[patch 00/38] s390 patches for 2.6.22

[patch 00/38] s390 patches for 2.6.22

[Martin Schwidefsky]

This is the s390 patchset for the merge window of 2.6.22.
Most of it is cleanup and bug-fixes, but there are some
new features as well:
 - Channel path configure function.
 - Improved oops output with an in-kernel disassembler
 - Make use of generic bug.
 - zfcpdump support.
 - Processor degradation notification.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 01/38] memory detection: stop at first memory hole.

[Martin Schwidefsky]

[-- Attachment #1: 001-memory-hole.diff --]
[-- Type: text/plain, Size: 1137 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

If both sclp and diag memory detection don't work stop at the first
memory hole. Otherwise the code might loop forever...

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/kernel/early.c |    7 +++----
 1 files changed, 3 insertions(+), 4 deletions(-)

Index: quilt-2.6/arch/s390/kernel/early.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/early.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/early.c	2007-04-27 16:04:52.000000000 +0200
@@ -253,11 +253,10 @@
 			break;
 #endif
 		/*
-		 * Finish memory detection at the first hole, unless
-		 * - we reached the hsa -> skip it.
-		 * - we know there must be more.
+		 * Finish memory detection at the first hole
+		 * if storage size is unknown.
 		 */
-		if (cc == -1UL && !memsize && old_addr != ADDR2G)
+		if (cc == -1UL && !memsize)
 			break;
 		if (memsize && addr >= memsize)
 			break;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 02/38] cio/ipl: Clean interface between cio and ipl code.

[Martin Schwidefsky]

[-- Attachment #1: 100-cio-ipl.diff --]
[-- Type: text/plain, Size: 4443 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

Clean interface between cio and ipl code, so Peter stops complaining.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/kernel/ipl.c |   23 ++++++++++++++++++++++-
 drivers/s390/cio/cio.c |   38 ++++++++++----------------------------
 include/asm-s390/cio.h |    7 +++++++
 include/asm-s390/ipl.h |    3 +--
 4 files changed, 40 insertions(+), 31 deletions(-)

Index: quilt-2.6/arch/s390/kernel/ipl.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/ipl.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/ipl.c	2007-04-27 16:04:52.000000000 +0200
@@ -47,7 +47,7 @@
  * Must be in data section since the bss section
  * is not cleared when these are accessed.
  */
-u16 ipl_devno __attribute__((__section__(".data"))) = 0;
+static u16 ipl_devno __attribute__((__section__(".data"))) = 0;
 u32 ipl_flags __attribute__((__section__(".data"))) = 0;
 
 static char *ipl_type_str(enum ipl_type type)
@@ -1038,6 +1038,27 @@
 
 __initcall(s390_ipl_init);
 
+void __init ipl_save_parameters(void)
+{
+	struct cio_iplinfo iplinfo;
+	unsigned int *ipl_ptr;
+	void *src, *dst;
+
+	if (cio_get_iplinfo(&iplinfo))
+		return;
+
+	ipl_devno = iplinfo.devno;
+	ipl_flags |= IPL_DEVNO_VALID;
+	if (!iplinfo.is_qdio)
+		return;
+	ipl_flags |= IPL_PARMBLOCK_VALID;
+	ipl_ptr = (unsigned int *)__LC_IPL_PARMBLOCK_PTR;
+	src = (void *)(unsigned long)*ipl_ptr;
+	dst = (void *)IPL_PARMBLOCK_ORIGIN;
+	memmove(dst, src, PAGE_SIZE);
+	*ipl_ptr = IPL_PARMBLOCK_ORIGIN;
+}
+
 static LIST_HEAD(rcall);
 static DEFINE_MUTEX(rcall_mutex);
 
Index: quilt-2.6/drivers/s390/cio/cio.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/cio.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/cio.c	2007-04-27 16:04:52.000000000 +0200
@@ -1048,37 +1048,19 @@
 	do_reipl_asm(*((__u32*)&schid));
 }
 
-static struct schib __initdata ipl_schib;
-
-/*
- * ipl_save_parameters gets called very early. It is not allowed to access
- * anything in the bss section at all. The bss section is not cleared yet,
- * but may contain some ipl parameters written by the firmware.
- * These parameters (if present) are copied to 0x2000.
- * To avoid corruption of the ipl parameters, all variables used by this
- * function must reside on the stack or in the data section.
- */
-void ipl_save_parameters(void)
+int __init cio_get_iplinfo(struct cio_iplinfo *iplinfo)
 {
 	struct subchannel_id schid;
-	unsigned int *ipl_ptr;
-	void *src, *dst;
+	struct schib schib;
 
 	schid = *(struct subchannel_id *)__LC_SUBCHANNEL_ID;
 	if (!schid.one)
-		return;
-	if (stsch(schid, &ipl_schib))
-		return;
-	if (!ipl_schib.pmcw.dnv)
-		return;
-	ipl_devno = ipl_schib.pmcw.dev;
-	ipl_flags |= IPL_DEVNO_VALID;
-	if (!ipl_schib.pmcw.qf)
-		return;
-	ipl_flags |= IPL_PARMBLOCK_VALID;
-	ipl_ptr = (unsigned int *)__LC_IPL_PARMBLOCK_PTR;
-	src = (void *)(unsigned long)*ipl_ptr;
-	dst = (void *)IPL_PARMBLOCK_ORIGIN;
-	memmove(dst, src, PAGE_SIZE);
-	*ipl_ptr = IPL_PARMBLOCK_ORIGIN;
+		return -ENODEV;
+	if (stsch(schid, &schib))
+		return -ENODEV;
+	if (!schib.pmcw.dnv)
+		return -ENODEV;
+	iplinfo->devno = schib.pmcw.dev;
+	iplinfo->is_qdio = schib.pmcw.qf;
+	return 0;
 }
Index: quilt-2.6/include/asm-s390/cio.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/cio.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/cio.h	2007-04-27 16:04:52.000000000 +0200
@@ -292,6 +292,13 @@
 
 extern void reipl_ccw_dev(struct ccw_dev_id *id);
 
+struct cio_iplinfo {
+	u16 devno;
+	int is_qdio;
+};
+
+extern int cio_get_iplinfo(struct cio_iplinfo *iplinfo);
+
 #endif
 
 #endif
Index: quilt-2.6/include/asm-s390/ipl.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/ipl.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/ipl.h	2007-04-27 16:04:52.000000000 +0200
@@ -74,10 +74,9 @@
 } __attribute__((packed));
 
 /*
- * IPL validity flags and parameters as detected in head.S
+ * IPL validity flags
  */
 extern u32 ipl_flags;
-extern u16 ipl_devno;
 
 extern u32 dump_prefix_page;
 extern void do_reipl(void);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 03/38] cio: Introduce struct chp_id.

[Martin Schwidefsky]

[-- Attachment #1: 101-cio-chpid.diff --]
[-- Type: text/plain, Size: 17929 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Introduce data type for channel-path IDs. 

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 drivers/s390/cio/chpid.h      |   51 ++++++++++++
 drivers/s390/cio/chsc.c       |  178 +++++++++++++++++++++++-------------------
 drivers/s390/cio/chsc.h       |    8 +
 drivers/s390/cio/cio.c        |    5 -
 drivers/s390/cio/device_fsm.c |    6 +
 drivers/s390/cio/ioasm.h      |    5 -
 6 files changed, 167 insertions(+), 86 deletions(-)

Index: quilt-2.6/drivers/s390/cio/chpid.h
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/cio/chpid.h	2007-04-27 16:04:52.000000000 +0200
@@ -0,0 +1,51 @@
+/*
+ *  drivers/s390/cio/chpid.h
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#ifndef S390_CHP_ID_H
+#define S390_CHP_ID_H S390_CHP_ID_H
+
+#include <linux/string.h>
+#include <asm/types.h>
+#include "css.h"
+
+struct chp_id {
+	u8 reserved1;
+	u8 cssid;
+	u8 reserved2;
+	u8 id;
+} __attribute__((packed));
+
+static inline void chp_id_init(struct chp_id *chpid)
+{
+	memset(chpid, 0, sizeof(struct chp_id));
+}
+
+static inline int chp_id_is_equal(struct chp_id *a, struct chp_id *b)
+{
+	return (a->id == b->id) && (a->cssid == b->cssid);
+}
+
+static inline void chp_id_next(struct chp_id *chpid)
+{
+	if (chpid->id < __MAX_CHPID)
+		chpid->id++;
+	else {
+		chpid->id = 0;
+		chpid->cssid++;
+	}
+}
+
+static inline int chp_id_is_valid(struct chp_id *chpid)
+{
+	return (chpid->cssid <= __MAX_CSSID);
+}
+
+
+#define chp_id_for_each(c) \
+	for (chp_id_init(c); chp_id_is_valid(c); chp_id_next(c))
+
+#endif /* S390_CHP_ID_H */
Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:52.000000000 +0200
@@ -20,43 +20,48 @@
 #include "cio.h"
 #include "cio_debug.h"
 #include "ioasm.h"
+#include "chpid.h"
 #include "chsc.h"
 
 static void *sei_page;
 
-static int new_channel_path(int chpid);
+static int new_channel_path(struct chp_id chpid);
 
-static inline void
-set_chp_logically_online(int chp, int onoff)
+static inline struct channel_path *chpid_to_chp(struct chp_id chpid)
 {
-	css[0]->chps[chp]->state = onoff;
+	return css[chpid.cssid]->chps[chpid.id];
 }
 
-static int
-get_chp_status(int chp)
+static void set_chp_logically_online(struct chp_id chpid, int onoff)
 {
-	return (css[0]->chps[chp] ? css[0]->chps[chp]->state : -ENODEV);
+	chpid_to_chp(chpid)->state = onoff;
 }
 
-void
-chsc_validate_chpids(struct subchannel *sch)
+static int get_chp_status(struct chp_id chpid)
+{
+	return (chpid_to_chp(chpid) ? chpid_to_chp(chpid)->state : -ENODEV);
+}
+
+void chsc_validate_chpids(struct subchannel *sch)
 {
 	int mask, chp;
+	struct chp_id chpid;
 
+	chp_id_init(&chpid);
 	for (chp = 0; chp <= 7; chp++) {
 		mask = 0x80 >> chp;
-		if (!get_chp_status(sch->schib.pmcw.chpid[chp]))
+		chpid.id = sch->schib.pmcw.chpid[chp];
+		if (!get_chp_status(chpid))
 			/* disable using this path */
 			sch->opm &= ~mask;
 	}
 }
 
-void
-chpid_is_actually_online(int chp)
+void chpid_is_actually_online(struct chp_id chpid)
 {
 	int state;
 
-	state = get_chp_status(chp);
+	state = get_chp_status(chpid);
 	if (state < 0) {
 		need_rescan = 1;
 		queue_work(slow_path_wq, &slow_path_work);
@@ -200,11 +205,14 @@
 	spin_unlock_irq(sch->lock);
 	free_page((unsigned long)page);
 	if (!ret) {
-		int j, chpid, mask;
+		int j, mask;
+		struct chp_id chpid;
+
+		chp_id_init(&chpid);
 		/* Allocate channel path structures, if needed. */
 		for (j = 0; j < 8; j++) {
 			mask = 0x80 >> j;
-			chpid = sch->ssd_info.chpid[j];
+			chpid.id = sch->ssd_info.chpid[j];
 			if ((sch->schib.pmcw.pim & mask) &&
 			    (get_chp_status(chpid) < 0))
 			    new_channel_path(chpid);
@@ -227,7 +235,7 @@
 	for (j = 0; j < 8; j++) {
 		mask = 0x80 >> j;
 		if ((sch->schib.pmcw.pim & mask) &&
-		    (sch->schib.pmcw.chpid[j] == chpid->id))
+		    (sch->schib.pmcw.chpid[j] == chpid->chpid.id))
 			break;
 	}
 	if (j >= 8)
@@ -277,18 +285,17 @@
 	return 0;
 }
 
-static void
-s390_set_chpid_offline( __u8 chpid)
+static void s390_set_chpid_offline(struct chp_id chpid)
 {
 	char dbf_txt[15];
 	struct device *dev;
 
-	sprintf(dbf_txt, "chpr%x", chpid);
+	sprintf(dbf_txt, "chpr%x.%02x", chpid.cssid, chpid.id);
 	CIO_TRACE_EVENT(2, dbf_txt);
 
 	if (get_chp_status(chpid) <= 0)
 		return;
-	dev = get_device(&css[0]->chps[chpid]->dev);
+	dev = get_device(&(chpid_to_chp(chpid)->dev));
 	bus_for_each_dev(&css_bus_type, NULL, to_channelpath(dev),
 			 s390_subchannel_remove_chpid);
 
@@ -316,7 +323,7 @@
 		 * check if chpid is in information updated by ssd
 		 */
 		if (sch->ssd_info.valid &&
-		    sch->ssd_info.chpid[chp] == res_data->chp->id &&
+		    sch->ssd_info.chpid[chp] == res_data->chp->chpid.id &&
 		    (sch->ssd_info.fla[chp] & res_data->fla_mask)
 		    == res_data->fla) {
 			found = 1;
@@ -409,7 +416,8 @@
 	int rc;
 	char dbf_txt[15];
 
-	sprintf(dbf_txt, "accpr%x", res_data->chp->id);
+	sprintf(dbf_txt, "accpr%x.%02x", res_data->chp->chpid.cssid,
+		res_data->chp->chpid.id);
 	CIO_TRACE_EVENT( 2, dbf_txt);
 	if (res_data->fla != 0) {
 		sprintf(dbf_txt, "fla%x", res_data->fla);
@@ -482,17 +490,21 @@
 
 static int chsc_process_sei_link_incident(struct chsc_sei_area *sei_area)
 {
-	int chpid;
+	struct chp_id chpid;
+	int id;
 
 	CIO_CRW_EVENT(4, "chsc: link incident (rs=%02x, rs_id=%04x)\n",
 		      sei_area->rs, sei_area->rsid);
 	if (sei_area->rs != 4)
 		return 0;
-	chpid = __get_chpid_from_lir(sei_area->ccdf);
-	if (chpid < 0)
+	id = __get_chpid_from_lir(sei_area->ccdf);
+	if (id < 0)
 		CIO_CRW_EVENT(4, "chsc: link incident - invalid LIR\n");
-	else
+	else {
+		chp_id_init(&chpid);
+		chpid.id = id;
 		s390_set_chpid_offline(chpid);
+	}
 
 	return 0;
 }
@@ -501,6 +513,7 @@
 {
 	struct res_acc_data res_data;
 	struct device *dev;
+	struct chp_id chpid;
 	int status;
 	int rc;
 
@@ -508,13 +521,15 @@
 		      "rs_id=%04x)\n", sei_area->rs, sei_area->rsid);
 	if (sei_area->rs != 4)
 		return 0;
+	chp_id_init(&chpid);
+	chpid.id = sei_area->rsid;
 	/* allocate a new channel path structure, if needed */
-	status = get_chp_status(sei_area->rsid);
+	status = get_chp_status(chpid);
 	if (status < 0)
-		new_channel_path(sei_area->rsid);
+		new_channel_path(chpid);
 	else if (!status)
 		return 0;
-	dev = get_device(&css[0]->chps[sei_area->rsid]->dev);
+	dev = get_device(&(chpid_to_chp(chpid)->dev));
 	memset(&res_data, 0, sizeof(struct res_acc_data));
 	res_data.chp = to_channelpath(dev);
 	if ((sei_area->vf & 0xc0) != 0) {
@@ -631,7 +646,7 @@
 	for (i=0; i<8; i++) {
 		mask = 0x80 >> i;
 		if ((sch->schib.pmcw.pim & mask) &&
-		    (sch->schib.pmcw.chpid[i] == chp->id)) {
+		    (sch->schib.pmcw.chpid[i] == chp->chpid.id)) {
 			if (stsch(sch->schid, &sch->schib) != 0) {
 				/* Endgame. */
 				spin_unlock_irq(sch->lock);
@@ -657,8 +672,7 @@
 	return 0;
 }
 
-static int
-chp_add(int chpid)
+static int chp_add(struct chp_id chpid)
 {
 	int rc;
 	char dbf_txt[15];
@@ -667,10 +681,10 @@
 	if (!get_chp_status(chpid))
 		return 0; /* no need to do the rest */
 	
-	sprintf(dbf_txt, "cadd%x", chpid);
+	sprintf(dbf_txt, "cadd%x.%02x", chpid.cssid, chpid.id);
 	CIO_TRACE_EVENT(2, dbf_txt);
 
-	dev = get_device(&css[0]->chps[chpid]->dev);
+	dev = get_device(&(chpid_to_chp(chpid)->dev));
 	rc = for_each_subchannel(__chp_add, to_channelpath(dev));
 	if (css_slow_subchannels_exist())
 		rc = -EAGAIN;
@@ -683,9 +697,12 @@
 /* 
  * Handling of crw machine checks with channel path source.
  */
-int
-chp_process_crw(int chpid, int on)
+int chp_process_crw(int id, int on)
 {
+	struct chp_id chpid;
+
+	chp_id_init(&chpid);
+	chpid.id = id;
 	if (on == 0) {
 		/* Path has gone. We use the link incident routine.*/
 		s390_set_chpid_offline(chpid);
@@ -733,8 +750,8 @@
 		sch->driver->termination(&sch->dev);
 }
 
-static void
-__s390_subchannel_vary_chpid(struct subchannel *sch, __u8 chpid, int on)
+static void __s390_subchannel_vary_chpid(struct subchannel *sch,
+					 struct chp_id chpid, int on)
 {
 	int chp, old_lpm;
 	unsigned long flags;
@@ -745,7 +762,7 @@
 	spin_lock_irqsave(sch->lock, flags);
 	old_lpm = sch->lpm;
 	for (chp = 0; chp < 8; chp++) {
-		if (sch->ssd_info.chpid[chp] != chpid)
+		if (sch->ssd_info.chpid[chp] != chpid.id)
 			continue;
 
 		if (on) {
@@ -780,11 +797,10 @@
 	spin_unlock_irqrestore(sch->lock, flags);
 }
 
-static int
-s390_subchannel_vary_chpid_off(struct device *dev, void *data)
+static int s390_subchannel_vary_chpid_off(struct device *dev, void *data)
 {
 	struct subchannel *sch;
-	__u8 *chpid;
+	struct chp_id *chpid;
 
 	sch = to_subchannel(dev);
 	chpid = data;
@@ -793,11 +809,10 @@
 	return 0;
 }
 
-static int
-s390_subchannel_vary_chpid_on(struct device *dev, void *data)
+static int s390_subchannel_vary_chpid_on(struct device *dev, void *data)
 {
 	struct subchannel *sch;
-	__u8 *chpid;
+	struct chp_id *chpid;
 
 	sch = to_subchannel(dev);
 	chpid = data;
@@ -833,23 +848,25 @@
  * Function: s390_vary_chpid
  * Varies the specified chpid online or offline
  */
-static int
-s390_vary_chpid( __u8 chpid, int on)
+static int s390_vary_chpid(struct chp_id chpid, int on)
 {
 	char dbf_text[15];
 	int status;
 
-	sprintf(dbf_text, on?"varyon%x":"varyoff%x", chpid);
+	sprintf(dbf_text, on?"varyon%x.%02x":"varyoff%x.%02x", chpid.cssid,
+		chpid.id);
 	CIO_TRACE_EVENT( 2, dbf_text);
 
 	status = get_chp_status(chpid);
 	if (status < 0) {
-		printk(KERN_ERR "Can't vary unknown chpid %02X\n", chpid);
+		printk(KERN_ERR "Can't vary unknown chpid %x.%02x\n",
+		       chpid.cssid, chpid.id);
 		return -EINVAL;
 	}
 
 	if (!on && !status) {
-		printk(KERN_ERR "chpid %x is already offline\n", chpid);
+		printk(KERN_ERR "chpid %x.%02x is already offline\n",
+		       chpid.cssid, chpid.id);
 		return -EINVAL;
 	}
 
@@ -904,20 +921,19 @@
 	.read = chp_measurement_chars_read,
 };
 
-static void
-chp_measurement_copy_block(struct cmg_entry *buf,
-			   struct channel_subsystem *css, int chpid)
+static void chp_measurement_copy_block(struct cmg_entry *buf,
+			struct channel_subsystem *css, struct chp_id chpid)
 {
 	void *area;
 	struct cmg_entry *entry, reference_buf;
 	int idx;
 
-	if (chpid < 128) {
+	if (chpid.id < 128) {
 		area = css->cub_addr1;
-		idx = chpid;
+		idx = chpid.id;
 	} else {
 		area = css->cub_addr2;
-		idx = chpid - 128;
+		idx = chpid.id - 128;
 	}
 	entry = area + (idx * sizeof(struct cmg_entry));
 	do {
@@ -941,7 +957,7 @@
 	/* Only allow single reads. */
 	if (off || count < size)
 		return 0;
-	chp_measurement_copy_block((struct cmg_entry *)buf, css, chp->id);
+	chp_measurement_copy_block((struct cmg_entry *)buf, css, chp->chpid);
 	count = size;
 	return count;
 }
@@ -1137,7 +1153,7 @@
 
 	if (!chp)
 		return 0;
-	return (get_chp_status(chp->id) ? sprintf(buf, "online\n") :
+	return (get_chp_status(chp->chpid) ? sprintf(buf, "online\n") :
 		sprintf(buf, "offline\n"));
 }
 
@@ -1154,9 +1170,9 @@
 		return count;
 
 	if (!strnicmp(cmd, "on", 2))
-		error = s390_vary_chpid(cp->id, 1);
+		error = s390_vary_chpid(cp->chpid, 1);
 	else if (!strnicmp(cmd, "off", 3))
-		error = s390_vary_chpid(cp->id, 0);
+		error = s390_vary_chpid(cp->chpid, 0);
 	else
 		error = -EINVAL;
 
@@ -1227,9 +1243,8 @@
 	kfree(cp);
 }
 
-static int
-chsc_determine_channel_path_description(int chpid,
-					struct channel_path_desc *desc)
+static int chsc_determine_channel_path_description(struct chp_id chpid,
+						struct channel_path_desc *desc)
 {
 	int ccode, ret;
 
@@ -1252,8 +1267,8 @@
 	scpd_area->request.length = 0x0010;
 	scpd_area->request.code = 0x0002;
 
-	scpd_area->first_chpid = chpid;
-	scpd_area->last_chpid = chpid;
+	scpd_area->first_chpid = chpid.id;
+	scpd_area->last_chpid = chpid.id;
 
 	ccode = chsc(scpd_area);
 	if (ccode > 0) {
@@ -1349,8 +1364,8 @@
 	scmc_area->request.length = 0x0010;
 	scmc_area->request.code = 0x0022;
 
-	scmc_area->first_chpid = chp->id;
-	scmc_area->last_chpid = chp->id;
+	scmc_area->first_chpid = chp->chpid.id;
+	scmc_area->last_chpid = chp->chpid.id;
 
 	ccode = chsc(scmc_area);
 	if (ccode > 0) {
@@ -1396,8 +1411,7 @@
  * Entries for chpids on the system bus.
  * This replaces /proc/chpids.
  */
-static int
-new_channel_path(int chpid)
+static int new_channel_path(struct chp_id chpid)
 {
 	struct channel_path *chp;
 	int ret;
@@ -1407,11 +1421,12 @@
 		return -ENOMEM;
 
 	/* fill in status, etc. */
-	chp->id = chpid;
+	chp->chpid = chpid;
 	chp->state = 1;
-	chp->dev.parent = &css[0]->device;
+	chp->dev.parent = &css[chpid.cssid]->device;
 	chp->dev.release = chp_release;
-	snprintf(chp->dev.bus_id, BUS_ID_SIZE, "chp0.%x", chpid);
+	snprintf(chp->dev.bus_id, BUS_ID_SIZE, "chp%x.%02x", chpid.cssid,
+		 chpid.id);
 
 	/* Obtain channel path description and fill it in. */
 	ret = chsc_determine_channel_path_description(chpid, &chp->desc);
@@ -1437,8 +1452,8 @@
 	/* make it known to the system */
 	ret = device_register(&chp->dev);
 	if (ret) {
-		printk(KERN_WARNING "%s: could not register %02x\n",
-		       __func__, chpid);
+		printk(KERN_WARNING "%s: could not register %x.%02x\n",
+		       __func__, chpid.cssid, chpid.id);
 		goto out_free;
 	}
 	ret = sysfs_create_group(&chp->dev.kobj, &chp_attr_group);
@@ -1446,18 +1461,18 @@
 		device_unregister(&chp->dev);
 		goto out_free;
 	}
-	mutex_lock(&css[0]->mutex);
-	if (css[0]->cm_enabled) {
+	mutex_lock(&css[chpid.cssid]->mutex);
+	if (css[chpid.cssid]->cm_enabled) {
 		ret = chsc_add_chp_cmg_attr(chp);
 		if (ret) {
 			sysfs_remove_group(&chp->dev.kobj, &chp_attr_group);
 			device_unregister(&chp->dev);
-			mutex_unlock(&css[0]->mutex);
+			mutex_unlock(&css[chpid.cssid]->mutex);
 			goto out_free;
 		}
 	}
-	css[0]->chps[chpid] = chp;
-	mutex_unlock(&css[0]->mutex);
+	css[chpid.cssid]->chps[chpid.id] = chp;
+	mutex_unlock(&css[chpid.cssid]->mutex);
 	return ret;
 out_free:
 	kfree(chp);
@@ -1469,8 +1484,11 @@
 {
 	struct channel_path *chp;
 	struct channel_path_desc *desc;
+	struct chp_id chpid;
 
-	chp = css[0]->chps[sch->schib.pmcw.chpid[chp_no]];
+	chp_id_init(&chpid);
+	chpid.id = sch->schib.pmcw.chpid[chp_no];
+	chp = chpid_to_chp(chpid);
 	if (!chp)
 		return NULL;
 	desc = kmalloc(sizeof(struct channel_path_desc), GFP_KERNEL);
Index: quilt-2.6/drivers/s390/cio/chsc.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.h	2007-04-27 16:04:52.000000000 +0200
@@ -1,6 +1,10 @@
 #ifndef S390_CHSC_H
 #define S390_CHSC_H
 
+#include <linux/types.h>
+#include <linux/device.h>
+#include "chpid.h"
+
 #define CHSC_SEI_ACC_CHPID        1
 #define CHSC_SEI_ACC_LINKADDR     2
 #define CHSC_SEI_ACC_FULLLINKADDR 3
@@ -34,7 +38,7 @@
 } __attribute__ ((packed));
 
 struct channel_path {
-	int id;
+	struct chp_id chpid;
 	int state;
 	struct channel_path_desc desc;
 	/* Channel-measurement related stuff: */
@@ -46,7 +50,7 @@
 
 extern void s390_process_css( void );
 extern void chsc_validate_chpids(struct subchannel *);
-extern void chpid_is_actually_online(int);
+extern void chpid_is_actually_online(struct chp_id);
 extern int css_get_ssd_info(struct subchannel *);
 extern int chsc_process_crw(void);
 extern int chp_process_crw(int, int);
Index: quilt-2.6/drivers/s390/cio/cio.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/cio.c	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/cio.c	2007-04-27 16:04:52.000000000 +0200
@@ -954,6 +954,7 @@
 {
 	int i, ret;
 	unsigned long long timeout;
+	struct chp_id chpid;
 
 	/* Reset subchannels. */
 	for_each_subchannel(__shutdown_subchannel_easy,  NULL);
@@ -963,8 +964,10 @@
 	__ctl_set_bit(14, 28);
 	/* Temporarily reenable machine checks. */
 	local_mcck_enable();
+	chp_id_init(&chpid);
 	for (i = 0; i <= __MAX_CHPID; i++) {
-		ret = rchp(i);
+		chpid.id = i;
+		ret = rchp(chpid);
 		if ((ret == 0) || (ret == 2))
 			/*
 			 * rchp either succeeded, or another rchp is already
Index: quilt-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_fsm.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:52.000000000 +0200
@@ -22,6 +22,7 @@
 #include "device.h"
 #include "chsc.h"
 #include "ioasm.h"
+#include "chpid.h"
 
 int
 device_is_online(struct subchannel *sch)
@@ -210,14 +211,17 @@
 __recover_lost_chpids(struct subchannel *sch, int old_lpm)
 {
 	int mask, i;
+	struct chp_id chpid;
 
+	chp_id_init(&chpid);
 	for (i = 0; i<8; i++) {
 		mask = 0x80 >> i;
 		if (!(sch->lpm & mask))
 			continue;
 		if (old_lpm & mask)
 			continue;
-		chpid_is_actually_online(sch->schib.pmcw.chpid[i]);
+		chpid.id = sch->schib.pmcw.chpid[i];
+		chpid_is_actually_online(chpid);
 	}
 }
 
Index: quilt-2.6/drivers/s390/cio/ioasm.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/ioasm.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/ioasm.h	2007-04-27 16:04:52.000000000 +0200
@@ -2,6 +2,7 @@
 #define S390_CIO_IOASM_H
 
 #include "schid.h"
+#include "chpid.h"
 
 /*
  * TPI info structure
@@ -189,9 +190,9 @@
 	return cc;
 }
 
-static inline int rchp(int chpid)
+static inline int rchp(struct chp_id chpid)
 {
-	register unsigned int reg1 asm ("1") = chpid;
+	register struct chp_id reg1 asm ("1") = chpid;
 	int ccode;
 
 	asm volatile(

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 04/38] cio: Allow 0 and 1 as input for channel path status attribute.

