From 26916264c17b5af8a3eaaf83b128f85cf1107cff Mon Sep 17 00:00:00 2001
From: Christian Borntraeger <cborntra@de.ibm.com>
Date: Mon, 4 Dec 2006 15:39:47 +0100
Subject: [S390] remove salipl memory detection.

The SALIPL entry point has an needless memory detection routine as we
later check the memory size again. The SALIPL code also uses diagnose
0x060 if we are running under VM, but this diagnose is not compatible
with the 64 bit addressing mode. The solution is to get rid of this
code and rely on the memory detection in the startup code.

Signed-off-by: Christian Borntraeger <cborntra@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/head.S | 21 ---------------------
 1 file changed, 21 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/head.S b/arch/s390/kernel/head.S
index 0cf59bb..8f8c802 100644
--- a/arch/s390/kernel/head.S
+++ b/arch/s390/kernel/head.S
@@ -418,24 +418,6 @@ start:
 .gotr:
 	l	%r10,.tbl		# EBCDIC to ASCII table
 	tr	0(240,%r8),0(%r10)
-	stidp	__LC_CPUID		# Are we running on VM maybe
-	cli	__LC_CPUID,0xff
-	bnz	.test
-	.long	0x83300060		# diag 3,0,x'0060' - storage size
-	b	.done
-.test:
-	mvc	0x68(8),.pgmnw		# set up pgm check handler
-	l	%r2,.fourmeg
-	lr	%r3,%r2
-	bctr	%r3,%r0			# 4M-1
-.loop:	iske	%r0,%r3
-	ar	%r3,%r2
-.pgmx:
-	sr	%r3,%r2
-	la	%r3,1(%r3)
-.done:
-	l	%r1,.memsize
-	st	%r3,ARCH_OFFSET(%r1)
 	slr	%r0,%r0
 	st	%r0,INITRD_SIZE+ARCH_OFFSET-PARMAREA(%r11)
 	st	%r0,INITRD_START+ARCH_OFFSET-PARMAREA(%r11)
@@ -443,9 +425,6 @@ start:
 .tbl:	.long	_ebcasc			# translate table
 .cmd:	.long	COMMAND_LINE		# address of command line buffer
 .parm:	.long	PARMAREA
-.memsize: .long memory_size
-.fourmeg: .long 0x00400000      	# 4M
-.pgmnw:	.long	0x00080000,.pgmx
 .lowcase:
 	.byte 0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07
 	.byte 0x08,0x09,0x0a,0x0b,0x0c,0x0d,0x0e,0x0f
-- 
cgit v1.1


From 03a4d2087644f5477d9a9742e75a329f23b279e6 Mon Sep 17 00:00:00 2001
From: Michael Holzheu <holzheu@de.ibm.com>
Date: Mon, 4 Dec 2006 15:39:58 +0100
Subject: [S390] Add ipl/reipl loadparm attribute.

If multiple kernel images are installed on one DASD, the loadparm can be used
to select the boot configuration. This patch introduces the following two new
sysfs attributes:

/sys/firmware/ipl/loadparm: shows loadparm of current system (ro)
/sys/firmware/reipl/ccw/loadparm: loadparm used for next reboot (rw)

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/ipl.c | 98 +++++++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 94 insertions(+), 4 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index 1f5e782..36f0d30 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -13,12 +13,20 @@
 #include <linux/device.h>
 #include <linux/delay.h>
 #include <linux/reboot.h>
+#include <linux/ctype.h>
 #include <asm/smp.h>
 #include <asm/setup.h>
 #include <asm/cpcmd.h>
 #include <asm/cio.h>
+#include <asm/ebcdic.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])
 
 enum ipl_type {
 	IPL_TYPE_NONE	 = 1,
@@ -289,9 +297,25 @@ static struct attribute_group ipl_fcp_attr_group = {
 
 /* CCW ipl device attributes */
 
+static ssize_t ipl_ccw_loadparm_show(struct subsystem *subsys, char *page)
+{
+	char loadparm[LOADPARM_LEN + 1] = {};
+
+	if (!SCCB_VALID)
+		return sprintf(page, "#unknown#\n");
+	memcpy(loadparm, SCCB_LOADPARM, LOADPARM_LEN);
+	EBCASC(loadparm, LOADPARM_LEN);
+	strstrip(loadparm);
+	return sprintf(page, "%s\n", loadparm);
+}
+
+static struct subsys_attribute sys_ipl_ccw_loadparm_attr =
+	__ATTR(loadparm, 0444, ipl_ccw_loadparm_show, NULL);
+
 static struct attribute *ipl_ccw_attrs[] = {
 	&sys_ipl_type_attr.attr,
 	&sys_ipl_device_attr.attr,
+	&sys_ipl_ccw_loadparm_attr.attr,
 	NULL,
 };
 
@@ -348,8 +372,57 @@ static struct attribute_group reipl_fcp_attr_group = {
 DEFINE_IPL_ATTR_RW(reipl_ccw, device, "0.0.%04llx\n", "0.0.%llx\n",
 	reipl_block_ccw->ipl_info.ccw.devno);
 
+static void reipl_get_ascii_loadparm(char *loadparm)
+{
+	memcpy(loadparm, &reipl_block_ccw->ipl_info.ccw.load_param,
+	       LOADPARM_LEN);
+	EBCASC(loadparm, LOADPARM_LEN);
+	loadparm[LOADPARM_LEN] = 0;
+	strstrip(loadparm);
+}
+
+static ssize_t reipl_ccw_loadparm_show(struct subsystem *subsys, char *page)
+{
+	char buf[LOADPARM_LEN + 1];
+
+	reipl_get_ascii_loadparm(buf);
+	return sprintf(page, "%s\n", buf);
+}
+
+static ssize_t reipl_ccw_loadparm_store(struct subsystem *subsys,
+					const char *buf, size_t len)
+{
+	int i, lp_len;
+
+	/* ignore trailing newline */
+	lp_len = len;
+	if ((len > 0) && (buf[len - 1] == '\n'))
+		lp_len--;
+	/* loadparm can have max 8 characters and must not start with a blank */
+	if ((lp_len > LOADPARM_LEN) || ((lp_len > 0) && (buf[0] == ' ')))
+		return -EINVAL;
+	/* loadparm can only contain "a-z,A-Z,0-9,SP,." */
+	for (i = 0; i < lp_len; i++) {
+		if (isalpha(buf[i]) || isdigit(buf[i]) || (buf[i] == ' ') ||
+		    (buf[i] == '.'))
+			continue;
+		return -EINVAL;
+	}
+	/* initialize loadparm with blanks */
+	memset(&reipl_block_ccw->ipl_info.ccw.load_param, ' ', LOADPARM_LEN);
+	/* copy and convert to ebcdic */
+	memcpy(&reipl_block_ccw->ipl_info.ccw.load_param, buf, lp_len);
+	ASCEBC(reipl_block_ccw->ipl_info.ccw.load_param, LOADPARM_LEN);
+	return len;
+}
+
+static struct subsys_attribute sys_reipl_ccw_loadparm_attr =
+	__ATTR(loadparm, 0644, reipl_ccw_loadparm_show,
+	       reipl_ccw_loadparm_store);
+
 static struct attribute *reipl_ccw_attrs[] = {
 	&sys_reipl_ccw_device_attr.attr,
+	&sys_reipl_ccw_loadparm_attr.attr,
 	NULL,
 };
 
