%r4
-	sr	%r5,%r5			# set src,length and pad to zero
-	sr	%r0,%r0
-	mvcle	%r2,%r4,0		# clear mem
-	jo	.-4			# branch back, if not finish
-
-	l	%r2,.Lrcp-.LPG1(%r13)	# Read SCP forced command word
-.Lservicecall:
-	stosm	.Lpmask-.LPG1(%r13),0x01	# authorize ext interrupts
-
-	stctl	%r0, %r0,.Lcr-.LPG1(%r13)	# get cr0
-	la	%r1,0x200		# set bit 22
-	o	%r1,.Lcr-.LPG1(%r13)	# or old cr0 with r1
-	st	%r1,.Lcr-.LPG1(%r13)
-	lctl	%r0, %r0,.Lcr-.LPG1(%r13)	# load modified cr0
-
-	mvc	__LC_EXT_NEW_PSW(8),.Lpcext-.LPG1(%r13) # set postcall psw
-	la	%r1, .Lsclph-.LPG1(%r13)
-	a	%r1,__LC_EXT_NEW_PSW+4	# set handler
-	st	%r1,__LC_EXT_NEW_PSW+4
-
-	l	%r4,.Lsccbaddr-.LPG1(%r13) # %r4 is our index for sccb stuff
-	lr	%r1,%r4			# our sccb
-	.insn	rre,0xb2200000,%r2,%r1	# service call
-	ipm	%r1
-	srl	%r1,28			# get cc code
-	xr	%r3, %r3
-	chi	%r1,3
-	be	.Lfchunk-.LPG1(%r13)	# leave
-	chi	%r1,2
-	be	.Lservicecall-.LPG1(%r13)
-	lpsw	.Lwaitsclp-.LPG1(%r13)
-.Lsclph:
-	lh	%r1,.Lsccbr-.Lsccb(%r4)
-	chi	%r1,0x10		# 0x0010 is the sucess code
-	je	.Lprocsccb		# let's process the sccb
-	chi	%r1,0x1f0
-	bne	.Lfchunk-.LPG1(%r13)	# unhandled error code
-	c	%r2, .Lrcp-.LPG1(%r13)	# Did we try Read SCP forced
-	bne	.Lfchunk-.LPG1(%r13)	# if no, give up
-	l	%r2, .Lrcp2-.LPG1(%r13)	# try with Read SCP
-	b	.Lservicecall-.LPG1(%r13)
-.Lprocsccb:
-	lhi	%r1,0
-	icm	%r1,3,.Lscpincr1-.Lsccb(%r4) # use this one if != 0
-	jnz	.Lscnd
-	lhi	%r1,0x800		# otherwise report 2GB
-.Lscnd:
-	lhi	%r3,0x800		# limit reported memory size to 2GB
-	cr	%r1,%r3
-	jl	.Lno2gb
-	lr	%r1,%r3
-.Lno2gb:
-	xr	%r3,%r3			# same logic
-	ic	%r3,.Lscpa1-.Lsccb(%r4)
-	chi	%r3,0x00
-	jne	.Lcompmem
-	l	%r3,.Lscpa2-.Lsccb(%r4)
-.Lcompmem:
-	mr	%r2,%r1			# mem in MB on 128-bit
-	l	%r1,.Lonemb-.LPG1(%r13)
-	mr	%r2,%r1			# mem size in bytes in %r3
-	b	.Lfchunk-.LPG1(%r13)
-
-	.align 4
-.Lipl_save_parameters:
-	.long	ipl_save_parameters
-.Linittu:
-	.long	init_thread_union
-.Lpmask:
-	.byte	0
-	.align	8
-.Lpcext:.long	0x00080000,0x80000000
-.Lcr:
-	.long	0x00			# place holder for cr0
-	.align	8
-.Lwaitsclp:
-	.long 0x010a0000,0x80000000 + .Lsclph
-.Lrcp:
-	.int	0x00120001		# Read SCP forced code
-.Lrcp2:
-	.int	0x00020001		# Read SCP code
-.Lonemb:
-	.int	0x100000
-.Lfchunk:
+	l	%r14,.Lstartup_init-.LPG1(%r13)
+	basr	%r14,%r14
 
-#
-# find memory chunks.
-#
-	lr	%r9,%r3			# end of mem
-	mvc	__LC_PGM_NEW_PSW(8),.Lpcmem-.LPG1(%r13)
-	la	%r1,1			# test in increments of 128KB
-	sll	%r1,17
-	l	%r3,.Lmchunk-.LPG1(%r13) # get pointer to memory_chunk array
-	slr	%r4,%r4			# set start of chunk to zero
-	slr	%r5,%r5			# set end of chunk to zero
-	slr	%r6,%r6			# set access code to zero
-	la	%r10,MEMORY_CHUNKS	# number of chunks
-.Lloop:
-	tprot	0(%r5),0		# test protection of first byte
-	ipm	%r7
-	srl	%r7,28
-	clr	%r6,%r7			# compare cc with last access code
-	be	.Lsame-.LPG1(%r13)
-	lhi	%r8,0			# no program checks
-	b	.Lsavchk-.LPG1(%r13)
-.Lsame:
-	ar	%r5,%r1			# add 128KB to end of chunk
-	bno	.Lloop-.LPG1(%r13)	# r1 < 0x80000000 -> loop
-.Lchkmem:				# > 2GB or tprot got a program check
-	lhi	%r8,1			# set program check flag
-.Lsavchk:
-	clr	%r4,%r5			# chunk size > 0?
-	be	.Lchkloop-.LPG1(%r13)
-	st	%r4,0(%r3)		# store start address of chunk
-	lr	%r0,%r5
-	slr	%r0,%r4
-	st	%r0,4(%r3)		# store size of chunk
-	st	%r6,8(%r3)		# store type of chunk
-	la	%r3,12(%r3)
-	ahi	%r10,-1			# update chunk number
-.Lchkloop:
-	lr	%r6,%r7			# set access code to last cc
-	# we got an exception or we're starting a new
-	# chunk , we must check if we should
-	# still try to find valid memory (if we detected
-	# the amount of available storage), and if we
-	# have chunks left
-	xr	%r0,%r0
-	clr	%r0,%r9			# did we detect memory?
-	je	.Ldonemem		# if not, leave
-	chi	%r10,0			# do we have chunks left?
-	je	.Ldonemem
-	chi	%r8,1			# program check ?
-	je	.Lpgmchk
-	lr	%r4,%r5			# potential new chunk
-	alr	%r5,%r1			# add 128KB to end of chunk
-	j	.Llpcnt
-.Lpgmchk:
-	alr	%r5,%r1			# add 128KB to end of chunk
-	lr	%r4,%r5			# potential new chunk
-.Llpcnt:
-	clr	%r5,%r9			# should we go on?
-	jl	.Lloop
-.Ldonemem:
 	l	%r12,.Lmflags-.LPG1(%r13) # get address of machine_flags
-#
-# find out if we are running under VM
-#
-	stidp	__LC_CPUID		# store cpuid
-	tm	__LC_CPUID,0xff		# running under VM ?
-	bno	.Lnovm-.LPG1(%r13)
-	oi	3(%r12),1		# set VM flag
-.Lnovm:
-	lh	%r0,__LC_CPUID+4	# get cpu version
-	chi	%r0,0x7490		# running on a P/390 ?
-	bne	.Lnop390-.LPG1(%r13)
-	oi	3(%r12),4		# set P/390 flag
-.Lnop390:
-
 #
 # find out if we have an IEEE fpu
 #
@@ -295,7 +134,6 @@ startup_continue:
 	.long	0			# cr15: linkage stack operations
 .Lduct:	.long	0,0,0,0,0,0,0,0
 	.long	0,0,0,0,0,0,0,0
-.Lpcmem:.long	0x00080000,0x80000000 + .Lchkmem
 .Lpcfpu:.long	0x00080000,0x80000000 + .Lchkfpu
 .Lpccsp:.long	0x00080000,0x80000000 + .Lchkcsp
 .Lpcmvpg:.long	0x00080000,0x80000000 + .Lchkmvpg
@@ -306,7 +144,9 @@ startup_continue:
 .Lbss_bgn:  .long __bss_start
 .Lbss_end:  .long _end
 .Lparmaddr: .long PARMAREA
-.Lsccbaddr: .long .Lsccb
+.Linittu:   .long init_thread_union
+.Lstartup_init:
+	    .long startup_init
 
 	.globl ipl_schib
 ipl_schib:
@@ -322,26 +162,6 @@ ipl_devno:
 	.word 0
 
 	.org	0x12000
-.globl s390_readinfo_sccb
-s390_readinfo_sccb:
-.Lsccb:
-	.hword	0x1000			# length, one page
-	.byte	0x00,0x00,0x00
-	.byte	0x80			# variable response bit set
-.Lsccbr:
-	.hword	0x00			# response code
-.Lscpincr1:
-	.hword	0x00
-.Lscpa1:
-	.byte	0x00
-	.fill	89,1,0
-.Lscpa2:
-	.int	0x00
-.Lscpincr2:
-	.quad	0x00
-	.fill	3984,1,0
-	.org	0x13000
-
 #ifdef CONFIG_SHARED_KERNEL
 	.org	0x100000
 #endif

arch/s390/kernel/head64.S

@@ -58,182 +58,14 @@ startup_continue:
 	stg	%r15,__LC_KERNEL_STACK	# set end of kernel stack
 	aghi	%r15,-160
 	xc	__SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15) # clear backchain
-
-	brasl	%r14,ipl_save_parameters
 #
-# clear bss memory
+# Save ipl parameters, clear bss memory, initialize storage key for kernel pages,
+# and create a kernel NSS if the SAVESYS= parm is defined
 #
-	larl	%r2,__bss_start 	# start of bss segment
-	larl	%r3,_end		# end of bss segment
-	sgr	%r3,%r2 		# length of bss
-	sgr	%r4,%r4 		#
-	sgr	%r5,%r5 		# set src,length and pad to zero
-	mvcle	%r2,%r4,0		# clear mem
-	jo	.-4			# branch back, if not finish
+	brasl	%r14,startup_init
 					# set program check new psw mask
 	mvc	__LC_PGM_NEW_PSW(8),.Lpcmsk-.LPG1(%r13)
-	larl	%r1,.Lslowmemdetect	# set program check address
-	stg	%r1,__LC_PGM_NEW_PSW+8
-	lghi	%r1,0xc
-	diag	%r0,%r1,0x260		# get memory size of virtual machine
-	cgr	%r0,%r1			# different? -> old detection routine
-	jne	.Lslowmemdetect
-	aghi	%r1,1			# size is one more than end
-	larl	%r2,memory_chunk
-	stg	%r1,8(%r2)		# store size of chunk
-	j	.Ldonemem
-
-.Lslowmemdetect:
-	l	%r2,.Lrcp-.LPG1(%r13)	# Read SCP forced command word
-.Lservicecall:
-	stosm	.Lpmask-.LPG1(%r13),0x01	# authorize ext interrupts
-
-	stctg	%r0,%r0,.Lcr-.LPG1(%r13)	# get cr0
-	la	%r1,0x200		# set bit 22
-	og	%r1,.Lcr-.LPG1(%r13)	# or old cr0 with r1
-	stg	%r1,.Lcr-.LPG1(%r13)
-	lctlg	%r0,%r0,.Lcr-.LPG1(%r13)	# load modified cr0
-
-	mvc	__LC_EXT_NEW_PSW(8),.Lpcmsk-.LPG1(%r13) # set postcall psw
-	larl	%r1,.Lsclph
-	stg	%r1,__LC_EXT_NEW_PSW+8	# set handler
-
-	larl	%r4,.Lsccb		# %r4 is our index for sccb stuff
-	lgr	%r1,%r4			# our sccb
-	.insn	rre,0xb2200000,%r2,%r1	# service call
-	ipm	%r1
-	srl	%r1,28			# get cc code
-	xr	%r3,%r3
-	chi	%r1,3
-	be	.Lfchunk-.LPG1(%r13)	# leave
-	chi	%r1,2
-	be	.Lservicecall-.LPG1(%r13)
-	lpswe	.Lwaitsclp-.LPG1(%r13)
-.Lsclph:
-	lh	%r1,.Lsccbr-.Lsccb(%r4)
-	chi	%r1,0x10		# 0x0010 is the sucess code
-	je	.Lprocsccb		# let's process the sccb
-	chi	%r1,0x1f0
-	bne	.Lfchunk-.LPG1(%r13)	# unhandled error code
-	c	%r2,.Lrcp-.LPG1(%r13)	# Did we try Read SCP forced
-	bne	.Lfchunk-.LPG1(%r13)	# if no, give up
-	l	%r2,.Lrcp2-.LPG1(%r13)	# try with Read SCP
-	b	.Lservicecall-.LPG1(%r13)
-.Lprocsccb:
-	lghi	%r1,0
-	icm	%r1,3,.Lscpincr1-.Lsccb(%r4)	# use this one if != 0
-	jnz	.Lscnd
-	lg	%r1,.Lscpincr2-.Lsccb(%r4)	# otherwise use this one
-.Lscnd:
-	xr	%r3,%r3			# same logic
-	ic	%r3,.Lscpa1-.Lsccb(%r4)
-	chi	%r3,0x00
-	jne	.Lcompmem
-	l	%r3,.Lscpa2-.Lsccb(%r4)
-.Lcompmem:
-	mlgr	%r2,%r1			# mem in MB on 128-bit
-	l	%r1,.Lonemb-.LPG1(%r13)
-	mlgr	%r2,%r1			# mem size in bytes in %r3
-	b	.Lfchunk-.LPG1(%r13)
-
-	.align	4
-.Lpmask:
-	.byte	0
-	.align	8
-.Lcr:
-	.quad	0x00  # place holder for cr0
-.Lwaitsclp:
-	.quad	0x0102000180000000,.Lsclph
-.Lrcp:
-	.int	0x00120001 # Read SCP forced code
-.Lrcp2:
-	.int	0x00020001 # Read SCP code
-.Lonemb:
-	.int	0x100000
-
-.Lfchunk:
-
-#
-# find memory chunks.
-#
-	lgr	%r9,%r3			# end of mem
-	larl	%r1,.Lchkmem		# set program check address
-	stg	%r1,__LC_PGM_NEW_PSW+8
-	la	%r1,1			# test in increments of 128KB
-	sllg	%r1,%r1,17
-	larl	%r3,memory_chunk
-	slgr	%r4,%r4 		# set start of chunk to zero
-	slgr	%r5,%r5 		# set end of chunk to zero
-	slr	%r6,%r6			# set access code to zero
-	la	%r10,MEMORY_CHUNKS	# number of chunks
-.Lloop:
-	tprot	0(%r5),0		# test protection of first byte
-	ipm	%r7
-	srl	%r7,28
-	clr	%r6,%r7			# compare cc with last access code
-	je	.Lsame
-	lghi	%r8,0			# no program checks
-	j	.Lsavchk
-.Lsame:
-	algr	%r5,%r1			# add 128KB to end of chunk
-					# no need to check here,
-	brc	12,.Lloop		# this is the same chunk
-.Lchkmem:				# > 16EB or tprot got a program check
-	lghi	%r8,1			# set program check flag
-.Lsavchk:
-	clgr	%r4,%r5			# chunk size > 0?
-	je	.Lchkloop
-	stg	%r4,0(%r3)		# store start address of chunk
-	lgr	%r0,%r5
-	slgr	%r0,%r4
-	stg	%r0,8(%r3)		# store size of chunk
-	st	%r6,20(%r3)		# store type of chunk
-	la	%r3,24(%r3)
-	ahi	%r10,-1			# update chunk number
-.Lchkloop:
-	lr	%r6,%r7			# set access code to last cc
-	# we got an exception or we're starting a new
-	# chunk , we must check if we should
-	# still try to find valid memory (if we detected
-	# the amount of available storage), and if we
-	# have chunks left
-	lghi	%r4,1
-	sllg	%r4,%r4,31
-	clgr	%r5,%r4
-	je	.Lhsaskip
-	xr	%r0, %r0
-	clgr	%r0, %r9		# did we detect memory?
-	je	.Ldonemem		# if not, leave
-	chi	%r10, 0			# do we have chunks left?
-	je	.Ldonemem
-.Lhsaskip:
-	chi	%r8,1			# program check ?
-	je	.Lpgmchk
-	lgr	%r4,%r5			# potential new chunk
-	algr	%r5,%r1			# add 128KB to end of chunk
-	j	.Llpcnt
-.Lpgmchk:
-	algr	%r5,%r1			# add 128KB to end of chunk
-	lgr	%r4,%r5			# potential new chunk
-.Llpcnt:
-	clgr	%r5,%r9			# should we go on?
-	jl	.Lloop
-.Ldonemem:
-
 	larl	%r12,machine_flags
-#
-# find out if we are running under VM
-#
-	stidp	__LC_CPUID		# store cpuid
-	tm	__LC_CPUID,0xff 	# running under VM ?
-	bno	0f-.LPG1(%r13)
-	oi	7(%r12),1		# set VM flag
-0:	lh	%r0,__LC_CPUID+4	# get cpu version
-	chi	%r0,0x7490		# running on a P/390 ?
-	bne	1f-.LPG1(%r13)
-	oi	7(%r12),4		# set P/390 flag
-1:
-
 #
 # find out if we have the MVPG instruction
 #
@@ -336,25 +168,6 @@ ipl_devno:
 	.word 0
 
 	.org	0x12000
-.globl s390_readinfo_sccb
-s390_readinfo_sccb:
-.Lsccb:
-	.hword	0x1000			# length, one page
-	.byte	0x00,0x00,0x00
-	.byte	0x80			# variable response bit set
-.Lsccbr:
-	.hword	0x00			# response code
-.Lscpincr1:
-	.hword	0x00
-.Lscpa1:
-	.byte	0x00
-	.fill	89,1,0
-.Lscpa2:
-	.int	0x00
-.Lscpincr2:
-	.quad	0x00
-	.fill	3984,1,0
-	.org	0x13000
 
 #ifdef CONFIG_SHARED_KERNEL
 	.org	0x100000

arch/s390/kernel/ipl.c

@@ -20,26 +20,27 @@
 #include <asm/cio.h>
 #include <asm/ebcdic.h>
 #include <asm/reset.h>
+#include <asm/sclp.h>
 
 #define IPL_PARM_BLOCK_VERSION 0
-#define LOADPARM_LEN 8
 
-extern char s390_readinfo_sccb[];
-#define SCCB_VALID (*((__u16*)&s390_readinfo_sccb[6]) == 0x0010)
-#define SCCB_LOADPARM (&s390_readinfo_sccb[24])
-#define SCCB_FLAG (s390_readinfo_sccb[91])
+#define SCCB_VALID (s390_readinfo_sccb.header.response_code == 0x10)
+#define SCCB_LOADPARM (&s390_readinfo_sccb.loadparm)
+#define SCCB_FLAG (s390_readinfo_sccb.flags)
 
 enum ipl_type {
 	IPL_TYPE_NONE	 = 1,
 	IPL_TYPE_UNKNOWN = 2,
 	IPL_TYPE_CCW	 = 4,
 	IPL_TYPE_FCP	 = 8,
+	IPL_TYPE_NSS	 = 16,
 };
 
 #define IPL_NONE_STR	 "none"
 #define IPL_UNKNOWN_STR  "unknown"
 #define IPL_CCW_STR	 "ccw"
 #define IPL_FCP_STR	 "fcp"
+#define IPL_NSS_STR	 "nss"
 
 static char *ipl_type_str(enum ipl_type type)
 {
@@ -50,6 +51,8 @@ static char *ipl_type_str(enum ipl_type type)
 		return IPL_CCW_STR;
 	case IPL_TYPE_FCP:
 		return IPL_FCP_STR;
+	case IPL_TYPE_NSS:
+		return IPL_NSS_STR;
 	case IPL_TYPE_UNKNOWN:
 	default:
 		return IPL_UNKNOWN_STR;
@@ -64,6 +67,7 @@ enum ipl_method {
 	IPL_METHOD_FCP_RO_DIAG,
 	IPL_METHOD_FCP_RW_DIAG,
 	IPL_METHOD_FCP_RO_VM,
+	IPL_METHOD_NSS,
 };
 
 enum shutdown_action {
@@ -114,11 +118,14 @@ enum diag308_rc {
 static int diag308_set_works = 0;
 
 static int reipl_capabilities = IPL_TYPE_UNKNOWN;
+
 static enum ipl_type reipl_type = IPL_TYPE_UNKNOWN;
 static enum ipl_method reipl_method = IPL_METHOD_NONE;
 static struct ipl_parameter_block *reipl_block_fcp;
 static struct ipl_parameter_block *reipl_block_ccw;
 
+static char reipl_nss_name[NSS_NAME_SIZE + 1];
+
 static int dump_capabilities = IPL_TYPE_NONE;
 static enum ipl_type dump_type = IPL_TYPE_NONE;
 static enum ipl_method dump_method = IPL_METHOD_NONE;
@@ -173,6 +180,24 @@ static struct subsys_attribute sys_##_prefix##_##_name##_attr =		\
 			sys_##_prefix##_##_name##_show,			\
 			sys_##_prefix##_##_name##_store);
 
+#define DEFINE_IPL_ATTR_STR_RW(_prefix, _name, _fmt_out, _fmt_in, _value)\
+static ssize_t sys_##_prefix##_##_name##_show(struct subsystem *subsys,	\
+		char *page)						\
+{									\
+	return sprintf(page, _fmt_out, _value);				\
+}									\
+static ssize_t sys_##_prefix##_##_name##_store(struct subsystem *subsys,\
+		const char *buf, size_t len)				\
+{									\
+	if (sscanf(buf, _fmt_in, _value) != 1)				\
+		return -EINVAL;						\
+	return len;							\
+}									\
+static struct subsys_attribute sys_##_prefix##_##_name##_attr =		\
+	__ATTR(_name,(S_IRUGO | S_IWUSR),				\
+			sys_##_prefix##_##_name##_show,			\
+			sys_##_prefix##_##_name##_store);
+
 static void make_attrs_ro(struct attribute **attrs)
 {
 	while (*attrs) {
@@ -189,6 +214,8 @@ static enum ipl_type ipl_get_type(void)
 {
 	struct ipl_parameter_block *ipl = IPL_PARMBLOCK_START;
 
+	if (ipl_flags & IPL_NSS_VALID)
+		return IPL_TYPE_NSS;
 	if (!(ipl_flags & IPL_DEVNO_VALID))
 		return IPL_TYPE_UNKNOWN;
 	if (!(ipl_flags & IPL_PARMBLOCK_VALID))
@@ -324,6 +351,20 @@ static struct attribute_group ipl_ccw_attr_group = {
 	.attrs = ipl_ccw_attrs,
 };
 
+/* NSS ipl device attributes */
+
+DEFINE_IPL_ATTR_RO(ipl_nss, name, "%s\n", kernel_nss_name);
+
+static struct attribute *ipl_nss_attrs[] = {
+	&sys_ipl_type_attr.attr,
+	&sys_ipl_nss_name_attr.attr,
+	NULL,
+};
+
+static struct attribute_group ipl_nss_attr_group = {
+	.attrs = ipl_nss_attrs,
+};
+
 /* UNKNOWN ipl device attributes */
 
 static struct attribute *ipl_unknown_attrs[] = {
@@ -432,6 +473,21 @@ static struct attribute_group reipl_ccw_attr_group = {
 	.attrs = reipl_ccw_attrs,
 };
 
+
+/* NSS reipl device attributes */
+
+DEFINE_IPL_ATTR_STR_RW(reipl_nss, name, "%s\n", "%s\n", reipl_nss_name);
+
+static struct attribute *reipl_nss_attrs[] = {
+	&sys_reipl_nss_name_attr.attr,
+	NULL,
+};
+
+static struct attribute_group reipl_nss_attr_group = {
+	.name  = IPL_NSS_STR,
+	.attrs = reipl_nss_attrs,
+};
+
 /* reipl type */
 
 static int reipl_set_type(enum ipl_type type)
@@ -454,6 +510,9 @@ static int reipl_set_type(enum ipl_type type)
 		else
 			reipl_method = IPL_METHOD_FCP_RO_DIAG;
 		break;
+	case IPL_TYPE_NSS:
+		reipl_method = IPL_METHOD_NSS;
+		break;
 	default:
 		reipl_method = IPL_METHOD_NONE;
 	}
@@ -475,6 +534,8 @@ static ssize_t reipl_type_store(struct subsystem *subsys, const char *buf,
 		rc = reipl_set_type(IPL_TYPE_CCW);
 	else if (strncmp(buf, IPL_FCP_STR, strlen(IPL_FCP_STR)) == 0)
 		rc = reipl_set_type(IPL_TYPE_FCP);
+	else if (strncmp(buf, IPL_NSS_STR, strlen(IPL_NSS_STR)) == 0)
+		rc = reipl_set_type(IPL_TYPE_NSS);
 	return (rc != 0) ? rc : len;
 }
 
@@ -647,6 +708,10 @@ void do_reipl(void)
 	case IPL_METHOD_FCP_RO_VM:
 		__cpcmd("IPL", NULL, 0, NULL);
 		break;
+	case IPL_METHOD_NSS:
+		sprintf(buf, "IPL %s", reipl_nss_name);
+		__cpcmd(buf, NULL, 0, NULL);
+		break;
 	case IPL_METHOD_NONE:
 	default:
 		if (MACHINE_IS_VM)
@@ -733,6 +798,10 @@ static int __init ipl_init(void)
 	case IPL_TYPE_FCP:
 		rc = ipl_register_fcp_files();
 		break;
+	case IPL_TYPE_NSS:
+		rc = sysfs_create_group(&ipl_subsys.kset.kobj,
+					&ipl_nss_attr_group);
+		break;
 	default:
 		rc = sysfs_create_group(&ipl_subsys.kset.kobj,
 					&ipl_unknown_attr_group);
@@ -755,6 +824,20 @@ static void __init reipl_probe(void)
 	free_page((unsigned long)buffer);
 }
 
+static int __init reipl_nss_init(void)
+{
+	int rc;
+
+	if (!MACHINE_IS_VM)
+		return 0;
+	rc = sysfs_create_group(&reipl_subsys.kset.kobj, &reipl_nss_attr_group);
+	if (rc)
+		return rc;
+	strncpy(reipl_nss_name, kernel_nss_name, NSS_NAME_SIZE + 1);
+	reipl_capabilities |= IPL_TYPE_NSS;
+	return 0;
+}
+
 static int __init reipl_ccw_init(void)
 {
 	int rc;
@@ -835,6 +918,9 @@ static int __init reipl_init(void)
 	if (rc)
 		return rc;
 	rc = reipl_fcp_init();
+	if (rc)
+		return rc;
+	rc = reipl_nss_init();
 	if (rc)
 		return rc;
 	rc = reipl_set_type(ipl_get_type());
@@ -993,8 +1079,6 @@ static void do_reset_calls(void)
 		reset->fn();
 }
 