[Martin Schwidefsky]

[-- Attachment #1: 102-cio-chpid01.diff --]
[-- Type: text/plain, Size: 1170 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Channel path status can now be modified by writing '0' and '1'
to the sysfs status attribute in addition to 'offline' and
'online' respectively.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 drivers/s390/cio/chsc.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:53.000000000 +0200
@@ -1169,9 +1169,9 @@
 	if (!num_args)
 		return count;
 
-	if (!strnicmp(cmd, "on", 2))
+	if (!strnicmp(cmd, "on", 2) || !strcmp(cmd, "1"))
 		error = s390_vary_chpid(cp->chpid, 1);
-	else if (!strnicmp(cmd, "off", 3))
+	else if (!strnicmp(cmd, "off", 3) || !strcmp(cmd, "0"))
 		error = s390_vary_chpid(cp->chpid, 0);
 	else
 		error = -EINVAL;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 05/38] cio: Introduce separate files for channel-path related code.

[Martin Schwidefsky]

[-- Attachment #1: 103-cio-chpid02.diff --]
[-- Type: text/plain, Size: 35005 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 drivers/s390/cio/Makefile     |    2 
 drivers/s390/cio/chp.c        |  433 ++++++++++++++++++++++++++++++++++++++
 drivers/s390/cio/chp.h        |   37 +++
 drivers/s390/cio/chsc.c       |  473 +++---------------------------------------
 drivers/s390/cio/chsc.h       |   30 --
 drivers/s390/cio/cio.c        |    8 
 drivers/s390/cio/css.h        |    1 
 drivers/s390/cio/device_fsm.c |    6 
 drivers/s390/cio/device_ops.c |    7 
 drivers/s390/s390mach.c       |    1 
 10 files changed, 535 insertions(+), 463 deletions(-)

Index: quilt-2.6/drivers/s390/cio/chp.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/cio/chp.c	2007-04-27 16:04:53.000000000 +0200
@@ -0,0 +1,433 @@
+/*
+ *  drivers/s390/cio/chp.c
+ *
+ *    Copyright IBM Corp. 1999,2007
+ *    Author(s): Cornelia Huck (cornelia.huck@de.ibm.com)
+ *		 Arnd Bergmann (arndb@de.ibm.com)
+ *		 Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#include <linux/bug.h>
+#include <linux/workqueue.h>
+#include <linux/spinlock.h>
+#include <asm/errno.h>
+
+#include "chpid.h"
+#include "cio.h"
+#include "css.h"
+#include "ioasm.h"
+#include "cio_debug.h"
+#include "chp.h"
+
+#define to_channelpath(device) container_of(device, struct channel_path, dev)
+
+/* Return channel_path struct for given chpid. */
+static inline struct channel_path *chpid_to_chp(struct chp_id chpid)
+{
+	return css[chpid.cssid]->chps[chpid.id];
+}
+
+/* Set vary state for given chpid. */
+static void set_chp_logically_online(struct chp_id chpid, int onoff)
+{
+	chpid_to_chp(chpid)->state = onoff;
+}
+
+/* On succes return 0 if channel-path is varied offline, 1 if it is varied
+ * online. Return -ENODEV if channel-path is not registered. */
+int chp_get_status(struct chp_id chpid)
+{
+	return (chpid_to_chp(chpid) ? chpid_to_chp(chpid)->state : -ENODEV);
+}
+
+/**
+ * chp_get_sch_opm - return opm for subchannel
+ * @sch: subchannel
+ *
+ * Calculate and return the operational path mask (opm) based on the chpids
+ * used by the subchannel and the status of the associated channel-paths.
+ */
+u8 chp_get_sch_opm(struct subchannel *sch)
+{
+	struct chp_id chpid;
+	int opm;
+	int i;
+
+	opm = 0;
+	chp_id_init(&chpid);
+	for (i=0; i < 8; i++) {
+		opm <<= 1;
+		chpid.id = sch->schib.pmcw.chpid[i];
+		if (chp_get_status(chpid) != 0)
+			opm |= 1;
+	}
+	return opm;
+}
+
+/**
+ * chp_is_registered - check if a channel-path is registered
+ * @chpid: channel-path ID
+ *
+ * Return non-zero if a channel-path with the given chpid is registered,
+ * zero otherwise.
+ */
+int chp_is_registered(struct chp_id chpid)
+{
+	return chpid_to_chp(chpid) != NULL;
+}
+
+/*
+ * Function: s390_vary_chpid
+ * Varies the specified chpid online or offline
+ */
+static int s390_vary_chpid(struct chp_id chpid, int on)
+{
+	char dbf_text[15];
+	int status;
+
+	sprintf(dbf_text, on?"varyon%x.%02x":"varyoff%x.%02x", chpid.cssid,
+		chpid.id);
+	CIO_TRACE_EVENT( 2, dbf_text);
+
+	status = chp_get_status(chpid);
+	if (status < 0) {
+		printk(KERN_ERR "Can't vary unknown chpid %x.%02x\n",
+		       chpid.cssid, chpid.id);
+		return -EINVAL;
+	}
+
+	if (!on && !status) {
+		printk(KERN_ERR "chpid %x.%02x is already offline\n",
+		       chpid.cssid, chpid.id);
+		return -EINVAL;
+	}
+
+	set_chp_logically_online(chpid, on);
+	chsc_chp_vary(chpid, on);
+	return 0;
+}
+
+/*
+ * Channel measurement related functions
+ */
+static ssize_t chp_measurement_chars_read(struct kobject *kobj, char *buf,
+					  loff_t off, size_t count)
+{
+	struct channel_path *chp;
+	unsigned int size;
+
+	chp = to_channelpath(container_of(kobj, struct device, kobj));
+	if (!chp->cmg_chars)
+		return 0;
+
+	size = sizeof(struct cmg_chars);
+
+	if (off > size)
+		return 0;
+	if (off + count > size)
+		count = size - off;
+	memcpy(buf, chp->cmg_chars + off, count);
+	return count;
+}
+
+static struct bin_attribute chp_measurement_chars_attr = {
+	.attr = {
+		.name = "measurement_chars",
+		.mode = S_IRUSR,
+		.owner = THIS_MODULE,
+	},
+	.size = sizeof(struct cmg_chars),
+	.read = chp_measurement_chars_read,
+};
+
+static void chp_measurement_copy_block(struct cmg_entry *buf,
+				       struct channel_subsystem *css,
+				       struct chp_id chpid)
+{
+	void *area;
+	struct cmg_entry *entry, reference_buf;
+	int idx;
+
+	if (chpid.id < 128) {
+		area = css->cub_addr1;
+		idx = chpid.id;
+	} else {
+		area = css->cub_addr2;
+		idx = chpid.id - 128;
+	}
+	entry = area + (idx * sizeof(struct cmg_entry));
+	do {
+		memcpy(buf, entry, sizeof(*entry));
+		memcpy(&reference_buf, entry, sizeof(*entry));
+	} while (reference_buf.values[0] != buf->values[0]);
+}
+
+static ssize_t chp_measurement_read(struct kobject *kobj, char *buf,
+				    loff_t off, size_t count)
+{
+	struct channel_path *chp;
+	struct channel_subsystem *css;
+	unsigned int size;
+
+	chp = to_channelpath(container_of(kobj, struct device, kobj));
+	css = to_css(chp->dev.parent);
+
+	size = sizeof(struct cmg_entry);
+
+	/* Only allow single reads. */
+	if (off || count < size)
+		return 0;
+	chp_measurement_copy_block((struct cmg_entry *)buf, css, chp->chpid);
+	count = size;
+	return count;
+}
+
+static struct bin_attribute chp_measurement_attr = {
+	.attr = {
+		.name = "measurement",
+		.mode = S_IRUSR,
+		.owner = THIS_MODULE,
+	},
+	.size = sizeof(struct cmg_entry),
+	.read = chp_measurement_read,
+};
+
+void chp_remove_cmg_attr(struct channel_path *chp)
+{
+	device_remove_bin_file(&chp->dev, &chp_measurement_chars_attr);
+	device_remove_bin_file(&chp->dev, &chp_measurement_attr);
+}
+
+int chp_add_cmg_attr(struct channel_path *chp)
+{
+	int ret;
+
+	ret = device_create_bin_file(&chp->dev, &chp_measurement_chars_attr);
+	if (ret)
+		return ret;
+	ret = device_create_bin_file(&chp->dev, &chp_measurement_attr);
+	if (ret)
+		device_remove_bin_file(&chp->dev, &chp_measurement_chars_attr);
+	return ret;
+}
+
+/*
+ * Files for the channel path entries.
+ */
+static ssize_t chp_status_show(struct device *dev,
+			       struct device_attribute *attr, char *buf)
+{
+	struct channel_path *chp = container_of(dev, struct channel_path, dev);
+
+	if (!chp)
+		return 0;
+	return (chp_get_status(chp->chpid) ? sprintf(buf, "online\n") :
+		sprintf(buf, "offline\n"));
+}
+
+static ssize_t chp_status_write(struct device *dev,
+				struct device_attribute *attr,
+				const char *buf, size_t count)
+{
+	struct channel_path *cp = container_of(dev, struct channel_path, dev);
+	char cmd[10];
+	int num_args;
+	int error;
+
+	num_args = sscanf(buf, "%5s", cmd);
+	if (!num_args)
+		return count;
+
+	if (!strnicmp(cmd, "on", 2) || !strcmp(cmd, "1"))
+		error = s390_vary_chpid(cp->chpid, 1);
+	else if (!strnicmp(cmd, "off", 3) || !strcmp(cmd, "0"))
+		error = s390_vary_chpid(cp->chpid, 0);
+	else
+		error = -EINVAL;
+
+	return error < 0 ? error : count;
+
+}
+
+static DEVICE_ATTR(status, 0644, chp_status_show, chp_status_write);
+
+static ssize_t chp_type_show(struct device *dev, struct device_attribute *attr,
+			     char *buf)
+{
+	struct channel_path *chp = container_of(dev, struct channel_path, dev);
+
+	if (!chp)
+		return 0;
+	return sprintf(buf, "%x\n", chp->desc.desc);
+}
+
+static DEVICE_ATTR(type, 0444, chp_type_show, NULL);
+
+static ssize_t chp_cmg_show(struct device *dev, struct device_attribute *attr,
+			    char *buf)
+{
+	struct channel_path *chp = to_channelpath(dev);
+
+	if (!chp)
+		return 0;
+	if (chp->cmg == -1) /* channel measurements not available */
+		return sprintf(buf, "unknown\n");
+	return sprintf(buf, "%x\n", chp->cmg);
+}
+
+static DEVICE_ATTR(cmg, 0444, chp_cmg_show, NULL);
+
+static ssize_t chp_shared_show(struct device *dev,
+			       struct device_attribute *attr, char *buf)
+{
+	struct channel_path *chp = to_channelpath(dev);
+
+	if (!chp)
+		return 0;
+	if (chp->shared == -1) /* channel measurements not available */
+		return sprintf(buf, "unknown\n");
+	return sprintf(buf, "%x\n", chp->shared);
+}
+
+static DEVICE_ATTR(shared, 0444, chp_shared_show, NULL);
+
+static struct attribute * chp_attrs[] = {
+	&dev_attr_status.attr,
+	&dev_attr_type.attr,
+	&dev_attr_cmg.attr,
+	&dev_attr_shared.attr,
+	NULL,
+};
+
+static struct attribute_group chp_attr_group = {
+	.attrs = chp_attrs,
+};
+
+static void chp_release(struct device *dev)
+{
+	struct channel_path *cp;
+
+	cp = container_of(dev, struct channel_path, dev);
+	kfree(cp);
+}
+
+/**
+ * chp_new - register a new channel-path
+ * @chpid - channel-path ID
+ *
+ * Create and register data structure representing new channel-path. Return
+ * zero on success, non-zero otherwise.
+ */
+int chp_new(struct chp_id chpid)
+{
+	struct channel_path *chp;
+	int ret;
+
+	chp = kzalloc(sizeof(struct channel_path), GFP_KERNEL);
+	if (!chp)
+		return -ENOMEM;
+
+	/* fill in status, etc. */
+	chp->chpid = chpid;
+	chp->state = 1;
+	chp->dev.parent = &css[chpid.cssid]->device;
+	chp->dev.release = chp_release;
+	snprintf(chp->dev.bus_id, BUS_ID_SIZE, "chp%x.%02x", chpid.cssid,
+		 chpid.id);
+
+	/* Obtain channel path description and fill it in. */
+	ret = chsc_determine_channel_path_description(chpid, &chp->desc);
+	if (ret)
+		goto out_free;
+	/* Get channel-measurement characteristics. */
+	if (css_characteristics_avail && css_chsc_characteristics.scmc
+	    && css_chsc_characteristics.secm) {
+		ret = chsc_get_channel_measurement_chars(chp);
+		if (ret)
+			goto out_free;
+	} else {
+		static int msg_done;
+
+		if (!msg_done) {
+			printk(KERN_WARNING "cio: Channel measurements not "
+			       "available, continuing.\n");
+			msg_done = 1;
+		}
+		chp->cmg = -1;
+	}
+
+	/* make it known to the system */
+	ret = device_register(&chp->dev);
+	if (ret) {
+		printk(KERN_WARNING "%s: could not register %x.%02x\n",
+		       __func__, chpid.cssid, chpid.id);
+		goto out_free;
+	}
+	ret = sysfs_create_group(&chp->dev.kobj, &chp_attr_group);
+	if (ret) {
+		device_unregister(&chp->dev);
+		goto out_free;
+	}
+	mutex_lock(&css[chpid.cssid]->mutex);
+	if (css[chpid.cssid]->cm_enabled) {
+		ret = chp_add_cmg_attr(chp);
+		if (ret) {
+			sysfs_remove_group(&chp->dev.kobj, &chp_attr_group);
+			device_unregister(&chp->dev);
+			mutex_unlock(&css[chpid.cssid]->mutex);
+			goto out_free;
+		}
+	}
+	css[chpid.cssid]->chps[chpid.id] = chp;
+	mutex_unlock(&css[chpid.cssid]->mutex);
+	return ret;
+out_free:
+	kfree(chp);
+	return ret;
+}
+
+/**
+ * chp_get_chp_desc - return newly allocated channel-path description
+ * @chpid: channel-path ID
+ *
+ * On success return a newly allocated copy of the channel-path description
+ * data associated with the given channel-path ID. Return %NULL on error.
+ */
+void *chp_get_chp_desc(struct chp_id chpid)
+{
+	struct channel_path *chp;
+	struct channel_path_desc *desc;
+
+	chp = chpid_to_chp(chpid);
+	if (!chp)
+		return NULL;
+	desc = kmalloc(sizeof(struct channel_path_desc), GFP_KERNEL);
+	if (!desc)
+		return NULL;
+	memcpy(desc, &chp->desc, sizeof(struct channel_path_desc));
+	return desc;
+}
+
+/**
+ * chp_process_crw - process channel-path status change
+ * @id: channel-path ID number
+ * @status: non-zero if channel-path has become available, zero otherwise
+ *
+ * Handle channel-report-words indicating that the status of a channel-path
+ * has changed.
+ */
+int chp_process_crw(int id, int status)
+{
+	struct chp_id chpid;
+
+	chp_id_init(&chpid);
+	chpid.id = id;
+	if (status) {
+		if (!chp_is_registered(chpid))
+			chp_new(chpid);
+		return chsc_chp_online(chpid);
+	} else {
+		chsc_chp_offline(chpid);
+		return 0;
+	}
+}
Index: quilt-2.6/drivers/s390/cio/chp.h
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/cio/chp.h	2007-04-27 16:04:53.000000000 +0200
@@ -0,0 +1,37 @@
+/*
+ *  drivers/s390/cio/chp.h
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#ifndef S390_CHP_H
+#define S390_CHP_H S390_CHP_H
+
+#include <linux/types.h>
+#include <linux/device.h>
+
+#include "chpid.h"
+#include "chsc.h"
+
+struct channel_path {
+	struct chp_id chpid;
+	int state;
+	struct channel_path_desc desc;
+	/* Channel-measurement related stuff: */
+	int cmg;
+	int shared;
+	void *cmg_chars;
+	struct device dev;
+};
+
+int chp_get_status(struct chp_id chpid);
+u8 chp_get_sch_opm(struct subchannel *sch);
+int chp_is_registered(struct chp_id chpid);
+void *chp_get_chp_desc(struct chp_id chpid);
+int chp_process_crw(int id, int available);
+void chp_remove_cmg_attr(struct channel_path *chp);
+int chp_add_cmg_attr(struct channel_path *chp);
+int chp_new(struct chp_id chpid);
+
+#endif /* S390_CHP_H */
Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:53.000000000 +0200
@@ -21,54 +21,11 @@
 #include "cio_debug.h"
 #include "ioasm.h"
 #include "chpid.h"
+#include "chp.h"
 #include "chsc.h"
 
 static void *sei_page;
 
-static int new_channel_path(struct chp_id chpid);
-
-static inline struct channel_path *chpid_to_chp(struct chp_id chpid)
-{
-	return css[chpid.cssid]->chps[chpid.id];
-}
-
-static void set_chp_logically_online(struct chp_id chpid, int onoff)
-{
-	chpid_to_chp(chpid)->state = onoff;
-}
-
-static int get_chp_status(struct chp_id chpid)
-{
-	return (chpid_to_chp(chpid) ? chpid_to_chp(chpid)->state : -ENODEV);
-}
-
-void chsc_validate_chpids(struct subchannel *sch)
-{
-	int mask, chp;
-	struct chp_id chpid;
-
-	chp_id_init(&chpid);
-	for (chp = 0; chp <= 7; chp++) {
-		mask = 0x80 >> chp;
-		chpid.id = sch->schib.pmcw.chpid[chp];
-		if (!get_chp_status(chpid))
-			/* disable using this path */
-			sch->opm &= ~mask;
-	}
-}
-
-void chpid_is_actually_online(struct chp_id chpid)
-{
-	int state;
-
-	state = get_chp_status(chpid);
-	if (state < 0) {
-		need_rescan = 1;
-		queue_work(slow_path_wq, &slow_path_work);
-	} else
-		WARN_ON(!state);
-}
-
 /* FIXME: this is _always_ called for every subchannel. shouldn't we
  *	  process more than one at a time? */
 static int
@@ -214,8 +171,8 @@
 			mask = 0x80 >> j;
 			chpid.id = sch->ssd_info.chpid[j];
 			if ((sch->schib.pmcw.pim & mask) &&
-			    (get_chp_status(chpid) < 0))
-			    new_channel_path(chpid);
+			    !chp_is_registered(chpid))
+				chp_new(chpid);
 		}
 	}
 	return ret;
@@ -227,7 +184,7 @@
 	int j;
 	int mask;
 	struct subchannel *sch;
-	struct channel_path *chpid;
+	struct chp_id *chpid;
 	struct schib schib;
 
 	sch = to_subchannel(dev);
@@ -235,7 +192,7 @@
 	for (j = 0; j < 8; j++) {
 		mask = 0x80 >> j;
 		if ((sch->schib.pmcw.pim & mask) &&
-		    (sch->schib.pmcw.chpid[j] == chpid->chpid.id))
+		    (sch->schib.pmcw.chpid[j] == chpid->id))
 			break;
 	}
 	if (j >= 8)
@@ -285,51 +242,48 @@
 	return 0;
 }
 
-static void s390_set_chpid_offline(struct chp_id chpid)
+void chsc_chp_offline(struct chp_id chpid)
 {
 	char dbf_txt[15];
-	struct device *dev;
 
 	sprintf(dbf_txt, "chpr%x.%02x", chpid.cssid, chpid.id);
 	CIO_TRACE_EVENT(2, dbf_txt);
 
-	if (get_chp_status(chpid) <= 0)
+	if (chp_get_status(chpid) <= 0)
 		return;
-	dev = get_device(&(chpid_to_chp(chpid)->dev));
-	bus_for_each_dev(&css_bus_type, NULL, to_channelpath(dev),
+	bus_for_each_dev(&css_bus_type, NULL, &chpid,
 			 s390_subchannel_remove_chpid);
 
 	if (need_rescan || css_slow_subchannels_exist())
 		queue_work(slow_path_wq, &slow_path_work);
-	put_device(dev);
 }
 
 struct res_acc_data {
-	struct channel_path *chp;
+	struct chp_id chpid;
 	u32 fla_mask;
 	u16 fla;
 };
 
-static int
-s390_process_res_acc_sch(struct res_acc_data *res_data, struct subchannel *sch)
+static int s390_process_res_acc_sch(struct res_acc_data *res_data,
+				    struct subchannel *sch)
 {
 	int found;
 	int chp;
 	int ccode;
-	
+
 	found = 0;
 	for (chp = 0; chp <= 7; chp++)
 		/*
 		 * check if chpid is in information updated by ssd
 		 */
 		if (sch->ssd_info.valid &&
-		    sch->ssd_info.chpid[chp] == res_data->chp->chpid.id &&
+		    sch->ssd_info.chpid[chp] == res_data->chpid.id &&
 		    (sch->ssd_info.fla[chp] & res_data->fla_mask)
 		    == res_data->fla) {
 			found = 1;
 			break;
 		}
-	
+
 	if (found == 0)
 		return 0;
 
@@ -416,8 +370,8 @@
 	int rc;
 	char dbf_txt[15];
 