@@ -571,11 +644,14 @@ void do_reipl(void)
 {
 	struct ccw_dev_id devid;
 	static char buf[100];
+	char loadparm[LOADPARM_LEN + 1];
 
 	switch (reipl_type) {
 	case IPL_TYPE_CCW:
+		reipl_get_ascii_loadparm(loadparm);
 		printk(KERN_EMERG "reboot on ccw device: 0.0.%04x\n",
 			reipl_block_ccw->ipl_info.ccw.devno);
+		printk(KERN_EMERG "loadparm = '%s'\n", loadparm);
 		break;
 	case IPL_TYPE_FCP:
 		printk(KERN_EMERG "reboot on fcp device:\n");
@@ -592,7 +668,12 @@ void do_reipl(void)
 		reipl_ccw_dev(&devid);
 		break;
 	case IPL_METHOD_CCW_VM:
-		sprintf(buf, "IPL %X", reipl_block_ccw->ipl_info.ccw.devno);
+		if (strlen(loadparm) == 0)
+			sprintf(buf, "IPL %X",
+				reipl_block_ccw->ipl_info.ccw.devno);
+		else
+			sprintf(buf, "IPL %X LOADPARM '%s'",
+				reipl_block_ccw->ipl_info.ccw.devno, loadparm);
 		cpcmd(buf, NULL, 0, NULL);
 		break;
 	case IPL_METHOD_CCW_DIAG:
@@ -746,6 +827,17 @@ static int __init reipl_ccw_init(void)
 	reipl_block_ccw->hdr.version = IPL_PARM_BLOCK_VERSION;
 	reipl_block_ccw->hdr.blk0_len = sizeof(reipl_block_ccw->ipl_info.ccw);
 	reipl_block_ccw->hdr.pbt = DIAG308_IPL_TYPE_CCW;
+	/* check if read scp info worked and set loadparm */
+	if (SCCB_VALID)
+		memcpy(reipl_block_ccw->ipl_info.ccw.load_param,
+		       SCCB_LOADPARM, LOADPARM_LEN);
+	else
+		/* read scp info failed: set empty loadparm (EBCDIC blanks) */
+		memset(reipl_block_ccw->ipl_info.ccw.load_param, 0x40,
+		       LOADPARM_LEN);
+	/* FIXME: check for diag308_set_works when enabling diag ccw reipl */
+	if (!MACHINE_IS_VM)
+		sys_reipl_ccw_loadparm_attr.attr.mode = S_IRUGO;
 	if (ipl_get_type() == IPL_TYPE_CCW)
 		reipl_block_ccw->ipl_info.ccw.devno = ipl_devno;
 	reipl_capabilities |= IPL_TYPE_CCW;
@@ -827,13 +919,11 @@ static int __init dump_ccw_init(void)
 	return 0;
 }
 
-extern char s390_readinfo_sccb[];
-
 static int __init dump_fcp_init(void)
 {
 	int rc;
 
-	if(!(s390_readinfo_sccb[91] & 0x2))
+	if(!(SCCB_FLAG & 0x2) || !SCCB_VALID)
 		return 0; /* LDIPL DUMP is not installed */
 	if (!diag308_set_works)
 		return 0;
-- 
cgit v1.1


From 3902e47628dcaf79b2c7a6a59f6978d968eff288 Mon Sep 17 00:00:00 2001
From: Michael Holzheu <holzheu@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:05 +0100
Subject: [S390] No panic for failed reboot

If reboot fails (e.g. because wrong devno has been specified by the user),
we should just stop all cpus, but should not trigger a kernel panic.

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/ipl.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index 36f0d30..60ba145 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -697,7 +697,8 @@ void do_reipl(void)
 		diag308(DIAG308_IPL, NULL);
 		break;
 	}
-	panic("reipl failed!\n");
+	printk(KERN_EMERG "reboot failed!\n");
+	signal_processor(smp_processor_id(), sigp_stop_and_store_status);
 }
 
 static void do_dump(void)
-- 
cgit v1.1


From 5986b0e845bc52a53ff3cafd74d341e81c95658d Mon Sep 17 00:00:00 2001
From: Michael Holzheu <holzheu@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:13 +0100
Subject: [S390] Use diag instead of ccw reipl.

Since the diag 308 reipl method is superior to the ccw method, we should
use it whenever it is possible. We can do that, if the user has not
specified a new reipl ccw device and the system has been ipled from
a ccw device.

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/ipl.c | 2 ++
 1 file changed, 2 insertions(+)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index 60ba145..ec12782 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -664,6 +664,8 @@ void do_reipl(void)
 	switch (reipl_method) {
 	case IPL_METHOD_CCW_CIO:
 		devid.devno = reipl_block_ccw->ipl_info.ccw.devno;
+		if (ipl_get_type() == IPL_TYPE_CCW && devid.devno == ipl_devno)
+			diag308(DIAG308_IPL, NULL);
 		devid.ssid  = 0;
 		reipl_ccw_dev(&devid);
 		break;
-- 
cgit v1.1


From 15e9b586e0bd3692e2a21c5be178810d9d32214e Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:26 +0100
Subject: [S390] Reset infrastructure for re-IPL.

In case of re-IPL and diag308 doesn't work we have to reset all devices
manually and wait synchronously that each reset finished.
This patch adds the necessary infrastucture and the first exploiter of it.

Subsystems that need to add a function that needs to be called at re-IPL
may register/unregister this function via

struct reset_call {
	struct reset_call *next;
	void (*fn)(void);
};

void register_reset_call(struct reset_call *reset);
void unregister_reset_call(struct reset_call *reset);

When the registered function get called the context is:

- all cpus beside the current one are stopped
- all machine checks and interrupts are disabled
- prefixing is disabled
- a default machine check handler is available for use

The registered functions may not take any locks are sleep.

For the common I/O layer part of this patch:

Introduce a reset_call css_reset that does the following:
- clear all subchannels
- perform a rchp on all channel paths and wait for the resulting
  machine checks
This replaces the calls to clear_all_subchannels() and
cio_reset_channel_paths() for kexec and ccw reipl. reipl_ccw_dev() now
uses reipl_find_schid() to determine the subchannel id for a given
device id.
Also remove cio_reset_channel_paths() and friends since they are not
needed anymore.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/Makefile            |  2 +-
 arch/s390/kernel/ipl.c               | 54 ++++++++++++++++++++++++++++++++++++
 arch/s390/kernel/machine_kexec.c     |  9 ++----
 arch/s390/kernel/reipl.S             | 17 ++----------
 arch/s390/kernel/reipl64.S           | 16 ++++-------
 arch/s390/kernel/relocate_kernel.S   |  5 +---
 arch/s390/kernel/relocate_kernel64.S |  5 +---
 arch/s390/kernel/reset.S             | 48 ++++++++++++++++++++++++++++++++
 8 files changed, 115 insertions(+), 41 deletions(-)
 create mode 100644 arch/s390/kernel/reset.S

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile
index aa97897..a81881c 100644
--- a/arch/s390/kernel/Makefile
+++ b/arch/s390/kernel/Makefile
@@ -4,7 +4,7 @@
 
 EXTRA_AFLAGS	:= -traditional
 
-obj-y	:=  bitmap.o traps.o time.o process.o \
+obj-y	:=  bitmap.o traps.o time.o process.o reset.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
 
diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index ec12782..77f4aff 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -19,6 +19,7 @@
 #include <asm/cpcmd.h>
 #include <asm/cio.h>
 #include <asm/ebcdic.h>
+#include <asm/reset.h>
 