-extern void reset_mcck_handler(void);
-extern void reset_pgm_handler(void);
 extern __u32 dump_prefix_page;
 
 void s390_reset_system(void)
@@ -1016,14 +1100,14 @@ void s390_reset_system(void)
 	__ctl_clear_bit(0,28);
 
 	/* Set new machine check handler */
-	S390_lowcore.mcck_new_psw.mask = PSW_KERNEL_BITS & ~PSW_MASK_MCHECK;
+	S390_lowcore.mcck_new_psw.mask = psw_kernel_bits & ~PSW_MASK_MCHECK;
 	S390_lowcore.mcck_new_psw.addr =
-		PSW_ADDR_AMODE | (unsigned long) &reset_mcck_handler;
+		PSW_ADDR_AMODE | (unsigned long) s390_base_mcck_handler;
 
 	/* Set new program check handler */
-	S390_lowcore.program_new_psw.mask = PSW_KERNEL_BITS & ~PSW_MASK_MCHECK;
+	S390_lowcore.program_new_psw.mask = psw_kernel_bits & ~PSW_MASK_MCHECK;
 	S390_lowcore.program_new_psw.addr =
-		PSW_ADDR_AMODE | (unsigned long) &reset_pgm_handler;
+		PSW_ADDR_AMODE | (unsigned long) s390_base_pgm_handler;
 
 	do_reset_calls();
 }

arch/s390/kernel/irq.c

@@ -1,9 +1,9 @@
 /*
  *  arch/s390/kernel/irq.c
  *
- *  S390 version
- *    Copyright (C) 2004 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Copyright IBM Corp. 2004,2007
  *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
+ *		 Thomas Spatzier (tspat@de.ibm.com)
  *
  * This file contains interrupt related functions.
  */
@@ -14,6 +14,8 @@
 #include <linux/interrupt.h>
 #include <linux/seq_file.h>
 #include <linux/cpu.h>
+#include <linux/proc_fs.h>
+#include <linux/profile.h>
 
 /*
  * show_interrupts is needed by /proc/interrupts.
@@ -93,5 +95,12 @@ asmlinkage void do_softirq(void)
 
 	local_irq_restore(flags);
 }
-
 EXPORT_SYMBOL(do_softirq);
+
+void init_irq_proc(void)
+{
+	struct proc_dir_entry *root_irq_dir;
+
+	root_irq_dir = proc_mkdir("irq", NULL);
+	create_prof_cpu_mask(root_irq_dir);
+}

arch/s390/kernel/kprobes.c

@@ -155,15 +155,34 @@ void __kprobes get_instruction_type(struct arch_specific_insn *ainsn)
 static int __kprobes swap_instruction(void *aref)
 {
 	struct ins_replace_args *args = aref;
+	u32 *addr;
+	u32 instr;
 	int err = -EFAULT;
 
+	/*
+	 * Text segment is read-only, hence we use stura to bypass dynamic
+	 * address translation to exchange the instruction. Since stura
+	 * always operates on four bytes, but we only want to exchange two
+	 * bytes do some calculations to get things right. In addition we
+	 * shall not cross any page boundaries (vmalloc area!) when writing
+	 * the new instruction.
+	 */
+	addr = (u32 *)ALIGN((unsigned long)args->ptr, 4);
+	if ((unsigned long)args->ptr & 2)
+		instr = ((*addr) & 0xffff0000) | args->new;
+	else
+		instr = ((*addr) & 0x0000ffff) | args->new << 16;
+
 	asm volatile(
-		"0: mvc  0(2,%2),0(%3)\n"
-		"1: la   %0,0\n"
+		"	lra	%1,0(%1)\n"
+		"0:	stura	%2,%1\n"
+		"1:	la	%0,0\n"
 		"2:\n"
 		EX_TABLE(0b,2b)
-		: "+d" (err), "=m" (*args->ptr)
-		: "a" (args->ptr), "a" (&args->new), "m" (args->new));
+		: "+d" (err)
+		: "a" (addr), "d" (instr)
+		: "memory", "cc");
+
 	return err;
 }
 
@@ -356,7 +375,7 @@ no_kprobe:
  *	- When the probed function returns, this probe
  *		causes the handlers to fire
  */
-void __kprobes kretprobe_trampoline_holder(void)
+void kretprobe_trampoline_holder(void)
 {
 	asm volatile(".global kretprobe_trampoline\n"
 		     "kretprobe_trampoline: bcr 0,0\n");
@@ -365,7 +384,8 @@ void __kprobes kretprobe_trampoline_holder(void)
 /*
  * Called when the probe at kretprobe trampoline is hit
  */
-int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
+static int __kprobes trampoline_probe_handler(struct kprobe *p,
+					      struct pt_regs *regs)
 {
 	struct kretprobe_instance *ri = NULL;
 	struct hlist_head *head, empty_rp;

arch/s390/kernel/machine_kexec.c

@@ -11,6 +11,7 @@
 #include <linux/mm.h>
 #include <linux/kexec.h>
 #include <linux/delay.h>
+#include <linux/reboot.h>
 #include <asm/cio.h>
 #include <asm/setup.h>
 #include <asm/pgtable.h>

arch/s390/kernel/module.c

@@ -30,6 +30,7 @@
 #include <linux/fs.h>
 #include <linux/string.h>
 #include <linux/kernel.h>
+#include <linux/moduleloader.h>
 
 #if 0
 #define DEBUGP printk
@@ -58,7 +59,7 @@ void module_free(struct module *mod, void *module_region)
            table entries. */
 }
 
-static inline void
+static void
 check_rela(Elf_Rela *rela, struct module *me)
 {
 	struct mod_arch_syminfo *info;
@@ -181,7 +182,7 @@ apply_relocate(Elf_Shdr *sechdrs, const char *strtab, unsigned int symindex,
 	return -ENOEXEC;
 }
 
-static inline int
+static int
 apply_rela(Elf_Rela *rela, Elf_Addr base, Elf_Sym *symtab, 
 	   struct module *me)
 {

arch/s390/kernel/process.c

@@ -144,7 +144,7 @@ static void default_idle(void)
 
 	trace_hardirqs_on();
 	/* Wait for external, I/O or machine check interrupt. */
-	__load_psw_mask(PSW_KERNEL_BITS | PSW_MASK_WAIT |
+	__load_psw_mask(psw_kernel_bits | PSW_MASK_WAIT |
 			PSW_MASK_IO | PSW_MASK_EXT);
 }
 
@@ -190,7 +190,7 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
 	struct pt_regs regs;
 
 	memset(&regs, 0, sizeof(regs));
-	regs.psw.mask = PSW_KERNEL_BITS | PSW_MASK_IO | PSW_MASK_EXT;
+	regs.psw.mask = psw_kernel_bits | PSW_MASK_IO | PSW_MASK_EXT;
 	regs.psw.addr = (unsigned long) kernel_thread_starter | PSW_ADDR_AMODE;
 	regs.gprs[9] = (unsigned long) fn;
 	regs.gprs[10] = (unsigned long) arg;

arch/s390/kernel/profile.c

@@ -1,20 +0,0 @@
-/*
- * arch/s390/kernel/profile.c
- *
- * Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
- * Author(s): Thomas Spatzier (tspat@de.ibm.com)
- *
- */
-#include <linux/proc_fs.h>
-#include <linux/profile.h>
-
-static struct proc_dir_entry * root_irq_dir;
-
-void init_irq_proc(void)
-{
-	/* create /proc/irq */
-	root_irq_dir = proc_mkdir("irq", NULL);
-
-	/* create /proc/irq/prof_cpu_mask */
-	create_prof_cpu_mask(root_irq_dir);
-}

arch/s390/kernel/ptrace.c

@@ -86,15 +86,13 @@ FixPerRegisters(struct task_struct *task)
 		per_info->control_regs.bits.storage_alt_space_ctl = 0;
 }
 
-void
-set_single_step(struct task_struct *task)
+static void set_single_step(struct task_struct *task)
 {
 	task->thread.per_info.single_step = 1;
 	FixPerRegisters(task);
 }
 
-void
-clear_single_step(struct task_struct *task)
+static void clear_single_step(struct task_struct *task)
 {
 	task->thread.per_info.single_step = 0;
 	FixPerRegisters(task);
@@ -232,9 +230,9 @@ poke_user(struct task_struct *child, addr_t addr, addr_t data)
 		 */
 		if (addr == (addr_t) &dummy->regs.psw.mask &&
 #ifdef CONFIG_COMPAT
-		    data != PSW_MASK_MERGE(PSW_USER32_BITS, data) &&
+		    data != PSW_MASK_MERGE(psw_user32_bits, data) &&
 #endif
-		    data != PSW_MASK_MERGE(PSW_USER_BITS, data))
+		    data != PSW_MASK_MERGE(psw_user_bits, data))
 			/* Invalid psw mask. */
 			return -EINVAL;
 #ifndef CONFIG_64BIT
@@ -309,7 +307,7 @@ do_ptrace_normal(struct task_struct *child, long request, long addr, long data)
 		copied = access_process_vm(child, addr, &tmp, sizeof(tmp), 0);
 		if (copied != sizeof(tmp))
 			return -EIO;
-		return put_user(tmp, (unsigned long __user *) data);
+		return put_user(tmp, (unsigned long __force __user *) data);
 
 	case PTRACE_PEEKUSR:
 		/* read the word at location addr in the USER area. */
@@ -331,7 +329,7 @@ do_ptrace_normal(struct task_struct *child, long request, long addr, long data)
 
 	case PTRACE_PEEKUSR_AREA:
 	case PTRACE_POKEUSR_AREA:
-		if (copy_from_user(&parea, (void __user *) addr,
+		if (copy_from_user(&parea, (void __force __user *) addr,
 							sizeof(parea)))
 			return -EFAULT;
 		addr = parea.kernel_addr;
@@ -341,10 +339,11 @@ do_ptrace_normal(struct task_struct *child, long request, long addr, long data)
 			if (request == PTRACE_PEEKUSR_AREA)
 				ret = peek_user(child, addr, data);
 			else {
-				addr_t tmp;
-				if (get_user (tmp, (addr_t __user *) data))
+				addr_t utmp;
+				if (get_user(utmp,
+					     (addr_t __force __user *) data))
 					return -EFAULT;
-				ret = poke_user(child, addr, tmp);
+				ret = poke_user(child, addr, utmp);
 			}
 			if (ret)
 				return ret;
@@ -394,7 +393,7 @@ peek_user_emu31(struct task_struct *child, addr_t addr, addr_t data)
 		if (addr == (addr_t) &dummy32->regs.psw.mask) {
 			/* Fake a 31 bit psw mask. */
 			tmp = (__u32)(task_pt_regs(child)->psw.mask >> 32);
-			tmp = PSW32_MASK_MERGE(PSW32_USER_BITS, tmp);
+			tmp = PSW32_MASK_MERGE(psw32_user_bits, tmp);
 		} else if (addr == (addr_t) &dummy32->regs.psw.addr) {
 			/* Fake a 31 bit psw address. */
 			tmp = (__u32) task_pt_regs(child)->psw.addr |
@@ -469,11 +468,11 @@ poke_user_emu31(struct task_struct *child, addr_t addr, addr_t data)
 		 */
 		if (addr == (addr_t) &dummy32->regs.psw.mask) {
 			/* Build a 64 bit psw mask from 31 bit mask. */
-			if (tmp != PSW32_MASK_MERGE(PSW32_USER_BITS, tmp))
+			if (tmp != PSW32_MASK_MERGE(psw32_user_bits, tmp))
 				/* Invalid psw mask. */
 				return -EINVAL;
 			task_pt_regs(child)->psw.mask =
-				PSW_MASK_MERGE(PSW_USER32_BITS, (__u64) tmp << 32);
+				PSW_MASK_MERGE(psw_user32_bits, (__u64) tmp << 32);
 		} else if (addr == (addr_t) &dummy32->regs.psw.addr) {
 			/* Build a 64 bit psw address from 31 bit address. */
 			task_pt_regs(child)->psw.addr =
@@ -550,7 +549,7 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 		copied = access_process_vm(child, addr, &tmp, sizeof(tmp), 0);
 		if (copied != sizeof(tmp))
 			return -EIO;
-		return put_user(tmp, (unsigned int __user *) data);
+		return put_user(tmp, (unsigned int __force __user *) data);
 
 	case PTRACE_PEEKUSR:
 		/* read the word at location addr in the USER area. */
@@ -571,7 +570,7 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 
 	case PTRACE_PEEKUSR_AREA:
 	case PTRACE_POKEUSR_AREA:
-		if (copy_from_user(&parea, (void __user *) addr,
+		if (copy_from_user(&parea, (void __force __user *) addr,
 							sizeof(parea)))
 			return -EFAULT;
 		addr = parea.kernel_addr;
@@ -581,10 +580,11 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 			if (request == PTRACE_PEEKUSR_AREA)
 				ret = peek_user_emu31(child, addr, data);
 			else {
-				__u32 tmp;
-				if (get_user (tmp, (__u32 __user *) data))
+				__u32 utmp;
+				if (get_user(utmp,
+					     (__u32 __force __user *) data))
 					return -EFAULT;
-				ret = poke_user_emu31(child, addr, tmp);
+				ret = poke_user_emu31(child, addr, utmp);
 			}
 			if (ret)
 				return ret;
@@ -595,17 +595,19 @@ do_ptrace_emu31(struct task_struct *child, long request, long addr, long data)
 		return 0;
 	case PTRACE_GETEVENTMSG:
 		return put_user((__u32) child->ptrace_message,
-				(unsigned int __user *) data);
+				(unsigned int __force __user *) data);
 	case PTRACE_GETSIGINFO:
 		if (child->last_siginfo == NULL)
 			return -EINVAL;
-		return copy_siginfo_to_user32((compat_siginfo_t __user *) data,
+		return copy_siginfo_to_user32((compat_siginfo_t
+					       __force __user *) data,
 					      child->last_siginfo);
 	case PTRACE_SETSIGINFO:
 		if (child->last_siginfo == NULL)
 			return -EINVAL;
 		return copy_siginfo_from_user32(child->last_siginfo,
-						(compat_siginfo_t __user *) data);
+						(compat_siginfo_t
+						 __force __user *) data);
 	}
 	return ptrace_request(child, request, addr, data);
 }

arch/s390/kernel/reset.S

@@ -1,90 +0,0 @@
-/*
- *  arch/s390/kernel/reset.S
- *
- *    Copyright (C) IBM Corp. 2006
- *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
- *		 Michael Holzheu <holzheu@de.ibm.com>
- */
-
-#include <asm/ptrace.h>
-#include <asm/lowcore.h>
-
-#ifdef CONFIG_64BIT
-
-	.globl	reset_mcck_handler
-reset_mcck_handler:
-	basr	%r13,0
-0:	lg	%r15,__LC_PANIC_STACK	# load panic stack
-	aghi	%r15,-STACK_FRAME_OVERHEAD
-	lg	%r1,s390_reset_mcck_handler-0b(%r13)
-	ltgr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-1:	la	%r1,4095
-	lmg	%r0,%r15,__LC_GPREGS_SAVE_AREA-4095(%r1)
-	lpswe	__LC_MCK_OLD_PSW
-
-	.globl	s390_reset_mcck_handler
-s390_reset_mcck_handler:
-	.quad	0
-
-	.globl	reset_pgm_handler
-reset_pgm_handler:
-	stmg	%r0,%r15,__LC_SAVE_AREA
-	basr	%r13,0
-0:	lg	%r15,__LC_PANIC_STACK	# load panic stack
-	aghi	%r15,-STACK_FRAME_OVERHEAD
-	lg	%r1,s390_reset_pgm_handler-0b(%r13)
-	ltgr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-	lmg	%r0,%r15,__LC_SAVE_AREA
-	lpswe	__LC_PGM_OLD_PSW
-1:	lpswe	disabled_wait_psw-0b(%r13)
-	.globl s390_reset_pgm_handler
-s390_reset_pgm_handler:
-	.quad	0
-	.align	8
-disabled_wait_psw:
-	.quad	0x0002000180000000,0x0000000000000000 + reset_pgm_handler
-
-#else /* CONFIG_64BIT */
-
-	.globl	reset_mcck_handler
-reset_mcck_handler:
-	basr	%r13,0
-0:	l	%r15,__LC_PANIC_STACK	# load panic stack
-	ahi	%r15,-STACK_FRAME_OVERHEAD
-	l	%r1,s390_reset_mcck_handler-0b(%r13)
-	ltr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-1:	lm	%r0,%r15,__LC_GPREGS_SAVE_AREA
-	lpsw	__LC_MCK_OLD_PSW
-
-	.globl	s390_reset_mcck_handler
-s390_reset_mcck_handler:
-	.long	0
-
-	.globl	reset_pgm_handler
-reset_pgm_handler:
-	stm	%r0,%r15,__LC_SAVE_AREA
-	basr	%r13,0
-0:	l	%r15,__LC_PANIC_STACK	# load panic stack
-	ahi	%r15,-STACK_FRAME_OVERHEAD
-	l	%r1,s390_reset_pgm_handler-0b(%r13)
-	ltr	%r1,%r1
-	jz	1f
-	basr	%r14,%r1
-	lm	%r0,%r15,__LC_SAVE_AREA
-	lpsw	__LC_PGM_OLD_PSW
-
-1:	lpsw	disabled_wait_psw-0b(%r13)
-	.globl	s390_reset_pgm_handler
-s390_reset_pgm_handler:
-	.long	0
-disabled_wait_psw:
-	.align 8
-	.long	0x000a0000,0x00000000 + reset_pgm_handler
-
-#endif /* CONFIG_64BIT */

arch/s390/kernel/s390_ext.c

@@ -125,14 +125,12 @@ void do_extint(struct pt_regs *regs, unsigned short code)
 		 * Make sure that the i/o interrupt did not "overtake"
 		 * the last HZ timer interrupt.
 		 */
-		account_ticks();
+		account_ticks(S390_lowcore.int_clock);
 	kstat_cpu(smp_processor_id()).irqs[EXTERNAL_INTERRUPT]++;
         index = ext_hash(code);
 	for (p = ext_int_hash[index]; p; p = p->next) {
-		if (likely(p->code == code)) {
-			if (likely(p->handler))
-				p->handler(code);
-		}
+		if (likely(p->code == code))
+			p->handler(code);
 	}
 	irq_exit();
 	set_irq_regs(old_regs);

arch/s390/kernel/setup.c

@@ -38,6 +38,8 @@
 #include <linux/device.h>
 #include <linux/notifier.h>
 #include <linux/pfn.h>
+#include <linux/ctype.h>
+#include <linux/reboot.h>
 
 #include <asm/uaccess.h>
 #include <asm/system.h>
@@ -49,6 +51,14 @@
 #include <asm/page.h>
 #include <asm/ptrace.h>
 #include <asm/sections.h>
+#include <asm/ebcdic.h>
+#include <asm/compat.h>
+
+long psw_kernel_bits	= (PSW_BASE_BITS | PSW_MASK_DAT | PSW_ASC_PRIMARY |
+			   PSW_MASK_MCHECK | PSW_DEFAULT_KEY);
+long psw_user_bits	= (PSW_BASE_BITS | PSW_MASK_DAT | PSW_ASC_HOME |
+			   PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			   PSW_MASK_PSTATE | PSW_DEFAULT_KEY);
 
 /*
  * User copy operations.
@@ -117,9 +127,9 @@ void __devinit cpu_init (void)
  */
 char vmhalt_cmd[128] = "";
 char vmpoff_cmd[128] = "";
-char vmpanic_cmd[128] = "";
+static char vmpanic_cmd[128] = "";
 
-static inline void strncpy_skip_quote(char *dst, char *src, int n)
+static void strncpy_skip_quote(char *dst, char *src, int n)
 {
         int sx, dx;
 
@@ -275,10 +285,6 @@ static void __init conmode_default(void)
 }
 
 #ifdef CONFIG_SMP
-extern void machine_restart_smp(char *);
-extern void machine_halt_smp(void);
-extern void machine_power_off_smp(void);
-
 void (*_machine_restart)(char *command) = machine_restart_smp;
 void (*_machine_halt)(void) = machine_halt_smp;
 void (*_machine_power_off)(void) = machine_power_off_smp;
@@ -386,6 +392,84 @@ static int __init early_parse_ipldelay(char *p)
 }
 early_param("ipldelay", early_parse_ipldelay);
 
+#ifdef CONFIG_S390_SWITCH_AMODE
+unsigned int switch_amode = 0;
+EXPORT_SYMBOL_GPL(switch_amode);
+
+static void set_amode_and_uaccess(unsigned long user_amode,
+				  unsigned long user32_amode)
+{
+	psw_user_bits = PSW_BASE_BITS | PSW_MASK_DAT | user_amode |
+			PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			PSW_MASK_PSTATE | PSW_DEFAULT_KEY;
+#ifdef CONFIG_COMPAT
+	psw_user32_bits = PSW_BASE32_BITS | PSW_MASK_DAT | user_amode |
+			  PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK |
+			  PSW_MASK_PSTATE | PSW_DEFAULT_KEY;
+	psw32_user_bits = PSW32_BASE_BITS | PSW32_MASK_DAT | user32_amode |
+			  PSW32_MASK_IO | PSW32_MASK_EXT | PSW32_MASK_MCHECK |
+			  PSW32_MASK_PSTATE;
+#endif
+	psw_kernel_bits = PSW_BASE_BITS | PSW_MASK_DAT | PSW_ASC_HOME |
+			  PSW_MASK_MCHECK | PSW_DEFAULT_KEY;
+
+	if (MACHINE_HAS_MVCOS) {
+		printk("mvcos available.\n");
+		memcpy(&uaccess, &uaccess_mvcos_switch, sizeof(uaccess));
+	} else {
+		printk("mvcos not available.\n");
+		memcpy(&uaccess, &uaccess_pt, sizeof(uaccess));
+	}
+}
+
+/*
+ * Switch kernel/user addressing modes?
+ */
+static int __init early_parse_switch_amode(char *p)
+{
+	switch_amode = 1;
+	return 0;
+}
+early_param("switch_amode", early_parse_switch_amode);
+
+#else /* CONFIG_S390_SWITCH_AMODE */
+static inline void set_amode_and_uaccess(unsigned long user_amode,
+					 unsigned long user32_amode)
+{
+}
+#endif /* CONFIG_S390_SWITCH_AMODE */
+
+#ifdef CONFIG_S390_EXEC_PROTECT
+unsigned int s390_noexec = 0;
+EXPORT_SYMBOL_GPL(s390_noexec);
+
+/*
+ * Enable execute protection?
+ */
+static int __init early_parse_noexec(char *p)
+{
+	if (!strncmp(p, "off", 3))
+		return 0;
+	switch_amode = 1;
+	s390_noexec = 1;
+	return 0;
+}
+early_param("noexec", early_parse_noexec);
+#endif /* CONFIG_S390_EXEC_PROTECT */
+
+static void setup_addressing_mode(void)
+{
+	if (s390_noexec) {
+		printk("S390 execute protection active, ");
+		set_amode_and_uaccess(PSW_ASC_SECONDARY, PSW32_ASC_SECONDARY);
+		return;
+	}
+	if (switch_amode) {
+		printk("S390 address spaces switched, ");
+		set_amode_and_uaccess(PSW_ASC_PRIMARY, PSW32_ASC_PRIMARY);
+	}
+}
+
 static void __init
 setup_lowcore(void)
 {
@@ -402,19 +486,21 @@ setup_lowcore(void)
 	lc->restart_psw.mask = PSW_BASE_BITS | PSW_DEFAULT_KEY;
 	lc->restart_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long) restart_int_handler;
-	lc->external_new_psw.mask = PSW_KERNEL_BITS;
+	if (switch_amode)
+		lc->restart_psw.mask |= PSW_ASC_HOME;
+	lc->external_new_psw.mask = psw_kernel_bits;
 	lc->external_new_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long) ext_int_handler;
-	lc->svc_new_psw.mask = PSW_KERNEL_BITS | PSW_MASK_IO | PSW_MASK_EXT;
+	lc->svc_new_psw.mask = psw_kernel_bits | PSW_MASK_IO | PSW_MASK_EXT;
 	lc->svc_new_psw.addr = PSW_ADDR_AMODE | (unsigned long) system_call;
-	lc->program_new_psw.mask = PSW_KERNEL_BITS;
+	lc->program_new_psw.mask = psw_kernel_bits;
 	lc->program_new_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long)pgm_check_handler;
 	lc->mcck_new_psw.mask =
-		PSW_KERNEL_BITS & ~PSW_MASK_MCHECK & ~PSW_MASK_DAT;
+		psw_kernel_bits & ~PSW_MASK_MCHECK & ~PSW_MASK_DAT;
 	lc->mcck_new_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long) mcck_int_handler;
-	lc->io_new_psw.mask = PSW_KERNEL_BITS;
+	lc->io_new_psw.mask = psw_kernel_bits;
 	lc->io_new_psw.addr = PSW_ADDR_AMODE | (unsigned long) io_int_handler;
 	lc->ipl_device = S390_lowcore.ipl_device;
 	lc->jiffy_timer = -1LL;
@@ -439,7 +525,7 @@ setup_lowcore(void)
 static void __init
 setup_resources(void)
 {
-	struct resource *res;
+	struct resource *res, *sub_res;
 	int i;
 
 	code_resource.start = (unsigned long) &_text;
@@ -464,8 +550,38 @@ setup_resources(void)
 		res->start = memory_chunk[i].addr;
 		res->end = memory_chunk[i].addr +  memory_chunk[i].size - 1;
 		request_resource(&iomem_resource, res);
-		request_resource(res, &code_resource);
-		request_resource(res, &data_resource);
+
+		if (code_resource.start >= res->start  &&
+			code_resource.start <= res->end &&
+			code_resource.end > res->end) {
+			sub_res = alloc_bootmem_low(sizeof(struct resource));
+			memcpy(sub_res, &code_resource,
+				sizeof(struct resource));
+			sub_res->end = res->end;
+			code_resource.start = res->end + 1;
+			request_resource(res, sub_res);
+		}
+
+		if (code_resource.start >= res->start &&
+			code_resource.start <= res->end &&
+			code_resource.end <= res->end)
+			request_resource(res, &code_resource);
+
+		if (data_resource.start >= res->start &&
+			data_resource.start <= res->end &&
+			data_resource.end > res->end) {
+			sub_res = alloc_bootmem_low(sizeof(struct resource));
+			memcpy(sub_res, &data_resource,
+				sizeof(struct resource));
+			sub_res->end = res->end;
+			data_resource.start = res->end + 1;
+			request_resource(res, sub_res);
+		}
+
+		if (data_resource.start >= res->start &&
+			data_resource.start <= res->end &&
+			data_resource.end <= res->end)
+			request_resource(res, &data_resource);
 	}
 }
 