-	sprintf(dbf_txt, "accpr%x.%02x", res_data->chp->chpid.cssid,
-		res_data->chp->chpid.id);
+	sprintf(dbf_txt, "accpr%x.%02x", res_data->chpid.cssid,
+		res_data->chpid.id);
 	CIO_TRACE_EVENT( 2, dbf_txt);
 	if (res_data->fla != 0) {
 		sprintf(dbf_txt, "fla%x", res_data->fla);
@@ -503,7 +457,7 @@
 	else {
 		chp_id_init(&chpid);
 		chpid.id = id;
-		s390_set_chpid_offline(chpid);
+		chsc_chp_offline(chpid);
 	}
 
 	return 0;
@@ -512,7 +466,6 @@
 static int chsc_process_sei_res_acc(struct chsc_sei_area *sei_area)
 {
 	struct res_acc_data res_data;
-	struct device *dev;
 	struct chp_id chpid;
 	int status;
 	int rc;
@@ -524,14 +477,13 @@
 	chp_id_init(&chpid);
 	chpid.id = sei_area->rsid;
 	/* allocate a new channel path structure, if needed */
-	status = get_chp_status(chpid);
+	status = chp_get_status(chpid);
 	if (status < 0)
-		new_channel_path(chpid);
+		chp_new(chpid);
 	else if (!status)
 		return 0;
-	dev = get_device(&(chpid_to_chp(chpid)->dev));
 	memset(&res_data, 0, sizeof(struct res_acc_data));
-	res_data.chp = to_channelpath(dev);
+	res_data.chpid = chpid;
 	if ((sei_area->vf & 0xc0) != 0) {
 		res_data.fla = sei_area->fla;
 		if ((sei_area->vf & 0xc0) == 0xc0)
@@ -542,7 +494,6 @@
 			res_data.fla_mask = 0xff00;
 	}
 	rc = s390_process_res_acc(&res_data);
-	put_device(dev);
 
 	return rc;
 }
@@ -634,10 +585,10 @@
 __chp_add(struct subchannel_id schid, void *data)
 {
 	int i, mask;
-	struct channel_path *chp;
+	struct chp_id *chpid;
 	struct subchannel *sch;
 
-	chp = data;
+	chpid = data;
 	sch = get_subchannel_by_schid(schid);
 	if (!sch)
 		/* Check if the subchannel is now available. */
@@ -646,7 +597,7 @@
 	for (i=0; i<8; i++) {
 		mask = 0x80 >> i;
 		if ((sch->schib.pmcw.pim & mask) &&
-		    (sch->schib.pmcw.chpid[i] == chp->chpid.id)) {
+		    (sch->schib.pmcw.chpid[i] == chpid->id)) {
 			if (stsch(sch->schid, &sch->schib) != 0) {
 				/* Endgame. */
 				spin_unlock_irq(sch->lock);
@@ -672,52 +623,24 @@
 	return 0;
 }
 
-static int chp_add(struct chp_id chpid)
+int chsc_chp_online(struct chp_id chpid)
 {
 	int rc;
 	char dbf_txt[15];
-	struct device *dev;
 
-	if (!get_chp_status(chpid))
-		return 0; /* no need to do the rest */
-	
 	sprintf(dbf_txt, "cadd%x.%02x", chpid.cssid, chpid.id);
 	CIO_TRACE_EVENT(2, dbf_txt);
 
-	dev = get_device(&(chpid_to_chp(chpid)->dev));
-	rc = for_each_subchannel(__chp_add, to_channelpath(dev));
+	if (chp_get_status(chpid) == 0)
+		return 0;
+	rc = for_each_subchannel(__chp_add, &chpid);
 	if (css_slow_subchannels_exist())
 		rc = -EAGAIN;
 	if (rc != -EAGAIN)
 		rc = 0;
-	put_device(dev);
 	return rc;
 }
 
-/* 
- * Handling of crw machine checks with channel path source.
- */
-int chp_process_crw(int id, int on)
-{
-	struct chp_id chpid;
-
-	chp_id_init(&chpid);
-	chpid.id = id;
-	if (on == 0) {
-		/* Path has gone. We use the link incident routine.*/
-		s390_set_chpid_offline(chpid);
-		return 0; /* De-register is async anyway. */
-	}
-	/*
-	 * Path has come. Allocate a new channel path structure,
-	 * if needed.
-	 */
-	if (get_chp_status(chpid) < 0)
-		new_channel_path(chpid);
-	/* Avoid the extra overhead in process_rec_acc. */
-	return chp_add(chpid);
-}
-
 static int check_for_io_on_path(struct subchannel *sch, int index)
 {
 	int cc;
@@ -844,34 +767,13 @@
 	return 0;
 }
 
-/*
- * Function: s390_vary_chpid
- * Varies the specified chpid online or offline
+/**
+ * chsc_chp_vary - propagate channel-path vary operation to subchannels
+ * @chpid: channl-path ID
+ * @on: non-zero for vary online, zero for vary offline
  */
-static int s390_vary_chpid(struct chp_id chpid, int on)
+int chsc_chp_vary(struct chp_id chpid, int on)
 {
-	char dbf_text[15];
-	int status;
-
-	sprintf(dbf_text, on?"varyon%x.%02x":"varyoff%x.%02x", chpid.cssid,
-		chpid.id);
-	CIO_TRACE_EVENT( 2, dbf_text);
-
-	status = get_chp_status(chpid);
-	if (status < 0) {
-		printk(KERN_ERR "Can't vary unknown chpid %x.%02x\n",
-		       chpid.cssid, chpid.id);
-		return -EINVAL;
-	}
-
-	if (!on && !status) {
-		printk(KERN_ERR "chpid %x.%02x is already offline\n",
-		       chpid.cssid, chpid.id);
-		return -EINVAL;
-	}
-
-	set_chp_logically_online(chpid, on);
-
 	/*
 	 * Redo PathVerification on the devices the chpid connects to
 	 */
@@ -887,112 +789,6 @@
 	return 0;
 }
 
-/*
- * Channel measurement related functions
- */
-static ssize_t
-chp_measurement_chars_read(struct kobject *kobj, char *buf, loff_t off,
-			   size_t count)
-{
-	struct channel_path *chp;
-	unsigned int size;
-
-	chp = to_channelpath(container_of(kobj, struct device, kobj));
-	if (!chp->cmg_chars)
-		return 0;
-
-	size = sizeof(struct cmg_chars);
-
-	if (off > size)
-		return 0;
-	if (off + count > size)
-		count = size - off;
-	memcpy(buf, chp->cmg_chars + off, count);
-	return count;
-}
-
-static struct bin_attribute chp_measurement_chars_attr = {
-	.attr = {
-		.name = "measurement_chars",
-		.mode = S_IRUSR,
-		.owner = THIS_MODULE,
-	},
-	.size = sizeof(struct cmg_chars),
-	.read = chp_measurement_chars_read,
-};
-
-static void chp_measurement_copy_block(struct cmg_entry *buf,
-			struct channel_subsystem *css, struct chp_id chpid)
-{
-	void *area;
-	struct cmg_entry *entry, reference_buf;
-	int idx;
-
-	if (chpid.id < 128) {
-		area = css->cub_addr1;
-		idx = chpid.id;
-	} else {
-		area = css->cub_addr2;
-		idx = chpid.id - 128;
-	}
-	entry = area + (idx * sizeof(struct cmg_entry));
-	do {
-		memcpy(buf, entry, sizeof(*entry));
-		memcpy(&reference_buf, entry, sizeof(*entry));
-	} while (reference_buf.values[0] != buf->values[0]);
-}
-
-static ssize_t
-chp_measurement_read(struct kobject *kobj, char *buf, loff_t off, size_t count)
-{
-	struct channel_path *chp;
-	struct channel_subsystem *css;
-	unsigned int size;
-
-	chp = to_channelpath(container_of(kobj, struct device, kobj));
-	css = to_css(chp->dev.parent);
-
-	size = sizeof(struct cmg_entry);
-
-	/* Only allow single reads. */
-	if (off || count < size)
-		return 0;
-	chp_measurement_copy_block((struct cmg_entry *)buf, css, chp->chpid);
-	count = size;
-	return count;
-}
-
-static struct bin_attribute chp_measurement_attr = {
-	.attr = {
-		.name = "measurement",
-		.mode = S_IRUSR,
-		.owner = THIS_MODULE,
-	},
-	.size = sizeof(struct cmg_entry),
-	.read = chp_measurement_read,
-};
-
-static void
-chsc_remove_chp_cmg_attr(struct channel_path *chp)
-{
-	device_remove_bin_file(&chp->dev, &chp_measurement_chars_attr);
-	device_remove_bin_file(&chp->dev, &chp_measurement_attr);
-}
-
-static int
-chsc_add_chp_cmg_attr(struct channel_path *chp)
-{
-	int ret;
-
-	ret = device_create_bin_file(&chp->dev, &chp_measurement_chars_attr);
-	if (ret)
-		return ret;
-	ret = device_create_bin_file(&chp->dev, &chp_measurement_attr);
-	if (ret)
-		device_remove_bin_file(&chp->dev, &chp_measurement_chars_attr);
-	return ret;
-}
-
 static void
 chsc_remove_cmg_attr(struct channel_subsystem *css)
 {
@@ -1001,7 +797,7 @@
 	for (i = 0; i <= __MAX_CHPID; i++) {
 		if (!css->chps[i])
 			continue;
-		chsc_remove_chp_cmg_attr(css->chps[i]);
+		chp_remove_cmg_attr(css->chps[i]);
 	}
 }
 
@@ -1014,7 +810,7 @@
 	for (i = 0; i <= __MAX_CHPID; i++) {
 		if (!css->chps[i])
 			continue;
-		ret = chsc_add_chp_cmg_attr(css->chps[i]);
+		ret = chp_add_cmg_attr(css->chps[i]);
 		if (ret)
 			goto cleanup;
 	}
@@ -1023,12 +819,11 @@
 	for (--i; i >= 0; i--) {
 		if (!css->chps[i])
 			continue;
-		chsc_remove_chp_cmg_attr(css->chps[i]);
+		chp_remove_cmg_attr(css->chps[i]);
 	}
 	return ret;
 }
 
-
 static int
 __chsc_do_secm(struct channel_subsystem *css, int enable, void *page)
 {
@@ -1143,108 +938,8 @@
 	return ret;
 }
 
-/*
- * Files for the channel path entries.
- */
-static ssize_t
-chp_status_show(struct device *dev, struct device_attribute *attr, char *buf)
-{
-	struct channel_path *chp = container_of(dev, struct channel_path, dev);
-
-	if (!chp)
-		return 0;
-	return (get_chp_status(chp->chpid) ? sprintf(buf, "online\n") :
-		sprintf(buf, "offline\n"));
-}
-
-static ssize_t
-chp_status_write(struct device *dev, struct device_attribute *attr, const char *buf, size_t count)
-{
-	struct channel_path *cp = container_of(dev, struct channel_path, dev);
-	char cmd[10];
-	int num_args;
-	int error;
-
-	num_args = sscanf(buf, "%5s", cmd);
-	if (!num_args)
-		return count;
-
-	if (!strnicmp(cmd, "on", 2) || !strcmp(cmd, "1"))
-		error = s390_vary_chpid(cp->chpid, 1);
-	else if (!strnicmp(cmd, "off", 3) || !strcmp(cmd, "0"))
-		error = s390_vary_chpid(cp->chpid, 0);
-	else
-		error = -EINVAL;
-
-	return error < 0 ? error : count;
-
-}
-
-static DEVICE_ATTR(status, 0644, chp_status_show, chp_status_write);
-
-static ssize_t
-chp_type_show(struct device *dev, struct device_attribute *attr, char *buf)
-{
-	struct channel_path *chp = container_of(dev, struct channel_path, dev);
-
-	if (!chp)
-		return 0;
-	return sprintf(buf, "%x\n", chp->desc.desc);
-}
-
-static DEVICE_ATTR(type, 0444, chp_type_show, NULL);
-
-static ssize_t
-chp_cmg_show(struct device *dev, struct device_attribute *attr, char *buf)
-{
-	struct channel_path *chp = to_channelpath(dev);
-
-	if (!chp)
-		return 0;
-	if (chp->cmg == -1) /* channel measurements not available */
-		return sprintf(buf, "unknown\n");
-	return sprintf(buf, "%x\n", chp->cmg);
-}
-
-static DEVICE_ATTR(cmg, 0444, chp_cmg_show, NULL);
-
-static ssize_t
-chp_shared_show(struct device *dev, struct device_attribute *attr, char *buf)
-{
-	struct channel_path *chp = to_channelpath(dev);
-
-	if (!chp)
-		return 0;
-	if (chp->shared == -1) /* channel measurements not available */
-		return sprintf(buf, "unknown\n");
-	return sprintf(buf, "%x\n", chp->shared);
-}
-
-static DEVICE_ATTR(shared, 0444, chp_shared_show, NULL);
-
-static struct attribute * chp_attrs[] = {
-	&dev_attr_status.attr,
-	&dev_attr_type.attr,
-	&dev_attr_cmg.attr,
-	&dev_attr_shared.attr,
-	NULL,
-};
-
-static struct attribute_group chp_attr_group = {
-	.attrs = chp_attrs,
-};
-
-static void
-chp_release(struct device *dev)
-{
-	struct channel_path *cp;
-	
-	cp = container_of(dev, struct channel_path, dev);
-	kfree(cp);
-}
-
-static int chsc_determine_channel_path_description(struct chp_id chpid,
-						struct channel_path_desc *desc)
+int chsc_determine_channel_path_description(struct chp_id chpid,
+					    struct channel_path_desc *desc)
 {
 	int ccode, ret;
 
@@ -1331,8 +1026,7 @@
 	}
 }
 
-static int
-chsc_get_channel_measurement_chars(struct channel_path *chp)
+int chsc_get_channel_measurement_chars(struct channel_path *chp)
 {
 	int ccode, ret;
 
@@ -1407,97 +1101,6 @@
 	return ret;
 }
 
-/*
- * Entries for chpids on the system bus.
- * This replaces /proc/chpids.
- */
-static int new_channel_path(struct chp_id chpid)
-{
-	struct channel_path *chp;
-	int ret;
-
-	chp = kzalloc(sizeof(struct channel_path), GFP_KERNEL);
-	if (!chp)
-		return -ENOMEM;
-
-	/* fill in status, etc. */
-	chp->chpid = chpid;
-	chp->state = 1;
-	chp->dev.parent = &css[chpid.cssid]->device;
-	chp->dev.release = chp_release;
-	snprintf(chp->dev.bus_id, BUS_ID_SIZE, "chp%x.%02x", chpid.cssid,
-		 chpid.id);
-
-	/* Obtain channel path description and fill it in. */
-	ret = chsc_determine_channel_path_description(chpid, &chp->desc);
-	if (ret)
-		goto out_free;
-	/* Get channel-measurement characteristics. */
-	if (css_characteristics_avail && css_chsc_characteristics.scmc
-	    && css_chsc_characteristics.secm) {
-		ret = chsc_get_channel_measurement_chars(chp);
-		if (ret)
-			goto out_free;
-	} else {
-		static int msg_done;
-
-		if (!msg_done) {
-			printk(KERN_WARNING "cio: Channel measurements not "
-			       "available, continuing.\n");
-			msg_done = 1;
-		}
-		chp->cmg = -1;
-	}
-
-	/* make it known to the system */
-	ret = device_register(&chp->dev);
-	if (ret) {
-		printk(KERN_WARNING "%s: could not register %x.%02x\n",
-		       __func__, chpid.cssid, chpid.id);
-		goto out_free;
-	}
-	ret = sysfs_create_group(&chp->dev.kobj, &chp_attr_group);
-	if (ret) {
-		device_unregister(&chp->dev);
-		goto out_free;
-	}
-	mutex_lock(&css[chpid.cssid]->mutex);
-	if (css[chpid.cssid]->cm_enabled) {
-		ret = chsc_add_chp_cmg_attr(chp);
-		if (ret) {
-			sysfs_remove_group(&chp->dev.kobj, &chp_attr_group);
-			device_unregister(&chp->dev);
-			mutex_unlock(&css[chpid.cssid]->mutex);
-			goto out_free;
-		}
-	}
-	css[chpid.cssid]->chps[chpid.id] = chp;
-	mutex_unlock(&css[chpid.cssid]->mutex);
-	return ret;
-out_free:
-	kfree(chp);
-	return ret;
-}
-
-void *
-chsc_get_chp_desc(struct subchannel *sch, int chp_no)
-{
-	struct channel_path *chp;
-	struct channel_path_desc *desc;
-	struct chp_id chpid;
-
-	chp_id_init(&chpid);
-	chpid.id = sch->schib.pmcw.chpid[chp_no];
-	chp = chpid_to_chp(chpid);
-	if (!chp)
-		return NULL;
-	desc = kmalloc(sizeof(struct channel_path_desc), GFP_KERNEL);
-	if (!desc)
-		return NULL;
-	memcpy(desc, &chp->desc, sizeof(struct channel_path_desc));
-	return desc;
-}
-
 static int __init
 chsc_alloc_sei_area(void)
 {
Index: quilt-2.6/drivers/s390/cio/chsc.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.h	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.h	2007-04-27 16:04:53.000000000 +0200
@@ -5,10 +5,6 @@
 #include <linux/device.h>
 #include "chpid.h"
 
-#define CHSC_SEI_ACC_CHPID        1
-#define CHSC_SEI_ACC_LINKADDR     2
-#define CHSC_SEI_ACC_FULLLINKADDR 3
-
 #define CHSC_SDA_OC_MSS   0x2
 
 struct chsc_header {
@@ -37,23 +33,10 @@
 	u8 chpp;
 } __attribute__ ((packed));
 
-struct channel_path {
-	struct chp_id chpid;
-	int state;
-	struct channel_path_desc desc;
-	/* Channel-measurement related stuff: */
-	int cmg;
-	int shared;
-	void *cmg_chars;
-	struct device dev;
-};
-
-extern void s390_process_css( void );
-extern void chsc_validate_chpids(struct subchannel *);
-extern void chpid_is_actually_online(struct chp_id);
+struct channel_path;
+
 extern int css_get_ssd_info(struct subchannel *);
 extern int chsc_process_crw(void);
-extern int chp_process_crw(int, int);
 
 struct css_general_char {
 	u64 : 41;
@@ -89,12 +72,15 @@
 extern int chsc_determine_css_characteristics(void);
 extern int css_characteristics_avail;
 
-extern void *chsc_get_chp_desc(struct subchannel*, int);
-
 extern int chsc_enable_facility(int);
 struct channel_subsystem;
 extern int chsc_secm(struct channel_subsystem *, int);
 
-#define to_channelpath(device) container_of(device, struct channel_path, dev)
+int chsc_chp_vary(struct chp_id chpid, int on);
+int chsc_determine_channel_path_description(struct chp_id chpid,
+					    struct channel_path_desc *desc);
+int chsc_chp_online(struct chp_id chpid);
+void chsc_chp_offline(struct chp_id chpid);
+int chsc_get_channel_measurement_chars(struct channel_path *chp);
 
 #endif
Index: quilt-2.6/drivers/s390/cio/cio.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/cio.c	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/cio.c	2007-04-27 16:04:53.000000000 +0200
@@ -29,6 +29,7 @@
 #include "ioasm.h"
 #include "blacklist.h"
 #include "cio_debug.h"
+#include "chp.h"
 #include "../s390mach.h"
 
 debug_info_t *cio_debug_msg_id;
@@ -592,9 +593,10 @@
 		err = -ENODEV;
 		goto out;
 	}
-	sch->opm = 0xff;
-	if (!cio_is_console(sch->schid))
-		chsc_validate_chpids(sch);
+	if (cio_is_console(sch->schid))
+		sch->opm = 0xff;
+	else
+		sch->opm = chp_get_sch_opm(sch);
 	sch->lpm = sch->schib.pmcw.pam & sch->opm;
 
 	CIO_DEBUG(KERN_INFO, 0,
Index: quilt-2.6/drivers/s390/cio/css.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/css.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/css.h	2007-04-27 16:04:53.000000000 +0200
@@ -4,6 +4,7 @@
 #include <linux/mutex.h>
 #include <linux/wait.h>
 #include <linux/workqueue.h>
+#include <linux/device.h>
 
 #include <asm/cio.h>
 
Index: quilt-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:53.000000000 +0200
@@ -23,6 +23,7 @@
 #include "chsc.h"
 #include "ioasm.h"
 #include "chpid.h"
+#include "chp.h"
 
 int
 device_is_online(struct subchannel *sch)
@@ -221,7 +222,10 @@
 		if (old_lpm & mask)
 			continue;
 		chpid.id = sch->schib.pmcw.chpid[i];
-		chpid_is_actually_online(chpid);
+		if (!chp_is_registered(chpid)) {
+			need_rescan = 1;
+			queue_work(slow_path_wq, &slow_path_work);
+		}
 	}
 }
 
Index: quilt-2.6/drivers/s390/cio/device_ops.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_ops.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_ops.c	2007-04-27 16:04:53.000000000 +0200
@@ -22,6 +22,8 @@
 #include "css.h"
 #include "chsc.h"
 #include "device.h"
+#include "chpid.h"
+#include "chp.h"
 
 int ccw_device_set_options_mask(struct ccw_device *cdev, unsigned long flags)
 {
@@ -606,9 +608,12 @@
 ccw_device_get_chp_desc(struct ccw_device *cdev, int chp_no)
 {
 	struct subchannel *sch;
+	struct chp_id chpid;
 
 	sch = to_subchannel(cdev->dev.parent);
-	return chsc_get_chp_desc(sch, chp_no);
+	chp_id_init(&chpid);
+	chpid.id = sch->schib.pmcw.chpid[chp_no];
+	return chp_get_chp_desc(chpid);
 }
 
 // FIXME: these have to go:
Index: quilt-2.6/drivers/s390/cio/Makefile
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/Makefile	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/Makefile	2007-04-27 16:04:53.000000000 +0200
@@ -2,7 +2,7 @@
 # Makefile for the S/390 common i/o drivers
 #
 
-obj-y += airq.o blacklist.o chsc.o cio.o css.o
+obj-y += airq.o blacklist.o chsc.o cio.o css.o chp.o
 ccw_device-objs += device.o device_fsm.o device_ops.o
 ccw_device-objs += device_id.o device_pgid.o device_status.o
 obj-y += ccw_device.o cmf.o
Index: quilt-2.6/drivers/s390/s390mach.c
===================================================================
--- quilt-2.6.orig/drivers/s390/s390mach.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/s390mach.c	2007-04-27 16:04:53.000000000 +0200
@@ -21,6 +21,7 @@
 #include "cio/cio.h"
 #include "cio/chsc.h"
 #include "cio/css.h"
+#include "cio/chp.h"
 #include "s390mach.h"
 
 static struct semaphore m_sem;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 06/38] cio: observe chpid valid flag

[Martin Schwidefsky]

[-- Attachment #1: 104-cio-chpid-valid.diff --]
[-- Type: text/plain, Size: 1051 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Check validity flag of CHPID description data before continuing with
channel-path initialization.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/chp.c |    4 ++++
 1 files changed, 4 insertions(+)

Index: quilt-2.6/drivers/s390/cio/chp.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chp.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chp.c	2007-04-27 16:04:53.000000000 +0200
@@ -339,6 +339,10 @@
 	ret = chsc_determine_channel_path_description(chpid, &chp->desc);
 	if (ret)
 		goto out_free;
+	if ((chp->desc.flags & 0x80) == 0) {
+		ret = -ENODEV;
+		goto out_free;
+	}
 	/* Get channel-measurement characteristics. */
 	if (css_characteristics_avail && css_chsc_characteristics.scmc
 	    && css_chsc_characteristics.secm) {

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 07/38] cio: Clean up online_store.

[Martin Schwidefsky]

[-- Attachment #1: 105-cio-cleanup-online.diff --]
[-- Type: text/plain, Size: 4251 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Detangle the online_store code and make it more readable.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/device.c |  109 ++++++++++++++++++++++++++--------------------
 1 files changed, 62 insertions(+), 47 deletions(-)

Index: quilt-2.6/drivers/s390/cio/device.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device.c	2007-04-27 16:04:53.000000000 +0200
@@ -413,11 +413,60 @@
 	return (ret == 0) ? -ENODEV : ret;
 }
 
-static ssize_t
-online_store (struct device *dev, struct device_attribute *attr, const char *buf, size_t count)
+static void online_store_handle_offline(struct ccw_device *cdev)
+{
+	if (cdev->private->state == DEV_STATE_DISCONNECTED)
+		ccw_device_remove_disconnected(cdev);
+	else if (cdev->drv && cdev->drv->set_offline)
+		ccw_device_set_offline(cdev);
+}
+
+static int online_store_recog_and_online(struct ccw_device *cdev)
+{
+	int ret;
+
+	/* Do device recognition, if needed. */
+	if (cdev->id.cu_type == 0) {
+		ret = ccw_device_recognition(cdev);
+		if (ret) {
+			printk(KERN_WARNING"Couldn't start recognition "
+			       "for device %s (ret=%d)\n",
+			       cdev->dev.bus_id, ret);
+			return ret;
+		}
+		wait_event(cdev->private->wait_q,
+			   cdev->private->flags.recog_done);
+	}
+	if (cdev->drv && cdev->drv->set_online)
+		ccw_device_set_online(cdev);
+	return 0;
+}
+static void online_store_handle_online(struct ccw_device *cdev, int force)
+{
+	int ret;
+
+	ret = online_store_recog_and_online(cdev);
+	if (ret)
+		return;
+	if (force && cdev->private->state == DEV_STATE_BOXED) {
+		ret = ccw_device_stlck(cdev);
+		if (ret) {
+			printk(KERN_WARNING"ccw_device_stlck for device %s "
+			       "returned %d!\n", cdev->dev.bus_id, ret);
+			return;
+		}
+		if (cdev->id.cu_type == 0)
+			cdev->private->state = DEV_STATE_NOT_OPER;
+		online_store_recog_and_online(cdev);
+	}
+
+}
+
+static ssize_t online_store (struct device *dev, struct device_attribute *attr,
+			     const char *buf, size_t count)
 {
 	struct ccw_device *cdev = to_ccwdev(dev);
-	int i, force, ret;
+	int i, force;
 	char *tmp;
 
 	if (atomic_cmpxchg(&cdev->private->onoff, 0, 1) != 0)
@@ -434,51 +483,17 @@
 		force = 0;
 		i = simple_strtoul(buf, &tmp, 16);
 	}
-	if (i == 1) {
-		/* Do device recognition, if needed. */
-		if (cdev->id.cu_type == 0) {
-			ret = ccw_device_recognition(cdev);
-			if (ret) {
-				printk(KERN_WARNING"Couldn't start recognition "
-				       "for device %s (ret=%d)\n",
-				       cdev->dev.bus_id, ret);
-				goto out;
-			}
-			wait_event(cdev->private->wait_q,
-				   cdev->private->flags.recog_done);
-		}
-		if (cdev->drv && cdev->drv->set_online)
-			ccw_device_set_online(cdev);
-	} else if (i == 0) {
-		if (cdev->private->state == DEV_STATE_DISCONNECTED)
-			ccw_device_remove_disconnected(cdev);
-		else if (cdev->drv && cdev->drv->set_offline)
-			ccw_device_set_offline(cdev);
-	}
-	if (force && cdev->private->state == DEV_STATE_BOXED) {
-		ret = ccw_device_stlck(cdev);
-		if (ret) {
-			printk(KERN_WARNING"ccw_device_stlck for device %s "
-			       "returned %d!\n", cdev->dev.bus_id, ret);
-			goto out;
-		}
-		/* Do device recognition, if needed. */
-		if (cdev->id.cu_type == 0) {
-			cdev->private->state = DEV_STATE_NOT_OPER;
-			ret = ccw_device_recognition(cdev);
-			if (ret) {
-				printk(KERN_WARNING"Couldn't start recognition "
-				       "for device %s (ret=%d)\n",
-				       cdev->dev.bus_id, ret);
-				goto out;
-			}
-			wait_event(cdev->private->wait_q,
-				   cdev->private->flags.recog_done);
-		}
-		if (cdev->drv && cdev->drv->set_online)
-			ccw_device_set_online(cdev);
+
+	switch (i) {
+	case 0:
+		online_store_handle_offline(cdev);
+		break;
+	case 1:
+		online_store_handle_online(cdev, force);
+		break;
+	default:
+		count = -EINVAL;
 	}
-	out:
 	if (cdev->drv)
 		module_put(cdev->drv->owner);
 	atomic_set(&cdev->private->onoff, 0);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 08/38] cio: Channel-path configure function.