 #define IPL_PARM_BLOCK_VERSION 0
 #define LOADPARM_LEN 8
@@ -1024,3 +1025,56 @@ static int __init s390_ipl_init(void)
 }
 
 __initcall(s390_ipl_init);
+
+static LIST_HEAD(rcall);
+static DEFINE_MUTEX(rcall_mutex);
+
+void register_reset_call(struct reset_call *reset)
+{
+	mutex_lock(&rcall_mutex);
+	list_add(&reset->list, &rcall);
+	mutex_unlock(&rcall_mutex);
+}
+EXPORT_SYMBOL_GPL(register_reset_call);
+
+void unregister_reset_call(struct reset_call *reset)
+{
+	mutex_lock(&rcall_mutex);
+	list_del(&reset->list);
+	mutex_unlock(&rcall_mutex);
+}
+EXPORT_SYMBOL_GPL(unregister_reset_call);
+
+static void do_reset_calls(void)
+{
+	struct reset_call *reset;
+
+	list_for_each_entry(reset, &rcall, list)
+		reset->fn();
+}
+
+extern void reset_mcck_handler(void);
+
+void s390_reset_system(void)
+{
+	struct _lowcore *lc;
+
+	/* Disable all interrupts/machine checks */
+	__load_psw_mask(PSW_KERNEL_BITS & ~PSW_MASK_MCHECK);
+
+	/* Stack for interrupt/machine check handler */
+	lc = (struct _lowcore *)(unsigned long) store_prefix();
+	lc->panic_stack = S390_lowcore.panic_stack;
+
+	/* Disable prefixing */
+	set_prefix(0);
+
+	/* Disable lowcore protection */
+	__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.addr =
+		PSW_ADDR_AMODE | (unsigned long) &reset_mcck_handler;
+	do_reset_calls();
+}
diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c
index 60b1ea9..202bf1fd 100644
--- a/arch/s390/kernel/machine_kexec.c
+++ b/arch/s390/kernel/machine_kexec.c
@@ -22,6 +22,7 @@
 #include <asm/pgalloc.h>
 #include <asm/system.h>
 #include <asm/smp.h>
+#include <asm/reset.h>
 
 static void kexec_halt_all_cpus(void *);
 
@@ -62,12 +63,6 @@ machine_shutdown(void)
 NORET_TYPE void
 machine_kexec(struct kimage *image)
 {
-	clear_all_subchannels();
-	cio_reset_channel_paths();
-
-	/* Disable lowcore protection */
-	ctl_clear_bit(0,28);
-
 	on_each_cpu(kexec_halt_all_cpus, image, 0, 0);
 	for (;;);
 }
@@ -98,6 +93,8 @@ kexec_halt_all_cpus(void *kernel_image)
 			cpu_relax();
 	}
 
+	s390_reset_system();
+
 	image = (struct kimage *) kernel_image;
 	data_mover = (relocate_kernel_t)
 		(page_to_pfn(image->control_code_page) << PAGE_SHIFT);
diff --git a/arch/s390/kernel/reipl.S b/arch/s390/kernel/reipl.S
index 0340477..f9434d4 100644
--- a/arch/s390/kernel/reipl.S
+++ b/arch/s390/kernel/reipl.S
@@ -11,19 +11,10 @@
 		.globl	do_reipl_asm
 do_reipl_asm:	basr	%r13,0
 .Lpg0:		lpsw	.Lnewpsw-.Lpg0(%r13)
-
-		# switch off lowcore protection
-
-.Lpg1:		stctl	%c0,%c0,.Lctlsave1-.Lpg0(%r13)
-		stctl	%c0,%c0,.Lctlsave2-.Lpg0(%r13)
-		ni	.Lctlsave1-.Lpg0(%r13),0xef
-		lctl	%c0,%c0,.Lctlsave1-.Lpg0(%r13)
-
-		# do store status of all registers
+.Lpg1:		# do store status of all registers
 
 		stm	%r0,%r15,__LC_GPREGS_SAVE_AREA
 		stctl	%c0,%c15,__LC_CREGS_SAVE_AREA
-		mvc	__LC_CREGS_SAVE_AREA(4),.Lctlsave2-.Lpg0(%r13)
 		stam	%a0,%a15,__LC_AREGS_SAVE_AREA
 		stpx	__LC_PREFIX_SAVE_AREA
 		stckc	.Lclkcmp-.Lpg0(%r13)
@@ -56,8 +47,7 @@ do_reipl_asm:	basr	%r13,0
 .L002:		tm	.Liplirb+8-.Lpg0(%r13),0xf3
 		jz	.L003
 		bas	%r14,.Ldisab-.Lpg0(%r13)
-.L003:		spx	.Lnull-.Lpg0(%r13)
-		st	%r1,__LC_SUBCHANNEL_ID
+.L003:		st	%r1,__LC_SUBCHANNEL_ID
 		lpsw	0
 		sigp	0,0,0(6)
 .Ldisab:	st	%r14,.Ldispsw+4-.Lpg0(%r13)
@@ -65,9 +55,6 @@ do_reipl_asm:	basr	%r13,0
 		.align	8
 .Lclkcmp:	.quad	0x0000000000000000
 .Lall:		.long	0xff000000
-.Lnull:		.long	0x00000000
-.Lctlsave1:	.long	0x00000000
-.Lctlsave2:	.long	0x00000000
 		.align	8
 .Lnewpsw:	.long	0x00080000,0x80000000+.Lpg1
 .Lpcnew:	.long	0x00080000,0x80000000+.Lecs
diff --git a/arch/s390/kernel/reipl64.S b/arch/s390/kernel/reipl64.S
index de74350..f18ef26 100644
--- a/arch/s390/kernel/reipl64.S
+++ b/arch/s390/kernel/reipl64.S
@@ -10,10 +10,10 @@
 #include <asm/lowcore.h>
 		.globl	do_reipl_asm
 do_reipl_asm:	basr	%r13,0
+.Lpg0:		lpswe	.Lnewpsw-.Lpg0(%r13)
+.Lpg1:		# do store status of all registers
 
-		# do store status of all registers
-
-.Lpg0:		stg	%r1,.Lregsave-.Lpg0(%r13)
+		stg	%r1,.Lregsave-.Lpg0(%r13)
 		lghi	%r1,0x1000
 		stmg	%r0,%r15,__LC_GPREGS_SAVE_AREA-0x1000(%r1)
 		lg	%r0,.Lregsave-.Lpg0(%r13)
@@ -27,11 +27,7 @@ do_reipl_asm:	basr	%r13,0
 		stpt	__LC_CPU_TIMER_SAVE_AREA-0x1000(%r1)
 		stg	%r13, __LC_PSW_SAVE_AREA-0x1000+8(%r1)
 