@@ -495,16 +611,13 @@ static void __init setup_memory_end(void)
 	}
 	if (!memory_end)
 		memory_end = memory_size;
-	if (real_size > memory_end)
-		printk("More memory detected than supported. Unused: %luk\n",
-		       (real_size - memory_end) >> 10);
 }
 
 static void __init
 setup_memory(void)
 {
         unsigned long bootmap_size;
-	unsigned long start_pfn, end_pfn, init_pfn;
+	unsigned long start_pfn, end_pfn;
 	int i;
 
 	/*
@@ -514,10 +627,6 @@ setup_memory(void)
 	start_pfn = PFN_UP(__pa(&_end));
 	end_pfn = max_pfn = PFN_DOWN(memory_end);
 
-	/* Initialize storage key for kernel pages */
-	for (init_pfn = 0 ; init_pfn < start_pfn; init_pfn++)
-		page_set_storage_key(init_pfn << PAGE_SHIFT, PAGE_DEFAULT_KEY);
-
 #ifdef CONFIG_BLK_DEV_INITRD
 	/*
 	 * Move the initrd in case the bitmap of the bootmem allocater
@@ -651,6 +760,7 @@ setup_arch(char **cmdline_p)
 	parse_early_param();
 
 	setup_memory_end();
+	setup_addressing_mode();
 	setup_memory();
 	setup_resources();
 	setup_lowcore();
@@ -694,6 +804,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
         struct cpuinfo_S390 *cpuinfo;
 	unsigned long n = (unsigned long) v - 1;
 
+	s390_adjust_jiffies();
 	preempt_disable();
 	if (!n) {
 		seq_printf(m, "vendor_id       : IBM/S390\n"

arch/s390/kernel/signal.c

@@ -119,7 +119,7 @@ static int save_sigregs(struct pt_regs *regs, _sigregs __user *sregs)
 
 	/* Copy a 'clean' PSW mask to the user to avoid leaking
 	   information about whether PER is currently on.  */
-	user_sregs.regs.psw.mask = PSW_MASK_MERGE(PSW_USER_BITS, regs->psw.mask);
+	user_sregs.regs.psw.mask = PSW_MASK_MERGE(psw_user_bits, regs->psw.mask);
 	user_sregs.regs.psw.addr = regs->psw.addr;
 	memcpy(&user_sregs.regs.gprs, &regs->gprs, sizeof(sregs->regs.gprs));
 	memcpy(&user_sregs.regs.acrs, current->thread.acrs,

arch/s390/kernel/smp.c

@@ -22,23 +22,23 @@
 
 #include <linux/module.h>
 #include <linux/init.h>
-
 #include <linux/mm.h>
 #include <linux/spinlock.h>
 #include <linux/kernel_stat.h>
 #include <linux/smp_lock.h>
-
 #include <linux/delay.h>
 #include <linux/cache.h>
 #include <linux/interrupt.h>
 #include <linux/cpu.h>
-
+#include <linux/timex.h>
+#include <asm/setup.h>
 #include <asm/sigp.h>
 #include <asm/pgalloc.h>
 #include <asm/irq.h>
 #include <asm/s390_ext.h>
 #include <asm/cpcmd.h>
 #include <asm/tlbflush.h>
+#include <asm/timer.h>
 
 extern volatile int __cpu_logical_map[];
 
@@ -53,12 +53,6 @@ cpumask_t cpu_possible_map = CPU_MASK_NONE;
 
 static struct task_struct *current_set[NR_CPUS];
 
-/*
- * Reboot, halt and power_off routines for SMP.
- */
-extern char vmhalt_cmd[];
-extern char vmpoff_cmd[];
-
 static void smp_ext_bitcall(int, ec_bit_sig);
 static void smp_ext_bitcall_others(ec_bit_sig);
 
@@ -200,7 +194,7 @@ int smp_call_function_on(void (*func) (void *info), void *info,
 }
 EXPORT_SYMBOL(smp_call_function_on);
 
-static inline void do_send_stop(void)
+static void do_send_stop(void)
 {
         int cpu, rc;
 
@@ -214,7 +208,7 @@ static inline void do_send_stop(void)
 	}
 }
 
-static inline void do_store_status(void)
+static void do_store_status(void)
 {
         int cpu, rc;
 
@@ -230,7 +224,7 @@ static inline void do_store_status(void)
         }
 }
 
-static inline void do_wait_for_stop(void)
+static void do_wait_for_stop(void)
 {
 	int cpu;
 
@@ -250,7 +244,7 @@ static inline void do_wait_for_stop(void)
 void smp_send_stop(void)
 {
 	/* Disable all interrupts/machine checks */
-	__load_psw_mask(PSW_KERNEL_BITS & ~PSW_MASK_MCHECK);
+	__load_psw_mask(psw_kernel_bits & ~PSW_MASK_MCHECK);
 
         /* write magic number to zero page (absolute 0) */
 	lowcore_ptr[smp_processor_id()]->panic_magic = __PANIC_MAGIC;
@@ -298,7 +292,7 @@ void machine_power_off_smp(void)
  * cpus are handled.
  */
 
-void do_ext_call_interrupt(__u16 code)
+static void do_ext_call_interrupt(__u16 code)
 {
         unsigned long bits;
 
@@ -385,7 +379,7 @@ struct ec_creg_mask_parms {
 /*
  * callback for setting/clearing control bits
  */
-void smp_ctl_bit_callback(void *info) {
+static void smp_ctl_bit_callback(void *info) {
 	struct ec_creg_mask_parms *pp = info;
 	unsigned long cregs[16];
 	int i;
@@ -458,17 +452,15 @@ __init smp_count_cpus(void)
 /*
  *      Activate a secondary processor.
  */
-extern void init_cpu_timer(void);
-extern void init_cpu_vtimer(void);
-
 int __devinit start_secondary(void *cpuvoid)
 {
         /* Setup the cpu */
         cpu_init();
 	preempt_disable();
-        /* init per CPU timer */
+	/* Enable TOD clock interrupts on the secondary cpu. */
         init_cpu_timer();
 #ifdef CONFIG_VIRT_TIMER
+	/* Enable cpu timer interrupts on the secondary cpu. */
         init_cpu_vtimer();
 #endif
 	/* Enable pfault pseudo page faults on this cpu. */
@@ -542,7 +534,7 @@ smp_put_cpu(int cpu)
 	spin_unlock_irqrestore(&smp_reserve_lock, flags);
 }
 
-static inline int
+static int
 cpu_stopped(int cpu)
 {
 	__u32 status;

arch/s390/kernel/stacktrace.c

@@ -11,11 +11,11 @@
 #include <linux/stacktrace.h>
 #include <linux/kallsyms.h>
 
-static inline unsigned long save_context_stack(struct stack_trace *trace,
-					       unsigned int *skip,
-					       unsigned long sp,
-					       unsigned long low,
-					       unsigned long high)
+static unsigned long save_context_stack(struct stack_trace *trace,
+					unsigned int *skip,
+					unsigned long sp,
+					unsigned long low,
+					unsigned long high)
 {
 	struct stack_frame *sf;
 	struct pt_regs *regs;

arch/s390/kernel/time.c

@@ -37,11 +37,15 @@
 #include <asm/irq.h>
 #include <asm/irq_regs.h>
 #include <asm/timer.h>
+#include <asm/etr.h>
 
 /* change this if you have some constant time drift */
 #define USECS_PER_JIFFY     ((unsigned long) 1000000/HZ)
 #define CLK_TICKS_PER_JIFFY ((unsigned long) USECS_PER_JIFFY << 12)
 
+/* The value of the TOD clock for 1.1.1970. */
+#define TOD_UNIX_EPOCH 0x7d91048bca000000ULL
+
 /*
  * Create a small time difference between the timer interrupts
  * on the different cpus to avoid lock contention.
@@ -51,6 +55,7 @@
 #define TICK_SIZE tick
 
 static ext_int_info_t ext_int_info_cc;
+static ext_int_info_t ext_int_etr_cc;
 static u64 init_timer_cc;
 static u64 jiffies_timer_cc;
 static u64 xtime_cc;
@@ -89,29 +94,21 @@ void tod_to_timeval(__u64 todval, struct timespec *xtime)
 #define s390_do_profile()	do { ; } while(0)
 #endif /* CONFIG_PROFILING */
 
-
 /*
- * timer_interrupt() needs to keep up the real-time clock,
- * as well as call the "do_timer()" routine every clocktick
+ * Advance the per cpu tick counter up to the time given with the
+ * "time" argument. The per cpu update consists of accounting
+ * the virtual cpu time, calling update_process_times and calling
+ * the profiling hook. If xtime is before time it is advanced as well.
  */
-void account_ticks(void)
+void account_ticks(u64 time)
 {
-	__u64 tmp;
 	__u32 ticks;
+	__u64 tmp;
 
 	/* Calculate how many ticks have passed. */
-	if (S390_lowcore.int_clock < S390_lowcore.jiffy_timer) {
-		/*
-		 * We have to program the clock comparator even if
-		 * no tick has passed. That happens if e.g. an i/o
-		 * interrupt wakes up an idle processor that has
-		 * switched off its hz timer.
-		 */
-		tmp = S390_lowcore.jiffy_timer + CPU_DEVIATION;
-		asm volatile ("SCKC %0" : : "m" (tmp));
+	if (time < S390_lowcore.jiffy_timer)
 		return;
-	}
-	tmp = S390_lowcore.int_clock - S390_lowcore.jiffy_timer;
+	tmp = time - S390_lowcore.jiffy_timer;
 	if (tmp >= 2*CLK_TICKS_PER_JIFFY) {  /* more than two ticks ? */
 		ticks = __div(tmp, CLK_TICKS_PER_JIFFY) + 1;
 		S390_lowcore.jiffy_timer +=
@@ -124,10 +121,6 @@ void account_ticks(void)
 		S390_lowcore.jiffy_timer += CLK_TICKS_PER_JIFFY;
 	}
 
-	/* set clock comparator for next tick */
-	tmp = S390_lowcore.jiffy_timer + CPU_DEVIATION;
-        asm volatile ("SCKC %0" : : "m" (tmp));
-
 #ifdef CONFIG_SMP
 	/*
 	 * Do not rely on the boot cpu to do the calls to do_timer.
@@ -173,7 +166,7 @@ int sysctl_hz_timer = 1;
  * Stop the HZ tick on the current CPU.
  * Only cpu_idle may call this function.
  */
-static inline void stop_hz_timer(void)
+static void stop_hz_timer(void)
 {
 	unsigned long flags;
 	unsigned long seq, next;
@@ -210,20 +203,21 @@ static inline void stop_hz_timer(void)
 		if (timer >= jiffies_timer_cc)
 			todval = timer;
 	}
-	asm volatile ("SCKC %0" : : "m" (todval));
+	set_clock_comparator(todval);
 }
 
 /*
  * Start the HZ tick on the current CPU.
  * Only cpu_idle may call this function.
  */
-static inline void start_hz_timer(void)
+static void start_hz_timer(void)
 {
 	BUG_ON(!in_interrupt());
 
 	if (!cpu_isset(smp_processor_id(), nohz_cpu_mask))
 		return;
-	account_ticks();
+	account_ticks(get_clock());
+	set_clock_comparator(S390_lowcore.jiffy_timer + CPU_DEVIATION);
 	cpu_clear(smp_processor_id(), nohz_cpu_mask);
 }
 
@@ -245,7 +239,7 @@ static struct notifier_block nohz_idle_nb = {
 	.notifier_call = nohz_idle_notify,
 };
 
-void __init nohz_init(void)
+static void __init nohz_init(void)
 {
 	if (register_idle_notifier(&nohz_idle_nb))
 		panic("Couldn't register idle notifier");
@@ -254,24 +248,57 @@ void __init nohz_init(void)
 #endif
 
 /*
- * Start the clock comparator on the current CPU.
+ * Set up per cpu jiffy timer and set the clock comparator.
+ */
+static void setup_jiffy_timer(void)
+{
+	/* Set up clock comparator to next jiffy. */
+	S390_lowcore.jiffy_timer =
+		jiffies_timer_cc + (jiffies_64 + 1) * CLK_TICKS_PER_JIFFY;
+	set_clock_comparator(S390_lowcore.jiffy_timer + CPU_DEVIATION);
+}
+
+/*
+ * Set up lowcore and control register of the current cpu to
+ * enable TOD clock and clock comparator interrupts.
  */
 void init_cpu_timer(void)
 {
-	unsigned long cr0;
-	__u64 timer;
+	setup_jiffy_timer();
 
-	timer = jiffies_timer_cc + jiffies_64 * CLK_TICKS_PER_JIFFY;
-	S390_lowcore.jiffy_timer = timer + CLK_TICKS_PER_JIFFY;
-	timer += CLK_TICKS_PER_JIFFY + CPU_DEVIATION;
-	asm volatile ("SCKC %0" : : "m" (timer));
-        /* allow clock comparator timer interrupt */
-	__ctl_store(cr0, 0, 0);
-        cr0 |= 0x800;
-	__ctl_load(cr0, 0, 0);
+	/* Enable clock comparator timer interrupt. */
+	__ctl_set_bit(0,11);
+
+	/* Always allow ETR external interrupts, even without an ETR. */
+	__ctl_set_bit(0, 4);
 }
 
-extern void vtime_init(void);
+static void clock_comparator_interrupt(__u16 code)
+{
+	/* set clock comparator for next tick */
+	set_clock_comparator(S390_lowcore.jiffy_timer + CPU_DEVIATION);
+}
+
+static void etr_reset(void);
+static void etr_init(void);
+static void etr_ext_handler(__u16);
+
+/*
+ * Get the TOD clock running.
+ */
+static u64 __init reset_tod_clock(void)
+{
+	u64 time;
+
+	etr_reset();
+	if (store_clock(&time) == 0)
+		return time;
+	/* TOD clock not running. Set the clock to Unix Epoch. */
+	if (set_clock(TOD_UNIX_EPOCH) != 0 || store_clock(&time) != 0)
+		panic("TOD clock not operational.");
+
+	return TOD_UNIX_EPOCH;
+}
 
 static cycle_t read_tod_clock(void)
 {
@@ -295,48 +322,31 @@ static struct clocksource clocksource_tod = {
  */
 void __init time_init(void)
 {
-	__u64 set_time_cc;
-	int cc;
-
-        /* kick the TOD clock */
-	asm volatile(
-		"	stck	0(%2)\n"
-		"	ipm	%0\n"
-		"	srl	%0,28"
-		: "=d" (cc), "=m" (init_timer_cc)
-		: "a" (&init_timer_cc) : "cc");
-        switch (cc) {
-        case 0: /* clock in set state: all is fine */
-                break;
-        case 1: /* clock in non-set state: FIXME */
-                printk("time_init: TOD clock in non-set state\n");
-                break;
-        case 2: /* clock in error state: FIXME */
-                printk("time_init: TOD clock in error state\n");
-                break;
-        case 3: /* clock in stopped or not-operational state: FIXME */
-                printk("time_init: TOD clock stopped/non-operational\n");
-                break;
-        }
+	init_timer_cc = reset_tod_clock();
+	xtime_cc = init_timer_cc + CLK_TICKS_PER_JIFFY;
 	jiffies_timer_cc = init_timer_cc - jiffies_64 * CLK_TICKS_PER_JIFFY;
 
 	/* set xtime */
-	xtime_cc = init_timer_cc + CLK_TICKS_PER_JIFFY;
-	set_time_cc = init_timer_cc - 0x8126d60e46000000LL +
-		(0x3c26700LL*1000000*4096);
-        tod_to_timeval(set_time_cc, &xtime);
+	tod_to_timeval(init_timer_cc - TOD_UNIX_EPOCH, &xtime);
         set_normalized_timespec(&wall_to_monotonic,
                                 -xtime.tv_sec, -xtime.tv_nsec);
 
 	/* request the clock comparator external interrupt */
-	if (register_early_external_interrupt(0x1004, NULL,
+	if (register_early_external_interrupt(0x1004,
+					      clock_comparator_interrupt,
 					      &ext_int_info_cc) != 0)
                 panic("Couldn't request external interrupt 0x1004");
 
 	if (clocksource_register(&clocksource_tod) != 0)
 		panic("Could not register TOD clock source");
 
-        init_cpu_timer();
+	/* request the etr external interrupt */
+	if (register_early_external_interrupt(0x1406, etr_ext_handler,
+					      &ext_int_etr_cc) != 0)
+		panic("Couldn't request external interrupt 0x1406");
+
+	/* Enable TOD clock interrupts on the boot cpu. */
+	init_cpu_timer();
 