[Martin Schwidefsky]

[-- Attachment #1: 106-cio-chp-configure.diff --]
[-- Type: text/plain, Size: 24921 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Add a new attribute to the channel-path sysfs directory through which
channel-path configure operations can be triggered. Also listen for
hardware events requesting channel-path configure operations and
process them accordingly.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 drivers/s390/char/Makefile    |    2 
 drivers/s390/char/sclp_chp.c  |  196 ++++++++++++++++++++++++++++++++
 drivers/s390/cio/chp.c        |  250 +++++++++++++++++++++++++++++++++++++++++-
 drivers/s390/cio/chp.h        |   20 +++
 drivers/s390/cio/chpid.h      |   51 --------
 drivers/s390/cio/chsc.c       |   44 +++++++
 drivers/s390/cio/chsc.h       |    2 
 drivers/s390/cio/cio.c        |    1 
 drivers/s390/cio/cio.h        |    1 
 drivers/s390/cio/css.h        |    4 
 drivers/s390/cio/device_fsm.c |    2 
 drivers/s390/cio/device_ops.c |    2 
 drivers/s390/cio/ioasm.h      |    2 
 include/asm-s390/chpid.h      |   53 ++++++++
 include/asm-s390/cio.h        |    1 
 include/asm-s390/sclp.h       |   12 ++
 16 files changed, 581 insertions(+), 62 deletions(-)

Index: quilt-2.6/drivers/s390/char/Makefile
===================================================================
--- quilt-2.6.orig/drivers/s390/char/Makefile	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/Makefile	2007-04-27 16:04:54.000000000 +0200
@@ -3,7 +3,7 @@
 #
 
 obj-y += ctrlchar.o keyboard.o defkeymap.o sclp.o sclp_rw.o sclp_quiesce.o \
-	 sclp_info.o
+	 sclp_info.o sclp_chp.o
 
 obj-$(CONFIG_TN3270) += raw3270.o
 obj-$(CONFIG_TN3270_CONSOLE) += con3270.o
Index: quilt-2.6/drivers/s390/char/sclp_chp.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/char/sclp_chp.c	2007-04-27 16:04:54.000000000 +0200
@@ -0,0 +1,196 @@
+/*
+ *  drivers/s390/char/sclp_chp.c
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#include <linux/types.h>
+#include <linux/gfp.h>
+#include <linux/errno.h>
+#include <linux/completion.h>
+#include <asm/sclp.h>
+#include <asm/chpid.h>
+
+#include "sclp.h"
+
+#define TAG	"sclp_chp: "
+
+#define SCLP_CMDW_CONFIGURE_CHANNEL_PATH	0x000f0001
+#define SCLP_CMDW_DECONFIGURE_CHANNEL_PATH	0x000e0001
+#define SCLP_CMDW_READ_CHANNEL_PATH_INFORMATION	0x00030001
+
+static inline sclp_cmdw_t get_configure_cmdw(struct chp_id chpid)
+{
+	return SCLP_CMDW_CONFIGURE_CHANNEL_PATH | chpid.id << 8;
+}
+
+static inline sclp_cmdw_t get_deconfigure_cmdw(struct chp_id chpid)
+{
+	return SCLP_CMDW_DECONFIGURE_CHANNEL_PATH | chpid.id << 8;
+}
+
+static void chp_callback(struct sclp_req *req, void *data)
+{
+	struct completion *completion = data;
+
+	complete(completion);
+}
+
+struct chp_cfg_sccb {
+	struct sccb_header header;
+	u8 ccm;
+	u8 reserved[6];
+	u8 cssid;
+} __attribute__((packed));
+
+struct chp_cfg_data {
+	struct chp_cfg_sccb sccb;
+	struct sclp_req req;
+	struct completion completion;
+} __attribute__((packed));
+
+static int do_configure(sclp_cmdw_t cmd)
+{
+	struct chp_cfg_data *data;
+	int rc;
+
+	/* Prepare sccb. */
+	data = (struct chp_cfg_data *) get_zeroed_page(GFP_KERNEL | GFP_DMA);
+	if (!data)
+		return -ENOMEM;
+	data->sccb.header.length = sizeof(struct chp_cfg_sccb);
+	data->req.command = cmd;
+	data->req.sccb = &(data->sccb);
+	data->req.status = SCLP_REQ_FILLED;
+	data->req.callback = chp_callback;
+	data->req.callback_data = &(data->completion);
+	init_completion(&data->completion);
+
+	/* Perform sclp request. */
+	rc = sclp_add_request(&(data->req));
+	if (rc)
+		goto out;
+	wait_for_completion(&data->completion);
+
+	/* Check response .*/
+	if (data->req.status != SCLP_REQ_DONE) {
+		printk(KERN_WARNING TAG "configure channel-path request failed "
+		       "(status=0x%02x)\n", data->req.status);
+		rc = -EIO;
+		goto out;
+	}
+	switch (data->sccb.header.response_code) {
+	case 0x0020:
+	case 0x0120:
+	case 0x0440:
+	case 0x0450:
+		break;
+	default:
+		printk(KERN_WARNING TAG "configure channel-path failed "
+		       "(cmd=0x%08x, response=0x%04x)\n", cmd,
+		       data->sccb.header.response_code);
+		rc = -EIO;
+		break;
+	}
+out:
+	free_page((unsigned long) data);
+
+	return rc;
+}
+
+/**
+ * sclp_chp_configure - perform configure channel-path sclp command
+ * @chpid: channel-path ID
+ *
+ * Perform configure channel-path command sclp command for specified chpid.
+ * Return 0 after command successfully finished, non-zero otherwise.
+ */
+int sclp_chp_configure(struct chp_id chpid)
+{
+	return do_configure(get_configure_cmdw(chpid));
+}
+
+/**
+ * sclp_chp_deconfigure - perform deconfigure channel-path sclp command
+ * @chpid: channel-path ID
+ *
+ * Perform deconfigure channel-path command sclp command for specified chpid
+ * and wait for completion. On success return 0. Return non-zero otherwise.
+ */
+int sclp_chp_deconfigure(struct chp_id chpid)
+{
+	return do_configure(get_deconfigure_cmdw(chpid));
+}
+
+struct chp_info_sccb {
+	struct sccb_header header;
+	u8 recognized[SCLP_CHP_INFO_MASK_SIZE];
+	u8 standby[SCLP_CHP_INFO_MASK_SIZE];
+	u8 configured[SCLP_CHP_INFO_MASK_SIZE];
+	u8 ccm;
+	u8 reserved[6];
+	u8 cssid;
+} __attribute__((packed));
+
+struct chp_info_data {
+	struct chp_info_sccb sccb;
+	struct sclp_req req;
+	struct completion completion;
+} __attribute__((packed));
+
+/**
+ * sclp_chp_read_info - perform read channel-path information sclp command
+ * @info: resulting channel-path information data
+ *
+ * Perform read channel-path information sclp command and wait for completion.
+ * On success, store channel-path information in @info and return 0. Return
+ * non-zero otherwise.
+ */
+int sclp_chp_read_info(struct sclp_chp_info *info)
+{
+	struct chp_info_data *data;
+	int rc;
+
+	/* Prepare sccb. */
+	data = (struct chp_info_data *) get_zeroed_page(GFP_KERNEL | GFP_DMA);
+	if (!data)
+		return -ENOMEM;
+	data->sccb.header.length = sizeof(struct chp_info_sccb);
+	data->req.command = SCLP_CMDW_READ_CHANNEL_PATH_INFORMATION;
+	data->req.sccb = &(data->sccb);
+	data->req.status = SCLP_REQ_FILLED;
+	data->req.callback = chp_callback;
+	data->req.callback_data = &(data->completion);
+	init_completion(&data->completion);
+
+	/* Perform sclp request. */
+	rc = sclp_add_request(&(data->req));
+	if (rc)
+		goto out;
+	wait_for_completion(&data->completion);
+
+	/* Check response .*/
+	if (data->req.status != SCLP_REQ_DONE) {
+		printk(KERN_WARNING TAG "read channel-path info request failed "
+		       "(status=0x%02x)\n", data->req.status);
+		rc = -EIO;
+		goto out;
+	}
+	if (data->sccb.header.response_code != 0x0010) {
+		printk(KERN_WARNING TAG "read channel-path info failed "
+		       "(response=0x%04x)\n", data->sccb.header.response_code);
+		rc = -EIO;
+		goto out;
+	}
+	memcpy(info->recognized, data->sccb.recognized,
+	       SCLP_CHP_INFO_MASK_SIZE);
+	memcpy(info->standby, data->sccb.standby,
+	       SCLP_CHP_INFO_MASK_SIZE);
+	memcpy(info->configured, data->sccb.configured,
+	       SCLP_CHP_INFO_MASK_SIZE);
+out:
+	free_page((unsigned long) data);
+
+	return rc;
+}
Index: quilt-2.6/drivers/s390/cio/chp.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chp.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chp.c	2007-04-27 16:04:54.000000000 +0200
@@ -10,9 +10,14 @@
 #include <linux/bug.h>
 #include <linux/workqueue.h>
 #include <linux/spinlock.h>
+#include <linux/init.h>
+#include <linux/jiffies.h>
+#include <linux/wait.h>
+#include <linux/mutex.h>
 #include <asm/errno.h>
+#include <asm/chpid.h>
+#include <asm/sclp.h>
 
-#include "chpid.h"
 #include "cio.h"
 #include "css.h"
 #include "ioasm.h"
@@ -20,6 +25,32 @@
 #include "chp.h"
 
 #define to_channelpath(device) container_of(device, struct channel_path, dev)
+#define CHP_INFO_UPDATE_INTERVAL	1*HZ
+
+enum cfg_task_t {
+	cfg_none,
+	cfg_configure,
+	cfg_deconfigure
+};
+
+/* Map for pending configure tasks. */
+static enum cfg_task_t chp_cfg_task[__MAX_CSSID + 1][__MAX_CHPID + 1];
+static DEFINE_MUTEX(cfg_lock);
+static int cfg_busy;
+
+/* Map for channel-path status. */
+static struct sclp_chp_info chp_info;
+static DEFINE_MUTEX(info_lock);
+
+/* Time after which channel-path status may be outdated. */
+static unsigned long chp_info_expires;
+
+/* Workqueue to perform pending configure tasks. */
+static struct workqueue_struct *chp_wq;
+static struct work_struct cfg_work;
+
+/* Wait queue for configure completion events. */
+static wait_queue_head_t cfg_wait_queue;
 
 /* Return channel_path struct for given chpid. */
 static inline struct channel_path *chpid_to_chp(struct chp_id chpid)
@@ -251,6 +282,43 @@
 
 static DEVICE_ATTR(status, 0644, chp_status_show, chp_status_write);
 
+static ssize_t chp_configure_show(struct device *dev,
+				  struct device_attribute *attr, char *buf)
+{
+	struct channel_path *cp;
+	int status;
+
+	cp = container_of(dev, struct channel_path, dev);
+	status = chp_info_get_status(cp->chpid);
+	if (status < 0)
+		return status;
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", status);
+}
+
+static int cfg_wait_idle(void);
+
+static ssize_t chp_configure_write(struct device *dev,
+				   struct device_attribute *attr,
+				   const char *buf, size_t count)
+{
+	struct channel_path *cp;
+	int val;
+	char delim;
+
+	if (sscanf(buf, "%d %c", &val, &delim) != 1)
+		return -EINVAL;
+	if (val != 0 && val != 1)
+		return -EINVAL;
+	cp = container_of(dev, struct channel_path, dev);
+	chp_cfg_schedule(cp->chpid, val);
+	cfg_wait_idle();
+
+	return count;
+}
+
+static DEVICE_ATTR(configure, 0644, chp_configure_show, chp_configure_write);
+
 static ssize_t chp_type_show(struct device *dev, struct device_attribute *attr,
 			     char *buf)
 {
@@ -293,6 +361,7 @@
 
 static struct attribute * chp_attrs[] = {
 	&dev_attr_status.attr,
+	&dev_attr_configure.attr,
 	&dev_attr_type.attr,
 	&dev_attr_cmg.attr,
 	&dev_attr_shared.attr,
@@ -323,6 +392,8 @@
 	struct channel_path *chp;
 	int ret;
 
+	if (chp_is_registered(chpid))
+		return 0;
 	chp = kzalloc(sizeof(struct channel_path), GFP_KERNEL);
 	if (!chp)
 		return -ENOMEM;
@@ -435,3 +506,180 @@
 		return 0;
 	}
 }
+
+static inline int info_bit_num(struct chp_id id)
+{
+	return id.id + id.cssid * (__MAX_CHPID + 1);
+}
+
+/* Force chp_info refresh on next call to info_validate(). */
+static void info_expire(void)
+{
+	mutex_lock(&info_lock);
+	chp_info_expires = jiffies - 1;
+	mutex_unlock(&info_lock);
+}
+
+/* Ensure that chp_info is up-to-date. */
+static int info_update(void)
+{
+	int rc;
+
+	mutex_lock(&info_lock);
+	rc = 0;
+	if (time_after(jiffies, chp_info_expires)) {
+		/* Data is too old, update. */
+		rc = sclp_chp_read_info(&chp_info);
+		chp_info_expires = jiffies + CHP_INFO_UPDATE_INTERVAL ;
+	}
+	mutex_unlock(&info_lock);
+
+	return rc;
+}
+
+/**
+ * chp_info_get_status - retrieve configure status of a channel-path
+ * @chpid: channel-path ID
+ *
+ * On success, return 0 for standby, 1 for configured, 2 for reserved,
+ * 3 for not recognized. Return negative error code on error.
+ */
+int chp_info_get_status(struct chp_id chpid)
+{
+	int rc;
+	int bit;
+
+	rc = info_update();
+	if (rc)
+		return rc;
+
+	bit = info_bit_num(chpid);
+	mutex_lock(&info_lock);
+	if (!chp_test_bit(chp_info.recognized, bit))
+		rc = CHP_STATUS_NOT_RECOGNIZED;
+	else if (chp_test_bit(chp_info.configured, bit))
+		rc = CHP_STATUS_CONFIGURED;
+	else if (chp_test_bit(chp_info.standby, bit))
+		rc = CHP_STATUS_STANDBY;
+	else
+		rc = CHP_STATUS_RESERVED;
+	mutex_unlock(&info_lock);
+
+	return rc;
+}
+
+/* Return configure task for chpid. */
+static enum cfg_task_t cfg_get_task(struct chp_id chpid)
+{
+	return chp_cfg_task[chpid.cssid][chpid.id];
+}
+
+/* Set configure task for chpid. */
+static void cfg_set_task(struct chp_id chpid, enum cfg_task_t cfg)
+{
+	chp_cfg_task[chpid.cssid][chpid.id] = cfg;
+}
+
+/* Perform one configure/deconfigure request. Reschedule work function until
+ * last request. */
+static void cfg_func(struct work_struct *work)
+{
+	struct chp_id chpid;
+	enum cfg_task_t t;
+
+	mutex_lock(&cfg_lock);
+	t = cfg_none;
+	chp_id_for_each(&chpid) {
+		t = cfg_get_task(chpid);
+		if (t != cfg_none) {
+			cfg_set_task(chpid, cfg_none);
+			break;
+		}
+	}
+	mutex_unlock(&cfg_lock);
+
+	switch (t) {
+	case cfg_configure:
+		sclp_chp_configure(chpid);
+		info_expire();
+		chsc_chp_online(chpid);
+		break;
+	case cfg_deconfigure:
+		sclp_chp_deconfigure(chpid);
+		info_expire();
+		chsc_chp_offline(chpid);
+		break;
+	case cfg_none:
+		/* Get updated information after last change. */
+		info_update();
+		mutex_lock(&cfg_lock);
+		cfg_busy = 0;
+		mutex_unlock(&cfg_lock);
+		wake_up_interruptible(&cfg_wait_queue);
+		return;
+	}
+	queue_work(chp_wq, &cfg_work);
+}
+
+/**
+ * chp_cfg_schedule - schedule chpid configuration request
+ * @chpid - channel-path ID
+ * @configure - Non-zero for configure, zero for deconfigure
+ *
+ * Schedule a channel-path configuration/deconfiguration request.
+ */
+void chp_cfg_schedule(struct chp_id chpid, int configure)
+{
+	CIO_MSG_EVENT(2, "chp_cfg_sched%x.%02x=%d\n", chpid.cssid, chpid.id,
+		      configure);
+	mutex_lock(&cfg_lock);
+	cfg_set_task(chpid, configure ? cfg_configure : cfg_deconfigure);
+	cfg_busy = 1;
+	mutex_unlock(&cfg_lock);
+	queue_work(chp_wq, &cfg_work);
+}
+
+/**
+ * chp_cfg_cancel_deconfigure - cancel chpid deconfiguration request
+ * @chpid - channel-path ID
+ *
+ * Cancel an active channel-path deconfiguration request if it has not yet
+ * been performed.
+ */
+void chp_cfg_cancel_deconfigure(struct chp_id chpid)
+{
+	CIO_MSG_EVENT(2, "chp_cfg_cancel:%x.%02x\n", chpid.cssid, chpid.id);
+	mutex_lock(&cfg_lock);
+	if (cfg_get_task(chpid) == cfg_deconfigure)
+		cfg_set_task(chpid, cfg_none);
+	mutex_unlock(&cfg_lock);
+}
+
+static int cfg_wait_idle(void)
+{
+	if (wait_event_interruptible(cfg_wait_queue, !cfg_busy))
+		return -ERESTARTSYS;
+	return 0;
+}
+
+static int __init chp_init(void)
+{
+	struct chp_id chpid;
+
+	chp_wq = create_singlethread_workqueue("cio_chp");
+	if (!chp_wq)
+		return -ENOMEM;
+	INIT_WORK(&cfg_work, cfg_func);
+	init_waitqueue_head(&cfg_wait_queue);
+	if (info_update())
+		return 0;
+	/* Register available channel-paths. */
+	chp_id_for_each(&chpid) {
+		if (chp_info_get_status(chpid) != CHP_STATUS_NOT_RECOGNIZED)
+			chp_new(chpid);
+	}
+
+	return 0;
+}
+
+subsys_initcall(chp_init);
Index: quilt-2.6/drivers/s390/cio/chp.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chp.h	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chp.h	2007-04-27 16:04:54.000000000 +0200
@@ -10,10 +10,23 @@
 
 #include <linux/types.h>
 #include <linux/device.h>
-
-#include "chpid.h"
+#include <asm/chpid.h>
 #include "chsc.h"
 
+#define CHP_STATUS_STANDBY		0
+#define CHP_STATUS_CONFIGURED		1
+#define CHP_STATUS_RESERVED		2
+#define CHP_STATUS_NOT_RECOGNIZED	3
+
+static inline int chp_test_bit(u8 *bitmap, int num)
+{
+	int byte = num >> 3;
+	int mask = 128 >> (num & 7);
+
+	return (bitmap[byte] & mask) ? 1 : 0;
+}
+
+
 struct channel_path {
 	struct chp_id chpid;
 	int state;
@@ -33,5 +46,8 @@
 void chp_remove_cmg_attr(struct channel_path *chp);
 int chp_add_cmg_attr(struct channel_path *chp);
 int chp_new(struct chp_id chpid);
+void chp_cfg_schedule(struct chp_id chpid, int configure);
+void chp_cfg_cancel_deconfigure(struct chp_id chpid);
+int chp_info_get_status(struct chp_id chpid);
 
 #endif /* S390_CHP_H */
Index: quilt-2.6/drivers/s390/cio/chpid.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chpid.h	2007-04-27 16:04:52.000000000 +0200
+++ /dev/null	1970-01-01 00:00:00.000000000 +0000
@@ -1,51 +0,0 @@
-/*
- *  drivers/s390/cio/chpid.h
- *
- *    Copyright IBM Corp. 2007
- *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
- */
-
-#ifndef S390_CHP_ID_H
-#define S390_CHP_ID_H S390_CHP_ID_H
-
-#include <linux/string.h>
-#include <asm/types.h>
-#include "css.h"
-
-struct chp_id {
-	u8 reserved1;
-	u8 cssid;
-	u8 reserved2;
-	u8 id;
-} __attribute__((packed));
-
-static inline void chp_id_init(struct chp_id *chpid)
-{
-	memset(chpid, 0, sizeof(struct chp_id));
-}
-
-static inline int chp_id_is_equal(struct chp_id *a, struct chp_id *b)
-{
-	return (a->id == b->id) && (a->cssid == b->cssid);
-}
-
-static inline void chp_id_next(struct chp_id *chpid)
-{
-	if (chpid->id < __MAX_CHPID)
-		chpid->id++;
-	else {
-		chpid->id = 0;
-		chpid->cssid++;
-	}
-}
-
-static inline int chp_id_is_valid(struct chp_id *chpid)
-{
-	return (chpid->cssid <= __MAX_CSSID);
-}
-
-
-#define chp_id_for_each(c) \
-	for (chp_id_init(c); chp_id_is_valid(c); chp_id_next(c))
-
-#endif /* S390_CHP_ID_H */
Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:54.000000000 +0200
@@ -15,12 +15,12 @@
 #include <linux/device.h>
 
 #include <asm/cio.h>
+#include <asm/chpid.h>
 
 #include "css.h"
 #include "cio.h"
 #include "cio_debug.h"
 #include "ioasm.h"
-#include "chpid.h"
 #include "chp.h"
 #include "chsc.h"
 
@@ -498,6 +498,45 @@
 	return rc;
 }
 
+struct chp_config_data {
+	u8 map[32];
+	u8 op;
+	u8 pc;
+};
+
+static int chsc_process_sei_chp_config(struct chsc_sei_area *sei_area)
+{
+	struct chp_config_data *data;
+	struct chp_id chpid;
+	int num;
+
+	CIO_CRW_EVENT(4, "chsc: channel-path-configuration notification\n");
+	if (sei_area->rs != 0)
+		return 0;
+	data = (struct chp_config_data *) &(sei_area->ccdf);
+	chp_id_init(&chpid);
+	for (num = 0; num <= __MAX_CHPID; num++) {
+		if (!chp_test_bit(data->map, num))
+			continue;
+		chpid.id = num;
+		printk(KERN_WARNING "cio: processing configure event %d for "
+		       "chpid %x.%02x\n", data->op, chpid.cssid, chpid.id);
+		switch (data->op) {
+		case 0:
+			chp_cfg_schedule(chpid, 1);
+			break;
+		case 1:
+			chp_cfg_schedule(chpid, 0);
+			break;
+		case 2:
+			chp_cfg_cancel_deconfigure(chpid);
+			break;
+		}
+	}
+
+	return 0;
+}
+
 static int chsc_process_sei(struct chsc_sei_area *sei_area)
 {
 	int rc;
@@ -514,6 +553,9 @@
 	case 2: /* i/o resource accessibiliy */
 		rc = chsc_process_sei_res_acc(sei_area);
 		break;
+	case 8: /* channel-path-configuration notification */
+		rc = chsc_process_sei_chp_config(sei_area);
+		break;
 	default: /* other stuff */
 		CIO_CRW_EVENT(4, "chsc: unhandled sei content code %d\n",
 			      sei_area->cc);
Index: quilt-2.6/drivers/s390/cio/chsc.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.h	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.h	2007-04-27 16:04:54.000000000 +0200
@@ -3,7 +3,7 @@
 
 #include <linux/types.h>
 #include <linux/device.h>
-#include "chpid.h"
+#include <asm/chpid.h>
 