-		lpswe	.Lnewpsw-.Lpg0(%r13)
-.Lpg1:		lctlg	%c6,%c6,.Lall-.Lpg0(%r13)
-		stctg	%c0,%c0,.Lregsave-.Lpg0(%r13)
-		ni	.Lregsave+4-.Lpg0(%r13),0xef
-		lctlg	%c0,%c0,.Lregsave-.Lpg0(%r13)
+		lctlg	%c6,%c6,.Lall-.Lpg0(%r13)
 		lgr	%r1,%r2
 		mvc	__LC_PGM_NEW_PSW(16),.Lpcnew-.Lpg0(%r13)
 		stsch	.Lschib-.Lpg0(%r13)
@@ -56,8 +52,7 @@ do_reipl_asm:	basr	%r13,0
 .L002:		tm	.Liplirb+8-.Lpg0(%r13),0xf3
 		jz	.L003
 		bas	%r14,.Ldisab-.Lpg0(%r13)
-.L003:		spx	.Lnull-.Lpg0(%r13)
-		st	%r1,__LC_SUBCHANNEL_ID
+.L003:		st	%r1,__LC_SUBCHANNEL_ID
 		lhi	%r1,0		 # mode 0 = esa
 		slr	%r0,%r0 	 # set cpuid to zero
 		sigp	%r1,%r0,0x12	 # switch to esa mode
@@ -70,7 +65,6 @@ do_reipl_asm:	basr	%r13,0
 .Lclkcmp:	.quad	0x0000000000000000
 .Lall:		.quad	0x00000000ff000000
 .Lregsave:	.quad	0x0000000000000000
-.Lnull:		.long	0x0000000000000000
 		.align	16
 /*
  * These addresses have to be 31 bit otherwise
diff --git a/arch/s390/kernel/relocate_kernel.S b/arch/s390/kernel/relocate_kernel.S
index f9899ff..3b456b8 100644
--- a/arch/s390/kernel/relocate_kernel.S
+++ b/arch/s390/kernel/relocate_kernel.S
@@ -26,8 +26,7 @@
 	relocate_kernel:
 		basr	%r13,0		# base address
 	.base:
-		stnsm	sys_msk-.base(%r13),0xf8	# disable DAT and IRQ (external)
-		spx	zero64-.base(%r13)	# absolute addressing mode
+		stnsm	sys_msk-.base(%r13),0xfb	# disable DAT
 		stctl	%c0,%c15,ctlregs-.base(%r13)
 		stm	%r0,%r15,gprregs-.base(%r13)
 		la	%r1,load_psw-.base(%r13)
@@ -97,8 +96,6 @@
 		lpsw	0		# hopefully start new kernel...
 
 		.align	8
-	zero64:
-		.quad	0
 	load_psw:
 		.long	0x00080000,0x80000000
 	sys_msk:
diff --git a/arch/s390/kernel/relocate_kernel64.S b/arch/s390/kernel/relocate_kernel64.S
index 4fb4430..1f9ea20 100644
--- a/arch/s390/kernel/relocate_kernel64.S
+++ b/arch/s390/kernel/relocate_kernel64.S
@@ -27,8 +27,7 @@
 	relocate_kernel:
 		basr	%r13,0		# base address
 	.base:
-		stnsm	sys_msk-.base(%r13),0xf8	# disable DAT and IRQs
-		spx	zero64-.base(%r13)	# absolute addressing mode
+		stnsm	sys_msk-.base(%r13),0xfb	# disable DAT
 		stctg	%c0,%c15,ctlregs-.base(%r13)
 		stmg	%r0,%r15,gprregs-.base(%r13)
 		lghi	%r0,3
@@ -100,8 +99,6 @@
 		lpsw	0		# hopefully start new kernel...
 
 		.align	8
-	zero64:
-		.quad	0
 	load_psw:
 		.long	0x00080000,0x80000000
 	sys_msk:
diff --git a/arch/s390/kernel/reset.S b/arch/s390/kernel/reset.S
new file mode 100644
index 0000000..be8688c
--- /dev/null
+++ b/arch/s390/kernel/reset.S
@@ -0,0 +1,48 @@
+/*
+ *  arch/s390/kernel/reset.S
+ *
+ *    Copyright (C) IBM Corp. 2006
+ *    Author(s): Heiko Carstens <heiko.carstens@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
+
+#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
+
+#endif /* CONFIG_64BIT */
-- 
cgit v1.1


From 740b5706b9c4b3767f597b3ea76654c6f2a800b2 Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:30 +0100
Subject: [S390] cpcmd <-> __cpcmd calling issues

In case of reipl cpcmd gets called when all other cpus are not running
anymore. To prevent deadlocks change __cpcmd so that it doesn't take
any locks and call cpcmd or __cpcmd, whatever is correct in the current
context.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/cpcmd.c | 18 +++++++++++-------
 arch/s390/kernel/ipl.c   | 10 +++++-----
 arch/s390/kernel/setup.c | 10 +++++-----
 3 files changed, 21 insertions(+), 17 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/cpcmd.c b/arch/s390/kernel/cpcmd.c
index 1eae74e..a5972f1 100644
--- a/arch/s390/kernel/cpcmd.c
+++ b/arch/s390/kernel/cpcmd.c
@@ -21,14 +21,15 @@ static DEFINE_SPINLOCK(cpcmd_lock);
 static char cpcmd_buf[241];
 
 /*
- * the caller of __cpcmd has to ensure that the response buffer is below 2 GB
+ * __cpcmd has some restrictions over cpcmd
+ *  - the response buffer must reside below 2GB (if any)
+ *  - __cpcmd is unlocked and therefore not SMP-safe
  */
 int  __cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 {
-	unsigned long flags, cmdlen;
+	unsigned cmdlen;
 	int return_code, return_len;
 
-	spin_lock_irqsave(&cpcmd_lock, flags);
 	cmdlen = strlen(cmd);
 	BUG_ON(cmdlen > 240);
 	memcpy(cpcmd_buf, cmd, cmdlen);
@@ -74,7 +75,6 @@ int  __cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 			: "+d" (reg3) : "d" (reg2) : "cc");
 		return_code = (int) reg3;
         }
-	spin_unlock_irqrestore(&cpcmd_lock, flags);
 	if (response_code != NULL)
 		*response_code = return_code;
 	return return_len;
@@ -82,15 +82,18 @@ int  __cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 
 EXPORT_SYMBOL(__cpcmd);
 
-#ifdef CONFIG_64BIT
 int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 {
 	char *lowbuf;
 	int len;
+	unsigned long flags;
 
 	if ((rlen == 0) || (response == NULL)
-	    || !((unsigned long)response >> 31))
+	    || !((unsigned long)response >> 31)) {
+		spin_lock_irqsave(&cpcmd_lock, flags);
 		len = __cpcmd(cmd, response, rlen, response_code);
+		spin_unlock_irqrestore(&cpcmd_lock, flags);
+	}
 	else {
 		lowbuf = kmalloc(rlen, GFP_KERNEL | GFP_DMA);
 		if (!lowbuf) {
@@ -98,7 +101,9 @@ int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 				"cpcmd: could not allocate response buffer\n");
 			return -ENOMEM;
 		}
+		spin_lock_irqsave(&cpcmd_lock, flags);
 		len = __cpcmd(cmd, lowbuf, rlen, response_code);
+		spin_unlock_irqrestore(&cpcmd_lock, flags);
 		memcpy(response, lowbuf, rlen);
 		kfree(lowbuf);
 	}
@@ -106,4 +111,3 @@ int cpcmd(const char *cmd, char *response, int rlen, int *response_code)
 }
 