 #ifdef CONFIG_NO_IDLE_HZ
 	nohz_init();
@@ -345,5 +355,1048 @@ void __init time_init(void)
 #ifdef CONFIG_VIRT_TIMER
 	vtime_init();
 #endif
+	etr_init();
+}
+
+/*
+ * External Time Reference (ETR) code.
+ */
+static int etr_port0_online;
+static int etr_port1_online;
+
+static int __init early_parse_etr(char *p)
+{
+	if (strncmp(p, "off", 3) == 0)
+		etr_port0_online = etr_port1_online = 0;
+	else if (strncmp(p, "port0", 5) == 0)
+		etr_port0_online = 1;
+	else if (strncmp(p, "port1", 5) == 0)
+		etr_port1_online = 1;
+	else if (strncmp(p, "on", 2) == 0)
+		etr_port0_online = etr_port1_online = 1;
+	return 0;
+}
+early_param("etr", early_parse_etr);
+
+enum etr_event {
+	ETR_EVENT_PORT0_CHANGE,
+	ETR_EVENT_PORT1_CHANGE,
+	ETR_EVENT_PORT_ALERT,
+	ETR_EVENT_SYNC_CHECK,
+	ETR_EVENT_SWITCH_LOCAL,
+	ETR_EVENT_UPDATE,
+};
+
+enum etr_flags {
+	ETR_FLAG_ENOSYS,
+	ETR_FLAG_EACCES,
+	ETR_FLAG_STEAI,
+};
+
+/*
+ * Valid bit combinations of the eacr register are (x = don't care):
+ * e0 e1 dp p0 p1 ea es sl
+ *  0  0  x  0	0  0  0  0  initial, disabled state
+ *  0  0  x  0	1  1  0  0  port 1 online
+ *  0  0  x  1	0  1  0  0  port 0 online
+ *  0  0  x  1	1  1  0  0  both ports online
+ *  0  1  x  0	1  1  0  0  port 1 online and usable, ETR or PPS mode
+ *  0  1  x  0	1  1  0  1  port 1 online, usable and ETR mode
+ *  0  1  x  0	1  1  1  0  port 1 online, usable, PPS mode, in-sync
+ *  0  1  x  0	1  1  1  1  port 1 online, usable, ETR mode, in-sync
+ *  0  1  x  1	1  1  0  0  both ports online, port 1 usable
+ *  0  1  x  1	1  1  1  0  both ports online, port 1 usable, PPS mode, in-sync
+ *  0  1  x  1	1  1  1  1  both ports online, port 1 usable, ETR mode, in-sync
+ *  1  0  x  1	0  1  0  0  port 0 online and usable, ETR or PPS mode
+ *  1  0  x  1	0  1  0  1  port 0 online, usable and ETR mode
+ *  1  0  x  1	0  1  1  0  port 0 online, usable, PPS mode, in-sync
+ *  1  0  x  1	0  1  1  1  port 0 online, usable, ETR mode, in-sync
+ *  1  0  x  1	1  1  0  0  both ports online, port 0 usable
+ *  1  0  x  1	1  1  1  0  both ports online, port 0 usable, PPS mode, in-sync
+ *  1  0  x  1	1  1  1  1  both ports online, port 0 usable, ETR mode, in-sync
+ *  1  1  x  1	1  1  1  0  both ports online & usable, ETR, in-sync
+ *  1  1  x  1	1  1  1  1  both ports online & usable, ETR, in-sync
+ */
+static struct etr_eacr etr_eacr;
+static u64 etr_tolec;			/* time of last eacr update */
+static unsigned long etr_flags;
+static struct etr_aib etr_port0;
+static int etr_port0_uptodate;
+static struct etr_aib etr_port1;
+static int etr_port1_uptodate;
+static unsigned long etr_events;
+static struct timer_list etr_timer;
+static struct tasklet_struct etr_tasklet;
+static DEFINE_PER_CPU(atomic_t, etr_sync_word);
+
+static void etr_timeout(unsigned long dummy);
+static void etr_tasklet_fn(unsigned long dummy);
+
+/*
+ * The etr get_clock function. It will write the current clock value
+ * to the clock pointer and return 0 if the clock is in sync with the
+ * external time source. If the clock mode is local it will return
+ * -ENOSYS and -EAGAIN if the clock is not in sync with the external
+ * reference. This function is what ETR is all about..
+ */
+int get_sync_clock(unsigned long long *clock)
+{
+	atomic_t *sw_ptr;
+	unsigned int sw0, sw1;
+
+	sw_ptr = &get_cpu_var(etr_sync_word);
+	sw0 = atomic_read(sw_ptr);
+	*clock = get_clock();
+	sw1 = atomic_read(sw_ptr);
+	put_cpu_var(etr_sync_sync);
+	if (sw0 == sw1 && (sw0 & 0x80000000U))
+		/* Success: time is in sync. */
+		return 0;
+	if (test_bit(ETR_FLAG_ENOSYS, &etr_flags))
+		return -ENOSYS;
+	if (test_bit(ETR_FLAG_EACCES, &etr_flags))
+		return -EACCES;
+	return -EAGAIN;
+}
+EXPORT_SYMBOL(get_sync_clock);
+
+/*
+ * Make get_sync_clock return -EAGAIN.
+ */
+static void etr_disable_sync_clock(void *dummy)
+{
+	atomic_t *sw_ptr = &__get_cpu_var(etr_sync_word);
+	/*
+	 * Clear the in-sync bit 2^31. All get_sync_clock calls will
+	 * fail until the sync bit is turned back on. In addition
+	 * increase the "sequence" counter to avoid the race of an
+	 * etr event and the complete recovery against get_sync_clock.
+	 */
+	atomic_clear_mask(0x80000000, sw_ptr);
+	atomic_inc(sw_ptr);
+}
+
+/*
+ * Make get_sync_clock return 0 again.
+ * Needs to be called from a context disabled for preemption.
+ */
+static void etr_enable_sync_clock(void)
+{
+	atomic_t *sw_ptr = &__get_cpu_var(etr_sync_word);
+	atomic_set_mask(0x80000000, sw_ptr);
+}
+
+/*
+ * Reset ETR attachment.
+ */
+static void etr_reset(void)
+{
+	etr_eacr =  (struct etr_eacr) {
+		.e0 = 0, .e1 = 0, ._pad0 = 4, .dp = 0,
+		.p0 = 0, .p1 = 0, ._pad1 = 0, .ea = 0,
+		.es = 0, .sl = 0 };
+	if (etr_setr(&etr_eacr) == 0)
+		etr_tolec = get_clock();
+	else {
+		set_bit(ETR_FLAG_ENOSYS, &etr_flags);
+		if (etr_port0_online || etr_port1_online) {
+			printk(KERN_WARNING "Running on non ETR capable "
+			       "machine, only local mode available.\n");
+			etr_port0_online = etr_port1_online = 0;
+		}
+	}
+}
+
+static void etr_init(void)
+{
+	struct etr_aib aib;
+
+	if (test_bit(ETR_FLAG_ENOSYS, &etr_flags))
+		return;
+	/* Check if this machine has the steai instruction. */
+	if (etr_steai(&aib, ETR_STEAI_STEPPING_PORT) == 0)
+		set_bit(ETR_FLAG_STEAI, &etr_flags);
+	setup_timer(&etr_timer, etr_timeout, 0UL);
+	tasklet_init(&etr_tasklet, etr_tasklet_fn, 0);
+	if (!etr_port0_online && !etr_port1_online)
+		set_bit(ETR_FLAG_EACCES, &etr_flags);
+	if (etr_port0_online) {
+		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	}
+	if (etr_port1_online) {
+		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	}
+}
+
+/*
+ * Two sorts of ETR machine checks. The architecture reads:
+ * "When a machine-check niterruption occurs and if a switch-to-local or
+ *  ETR-sync-check interrupt request is pending but disabled, this pending
+ *  disabled interruption request is indicated and is cleared".
+ * Which means that we can get etr_switch_to_local events from the machine
+ * check handler although the interruption condition is disabled. Lovely..
+ */
+
+/*
+ * Switch to local machine check. This is called when the last usable
+ * ETR port goes inactive. After switch to local the clock is not in sync.
+ */
+void etr_switch_to_local(void)
+{
+	if (!etr_eacr.sl)
+		return;
+	etr_disable_sync_clock(NULL);
+	set_bit(ETR_EVENT_SWITCH_LOCAL, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+/*
+ * ETR sync check machine check. This is called when the ETR OTE and the
+ * local clock OTE are farther apart than the ETR sync check tolerance.
+ * After a ETR sync check the clock is not in sync. The machine check
+ * is broadcasted to all cpus at the same time.
+ */
+void etr_sync_check(void)
+{
+	if (!etr_eacr.es)
+		return;
+	etr_disable_sync_clock(NULL);
+	set_bit(ETR_EVENT_SYNC_CHECK, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+/*
+ * ETR external interrupt. There are two causes:
+ * 1) port state change, check the usability of the port
+ * 2) port alert, one of the ETR-data-validity bits (v1-v2 bits of the
+ *    sldr-status word) or ETR-data word 1 (edf1) or ETR-data word 3 (edf3)
+ *    or ETR-data word 4 (edf4) has changed.
+ */
+static void etr_ext_handler(__u16 code)
+{
+	struct etr_interruption_parameter *intparm =
+		(struct etr_interruption_parameter *) &S390_lowcore.ext_params;
+
+	if (intparm->pc0)
+		/* ETR port 0 state change. */
+		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
+	if (intparm->pc1)
+		/* ETR port 1 state change. */
+		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
+	if (intparm->eai)
+		/*
+		 * ETR port alert on either port 0, 1 or both.
+		 * Both ports are not up-to-date now.
+		 */
+		set_bit(ETR_EVENT_PORT_ALERT, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+static void etr_timeout(unsigned long dummy)
+{
+	set_bit(ETR_EVENT_UPDATE, &etr_events);
+	tasklet_hi_schedule(&etr_tasklet);
+}
+
+/*
+ * Check if the etr mode is pss.
+ */
+static inline int etr_mode_is_pps(struct etr_eacr eacr)
+{
+	return eacr.es && !eacr.sl;
+}
+
+/*
+ * Check if the etr mode is etr.
+ */
+static inline int etr_mode_is_etr(struct etr_eacr eacr)
+{
+	return eacr.es && eacr.sl;
+}
+
+/*
+ * Check if the port can be used for TOD synchronization.
+ * For PPS mode the port has to receive OTEs. For ETR mode
+ * the port has to receive OTEs, the ETR stepping bit has to
+ * be zero and the validity bits for data frame 1, 2, and 3
+ * have to be 1.
+ */
+static int etr_port_valid(struct etr_aib *aib, int port)
+{
+	unsigned int psc;
+
+	/* Check that this port is receiving OTEs. */
+	if (aib->tsp == 0)
+		return 0;
+
+	psc = port ? aib->esw.psc1 : aib->esw.psc0;
+	if (psc == etr_lpsc_pps_mode)
+		return 1;
+	if (psc == etr_lpsc_operational_step)
+		return !aib->esw.y && aib->slsw.v1 &&
+			aib->slsw.v2 && aib->slsw.v3;
+	return 0;
+}
+
+/*
+ * Check if two ports are on the same network.
+ */
+static int etr_compare_network(struct etr_aib *aib1, struct etr_aib *aib2)
+{
+	// FIXME: any other fields we have to compare?
+	return aib1->edf1.net_id == aib2->edf1.net_id;
+}
+
+/*
+ * Wrapper for etr_stei that converts physical port states
+ * to logical port states to be consistent with the output
+ * of stetr (see etr_psc vs. etr_lpsc).
+ */
+static void etr_steai_cv(struct etr_aib *aib, unsigned int func)
+{
+	BUG_ON(etr_steai(aib, func) != 0);
+	/* Convert port state to logical port state. */
+	if (aib->esw.psc0 == 1)
+		aib->esw.psc0 = 2;
+	else if (aib->esw.psc0 == 0 && aib->esw.p == 0)
+		aib->esw.psc0 = 1;
+	if (aib->esw.psc1 == 1)
+		aib->esw.psc1 = 2;
+	else if (aib->esw.psc1 == 0 && aib->esw.p == 1)
+		aib->esw.psc1 = 1;
+}
+
+/*
+ * Check if the aib a2 is still connected to the same attachment as
+ * aib a1, the etv values differ by one and a2 is valid.
+ */
+static int etr_aib_follows(struct etr_aib *a1, struct etr_aib *a2, int p)
+{
+	int state_a1, state_a2;
+
+	/* Paranoia check: e0/e1 should better be the same. */
+	if (a1->esw.eacr.e0 != a2->esw.eacr.e0 ||
+	    a1->esw.eacr.e1 != a2->esw.eacr.e1)
+		return 0;
+
+	/* Still connected to the same etr ? */
+	state_a1 = p ? a1->esw.psc1 : a1->esw.psc0;
+	state_a2 = p ? a2->esw.psc1 : a2->esw.psc0;
+	if (state_a1 == etr_lpsc_operational_step) {
+		if (state_a2 != etr_lpsc_operational_step ||
+		    a1->edf1.net_id != a2->edf1.net_id ||
+		    a1->edf1.etr_id != a2->edf1.etr_id ||
+		    a1->edf1.etr_pn != a2->edf1.etr_pn)
+			return 0;
+	} else if (state_a2 != etr_lpsc_pps_mode)
+		return 0;
+
+	/* The ETV value of a2 needs to be ETV of a1 + 1. */
+	if (a1->edf2.etv + 1 != a2->edf2.etv)
+		return 0;
+
+	if (!etr_port_valid(a2, p))
+		return 0;
+
+	return 1;
+}
+
+/*
+ * The time is "clock". xtime is what we think the time is.
+ * Adjust the value by a multiple of jiffies and add the delta to ntp.
+ * "delay" is an approximation how long the synchronization took. If
+ * the time correction is positive, then "delay" is subtracted from
+ * the time difference and only the remaining part is passed to ntp.
+ */
+static void etr_adjust_time(unsigned long long clock, unsigned long long delay)
+{
+	unsigned long long delta, ticks;
+	struct timex adjust;
+
+	/*
+	 * We don't have to take the xtime lock because the cpu
+	 * executing etr_adjust_time is running disabled in
+	 * tasklet context and all other cpus are looping in
+	 * etr_sync_cpu_start.
+	 */
+	if (clock > xtime_cc) {
+		/* It is later than we thought. */
+		delta = ticks = clock - xtime_cc;
+		delta = ticks = (delta < delay) ? 0 : delta - delay;
+		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
+		init_timer_cc = init_timer_cc + delta;
+		jiffies_timer_cc = jiffies_timer_cc + delta;
+		xtime_cc = xtime_cc + delta;
+		adjust.offset = ticks * (1000000 / HZ);
+	} else {
+		/* It is earlier than we thought. */
+		delta = ticks = xtime_cc - clock;
+		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
+		init_timer_cc = init_timer_cc - delta;
+		jiffies_timer_cc = jiffies_timer_cc - delta;
+		xtime_cc = xtime_cc - delta;
+		adjust.offset = -ticks * (1000000 / HZ);
+	}
+	if (adjust.offset != 0) {
+		printk(KERN_NOTICE "etr: time adjusted by %li micro-seconds\n",
+		       adjust.offset);
+		adjust.modes = ADJ_OFFSET_SINGLESHOT;
+		do_adjtimex(&adjust);
+	}
+}
+
+static void etr_sync_cpu_start(void *dummy)
+{
+	int *in_sync = dummy;
+
+	etr_enable_sync_clock();
+	/*
+	 * This looks like a busy wait loop but it isn't. etr_sync_cpus
+	 * is called on all other cpus while the TOD clocks is stopped.
+	 * __udelay will stop the cpu on an enabled wait psw until the
+	 * TOD is running again.
+	 */
+	while (*in_sync == 0)
+		__udelay(1);
+	if (*in_sync != 1)
+		/* Didn't work. Clear per-cpu in sync bit again. */
+		etr_disable_sync_clock(NULL);
+	/*
+	 * This round of TOD syncing is done. Set the clock comparator
+	 * to the next tick and let the processor continue.
+	 */
+	setup_jiffy_timer();
+}
+
+static void etr_sync_cpu_end(void *dummy)
+{
+}
+
+/*
+ * Sync the TOD clock using the port refered to by aibp. This port
+ * has to be enabled and the other port has to be disabled. The
+ * last eacr update has to be more than 1.6 seconds in the past.
+ */
+static int etr_sync_clock(struct etr_aib *aib, int port)
+{
+	struct etr_aib *sync_port;
+	unsigned long long clock, delay;
+	int in_sync, follows;
+	int rc;
+
+	/* Check if the current aib is adjacent to the sync port aib. */
+	sync_port = (port == 0) ? &etr_port0 : &etr_port1;
+	follows = etr_aib_follows(sync_port, aib, port);
+	memcpy(sync_port, aib, sizeof(*aib));
+	if (!follows)
+		return -EAGAIN;
+
+	/*
+	 * Catch all other cpus and make them wait until we have
+	 * successfully synced the clock. smp_call_function will
+	 * return after all other cpus are in etr_sync_cpu_start.
+	 */
+	in_sync = 0;
+	preempt_disable();
+	smp_call_function(etr_sync_cpu_start,&in_sync,0,0);
+	local_irq_disable();
+	etr_enable_sync_clock();
+
+	/* Set clock to next OTE. */
+	__ctl_set_bit(14, 21);
+	__ctl_set_bit(0, 29);
+	clock = ((unsigned long long) (aib->edf2.etv + 1)) << 32;
+	if (set_clock(clock) == 0) {
+		__udelay(1);	/* Wait for the clock to start. */
+		__ctl_clear_bit(0, 29);
+		__ctl_clear_bit(14, 21);
+		etr_stetr(aib);
+		/* Adjust Linux timing variables. */
+		delay = (unsigned long long)
+			(aib->edf2.etv - sync_port->edf2.etv) << 32;
+		etr_adjust_time(clock, delay);
+		setup_jiffy_timer();
+		/* Verify that the clock is properly set. */
+		if (!etr_aib_follows(sync_port, aib, port)) {
+			/* Didn't work. */
+			etr_disable_sync_clock(NULL);
+			in_sync = -EAGAIN;
+			rc = -EAGAIN;
+		} else {
+			in_sync = 1;
+			rc = 0;
+		}
+	} else {
+		/* Could not set the clock ?!? */
+		__ctl_clear_bit(0, 29);
+		__ctl_clear_bit(14, 21);
+		etr_disable_sync_clock(NULL);
+		in_sync = -EAGAIN;
+		rc = -EAGAIN;
+	}
+	local_irq_enable();
+	smp_call_function(etr_sync_cpu_end,NULL,0,0);
+	preempt_enable();
+	return rc;
+}
+
+/*
+ * Handle the immediate effects of the different events.
+ * The port change event is used for online/offline changes.
+ */
+static struct etr_eacr etr_handle_events(struct etr_eacr eacr)
+{
+	if (test_and_clear_bit(ETR_EVENT_SYNC_CHECK, &etr_events))
+		eacr.es = 0;
+	if (test_and_clear_bit(ETR_EVENT_SWITCH_LOCAL, &etr_events))
+		eacr.es = eacr.sl = 0;
+	if (test_and_clear_bit(ETR_EVENT_PORT_ALERT, &etr_events))
+		etr_port0_uptodate = etr_port1_uptodate = 0;
+
+	if (test_and_clear_bit(ETR_EVENT_PORT0_CHANGE, &etr_events)) {
+		if (eacr.e0)
+			/*
+			 * Port change of an enabled port. We have to
+			 * assume that this can have caused an stepping
+			 * port switch.
+			 */
+			etr_tolec = get_clock();
+		eacr.p0 = etr_port0_online;
+		if (!eacr.p0)
+			eacr.e0 = 0;
+		etr_port0_uptodate = 0;
+	}
+	if (test_and_clear_bit(ETR_EVENT_PORT1_CHANGE, &etr_events)) {
+		if (eacr.e1)
+			/*
+			 * Port change of an enabled port. We have to
+			 * assume that this can have caused an stepping
+			 * port switch.
+			 */
+			etr_tolec = get_clock();
+		eacr.p1 = etr_port1_online;
+		if (!eacr.p1)
+			eacr.e1 = 0;
+		etr_port1_uptodate = 0;
+	}
+	clear_bit(ETR_EVENT_UPDATE, &etr_events);
+	return eacr;
+}
+
+/*
+ * Set up a timer that expires after the etr_tolec + 1.6 seconds if
+ * one of the ports needs an update.
+ */
+static void etr_set_tolec_timeout(unsigned long long now)
+{
+	unsigned long micros;
+
+	if ((!etr_eacr.p0 || etr_port0_uptodate) &&
+	    (!etr_eacr.p1 || etr_port1_uptodate))
+		return;
+	micros = (now > etr_tolec) ? ((now - etr_tolec) >> 12) : 0;
+	micros = (micros > 1600000) ? 0 : 1600000 - micros;
+	mod_timer(&etr_timer, jiffies + (micros * HZ) / 1000000 + 1);
+}
+
+/*
+ * Set up a time that expires after 1/2 second.
+ */
+static void etr_set_sync_timeout(void)
+{
+	mod_timer(&etr_timer, jiffies + HZ/2);
+}
+
+/*
+ * Update the aib information for one or both ports.
+ */
+static struct etr_eacr etr_handle_update(struct etr_aib *aib,
+					 struct etr_eacr eacr)
+{
+	/* With both ports disabled the aib information is useless. */
+	if (!eacr.e0 && !eacr.e1)
+		return eacr;
+
+	/* Update port0 or port1 with aib stored in etr_tasklet_fn. */
+	if (aib->esw.q == 0) {
+		/* Information for port 0 stored. */
+		if (eacr.p0 && !etr_port0_uptodate) {
+			etr_port0 = *aib;
+			if (etr_port0_online)
+				etr_port0_uptodate = 1;
+		}
+	} else {
+		/* Information for port 1 stored. */
+		if (eacr.p1 && !etr_port1_uptodate) {
+			etr_port1 = *aib;
+			if (etr_port0_online)
+				etr_port1_uptodate = 1;
+		}
+	}
+
+	/*
+	 * Do not try to get the alternate port aib if the clock
+	 * is not in sync yet.
+	 */
+	if (!eacr.es)
+		return eacr;
+
+	/*
+	 * If steai is available we can get the information about
+	 * the other port immediately. If only stetr is available the
+	 * data-port bit toggle has to be used.
+	 */
+	if (test_bit(ETR_FLAG_STEAI, &etr_flags)) {
+		if (eacr.p0 && !etr_port0_uptodate) {
+			etr_steai_cv(&etr_port0, ETR_STEAI_PORT_0);
+			etr_port0_uptodate = 1;
+		}
+		if (eacr.p1 && !etr_port1_uptodate) {
+			etr_steai_cv(&etr_port1, ETR_STEAI_PORT_1);
+			etr_port1_uptodate = 1;
+		}
+	} else {
+		/*
+		 * One port was updated above, if the other
+		 * port is not uptodate toggle dp bit.
+		 */
+		if ((eacr.p0 && !etr_port0_uptodate) ||
+		    (eacr.p1 && !etr_port1_uptodate))
+			eacr.dp ^= 1;
+		else
+			eacr.dp = 0;
+	}
+	return eacr;
+}
+
+/*
+ * Write new etr control register if it differs from the current one.
+ * Return 1 if etr_tolec has been updated as well.
+ */
+static void etr_update_eacr(struct etr_eacr eacr)
+{
+	int dp_changed;
+
+	if (memcmp(&etr_eacr, &eacr, sizeof(eacr)) == 0)
+		/* No change, return. */
+		return;
+	/*
+	 * The disable of an active port of the change of the data port
+	 * bit can/will cause a change in the data port.
+	 */
+	dp_changed = etr_eacr.e0 > eacr.e0 || etr_eacr.e1 > eacr.e1 ||
+		(etr_eacr.dp ^ eacr.dp) != 0;
+	etr_eacr = eacr;
+	etr_setr(&etr_eacr);
+	if (dp_changed)
+		etr_tolec = get_clock();
+}
+
+/*
+ * ETR tasklet. In this function you'll find the main logic. In
+ * particular this is the only function that calls etr_update_eacr(),
+ * it "controls" the etr control register.
+ */
+static void etr_tasklet_fn(unsigned long dummy)
+{
+	unsigned long long now;
+	struct etr_eacr eacr;
+	struct etr_aib aib;
+	int sync_port;
+
+	/* Create working copy of etr_eacr. */
+	eacr = etr_eacr;
+
+	/* Check for the different events and their immediate effects. */
+	eacr = etr_handle_events(eacr);
+
+	/* Check if ETR is supposed to be active. */
+	eacr.ea = eacr.p0 || eacr.p1;
+	if (!eacr.ea) {
+		/* Both ports offline. Reset everything. */
+		eacr.dp = eacr.es = eacr.sl = 0;
+		on_each_cpu(etr_disable_sync_clock, NULL, 0, 1);
+		del_timer_sync(&etr_timer);
+		etr_update_eacr(eacr);
+		set_bit(ETR_FLAG_EACCES, &etr_flags);
+		return;
+	}
+
+	/* Store aib to get the current ETR status word. */
+	BUG_ON(etr_stetr(&aib) != 0);
+	etr_port0.esw = etr_port1.esw = aib.esw;	/* Copy status word. */
+	now = get_clock();
+
+	/*
+	 * Update the port information if the last stepping port change
+	 * or data port change is older than 1.6 seconds.
+	 */
+	if (now >= etr_tolec + (1600000 << 12))
+		eacr = etr_handle_update(&aib, eacr);
+
+	/*
+	 * Select ports to enable. The prefered synchronization mode is PPS.
+	 * If a port can be enabled depends on a number of things:
+	 * 1) The port needs to be online and uptodate. A port is not
+	 *    disabled just because it is not uptodate, but it is only
+	 *    enabled if it is uptodate.
+	 * 2) The port needs to have the same mode (pps / etr).
+	 * 3) The port needs to be usable -> etr_port_valid() == 1
+	 * 4) To enable the second port the clock needs to be in sync.
+	 * 5) If both ports are useable and are ETR ports, the network id
+	 *    has to be the same.
+	 * The eacr.sl bit is used to indicate etr mode vs. pps mode.
+	 */
+	if (eacr.p0 && aib.esw.psc0 == etr_lpsc_pps_mode) {
+		eacr.sl = 0;
+		eacr.e0 = 1;
+		if (!etr_mode_is_pps(etr_eacr))
+			eacr.es = 0;
+		if (!eacr.es || !eacr.p1 || aib.esw.psc1 != etr_lpsc_pps_mode)
+			eacr.e1 = 0;
+		// FIXME: uptodate checks ?
+		else if (etr_port0_uptodate && etr_port1_uptodate)
+			eacr.e1 = 1;
+		sync_port = (etr_port0_uptodate &&
+			     etr_port_valid(&etr_port0, 0)) ? 0 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else if (eacr.p1 && aib.esw.psc1 == etr_lpsc_pps_mode) {
+		eacr.sl = 0;
+		eacr.e0 = 0;
+		eacr.e1 = 1;
+		if (!etr_mode_is_pps(etr_eacr))
+			eacr.es = 0;
+		sync_port = (etr_port1_uptodate &&
+			     etr_port_valid(&etr_port1, 1)) ? 1 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else if (eacr.p0 && aib.esw.psc0 == etr_lpsc_operational_step) {
+		eacr.sl = 1;
+		eacr.e0 = 1;
+		if (!etr_mode_is_etr(etr_eacr))
+			eacr.es = 0;
+		if (!eacr.es || !eacr.p1 ||
+		    aib.esw.psc1 != etr_lpsc_operational_alt)
+			eacr.e1 = 0;
+		else if (etr_port0_uptodate && etr_port1_uptodate &&
+			 etr_compare_network(&etr_port0, &etr_port1))
+			eacr.e1 = 1;
+		sync_port = (etr_port0_uptodate &&
+			     etr_port_valid(&etr_port0, 0)) ? 0 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else if (eacr.p1 && aib.esw.psc1 == etr_lpsc_operational_step) {
+		eacr.sl = 1;
+		eacr.e0 = 0;
+		eacr.e1 = 1;
+		if (!etr_mode_is_etr(etr_eacr))
+			eacr.es = 0;
+		sync_port = (etr_port1_uptodate &&
+			     etr_port_valid(&etr_port1, 1)) ? 1 : -1;
+		clear_bit(ETR_FLAG_EACCES, &etr_flags);
+	} else {
+		/* Both ports not usable. */
+		eacr.es = eacr.sl = 0;
+		sync_port = -1;
+		set_bit(ETR_FLAG_EACCES, &etr_flags);
+	}
+
+	/*
+	 * If the clock is in sync just update the eacr and return.
+	 * If there is no valid sync port wait for a port update.
+	 */
+	if (eacr.es || sync_port < 0) {
+		etr_update_eacr(eacr);
+		etr_set_tolec_timeout(now);
+		return;
+	}
+
+	/*
+	 * Prepare control register for clock syncing
+	 * (reset data port bit, set sync check control.
+	 */
+	eacr.dp = 0;
+	eacr.es = 1;
+
+	/*
+	 * Update eacr and try to synchronize the clock. If the update
+	 * of eacr caused a stepping port switch (or if we have to
+	 * assume that a stepping port switch has occured) or the
+	 * clock syncing failed, reset the sync check control bit
+	 * and set up a timer to try again after 0.5 seconds
+	 */
+	etr_update_eacr(eacr);
+	if (now < etr_tolec + (1600000 << 12) ||
+	    etr_sync_clock(&aib, sync_port) != 0) {
+		/* Sync failed. Try again in 1/2 second. */
+		eacr.es = 0;
+		etr_update_eacr(eacr);
+		etr_set_sync_timeout();
+	} else
+		etr_set_tolec_timeout(now);
+}
+
+/*
+ * Sysfs interface functions
+ */
+static struct sysdev_class etr_sysclass = {
+	set_kset_name("etr")
+};
+
+static struct sys_device etr_port0_dev = {
+	.id	= 0,
+	.cls	= &etr_sysclass,
+};
+
+static struct sys_device etr_port1_dev = {
+	.id	= 1,
+	.cls	= &etr_sysclass,
+};
+
+/*
+ * ETR class attributes
+ */
+static ssize_t etr_stepping_port_show(struct sysdev_class *class, char *buf)
+{
+	return sprintf(buf, "%i\n", etr_port0.esw.p);
+}
+
+static SYSDEV_CLASS_ATTR(stepping_port, 0400, etr_stepping_port_show, NULL);
+
+static ssize_t etr_stepping_mode_show(struct sysdev_class *class, char *buf)
+{
+	char *mode_str;
+
+	if (etr_mode_is_pps(etr_eacr))
+		mode_str = "pps";
+	else if (etr_mode_is_etr(etr_eacr))
+		mode_str = "etr";
+	else
+		mode_str = "local";
+	return sprintf(buf, "%s\n", mode_str);
+}
+
+static SYSDEV_CLASS_ATTR(stepping_mode, 0400, etr_stepping_mode_show, NULL);
+
+/*
+ * ETR port attributes
+ */
+static inline struct etr_aib *etr_aib_from_dev(struct sys_device *dev)
+{
+	if (dev == &etr_port0_dev)
+		return etr_port0_online ? &etr_port0 : NULL;
+	else
+		return etr_port1_online ? &etr_port1 : NULL;
+}
+
+static ssize_t etr_online_show(struct sys_device *dev, char *buf)
+{
+	unsigned int online;
+
+	online = (dev == &etr_port0_dev) ? etr_port0_online : etr_port1_online;
+	return sprintf(buf, "%i\n", online);
+}
+
+static ssize_t etr_online_store(struct sys_device *dev,
+			      const char *buf, size_t count)
+{
+	unsigned int value;
+
+	value = simple_strtoul(buf, NULL, 0);
+	if (value != 0 && value != 1)
+		return -EINVAL;
+	if (test_bit(ETR_FLAG_ENOSYS, &etr_flags))
+		return -ENOSYS;
+	if (dev == &etr_port0_dev) {
+		if (etr_port0_online == value)
+			return count;	/* Nothing to do. */
+		etr_port0_online = value;
+		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	} else {
+		if (etr_port1_online == value)
+			return count;	/* Nothing to do. */
+		etr_port1_online = value;
+		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
+		tasklet_hi_schedule(&etr_tasklet);
+	}
+	return count;
+}
+
+static SYSDEV_ATTR(online, 0600, etr_online_show, etr_online_store);
+
+static ssize_t etr_stepping_control_show(struct sys_device *dev, char *buf)
+{
+	return sprintf(buf, "%i\n", (dev == &etr_port0_dev) ?
+		       etr_eacr.e0 : etr_eacr.e1);
+}
+
+static SYSDEV_ATTR(stepping_control, 0400, etr_stepping_control_show, NULL);
+
+static ssize_t etr_mode_code_show(struct sys_device *dev, char *buf)
+{
+	if (!etr_port0_online && !etr_port1_online)
+		/* Status word is not uptodate if both ports are offline. */
+		return -ENODATA;
+	return sprintf(buf, "%i\n", (dev == &etr_port0_dev) ?
+		       etr_port0.esw.psc0 : etr_port0.esw.psc1);
+}
+
+static SYSDEV_ATTR(state_code, 0400, etr_mode_code_show, NULL);
+
+static ssize_t etr_untuned_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.u);
+}
+
+static SYSDEV_ATTR(untuned, 0400, etr_untuned_show, NULL);
+
+static ssize_t etr_network_id_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.net_id);
+}
+
+static SYSDEV_ATTR(network, 0400, etr_network_id_show, NULL);
+
+static ssize_t etr_id_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.etr_id);
+}
+
+static SYSDEV_ATTR(id, 0400, etr_id_show, NULL);
+
+static ssize_t etr_port_number_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v1)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf1.etr_pn);
+}
+
+static SYSDEV_ATTR(port, 0400, etr_port_number_show, NULL);
+
+static ssize_t etr_coupled_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v3)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf3.c);
+}
+
+static SYSDEV_ATTR(coupled, 0400, etr_coupled_show, NULL);
+
+static ssize_t etr_local_time_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v3)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf3.blto);
+}
+
+static SYSDEV_ATTR(local_time, 0400, etr_local_time_show, NULL);
+
+static ssize_t etr_utc_offset_show(struct sys_device *dev, char *buf)
+{
+	struct etr_aib *aib = etr_aib_from_dev(dev);
+
+	if (!aib || !aib->slsw.v3)
+		return -ENODATA;
+	return sprintf(buf, "%i\n", aib->edf3.buo);
+}
+
+static SYSDEV_ATTR(utc_offset, 0400, etr_utc_offset_show, NULL);
+
+static struct sysdev_attribute *etr_port_attributes[] = {
+	&attr_online,
+	&attr_stepping_control,
+	&attr_state_code,
+	&attr_untuned,
+	&attr_network,
+	&attr_id,
+	&attr_port,
+	&attr_coupled,
+	&attr_local_time,
+	&attr_utc_offset,
+	NULL
+};
+
+static int __init etr_register_port(struct sys_device *dev)
+{
+	struct sysdev_attribute **attr;
+	int rc;
+
+	rc = sysdev_register(dev);
+	if (rc)
+		goto out;
+	for (attr = etr_port_attributes; *attr; attr++) {
+		rc = sysdev_create_file(dev, *attr);
+		if (rc)
+			goto out_unreg;
+	}
+	return 0;
+out_unreg:
+	for (; attr >= etr_port_attributes; attr--)
+		sysdev_remove_file(dev, *attr);
+	sysdev_unregister(dev);
+out:
+	return rc;
+}
+
+static void __init etr_unregister_port(struct sys_device *dev)
+{
+	struct sysdev_attribute **attr;
+
+	for (attr = etr_port_attributes; *attr; attr++)
+		sysdev_remove_file(dev, *attr);
+	sysdev_unregister(dev);
+}
+
+static int __init etr_init_sysfs(void)
+{
+	int rc;
+
+	rc = sysdev_class_register(&etr_sysclass);
+	if (rc)
+		goto out;
+	rc = sysdev_class_create_file(&etr_sysclass, &attr_stepping_port);
+	if (rc)
+		goto out_unreg_class;
+	rc = sysdev_class_create_file(&etr_sysclass, &attr_stepping_mode);
+	if (rc)
+		goto out_remove_stepping_port;
+	rc = etr_register_port(&etr_port0_dev);
+	if (rc)
+		goto out_remove_stepping_mode;
+	rc = etr_register_port(&etr_port1_dev);
+	if (rc)
+		goto out_remove_port0;
+	return 0;
+
+out_remove_port0:
+	etr_unregister_port(&etr_port0_dev);
+out_remove_stepping_mode:
+	sysdev_class_remove_file(&etr_sysclass, &attr_stepping_mode);
+out_remove_stepping_port:
+	sysdev_class_remove_file(&etr_sysclass, &attr_stepping_port);
+out_unreg_class:
+	sysdev_class_unregister(&etr_sysclass);
+out:
+	return rc;
 }
 