 #define CHSC_SDA_OC_MSS   0x2
 
Index: quilt-2.6/drivers/s390/cio/cio.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/cio.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/cio.c	2007-04-27 16:04:54.000000000 +0200
@@ -22,6 +22,7 @@
 #include <asm/setup.h>
 #include <asm/reset.h>
 #include <asm/ipl.h>
+#include <asm/chpid.h>
 #include "airq.h"
 #include "cio.h"
 #include "css.h"
Index: quilt-2.6/drivers/s390/cio/cio.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/cio.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/cio.h	2007-04-27 16:04:54.000000000 +0200
@@ -3,6 +3,7 @@
 
 #include "schid.h"
 #include <linux/mutex.h>
+#include <linux/device.h>
 
 /*
  * where we put the ssd info
Index: quilt-2.6/drivers/s390/cio/css.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/css.h	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/css.h	2007-04-27 16:04:54.000000000 +0200
@@ -5,8 +5,10 @@
 #include <linux/wait.h>
 #include <linux/workqueue.h>
 #include <linux/device.h>
+#include <linux/types.h>
 
 #include <asm/cio.h>
+#include <asm/chpid.h>
 
 #include "schid.h"
 
@@ -149,8 +151,6 @@
 
 #define __MAX_SUBCHANNEL 65535
 #define __MAX_SSID 3
-#define __MAX_CHPID 255
-#define __MAX_CSSID 0
 
 struct channel_subsystem {
 	u8 cssid;
Index: quilt-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:54.000000000 +0200
@@ -15,6 +15,7 @@
 
 #include <asm/ccwdev.h>
 #include <asm/cio.h>
+#include <asm/chpid.h>
 
 #include "cio.h"
 #include "cio_debug.h"
@@ -22,7 +23,6 @@
 #include "device.h"
 #include "chsc.h"
 #include "ioasm.h"
-#include "chpid.h"
 #include "chp.h"
 
 int
Index: quilt-2.6/drivers/s390/cio/device_ops.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_ops.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_ops.c	2007-04-27 16:04:54.000000000 +0200
@@ -16,13 +16,13 @@
 
 #include <asm/ccwdev.h>
 #include <asm/idals.h>
+#include <asm/chpid.h>
 
 #include "cio.h"
 #include "cio_debug.h"
 #include "css.h"
 #include "chsc.h"
 #include "device.h"
-#include "chpid.h"
 #include "chp.h"
 
 int ccw_device_set_options_mask(struct ccw_device *cdev, unsigned long flags)
Index: quilt-2.6/drivers/s390/cio/ioasm.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/ioasm.h	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/ioasm.h	2007-04-27 16:04:54.000000000 +0200
@@ -1,8 +1,8 @@
 #ifndef S390_CIO_IOASM_H
 #define S390_CIO_IOASM_H
 
+#include <asm/chpid.h>
 #include "schid.h"
-#include "chpid.h"
 
 /*
  * TPI info structure
Index: quilt-2.6/include/asm-s390/chpid.h
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/include/asm-s390/chpid.h	2007-04-27 16:04:54.000000000 +0200
@@ -0,0 +1,53 @@
+/*
+ *  drivers/s390/cio/chpid.h
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#ifndef _ASM_S390_CHPID_H
+#define _ASM_S390_CHPID_H _ASM_S390_CHPID_H
+
+#include <linux/string.h>
+#include <asm/types.h>
+#include <asm/cio.h>
+
+#define __MAX_CHPID 255
+
+struct chp_id {
+	u8 reserved1;
+	u8 cssid;
+	u8 reserved2;
+	u8 id;
+} __attribute__((packed));
+
+static inline void chp_id_init(struct chp_id *chpid)
+{
+	memset(chpid, 0, sizeof(struct chp_id));
+}
+
+static inline int chp_id_is_equal(struct chp_id *a, struct chp_id *b)
+{
+	return (a->id == b->id) && (a->cssid == b->cssid);
+}
+
+static inline void chp_id_next(struct chp_id *chpid)
+{
+	if (chpid->id < __MAX_CHPID)
+		chpid->id++;
+	else {
+		chpid->id = 0;
+		chpid->cssid++;
+	}
+}
+
+static inline int chp_id_is_valid(struct chp_id *chpid)
+{
+	return (chpid->cssid <= __MAX_CSSID);
+}
+
+
+#define chp_id_for_each(c) \
+	for (chp_id_init(c); chp_id_is_valid(c); chp_id_next(c))
+
+#endif /* _ASM_S390_CHPID_H */
Index: quilt-2.6/include/asm-s390/cio.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/cio.h	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/include/asm-s390/cio.h	2007-04-27 16:04:54.000000000 +0200
@@ -13,6 +13,7 @@
 #ifdef __KERNEL__
 
 #define LPM_ANYPATH 0xff
+#define __MAX_CSSID 0
 
 /*
  * subchannel status word
Index: quilt-2.6/include/asm-s390/sclp.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/sclp.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/sclp.h	2007-04-27 16:04:54.000000000 +0200
@@ -9,6 +9,7 @@
 #define _ASM_S390_SCLP_H
 
 #include <linux/types.h>
+#include <asm/chpid.h>
 
 struct sccb_header {
 	u16	length;
@@ -33,7 +34,18 @@
 	u8	_reserved3[4096 - 112];	/* 112-4095 */
 } __attribute__((packed, aligned(4096)));
 
+#define SCLP_CHP_INFO_MASK_SIZE		32
+
+struct sclp_chp_info {
+	u8 recognized[SCLP_CHP_INFO_MASK_SIZE];
+	u8 standby[SCLP_CHP_INFO_MASK_SIZE];
+	u8 configured[SCLP_CHP_INFO_MASK_SIZE];
+};
+
 extern struct sclp_readinfo_sccb s390_readinfo_sccb;
 extern void sclp_readinfo_early(void);
+extern int sclp_chp_configure(struct chp_id chpid);
+extern int sclp_chp_deconfigure(struct chp_id chpid);
+extern int sclp_chp_read_info(struct sclp_chp_info *info);
 
 #endif /* _ASM_S390_SCLP_H */

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 09/38] cio: Use add_uevent_var.

[Martin Schwidefsky]

[-- Attachment #1: 107-cio-adduevent.diff --]
[-- Type: text/plain, Size: 3778 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Convert ccw_uevent to use add_uevent_var and adapt snprint_alias.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/device.c |   66 ++++++++++++++++++----------------------------
 1 files changed, 26 insertions(+), 40 deletions(-)

Index: quilt-2.6/drivers/s390/cio/device.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device.c	2007-04-27 16:04:54.000000000 +0200
@@ -56,13 +56,12 @@
 /* Store modalias string delimited by prefix/suffix string into buffer with
  * specified size. Return length of resulting string (excluding trailing '\0')
  * even if string doesn't fit buffer (snprintf semantics). */
-static int snprint_alias(char *buf, size_t size, const char *prefix,
+static int snprint_alias(char *buf, size_t size,
 			 struct ccw_device_id *id, const char *suffix)
 {
 	int len;
 
-	len = snprintf(buf, size, "%sccw:t%04Xm%02X", prefix, id->cu_type,
-		       id->cu_model);
+	len = snprintf(buf, size, "ccw:t%04Xm%02X", id->cu_type, id->cu_model);
 	if (len > size)
 		return len;
 	buf += len;
@@ -85,53 +84,40 @@
 	struct ccw_device *cdev = to_ccwdev(dev);
 	struct ccw_device_id *id = &(cdev->id);
 	int i = 0;
-	int len;
+	int len = 0;
+	int ret;
+	char modalias_buf[30];
 
 	/* CU_TYPE= */
-	len = snprintf(buffer, buffer_size, "CU_TYPE=%04X", id->cu_type) + 1;
-	if (len > buffer_size || i >= num_envp)
-		return -ENOMEM;
-	envp[i++] = buffer;
-	buffer += len;
-	buffer_size -= len;
+	ret = add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &len,
+			     "CU_TYPE=%04X", id->cu_type);
+	if (ret)
+		return ret;
 
 	/* CU_MODEL= */
-	len = snprintf(buffer, buffer_size, "CU_MODEL=%02X", id->cu_model) + 1;
-	if (len > buffer_size || i >= num_envp)
-		return -ENOMEM;
-	envp[i++] = buffer;
-	buffer += len;
-	buffer_size -= len;
+	ret = add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &len,
+			     "CU_MODEL=%02X", id->cu_model);
+	if (ret)
+		return ret;
 
 	/* The next two can be zero, that's ok for us */
 	/* DEV_TYPE= */
-	len = snprintf(buffer, buffer_size, "DEV_TYPE=%04X", id->dev_type) + 1;
-	if (len > buffer_size || i >= num_envp)
-		return -ENOMEM;
-	envp[i++] = buffer;
-	buffer += len;
-	buffer_size -= len;
+	ret = add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &len,
+			     "DEV_TYPE=%04X", id->dev_type);
+	if (ret)
+		return ret;
 
 	/* DEV_MODEL= */
-	len = snprintf(buffer, buffer_size, "DEV_MODEL=%02X",
-			(unsigned char) id->dev_model) + 1;
-	if (len > buffer_size || i >= num_envp)
-		return -ENOMEM;
-	envp[i++] = buffer;
-	buffer += len;
-	buffer_size -= len;
+	ret = add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &len,
+			     "DEV_MODEL=%02X", id->dev_model);
+	if (ret)
+		return ret;
 
 	/* MODALIAS=  */
-	len = snprint_alias(buffer, buffer_size, "MODALIAS=", id, "") + 1;
-	if (len > buffer_size || i >= num_envp)
-		return -ENOMEM;
-	envp[i++] = buffer;
-	buffer += len;
-	buffer_size -= len;
-
-	envp[i] = NULL;
-
-	return 0;
+	snprint_alias(modalias_buf, sizeof(modalias_buf), id, "");
+	ret = add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &len,
+			     "MODALIAS=%s", modalias_buf);
+	return ret;
 }
 
 struct bus_type ccw_bus_type;
@@ -280,7 +266,7 @@
 	struct ccw_device_id *id = &(cdev->id);
 	int len;
 
-	len = snprint_alias(buf, PAGE_SIZE, "", id, "\n") + 1;
+	len = snprint_alias(buf, PAGE_SIZE, id, "\n") + 1;
 
 	return len > PAGE_SIZE ? PAGE_SIZE : len;
 }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 10/38] cio: Re-start path verification after aborting internal I/O.

[Martin Schwidefsky]

[-- Attachment #1: 108-cio-path-verify.diff --]
[-- Type: text/plain, Size: 4711 bytes --]

From: Stefan Bader <shbader@de.ibm.com>

Path verification triggered by changes to the available CHPIDs will be
interrupted by another change but not re-started. This results in an
invalid path mask.
To solve this make sure to completely re-start path verification when
changing the available paths.

Signed-off-by: Stefan Bader <shbader@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/chsc.c |  108 ++++++++++++++++++++++++------------------------
 1 files changed, 54 insertions(+), 54 deletions(-)

Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:54.000000000 +0200
@@ -178,6 +178,38 @@
 	return ret;
 }
 
+static int check_for_io_on_path(struct subchannel *sch, int mask)
+{
+	int cc;
+
+	cc = stsch(sch->schid, &sch->schib);
+	if (cc)
+		return 0;
+	if (sch->schib.scsw.actl && sch->schib.pmcw.lpum == mask)
+		return 1;
+	return 0;
+}
+
+static void terminate_internal_io(struct subchannel *sch)
+{
+	if (cio_clear(sch)) {
+		/* Recheck device in case clear failed. */
+		sch->lpm = 0;
+		if (device_trigger_verify(sch) != 0) {
+			if(css_enqueue_subchannel_slow(sch->schid)) {
+				css_clear_subchannel_slow_list();
+				need_rescan = 1;
+			}
+		}
+		return;
+	}
+	/* Request retry of internal operation. */
+	device_set_intretry(sch);
+	/* Call handler. */
+	if (sch->driver && sch->driver->termination)
+		sch->driver->termination(&sch->dev);
+}
+
 static int
 s390_subchannel_remove_chpid(struct device *dev, void *data)
 {
@@ -208,37 +240,33 @@
 	if (sch->schib.pmcw.pim == 0x80)
 		goto out_unreg;
 
-	if ((sch->schib.scsw.actl & SCSW_ACTL_DEVACT) &&
-	    (sch->schib.scsw.actl & SCSW_ACTL_SCHACT) &&
-	    (sch->schib.pmcw.lpum == mask)) {
-		int cc;
-
-		cc = cio_clear(sch);
-		if (cc == -ENODEV)
+	if (check_for_io_on_path(sch, mask)) {
+		if (device_is_online(sch))
+			device_kill_io(sch);
+		else {
+			terminate_internal_io(sch);
+			/* Re-start path verification. */
+			if (sch->driver && sch->driver->verify)
+				sch->driver->verify(&sch->dev);
+		}
+	} else {
+		/* trigger path verification. */
+		if (sch->driver && sch->driver->verify)
+			sch->driver->verify(&sch->dev);
+		else if (sch->lpm == mask)
 			goto out_unreg;
-		/* Request retry of internal operation. */
-		device_set_intretry(sch);
-		/* Call handler. */
-		if (sch->driver && sch->driver->termination)
-			sch->driver->termination(&sch->dev);
-		goto out_unlock;
 	}
 
-	/* trigger path verification. */
-	if (sch->driver && sch->driver->verify)
-		sch->driver->verify(&sch->dev);
-	else if (sch->lpm == mask)
-		goto out_unreg;
-out_unlock:
 	spin_unlock_irq(sch->lock);
 	return 0;
+
 out_unreg:
-	spin_unlock_irq(sch->lock);
 	sch->lpm = 0;
 	if (css_enqueue_subchannel_slow(sch->schid)) {
 		css_clear_subchannel_slow_list();
 		need_rescan = 1;
 	}
+	spin_unlock_irq(sch->lock);
 	return 0;
 }
 
@@ -683,38 +711,6 @@
 	return rc;
 }
 
-static int check_for_io_on_path(struct subchannel *sch, int index)
-{
-	int cc;
-
-	cc = stsch(sch->schid, &sch->schib);
-	if (cc)
-		return 0;
-	if (sch->schib.scsw.actl && sch->schib.pmcw.lpum == (0x80 >> index))
-		return 1;
-	return 0;
-}
-
-static void terminate_internal_io(struct subchannel *sch)
-{
-	if (cio_clear(sch)) {
-		/* Recheck device in case clear failed. */
-		sch->lpm = 0;
-		if (device_trigger_verify(sch) != 0) {
-			if(css_enqueue_subchannel_slow(sch->schid)) {
-				css_clear_subchannel_slow_list();
-				need_rescan = 1;
-			}
-		}
-		return;
-	}
-	/* Request retry of internal operation. */
-	device_set_intretry(sch);
-	/* Call handler. */
-	if (sch->driver && sch->driver->termination)
-		sch->driver->termination(&sch->dev);
-}
-
 static void __s390_subchannel_vary_chpid(struct subchannel *sch,
 					 struct chp_id chpid, int on)
 {
@@ -741,13 +737,17 @@
 		}
 		sch->opm &= ~(0x80 >> chp);
 		sch->lpm &= ~(0x80 >> chp);
-		if (check_for_io_on_path(sch, chp)) {
+		if (check_for_io_on_path(sch, (0x80 >> chp))) {
 			if (device_is_online(sch))
 				/* Path verification is done after killing. */
 				device_kill_io(sch);
-			else
+			else {
 				/* Kill and retry internal I/O. */
 				terminate_internal_io(sch);
+				/* Re-start path verification. */
+				if (sch->driver && sch->driver->verify)
+					sch->driver->verify(&sch->dev);
+			}
 		} else if (!sch->lpm) {
 			if (device_trigger_verify(sch) != 0) {
 				if (css_enqueue_subchannel_slow(sch->schid)) {

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 11/38] cio: replace subchannel evaluation queue with bitmap

[Martin Schwidefsky]

[-- Attachment #1: 109-cio-idset.diff --]
[-- Type: text/plain, Size: 25847 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Use a bitmap for indicating which subchannels require evaluation
instead of allocating memory for each evaluation request. This
approach reduces memory consumption during recovery in case of
massive evaluation request occurrence and removes the need for
memory allocation failure handling.

Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/Makefile     |    2 
 drivers/s390/cio/chp.c        |    8 --
 drivers/s390/cio/chp.h        |    2 
 drivers/s390/cio/chsc.c       |  130 +++++++++---------------------------
 drivers/s390/cio/chsc.h       |    4 -
 drivers/s390/cio/css.c        |  148 +++++++++++++-----------------------------
 drivers/s390/cio/css.h        |   10 --
 drivers/s390/cio/device_fsm.c |    6 -
 drivers/s390/cio/idset.c      |  112 +++++++++++++++++++++++++++++++
 drivers/s390/cio/idset.h      |   25 +++++++
 drivers/s390/s390mach.c       |   24 +-----
 11 files changed, 235 insertions(+), 236 deletions(-)

Index: quilt-2.6/drivers/s390/cio/chp.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chp.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chp.c	2007-04-27 16:04:54.000000000 +0200
@@ -491,7 +491,7 @@
  * Handle channel-report-words indicating that the status of a channel-path
  * has changed.
  */
-int chp_process_crw(int id, int status)
+void chp_process_crw(int id, int status)
 {
 	struct chp_id chpid;
 
@@ -500,11 +500,9 @@
 	if (status) {
 		if (!chp_is_registered(chpid))
 			chp_new(chpid);
-		return chsc_chp_online(chpid);
-	} else {
+		chsc_chp_online(chpid);
+	} else
 		chsc_chp_offline(chpid);
-		return 0;
-	}
 }
 
 static inline int info_bit_num(struct chp_id id)
Index: quilt-2.6/drivers/s390/cio/chp.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chp.h	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chp.h	2007-04-27 16:04:54.000000000 +0200
@@ -42,7 +42,7 @@
 u8 chp_get_sch_opm(struct subchannel *sch);
 int chp_is_registered(struct chp_id chpid);
 void *chp_get_chp_desc(struct chp_id chpid);
-int chp_process_crw(int id, int available);
+void chp_process_crw(int id, int available);
 void chp_remove_cmg_attr(struct channel_path *chp);
 int chp_add_cmg_attr(struct channel_path *chp);
 int chp_new(struct chp_id chpid);
Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:54.000000000 +0200
@@ -195,12 +195,8 @@
 	if (cio_clear(sch)) {
 		/* Recheck device in case clear failed. */
 		sch->lpm = 0;
-		if (device_trigger_verify(sch) != 0) {
-			if(css_enqueue_subchannel_slow(sch->schid)) {
-				css_clear_subchannel_slow_list();
-				need_rescan = 1;
-			}
-		}
+		if (device_trigger_verify(sch) != 0)
+			css_schedule_eval(sch->schid);
 		return;
 	}
 	/* Request retry of internal operation. */
@@ -262,11 +258,8 @@
 
 out_unreg:
 	sch->lpm = 0;
-	if (css_enqueue_subchannel_slow(sch->schid)) {
-		css_clear_subchannel_slow_list();
-		need_rescan = 1;
-	}
 	spin_unlock_irq(sch->lock);
+	css_schedule_eval(sch->schid);
 	return 0;
 }
 
@@ -281,9 +274,6 @@
 		return;
 	bus_for_each_dev(&css_bus_type, NULL, &chpid,
 			 s390_subchannel_remove_chpid);
-
-	if (need_rescan || css_slow_subchannels_exist())
-		queue_work(slow_path_wq, &slow_path_work);
 }
 
 struct res_acc_data {
@@ -331,7 +321,6 @@
 s390_process_res_acc_new_sch(struct subchannel_id schid)
 {
 	struct schib schib;
-	int ret;
 	/*
 	 * We don't know the device yet, but since a path
 	 * may be available now to the device we'll have
@@ -342,15 +331,10 @@
 	 */
 	if (stsch_err(schid, &schib))
 		/* We're through */
-		return need_rescan ? -EAGAIN : -ENXIO;
+		return -ENXIO;
 
 	/* Put it on the slow path. */
-	ret = css_enqueue_subchannel_slow(schid);
-	if (ret) {
-		css_clear_subchannel_slow_list();
-		need_rescan = 1;
-		return -EAGAIN;
-	}
+	css_schedule_eval(schid);
 	return 0;
 }
 
@@ -392,10 +376,8 @@
 }
 
 
-static int
-s390_process_res_acc (struct res_acc_data *res_data)
+static void s390_process_res_acc (struct res_acc_data *res_data)
 {
-	int rc;
 	char dbf_txt[15];
 
 	sprintf(dbf_txt, "accpr%x.%02x", res_data->chpid.cssid,
@@ -413,12 +395,7 @@
 	 * The more information we have (info), the less scanning
 	 * will we have to do.
 	 */
-	rc = for_each_subchannel(__s390_process_res_acc, res_data);
-	if (css_slow_subchannels_exist())
-		rc = -EAGAIN;
-	else if (rc != -EAGAIN)
-		rc = 0;
-	return rc;
+	for_each_subchannel(__s390_process_res_acc, res_data);
 }
 
 static int
@@ -470,7 +447,7 @@
 	/* ccdf has to be big enough for a link-incident record */
 } __attribute__ ((packed));
 
-static int chsc_process_sei_link_incident(struct chsc_sei_area *sei_area)
+static void chsc_process_sei_link_incident(struct chsc_sei_area *sei_area)
 {
 	struct chp_id chpid;
 	int id;
@@ -478,7 +455,7 @@
 	CIO_CRW_EVENT(4, "chsc: link incident (rs=%02x, rs_id=%04x)\n",
 		      sei_area->rs, sei_area->rsid);
 	if (sei_area->rs != 4)
-		return 0;
+		return;
 	id = __get_chpid_from_lir(sei_area->ccdf);
 	if (id < 0)
 		CIO_CRW_EVENT(4, "chsc: link incident - invalid LIR\n");
@@ -487,21 +464,18 @@
 		chpid.id = id;
 		chsc_chp_offline(chpid);
 	}
-
-	return 0;
 }
 
-static int chsc_process_sei_res_acc(struct chsc_sei_area *sei_area)
+static void chsc_process_sei_res_acc(struct chsc_sei_area *sei_area)
 {
 	struct res_acc_data res_data;
 	struct chp_id chpid;
 	int status;
-	int rc;
 
 	CIO_CRW_EVENT(4, "chsc: resource accessibility event (rs=%02x, "
 		      "rs_id=%04x)\n", sei_area->rs, sei_area->rsid);
 	if (sei_area->rs != 4)
-		return 0;
+		return;
 	chp_id_init(&chpid);
 	chpid.id = sei_area->rsid;
 	/* allocate a new channel path structure, if needed */
@@ -509,7 +483,7 @@
 	if (status < 0)
 		chp_new(chpid);
 	else if (!status)
-		return 0;
+		return;
 	memset(&res_data, 0, sizeof(struct res_acc_data));
 	res_data.chpid = chpid;
 	if ((sei_area->vf & 0xc0) != 0) {
@@ -521,9 +495,7 @@
 			/* link address */
 			res_data.fla_mask = 0xff00;
 	}
-	rc = s390_process_res_acc(&res_data);
-
-	return rc;
+	s390_process_res_acc(&res_data);
 }
 
 struct chp_config_data {
@@ -532,7 +504,7 @@
 	u8 pc;
 };
 
-static int chsc_process_sei_chp_config(struct chsc_sei_area *sei_area)
+static void chsc_process_sei_chp_config(struct chsc_sei_area *sei_area)
 {
 	struct chp_config_data *data;
 	struct chp_id chpid;
@@ -540,7 +512,7 @@
 
 	CIO_CRW_EVENT(4, "chsc: channel-path-configuration notification\n");
 	if (sei_area->rs != 0)
-		return 0;
+		return;
 	data = (struct chp_config_data *) &(sei_area->ccdf);
 	chp_id_init(&chpid);
 	for (num = 0; num <= __MAX_CHPID; num++) {
@@ -561,52 +533,44 @@
 			break;
 		}
 	}
-
-	return 0;
 }
 
-static int chsc_process_sei(struct chsc_sei_area *sei_area)
+static void chsc_process_sei(struct chsc_sei_area *sei_area)
 {
-	int rc;
-
 	/* Check if we might have lost some information. */
-	if (sei_area->flags & 0x40)
+	if (sei_area->flags & 0x40) {
 		CIO_CRW_EVENT(2, "chsc: event overflow\n");
+		css_schedule_eval_all();
+	}
 	/* which kind of information was stored? */
-	rc = 0;
 	switch (sei_area->cc) {
 	case 1: /* link incident*/
-		rc = chsc_process_sei_link_incident(sei_area);
+		chsc_process_sei_link_incident(sei_area);
 		break;
 	case 2: /* i/o resource accessibiliy */
-		rc = chsc_process_sei_res_acc(sei_area);
+		chsc_process_sei_res_acc(sei_area);
 		break;
 	case 8: /* channel-path-configuration notification */
-		rc = chsc_process_sei_chp_config(sei_area);
+		chsc_process_sei_chp_config(sei_area);
 		break;
 	default: /* other stuff */
 		CIO_CRW_EVENT(4, "chsc: unhandled sei content code %d\n",
 			      sei_area->cc);
 		break;
 	}
-
-	return rc;
 }
 
-int chsc_process_crw(void)
+void chsc_process_crw(void)
 {
 	struct chsc_sei_area *sei_area;
-	int ret;
-	int rc;
 
 	if (!sei_page)
-		return 0;
+		return;
 	/* Access to sei_page is serialized through machine check handler
 	 * thread, so no need for locking. */
 	sei_area = sei_page;
 