 EXPORT_SYMBOL(cpcmd);
-#endif		/* CONFIG_64BIT */
diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index 77f4aff..101b003 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -677,7 +677,7 @@ void do_reipl(void)
 		else
 			sprintf(buf, "IPL %X LOADPARM '%s'",
 				reipl_block_ccw->ipl_info.ccw.devno, loadparm);
-		cpcmd(buf, NULL, 0, NULL);
+		__cpcmd(buf, NULL, 0, NULL);
 		break;
 	case IPL_METHOD_CCW_DIAG:
 		diag308(DIAG308_SET, reipl_block_ccw);
@@ -691,12 +691,12 @@ void do_reipl(void)
 		diag308(DIAG308_IPL, NULL);
 		break;
 	case IPL_METHOD_FCP_RO_VM:
-		cpcmd("IPL", NULL, 0, NULL);
+		__cpcmd("IPL", NULL, 0, NULL);
 		break;
 	case IPL_METHOD_NONE:
 	default:
 		if (MACHINE_IS_VM)
-			cpcmd("IPL", NULL, 0, NULL);
+			__cpcmd("IPL", NULL, 0, NULL);
 		diag308(DIAG308_IPL, NULL);
 		break;
 	}
@@ -732,9 +732,9 @@ static void do_dump(void)
 	case IPL_METHOD_CCW_VM:
 		dump_smp_stop_all();
 		sprintf(buf, "STORE STATUS");
-		cpcmd(buf, NULL, 0, NULL);
+		__cpcmd(buf, NULL, 0, NULL);
 		sprintf(buf, "IPL %X", dump_block_ccw->ipl_info.ccw.devno);
-		cpcmd(buf, NULL, 0, NULL);
+		__cpcmd(buf, NULL, 0, NULL);
 		break;
 	case IPL_METHOD_CCW_DIAG:
 		diag308(DIAG308_SET, dump_block_ccw);
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index 2aa13e8..9bbef0c 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -229,11 +229,11 @@ static void __init conmode_default(void)
 	char *ptr;
 
         if (MACHINE_IS_VM) {
-		__cpcmd("QUERY CONSOLE", query_buffer, 1024, NULL);
+		cpcmd("QUERY CONSOLE", query_buffer, 1024, NULL);
 		console_devno = simple_strtoul(query_buffer + 5, NULL, 16);
 		ptr = strstr(query_buffer, "SUBCHANNEL =");
 		console_irq = simple_strtoul(ptr + 13, NULL, 16);
-		__cpcmd("QUERY TERM", query_buffer, 1024, NULL);
+		cpcmd("QUERY TERM", query_buffer, 1024, NULL);
 		ptr = strstr(query_buffer, "CONMODE");
 		/*
 		 * Set the conmode to 3215 so that the device recognition 
@@ -242,7 +242,7 @@ static void __init conmode_default(void)
 		 * 3215 and the 3270 driver will try to access the console
 		 * device (3215 as console and 3270 as normal tty).
 		 */
-		__cpcmd("TERM CONMODE 3215", NULL, 0, NULL);
+		cpcmd("TERM CONMODE 3215", NULL, 0, NULL);
 		if (ptr == NULL) {
 #if defined(CONFIG_SCLP_CONSOLE)
 			SET_CONSOLE_SCLP;
@@ -299,14 +299,14 @@ static void do_machine_restart_nonsmp(char * __unused)
 static void do_machine_halt_nonsmp(void)
 {
         if (MACHINE_IS_VM && strlen(vmhalt_cmd) > 0)
-                cpcmd(vmhalt_cmd, NULL, 0, NULL);
+		__cpcmd(vmhalt_cmd, NULL, 0, NULL);
         signal_processor(smp_processor_id(), sigp_stop_and_store_status);
 }
 
 static void do_machine_power_off_nonsmp(void)
 {
         if (MACHINE_IS_VM && strlen(vmpoff_cmd) > 0)
-                cpcmd(vmpoff_cmd, NULL, 0, NULL);
+		__cpcmd(vmpoff_cmd, NULL, 0, NULL);
         signal_processor(smp_processor_id(), sigp_stop_and_store_status);
 }
 
-- 
cgit v1.1


From c6b5b847a7cf11f131c43fe0041443ec11697fc7 Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:33 +0100
Subject: [S390] cpu shutdown rework

Let one master cpu kill all other cpus instead of sending an external
interrupt to all other cpus so they can kill themselves.
Simplifies reipl/shutdown functions a lot.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/ipl.c           |  24 +--------
 arch/s390/kernel/machine_kexec.c |  65 ++++++------------------
 arch/s390/kernel/smp.c           | 104 ++++++++++++---------------------------
 3 files changed, 49 insertions(+), 144 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index 101b003..a36bea1 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -576,23 +576,6 @@ static struct subsys_attribute dump_type_attr =
 
 static decl_subsys(dump, NULL, NULL);
 
-#ifdef CONFIG_SMP
-static void dump_smp_stop_all(void)
-{
-	int cpu;
-	preempt_disable();
-	for_each_online_cpu(cpu) {
-		if (cpu == smp_processor_id())
-			continue;
-		while (signal_processor(cpu, sigp_stop) == sigp_busy)
-			udelay(10);
-	}
-	preempt_enable();
-}
-#else
-#define dump_smp_stop_all() do { } while (0)
-#endif
-
 /*
  * Shutdown actions section
  */
@@ -724,13 +707,13 @@ static void do_dump(void)
 
 	switch (dump_method) {
 	case IPL_METHOD_CCW_CIO:
-		dump_smp_stop_all();
+		smp_send_stop();
 		devid.devno = dump_block_ccw->ipl_info.ccw.devno;
 		devid.ssid  = 0;
 		reipl_ccw_dev(&devid);
 		break;
 	case IPL_METHOD_CCW_VM:
-		dump_smp_stop_all();
+		smp_send_stop();
 		sprintf(buf, "STORE STATUS");
 		__cpcmd(buf, NULL, 0, NULL);
 		sprintf(buf, "IPL %X", dump_block_ccw->ipl_info.ccw.devno);
@@ -1059,9 +1042,6 @@ void s390_reset_system(void)
 {
 	struct _lowcore *lc;
 
-	/* Disable all interrupts/machine checks */
-	__load_psw_mask(PSW_KERNEL_BITS & ~PSW_MASK_MCHECK);
-
 	/* Stack for interrupt/machine check handler */
 	lc = (struct _lowcore *)(unsigned long) store_prefix();
 	lc->panic_stack = S390_lowcore.panic_stack;
diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c
index 202bf1fd..def5caf 100644
--- a/arch/s390/kernel/machine_kexec.c
+++ b/arch/s390/kernel/machine_kexec.c
@@ -1,15 +1,10 @@
 /*
  * arch/s390/kernel/machine_kexec.c
  *
- * (C) Copyright IBM Corp. 2005
+ * Copyright IBM Corp. 2005,2006
  *
- * Author(s): Rolf Adelsberger <adelsberger@de.ibm.com>
- *
- */
-
-/*
- * s390_machine_kexec.c - handle the transition of Linux booting another kernel
- * on the S390 architecture.
+ * Author(s): Rolf Adelsberger,
+ *	      Heiko Carstens <heiko.carstens@de.ibm.com>
  */
 
 #include <linux/device.h>
@@ -24,81 +19,53 @@
 #include <asm/smp.h>
 #include <asm/reset.h>
 
-static void kexec_halt_all_cpus(void *);
-
-typedef void (*relocate_kernel_t) (kimage_entry_t *, unsigned long);
+typedef void (*relocate_kernel_t)(kimage_entry_t *, unsigned long);
 
 extern const unsigned char relocate_kernel[];
 extern const unsigned long long relocate_kernel_len;
 
-int
-machine_kexec_prepare(struct kimage *image)
+int machine_kexec_prepare(struct kimage *image)
 {
-	unsigned long reboot_code_buffer;
+	void *reboot_code_buffer;
 