+device_initcall(etr_init_sysfs);

arch/s390/kernel/traps.c

@@ -283,7 +283,7 @@ char *task_show_regs(struct task_struct *task, char *buffer)
 	return buffer;
 }
 
-DEFINE_SPINLOCK(die_lock);
+static DEFINE_SPINLOCK(die_lock);
 
 void die(const char * str, struct pt_regs * regs, long err)
 {
@@ -364,8 +364,7 @@ void __kprobes do_single_step(struct pt_regs *regs)
 		force_sig(SIGTRAP, current);
 }
 
-asmlinkage void
-default_trap_handler(struct pt_regs * regs, long interruption_code)
+static void default_trap_handler(struct pt_regs * regs, long interruption_code)
 {
         if (regs->psw.mask & PSW_MASK_PSTATE) {
 		local_irq_enable();
@@ -376,7 +375,7 @@ default_trap_handler(struct pt_regs * regs, long interruption_code)
 }
 
 #define DO_ERROR_INFO(signr, str, name, sicode, siaddr) \
-asmlinkage void name(struct pt_regs * regs, long interruption_code) \
+static void name(struct pt_regs * regs, long interruption_code) \
 { \
         siginfo_t info; \
         info.si_signo = signr; \
@@ -442,7 +441,7 @@ do_fp_trap(struct pt_regs *regs, void __user *location,
 		"floating point exception", regs, &si);
 }
 
-asmlinkage void illegal_op(struct pt_regs * regs, long interruption_code)
+static void illegal_op(struct pt_regs * regs, long interruption_code)
 {
 	siginfo_t info;
         __u8 opcode[6];
@@ -491,8 +490,15 @@ asmlinkage void illegal_op(struct pt_regs * regs, long interruption_code)
 #endif
 		} else
 			signal = SIGILL;
-	} else
-		signal = SIGILL;
+	} else {
+		/*
+		 * If we get an illegal op in kernel mode, send it through the
+		 * kprobes notifier. If kprobes doesn't pick it up, SIGILL
+		 */
+		if (notify_die(DIE_BPT, "bpt", regs, interruption_code,
+			       3, SIGTRAP) != NOTIFY_STOP)
+			signal = SIGILL;
+	}
 
 #ifdef CONFIG_MATHEMU
         if (signal == SIGFPE)
@@ -585,7 +591,7 @@ DO_ERROR_INFO(SIGILL, "specification exception", specification_exception,
 	      ILL_ILLOPN, get_check_address(regs));
 #endif
 
-asmlinkage void data_exception(struct pt_regs * regs, long interruption_code)
+static void data_exception(struct pt_regs * regs, long interruption_code)
 {
 	__u16 __user *location;
 	int signal = 0;
@@ -675,7 +681,7 @@ asmlinkage void data_exception(struct pt_regs * regs, long interruption_code)
 	}
 }
 
-asmlinkage void space_switch_exception(struct pt_regs * regs, long int_code)
+static void space_switch_exception(struct pt_regs * regs, long int_code)
 {
         siginfo_t info;
 

arch/s390/kernel/vmlinux.lds.S

@@ -31,18 +31,19 @@ SECTIONS
 
   _etext = .;			/* End of text section */
 
-  . = ALIGN(16);		/* Exception table */
-  __start___ex_table = .;
-  __ex_table : { *(__ex_table) }
-  __stop___ex_table = .;
-
   RODATA
 
 #ifdef CONFIG_SHARED_KERNEL
   . = ALIGN(1048576);		/* VM shared segments are 1MB aligned */
+#endif
 
+  . = ALIGN(4096);
   _eshared = .;			/* End of shareable data */
-#endif
+
+  . = ALIGN(16);		/* Exception table */
+  __start___ex_table = .;
+  __ex_table : { *(__ex_table) }
+  __stop___ex_table = .;
 
   .data : {			/* Data */
 	*(.data)

arch/s390/kernel/vtime.c

@@ -25,7 +25,7 @@
 #include <asm/irq_regs.h>
 
 static ext_int_info_t ext_int_info_timer;
-DEFINE_PER_CPU(struct vtimer_queue, virt_cpu_timer);
+static DEFINE_PER_CPU(struct vtimer_queue, virt_cpu_timer);
 
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 /*
@@ -524,16 +524,15 @@ EXPORT_SYMBOL(del_virt_timer);
 void init_cpu_vtimer(void)
 {
 	struct vtimer_queue *vt_list;
-	unsigned long cr0;
 
 	/* kick the virtual timer */
 	S390_lowcore.exit_timer = VTIMER_MAX_SLICE;
 	S390_lowcore.last_update_timer = VTIMER_MAX_SLICE;
 	asm volatile ("SPT %0" : : "m" (S390_lowcore.last_update_timer));
 	asm volatile ("STCK %0" : "=m" (S390_lowcore.last_update_clock));
-	__ctl_store(cr0, 0, 0);
-	cr0 |= 0x400;
-	__ctl_load(cr0, 0, 0);
+
+	/* enable cpu timer interrupts */
+	__ctl_set_bit(0,10);
 
 	vt_list = &per_cpu(virt_cpu_timer, smp_processor_id());
 	INIT_LIST_HEAD(&vt_list->list);
@@ -572,6 +571,7 @@ void __init vtime_init(void)
 	if (register_idle_notifier(&vtimer_idle_nb))
 		panic("Couldn't register idle notifier");
 
+	/* Enable cpu timer interrupts on the boot cpu. */
 	init_cpu_vtimer();
 }
 

arch/s390/lib/Makefile

@@ -4,7 +4,7 @@
 
 EXTRA_AFLAGS := -traditional
 
-lib-y += delay.o string.o uaccess_std.o uaccess_pt.o
+lib-y += delay.o string.o uaccess_std.o uaccess_pt.o qrnnd.o
 lib-$(CONFIG_32BIT) += div64.o
 lib-$(CONFIG_64BIT) += uaccess_mvcos.o
 lib-$(CONFIG_SMP) += spinlock.o

arch/s390/lib/delay.c

@@ -1,5 +1,5 @@
 /*
- *  arch/s390/kernel/delay.c
+ *  arch/s390/lib/delay.c
  *    Precise Delay Loops for S390
  *
  *  S390 version
@@ -13,10 +13,8 @@
 
 #include <linux/sched.h>
 #include <linux/delay.h>
-
-#ifdef CONFIG_SMP
-#include <asm/smp.h>
-#endif
+#include <linux/timex.h>
+#include <linux/irqflags.h>
 
 void __delay(unsigned long loops)
 {
@@ -31,17 +29,39 @@ void __delay(unsigned long loops)
 }
 
 /*
- * Waits for 'usecs' microseconds using the tod clock, giving up the time slice
- * of the virtual PU inbetween to avoid congestion.
+ * Waits for 'usecs' microseconds using the TOD clock comparator.
  */
 void __udelay(unsigned long usecs)
 {
-	uint64_t start_cc;
+	u64 end, time, jiffy_timer = 0;
+	unsigned long flags, cr0, mask, dummy;
+
+	local_irq_save(flags);
+	if (raw_irqs_disabled_flags(flags)) {
+		jiffy_timer = S390_lowcore.jiffy_timer;
+		S390_lowcore.jiffy_timer = -1ULL - (4096 << 12);
+		__ctl_store(cr0, 0, 0);
+		dummy = (cr0 & 0xffff00e0) | 0x00000800;
+		__ctl_load(dummy , 0, 0);
+		mask = psw_kernel_bits | PSW_MASK_WAIT | PSW_MASK_EXT;
+	} else
+		mask = psw_kernel_bits | PSW_MASK_WAIT |
+			PSW_MASK_EXT | PSW_MASK_IO;
+
+	end = get_clock() + ((u64) usecs << 12);
+	do {
+		time = end < S390_lowcore.jiffy_timer ?
+			end : S390_lowcore.jiffy_timer;
+		set_clock_comparator(time);
+		trace_hardirqs_on();
+		__load_psw_mask(mask);
+		local_irq_disable();
+	} while (get_clock() < end);
 
-        if (usecs == 0)
-                return;
-	start_cc = get_clock();
-        do {
-		cpu_relax();
-	} while (((get_clock() - start_cc)/4096) < usecs);
+	if (raw_irqs_disabled_flags(flags)) {
+		__ctl_load(cr0, 0, 0);
+		S390_lowcore.jiffy_timer = jiffy_timer;
+	}
+	set_clock_comparator(S390_lowcore.jiffy_timer);
+	local_irq_restore(flags);
 }

arch/s390/lib/qrnnd.S

@@ -0,0 +1,77 @@
+# S/390 __udiv_qrnnd
+
+# r2 : &__r
+# r3 : upper half of 64 bit word n
+# r4 : lower half of 64 bit word n
+# r5 : divisor d
+# the reminder r of the division is to be stored to &__r and
+# the quotient q is to be returned
+
+	.text
+	.globl __udiv_qrnnd
+__udiv_qrnnd:
+	st    %r2,24(%r15)	  # store pointer to reminder for later
+	lr    %r0,%r3		  # reload n
+	lr    %r1,%r4
+	ltr   %r2,%r5		  # reload and test divisor
+	jp    5f
+	# divisor >= 0x80000000
+	srdl  %r0,2		  # n/4
+	srl   %r2,1		  # d/2
+	slr   %r1,%r2		  # special case if last bit of d is set
+	brc   3,0f		  #  (n/4) div (n/2) can overflow by 1
+	ahi   %r0,-1		  #  trick: subtract n/2, then divide
+0:	dr    %r0,%r2		  # signed division
+	ahi   %r1,1		  #  trick part 2: add 1 to the quotient
+	# now (n >> 2) = (d >> 1) * %r1 + %r0
+	lhi   %r3,1
+	nr    %r3,%r1		  # test last bit of q
+	jz    1f
+	alr   %r0,%r2		  # add (d>>1) to r
+1:	srl   %r1,1		  # q >>= 1
+	# now (n >> 2) = (d&-2) * %r1 + %r0
+	lhi   %r3,1
+	nr    %r3,%r5		  # test last bit of d
+	jz    2f
+	slr   %r0,%r1		  # r -= q
+	brc   3,2f		  # borrow ?
+	alr   %r0,%r5		  # r += d
+	ahi   %r1,-1
+2:	# now (n >> 2) = d * %r1 + %r0
+	alr   %r1,%r1		  # q <<= 1
+	alr   %r0,%r0		  # r <<= 1
+	brc   12,3f		  # overflow on r ?
+	slr   %r0,%r5		  # r -= d
+	ahi   %r1,1		  # q += 1
+3:	lhi   %r3,2
+	nr    %r3,%r4		  # test next to last bit of n
+	jz    4f
+	ahi   %r0,1		  # r += 1
+4:	clr   %r0,%r5		  # r >= d ?
+	jl    6f
+	slr   %r0,%r5		  # r -= d
+	ahi   %r1,1		  # q += 1
+	# now (n >> 1) = d * %r1 + %r0
+	j     6f
+5:	# divisor < 0x80000000
+	srdl  %r0,1
+	dr    %r0,%r2		  # signed division
+	# now (n >> 1) = d * %r1 + %r0
+6:	alr   %r1,%r1		  # q <<= 1
+	alr   %r0,%r0		  # r <<= 1
+	brc   12,7f		  # overflow on r ?
+	slr   %r0,%r5		  # r -= d
+	ahi   %r1,1		  # q += 1
+7:	lhi   %r3,1
+	nr    %r3,%r4		  # isolate last bit of n
+	alr   %r0,%r3		  # r += (n & 1)
+	clr   %r0,%r5		  # r >= d ?
+	jl    8f
+	slr   %r0,%r5		  # r -= d
+	ahi   %r1,1		  # q += 1
+8:	# now n = d * %r1 + %r0
+	l     %r2,24(%r15)
+	st    %r0,0(%r2)
+	lr    %r2,%r1
+	br    %r14
+	.end	__udiv_qrnnd

arch/s390/lib/uaccess.h

@@ -0,0 +1,23 @@
+/*
+ *  arch/s390/uaccess.h
+ *
+ *    Copyright IBM Corp. 2007
+ *
+ */
+
+#ifndef __ARCH_S390_LIB_UACCESS_H
+#define __ARCH_S390_LIB_UACCESS_H
+
+extern size_t copy_from_user_std(size_t, const void __user *, void *);
+extern size_t copy_to_user_std(size_t, void __user *, const void *);
+extern size_t strnlen_user_std(size_t, const char __user *);
+extern size_t strncpy_from_user_std(size_t, const char __user *, char *);
+extern int futex_atomic_cmpxchg_std(int __user *, int, int);
+extern int futex_atomic_op_std(int, int __user *, int, int *);
+
+extern size_t copy_from_user_pt(size_t, const void __user *, void *);
+extern size_t copy_to_user_pt(size_t, void __user *, const void *);
+extern int futex_atomic_op_pt(int, int __user *, int, int *);
+extern int futex_atomic_cmpxchg_pt(int __user *, int, int);
+
+#endif /* __ARCH_S390_LIB_UACCESS_H */

arch/s390/lib/uaccess_mvcos.c

@@ -12,6 +12,7 @@
 #include <linux/mm.h>
 #include <asm/uaccess.h>
 #include <asm/futex.h>
+#include "uaccess.h"
 
 #ifndef __s390x__
 #define AHI	"ahi"
@@ -27,10 +28,7 @@
 #define SLR	"slgr"
 #endif
 
-extern size_t copy_from_user_std(size_t, const void __user *, void *);
-extern size_t copy_to_user_std(size_t, void __user *, const void *);
-
-size_t copy_from_user_mvcos(size_t size, const void __user *ptr, void *x)
+static size_t copy_from_user_mvcos(size_t size, const void __user *ptr, void *x)
 {
 	register unsigned long reg0 asm("0") = 0x81UL;
 	unsigned long tmp1, tmp2;
@@ -69,14 +67,14 @@ size_t copy_from_user_mvcos(size_t size, const void __user *ptr, void *x)
 	return size;
 }
 
-size_t copy_from_user_mvcos_check(size_t size, const void __user *ptr, void *x)
+static size_t copy_from_user_mvcos_check(size_t size, const void __user *ptr, void *x)
 {
 	if (size <= 256)
 		return copy_from_user_std(size, ptr, x);
 	return copy_from_user_mvcos(size, ptr, x);
 }
 
-size_t copy_to_user_mvcos(size_t size, void __user *ptr, const void *x)
+static size_t copy_to_user_mvcos(size_t size, void __user *ptr, const void *x)
 {
 	register unsigned long reg0 asm("0") = 0x810000UL;
 	unsigned long tmp1, tmp2;
@@ -105,14 +103,16 @@ size_t copy_to_user_mvcos(size_t size, void __user *ptr, const void *x)
 	return size;
 }
 
-size_t copy_to_user_mvcos_check(size_t size, void __user *ptr, const void *x)
+static size_t copy_to_user_mvcos_check(size_t size, void __user *ptr,
+				       const void *x)
 {
 	if (size <= 256)
 		return copy_to_user_std(size, ptr, x);
 	return copy_to_user_mvcos(size, ptr, x);
 }
 
-size_t copy_in_user_mvcos(size_t size, void __user *to, const void __user *from)
+static size_t copy_in_user_mvcos(size_t size, void __user *to,
+				 const void __user *from)
 {
 	register unsigned long reg0 asm("0") = 0x810081UL;
 	unsigned long tmp1, tmp2;
@@ -134,7 +134,7 @@ size_t copy_in_user_mvcos(size_t size, void __user *to, const void __user *from)
 	return size;
 }
 
-size_t clear_user_mvcos(size_t size, void __user *to)
+static size_t clear_user_mvcos(size_t size, void __user *to)
 {
 	register unsigned long reg0 asm("0") = 0x810000UL;
 	unsigned long tmp1, tmp2;
@@ -162,10 +162,43 @@ size_t clear_user_mvcos(size_t size, void __user *to)
 	return size;
 }
 
-extern size_t strnlen_user_std(size_t, const char __user *);
-extern size_t strncpy_from_user_std(size_t, const char __user *, char *);
-extern int futex_atomic_op(int, int __user *, int, int *);
-extern int futex_atomic_cmpxchg(int __user *, int, int);
+static size_t strnlen_user_mvcos(size_t count, const char __user *src)
+{
+	char buf[256];
+	int rc;
+	size_t done, len, len_str;
+
+	done = 0;
+	do {
+		len = min(count - done, (size_t) 256);
+		rc = uaccess.copy_from_user(len, src + done, buf);
+		if (unlikely(rc == len))
+			return 0;
+		len -= rc;
+		len_str = strnlen(buf, len);
+		done += len_str;
+	} while ((len_str == len) && (done < count));
+	return done + 1;
+}
+
+static size_t strncpy_from_user_mvcos(size_t count, const char __user *src,
+				      char *dst)
+{
+	int rc;
+	size_t done, len, len_str;
+
+	done = 0;
+	do {
+		len = min(count - done, (size_t) 4096);
+		rc = uaccess.copy_from_user(len, src + done, dst);
+		if (unlikely(rc == len))
+			return -EFAULT;
+		len -= rc;
+		len_str = strnlen(dst, len);
+		done += len_str;
+	} while ((len_str == len) && (done < count));
+	return done;
+}
 
 struct uaccess_ops uaccess_mvcos = {
 	.copy_from_user = copy_from_user_mvcos_check,
@@ -176,6 +209,21 @@ struct uaccess_ops uaccess_mvcos = {
 	.clear_user = clear_user_mvcos,
 	.strnlen_user = strnlen_user_std,
 	.strncpy_from_user = strncpy_from_user_std,
-	.futex_atomic_op = futex_atomic_op,
-	.futex_atomic_cmpxchg = futex_atomic_cmpxchg,
+	.futex_atomic_op = futex_atomic_op_std,
+	.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,
+	.copy_to_user = copy_to_user_mvcos,
+	.copy_to_user_small = copy_to_user_mvcos,
+	.copy_in_user = copy_in_user_mvcos,
+	.clear_user = clear_user_mvcos,
+	.strnlen_user = strnlen_user_mvcos,
+	.strncpy_from_user = strncpy_from_user_mvcos,
+	.futex_atomic_op = futex_atomic_op_pt,
+	.futex_atomic_cmpxchg = futex_atomic_cmpxchg_pt,
 };
+#endif

arch/s390/lib/uaccess_pt.c

@@ -1,7 +1,8 @@
 /*
  *  arch/s390/lib/uaccess_pt.c
  *
- *  User access functions based on page table walks.
+ *  User access functions based on page table walks for enhanced
+ *  system layout without hardware support.
  *
  *    Copyright IBM Corp. 2006
  *    Author(s): Gerald Schaefer (gerald.schaefer@de.ibm.com)
@@ -12,9 +13,10 @@
 #include <linux/mm.h>
 #include <asm/uaccess.h>
 #include <asm/futex.h>
+#include "uaccess.h"
 
-static inline int __handle_fault(struct mm_struct *mm, unsigned long address,
-				 int write_access)
+static int __handle_fault(struct mm_struct *mm, unsigned long address,
+			  int write_access)
 {
 	struct vm_area_struct *vma;
 	int ret = -EFAULT;
@@ -79,8 +81,8 @@ out_sigbus:
 	return ret;
 }
 
-static inline size_t __user_copy_pt(unsigned long uaddr, void *kptr,
-				    size_t n, int write_user)
+static size_t __user_copy_pt(unsigned long uaddr, void *kptr,
+			     size_t n, int write_user)
 {
 	struct mm_struct *mm = current->mm;
 	unsigned long offset, pfn, done, size;
@@ -133,6 +135,49 @@ fault:
 	goto retry;
 }
 
+/*
+ * 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)
+{
+	struct mm_struct *mm = current->mm;
+	unsigned long pfn, ret;
+	pgd_t *pgd;
+	pmd_t *pmd;
+	pte_t *pte;
+	int rc;
+
+	ret = 0;
+retry:
+	pgd = pgd_offset(mm, uaddr);
+	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
+		goto fault;
+
+	pmd = pmd_offset(pgd, uaddr);
+	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
+		goto fault;
+
+	pte = pte_offset_map(pmd, uaddr);
+	if (!pte || !pte_present(*pte))
+		goto fault;
+
+	pfn = pte_pfn(*pte);
+	if (!pfn_valid(pfn))
+		goto out;
+
+	ret = (pfn << PAGE_SHIFT) + (uaddr & (PAGE_SIZE - 1));
+out:
+	return ret;
+fault:
+	spin_unlock(&mm->page_table_lock);
+	rc = __handle_fault(mm, uaddr, 0);
+	spin_lock(&mm->page_table_lock);
+	if (rc)
+		goto out;
+	goto retry;
+}
+
 size_t copy_from_user_pt(size_t n, const void __user *from, void *to)
 {
 	size_t rc;
@@ -155,3 +200,277 @@ size_t copy_to_user_pt(size_t n, void __user *to, const void *from)
 	}
 	return __user_copy_pt((unsigned long) to, (void *) from, n, 1);
 }
+
+static size_t clear_user_pt(size_t n, void __user *to)
+{
+	long done, size, ret;
+
+	if (segment_eq(get_fs(), KERNEL_DS)) {
+		memset((void __kernel __force *) to, 0, n);
+		return 0;
+	}
+	done = 0;
+	do {
+		if (n - done > PAGE_SIZE)
+			size = PAGE_SIZE;
+		else
+			size = n - done;
+		ret = __user_copy_pt((unsigned long) to + done,
+				      &empty_zero_page, size, 1);
+		done += size;
+		if (ret)
+			return ret + n - done;
+	} while (done < n);
+	return 0;
+}
+
+static size_t strnlen_user_pt(size_t count, const char __user *src)
+{
+	char *addr;
+	unsigned long uaddr = (unsigned long) src;
+	struct mm_struct *mm = current->mm;
+	unsigned long offset, pfn, done, len;
+	pgd_t *pgd;
+	pmd_t *pmd;
+	pte_t *pte;
+	size_t len_str;
+
+	if (segment_eq(get_fs(), KERNEL_DS))
+		return strnlen((const char __kernel __force *) src, count) + 1;
+	done = 0;
+retry:
+	spin_lock(&mm->page_table_lock);
+	do {
+		pgd = pgd_offset(mm, uaddr);
+		if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
+			goto fault;
+
+		pmd = pmd_offset(pgd, uaddr);
+		if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
+			goto fault;
+
+		pte = pte_offset_map(pmd, uaddr);
+		if (!pte || !pte_present(*pte))
+			goto fault;
+
+		pfn = pte_pfn(*pte);
+		if (!pfn_valid(pfn)) {
+			done = -1;
+			goto out;
+		}
+
+		offset = uaddr & (PAGE_SIZE-1);
+		addr = (char *)(pfn << PAGE_SHIFT) + offset;
+		len = min(count - done, PAGE_SIZE - offset);
+		len_str = strnlen(addr, len);
+		done += len_str;
+		uaddr += len_str;
+	} while ((len_str == len) && (done < count));
+out:
+	spin_unlock(&mm->page_table_lock);
+	return done + 1;
+fault:
+	spin_unlock(&mm->page_table_lock);
+	if (__handle_fault(mm, uaddr, 0)) {
+		return 0;
+	}
+	goto retry;
+}
+
+static size_t strncpy_from_user_pt(size_t count, const char __user *src,
+				   char *dst)
+{
+	size_t n = strnlen_user_pt(count, src);
+
+	if (!n)
+		return -EFAULT;
+	if (n > count)
+		n = count;
+	if (segment_eq(get_fs(), KERNEL_DS)) {
+		memcpy(dst, (const char __kernel __force *) src, n);
+		if (dst[n-1] == '\0')
+			return n-1;
+		else
+			return n;
+	}
+	if (__user_copy_pt((unsigned long) src, dst, n, 0))
+		return -EFAULT;
+	if (dst[n-1] == '\0')
+		return n-1;
+	else
+		return n;
+}
+
+static size_t copy_in_user_pt(size_t n, void __user *to,
+			      const void __user *from)
+{
+	struct mm_struct *mm = current->mm;
+	unsigned long offset_from, offset_to, offset_max, pfn_from, pfn_to,
+		      uaddr, done, size;
+	unsigned long uaddr_from = (unsigned long) from;
+	unsigned long uaddr_to = (unsigned long) to;
+	pgd_t *pgd_from, *pgd_to;
+	pmd_t *pmd_from, *pmd_to;
+	pte_t *pte_from, *pte_to;
+	int write_user;
+
+	done = 0;
+retry:
+	spin_lock(&mm->page_table_lock);
+	do {
+		pgd_from = pgd_offset(mm, uaddr_from);
+		if (pgd_none(*pgd_from) || unlikely(pgd_bad(*pgd_from))) {
+			uaddr = uaddr_from;
+			write_user = 0;
+			goto fault;
+		}
+		pgd_to = pgd_offset(mm, uaddr_to);
+		if (pgd_none(*pgd_to) || unlikely(pgd_bad(*pgd_to))) {
+			uaddr = uaddr_to;
+			write_user = 1;
+			goto fault;
+		}
+
+		pmd_from = pmd_offset(pgd_from, uaddr_from);
+		if (pmd_none(*pmd_from) || unlikely(pmd_bad(*pmd_from))) {
+			uaddr = uaddr_from;
+			write_user = 0;
+			goto fault;
+		}
+		pmd_to = pmd_offset(pgd_to, uaddr_to);
+		if (pmd_none(*pmd_to) || unlikely(pmd_bad(*pmd_to))) {
+			uaddr = uaddr_to;
+			write_user = 1;
+			goto fault;
+		}
+
+		pte_from = pte_offset_map(pmd_from, uaddr_from);
+		if (!pte_from || !pte_present(*pte_from)) {
+			uaddr = uaddr_from;
+			write_user = 0;
+			goto fault;
+		}
+		pte_to = pte_offset_map(pmd_to, uaddr_to);
+		if (!pte_to || !pte_present(*pte_to) || !pte_write(*pte_to)) {
+			uaddr = uaddr_to;
+			write_user = 1;
+			goto fault;
+		}
+
+		pfn_from = pte_pfn(*pte_from);
+		if (!pfn_valid(pfn_from))
+			goto out;
+		pfn_to = pte_pfn(*pte_to);
+		if (!pfn_valid(pfn_to))
+			goto out;
+
+		offset_from = uaddr_from & (PAGE_SIZE-1);
+		offset_to = uaddr_from & (PAGE_SIZE-1);
+		offset_max = max(offset_from, offset_to);
+		size = min(n - done, PAGE_SIZE - offset_max);
+
+		memcpy((void *)(pfn_to << PAGE_SHIFT) + offset_to,
+		       (void *)(pfn_from << PAGE_SHIFT) + offset_from, size);
+		done += size;
+		uaddr_from += size;
+		uaddr_to += size;
+	} while (done < n);
+out:
+	spin_unlock(&mm->page_table_lock);
+	return n - done;
+fault:
+	spin_unlock(&mm->page_table_lock);
+	if (__handle_fault(mm, uaddr, write_user))
+		return n - done;
+	goto retry;
+}
+
+#define __futex_atomic_op(insn, ret, oldval, newval, uaddr, oparg)	\
+	asm volatile("0: l   %1,0(%6)\n"				\
+		     "1: " insn						\
+		     "2: cs  %1,%2,0(%6)\n"				\
+		     "3: jl  1b\n"					\
+		     "   lhi %0,0\n"					\
+		     "4:\n"						\
+		     EX_TABLE(0b,4b) EX_TABLE(2b,4b) EX_TABLE(3b,4b)	\
+		     : "=d" (ret), "=&d" (oldval), "=&d" (newval),	\
+		       "=m" (*uaddr)					\
+		     : "0" (-EFAULT), "d" (oparg), "a" (uaddr),		\
+		       "m" (*uaddr) : "cc" );
+
+int futex_atomic_op_pt(int op, int __user *uaddr, int oparg, int *old)
+{
+	int oldval = 0, newval, ret;
+
+	spin_lock(&current->mm->page_table_lock);
+	uaddr = (int __user *) __dat_user_addr((unsigned long) uaddr);
+	if (!uaddr) {
+		spin_unlock(&current->mm->page_table_lock);
+		return -EFAULT;
+	}
+	get_page(virt_to_page(uaddr));
+	spin_unlock(&current->mm->page_table_lock);
+	switch (op) {
+	case FUTEX_OP_SET:
+		__futex_atomic_op("lr %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_ADD:
+		__futex_atomic_op("lr %2,%1\nar %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_OR:
+		__futex_atomic_op("lr %2,%1\nor %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_ANDN:
+		__futex_atomic_op("lr %2,%1\nnr %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_XOR:
+		__futex_atomic_op("lr %2,%1\nxr %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	default:
+		ret = -ENOSYS;
+	}
+	put_page(virt_to_page(uaddr));
+	*old = oldval;
+	return ret;
+}
+
+int futex_atomic_cmpxchg_pt(int __user *uaddr, int oldval, int newval)
+{
+	int ret;
+
+	spin_lock(&current->mm->page_table_lock);
+	uaddr = (int __user *) __dat_user_addr((unsigned long) uaddr);
+	if (!uaddr) {
+		spin_unlock(&current->mm->page_table_lock);
+		return -EFAULT;
+	}
+	get_page(virt_to_page(uaddr));
+	spin_unlock(&current->mm->page_table_lock);
+	asm volatile("   cs   %1,%4,0(%5)\n"
+		     "0: lr   %0,%1\n"
+		     "1:\n"
+		     EX_TABLE(0b,1b)
+		     : "=d" (ret), "+d" (oldval), "=m" (*uaddr)
+		     : "0" (-EFAULT), "d" (newval), "a" (uaddr), "m" (*uaddr)
+		     : "cc", "memory" );
+	put_page(virt_to_page(uaddr));
+	return ret;
+}
+
+struct uaccess_ops uaccess_pt = {
+	.copy_from_user		= copy_from_user_pt,
+	.copy_from_user_small	= copy_from_user_pt,
+	.copy_to_user		= copy_to_user_pt,
+	.copy_to_user_small	= copy_to_user_pt,
+	.copy_in_user		= copy_in_user_pt,
+	.clear_user		= clear_user_pt,
+	.strnlen_user		= strnlen_user_pt,
+	.strncpy_from_user	= strncpy_from_user_pt,
+	.futex_atomic_op	= futex_atomic_op_pt,
+	.futex_atomic_cmpxchg	= futex_atomic_cmpxchg_pt,
+};