 	CIO_TRACE_EVENT( 2, "prcss");
-	ret = 0;
 	do {
 		memset(sei_area, 0, sizeof(*sei_area));
 		sei_area->request.length = 0x0010;
@@ -616,37 +580,26 @@
 
 		if (sei_area->response.code == 0x0001) {
 			CIO_CRW_EVENT(4, "chsc: sei successful\n");
-			rc = chsc_process_sei(sei_area);
-			if (rc)
-				ret = rc;
+			chsc_process_sei(sei_area);
 		} else {
 			CIO_CRW_EVENT(2, "chsc: sei failed (rc=%04x)\n",
 				      sei_area->response.code);
-			ret = 0;
 			break;
 		}
 	} while (sei_area->flags & 0x80);
-
-	return ret;
 }
 
 static int
 __chp_add_new_sch(struct subchannel_id schid)
 {
 	struct schib schib;
-	int ret;
 
 	if (stsch_err(schid, &schib))
 		/* We're through */
-		return need_rescan ? -EAGAIN : -ENXIO;
+		return -ENXIO;
 
 	/* Put it on the slow path. */
-	ret = css_enqueue_subchannel_slow(schid);
-	if (ret) {
-		css_clear_subchannel_slow_list();
-		need_rescan = 1;
-		return -EAGAIN;
-	}
+	css_schedule_eval(schid);
 	return 0;
 }
 
@@ -693,22 +646,15 @@
 	return 0;
 }
 
-int chsc_chp_online(struct chp_id chpid)
+void chsc_chp_online(struct chp_id chpid)
 {
-	int rc;
 	char dbf_txt[15];
 
 	sprintf(dbf_txt, "cadd%x.%02x", chpid.cssid, chpid.id);
 	CIO_TRACE_EVENT(2, dbf_txt);
 
-	if (chp_get_status(chpid) == 0)
-		return 0;
-	rc = for_each_subchannel(__chp_add, &chpid);
-	if (css_slow_subchannels_exist())
-		rc = -EAGAIN;
-	if (rc != -EAGAIN)
-		rc = 0;
-	return rc;
+	if (chp_get_status(chpid) != 0)
+		for_each_subchannel(__chp_add, &chpid);
 }
 
 static void __s390_subchannel_vary_chpid(struct subchannel *sch,
@@ -749,12 +695,8 @@
 					sch->driver->verify(&sch->dev);
 			}
 		} else if (!sch->lpm) {
-			if (device_trigger_verify(sch) != 0) {
-				if (css_enqueue_subchannel_slow(sch->schid)) {
-					css_clear_subchannel_slow_list();
-					need_rescan = 1;
-				}
-			}
+			if (device_trigger_verify(sch) != 0)
+				css_schedule_eval(sch->schid);
 		} else if (sch->driver && sch->driver->verify)
 			sch->driver->verify(&sch->dev);
 		break;
@@ -801,11 +743,7 @@
 		/* We're through */
 		return -ENXIO;
 	/* Put it on the slow path. */
-	if (css_enqueue_subchannel_slow(schid)) {
-		css_clear_subchannel_slow_list();
-		need_rescan = 1;
-		return -EAGAIN;
-	}
+	css_schedule_eval(schid);
 	return 0;
 }
 
@@ -826,8 +764,6 @@
 	if (on)
 		/* Scan for new devices on varied on path. */
 		for_each_subchannel(__s390_vary_chpid_on, NULL);
-	if (need_rescan || css_slow_subchannels_exist())
-		queue_work(slow_path_wq, &slow_path_work);
 	return 0;
 }
 
Index: quilt-2.6/drivers/s390/cio/chsc.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.h	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.h	2007-04-27 16:04:54.000000000 +0200
@@ -36,7 +36,7 @@
 struct channel_path;
 
 extern int css_get_ssd_info(struct subchannel *);
-extern int chsc_process_crw(void);
+extern void chsc_process_crw(void);
 
 struct css_general_char {
 	u64 : 41;
@@ -79,7 +79,7 @@
 int chsc_chp_vary(struct chp_id chpid, int on);
 int chsc_determine_channel_path_description(struct chp_id chpid,
 					    struct channel_path_desc *desc);
-int chsc_chp_online(struct chp_id chpid);
+void chsc_chp_online(struct chp_id chpid);
 void chsc_chp_offline(struct chp_id chpid);
 int chsc_get_channel_measurement_chars(struct channel_path *chp);
 
Index: quilt-2.6/drivers/s390/cio/css.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/css.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/css.c	2007-04-27 16:04:54.000000000 +0200
@@ -20,8 +20,8 @@
 #include "ioasm.h"
 #include "chsc.h"
 #include "device.h"
+#include "idset.h"
 
-int need_rescan = 0;
 int css_init_done = 0;
 static int need_reprobe = 0;
 static int max_ssid = 0;
@@ -306,7 +306,7 @@
 	return css_probe_device(schid);
 }
 
-static int css_evaluate_subchannel(struct subchannel_id schid, int slow)
+static void css_evaluate_subchannel(struct subchannel_id schid, int slow)
 {
 	struct subchannel *sch;
 	int ret;
@@ -317,53 +317,66 @@
 		put_device(&sch->dev);
 	} else
 		ret = css_evaluate_new_subchannel(schid, slow);
-
-	return ret;
+	if (ret == -EAGAIN)
+		css_schedule_eval(schid);
 }
 
-static int
-css_rescan_devices(struct subchannel_id schid, void *data)
+static struct idset *slow_subchannel_set;
+static spinlock_t slow_subchannel_lock;
+
+static int __init slow_subchannel_init(void)
 {
-	return css_evaluate_subchannel(schid, 1);
+	spin_lock_init(&slow_subchannel_lock);
+	slow_subchannel_set = idset_sch_new();
+	if (!slow_subchannel_set) {
+		printk(KERN_WARNING "cio: could not allocate slow subchannel "
+		       "set\n");
+		return -ENOMEM;
+	}
+	return 0;
 }
 
-struct slow_subchannel {
-	struct list_head slow_list;
-	struct subchannel_id schid;
-};
-
-static LIST_HEAD(slow_subchannels_head);
-static DEFINE_SPINLOCK(slow_subchannel_lock);
+subsys_initcall(slow_subchannel_init);
 
-static void
-css_trigger_slow_path(struct work_struct *unused)
+static void css_slow_path_func(struct work_struct *unused)
 {
-	CIO_TRACE_EVENT(4, "slowpath");
-
-	if (need_rescan) {
-		need_rescan = 0;
-		for_each_subchannel(css_rescan_devices, NULL);
-		return;
-	}
+	struct subchannel_id schid;
 
+	CIO_TRACE_EVENT(4, "slowpath");
 	spin_lock_irq(&slow_subchannel_lock);
-	while (!list_empty(&slow_subchannels_head)) {
-		struct slow_subchannel *slow_sch =
-			list_entry(slow_subchannels_head.next,
-				   struct slow_subchannel, slow_list);
-
-		list_del_init(slow_subchannels_head.next);
+	init_subchannel_id(&schid);
+	while (idset_sch_get_first(slow_subchannel_set, &schid)) {
+		idset_sch_del(slow_subchannel_set, schid);
 		spin_unlock_irq(&slow_subchannel_lock);
-		css_evaluate_subchannel(slow_sch->schid, 1);
+		css_evaluate_subchannel(schid, 1);
 		spin_lock_irq(&slow_subchannel_lock);
-		kfree(slow_sch);
 	}
 	spin_unlock_irq(&slow_subchannel_lock);
 }
 
-DECLARE_WORK(slow_path_work, css_trigger_slow_path);
+static DECLARE_WORK(slow_path_work, css_slow_path_func);
 struct workqueue_struct *slow_path_wq;
 
+void css_schedule_eval(struct subchannel_id schid)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&slow_subchannel_lock, flags);
+	idset_sch_add(slow_subchannel_set, schid);
+	queue_work(slow_path_wq, &slow_path_work);
+	spin_unlock_irqrestore(&slow_subchannel_lock, flags);
+}
+
+void css_schedule_eval_all(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&slow_subchannel_lock, flags);
+	idset_fill(slow_subchannel_set);
+	queue_work(slow_path_wq, &slow_path_work);
+	spin_unlock_irqrestore(&slow_subchannel_lock, flags);
+}
+
 /* Reprobe subchannel if unregistered. */
 static int reprobe_subchannel(struct subchannel_id schid, void *data)
 {
@@ -426,33 +439,14 @@
 EXPORT_SYMBOL_GPL(css_schedule_reprobe);
 
 /*
- * Rescan for new devices. FIXME: This is slow.
- * This function is called when we have lost CRWs due to overflows and we have
- * to do subchannel housekeeping.
- */
-void
-css_reiterate_subchannels(void)
-{
-	css_clear_subchannel_slow_list();
-	need_rescan = 1;
-}
-
-/*
  * Called from the machine check handler for subchannel report words.
  */
-int
-css_process_crw(int rsid1, int rsid2)
+void css_process_crw(int rsid1, int rsid2)
 {
-	int ret;
 	struct subchannel_id mchk_schid;
 
 	CIO_CRW_EVENT(2, "source is subchannel %04X, subsystem id %x\n",
 		      rsid1, rsid2);
-
-	if (need_rescan)
-		/* We need to iterate all subchannels anyway. */
-		return -EAGAIN;
-
 	init_subchannel_id(&mchk_schid);
 	mchk_schid.sch_no = rsid1;
 	if (rsid2 != 0)
@@ -463,14 +457,7 @@
 	 * use stsch() to find out if the subchannel in question has come
 	 * or gone.
 	 */
-	ret = css_evaluate_subchannel(mchk_schid, 0);
-	if (ret == -EAGAIN) {
-		if (css_enqueue_subchannel_slow(mchk_schid)) {
-			css_clear_subchannel_slow_list();
-			need_rescan = 1;
-		}
-	}
-	return ret;
+	css_evaluate_subchannel(mchk_schid, 0);
 }
 
 static int __init
@@ -745,47 +732,6 @@
 
 subsys_initcall(init_channel_subsystem);
 
-int
-css_enqueue_subchannel_slow(struct subchannel_id schid)
-{
-	struct slow_subchannel *new_slow_sch;
-	unsigned long flags;
-
-	new_slow_sch = kzalloc(sizeof(struct slow_subchannel), GFP_ATOMIC);
-	if (!new_slow_sch)
-		return -ENOMEM;
-	new_slow_sch->schid = schid;
-	spin_lock_irqsave(&slow_subchannel_lock, flags);
-	list_add_tail(&new_slow_sch->slow_list, &slow_subchannels_head);
-	spin_unlock_irqrestore(&slow_subchannel_lock, flags);
-	return 0;
-}
-
-void
-css_clear_subchannel_slow_list(void)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&slow_subchannel_lock, flags);
-	while (!list_empty(&slow_subchannels_head)) {
-		struct slow_subchannel *slow_sch =
-			list_entry(slow_subchannels_head.next,
-				   struct slow_subchannel, slow_list);
-
-		list_del_init(slow_subchannels_head.next);
-		kfree(slow_sch);
-	}
-	spin_unlock_irqrestore(&slow_subchannel_lock, flags);
-}
-
-
-
-int
-css_slow_subchannels_exist(void)
-{
-	return (!list_empty(&slow_subchannels_head));
-}
-
 MODULE_LICENSE("GPL");
 EXPORT_SYMBOL(css_bus_type);
 EXPORT_SYMBOL_GPL(css_characteristics_avail);
Index: quilt-2.6/drivers/s390/cio/css.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/css.h	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/css.h	2007-04-27 16:04:54.000000000 +0200
@@ -146,7 +146,7 @@
 extern struct subchannel * get_subchannel_by_schid(struct subchannel_id);
 extern int css_init_done;
 extern int for_each_subchannel(int(*fn)(struct subchannel_id, void *), void *);
-extern int css_process_crw(int, int);
+extern void css_process_crw(int, int);
 extern void css_reiterate_subchannels(void);
 
 #define __MAX_SUBCHANNEL 65535
@@ -186,16 +186,12 @@
 void device_kill_pending_timer(struct subchannel *);
 
 /* Helper functions to build lists for the slow path. */
-extern int css_enqueue_subchannel_slow(struct subchannel_id schid);
-void css_walk_subchannel_slow_list(void (*fn)(unsigned long));
-void css_clear_subchannel_slow_list(void);
-int css_slow_subchannels_exist(void);
-extern int need_rescan;
+void css_schedule_eval(struct subchannel_id schid);
+void css_schedule_eval_all(void);
 
 int sch_is_pseudo_sch(struct subchannel *);
 
 extern struct workqueue_struct *slow_path_wq;
-extern struct work_struct slow_path_work;
 
 int subchannel_add_files (struct device *);
 extern struct attribute_group *subch_attr_groups[];
Index: quilt-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:54.000000000 +0200
@@ -222,10 +222,8 @@
 		if (old_lpm & mask)
 			continue;
 		chpid.id = sch->schib.pmcw.chpid[i];
-		if (!chp_is_registered(chpid)) {
-			need_rescan = 1;
-			queue_work(slow_path_wq, &slow_path_work);
-		}
+		if (!chp_is_registered(chpid))
+			css_schedule_eval_all();
 	}
 }
 
Index: quilt-2.6/drivers/s390/cio/idset.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/cio/idset.c	2007-04-27 16:04:54.000000000 +0200
@@ -0,0 +1,112 @@
+/*
+ *  drivers/s390/cio/idset.c
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#include <linux/slab.h>
+#include <asm/bitops.h>
+#include "idset.h"
+#include "css.h"
+
+struct idset {
+	int num_ssid;
+	int num_id;
+	unsigned long bitmap[0];
+};
+
+static inline unsigned long bitmap_size(int num_ssid, int num_id)
+{
+	return __BITOPS_WORDS(num_ssid * num_id) * sizeof(unsigned long);
+}
+
+static struct idset *idset_new(int num_ssid, int num_id)
+{
+	struct idset *set;
+
+	set = kzalloc(sizeof(struct idset) + bitmap_size(num_ssid, num_id),
+		      GFP_KERNEL);
+	if (set) {
+		set->num_ssid = num_ssid;
+		set->num_id = num_id;
+	}
+	return set;
+}
+
+void idset_free(struct idset *set)
+{
+	kfree(set);
+}
+
+void idset_clear(struct idset *set)
+{
+	memset(set->bitmap, 0, bitmap_size(set->num_ssid, set->num_id));
+}
+
+void idset_fill(struct idset *set)
+{
+	memset(set->bitmap, 0xff, bitmap_size(set->num_ssid, set->num_id));
+}
+
+static inline void idset_add(struct idset *set, int ssid, int id)
+{
+	set_bit(ssid * set->num_id + id, set->bitmap);
+}
+
+static inline void idset_del(struct idset *set, int ssid, int id)
+{
+	clear_bit(ssid * set->num_id + id, set->bitmap);
+}
+
+static inline int idset_contains(struct idset *set, int ssid, int id)
+{
+	return test_bit(ssid * set->num_id + id, set->bitmap);
+}
+
+static inline int idset_get_first(struct idset *set, int *ssid, int *id)
+{
+	int bitnum;
+
+	bitnum = find_first_bit(set->bitmap, set->num_ssid * set->num_id);
+	if (bitnum >= set->num_ssid * set->num_id)
+		return 0;
+	*ssid = bitnum / set->num_id;
+	*id = bitnum % set->num_id;
+	return 1;
+}
+
+struct idset *idset_sch_new(void)
+{
+	return idset_new(__MAX_SSID + 1, __MAX_SUBCHANNEL + 1);
+}
+
+void idset_sch_add(struct idset *set, struct subchannel_id schid)
+{
+	idset_add(set, schid.ssid, schid.sch_no);
+}
+
+void idset_sch_del(struct idset *set, struct subchannel_id schid)
+{
+	idset_del(set, schid.ssid, schid.sch_no);
+}
+
+int idset_sch_contains(struct idset *set, struct subchannel_id schid)
+{
+	return idset_contains(set, schid.ssid, schid.sch_no);
+}
+
+int idset_sch_get_first(struct idset *set, struct subchannel_id *schid)
+{
+	int ssid = 0;
+	int id = 0;
+	int rc;
+
+	rc = idset_get_first(set, &ssid, &id);
+	if (rc) {
+		init_subchannel_id(schid);
+		schid->ssid = ssid;
+		schid->sch_no = id;
+	}
+	return rc;
+}
Index: quilt-2.6/drivers/s390/cio/idset.h
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/cio/idset.h	2007-04-27 16:04:54.000000000 +0200
@@ -0,0 +1,25 @@
+/*
+ *  drivers/s390/cio/idset.h
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#ifndef S390_IDSET_H
+#define S390_IDSET_H S390_IDSET_H
+
+#include "schid.h"
+
+struct idset;
+
+void idset_free(struct idset *set);
+void idset_clear(struct idset *set);
+void idset_fill(struct idset *set);
+
+struct idset *idset_sch_new(void);
+void idset_sch_add(struct idset *set, struct subchannel_id id);
+void idset_sch_del(struct idset *set, struct subchannel_id id);
+int idset_sch_contains(struct idset *set, struct subchannel_id id);
+int idset_sch_get_first(struct idset *set, struct subchannel_id *id);
+
+#endif /* S390_IDSET_H */
Index: quilt-2.6/drivers/s390/cio/Makefile
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/Makefile	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/Makefile	2007-04-27 16:04:54.000000000 +0200
@@ -2,7 +2,7 @@
 # Makefile for the S/390 common i/o drivers
 #
 
-obj-y += airq.o blacklist.o chsc.o cio.o css.o chp.o
+obj-y += airq.o blacklist.o chsc.o cio.o css.o chp.o idset.o
 ccw_device-objs += device.o device_fsm.o device_ops.o
 ccw_device-objs += device_id.o device_pgid.o device_status.o
 obj-y += ccw_device.o cmf.o
Index: quilt-2.6/drivers/s390/s390mach.c
===================================================================
--- quilt-2.6.orig/drivers/s390/s390mach.c	2007-04-27 16:04:53.000000000 +0200
+++ quilt-2.6/drivers/s390/s390mach.c	2007-04-27 16:04:54.000000000 +0200
@@ -45,14 +45,13 @@
 s390_collect_crw_info(void *param)
 {
 	struct crw crw[2];
-	int ccode, ret, slow;
+	int ccode;
 	struct semaphore *sem;
 	unsigned int chain;
 
 	sem = (struct semaphore *)param;
 repeat:
 	down_interruptible(sem);
-	slow = 0;
 	chain = 0;
 	while (1) {
 		if (unlikely(chain > 1)) {
@@ -85,9 +84,8 @@
 		/* Check for overflows. */
 		if (crw[chain].oflw) {
 			pr_debug("%s: crw overflow detected!\n", __FUNCTION__);
-			css_reiterate_subchannels();
+			css_schedule_eval_all();
 			chain = 0;
-			slow = 1;
 			continue;
 		}
 		switch (crw[chain].rsc) {
@@ -95,10 +93,7 @@
 			if (crw[0].chn && !chain)
 				break;
 			pr_debug("source is subchannel %04X\n", crw[0].rsid);
-			ret = css_process_crw (crw[0].rsid,
-					       chain ? crw[1].rsid : 0);
-			if (ret == -EAGAIN)
-				slow = 1;
+			css_process_crw(crw[0].rsid, chain ? crw[1].rsid : 0);
 			break;
 		case CRW_RSC_MONITOR:
 			pr_debug("source is monitoring facility\n");
@@ -117,28 +112,23 @@
 			}
 			switch (crw[0].erc) {
 			case CRW_ERC_IPARM: /* Path has come. */
-				ret = chp_process_crw(crw[0].rsid, 1);
+				chp_process_crw(crw[0].rsid, 1);
 				break;
 			case CRW_ERC_PERRI: /* Path has gone. */
 			case CRW_ERC_PERRN:
-				ret = chp_process_crw(crw[0].rsid, 0);
+				chp_process_crw(crw[0].rsid, 0);
 				break;
 			default:
 				pr_debug("Don't know how to handle erc=%x\n",
 					 crw[0].erc);
-				ret = 0;
 			}
-			if (ret == -EAGAIN)
-				slow = 1;
 			break;
 		case CRW_RSC_CONFIG:
 			pr_debug("source is configuration-alert facility\n");
 			break;
 		case CRW_RSC_CSS:
 			pr_debug("source is channel subsystem\n");
-			ret = chsc_process_crw();
-			if (ret == -EAGAIN)
-				slow = 1;
+			chsc_process_crw();
 			break;
 		default:
 			pr_debug("unknown source\n");
@@ -147,8 +137,6 @@
 		/* chain is always 0 or 1 here. */
 		chain = crw[chain].chn ? chain + 1 : 0;
 	}
-	if (slow)
-		queue_work(slow_path_wq, &slow_path_work);
 	goto repeat;
 	return 0;
 }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 12/38] cio: fix subchannel channel-path data usage

[Martin Schwidefsky]

[-- Attachment #1: 110-cio-chpid-data.diff --]
[-- Type: text/plain, Size: 16392 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Ensure that channel-path related subchannel data is only retrieved and
used when it is valid and that it is updated when it may have changed.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/chsc.c       |  303 ++++++++++++++----------------------------
 drivers/s390/cio/chsc.h       |   10 +
 drivers/s390/cio/cio.h        |   16 --
 drivers/s390/cio/css.c        |   53 ++++++-
 drivers/s390/cio/css.h        |    1 
 drivers/s390/cio/device.c     |   12 +
 drivers/s390/cio/device_fsm.c |    1 
 7 files changed, 179 insertions(+), 217 deletions(-)

Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:55.000000000 +0200
@@ -26,155 +26,84 @@
 
 static void *sei_page;
 
-/* FIXME: this is _always_ called for every subchannel. shouldn't we
- *	  process more than one at a time? */
-static int
-chsc_get_sch_desc_irq(struct subchannel *sch, void *page)
-{
-	int ccode, j;
-
-	struct {
-		struct chsc_header request;
-		u16 reserved1a:10;
-		u16 ssid:2;
-		u16 reserved1b:4;
-		u16 f_sch;	  /* first subchannel */
-		u16 reserved2;
-		u16 l_sch;	  /* last subchannel */
-		u32 reserved3;
-		struct chsc_header response;
-		u32 reserved4;
-		u8 sch_valid : 1;
-		u8 dev_valid : 1;
-		u8 st	     : 3; /* subchannel type */
-		u8 zeroes    : 3;
-		u8  unit_addr;	  /* unit address */
-		u16 devno;	  /* device number */
-		u8 path_mask;
-		u8 fla_valid_mask;
-		u16 sch;	  /* subchannel */
-		u8 chpid[8];	  /* chpids 0-7 */
-		u16 fla[8];	  /* full link addresses 0-7 */
-	} __attribute__ ((packed)) *ssd_area;
+struct chsc_ssd_area {
+	struct chsc_header request;
+	u16 :10;
+	u16 ssid:2;
+	u16 :4;
+	u16 f_sch;	  /* first subchannel */
+	u16 :16;
+	u16 l_sch;	  /* last subchannel */
+	u32 :32;
+	struct chsc_header response;
+	u32 :32;
+	u8 sch_valid : 1;
+	u8 dev_valid : 1;
+	u8 st	     : 3; /* subchannel type */
+	u8 zeroes    : 3;
+	u8  unit_addr;	  /* unit address */
+	u16 devno;	  /* device number */
+	u8 path_mask;
+	u8 fla_valid_mask;
+	u16 sch;	  /* subchannel */
+	u8 chpid[8];	  /* chpids 0-7 */
+	u16 fla[8];	  /* full link addresses 0-7 */
+} __attribute__ ((packed));
 
-	ssd_area = page;
+int chsc_get_ssd_info(struct subchannel_id schid, struct chsc_ssd_info *ssd)
+{
+	unsigned long page;
+	struct chsc_ssd_area *ssd_area;
+	int ccode;
+	int ret;
+	int i;
+	int mask;
 