 	/* We don't support anything but the default image type for now. */
 	if (image->type != KEXEC_TYPE_DEFAULT)
 		return -EINVAL;
 
 	/* Get the destination where the assembler code should be copied to.*/
-	reboot_code_buffer = page_to_pfn(image->control_code_page)<<PAGE_SHIFT;
+	reboot_code_buffer = (void *) page_to_phys(image->control_code_page);
 
 	/* Then copy it */
-	memcpy((void *) reboot_code_buffer, relocate_kernel,
-	       relocate_kernel_len);
+	memcpy(reboot_code_buffer, relocate_kernel, relocate_kernel_len);
 	return 0;
 }
 
-void
-machine_kexec_cleanup(struct kimage *image)
+void machine_kexec_cleanup(struct kimage *image)
 {
 }
 
-void
-machine_shutdown(void)
+void machine_shutdown(void)
 {
 	printk(KERN_INFO "kexec: machine_shutdown called\n");
 }
 
-NORET_TYPE void
-machine_kexec(struct kimage *image)
-{
-	on_each_cpu(kexec_halt_all_cpus, image, 0, 0);
-	for (;;);
-}
-
 extern void pfault_fini(void);
 
-static void
-kexec_halt_all_cpus(void *kernel_image)
+void machine_kexec(struct kimage *image)
 {
-	static atomic_t cpuid = ATOMIC_INIT(-1);
-	int cpu;
-	struct kimage *image;
 	relocate_kernel_t data_mover;
 
+	preempt_disable();
 #ifdef CONFIG_PFAULT
 	if (MACHINE_IS_VM)
 		pfault_fini();
 #endif
-
-	if (atomic_cmpxchg(&cpuid, -1, smp_processor_id()) != -1)
-		signal_processor(smp_processor_id(), sigp_stop);
-
-	/* Wait for all other cpus to enter stopped state */
-	for_each_online_cpu(cpu) {
-		if (cpu == smp_processor_id())
-			continue;
-		while (!smp_cpu_not_running(cpu))
-			cpu_relax();
-	}
-
+	smp_send_stop();
 	s390_reset_system();
 
-	image = (struct kimage *) kernel_image;
-	data_mover = (relocate_kernel_t)
-		(page_to_pfn(image->control_code_page) << PAGE_SHIFT);
+	data_mover = (relocate_kernel_t) page_to_phys(image->control_code_page);
 
 	/* Call the moving routine */
-	(*data_mover) (&image->head, image->start);
+	(*data_mover)(&image->head, image->start);
+	for (;;);
 }
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 6282224..b549a43 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -230,18 +230,37 @@ static inline void do_store_status(void)
         }
 }
 
+static inline void do_wait_for_stop(void)
+{
+	int cpu;
+
+	/* Wait for all other cpus to enter stopped state */
+	for_each_online_cpu(cpu) {
+		if (cpu == smp_processor_id())
+			continue;
+		while(!smp_cpu_not_running(cpu))
+			cpu_relax();
+	}
+}
+
 /*
  * this function sends a 'stop' sigp to all other CPUs in the system.
  * it goes straight through.
  */
 void smp_send_stop(void)
 {
+	/* Disable all interrupts/machine checks */
+	__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;
 
 	/* stop other processors. */
 	do_send_stop();
 
+	/* wait until other processors are stopped */
+	do_wait_for_stop();
+
 	/* store status of other processors. */
 	do_store_status();
 }
@@ -250,88 +269,28 @@ void smp_send_stop(void)
  * Reboot, halt and power_off routines for SMP.
  */
 
-static void do_machine_restart(void * __unused)
-{
-	int cpu;
-	static atomic_t cpuid = ATOMIC_INIT(-1);
-
-	if (atomic_cmpxchg(&cpuid, -1, smp_processor_id()) != -1)
-		signal_processor(smp_processor_id(), sigp_stop);
-
-	/* Wait for all other cpus to enter stopped state */
-	for_each_online_cpu(cpu) {
-		if (cpu == smp_processor_id())
-			continue;
-		while(!smp_cpu_not_running(cpu))
-			cpu_relax();
-	}
-
-	/* Store status of other cpus. */
-	do_store_status();
-
-	/*
-	 * Finally call reipl. Because we waited for all other
-	 * cpus to enter this function we know that they do
-	 * not hold any s390irq-locks (the cpus have been
-	 * interrupted by an external interrupt and s390irq
-	 * locks are always held disabled).
-	 */
-	do_reipl();
-}
-
 void machine_restart_smp(char * __unused) 
 {
-        on_each_cpu(do_machine_restart, NULL, 0, 0);
-}
-
-static void do_wait_for_stop(void)
-{
-	unsigned long cr[16];
-
-	__ctl_store(cr, 0, 15);
-	cr[0] &= ~0xffff;
-	cr[6] = 0;
-	__ctl_load(cr, 0, 15);
-	for (;;)
-		enabled_wait();
-}
-
-static void do_machine_halt(void * __unused)
-{
-	static atomic_t cpuid = ATOMIC_INIT(-1);
-
-	if (atomic_cmpxchg(&cpuid, -1, smp_processor_id()) == -1) {
-		smp_send_stop();
-		if (MACHINE_IS_VM && strlen(vmhalt_cmd) > 0)
-			cpcmd(vmhalt_cmd, NULL, 0, NULL);
-		signal_processor(smp_processor_id(),
-				 sigp_stop_and_store_status);
-	}
-	do_wait_for_stop();
+	smp_send_stop();
+	do_reipl();
 }
 
 void machine_halt_smp(void)
 {
-        on_each_cpu(do_machine_halt, NULL, 0, 0);
-}
-
-static void do_machine_power_off(void * __unused)
-{
-	static atomic_t cpuid = ATOMIC_INIT(-1);
-
-	if (atomic_cmpxchg(&cpuid, -1, smp_processor_id()) == -1) {
-		smp_send_stop();
-		if (MACHINE_IS_VM && strlen(vmpoff_cmd) > 0)
-			cpcmd(vmpoff_cmd, NULL, 0, NULL);
-		signal_processor(smp_processor_id(),
-				 sigp_stop_and_store_status);
-	}
-	do_wait_for_stop();
+	smp_send_stop();
+	if (MACHINE_IS_VM && strlen(vmhalt_cmd) > 0)
+		__cpcmd(vmhalt_cmd, NULL, 0, NULL);
+	signal_processor(smp_processor_id(), sigp_stop_and_store_status);
+	for (;;);
 }
 
 void machine_power_off_smp(void)
 {
-        on_each_cpu(do_machine_power_off, NULL, 0, 0);
+	smp_send_stop();
+	if (MACHINE_IS_VM && strlen(vmpoff_cmd) > 0)
+		__cpcmd(vmpoff_cmd, NULL, 0, NULL);
+	signal_processor(smp_processor_id(), sigp_stop_and_store_status);
+	for (;;);
 }
 
 /*
@@ -860,4 +819,3 @@ EXPORT_SYMBOL(smp_ctl_clear_bit);
 EXPORT_SYMBOL(smp_call_function);
 EXPORT_SYMBOL(smp_get_cpu);
 EXPORT_SYMBOL(smp_put_cpu);
-
-- 
cgit v1.1


From baf2aeb3d9e286add823bcaea5442ad4ee34f6e4 Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:36 +0100
Subject: [S390] Misaligned wait PSW at memory detection.