arch/s390/lib/uaccess_std.c

@@ -13,6 +13,7 @@
 #include <linux/mm.h>
 #include <linux/uaccess.h>
 #include <asm/futex.h>
+#include "uaccess.h"
 
 #ifndef __s390x__
 #define AHI	"ahi"
@@ -28,9 +29,6 @@
 #define SLR	"slgr"
 #endif
 
-extern size_t copy_from_user_pt(size_t n, const void __user *from, void *to);
-extern size_t copy_to_user_pt(size_t n, void __user *to, const void *from);
-
 size_t copy_from_user_std(size_t size, const void __user *ptr, void *x)
 {
 	unsigned long tmp1, tmp2;
@@ -72,7 +70,8 @@ size_t copy_from_user_std(size_t size, const void __user *ptr, void *x)
 	return size;
 }
 
-size_t copy_from_user_std_check(size_t size, const void __user *ptr, void *x)
+static size_t copy_from_user_std_check(size_t size, const void __user *ptr,
+				       void *x)
 {
 	if (size <= 1024)
 		return copy_from_user_std(size, ptr, x);
@@ -110,14 +109,16 @@ size_t copy_to_user_std(size_t size, void __user *ptr, const void *x)
 	return size;
 }
 
-size_t copy_to_user_std_check(size_t size, void __user *ptr, const void *x)
+static size_t copy_to_user_std_check(size_t size, void __user *ptr,
+				     const void *x)
 {
 	if (size <= 1024)
 		return copy_to_user_std(size, ptr, x);
 	return copy_to_user_pt(size, ptr, x);
 }
 
-size_t copy_in_user_std(size_t size, void __user *to, const void __user *from)
+static size_t copy_in_user_std(size_t size, void __user *to,
+			       const void __user *from)
 {
 	unsigned long tmp1;
 
@@ -148,7 +149,7 @@ size_t copy_in_user_std(size_t size, void __user *to, const void __user *from)
 	return size;
 }
 
-size_t clear_user_std(size_t size, void __user *to)
+static size_t clear_user_std(size_t size, void __user *to)
 {
 	unsigned long tmp1, tmp2;
 
@@ -254,7 +255,7 @@ size_t strncpy_from_user_std(size_t size, const char __user *src, char *dst)
 		: "0" (-EFAULT), "d" (oparg), "a" (uaddr),		\
 		  "m" (*uaddr) : "cc");
 
-int futex_atomic_op(int op, int __user *uaddr, int oparg, int *old)
+int futex_atomic_op_std(int op, int __user *uaddr, int oparg, int *old)
 {
 	int oldval = 0, newval, ret;
 
@@ -286,7 +287,7 @@ int futex_atomic_op(int op, int __user *uaddr, int oparg, int *old)
 	return ret;
 }
 
-int futex_atomic_cmpxchg(int __user *uaddr, int oldval, int newval)
+int futex_atomic_cmpxchg_std(int __user *uaddr, int oldval, int newval)
 {
 	int ret;
 
@@ -311,6 +312,6 @@ struct uaccess_ops uaccess_std = {
 	.clear_user = clear_user_std,
 	.strnlen_user = strnlen_user_std,
 	.strncpy_from_user = strncpy_from_user_std,
-	.futex_atomic_op = futex_atomic_op,
-	.futex_atomic_cmpxchg = futex_atomic_cmpxchg,
+	.futex_atomic_op = futex_atomic_op_std,
+	.futex_atomic_cmpxchg = futex_atomic_cmpxchg_std,
 };

arch/s390/math-emu/Makefile

@@ -2,7 +2,7 @@
 # Makefile for the FPU instruction emulation.
 #
 
-obj-$(CONFIG_MATHEMU) := math.o qrnnd.o
+obj-$(CONFIG_MATHEMU) := math.o
 
 EXTRA_CFLAGS := -I$(src) -Iinclude/math-emu -w
 EXTRA_AFLAGS := -traditional

arch/s390/math-emu/math.c

@@ -15,7 +15,7 @@
 #include <asm/uaccess.h>
 #include <asm/lowcore.h>
 
-#include "sfp-util.h"
+#include <asm/sfp-util.h>
 #include <math-emu/soft-fp.h>
 #include <math-emu/single.h>
 #include <math-emu/double.h>

arch/s390/math-emu/qrnnd.S

@@ -1,77 +0,0 @@
-# S/390 __udiv_qrnnd
-
-# r2 : &__r
-# r3 : upper half of 64 bit word n
-# r4 : lower half of 64 bit word n
-# r5 : divisor d
-# the reminder r of the division is to be stored to &__r and
-# the quotient q is to be returned
-
-        .text
-        .globl __udiv_qrnnd
-__udiv_qrnnd:
-        st    %r2,24(%r15)        # store pointer to reminder for later
-        lr    %r0,%r3             # reload n
-	lr    %r1,%r4
-	ltr   %r2,%r5             # reload and test divisor
-        jp    5f
-        # divisor >= 0x80000000
-	srdl  %r0,2               # n/4
-        srl   %r2,1               # d/2
-        slr   %r1,%r2             # special case if last bit of d is set
-        brc   3,0f                #  (n/4) div (n/2) can overflow by 1
-        ahi   %r0,-1              #  trick: subtract n/2, then divide
-0:      dr    %r0,%r2             # signed division
-        ahi   %r1,1               #  trick part 2: add 1 to the quotient
-        # now (n >> 2) = (d >> 1) * %r1 + %r0
-	lhi   %r3,1
-        nr    %r3,%r1             # test last bit of q
-        jz    1f
-        alr   %r0,%r2             # add (d>>1) to r
-1:      srl   %r1,1               # q >>= 1
-        # now (n >> 2) = (d&-2) * %r1 + %r0
-        lhi   %r3,1
-        nr    %r3,%r5             # test last bit of d
-        jz    2f
-        slr   %r0,%r1             # r -= q
-	brc   3,2f                # borrow ?
-	alr   %r0,%r5             # r += d
-	ahi   %r1,-1
-2:      # now (n >> 2) = d * %r1 + %r0
-        alr   %r1,%r1             # q <<= 1
-        alr   %r0,%r0             # r <<= 1
-        brc   12,3f               # overflow on r ?
-        slr   %r0,%r5             # r -= d
-        ahi   %r1,1               # q += 1
-3:      lhi   %r3,2
-        nr    %r3,%r4             # test next to last bit of n
-        jz    4f
-        ahi   %r0,1               # r += 1
-4:      clr   %r0,%r5             # r >= d ?
-        jl    6f
-        slr   %r0,%r5             # r -= d
-        ahi   %r1,1               # q += 1
-        # now (n >> 1) = d * %r1 + %r0
-        j     6f
-5:      # divisor < 0x80000000
-	srdl  %r0,1
-        dr    %r0,%r2             # signed division
-        # now (n >> 1) = d * %r1 + %r0
-6:      alr   %r1,%r1             # q <<= 1
-        alr   %r0,%r0             # r <<= 1
-        brc   12,7f               # overflow on r ?
-        slr   %r0,%r5             # r -= d
-        ahi   %r1,1               # q += 1
-7:      lhi   %r3,1
-        nr    %r3,%r4             # isolate last bit of n
-        alr   %r0,%r3             # r += (n & 1)
-        clr   %r0,%r5             # r >= d ?
-        jl    8f
-        slr   %r0,%r5             # r -= d
-        ahi   %r1,1               # q += 1
-8:      # now n = d * %r1 + %r0
-	l     %r2,24(%r15)
-        st    %r0,0(%r2)
-        lr    %r2,%r1
-        br    %r14
-	.end	__udiv_qrnnd

arch/s390/mm/cmm.c

@@ -245,7 +245,7 @@ cmm_set_timeout(long nr, long seconds)
 	cmm_set_timer();
 }
 
-static inline int
+static int
 cmm_skip_blanks(char *cp, char **endp)
 {
 	char *str;
@@ -414,7 +414,7 @@ cmm_smsg_target(char *from, char *msg)
 }
 #endif
 
-struct ctl_table_header *cmm_sysctl_header;
+static struct ctl_table_header *cmm_sysctl_header;
 
 static int
 cmm_init (void)

arch/s390/mm/extmem.c

@@ -15,6 +15,7 @@
 #include <linux/module.h>
 #include <linux/bootmem.h>
 #include <linux/ctype.h>
+#include <linux/ioport.h>
 #include <asm/page.h>
 #include <asm/pgtable.h>
 #include <asm/ebcdic.h>
@@ -70,6 +71,7 @@ struct qin64 {
 struct dcss_segment {
 	struct list_head list;
 	char dcss_name[8];
+	char res_name[15];
 	unsigned long start_addr;
 	unsigned long end;
 	atomic_t ref_count;
@@ -77,6 +79,7 @@ struct dcss_segment {
 	unsigned int vm_segtype;
 	struct qrange range[6];
 	int segcnt;
+	struct resource *res;
 };
 
 static DEFINE_MUTEX(dcss_lock);
@@ -88,7 +91,7 @@ static char *segtype_string[] = { "SW", "EW", "SR", "ER", "SN", "EN", "SC",
  * Create the 8 bytes, ebcdic VM segment name from
  * an ascii name.
  */
-static void inline
+static void
 dcss_mkname(char *name, char *dcss_name)
 {
 	int i;
@@ -303,6 +306,29 @@ __segment_load (char *name, int do_nonshared, unsigned long *addr, unsigned long
 		goto out_free;
 	}
 
+	seg->res = kzalloc(sizeof(struct resource), GFP_KERNEL);
+	if (seg->res == NULL) {
+		rc = -ENOMEM;
+		goto out_shared;
+	}
+	seg->res->flags = IORESOURCE_BUSY | IORESOURCE_MEM;
+	seg->res->start = seg->start_addr;
+	seg->res->end = seg->end;
+	memcpy(&seg->res_name, seg->dcss_name, 8);
+	EBCASC(seg->res_name, 8);
+	seg->res_name[8] = '\0';
+	strncat(seg->res_name, " (DCSS)", 7);
+	seg->res->name = seg->res_name;
+	rc = seg->vm_segtype;
+	if (rc == SEG_TYPE_SC ||
+	    ((rc == SEG_TYPE_SR || rc == SEG_TYPE_ER) && !do_nonshared))
+		seg->res->flags |= IORESOURCE_READONLY;
+	if (request_resource(&iomem_resource, seg->res)) {
+		rc = -EBUSY;
+		kfree(seg->res);
+		goto out_shared;
+	}
+
 	if (do_nonshared)
 		dcss_command = DCSS_LOADNSR;
 	else
@@ -316,12 +342,11 @@ __segment_load (char *name, int do_nonshared, unsigned long *addr, unsigned long
 		rc = dcss_diag_translate_rc (seg->end);
 		dcss_diag(DCSS_PURGESEG, seg->dcss_name,
 				&seg->start_addr, &seg->end);
-		goto out_shared;
+		goto out_resource;
 	}
 	seg->do_nonshared = do_nonshared;
 	atomic_set(&seg->ref_count, 1);
 	list_add(&seg->list, &dcss_list);
-	rc = seg->vm_segtype;
 	*addr = seg->start_addr;
 	*end  = seg->end;
 	if (do_nonshared)
@@ -329,12 +354,16 @@ __segment_load (char *name, int do_nonshared, unsigned long *addr, unsigned long
 				"type %s in non-shared mode\n", name,
 				(void*)seg->start_addr, (void*)seg->end,
 				segtype_string[seg->vm_segtype]);
-	else
+	else {
 		PRINT_INFO ("segment_load: loaded segment %s range %p .. %p "
 				"type %s in shared mode\n", name,
 				(void*)seg->start_addr, (void*)seg->end,
 				segtype_string[seg->vm_segtype]);
+	}
 	goto out;
+ out_resource:
+	release_resource(seg->res);
+	kfree(seg->res);
  out_shared:
 	remove_shared_memory(seg->start_addr, seg->end - seg->start_addr + 1);
  out_free:
@@ -401,6 +430,7 @@ segment_load (char *name, int do_nonshared, unsigned long *addr,
  * -ENOENT  : no such segment (segment gone!)
  * -EAGAIN  : segment is in use by other exploiters, try later
  * -EINVAL  : no segment with the given name is currently loaded - name invalid
+ * -EBUSY   : segment can temporarily not be used (overlaps with dcss)
  * 0	    : operation succeeded
  */
 int
@@ -428,12 +458,24 @@ segment_modify_shared (char *name, int do_nonshared)
 		rc = -EAGAIN;
 		goto out_unlock;
 	}
-	dcss_diag(DCSS_PURGESEG, seg->dcss_name,
-		  &dummy, &dummy);
-	if (do_nonshared)
+	release_resource(seg->res);
+	if (do_nonshared) {
 		dcss_command = DCSS_LOADNSR;
-	else
-	dcss_command = DCSS_LOADNOLY;
+		seg->res->flags &= ~IORESOURCE_READONLY;
+	} else {
+		dcss_command = DCSS_LOADNOLY;
+		if (seg->vm_segtype == SEG_TYPE_SR ||
+		    seg->vm_segtype == SEG_TYPE_ER)
+			seg->res->flags |= IORESOURCE_READONLY;
+	}
+	if (request_resource(&iomem_resource, seg->res)) {
+		PRINT_WARN("segment_modify_shared: could not reload segment %s"
+			   " - overlapping resources\n", name);
+		rc = -EBUSY;
+		kfree(seg->res);
+		goto out_del;
+	}
+	dcss_diag(DCSS_PURGESEG, seg->dcss_name, &dummy, &dummy);
 	diag_cc = dcss_diag(dcss_command, seg->dcss_name,
 			&seg->start_addr, &seg->end);
 	if (diag_cc > 1) {
@@ -446,9 +488,9 @@ segment_modify_shared (char *name, int do_nonshared)
 	rc = 0;
 	goto out_unlock;
  out_del:
+	remove_shared_memory(seg->start_addr, seg->end - seg->start_addr + 1);
 	list_del(&seg->list);
-	dcss_diag(DCSS_PURGESEG, seg->dcss_name,
-		  &dummy, &dummy);
+	dcss_diag(DCSS_PURGESEG, seg->dcss_name, &dummy, &dummy);
 	kfree(seg);
  out_unlock:
 	mutex_unlock(&dcss_lock);
@@ -478,6 +520,8 @@ segment_unload(char *name)
 	}
 	if (atomic_dec_return(&seg->ref_count) != 0)
 		goto out_unlock;
+	release_resource(seg->res);
+	kfree(seg->res);
 	remove_shared_memory(seg->start_addr, seg->end - seg->start_addr + 1);
 	list_del(&seg->list);
 	dcss_diag(DCSS_PURGESEG, seg->dcss_name, &dummy, &dummy);

arch/s390/mm/fault.c

@@ -52,7 +52,7 @@ extern int sysctl_userprocess_debug;
 extern void die(const char *,struct pt_regs *,long);
 
 #ifdef CONFIG_KPROBES
-ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
+static ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
 int register_page_fault_notifier(struct notifier_block *nb)
 {
 	return atomic_notifier_chain_register(&notify_page_fault_chain, nb);
@@ -137,7 +137,9 @@ static int __check_access_register(struct pt_regs *regs, int error_code)
 
 /*
  * Check which address space the address belongs to.
- * Returns 1 for user space and 0 for kernel space.
+ * May return 1 or 2 for user space and 0 for kernel space.
+ * Returns 2 for user space in primary addressing mode with
+ * CONFIG_S390_EXEC_PROTECT on and kernel parameter noexec=on.
  */
 static inline int check_user_space(struct pt_regs *regs, int error_code)
 {
@@ -154,7 +156,7 @@ static inline int check_user_space(struct pt_regs *regs, int error_code)
 		return __check_access_register(regs, error_code);
 	if (descriptor == 2)
 		return current->thread.mm_segment.ar4;
-	return descriptor != 0;
+	return ((descriptor != 0) ^ (switch_amode)) << s390_noexec;
 }
 
 /*
@@ -183,6 +185,77 @@ static void do_sigsegv(struct pt_regs *regs, unsigned long error_code,
 	force_sig_info(SIGSEGV, &si, current);
 }
 
+#ifdef CONFIG_S390_EXEC_PROTECT
+extern long sys_sigreturn(struct pt_regs *regs);
+extern long sys_rt_sigreturn(struct pt_regs *regs);
+extern long sys32_sigreturn(struct pt_regs *regs);
+extern long sys32_rt_sigreturn(struct pt_regs *regs);
+
+static inline void do_sigreturn(struct mm_struct *mm, struct pt_regs *regs,
+				int rt)
+{
+	up_read(&mm->mmap_sem);
+	clear_tsk_thread_flag(current, TIF_SINGLE_STEP);
+#ifdef CONFIG_COMPAT
+	if (test_tsk_thread_flag(current, TIF_31BIT)) {
+		if (rt)
+			sys32_rt_sigreturn(regs);
+		else
+			sys32_sigreturn(regs);
+		return;
+	}
+#endif /* CONFIG_COMPAT */
+	if (rt)
+		sys_rt_sigreturn(regs);
+	else
+		sys_sigreturn(regs);
+	return;
+}
+
+static int signal_return(struct mm_struct *mm, struct pt_regs *regs,
+			 unsigned long address, unsigned long error_code)
+{
+	pgd_t *pgd;
+	pmd_t *pmd;
+	pte_t *pte;
+	u16 *instruction;
+	unsigned long pfn, uaddr = regs->psw.addr;
+
+	spin_lock(&mm->page_table_lock);
+	pgd = pgd_offset(mm, uaddr);
+	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
+		goto out_fault;
+	pmd = pmd_offset(pgd, uaddr);
+	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
+		goto out_fault;
+	pte = pte_offset_map(pmd_offset(pgd_offset(mm, uaddr), uaddr), uaddr);
+	if (!pte || !pte_present(*pte))
+		goto out_fault;
+	pfn = pte_pfn(*pte);
+	if (!pfn_valid(pfn))
+		goto out_fault;
+	spin_unlock(&mm->page_table_lock);
+
+	instruction = (u16 *) ((pfn << PAGE_SHIFT) + (uaddr & (PAGE_SIZE-1)));
+	if (*instruction == 0x0a77)
+		do_sigreturn(mm, regs, 0);
+	else if (*instruction == 0x0aad)
+		do_sigreturn(mm, regs, 1);
+	else {
+		printk("- XXX - do_exception: task = %s, primary, NO EXEC "
+		       "-> SIGSEGV\n", current->comm);
+		up_read(&mm->mmap_sem);
+		current->thread.prot_addr = address;
+		current->thread.trap_no = error_code;
+		do_sigsegv(regs, error_code, SEGV_MAPERR, address);
+	}
+	return 0;
+out_fault:
+	spin_unlock(&mm->page_table_lock);
+	return -EFAULT;
+}
+#endif /* CONFIG_S390_EXEC_PROTECT */
+
 /*
  * This routine handles page faults.  It determines the address,
  * and the problem, and then passes it off to one of the appropriate
@@ -260,6 +333,17 @@ do_exception(struct pt_regs *regs, unsigned long error_code, int is_protection)
         vma = find_vma(mm, address);
         if (!vma)
                 goto bad_area;
+
+#ifdef CONFIG_S390_EXEC_PROTECT
+	if (unlikely((user_address == 2) && !(vma->vm_flags & VM_EXEC)))
+		if (!signal_return(mm, regs, address, error_code))
+			/*
+			 * signal_return() has done an up_read(&mm->mmap_sem)
+			 * if it returns 0.
+			 */
+			return;
+#endif
+
         if (vma->vm_start <= address) 
                 goto good_area;
         if (!(vma->vm_flags & VM_GROWSDOWN))
@@ -452,8 +536,7 @@ void pfault_fini(void)
 		: : "a" (&refbk), "m" (refbk) : "cc");
 }
 
-asmlinkage void
-pfault_interrupt(__u16 error_code)
+static void pfault_interrupt(__u16 error_code)
 {
 	struct task_struct *tsk;
 	__u16 subcode;

arch/s390/mm/init.c

@@ -25,7 +25,7 @@
 #include <linux/bootmem.h>
 #include <linux/pfn.h>
 #include <linux/poison.h>
-
+#include <linux/initrd.h>
 #include <asm/processor.h>
 #include <asm/system.h>
 #include <asm/uaccess.h>
@@ -95,20 +95,18 @@ static void __init setup_ro_region(void)
 	pte_t new_pte;
 	unsigned long address, end;
 
-	address = ((unsigned long)&__start_rodata) & PAGE_MASK;
-	end = PFN_ALIGN((unsigned long)&__end_rodata);
+	address = ((unsigned long)&_stext) & PAGE_MASK;
+	end = PFN_ALIGN((unsigned long)&_eshared);
 
 	for (; address < end; address += PAGE_SIZE) {
 		pgd = pgd_offset_k(address);
 		pmd = pmd_offset(pgd, address);
 		pte = pte_offset_kernel(pmd, address);
 		new_pte = mk_pte_phys(address, __pgprot(_PAGE_RO));
-		set_pte(pte, new_pte);
+		*pte = new_pte;
 	}
 }
 
-extern void vmem_map_init(void);
-
 /*
  * paging_init() sets up the page tables
  */
@@ -125,11 +123,11 @@ void __init paging_init(void)
 #ifdef CONFIG_64BIT
 	pgdir_k = (__pa(swapper_pg_dir) & PAGE_MASK) | _KERN_REGION_TABLE;
 	for (i = 0; i < PTRS_PER_PGD; i++)
-		pgd_clear(pg_dir + i);
+		pgd_clear_kernel(pg_dir + i);
 #else
 	pgdir_k = (__pa(swapper_pg_dir) & PAGE_MASK) | _KERNSEG_TABLE;
 	for (i = 0; i < PTRS_PER_PGD; i++)
-		pmd_clear((pmd_t *)(pg_dir + i));
+		pmd_clear_kernel((pmd_t *)(pg_dir + i));
 #endif
 	vmem_map_init();
 	setup_ro_region();
@@ -174,10 +172,8 @@ void __init mem_init(void)
                 datasize >>10,
                 initsize >> 10);
 	printk("Write protected kernel read-only data: %#lx - %#lx\n",
-	       (unsigned long)&__start_rodata,
-	       PFN_ALIGN((unsigned long)&__end_rodata) - 1);
-	printk("Virtual memmap size: %ldk\n",
-	       (max_pfn * sizeof(struct page)) >> 10);
+	       (unsigned long)&_stext,
+	       PFN_ALIGN((unsigned long)&_eshared) - 1);
 }
 
 void free_initmem(void)

arch/s390/mm/vmem.c

@@ -82,7 +82,7 @@ static inline pmd_t *vmem_pmd_alloc(void)
 	if (!pmd)
 		return NULL;
 	for (i = 0; i < PTRS_PER_PMD; i++)
-		pmd_clear(pmd + i);
+		pmd_clear_kernel(pmd + i);
 	return pmd;
 }
 
@@ -97,7 +97,7 @@ static inline pte_t *vmem_pte_alloc(void)
 		return NULL;
 	pte_val(empty_pte) = _PAGE_TYPE_EMPTY;
 	for (i = 0; i < PTRS_PER_PTE; i++)
-		set_pte(pte + i, empty_pte);
+		pte[i] = empty_pte;
 	return pte;
 }
 
@@ -119,7 +119,7 @@ static int vmem_add_range(unsigned long start, unsigned long size)
 			pm_dir = vmem_pmd_alloc();
 			if (!pm_dir)
 				goto out;
-			pgd_populate(&init_mm, pg_dir, pm_dir);
+			pgd_populate_kernel(&init_mm, pg_dir, pm_dir);
 		}
 
 		pm_dir = pmd_offset(pg_dir, address);
@@ -132,7 +132,7 @@ static int vmem_add_range(unsigned long start, unsigned long size)
 
 		pt_dir = pte_offset_kernel(pm_dir, address);
 		pte = pfn_pte(address >> PAGE_SHIFT, PAGE_KERNEL);
-		set_pte(pt_dir, pte);
+		*pt_dir = pte;
 	}
 	ret = 0;
 out:
@@ -161,7 +161,7 @@ static void vmem_remove_range(unsigned long start, unsigned long size)
 		if (pmd_none(*pm_dir))
 			continue;
 		pt_dir = pte_offset_kernel(pm_dir, address);
-		set_pte(pt_dir, pte);
+		*pt_dir = pte;
 	}
 	flush_tlb_kernel_range(start, start + size);
 }
@@ -191,7 +191,7 @@ static int vmem_add_mem_map(unsigned long start, unsigned long size)
 			pm_dir = vmem_pmd_alloc();
 			if (!pm_dir)
 				goto out;
-			pgd_populate(&init_mm, pg_dir, pm_dir);
+			pgd_populate_kernel(&init_mm, pg_dir, pm_dir);
 		}
 
 		pm_dir = pmd_offset(pg_dir, address);
@@ -210,7 +210,7 @@ static int vmem_add_mem_map(unsigned long start, unsigned long size)
 			if (!new_page)
 				goto out;
 			pte = pfn_pte(new_page >> PAGE_SHIFT, PAGE_KERNEL);
-			set_pte(pt_dir, pte);
+			*pt_dir = pte;
 		}
 	}
 	ret = 0;

crypto/Kconfig

@@ -74,14 +74,6 @@ config CRYPTO_SHA1
 	help
 	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2).
 