+	page = get_zeroed_page(GFP_KERNEL | GFP_DMA);
+	if (!page)
+		return -ENOMEM;
+	ssd_area = (struct chsc_ssd_area *) page;
 	ssd_area->request.length = 0x0010;
 	ssd_area->request.code = 0x0004;
-
-	ssd_area->ssid = sch->schid.ssid;
-	ssd_area->f_sch = sch->schid.sch_no;
-	ssd_area->l_sch = sch->schid.sch_no;
+	ssd_area->ssid = schid.ssid;
+	ssd_area->f_sch = schid.sch_no;
+	ssd_area->l_sch = schid.sch_no;
 
 	ccode = chsc(ssd_area);
+	/* Check response. */
 	if (ccode > 0) {
-		pr_debug("chsc returned with ccode = %d\n", ccode);
-		return (ccode == 3) ? -ENODEV : -EBUSY;
+		ret = (ccode == 3) ? -ENODEV : -EBUSY;
+		goto out_free;
 	}
-
-	switch (ssd_area->response.code) {
-	case 0x0001: /* everything ok */
-		break;
-	case 0x0002:
-		CIO_CRW_EVENT(2, "Invalid command!\n");
-		return -EINVAL;
-	case 0x0003:
-		CIO_CRW_EVENT(2, "Error in chsc request block!\n");
-		return -EINVAL;
-	case 0x0004:
-		CIO_CRW_EVENT(2, "Model does not provide ssd\n");
-		return -EOPNOTSUPP;
-	default:
-		CIO_CRW_EVENT(2, "Unknown CHSC response %d\n",
+	if (ssd_area->response.code != 0x0001) {
+		CIO_MSG_EVENT(2, "chsc: ssd failed for 0.%x.%04x (rc=%04x)\n",
+			      schid.ssid, schid.sch_no,
 			      ssd_area->response.code);
-		return -EIO;
+		ret = -EIO;
+		goto out_free;
 	}
-
-	/*
-	 * ssd_area->st stores the type of the detected
-	 * subchannel, with the following definitions:
-	 *
-	 * 0: I/O subchannel:	  All fields have meaning
-	 * 1: CHSC subchannel:	  Only sch_val, st and sch
-	 *			  have meaning
-	 * 2: Message subchannel: All fields except unit_addr
-	 *			  have meaning
-	 * 3: ADM subchannel:	  Only sch_val, st and sch
-	 *			  have meaning
-	 *
-	 * Other types are currently undefined.
-	 */
-	if (ssd_area->st > 3) { /* uhm, that looks strange... */
-		CIO_CRW_EVENT(0, "Strange subchannel type %d"
-			      " for sch 0.%x.%04x\n", ssd_area->st,
-			      sch->schid.ssid, sch->schid.sch_no);
-		/*
-		 * There may have been a new subchannel type defined in the
-		 * time since this code was written; since we don't know which
-		 * fields have meaning and what to do with it we just jump out
-		 */
-		return 0;
-	} else {
-		const char *type[4] = {"I/O", "chsc", "message", "ADM"};
-		CIO_CRW_EVENT(6, "ssd: sch 0.%x.%04x is %s subchannel\n",
-			      sch->schid.ssid, sch->schid.sch_no,
-			      type[ssd_area->st]);
-
-		sch->ssd_info.valid = 1;
-		sch->ssd_info.type = ssd_area->st;
-	}
-
-	if (ssd_area->st == 0 || ssd_area->st == 2) {
-		for (j = 0; j < 8; j++) {
-			if (!((0x80 >> j) & ssd_area->path_mask &
-			      ssd_area->fla_valid_mask))
-				continue;
-			sch->ssd_info.chpid[j] = ssd_area->chpid[j];
-			sch->ssd_info.fla[j]   = ssd_area->fla[j];
-		}
+	if (!ssd_area->sch_valid) {
+		ret = -ENODEV;
+		goto out_free;
 	}
-	return 0;
-}
-
-int
-css_get_ssd_info(struct subchannel *sch)
-{
-	int ret;
-	void *page;
-
-	page = (void *)get_zeroed_page(GFP_KERNEL | GFP_DMA);
-	if (!page)
-		return -ENOMEM;
-	spin_lock_irq(sch->lock);
-	ret = chsc_get_sch_desc_irq(sch, page);
-	if (ret) {
-		static int cio_chsc_err_msg;
-		
-		if (!cio_chsc_err_msg) {
-			printk(KERN_ERR
-			       "chsc_get_sch_descriptions:"
-			       " Error %d while doing chsc; "
-			       "processing some machine checks may "
-			       "not work\n", ret);
-			cio_chsc_err_msg = 1;
-		}
-	}
-	spin_unlock_irq(sch->lock);
-	free_page((unsigned long)page);
-	if (!ret) {
-		int j, mask;
-		struct chp_id chpid;
-
-		chp_id_init(&chpid);
-		/* Allocate channel path structures, if needed. */
-		for (j = 0; j < 8; j++) {
-			mask = 0x80 >> j;
-			chpid.id = sch->ssd_info.chpid[j];
-			if ((sch->schib.pmcw.pim & mask) &&
-			    !chp_is_registered(chpid))
-				chp_new(chpid);
+	/* Copy data */
+	ret = 0;
+	memset(ssd, 0, sizeof(struct chsc_ssd_info));
+	if ((ssd_area->st != 0) && (ssd_area->st != 2))
+		goto out_free;
+	ssd->path_mask = ssd_area->path_mask;
+	ssd->fla_valid_mask = ssd_area->fla_valid_mask;
+	for (i = 0; i < 8; i++) {
+		mask = 0x80 >> i;
+		if (ssd_area->path_mask & mask) {
+			chp_id_init(&ssd->chpid[i]);
+			ssd->chpid[i].id = ssd_area->chpid[i];
 		}
+		if (ssd_area->fla_valid_mask & mask)
+			ssd->fla[i] = ssd_area->fla[i];
 	}
+out_free:
+	free_page(page);
 	return ret;
 }
 
@@ -276,47 +205,6 @@
 			 s390_subchannel_remove_chpid);
 }
 
-struct res_acc_data {
-	struct chp_id chpid;
-	u32 fla_mask;
-	u16 fla;
-};
-
-static int s390_process_res_acc_sch(struct res_acc_data *res_data,
-				    struct subchannel *sch)
-{
-	int found;
-	int chp;
-	int ccode;
-
-	found = 0;
-	for (chp = 0; chp <= 7; chp++)
-		/*
-		 * check if chpid is in information updated by ssd
-		 */
-		if (sch->ssd_info.valid &&
-		    sch->ssd_info.chpid[chp] == res_data->chpid.id &&
-		    (sch->ssd_info.fla[chp] & res_data->fla_mask)
-		    == res_data->fla) {
-			found = 1;
-			break;
-		}
-
-	if (found == 0)
-		return 0;
-
-	/*
-	 * Do a stsch to update our subchannel structure with the
-	 * new path information and eventually check for logically
-	 * offline chpids.
-	 */
-	ccode = stsch(sch->schid, &sch->schib);
-	if (ccode > 0)
-		return 0;
-
-	return 0x80 >> chp;
-}
-
 static int
 s390_process_res_acc_new_sch(struct subchannel_id schid)
 {
@@ -338,6 +226,32 @@
 	return 0;
 }
 
+struct res_acc_data {
+	struct chp_id chpid;
+	u32 fla_mask;
+	u16 fla;
+};
+
+static int get_res_chpid_mask(struct chsc_ssd_info *ssd,
+			      struct res_acc_data *data)
+{
+	int i;
+	int mask;
+
+	for (i = 0; i < 8; i++) {
+		mask = 0x80 >> i;
+		if (!(ssd->path_mask & mask))
+			continue;
+		if (!chp_id_is_equal(&ssd->chpid[i], &data->chpid))
+			continue;
+		if ((ssd->fla_valid_mask & mask) &&
+		    ((ssd->fla[i] & data->fla_mask) != data->fla))
+			continue;
+		return mask;
+	}
+	return 0;
+}
+
 static int
 __s390_process_res_acc(struct subchannel_id schid, void *data)
 {
@@ -352,14 +266,11 @@
 		return s390_process_res_acc_new_sch(schid);
 
 	spin_lock_irq(sch->lock);
-
-	chp_mask = s390_process_res_acc_sch(res_data, sch);
-
-	if (chp_mask == 0) {
-		spin_unlock_irq(sch->lock);
-		put_device(&sch->dev);
-		return 0;
-	}
+	chp_mask = get_res_chpid_mask(&sch->ssd_info, res_data);
+	if (chp_mask == 0)
+		goto out;
+	if (stsch(sch->schid, &sch->schib))
+		goto out;
 	old_lpm = sch->lpm;
 	sch->lpm = ((sch->schib.pmcw.pim &
 		     sch->schib.pmcw.pam &
@@ -369,13 +280,12 @@
 		device_trigger_reprobe(sch);
 	else if (sch->driver && sch->driver->verify)
 		sch->driver->verify(&sch->dev);
-
+out:
 	spin_unlock_irq(sch->lock);
 	put_device(&sch->dev);
 	return 0;
 }
 
-
 static void s390_process_res_acc (struct res_acc_data *res_data)
 {
 	char dbf_txt[15];
@@ -661,29 +571,30 @@
 					 struct chp_id chpid, int on)
 {
 	int chp, old_lpm;
+	int mask;
 	unsigned long flags;
 
-	if (!sch->ssd_info.valid)
-		return;
-	
 	spin_lock_irqsave(sch->lock, flags);
 	old_lpm = sch->lpm;
 	for (chp = 0; chp < 8; chp++) {
-		if (sch->ssd_info.chpid[chp] != chpid.id)
+		mask = 0x80 >> chp;
+		if (!(sch->ssd_info.path_mask & mask))
+			continue;
+		if (!chp_id_is_equal(&sch->ssd_info.chpid[chp], &chpid))
 			continue;
 
 		if (on) {
-			sch->opm |= (0x80 >> chp);
-			sch->lpm |= (0x80 >> chp);
+			sch->opm |= mask;
+			sch->lpm |= mask;
 			if (!old_lpm)
 				device_trigger_reprobe(sch);
 			else if (sch->driver && sch->driver->verify)
 				sch->driver->verify(&sch->dev);
 			break;
 		}
-		sch->opm &= ~(0x80 >> chp);
-		sch->lpm &= ~(0x80 >> chp);
-		if (check_for_io_on_path(sch, (0x80 >> chp))) {
+		sch->opm &= ~mask;
+		sch->lpm &= ~mask;
+		if (check_for_io_on_path(sch, mask)) {
 			if (device_is_online(sch))
 				/* Path verification is done after killing. */
 				device_kill_io(sch);
Index: quilt-2.6/drivers/s390/cio/chsc.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.h	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.h	2007-04-27 16:04:55.000000000 +0200
@@ -4,6 +4,7 @@
 #include <linux/types.h>
 #include <linux/device.h>
 #include <asm/chpid.h>
+#include "schid.h"
 
 #define CHSC_SDA_OC_MSS   0x2
 
@@ -35,7 +36,6 @@
 
 struct channel_path;
 
-extern int css_get_ssd_info(struct subchannel *);
 extern void chsc_process_crw(void);
 
 struct css_general_char {
@@ -69,6 +69,14 @@
 extern struct css_general_char css_general_characteristics;
 extern struct css_chsc_char css_chsc_characteristics;
 
+struct chsc_ssd_info {
+	u8 path_mask;
+	u8 fla_valid_mask;
+	struct chp_id chpid[8];
+	u16 fla[8];
+};
+extern int chsc_get_ssd_info(struct subchannel_id schid,
+			     struct chsc_ssd_info *ssd);
 extern int chsc_determine_css_characteristics(void);
 extern int css_characteristics_avail;
 
Index: quilt-2.6/drivers/s390/cio/cio.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/cio.h	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/cio.h	2007-04-27 16:04:55.000000000 +0200
@@ -1,19 +1,11 @@
 #ifndef S390_CIO_H
 #define S390_CIO_H
 
-#include "schid.h"
 #include <linux/mutex.h>
 #include <linux/device.h>
-
-/*
- * where we put the ssd info
- */
-struct ssd_info {
-	__u8  valid:1;
-	__u8  type:7;		/* subchannel type */
-	__u8  chpid[8];		/* chpids */
-	__u16 fla[8];		/* full link addresses */
-} __attribute__ ((packed));
+#include <asm/chpid.h>
+#include "chsc.h"
+#include "schid.h"
 
 /*
  * path management control word
@@ -109,7 +101,7 @@
 	struct schib schib;	/* subchannel information block */
 	struct orb orb;		/* operation request block */
 	struct ccw1 sense_ccw;	/* static ccw for sense command */
-	struct ssd_info ssd_info;	/* subchannel description */
+	struct chsc_ssd_info ssd_info;	/* subchannel description */
 	struct device dev;	/* entry in device tree */
 	struct css_driver *driver;
 } __attribute__ ((aligned(8)));
Index: quilt-2.6/drivers/s390/cio/css.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/css.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/css.c	2007-04-27 16:04:55.000000000 +0200
@@ -21,6 +21,7 @@
 #include "chsc.h"
 #include "device.h"
 #include "idset.h"
+#include "chp.h"
 
 int css_init_done = 0;
 static int need_reprobe = 0;
@@ -125,8 +126,52 @@
 	mutex_unlock(&sch->reg_mutex);
 }
 
-static int
-css_register_subchannel(struct subchannel *sch)
+static void ssd_from_pmcw(struct chsc_ssd_info *ssd, struct pmcw *pmcw)
+{
+	int i;
+	int mask;
+
+	memset(ssd, 0, sizeof(struct chsc_ssd_info));
+	ssd->path_mask = pmcw->pim;
+	for (i = 0; i < 8; i++) {
+		mask = 0x80 >> i;
+		if (pmcw->pim & mask) {
+			chp_id_init(&ssd->chpid[i]);
+			ssd->chpid[i].id = pmcw->chpid[i];
+		}
+	}
+}
+
+static void ssd_register_chpids(struct chsc_ssd_info *ssd)
+{
+	int i;
+	int mask;
+
+	for (i = 0; i < 8; i++) {
+		mask = 0x80 >> i;
+		if (ssd->path_mask & mask)
+			if (!chp_is_registered(ssd->chpid[i]))
+				chp_new(ssd->chpid[i]);
+	}
+}
+
+void css_update_ssd_info(struct subchannel *sch)
+{
+	int ret;
+
+	if (cio_is_console(sch->schid)) {
+		/* Console is initialized too early for functions requiring
+		 * memory allocation. */
+		ssd_from_pmcw(&sch->ssd_info, &sch->schib.pmcw);
+	} else {
+		ret = chsc_get_ssd_info(sch->schid, &sch->ssd_info);
+		if (ret)
+			ssd_from_pmcw(&sch->ssd_info, &sch->schib.pmcw);
+		ssd_register_chpids(&sch->ssd_info);
+	}
+}
+
+static int css_register_subchannel(struct subchannel *sch)
 {
 	int ret;
 
@@ -135,9 +180,7 @@
 	sch->dev.bus = &css_bus_type;
 	sch->dev.release = &css_subchannel_release;
 	sch->dev.groups = subch_attr_groups;
-
-	css_get_ssd_info(sch);
-
+	css_update_ssd_info(sch);
 	/* make it known to the system */
 	ret = css_sch_device_register(sch);
 	if (ret) {
Index: quilt-2.6/drivers/s390/cio/css.h
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/css.h	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/css.h	2007-04-27 16:04:55.000000000 +0200
@@ -148,6 +148,7 @@
 extern int for_each_subchannel(int(*fn)(struct subchannel_id, void *), void *);
 extern void css_process_crw(int, int);
 extern void css_reiterate_subchannels(void);
+void css_update_ssd_info(struct subchannel *sch);
 
 #define __MAX_SUBCHANNEL 65535
 #define __MAX_SSID 3
Index: quilt-2.6/drivers/s390/cio/device.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device.c	2007-04-27 16:04:55.000000000 +0200
@@ -216,12 +216,18 @@
 chpids_show (struct device * dev, struct device_attribute *attr, char * buf)
 {
 	struct subchannel *sch = to_subchannel(dev);
-	struct ssd_info *ssd = &sch->ssd_info;
+	struct chsc_ssd_info *ssd = &sch->ssd_info;
 	ssize_t ret = 0;
 	int chp;
+	int mask;
 
-	for (chp = 0; chp < 8; chp++)
-		ret += sprintf (buf+ret, "%02x ", ssd->chpid[chp]);
+	for (chp = 0; chp < 8; chp++) {
+		mask = 0x80 >> chp;
+		if (ssd->path_mask & mask)
+			ret += sprintf(buf + ret, "%02x ", ssd->chpid[chp].id);
+		else
+			ret += sprintf(buf + ret, "00 ");
+	}
 	ret += sprintf (buf+ret, "\n");
 	return min((ssize_t)PAGE_SIZE, ret);
 }
Index: quilt-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:55.000000000 +0200
@@ -246,6 +246,7 @@
 	 */
 	old_lpm = sch->lpm;
 	stsch(sch->schid, &sch->schib);
+	css_update_ssd_info(sch);
 	sch->lpm = sch->schib.pmcw.pam & sch->opm;
 	/* Check since device may again have become not operational. */
 	if (!sch->schib.pmcw.dnv)

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 13/38] cio: Dont call css_update_ssd_info from interrupt context.

[Martin Schwidefsky]

[-- Attachment #1: 111-cio-update-ssd.diff --]
[-- Type: text/plain, Size: 1729 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/device.c     |    3 ++-
 drivers/s390/cio/device_fsm.c |    1 -
 2 files changed, 2 insertions(+), 2 deletions(-)

Index: quilt-2.6/drivers/s390/cio/device.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device.c	2007-04-27 16:04:55.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device.c	2007-04-27 16:04:55.000000000 +0200
@@ -743,6 +743,7 @@
 static void sch_attach_device(struct subchannel *sch,
 			      struct ccw_device *cdev)
 {
+	css_update_ssd_info(sch);
 	spin_lock_irq(sch->lock);
 	sch->dev.driver_data = cdev;
 	cdev->private->schid = sch->schid;
@@ -878,7 +879,7 @@
 	priv = container_of(work, struct ccw_device_private, kick_work);
 	cdev = priv->cdev;
 	sch = to_subchannel(cdev->dev.parent);
-
+	css_update_ssd_info(sch);
 	/*
 	 * io_subchannel_register() will also be called after device
 	 * recognition has been done for a boxed device (which will already
Index: quilt-2.6/drivers/s390/cio/device_fsm.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:55.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device_fsm.c	2007-04-27 16:04:55.000000000 +0200
@@ -246,7 +246,6 @@
 	 */
 	old_lpm = sch->lpm;
 	stsch(sch->schid, &sch->schib);
-	css_update_ssd_info(sch);
 	sch->lpm = sch->schib.pmcw.pam & sch->opm;
 	/* Check since device may again have become not operational. */
 	if (!sch->schib.pmcw.dnv)

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 14/38] cio: ccwgroup register vs. unregister.

[Martin Schwidefsky]

[-- Attachment #1: 112-cio-ccwgroup.diff --]
[-- Type: text/plain, Size: 3985 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

Introduce a mutex for struct ccwgroup to prevent simuntaneous
register/unregister on the same ccwgroup device.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/ccwgroup.c |   33 ++++++++++++++++++---------------
 include/asm-s390/ccwgroup.h |    1 +
 2 files changed, 19 insertions(+), 15 deletions(-)

Index: quilt-2.6/drivers/s390/cio/ccwgroup.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/ccwgroup.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/ccwgroup.c	2007-04-27 16:04:55.000000000 +0200
@@ -75,8 +75,10 @@
 {
 	struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
 
+	mutex_lock(&gdev->reg_mutex);
 	__ccwgroup_remove_symlinks(gdev);
 	device_unregister(dev);
+	mutex_unlock(&gdev->reg_mutex);
 }
 
 static ssize_t
@@ -173,7 +175,8 @@
 		return -ENOMEM;
 
 	atomic_set(&gdev->onoff, 0);
-
+	mutex_init(&gdev->reg_mutex);
+	mutex_lock(&gdev->reg_mutex);
 	for (i = 0; i < argc; i++) {
 		gdev->cdev[i] = get_ccwdev_by_busid(cdrv, argv[i]);
 
@@ -183,12 +186,12 @@
 		    || gdev->cdev[i]->id.driver_info !=
 		    gdev->cdev[0]->id.driver_info) {
 			rc = -EINVAL;
-			goto free_dev;
+			goto error;
 		}
 		/* Don't allow a device to belong to more than one group. */
 		if (gdev->cdev[i]->dev.driver_data) {
 			rc = -EINVAL;
-			goto free_dev;
+			goto error;
 		}
 		gdev->cdev[i]->dev.driver_data = gdev;
 	}
@@ -203,9 +206,8 @@
 			gdev->cdev[0]->dev.bus_id);
 
 	rc = device_register(&gdev->dev);
-	
 	if (rc)
-		goto free_dev;
+		goto error;
 	get_device(&gdev->dev);
 	rc = device_create_file(&gdev->dev, &dev_attr_ungroup);
 
@@ -216,6 +218,7 @@
 
 	rc = __ccwgroup_create_symlinks(gdev);
 	if (!rc) {
+		mutex_unlock(&gdev->reg_mutex);
 		put_device(&gdev->dev);
 		return 0;
 	}
@@ -224,19 +227,12 @@
 error:
 	for (i = 0; i < argc; i++)
 		if (gdev->cdev[i]) {
-			put_device(&gdev->cdev[i]->dev);
-			gdev->cdev[i]->dev.driver_data = NULL;
-		}
-	put_device(&gdev->dev);
-	return rc;
-free_dev:
-	for (i = 0; i < argc; i++)
-		if (gdev->cdev[i]) {
 			if (gdev->cdev[i]->dev.driver_data == gdev)
 				gdev->cdev[i]->dev.driver_data = NULL;
 			put_device(&gdev->cdev[i]->dev);
 		}
-	kfree(gdev);
+	mutex_unlock(&gdev->reg_mutex);
+	put_device(&gdev->dev);
 	return rc;
 }
 
@@ -422,8 +418,12 @@
 	get_driver(&cdriver->driver);
 	while ((dev = driver_find_device(&cdriver->driver, NULL, NULL,
 					 __ccwgroup_match_all))) {
-		__ccwgroup_remove_symlinks(to_ccwgroupdev(dev));
+		struct ccwgroup_device *gdev = to_ccwgroupdev(dev);
+
+		mutex_lock(&gdev->reg_mutex);
+		__ccwgroup_remove_symlinks(gdev);
 		device_unregister(dev);
+		mutex_unlock(&gdev->reg_mutex);
 		put_device(dev);
 	}
 	put_driver(&cdriver->driver);
@@ -444,8 +444,10 @@
 	if (cdev->dev.driver_data) {
 		gdev = (struct ccwgroup_device *)cdev->dev.driver_data;
 		if (get_device(&gdev->dev)) {
+			mutex_lock(&gdev->reg_mutex);
 			if (device_is_registered(&gdev->dev))
 				return gdev;
+			mutex_unlock(&gdev->reg_mutex);
 			put_device(&gdev->dev);
 		}
 		return NULL;
@@ -465,6 +467,7 @@
 	if (gdev) {
 		__ccwgroup_remove_symlinks(gdev);
 		device_unregister(&gdev->dev);
+		mutex_unlock(&gdev->reg_mutex);
 		put_device(&gdev->dev);
 	}
 }
Index: quilt-2.6/include/asm-s390/ccwgroup.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/ccwgroup.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/ccwgroup.h	2007-04-27 16:04:55.000000000 +0200
@@ -11,6 +11,7 @@
 		CCWGROUP_ONLINE,
 	} state;
 	atomic_t onoff;
+	struct mutex reg_mutex;
 	unsigned int count;		/* number of attached slave devices */
 	struct device	dev;		/* master device		    */
 	struct ccw_device *cdev[0];	/* variable number, allocate as needed */

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 15/38] cio: cm_enable memory leak.

[Martin Schwidefsky]

[-- Attachment #1: 113-cio-leak.diff --]
[-- Type: text/plain, Size: 982 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

We allocage two pages when channel path measurements are enabled
via cm_enable. We must not forget to free them again when
channel path measurements are disabled again.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/chsc.c |    2 +-
 1 files changed, 1 insertion(+), 1 deletion(-)

Index: quilt-2.6/drivers/s390/cio/chsc.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/chsc.c	2007-04-27 16:04:55.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/chsc.c	2007-04-27 16:04:55.000000000 +0200
@@ -818,7 +818,7 @@
 		} else
 			chsc_remove_cmg_attr(css);
 	}
-	if (enable && !css->cm_enabled) {
+	if (!css->cm_enabled) {
 		free_page((unsigned long)css->cub_addr1);
 		free_page((unsigned long)css->cub_addr2);
 	}

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 16/38] cio: Unregister ccw devices directly.

[Martin Schwidefsky]

[-- Attachment #1: 114-cio-unreg.diff --]
[-- Type: text/plain, Size: 4086 bytes --]

From: Cornelia Huck <cornelia.huck@de.ibm.com>

We used to unregister ccw devices not directly from the I/O
subchannel remove function in order to avoid lifelocks on the
css bus semaphore. This semaphore is gone, and there is no reason
to not unregister the ccw device directly (it is even better since
it is more in keeping with the goal of immediate disconnect).

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/cio/device.c |   56 +++++++++-------------------------------------
 1 files changed, 12 insertions(+), 44 deletions(-)

Index: quilt-2.6/drivers/s390/cio/device.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/device.c	2007-04-27 16:04:55.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/device.c	2007-04-27 16:04:56.000000000 +0200
@@ -290,16 +290,10 @@
 	return sch_is_pseudo_sch(to_subchannel(cdev->dev.parent));
 }
 
-static void ccw_device_unregister(struct work_struct *work)
+static void ccw_device_unregister(struct ccw_device *cdev)
 {
-	struct ccw_device_private *priv;
-	struct ccw_device *cdev;
-
-	priv = container_of(work, struct ccw_device_private, kick_work);
-	cdev = priv->cdev;
 	if (test_and_clear_bit(1, &cdev->private->registered))
-		device_unregister(&cdev->dev);
-	put_device(&cdev->dev);
+		device_del(&cdev->dev);
 }
 
 static void
@@ -316,11 +310,8 @@
 		spin_lock_irqsave(cdev->ccwlock, flags);
 		cdev->private->state = DEV_STATE_NOT_OPER;
 		spin_unlock_irqrestore(cdev->ccwlock, flags);
-		if (get_device(&cdev->dev)) {
-			PREPARE_WORK(&cdev->private->kick_work,
-				     ccw_device_unregister);
-			queue_work(ccw_device_work, &cdev->private->kick_work);
-		}
+		ccw_device_unregister(cdev);
+		put_device(&cdev->dev);
 		return ;
 	}
 	sch = to_subchannel(cdev->dev.parent);
@@ -555,17 +546,10 @@
 	.attrs = ccwdev_attrs,
 };
 
-static int
-device_add_files (struct device *dev)
-{
-	return sysfs_create_group(&dev->kobj, &ccwdev_attr_group);
-}
-
-static void
-device_remove_files(struct device *dev)
-{
-	sysfs_remove_group(&dev->kobj, &ccwdev_attr_group);
-}
+struct attribute_group *ccwdev_attr_groups[] = {
+	&ccwdev_attr_group,
+	NULL,
+};
 
 /* this is a simple abstraction for device_register that sets the
  * correct bus type and adds the bus specific files */
@@ -580,10 +564,6 @@
 		return ret;
 
 	set_bit(1, &cdev->private->registered);
-	if ((ret = device_add_files(dev))) {
-		if (test_and_clear_bit(1, &cdev->private->registered))
-			device_del(dev);
-	}
 	return ret;
 }
 
@@ -655,10 +635,6 @@
 		return;
 	}
 	set_bit(1, &cdev->private->registered);
-	if (device_add_files(&cdev->dev)) {
-		if (test_and_clear_bit(1, &cdev->private->registered))
-			device_unregister(&cdev->dev);
-	}
 }
 
 void ccw_device_do_unreg_rereg(struct work_struct *work)
@@ -671,9 +647,7 @@
 	cdev = priv->cdev;
 	sch = to_subchannel(cdev->dev.parent);
 
-	device_remove_files(&cdev->dev);
-	if (test_and_clear_bit(1, &cdev->private->registered))
-		device_del(&cdev->dev);
+	ccw_device_unregister(cdev);
 	PREPARE_WORK(&cdev->private->kick_work,
 		     ccw_device_add_changed);
 	queue_work(ccw_device_work, &cdev->private->kick_work);
@@ -712,6 +686,7 @@
 	cdev->dev.parent = &sch->dev;
 	cdev->dev.release = ccw_device_release;
 	INIT_LIST_HEAD(&cdev->private->kick_work.entry);
+	cdev->dev.groups = ccwdev_attr_groups;
 	/* Do first half of device_register. */
 	device_initialize(&cdev->dev);
 	if (!get_device(&sch->dev)) {
@@ -1141,15 +1116,8 @@
 	sch->dev.driver_data = NULL;
 	cdev->private->state = DEV_STATE_NOT_OPER;
 	spin_unlock_irqrestore(cdev->ccwlock, flags);
-	/*
-	 * Put unregistration on workqueue to avoid livelocks on the css bus
-	 * semaphore.
-	 */
-	if (get_device(&cdev->dev)) {
-		PREPARE_WORK(&cdev->private->kick_work,
-			     ccw_device_unregister);
-		queue_work(ccw_device_work, &cdev->private->kick_work);
-	}
+	ccw_device_unregister(cdev);
+	put_device(&cdev->dev);
 	return 0;
 }
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 17/38] System call cleanup.