If the memory detection code would ever reach the point where it would
load the wait psw, it would generate a specification exception and the
system would crash at ipl time. This is because of a misaligned wait
psw. It needs to be on a double word boundary instead of a word
boundary.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/head31.S | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/head31.S b/arch/s390/kernel/head31.S
index 0a2c929..adb082a 100644
--- a/arch/s390/kernel/head31.S
+++ b/arch/s390/kernel/head31.S
@@ -131,10 +131,11 @@ startup_continue:
 	.long	init_thread_union
 .Lpmask:
 	.byte	0
-.align 8
+	.align	8
 .Lpcext:.long	0x00080000,0x80000000
 .Lcr:
 	.long	0x00			# place holder for cr0
+	.align	8
 .Lwaitsclp:
 	.long 0x010a0000,0x80000000 + .Lsclph
 .Lrcp:
-- 
cgit v1.1


From 36a2bd425d9b3ba2a40b0653e08d17702c78558e Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:38 +0100
Subject: [S390] Cleanup memory_chunk array usage.

Need this at yet another file and don't want to add yet another
extern...

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/head31.S | 2 +-
 arch/s390/kernel/setup.c  | 7 ++-----
 2 files changed, 3 insertions(+), 6 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/head31.S b/arch/s390/kernel/head31.S
index adb082a..9817c32 100644
--- a/arch/s390/kernel/head31.S
+++ b/arch/s390/kernel/head31.S
@@ -157,7 +157,7 @@ startup_continue:
 	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
+	la	%r10,MEMORY_CHUNKS	# number of chunks
 .Lloop:
 	tprot	0(%r5),0		# test protection of first byte
 	ipm	%r7
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index 9bbef0c..b1a8ad9 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -64,11 +64,8 @@ unsigned int console_devno = -1;
 unsigned int console_irq = -1;
 unsigned long memory_size = 0;
 unsigned long machine_flags = 0;
-struct {
-	unsigned long addr, size, type;
-} memory_chunk[MEMORY_CHUNKS] = { { 0 } };
-#define CHUNK_READ_WRITE 0
-#define CHUNK_READ_ONLY 1
+
+struct mem_chunk memory_chunk[MEMORY_CHUNKS];
 volatile int __cpu_logical_map[NR_CPUS]; /* logical cpu to cpu address */
 unsigned long __initdata zholes_size[MAX_NR_ZONES];
 static unsigned long __initdata memory_end;
-- 
cgit v1.1


From 29b08d2bae854f66d3cfd5f57aaf2e7c2c7fce32 Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:40 +0100
Subject: [S390] pfault code cleanup.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/machine_kexec.c |  8 +-------
 arch/s390/kernel/smp.c           | 13 +++----------
 arch/s390/kernel/traps.c         | 25 +------------------------
 3 files changed, 5 insertions(+), 41 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c
index def5caf..f6d9bcc 100644
--- a/arch/s390/kernel/machine_kexec.c
+++ b/arch/s390/kernel/machine_kexec.c
@@ -49,18 +49,12 @@ void machine_shutdown(void)
 	printk(KERN_INFO "kexec: machine_shutdown called\n");
 }
 
-extern void pfault_fini(void);
-
 void machine_kexec(struct kimage *image)
 {
 	relocate_kernel_t data_mover;
 
-	preempt_disable();
-#ifdef CONFIG_PFAULT
-	if (MACHINE_IS_VM)
-		pfault_fini();
-#endif
 	smp_send_stop();
+	pfault_fini();
 	s390_reset_system();
 
 	data_mover = (relocate_kernel_t) page_to_phys(image->control_code_page);
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index b549a43..19090f7 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -460,8 +460,6 @@ __init smp_count_cpus(void)
  */
 extern void init_cpu_timer(void);
 extern void init_cpu_vtimer(void);
-extern int pfault_init(void);
-extern void pfault_fini(void);
 
 int __devinit start_secondary(void *cpuvoid)
 {
@@ -473,11 +471,9 @@ int __devinit start_secondary(void *cpuvoid)
 #ifdef CONFIG_VIRT_TIMER
         init_cpu_vtimer();
 #endif
-#ifdef CONFIG_PFAULT
 	/* Enable pfault pseudo page faults on this cpu. */
-	if (MACHINE_IS_VM)
-		pfault_init();
-#endif
+	pfault_init();
+
 	/* Mark this cpu as online */
 	cpu_set(smp_processor_id(), cpu_online_map);
 	/* Switch on interrupts */
@@ -667,11 +663,8 @@ __cpu_disable(void)
 	}
 	cpu_clear(cpu, cpu_online_map);
 
-#ifdef CONFIG_PFAULT
 	/* Disable pfault pseudo page faults on this cpu. */
-	if (MACHINE_IS_VM)
-		pfault_fini();
-#endif
+	pfault_fini();
 
 	memset(&cr_parms.orvals, 0, sizeof(cr_parms.orvals));
 	memset(&cr_parms.andvals, 0xff, sizeof(cr_parms.andvals));
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index 92ecffb..475dbb8 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -58,12 +58,6 @@ int sysctl_userprocess_debug = 0;
 
 extern pgm_check_handler_t do_protection_exception;
 extern pgm_check_handler_t do_dat_exception;
-#ifdef CONFIG_PFAULT
-extern int pfault_init(void);
-extern void pfault_fini(void);
-extern void pfault_interrupt(__u16 error_code);
-static ext_int_info_t ext_int_pfault;
-#endif
 extern pgm_check_handler_t do_monitor_call;
 
 #define stack_pointer ({ void **sp; asm("la %0,0(15)" : "=&d" (sp)); sp; })
@@ -739,22 +733,5 @@ void __init trap_init(void)
         pgm_check_table[0x1C] = &space_switch_exception;
         pgm_check_table[0x1D] = &hfp_sqrt_exception;
 	pgm_check_table[0x40] = &do_monitor_call;
-
-	if (MACHINE_IS_VM) {
-#ifdef CONFIG_PFAULT
-		/*
-		 * Try to get pfault pseudo page faults going.
-		 */
-		if (register_early_external_interrupt(0x2603, pfault_interrupt,
-						      &ext_int_pfault) != 0)
-			panic("Couldn't request external interrupt 0x2603");
-
-		if (pfault_init() == 0) 
-			return;
-		
-		/* Tough luck, no pfault. */
-		unregister_early_external_interrupt(0x2603, pfault_interrupt,
-						    &ext_int_pfault);
-#endif
-	}
+	pfault_irq_init();
 }
-- 
cgit v1.1


From d57de5a36791cb1b7285649c62f183b0d3505f7d Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:42 +0100
Subject: [S390] Use diag260 for memory size detection.