-config CRYPTO_SHA1_S390
-	tristate "SHA1 digest algorithm (s390)"
-	depends on S390
-	select CRYPTO_ALGAPI
-	help
-	  This is the s390 hardware accelerated implementation of the
-	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2).
-
 config CRYPTO_SHA256
 	tristate "SHA256 digest algorithm"
 	select CRYPTO_ALGAPI
@@ -91,17 +83,6 @@ config CRYPTO_SHA256
 	  This version of SHA implements a 256 bit hash with 128 bits of
 	  security against collision attacks.
 
-config CRYPTO_SHA256_S390
-	tristate "SHA256 digest algorithm (s390)"
-	depends on S390
-	select CRYPTO_ALGAPI
-	help
-	  This is the s390 hardware accelerated implementation of the
-	  SHA256 secure hash standard (DFIPS 180-2).
-
-	  This version of SHA implements a 256 bit hash with 128 bits of
-	  security against collision attacks.
-
 config CRYPTO_SHA512
 	tristate "SHA384 and SHA512 digest algorithms"
 	select CRYPTO_ALGAPI
@@ -187,14 +168,6 @@ config CRYPTO_DES
 	help
 	  DES cipher algorithm (FIPS 46-2), and Triple DES EDE (FIPS 46-3).
 
-config CRYPTO_DES_S390
-	tristate "DES and Triple DES cipher algorithms (s390)"
-	depends on S390
-	select CRYPTO_ALGAPI
-	select CRYPTO_BLKCIPHER
-	help
-	  DES cipher algorithm (FIPS 46-2), and Triple DES EDE (FIPS 46-3).
-
 config CRYPTO_BLOWFISH
 	tristate "Blowfish cipher algorithm"
 	select CRYPTO_ALGAPI
@@ -336,28 +309,6 @@ config CRYPTO_AES_X86_64
 
 	  See <http://csrc.nist.gov/encryption/aes/> for more information.
 
-config CRYPTO_AES_S390
-	tristate "AES cipher algorithms (s390)"
-	depends on S390
-	select CRYPTO_ALGAPI
-	select CRYPTO_BLKCIPHER
-	help
-	  This is the s390 hardware accelerated implementation of the
-	  AES cipher algorithms (FIPS-197). AES uses the Rijndael
-	  algorithm.
-
-	  Rijndael appears to be consistently a very good performer in
-	  both hardware and software across a wide range of computing
-	  environments regardless of its use in feedback or non-feedback
-	  modes. Its key setup time is excellent, and its key agility is
-	  good. Rijndael's very low memory requirements make it very well
-	  suited for restricted-space environments, in which it also
-	  demonstrates excellent performance. Rijndael's operations are
-	  among the easiest to defend against power and timing attacks.
-
-	  On s390 the System z9-109 currently only supports the key size
-	  of 128 bit.
-
 config CRYPTO_CAST5
 	tristate "CAST5 (CAST-128) cipher algorithm"
 	select CRYPTO_ALGAPI

drivers/crypto/Kconfig

@@ -51,6 +51,8 @@ config CRYPTO_DEV_PADLOCK_SHA
 	  If unsure say M. The compiled module will be
 	  called padlock-sha.ko
 
+source "arch/s390/crypto/Kconfig"
+
 config CRYPTO_DEV_GEODE
 	tristate "Support for the Geode LX AES engine"
 	depends on CRYPTO && X86_32 && PCI

drivers/s390/Kconfig

@@ -103,14 +103,8 @@ config CCW_CONSOLE
  	depends on TN3215_CONSOLE || TN3270_CONSOLE
  	default y
  
-config SCLP
-	bool "Support for SCLP"
-	help
-	  Include support for the SCLP interface to the service element.
-
 config SCLP_TTY
 	bool "Support for SCLP line mode terminal"
-	depends on SCLP
 	help
 	  Include support for IBM SCLP line-mode terminals.
 
@@ -123,7 +117,6 @@ config SCLP_CONSOLE
 
 config SCLP_VT220_TTY
 	bool "Support for SCLP VT220-compatible terminal"
-	depends on SCLP
 	help
 	  Include support for an IBM SCLP VT220-compatible terminal.
 
@@ -136,7 +129,6 @@ config SCLP_VT220_CONSOLE
 
 config SCLP_CPI
 	tristate "Control-Program Identification"
-	depends on SCLP
 	help
 	  This option enables the hardware console interface for system
 	  identification. This is commonly used for workload management and

drivers/s390/Makefile

@@ -2,6 +2,8 @@
 # Makefile for the S/390 specific device drivers
 #
 
+CFLAGS_sysinfo.o += -Iinclude/math-emu -Iarch/s390/math-emu -w
+
 obj-y += s390mach.o sysinfo.o s390_rdev.o
 obj-y += cio/ block/ char/ crypto/ net/ scsi/
 

drivers/s390/block/dasd.c

@@ -37,6 +37,7 @@
  */
 debug_info_t *dasd_debug_area;
 struct dasd_discipline *dasd_diag_discipline_pointer;
+void dasd_int_handler(struct ccw_device *, unsigned long, struct irb *);
 
 MODULE_AUTHOR("Holger Smolinski <Holger.Smolinski@de.ibm.com>");
 MODULE_DESCRIPTION("Linux on S/390 DASD device driver,"
@@ -51,7 +52,6 @@ static int  dasd_alloc_queue(struct dasd_device * device);
 static void dasd_setup_queue(struct dasd_device * device);
 static void dasd_free_queue(struct dasd_device * device);
 static void dasd_flush_request_queue(struct dasd_device *);
-static void dasd_int_handler(struct ccw_device *, unsigned long, struct irb *);
 static int dasd_flush_ccw_queue(struct dasd_device *, int);
 static void dasd_tasklet(struct dasd_device *);
 static void do_kick_device(struct work_struct *);
@@ -483,7 +483,7 @@ unsigned int dasd_profile_level = DASD_PROFILE_OFF;
 /*
  * Add profiling information for cqr before execution.
  */
-static inline void
+static void
 dasd_profile_start(struct dasd_device *device, struct dasd_ccw_req * cqr,
 		   struct request *req)
 {
@@ -505,7 +505,7 @@ dasd_profile_start(struct dasd_device *device, struct dasd_ccw_req * cqr,
 /*
  * Add profiling information for cqr after execution.
  */
-static inline void
+static void
 dasd_profile_end(struct dasd_device *device, struct dasd_ccw_req * cqr,
 		 struct request *req)
 {
@@ -1022,8 +1022,6 @@ dasd_int_handler(struct ccw_device *cdev, unsigned long intparm,
 		 irb->scsw.cstat == 0 &&
 		 !irb->esw.esw0.erw.cons)
 		era = dasd_era_none;
-	else if (!test_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags))
- 	        era = dasd_era_fatal; /* don't recover this request */
 	else if (irb->esw.esw0.erw.cons)
 		era = device->discipline->examine_error(cqr, irb);
 	else
@@ -1104,7 +1102,7 @@ __dasd_process_erp(struct dasd_device *device, struct dasd_ccw_req *cqr)
 /*
  * Process ccw request queue.
  */
-static inline void
+static void
 __dasd_process_ccw_queue(struct dasd_device * device,
 			 struct list_head *final_queue)
 {
@@ -1127,7 +1125,9 @@ restart:
 				cqr->status = DASD_CQR_FAILED;
 				cqr->stopclk = get_clock();
 			} else {
-				if (cqr->irb.esw.esw0.erw.cons) {
+				if (cqr->irb.esw.esw0.erw.cons &&
+				    test_bit(DASD_CQR_FLAGS_USE_ERP,
+					     &cqr->flags)) {
 					erp_fn = device->discipline->
 						erp_action(cqr);
 					erp_fn(cqr);
@@ -1181,7 +1181,7 @@ dasd_end_request_cb(struct dasd_ccw_req * cqr, void *data)
 /*
  * Fetch requests from the block device queue.
  */
-static inline void
+static void
 __dasd_process_blk_queue(struct dasd_device * device)
 {
 	request_queue_t *queue;
@@ -1232,6 +1232,19 @@ __dasd_process_blk_queue(struct dasd_device * device)
 		if (IS_ERR(cqr)) {
 			if (PTR_ERR(cqr) == -ENOMEM)
 				break;	/* terminate request queue loop */
+			if (PTR_ERR(cqr) == -EAGAIN) {
+				/*
+				 * The current request cannot be build right
+				 * now, we have to try later. If this request
+				 * is the head-of-queue we stop the device
+				 * for 1/2 second.
+				 */
+				if (!list_empty(&device->ccw_queue))
+					break;
+				device->stopped |= DASD_STOPPED_PENDING;
+				dasd_set_timer(device, HZ/2);
+				break;
+			}
 			DBF_DEV_EVENT(DBF_ERR, device,
 				      "CCW creation failed (rc=%ld) "
 				      "on request %p",
@@ -1254,7 +1267,7 @@ __dasd_process_blk_queue(struct dasd_device * device)
  * Take a look at the first request on the ccw queue and check
  * if it reached its expire time. If so, terminate the IO.
  */
-static inline void
+static void
 __dasd_check_expire(struct dasd_device * device)
 {
 	struct dasd_ccw_req *cqr;
@@ -1285,7 +1298,7 @@ __dasd_check_expire(struct dasd_device * device)
  * Take a look at the first request on the ccw queue and check
  * if it needs to be started.
  */
-static inline void
+static void
 __dasd_start_head(struct dasd_device * device)
 {
 	struct dasd_ccw_req *cqr;

drivers/s390/block/dasd_3990_erp.c

@@ -170,7 +170,6 @@ dasd_3990_erp_examine(struct dasd_ccw_req * cqr, struct irb * irb)
 	/* log the erp chain if fatal error occurred */
 	if ((era == dasd_era_fatal) && (device->state >= DASD_STATE_READY)) {
 		dasd_log_sense(cqr, irb);
-		dasd_log_ccw(cqr, 0, irb->scsw.cpa);
 	}
 
 	return era;
@@ -2640,7 +2639,6 @@ dasd_3990_erp_action(struct dasd_ccw_req * cqr)
 
 	struct dasd_ccw_req *erp = NULL;
 	struct dasd_device *device = cqr->device;
-	__u32 cpa = cqr->irb.scsw.cpa;
 	struct dasd_ccw_req *temp_erp = NULL;
 
 	if (device->features & DASD_FEATURE_ERPLOG) {
@@ -2706,9 +2704,6 @@ dasd_3990_erp_action(struct dasd_ccw_req * cqr)
 		}
 	}
 
-	if (erp->status == DASD_CQR_FAILED)
-		dasd_log_ccw(erp, 1, cpa);
-
 	/* enqueue added ERP request */
 	if (erp->status == DASD_CQR_FILLED) {
 		erp->status = DASD_CQR_QUEUED;

drivers/s390/block/dasd_devmap.c

@@ -136,7 +136,7 @@ __setup ("dasd=", dasd_call_setup);
 /*
  * Read a device busid/devno from a string.
  */
-static inline int
+static int
 dasd_busid(char **str, int *id0, int *id1, int *devno)
 {
 	int val, old_style;
@@ -182,7 +182,7 @@ dasd_busid(char **str, int *id0, int *id1, int *devno)
  * only one: "ro" for read-only devices. The default feature set
  * is empty (value 0).
  */
-static inline int
+static int
 dasd_feature_list(char *str, char **endp)
 {
 	int features, len, rc;
@@ -341,7 +341,7 @@ dasd_parse_range( char *parsestring ) {
 	return ERR_PTR(-EINVAL);
 }
 
-static inline char *
+static char *
 dasd_parse_next_element( char *parsestring ) {
 	char * residual_str;
 	residual_str = dasd_parse_keyword(parsestring);

drivers/s390/block/dasd_diag.c

@@ -43,7 +43,7 @@ MODULE_LICENSE("GPL");
 #define DIAG_MAX_RETRIES	32
 #define DIAG_TIMEOUT		50 * HZ
 
-struct dasd_discipline dasd_diag_discipline;
+static struct dasd_discipline dasd_diag_discipline;
 
 struct dasd_diag_private {
 	struct dasd_diag_characteristics rdc_data;
@@ -90,7 +90,7 @@ static inline int dia250(void *iob, int cmd)
  * block offset. On success, return zero and set end_block to contain the
  * number of blocks on the device minus the specified offset. Return non-zero
  * otherwise. */
-static __inline__ int
+static inline int
 mdsk_init_io(struct dasd_device *device, unsigned int blocksize,
 	     blocknum_t offset, blocknum_t *end_block)
 {
@@ -117,7 +117,7 @@ mdsk_init_io(struct dasd_device *device, unsigned int blocksize,
 
 /* Remove block I/O environment for device. Return zero on success, non-zero
  * otherwise. */
-static __inline__ int
+static inline int
 mdsk_term_io(struct dasd_device * device)
 {
 	struct dasd_diag_private *private;
@@ -576,7 +576,7 @@ dasd_diag_dump_sense(struct dasd_device *device, struct dasd_ccw_req * req,
 		    "dump sense not available for DIAG data");
 }
 
-struct dasd_discipline dasd_diag_discipline = {
+static struct dasd_discipline dasd_diag_discipline = {
 	.owner = THIS_MODULE,
 	.name = "DIAG",
 	.ebcname = "DIAG",

drivers/s390/block/dasd_eckd.c

@@ -134,44 +134,7 @@ ceil_quot(unsigned int d1, unsigned int d2)
 	return (d1 + (d2 - 1)) / d2;
 }
 
-static inline int
-bytes_per_record(struct dasd_eckd_characteristics *rdc, int kl, int dl)
-{
-	unsigned int fl1, fl2, int1, int2;
-	int bpr;
-
-	switch (rdc->formula) {
-	case 0x01:
-		fl1 = round_up_multiple(ECKD_F2(rdc) + dl, ECKD_F1(rdc));
-		fl2 = round_up_multiple(kl ? ECKD_F2(rdc) + kl : 0,
-					ECKD_F1(rdc));
-		bpr = fl1 + fl2;
-		break;
-	case 0x02:
-		int1 = ceil_quot(dl + ECKD_F6(rdc), ECKD_F5(rdc) << 1);
-		int2 = ceil_quot(kl + ECKD_F6(rdc), ECKD_F5(rdc) << 1);
-		fl1 = round_up_multiple(ECKD_F1(rdc) * ECKD_F2(rdc) + dl +
-					ECKD_F6(rdc) + ECKD_F4(rdc) * int1,
-					ECKD_F1(rdc));
-		fl2 = round_up_multiple(ECKD_F1(rdc) * ECKD_F3(rdc) + kl +
-					ECKD_F6(rdc) + ECKD_F4(rdc) * int2,
-					ECKD_F1(rdc));
-		bpr = fl1 + fl2;
-		break;
-	default:
-		bpr = 0;
-		break;
-	}
-	return bpr;
-}
-
-static inline unsigned int
-bytes_per_track(struct dasd_eckd_characteristics *rdc)
-{
-	return *(unsigned int *) (rdc->byte_per_track) >> 8;
-}
-
-static inline unsigned int
+static unsigned int
 recs_per_track(struct dasd_eckd_characteristics * rdc,
 	       unsigned int kl, unsigned int dl)
 {
@@ -204,37 +167,39 @@ recs_per_track(struct dasd_eckd_characteristics * rdc,
 	return 0;
 }
 
-static inline void
+static int
 check_XRC (struct ccw1         *de_ccw,
            struct DE_eckd_data *data,
            struct dasd_device  *device)
 {
         struct dasd_eckd_private *private;
+	int rc;
 
         private = (struct dasd_eckd_private *) device->private;
+	if (!private->rdc_data.facilities.XRC_supported)
+		return 0;
 
         /* switch on System Time Stamp - needed for XRC Support */
-        if (private->rdc_data.facilities.XRC_supported) {
-
-                data->ga_extended |= 0x08; /* switch on 'Time Stamp Valid'   */
-                data->ga_extended |= 0x02; /* switch on 'Extended Parameter' */
-
-                data->ep_sys_time = get_clock ();
-
-                de_ccw->count = sizeof (struct DE_eckd_data);
-		de_ccw->flags |= CCW_FLAG_SLI;
-        }
+	data->ga_extended |= 0x08; /* switch on 'Time Stamp Valid'   */
+	data->ga_extended |= 0x02; /* switch on 'Extended Parameter' */
 
-        return;
+	rc = get_sync_clock(&data->ep_sys_time);
+	/* Ignore return code if sync clock is switched off. */
+	if (rc == -ENOSYS || rc == -EACCES)
+		rc = 0;
 
-} /* end check_XRC */
+	de_ccw->count = sizeof (struct DE_eckd_data);
+	de_ccw->flags |= CCW_FLAG_SLI;
+	return rc;
+}
 
-static inline void
+static int
 define_extent(struct ccw1 * ccw, struct DE_eckd_data * data, int trk,
 	      int totrk, int cmd, struct dasd_device * device)
 {
 	struct dasd_eckd_private *private;
 	struct ch_t geo, beg, end;
+	int rc = 0;
 
 	private = (struct dasd_eckd_private *) device->private;
 
@@ -263,12 +228,12 @@ define_extent(struct ccw1 * ccw, struct DE_eckd_data * data, int trk,
 	case DASD_ECKD_CCW_WRITE_KD_MT:
 		data->mask.perm = 0x02;
 		data->attributes.operation = private->attrib.operation;
-                check_XRC (ccw, data, device);
+		rc = check_XRC (ccw, data, device);
 		break;
 	case DASD_ECKD_CCW_WRITE_CKD:
 	case DASD_ECKD_CCW_WRITE_CKD_MT:
 		data->attributes.operation = DASD_BYPASS_CACHE;
-                check_XRC (ccw, data, device);
+		rc = check_XRC (ccw, data, device);
 		break;
 	case DASD_ECKD_CCW_ERASE:
 	case DASD_ECKD_CCW_WRITE_HOME_ADDRESS:
@@ -276,7 +241,7 @@ define_extent(struct ccw1 * ccw, struct DE_eckd_data * data, int trk,
 		data->mask.perm = 0x3;
 		data->mask.auth = 0x1;
 		data->attributes.operation = DASD_BYPASS_CACHE;
-                check_XRC (ccw, data, device);
+		rc = check_XRC (ccw, data, device);
 		break;
 	default:
 		DEV_MESSAGE(KERN_ERR, device, "unknown opcode 0x%x", cmd);
@@ -312,9 +277,10 @@ define_extent(struct ccw1 * ccw, struct DE_eckd_data * data, int trk,
 	data->beg_ext.head = beg.head;
 	data->end_ext.cyl = end.cyl;
 	data->end_ext.head = end.head;
+	return rc;
 }
 
-static inline void
+static void
 locate_record(struct ccw1 *ccw, struct LO_eckd_data *data, int trk,
 	      int rec_on_trk, int no_rec, int cmd,
 	      struct dasd_device * device, int reclen)
@@ -548,7 +514,7 @@ dasd_eckd_read_conf(struct dasd_device *device)
 /*
  * Build CP for Perform Subsystem Function - SSC.
  */
-struct dasd_ccw_req *
+static struct dasd_ccw_req *
 dasd_eckd_build_psf_ssc(struct dasd_device *device)
 {
        struct dasd_ccw_req *cqr;
@@ -1200,7 +1166,12 @@ dasd_eckd_build_cp(struct dasd_device * device, struct request *req)
 		return cqr;
 	ccw = cqr->cpaddr;
 	/* First ccw is define extent. */
-	define_extent(ccw++, cqr->data, first_trk, last_trk, cmd, device);
+	if (define_extent(ccw++, cqr->data, first_trk,
+			  last_trk, cmd, device) == -EAGAIN) {
+		/* Clock not in sync and XRC is enabled. Try again later. */
+		dasd_sfree_request(cqr, device);
+		return ERR_PTR(-EAGAIN);
+	}
 	/* Build locate_record+read/write/ccws. */
 	idaws = (unsigned long *) (cqr->data + sizeof(struct DE_eckd_data));
 	LO_data = (struct LO_eckd_data *) (idaws + cidaw);
@@ -1380,7 +1351,7 @@ dasd_eckd_release(struct dasd_device *device)
 	cqr->device = device;
 	clear_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags);
 	set_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags);
-	cqr->retries = 0;
+	cqr->retries = 2;	/* set retry counter to enable basic ERP */
 	cqr->expires = 2 * HZ;
 	cqr->buildclk = get_clock();
 	cqr->status = DASD_CQR_FILLED;
@@ -1420,7 +1391,7 @@ dasd_eckd_reserve(struct dasd_device *device)
 	cqr->device = device;
 	clear_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags);
 	set_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags);
-	cqr->retries = 0;
+	cqr->retries = 2;	/* set retry counter to enable basic ERP */
 	cqr->expires = 2 * HZ;
 	cqr->buildclk = get_clock();
 	cqr->status = DASD_CQR_FILLED;
@@ -1459,7 +1430,7 @@ dasd_eckd_steal_lock(struct dasd_device *device)
 	cqr->device = device;
 	clear_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags);
 	set_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags);
-	cqr->retries = 0;
+	cqr->retries = 2;	/* set retry counter to enable basic ERP */
 	cqr->expires = 2 * HZ;
 	cqr->buildclk = get_clock();
 	cqr->status = DASD_CQR_FILLED;
@@ -1609,7 +1580,7 @@ dasd_eckd_ioctl(struct dasd_device *device, unsigned int cmd, void __user *argp)
  * Dump the range of CCWs into 'page' buffer
  * and return number of printed chars.
  */
-static inline int
+static int
 dasd_eckd_dump_ccw_range(struct ccw1 *from, struct ccw1 *to, char *page)
 {
 	int len, count;

drivers/s390/block/dasd_eer.c

@@ -658,18 +658,24 @@ static struct file_operations dasd_eer_fops = {
 	.owner		= THIS_MODULE,
 };
 
-static struct miscdevice dasd_eer_dev = {
-	.minor	    = MISC_DYNAMIC_MINOR,
-	.name	    = "dasd_eer",
-	.fops	    = &dasd_eer_fops,
-};
+static struct miscdevice *dasd_eer_dev = NULL;
 
 int __init dasd_eer_init(void)
 {
 	int rc;
 
-	rc = misc_register(&dasd_eer_dev);
+	dasd_eer_dev = kzalloc(sizeof(*dasd_eer_dev), GFP_KERNEL);
+	if (!dasd_eer_dev)
+		return -ENOMEM;
+
+	dasd_eer_dev->minor = MISC_DYNAMIC_MINOR;
+	dasd_eer_dev->name  = "dasd_eer";
+	dasd_eer_dev->fops  = &dasd_eer_fops;
+
+	rc = misc_register(dasd_eer_dev);
 	if (rc) {
+		kfree(dasd_eer_dev);
+		dasd_eer_dev = NULL;
 		MESSAGE(KERN_ERR, "%s", "dasd_eer_init could not "
 		       "register misc device");
 		return rc;
@@ -680,5 +686,9 @@ int __init dasd_eer_init(void)
 
 void dasd_eer_exit(void)
 {
-	WARN_ON(misc_deregister(&dasd_eer_dev) != 0);
+	if (dasd_eer_dev) {
+		WARN_ON(misc_deregister(dasd_eer_dev) != 0);
+		kfree(dasd_eer_dev);
+		dasd_eer_dev = NULL;
+	}
 }

drivers/s390/block/dasd_erp.c

@@ -152,25 +152,6 @@ dasd_default_erp_postaction(struct dasd_ccw_req * cqr)
 
 }				/* end default_erp_postaction */
 
-/*
- * Print the hex dump of the memory used by a request. This includes
- * all error recovery ccws that have been chained in from of the
- * real request.
- */
-static inline void
-hex_dump_memory(struct dasd_device *device, void *data, int len)
-{
-	int *pint;
-
-	pint = (int *) data;
-	while (len > 0) {
-		DEV_MESSAGE(KERN_ERR, device, "%p: %08x %08x %08x %08x",
-			    pint, pint[0], pint[1], pint[2], pint[3]);
-		pint += 4;
-		len -= 16;
-	}
-}
-
 void
 dasd_log_sense(struct dasd_ccw_req *cqr, struct irb *irb)
 {
@@ -182,69 +163,8 @@ dasd_log_sense(struct dasd_ccw_req *cqr, struct irb *irb)
 		device->discipline->dump_sense(device, cqr, irb);
 }
 
-void
-dasd_log_ccw(struct dasd_ccw_req * cqr, int caller, __u32 cpa)
-{
-	struct dasd_device *device;
-	struct dasd_ccw_req *lcqr;
-	struct ccw1 *ccw;
-	int cplength;
-
-	device = cqr->device;
-	/* log the channel program */
-	for (lcqr = cqr; lcqr != NULL; lcqr = lcqr->refers) {
-		DEV_MESSAGE(KERN_ERR, device,
-			    "(%s) ERP chain report for req: %p",
-			    caller == 0 ? "EXAMINE" : "ACTION", lcqr);
-		hex_dump_memory(device, lcqr, sizeof(struct dasd_ccw_req));
-
-		cplength = 1;
-		ccw = lcqr->cpaddr;
-		while (ccw++->flags & (CCW_FLAG_DC | CCW_FLAG_CC))
-			cplength++;
-
-		if (cplength > 40) {	/* log only parts of the CP */
-			DEV_MESSAGE(KERN_ERR, device, "%s",
-				    "Start of channel program:");
-			hex_dump_memory(device, lcqr->cpaddr,
-					40*sizeof(struct ccw1));
-
-			DEV_MESSAGE(KERN_ERR, device, "%s",
-				    "End of channel program:");
-			hex_dump_memory(device, lcqr->cpaddr + cplength - 10,
-					10*sizeof(struct ccw1));
-		} else {	/* log the whole CP */
-			DEV_MESSAGE(KERN_ERR, device, "%s",
-				    "Channel program (complete):");
-			hex_dump_memory(device, lcqr->cpaddr,
-					cplength*sizeof(struct ccw1));
-		}
-
-		if (lcqr != cqr)
-			continue;
-
-		/*
-		 * Log bytes arround failed CCW but only if we did
-		 * not log the whole CP of the CCW is outside the
-		 * logged CP.
-		 */
-		if (cplength > 40 ||
-		    ((addr_t) cpa < (addr_t) lcqr->cpaddr &&
-		     (addr_t) cpa > (addr_t) (lcqr->cpaddr + cplength + 4))) {
-
-			DEV_MESSAGE(KERN_ERR, device,
-				    "Failed CCW (%p) (area):",
-				    (void *) (long) cpa);
-			hex_dump_memory(device, cqr->cpaddr - 10,
-					20*sizeof(struct ccw1));
-		}
-	}
-
-}				/* end log_erp_chain */
-
 EXPORT_SYMBOL(dasd_default_erp_action);
 EXPORT_SYMBOL(dasd_default_erp_postaction);
 EXPORT_SYMBOL(dasd_alloc_erp_request);
 EXPORT_SYMBOL(dasd_free_erp_request);
 EXPORT_SYMBOL(dasd_log_sense);
-EXPORT_SYMBOL(dasd_log_ccw);

drivers/s390/block/dasd_fba.c

@@ -75,7 +75,7 @@ static struct ccw_driver dasd_fba_driver = {
 	.notify      = dasd_generic_notify,
 };
 
-static inline void
+static void
 define_extent(struct ccw1 * ccw, struct DE_fba_data *data, int rw,
 	      int blksize, int beg, int nr)
 {
@@ -95,7 +95,7 @@ define_extent(struct ccw1 * ccw, struct DE_fba_data *data, int rw,
 	data->ext_end = nr - 1;
 }
 
-static inline void
+static void
 locate_record(struct ccw1 * ccw, struct LO_fba_data *data, int rw,
 	      int block_nr, int block_ct)
 {

drivers/s390/block/dasd_genhd.c

@@ -147,7 +147,7 @@ dasd_destroy_partitions(struct dasd_device * device)
 	 */
 	memset(&bpart, 0, sizeof(struct blkpg_partition));
 	memset(&barg, 0, sizeof(struct blkpg_ioctl_arg));
-	barg.data = (void __user *) &bpart;
+	barg.data = (void __force __user *) &bpart;
 	barg.op = BLKPG_DEL_PARTITION;
 	for (bpart.pno = device->gdp->minors - 1; bpart.pno > 0; bpart.pno--)
 		ioctl_by_bdev(bdev, BLKPG, (unsigned long) &barg);

drivers/s390/block/dasd_int.h

@@ -559,7 +559,6 @@ struct dasd_ccw_req *dasd_alloc_erp_request(char *, int, int,
 					    struct dasd_device *);
 void dasd_free_erp_request(struct dasd_ccw_req *, struct dasd_device *);
 void dasd_log_sense(struct dasd_ccw_req *, struct irb *);
-void dasd_log_ccw(struct dasd_ccw_req *, int, __u32);
 
 /* externals in dasd_3370_erp.c */
 dasd_era_t dasd_3370_erp_examine(struct dasd_ccw_req *, struct irb *);

drivers/s390/block/dasd_proc.c

@@ -28,7 +28,7 @@ static struct proc_dir_entry *dasd_proc_root_entry = NULL;
 static struct proc_dir_entry *dasd_devices_entry = NULL;
 static struct proc_dir_entry *dasd_statistics_entry = NULL;
 
-static inline char *
+static char *
 dasd_get_user_string(const char __user *user_buf, size_t user_len)
 {
 	char *buffer;
@@ -154,7 +154,7 @@ static struct file_operations dasd_devices_file_ops = {
 	.release	= seq_release,
 };
 
-static inline int
+static int
 dasd_calc_metrics(char *page, char **start, off_t off,
 		  int count, int *eof, int len)
 {
@@ -167,8 +167,8 @@ dasd_calc_metrics(char *page, char **start, off_t off,
 	return len;
 }
 
-static inline char *
-dasd_statistics_array(char *str, int *array, int shift)
+static char *
+dasd_statistics_array(char *str, unsigned int *array, int shift)
 {
 	int i;
 

drivers/s390/block/dcssblk.c

@@ -102,7 +102,7 @@ dcssblk_release_segment(struct device *dev)
  * device needs to be enqueued before the semaphore is
  * freed.
  */
-static inline int
+static int
 dcssblk_assign_free_minor(struct dcssblk_dev_info *dev_info)
 {
 	int minor, found;
@@ -230,7 +230,7 @@ dcssblk_shared_store(struct device *dev, struct device_attribute *attr, const ch
 					   SEGMENT_SHARED);
 		if (rc < 0) {
 			BUG_ON(rc == -EINVAL);
-			if (rc == -EIO || rc == -ENOENT)
+			if (rc != -EAGAIN)
 				goto removeseg;
 		} else {
 			dev_info->is_shared = 1;
@@ -253,7 +253,7 @@ dcssblk_shared_store(struct device *dev, struct device_attribute *attr, const ch
 					   SEGMENT_EXCLUSIVE);
 		if (rc < 0) {
 			BUG_ON(rc == -EINVAL);
-			if (rc == -EIO || rc == -ENOENT)
+			if (rc != -EAGAIN)
 				goto removeseg;
 		} else {
 			dev_info->is_shared = 0;

drivers/s390/char/Makefile

@@ -2,7 +2,8 @@
 # S/390 character devices
 #
 
-obj-y += ctrlchar.o keyboard.o defkeymap.o
+obj-y += ctrlchar.o keyboard.o defkeymap.o sclp.o sclp_rw.o sclp_quiesce.o \
+	 sclp_info.o
 
 obj-$(CONFIG_TN3270) += raw3270.o
 obj-$(CONFIG_TN3270_CONSOLE) += con3270.o
@@ -11,7 +12,6 @@ obj-$(CONFIG_TN3270_FS) += fs3270.o
 
 obj-$(CONFIG_TN3215) += con3215.o
 
-obj-$(CONFIG_SCLP) += sclp.o sclp_rw.o sclp_quiesce.o
 obj-$(CONFIG_SCLP_TTY) += sclp_tty.o
 obj-$(CONFIG_SCLP_CONSOLE) += sclp_con.o
 obj-$(CONFIG_SCLP_VT220_TTY) += sclp_vt220.o

drivers/s390/char/con3215.c

@@ -1121,7 +1121,7 @@ static const struct tty_operations tty3215_ops = {
  * 3215 tty registration code called from tty_init().
  * Most kernel services (incl. kmalloc) are available at this poimt.
  */
-int __init
+static int __init
 tty3215_init(void)
 {
 	struct tty_driver *driver;

drivers/s390/char/con3270.c

@@ -69,8 +69,7 @@ static void con3270_update(struct con3270 *);
 /*
  * Setup timeout for a device. On timeout trigger an update.
  */
-void
-con3270_set_timer(struct con3270 *cp, int expires)
+static void con3270_set_timer(struct con3270 *cp, int expires)
 {
 	if (expires == 0) {
 		if (timer_pending(&cp->timer))

drivers/s390/char/defkeymap.c

@@ -5,6 +5,8 @@
 #include <linux/types.h>
 #include <linux/keyboard.h>
 #include <linux/kd.h>
+#include <linux/kbd_kern.h>
+#include <linux/kbd_diacr.h>
 
 u_short plain_map[NR_KEYS] = {
 	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,

drivers/s390/char/fs3270.c

@@ -23,7 +23,7 @@
 #include "raw3270.h"
 #include "ctrlchar.h"
 
-struct raw3270_fn fs3270_fn;
+static struct raw3270_fn fs3270_fn;
 
 struct fs3270 {
 	struct raw3270_view view;
@@ -401,7 +401,7 @@ fs3270_release(struct raw3270_view *view)
 }
 
 /* View to a 3270 device. Can be console, tty or fullscreen. */
-struct raw3270_fn fs3270_fn = {
+static struct raw3270_fn fs3270_fn = {
 	.activate = fs3270_activate,
 	.deactivate = fs3270_deactivate,
 	.intv = (void *) fs3270_irq,

drivers/s390/char/keyboard.c

@@ -148,6 +148,7 @@ kbd_ascebc(struct kbd_data *kbd, unsigned char *ascebc)
 	}
 }
 
+#if 0
 /*
  * Generate ebcdic -> ascii translation table from kbd_data.
  */
@@ -173,6 +174,7 @@ kbd_ebcasc(struct kbd_data *kbd, unsigned char *ebcasc)
 		}
 	}
 }
+#endif
 
 /*
  * We have a combining character DIACR here, followed by the character CH.

drivers/s390/char/monwriter.c

@@ -67,8 +67,8 @@ static int monwrite_diag(struct monwrite_hdr *myhdr, char *buffer, int fcn)
 	return -EINVAL;
 }
 
-static inline struct mon_buf *monwrite_find_hdr(struct mon_private *monpriv,
-						struct monwrite_hdr *monhdr)
+static struct mon_buf *monwrite_find_hdr(struct mon_private *monpriv,
+					 struct monwrite_hdr *monhdr)
 {
 	struct mon_buf *entry, *next;
 

drivers/s390/char/raw3270.c

@@ -29,7 +29,7 @@
 #include <linux/device.h>
 #include <linux/mutex.h>
 
-struct class *class3270;
+static struct class *class3270;
 
 /* The main 3270 data structure. */
 struct raw3270 {
@@ -86,7 +86,7 @@ DECLARE_WAIT_QUEUE_HEAD(raw3270_wait_queue);
 /*
  * Encode array for 12 bit 3270 addresses.
  */
-unsigned char raw3270_ebcgraf[64] =	{
+static unsigned char raw3270_ebcgraf[64] =	{
 	0x40, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7,
 	0xc8, 0xc9, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f,
 	0x50, 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7,

drivers/s390/char/sclp.c

@@ -59,7 +59,8 @@ static volatile enum sclp_init_state_t {
 /* Internal state: is a request active at the sclp? */
 static volatile enum sclp_running_state_t {
 	sclp_running_state_idle,
-	sclp_running_state_running
+	sclp_running_state_running,
+	sclp_running_state_reset_pending
 } sclp_running_state = sclp_running_state_idle;
 
 /* Internal state: is a read request pending? */
@@ -88,15 +89,15 @@ static volatile enum sclp_mask_state_t {
 
 /* Timeout intervals in seconds.*/
 #define SCLP_BUSY_INTERVAL	10
-#define SCLP_RETRY_INTERVAL	15
+#define SCLP_RETRY_INTERVAL	30
 
 static void sclp_process_queue(void);
 static int sclp_init_mask(int calculate);
 static int sclp_init(void);
 
 /* Perform service call. Return 0 on success, non-zero otherwise. */
-static int
-service_call(sclp_cmdw_t command, void *sccb)
+int
+sclp_service_call(sclp_cmdw_t command, void *sccb)
 {
 	int cc;
 
@@ -113,19 +114,17 @@ service_call(sclp_cmdw_t command, void *sccb)
 	return 0;
 }
 
-/* Request timeout handler. Restart the request queue. If DATA is non-zero,
- * force restart of running request. */
+static inline void __sclp_make_read_req(void);
+
 static void
-sclp_request_timeout(unsigned long data)
+__sclp_queue_read_req(void)
 {
-	unsigned long flags;
-
-	if (data) {
-		spin_lock_irqsave(&sclp_lock, flags);
-		sclp_running_state = sclp_running_state_idle;
-		spin_unlock_irqrestore(&sclp_lock, flags);
+	if (sclp_reading_state == sclp_reading_state_idle) {
+		sclp_reading_state = sclp_reading_state_reading;
+		__sclp_make_read_req();
+		/* Add request to head of queue */
+		list_add(&sclp_read_req.list, &sclp_req_queue);
 	}
-	sclp_process_queue();
 }
 
 /* Set up request retry timer. Called while sclp_lock is locked. */
@@ -140,6 +139,29 @@ __sclp_set_request_timer(unsigned long time, void (*function)(unsigned long),
 	add_timer(&sclp_request_timer);
 }
 
+/* Request timeout handler. Restart the request queue. If DATA is non-zero,
+ * force restart of running request. */
+static void
+sclp_request_timeout(unsigned long data)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	if (data) {
+		if (sclp_running_state == sclp_running_state_running) {
+			/* Break running state and queue NOP read event request
+			 * to get a defined interface state. */
+			__sclp_queue_read_req();
+			sclp_running_state = sclp_running_state_idle;
+		}
+	} else {
+		__sclp_set_request_timer(SCLP_BUSY_INTERVAL * HZ,
+					 sclp_request_timeout, 0);
+	}
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	sclp_process_queue();
+}
+
 /* Try to start a request. Return zero if the request was successfully
  * started or if it will be started at a later time. Return non-zero otherwise.
  * Called while sclp_lock is locked. */
@@ -151,7 +173,7 @@ __sclp_start_request(struct sclp_req *req)
 	if (sclp_running_state != sclp_running_state_idle)
 		return 0;
 	del_timer(&sclp_request_timer);
-	rc = service_call(req->command, req->sccb);
+	rc = sclp_service_call(req->command, req->sccb);
 	req->start_count++;
 
 	if (rc == 0) {
@@ -191,7 +213,15 @@ sclp_process_queue(void)
 		rc = __sclp_start_request(req);
 		if (rc == 0)
 			break;
-		/* Request failed. */
+		/* Request failed */
+		if (req->start_count > 1) {
+			/* Cannot abort already submitted request - could still
+			 * be active at the SCLP */
+			__sclp_set_request_timer(SCLP_BUSY_INTERVAL * HZ,
+						 sclp_request_timeout, 0);
+			break;
+		}
+		/* Post-processing for aborted request */
 		list_del(&req->list);
 		if (req->callback) {
 			spin_unlock_irqrestore(&sclp_lock, flags);
@@ -221,7 +251,8 @@ sclp_add_request(struct sclp_req *req)
 	list_add_tail(&req->list, &sclp_req_queue);
 	rc = 0;
 	/* Start if request is first in list */
-	if (req->list.prev == &sclp_req_queue) {
+	if (sclp_running_state == sclp_running_state_idle &&
+	    req->list.prev == &sclp_req_queue) {
 		rc = __sclp_start_request(req);
 		if (rc)
 			list_del(&req->list);
@@ -294,7 +325,7 @@ __sclp_make_read_req(void)
 	sccb = (struct sccb_header *) sclp_read_sccb;
 	clear_page(sccb);
 	memset(&sclp_read_req, 0, sizeof(struct sclp_req));
-	sclp_read_req.command = SCLP_CMDW_READDATA;
+	sclp_read_req.command = SCLP_CMDW_READ_EVENT_DATA;
 	sclp_read_req.status = SCLP_REQ_QUEUED;
 	sclp_read_req.start_count = 0;
 	sclp_read_req.callback = sclp_read_cb;
@@ -334,6 +365,8 @@ sclp_interrupt_handler(__u16 code)
 	finished_sccb = S390_lowcore.ext_params & 0xfffffff8;
 	evbuf_pending = S390_lowcore.ext_params & 0x3;
 	if (finished_sccb) {
+		del_timer(&sclp_request_timer);
+		sclp_running_state = sclp_running_state_reset_pending;
 		req = __sclp_find_req(finished_sccb);
 		if (req) {
 			/* Request post-processing */
@@ -348,13 +381,8 @@ sclp_interrupt_handler(__u16 code)
 		sclp_running_state = sclp_running_state_idle;
 	}
 	if (evbuf_pending && sclp_receive_mask != 0 &&
-	    sclp_reading_state == sclp_reading_state_idle &&
-	    sclp_activation_state == sclp_activation_state_active ) {
-		sclp_reading_state = sclp_reading_state_reading;
-		__sclp_make_read_req();
-		/* Add request to head of queue */
-		list_add(&sclp_read_req.list, &sclp_req_queue);
-	}
+	    sclp_activation_state == sclp_activation_state_active)
+		__sclp_queue_read_req();
 	spin_unlock(&sclp_lock);
 	sclp_process_queue();
 }
@@ -374,6 +402,7 @@ sclp_sync_wait(void)
 	unsigned long flags;
 	unsigned long cr0, cr0_sync;
 	u64 timeout;
+	int irq_context;
 
 	/* We'll be disabling timer interrupts, so we need a custom timeout
 	 * mechanism */
@@ -386,7 +415,9 @@ sclp_sync_wait(void)
 	}
 	local_irq_save(flags);
 	/* Prevent bottom half from executing once we force interrupts open */
-	local_bh_disable();
+	irq_context = in_interrupt();
+	if (!irq_context)
+		local_bh_disable();
 	/* Enable service-signal interruption, disable timer interrupts */
 	trace_hardirqs_on();
 	__ctl_store(cr0, 0, 0);
@@ -402,19 +433,19 @@ sclp_sync_wait(void)
 		    get_clock() > timeout &&
 		    del_timer(&sclp_request_timer))
 			sclp_request_timer.function(sclp_request_timer.data);
-		barrier();
 		cpu_relax();
 	}
 	local_irq_disable();
 	__ctl_load(cr0, 0, 0);
-	_local_bh_enable();
+	if (!irq_context)
+		_local_bh_enable();
 	local_irq_restore(flags);
 }
 
 EXPORT_SYMBOL(sclp_sync_wait);
 
 /* Dispatch changes in send and receive mask to registered listeners. */
-static inline void
+static void
 sclp_dispatch_state_change(void)
 {
 	struct list_head *l;
@@ -597,7 +628,7 @@ __sclp_make_init_req(u32 receive_mask, u32 send_mask)
 	sccb = (struct init_sccb *) sclp_init_sccb;
 	clear_page(sccb);
 	memset(&sclp_init_req, 0, sizeof(struct sclp_req));
-	sclp_init_req.command = SCLP_CMDW_WRITEMASK;
+	sclp_init_req.command = SCLP_CMDW_WRITE_EVENT_MASK;
 	sclp_init_req.status = SCLP_REQ_FILLED;
 	sclp_init_req.start_count = 0;
 	sclp_init_req.callback = NULL;
@@ -800,7 +831,7 @@ sclp_check_interface(void)
 	for (retry = 0; retry <= SCLP_INIT_RETRY; retry++) {
 		__sclp_make_init_req(0, 0);
 		sccb = (struct init_sccb *) sclp_init_req.sccb;
-		rc = service_call(sclp_init_req.command, sccb);
+		rc = sclp_service_call(sclp_init_req.command, sccb);
 		if (rc == -EIO)
 			break;
 		sclp_init_req.status = SCLP_REQ_RUNNING;

drivers/s390/char/sclp.h

@@ -12,7 +12,7 @@
 
 #include <linux/types.h>
 #include <linux/list.h>
-
+#include <asm/sclp.h>
 #include <asm/ebcdic.h>
 
 /* maximum number of pages concerning our own memory management */
@@ -49,9 +49,11 @@
 
 typedef unsigned int sclp_cmdw_t;
 
-#define SCLP_CMDW_READDATA	0x00770005
-#define SCLP_CMDW_WRITEDATA	0x00760005
-#define SCLP_CMDW_WRITEMASK	0x00780005
+#define SCLP_CMDW_READ_EVENT_DATA	0x00770005
+#define SCLP_CMDW_WRITE_EVENT_DATA	0x00760005
+#define SCLP_CMDW_WRITE_EVENT_MASK	0x00780005
+#define SCLP_CMDW_READ_SCP_INFO		0x00020001
+#define SCLP_CMDW_READ_SCP_INFO_FORCED	0x00120001
 
 #define GDS_ID_MDSMU		0x1310
 #define GDS_ID_MDSRouteInfo	0x1311
@@ -66,13 +68,6 @@ typedef unsigned int sclp_cmdw_t;
 
 typedef u32 sccb_mask_t;	/* ATTENTION: assumes 32bit mask !!! */
 
-struct sccb_header {
-	u16	length;
-	u8	function_code;
-	u8	control_mask[3];
-	u16	response_code;
-} __attribute__((packed));
-
 struct gds_subvector {
 	u8	length;
 	u8	key;
@@ -131,6 +126,7 @@ void sclp_unregister(struct sclp_register *reg);
 int sclp_remove_processed(struct sccb_header *sccb);
 int sclp_deactivate(void);
 int sclp_reactivate(void);
+int sclp_service_call(sclp_cmdw_t command, void *sccb);
 
 /* useful inlines */
 

drivers/s390/char/sclp_con.c

@@ -66,7 +66,7 @@ sclp_conbuf_callback(struct sclp_buffer *buffer, int rc)
 	} while (buffer && sclp_emit_buffer(buffer, sclp_conbuf_callback));
 }
 
-static inline void
+static void
 sclp_conbuf_emit(void)
 {
 	struct sclp_buffer* buffer;

drivers/s390/char/sclp_cpi.c

@@ -169,7 +169,7 @@ cpi_prepare_req(void)
 	}
 
 	/* prepare request data structure presented to SCLP driver */
-	req->command = SCLP_CMDW_WRITEDATA;
+	req->command = SCLP_CMDW_WRITE_EVENT_DATA;
 	req->sccb = sccb;
 	req->status = SCLP_REQ_FILLED;
 	req->callback = cpi_callback;

drivers/s390/char/sclp_info.c

@@ -0,0 +1,57 @@
+/*
+ *  drivers/s390/char/sclp_info.c
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ */
+
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <asm/sclp.h>
+#include "sclp.h"
+
+struct sclp_readinfo_sccb s390_readinfo_sccb;
+
+void __init sclp_readinfo_early(void)
+{
+	sclp_cmdw_t command;
+	struct sccb_header *sccb;
+	int ret;
+
+	__ctl_set_bit(0, 9); /* enable service signal subclass mask */
+
+	sccb = &s390_readinfo_sccb.header;
+	command = SCLP_CMDW_READ_SCP_INFO_FORCED;
+	while (1) {
+		u16 response;
+
+		memset(&s390_readinfo_sccb, 0, sizeof(s390_readinfo_sccb));
+		sccb->length = sizeof(s390_readinfo_sccb);
+		sccb->control_mask[2] = 0x80;
+
+		ret = sclp_service_call(command, &s390_readinfo_sccb);
+
+		if (ret == -EIO)
+			goto out;
+		if (ret == -EBUSY)
+			continue;
+
+		__load_psw_mask(PSW_BASE_BITS | PSW_MASK_EXT |
+				PSW_MASK_WAIT | PSW_DEFAULT_KEY);
+		local_irq_disable();
+		barrier();
+
+		response = sccb->response_code;
+
+		if (response == 0x10)
+			break;
+
+		if (response != 0x1f0 || command == SCLP_CMDW_READ_SCP_INFO)
+			break;
+
+		command = SCLP_CMDW_READ_SCP_INFO;
+	}
+out:
+	__ctl_clear_bit(0, 9); /* disable service signal subclass mask */
+}

drivers/s390/char/sclp_rw.c

@@ -460,7 +460,7 @@ sclp_emit_buffer(struct sclp_buffer *buffer,
 		sccb->msg_buf.header.type = EvTyp_PMsgCmd;
 	else
 		return -ENOSYS;
-	buffer->request.command = SCLP_CMDW_WRITEDATA;
+	buffer->request.command = SCLP_CMDW_WRITE_EVENT_DATA;
 	buffer->request.status = SCLP_REQ_FILLED;
 	buffer->request.callback = sclp_writedata_callback;
 	buffer->request.callback_data = buffer;

drivers/s390/char/sclp_tty.c

@@ -721,7 +721,7 @@ static const struct tty_operations sclp_ops = {
 	.ioctl = sclp_tty_ioctl,
 };
 
-int __init
+static int __init
 sclp_tty_init(void)
 {
 	struct tty_driver *driver;