Avoid the tprot loop if diag260 works and reports that there are no
holes in memory. The tprot instruction can lead to a significant delay
in the ipl process if the virtual guest has a lot of memory and the
host is under memory pressure.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/head64.S | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/head64.S b/arch/s390/kernel/head64.S
index 42f54d4..cc6dbc5 100644
--- a/arch/s390/kernel/head64.S
+++ b/arch/s390/kernel/head64.S
@@ -70,7 +70,22 @@ startup_continue:
 	sgr	%r5,%r5 		# set src,length and pad to zero
 	mvcle	%r2,%r4,0		# clear mem
 	jo	.-4			# branch back, if not finish
+					# 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
+	larl	%r2,memory_size
+	stg	%r1,0(%r2)		# set memory size
+	j	.Ldonemem
 
+.Lslowmemdetect:
 	l	%r2,.Lrcp-.LPG1(%r13)	# Read SCP forced command word
 .Lservicecall:
 	stosm	.Lpmask-.LPG1(%r13),0x01	# authorize ext interrupts
@@ -139,8 +154,6 @@ startup_continue:
 	.int	0x100000
 
 .Lfchunk:
-					# set program check new psw mask
-	mvc	__LC_PGM_NEW_PSW(8),.Lpcmsk-.LPG1(%r13)
 
 #
 # find memory chunks.
-- 
cgit v1.1


From 236257eee99a35d67d3feab0769bded83efdc3ec Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:47 +0100
Subject: [S390] lockdep: show held locks when showing a stackdump

Follow i386/x86_64:
lockdep can be used to print held locks when printing a
backtrace. This can be useful when debugging things like
'scheduling while atomic' asserts.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/traps.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index 475dbb8..3cbb0dc 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -129,7 +129,7 @@ __show_trace(unsigned long sp, unsigned long low, unsigned long high)
 	}
 }
 
-void show_trace(struct task_struct *task, unsigned long * stack)
+void show_trace(struct task_struct *task, unsigned long *stack)
 {
 	register unsigned long __r15 asm ("15");
 	unsigned long sp;
@@ -151,6 +151,9 @@ void show_trace(struct task_struct *task, unsigned long * stack)
 		__show_trace(sp, S390_lowcore.thread_info,
 			     S390_lowcore.thread_info + THREAD_SIZE);
 	printk("\n");
+	if (!task)
+		task = current;
+	debug_show_held_locks(task);
 }
 
 void show_stack(struct task_struct *task, unsigned long *sp)
-- 
cgit v1.1


From 8b62bc9642300471737bc3b77b2a4a2ead46dedb Mon Sep 17 00:00:00 2001
From: Heiko Carstens <heiko.carstens@de.ibm.com>
Date: Mon, 4 Dec 2006 15:40:56 +0100
Subject: [S390] Memory detection fixes.

VMALLOC_END on 31bit should be 0x8000000UL instead of 0x7fffffffL.
The page mask which is used to make sure memory_end is on 4MB/2MB
boundary is wrong and not needed. Therefore remove it.
Make sure a vmalloc area does also exist and work on (future)
machines with 4TB and more memory.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/s390/kernel/head31.S |  3 ---
 arch/s390/kernel/head64.S |  4 ----
 arch/s390/kernel/setup.c  | 49 +++++++++++++++++++++++++++++++----------------
 3 files changed, 32 insertions(+), 24 deletions(-)

(limited to 'arch/s390/kernel')

diff --git a/arch/s390/kernel/head31.S b/arch/s390/kernel/head31.S
index 9817c32..4388b33 100644
--- a/arch/s390/kernel/head31.S
+++ b/arch/s390/kernel/head31.S
@@ -177,8 +177,6 @@ startup_continue:
 	st	%r0,4(%r3)		# store size of chunk
 	st	%r6,8(%r3)		# store type of chunk
 	la	%r3,12(%r3)
-	l	%r4,.Lmemsize-.LPG1(%r13)	 # address of variable memory_size
-	st	%r5,0(%r4)		# store last end to memory size
 	ahi	%r10,-1			# update chunk number
 .Lchkloop:
 	lr	%r6,%r7			# set access code to last cc
@@ -293,7 +291,6 @@ startup_continue:
 .Lpcmvpg:.long	0x00080000,0x80000000 + .Lchkmvpg
 .Lpcidte:.long	0x00080000,0x80000000 + .Lchkidte
 .Lpcdiag9c:.long 0x00080000,0x80000000 + .Lchkdiag9c
-.Lmemsize:.long memory_size
 .Lmchunk:.long	memory_chunk
 .Lmflags:.long	machine_flags
 .Lbss_bgn:  .long __bss_start
diff --git a/arch/s390/kernel/head64.S b/arch/s390/kernel/head64.S
index cc6dbc5..c526279e1 100644
--- a/arch/s390/kernel/head64.S
+++ b/arch/s390/kernel/head64.S
@@ -81,8 +81,6 @@ startup_continue:
 	aghi	%r1,1			# size is one more than end
 	larl	%r2,memory_chunk
 	stg	%r1,8(%r2)		# store size of chunk
-	larl	%r2,memory_size
-	stg	%r1,0(%r2)		# set memory size
 	j	.Ldonemem
 
 .Lslowmemdetect:
@@ -188,8 +186,6 @@ startup_continue:
 	stg	%r0,8(%r3)		# store size of chunk
 	st	%r6,20(%r3)		# store type of chunk
 	la	%r3,24(%r3)
-	larl	%r8,memory_size
-	stg	%r5,0(%r8)		# store memory size
 	ahi	%r10,-1			# update chunk number
 .Lchkloop:
 	lr	%r6,%r7			# set access code to last cc
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index b1a8ad9..b928fec 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -62,7 +62,6 @@ EXPORT_SYMBOL_GPL(uaccess);
 unsigned int console_mode = 0;
 unsigned int console_devno = -1;
 unsigned int console_irq = -1;
-unsigned long memory_size = 0;
 unsigned long machine_flags = 0;
 
 struct mem_chunk memory_chunk[MEMORY_CHUNKS];
@@ -486,6 +485,37 @@ setup_resources(void)
 	}
 }
 
+static void __init setup_memory_end(void)
+{
+	unsigned long real_size, memory_size;
+	unsigned long max_mem, max_phys;
+	int i;
+
+	memory_size = real_size = 0;
+	max_phys = VMALLOC_END - VMALLOC_MIN_SIZE;
+	memory_end &= PAGE_MASK;
+
+	max_mem = memory_end ? min(max_phys, memory_end) : max_phys;
+
+	for (i = 0; i < MEMORY_CHUNKS; i++) {
+		struct mem_chunk *chunk = &memory_chunk[i];
+
+		real_size = max(real_size, chunk->addr + chunk->size);
+		if (chunk->addr >= max_mem) {
+			memset(chunk, 0, sizeof(*chunk));
+			continue;
+		}
+		if (chunk->addr + chunk->size > max_mem)
+			chunk->size = max_mem - chunk->addr;
+		memory_size = max(memory_size, chunk->addr + chunk->size);
+	}
+	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)
 {
@@ -642,8 +672,6 @@ setup_arch(char **cmdline_p)
 	init_mm.end_data = (unsigned long) &_edata;
 	init_mm.brk = (unsigned long) &_end;
 
-	memory_end = memory_size;
-
 	if (MACHINE_HAS_MVCOS)
 		memcpy(&uaccess, &uaccess_mvcos, sizeof(uaccess));
 	else
@@ -651,20 +679,7 @@ setup_arch(char **cmdline_p)
 
 	parse_early_param();
 
-#ifndef CONFIG_64BIT
-	memory_end &= ~0x400000UL;
-
-        /*
-         * We need some free virtual space to be able to do vmalloc.
-         * On a machine with 2GB memory we make sure that we have at
-         * least 128 MB free space for vmalloc.
-         */
-        if (memory_end > 1920*1024*1024)
-                memory_end = 1920*1024*1024;
-#else /* CONFIG_64BIT */
-	memory_end &= ~0x200000UL;
-#endif /* CONFIG_64BIT */
-
+	setup_memory_end();
 	setup_memory();
 	setup_resources();
 	setup_lowcore();
-- 
cgit v1.1