[Martin Schwidefsky]

[-- Attachment #1: 120-svc-glue.diff --]
[-- Type: text/plain, Size: 20603 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Remove system call glue for sys_clone, sys_fork, sys_vfork, sys_execve,
sys_sigreturn, sys_rt_sigreturn and sys_sigaltstack. Call do_execve from
kernel_execve directly, move pt_regs to the right place and branch to
sysc_return to start the user space program. This removes the last
in-kernel system call.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/kernel/compat_linux.c  |   62 +++++++++++++-----------
 arch/s390/kernel/compat_signal.c |   14 +++--
 arch/s390/kernel/entry.S         |   87 ++++++++++++---------------------
 arch/s390/kernel/entry64.S       |  100 ++++++++++-----------------------------
 arch/s390/kernel/process.c       |   82 ++++++++++++++++++-------------
 arch/s390/kernel/signal.c        |   10 ++-
 arch/s390/kernel/sys_s390.c      |   20 -------
 arch/s390/kernel/syscalls.S      |   14 ++---
 8 files changed, 163 insertions(+), 226 deletions(-)

Index: quilt-2.6/arch/s390/kernel/compat_linux.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/compat_linux.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/compat_linux.c	2007-04-27 16:04:56.000000000 +0200
@@ -495,29 +495,34 @@
  * sys32_execve() executes a new program after the asm stub has set
  * things up for us.  This should basically do what I want it to.
  */
-asmlinkage long
-sys32_execve(struct pt_regs regs)
+asmlinkage long sys32_execve(void)
 {
-        int error;
-        char * filename;
-
-        filename = getname(compat_ptr(regs.orig_gpr2));
-        error = PTR_ERR(filename);
-        if (IS_ERR(filename))
+	struct pt_regs *regs = task_pt_regs(current);
+	char *filename;
+	unsigned long result;
+	int rc;
+
+	filename = getname(compat_ptr(regs->orig_gpr2));
+	if (IS_ERR(filename)) {
+		result = PTR_ERR(filename);
                 goto out;
-        error = compat_do_execve(filename, compat_ptr(regs.gprs[3]),
-				 compat_ptr(regs.gprs[4]), &regs);
-	if (error == 0)
-	{
-		task_lock(current);
-		current->ptrace &= ~PT_DTRACE;
-		task_unlock(current);
-		current->thread.fp_regs.fpc=0;
-		asm volatile("sfpc %0,0" : : "d" (0));
 	}
+	rc = compat_do_execve(filename, compat_ptr(regs->gprs[3]),
+			      compat_ptr(regs->gprs[4]), regs);
+	if (rc) {
+		result = rc;
+		goto out_putname;
+	}
+	task_lock(current);
+	current->ptrace &= ~PT_DTRACE;
+	task_unlock(current);
+	current->thread.fp_regs.fpc=0;
+	asm volatile("sfpc %0,0" : : "d" (0));
+	result = regs->gprs[2];
+out_putname:
         putname(filename);
 out:
-        return error;
+	return result;
 }
 
 
@@ -918,19 +923,20 @@
 	return sys_write(fd, buf, count);
 }
 
-asmlinkage long sys32_clone(struct pt_regs regs)
+asmlinkage long sys32_clone(void)
 {
-        unsigned long clone_flags;
-        unsigned long newsp;
+	struct pt_regs *regs = task_pt_regs(current);
+	unsigned long clone_flags;
+	unsigned long newsp;
 	int __user *parent_tidptr, *child_tidptr;
 
-        clone_flags = regs.gprs[3] & 0xffffffffUL;
-        newsp = regs.orig_gpr2 & 0x7fffffffUL;
-	parent_tidptr = compat_ptr(regs.gprs[4]);
-	child_tidptr = compat_ptr(regs.gprs[5]);
-        if (!newsp)
-                newsp = regs.gprs[15];
-        return do_fork(clone_flags, newsp, &regs, 0,
+	clone_flags = regs->gprs[3] & 0xffffffffUL;
+	newsp = regs->orig_gpr2 & 0x7fffffffUL;
+	parent_tidptr = compat_ptr(regs->gprs[4]);
+	child_tidptr = compat_ptr(regs->gprs[5]);
+	if (!newsp)
+		newsp = regs->gprs[15];
+	return do_fork(clone_flags, newsp, regs, 0,
 		       parent_tidptr, child_tidptr);
 }
 
Index: quilt-2.6/arch/s390/kernel/compat_signal.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/compat_signal.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/compat_signal.c	2007-04-27 16:04:56.000000000 +0200
@@ -255,9 +255,9 @@
 }
 
 asmlinkage long
-sys32_sigaltstack(const stack_t32 __user *uss, stack_t32 __user *uoss,
-							struct pt_regs *regs)
+sys32_sigaltstack(const stack_t32 __user *uss, stack_t32 __user *uoss)
 {
+	struct pt_regs *regs = task_pt_regs(current);
 	stack_t kss, koss;
 	unsigned long ss_sp;
 	int ret, err = 0;
@@ -344,8 +344,9 @@
 	return 0;
 }
 
-asmlinkage long sys32_sigreturn(struct pt_regs *regs)
+asmlinkage long sys32_sigreturn(void)
 {
+	struct pt_regs *regs = task_pt_regs(current);
 	sigframe32 __user *frame = (sigframe32 __user *)regs->gprs[15];
 	sigset_t set;
 
@@ -370,8 +371,9 @@
 	return 0;
 }
 
-asmlinkage long sys32_rt_sigreturn(struct pt_regs *regs)
+asmlinkage long sys32_rt_sigreturn(void)
 {
+	struct pt_regs *regs = task_pt_regs(current);
 	rt_sigframe32 __user *frame = (rt_sigframe32 __user *)regs->gprs[15];
 	sigset_t set;
 	stack_t st;
@@ -407,8 +409,8 @@
 	return regs->gprs[2];
 
 badframe:
-        force_sig(SIGSEGV, current);
-        return 0;
+	force_sig(SIGSEGV, current);
+	return 0;
 }	
 
 /*
Index: quilt-2.6/arch/s390/kernel/entry64.S
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/entry64.S	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/entry64.S	2007-04-27 16:04:56.000000000 +0200
@@ -244,8 +244,6 @@
 	jnz	sysc_tracesys
 	basr	%r14,%r8	# call sys_xxxx
 	stg	%r2,SP_R2(%r15) # store return value (change R2 on stack)
-				# ATTENTION: check sys_execve_glue before
-				# changing anything here !!
 
 sysc_return:
 	tm	SP_PSW+1(%r15),0x01	# returning to user ?
@@ -371,77 +369,35 @@
 	j	sysc_return
 
 #
-# clone, fork, vfork, exec and sigreturn need glue,
-# because they all expect pt_regs as parameter,
-# but are called with different parameter.
-# return-address is set up above
+# kernel_execve function needs to deal with pt_regs that is not
+# at the usual place
 #
-sys_clone_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	jg	sys_clone		# branch to sys_clone
-
-#ifdef CONFIG_COMPAT
-sys32_clone_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	jg	sys32_clone		# branch to sys32_clone
-#endif
-
-sys_fork_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	jg	sys_fork		# branch to sys_fork
-
-sys_vfork_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	jg	sys_vfork		# branch to sys_vfork
-
-sys_execve_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	lgr	%r12,%r14		# save return address
-	brasl	%r14,sys_execve 	# call sys_execve
-	ltgr	%r2,%r2 		# check if execve failed
-	bnz	0(%r12) 		# it did fail -> store result in gpr2
-	b	6(%r12) 		# SKIP STG 2,SP_R2(15) in
-					# system_call/sysc_tracesys
-#ifdef CONFIG_COMPAT
-sys32_execve_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	lgr	%r12,%r14		# save return address
-	brasl	%r14,sys32_execve	# call sys32_execve
-	ltgr	%r2,%r2 		# check if execve failed
-	bnz	0(%r12) 		# it did fail -> store result in gpr2
-	b	6(%r12) 		# SKIP STG 2,SP_R2(15) in
-					# system_call/sysc_tracesys
-#endif
-
-sys_sigreturn_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs as parameter
-	jg	sys_sigreturn		# branch to sys_sigreturn
-
-#ifdef CONFIG_COMPAT
-sys32_sigreturn_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs as parameter
-	jg	sys32_sigreturn 	# branch to sys32_sigreturn
-#endif
-
-sys_rt_sigreturn_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs as parameter
-	jg	sys_rt_sigreturn	# branch to sys_sigreturn
-
-#ifdef CONFIG_COMPAT
-sys32_rt_sigreturn_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs as parameter
-	jg	sys32_rt_sigreturn	# branch to sys32_sigreturn
-#endif
-
-sys_sigaltstack_glue:
-	la	%r4,SP_PTREGS(%r15)	# load pt_regs as parameter
-	jg	sys_sigaltstack 	# branch to sys_sigreturn
-
-#ifdef CONFIG_COMPAT
-sys32_sigaltstack_glue:
-	la	%r4,SP_PTREGS(%r15)	# load pt_regs as parameter
-	jg	sys32_sigaltstack_wrapper # branch to sys_sigreturn
-#endif
+	.globl	kernel_execve
+kernel_execve:
+	stmg	%r12,%r15,96(%r15)
+	lgr	%r14,%r15
+	aghi	%r15,-SP_SIZE
+	stg	%r14,__SF_BACKCHAIN(%r15)
+	la	%r12,SP_PTREGS(%r15)
+	xc	0(__PT_SIZE,%r12),0(%r12)
+	lgr	%r5,%r12
+	brasl	%r14,do_execve
+	ltgfr	%r2,%r2
+	je	0f
+	aghi	%r15,SP_SIZE
+	lmg	%r12,%r15,96(%r15)
+	br	%r14
+	# execve succeeded.
+0:	stnsm	__SF_EMPTY(%r15),0xfc	# disable interrupts
+	lg	%r15,__LC_KERNEL_STACK	# load ksp
+	aghi	%r15,-SP_SIZE		# make room for registers & psw
+	lg	%r13,__LC_SVC_NEW_PSW+8
+	lg	%r9,__LC_THREAD_INFO
+	mvc	SP_PTREGS(__PT_SIZE,%r15),0(%r12)	# copy pt_regs
+	xc	__SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
+	stosm	__SF_EMPTY(%r15),0x03	# reenable interrupts
+	brasl	%r14,execve_tail
+	j	sysc_return
 
 /*
  * Program check handler routine
Index: quilt-2.6/arch/s390/kernel/entry.S
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/entry.S	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/entry.S	2007-04-27 16:04:56.000000000 +0200
@@ -249,8 +249,6 @@
 	bnz	BASED(sysc_tracesys)
 	basr	%r14,%r8	  # call sys_xxxx
 	st	%r2,SP_R2(%r15)   # store return value (change R2 on stack)
-				  # ATTENTION: check sys_execve_glue before
-				  # changing anything here !!
 
 sysc_return:
 	tm	SP_PSW+1(%r15),0x01	# returning to user ?
@@ -381,50 +379,37 @@
 	b	BASED(sysc_return)
 
 #
-# clone, fork, vfork, exec and sigreturn need glue,
-# because they all expect pt_regs as parameter,
-# but are called with different parameter.
-# return-address is set up above
+# kernel_execve function needs to deal with pt_regs that is not
+# at the usual place
 #
-sys_clone_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	l	%r1,BASED(.Lclone)
-	br	%r1			# branch to sys_clone
-
-sys_fork_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	l	%r1,BASED(.Lfork)
-	br	%r1			# branch to sys_fork
-
-sys_vfork_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	l	%r1,BASED(.Lvfork)
-	br	%r1			# branch to sys_vfork
-
-sys_execve_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs
-	l	%r1,BASED(.Lexecve)
-	lr	%r12,%r14		# save return address
-	basr	%r14,%r1		# call sys_execve
-	ltr	%r2,%r2			# check if execve failed
-	bnz	0(%r12)			# it did fail -> store result in gpr2
-	b	4(%r12)			# SKIP ST 2,SP_R2(15) after BASR 14,8
-					# in system_call/sysc_tracesys
-
-sys_sigreturn_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs as parameter
-	l	%r1,BASED(.Lsigreturn)
-	br	%r1			# branch to sys_sigreturn
-
-sys_rt_sigreturn_glue:
-	la	%r2,SP_PTREGS(%r15)	# load pt_regs as parameter
-	l	%r1,BASED(.Lrt_sigreturn)
-	br	%r1			# branch to sys_sigreturn
-
-sys_sigaltstack_glue:
-	la	%r4,SP_PTREGS(%r15)	# load pt_regs as parameter
-	l	%r1,BASED(.Lsigaltstack)
-	br	%r1			# branch to sys_sigreturn
+	.globl	kernel_execve
+kernel_execve:
+	stm	%r12,%r15,48(%r15)
+	lr	%r14,%r15
+	l	%r13,__LC_SVC_NEW_PSW+4
+	s	%r15,BASED(.Lc_spsize)
+	st	%r14,__SF_BACKCHAIN(%r15)
+	la	%r12,SP_PTREGS(%r15)
+	xc	0(__PT_SIZE,%r12),0(%r12)
+	l	%r1,BASED(.Ldo_execve)
+	lr	%r5,%r12
+	basr	%r14,%r1
+	ltr	%r2,%r2
+	be	BASED(0f)
+	a	%r15,BASED(.Lc_spsize)
+	lm	%r12,%r15,48(%r15)
+	br	%r14
+	# execve succeeded.
+0:	stnsm	__SF_EMPTY(%r15),0xfc	# disable interrupts
+	l	%r15,__LC_KERNEL_STACK	# load ksp
+	s	%r15,BASED(.Lc_spsize)	# make room for registers & psw
+	l	%r9,__LC_THREAD_INFO
+	mvc	SP_PTREGS(__PT_SIZE,%r15),0(%r12)	# copy pt_regs
+	xc	__SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15)
+	stosm	__SF_EMPTY(%r15),0x03	# reenable interrupts
+	l	%r1,BASED(.Lexecve_tail)
+	basr	%r14,%r1
+	b	BASED(sysc_return)
 
 /*
  * Program check handler routine
@@ -1031,19 +1016,11 @@
 .Ldo_extint:	.long	do_extint
 .Ldo_signal:	.long	do_signal
 .Lhandle_per:	.long	do_single_step
+.Ldo_execve:	.long	do_execve
+.Lexecve_tail:	.long	execve_tail
 .Ljump_table:	.long	pgm_check_table
 .Lschedule:	.long	schedule
-.Lclone:	.long	sys_clone
-.Lexecve:	.long	sys_execve
-.Lfork: 	.long	sys_fork
-.Lrt_sigreturn: .long	sys_rt_sigreturn
-.Lrt_sigsuspend:
-		.long	sys_rt_sigsuspend
-.Lsigreturn:	.long	sys_sigreturn
-.Lsigsuspend:	.long	sys_sigsuspend
-.Lsigaltstack:	.long	sys_sigaltstack
 .Ltrace:	.long	syscall_trace
-.Lvfork:	.long	sys_vfork
 .Lschedtail:	.long	schedule_tail
 .Lsysc_table:	.long	sys_call_table
 #ifdef CONFIG_TRACE_IRQFLAGS
Index: quilt-2.6/arch/s390/kernel/process.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/process.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/process.c	2007-04-27 16:04:56.000000000 +0200
@@ -280,24 +280,26 @@
         return 0;
 }
 
-asmlinkage long sys_fork(struct pt_regs regs)
+asmlinkage long sys_fork(void)
 {
-	return do_fork(SIGCHLD, regs.gprs[15], &regs, 0, NULL, NULL);
+	struct pt_regs *regs = task_pt_regs(current);
+	return do_fork(SIGCHLD, regs->gprs[15], regs, 0, NULL, NULL);
 }
 
-asmlinkage long sys_clone(struct pt_regs regs)
+asmlinkage long sys_clone(void)
 {
-        unsigned long clone_flags;
-        unsigned long newsp;
+	struct pt_regs *regs = task_pt_regs(current);
+	unsigned long clone_flags;
+	unsigned long newsp;
 	int __user *parent_tidptr, *child_tidptr;
 
-        clone_flags = regs.gprs[3];
-        newsp = regs.orig_gpr2;
-	parent_tidptr = (int __user *) regs.gprs[4];
-	child_tidptr = (int __user *) regs.gprs[5];
-        if (!newsp)
-                newsp = regs.gprs[15];
-        return do_fork(clone_flags, newsp, &regs, 0,
+	clone_flags = regs->gprs[3];
+	newsp = regs->orig_gpr2;
+	parent_tidptr = (int __user *) regs->gprs[4];
+	child_tidptr = (int __user *) regs->gprs[5];
+	if (!newsp)
+		newsp = regs->gprs[15];
+	return do_fork(clone_flags, newsp, regs, 0,
 		       parent_tidptr, child_tidptr);
 }
 
@@ -311,40 +313,52 @@
  * do not have enough call-clobbered registers to hold all
  * the information you need.
  */
-asmlinkage long sys_vfork(struct pt_regs regs)
+asmlinkage long sys_vfork(void)
 {
+	struct pt_regs *regs = task_pt_regs(current);
 	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD,
-		       regs.gprs[15], &regs, 0, NULL, NULL);
+		       regs->gprs[15], regs, 0, NULL, NULL);
+}
+
+asmlinkage void execve_tail(void)
+{
+	task_lock(current);
+	current->ptrace &= ~PT_DTRACE;
+	task_unlock(current);
+	current->thread.fp_regs.fpc = 0;
+	if (MACHINE_HAS_IEEE)
+		asm volatile("sfpc %0,%0" : : "d" (0));
 }
 
 /*
  * sys_execve() executes a new program.
  */
-asmlinkage long sys_execve(struct pt_regs regs)
+asmlinkage long sys_execve(void)
 {
-        int error;
-        char * filename;
-
-        filename = getname((char __user *) regs.orig_gpr2);
-        error = PTR_ERR(filename);
-        if (IS_ERR(filename))
-                goto out;
-        error = do_execve(filename, (char __user * __user *) regs.gprs[3],
-			  (char __user * __user *) regs.gprs[4], &regs);
-	if (error == 0) {
-		task_lock(current);
-		current->ptrace &= ~PT_DTRACE;
-		task_unlock(current);
-		current->thread.fp_regs.fpc = 0;
-		if (MACHINE_HAS_IEEE)
-			asm volatile("sfpc %0,%0" : : "d" (0));
+	struct pt_regs *regs = task_pt_regs(current);
+	char *filename;
+	unsigned long result;
+	int rc;
+
+	filename = getname((char __user *) regs->orig_gpr2);
+	if (IS_ERR(filename)) {
+		result = PTR_ERR(filename);
+		goto out;
+	}
+	rc = do_execve(filename, (char __user * __user *) regs->gprs[3],
+		       (char __user * __user *) regs->gprs[4], regs);
+	if (rc) {
+		result = rc;
+		goto out_putname;
 	}
-        putname(filename);
+	execve_tail();
+	result = regs->gprs[2];
+out_putname:
+	putname(filename);
 out:
-        return error;
+	return result;
 }
 
-
 /*
  * fill in the FPU structure for a core dump.
  */
Index: quilt-2.6/arch/s390/kernel/signal.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/signal.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/signal.c	2007-04-27 16:04:56.000000000 +0200
@@ -102,9 +102,9 @@
 }
 
 asmlinkage long
-sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss,
-					struct pt_regs *regs)
+sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss)
 {
+	struct pt_regs *regs = task_pt_regs(current);
 	return do_sigaltstack(uss, uoss, regs->gprs[15]);
 }
 
@@ -163,8 +163,9 @@
 	return 0;
 }
 
-asmlinkage long sys_sigreturn(struct pt_regs *regs)
+asmlinkage long sys_sigreturn(void)
 {
+	struct pt_regs *regs = task_pt_regs(current);
 	sigframe __user *frame = (sigframe __user *)regs->gprs[15];
 	sigset_t set;
 
@@ -189,8 +190,9 @@
 	return 0;
 }
 
-asmlinkage long sys_rt_sigreturn(struct pt_regs *regs)
+asmlinkage long sys_rt_sigreturn(void)
 {
+	struct pt_regs *regs = task_pt_regs(current);
 	rt_sigframe __user *frame = (rt_sigframe __user *)regs->gprs[15];
 	sigset_t set;
 
Index: quilt-2.6/arch/s390/kernel/syscalls.S
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/syscalls.S	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/syscalls.S	2007-04-27 16:04:56.000000000 +0200
@@ -10,7 +10,7 @@
 
 NI_SYSCALL							/* 0 */
 SYSCALL(sys_exit,sys_exit,sys32_exit_wrapper)
-SYSCALL(sys_fork_glue,sys_fork_glue,sys_fork_glue)
+SYSCALL(sys_fork,sys_fork,sys_fork)
 SYSCALL(sys_read,sys_read,sys32_read_wrapper)
 SYSCALL(sys_write,sys_write,sys32_write_wrapper)
 SYSCALL(sys_open,sys_open,sys32_open_wrapper)			/* 5 */
@@ -19,7 +19,7 @@
 SYSCALL(sys_creat,sys_creat,sys32_creat_wrapper)
 SYSCALL(sys_link,sys_link,sys32_link_wrapper)
 SYSCALL(sys_unlink,sys_unlink,sys32_unlink_wrapper)		/* 10 */
-SYSCALL(sys_execve_glue,sys_execve_glue,sys32_execve_glue)
+SYSCALL(sys_execve,sys_execve,sys32_execve)
 SYSCALL(sys_chdir,sys_chdir,sys32_chdir_wrapper)
 SYSCALL(sys_time,sys_ni_syscall,sys32_time_wrapper)		/* old time syscall */
 SYSCALL(sys_mknod,sys_mknod,sys32_mknod_wrapper)
@@ -127,8 +127,8 @@
 SYSCALL(sys_sysinfo,sys_sysinfo,compat_sys_sysinfo_wrapper)
 SYSCALL(sys_ipc,sys_ipc,sys32_ipc_wrapper)
 SYSCALL(sys_fsync,sys_fsync,sys32_fsync_wrapper)
-SYSCALL(sys_sigreturn_glue,sys_sigreturn_glue,sys32_sigreturn_glue)
-SYSCALL(sys_clone_glue,sys_clone_glue,sys32_clone_glue)		/* 120 */
+SYSCALL(sys_sigreturn,sys_sigreturn,sys32_sigreturn)
+SYSCALL(sys_clone,sys_clone,sys32_clone)			/* 120 */
 SYSCALL(sys_setdomainname,sys_setdomainname,sys32_setdomainname_wrapper)
 SYSCALL(sys_newuname,s390x_newuname,sys32_newuname_wrapper)
 NI_SYSCALL							/* modify_ldt for i386 */
@@ -181,7 +181,7 @@
 SYSCALL(sys_setresgid16,sys_ni_syscall,sys32_setresgid16_wrapper)	/* 170 old setresgid16 syscall */
 SYSCALL(sys_getresgid16,sys_ni_syscall,sys32_getresgid16_wrapper)	/* old getresgid16 syscall */
 SYSCALL(sys_prctl,sys_prctl,sys32_prctl_wrapper)
-SYSCALL(sys_rt_sigreturn_glue,sys_rt_sigreturn_glue,sys32_rt_sigreturn_glue)
+SYSCALL(sys_rt_sigreturn,sys_rt_sigreturn,sys32_rt_sigreturn)
 SYSCALL(sys_rt_sigaction,sys_rt_sigaction,sys32_rt_sigaction_wrapper)
 SYSCALL(sys_rt_sigprocmask,sys_rt_sigprocmask,sys32_rt_sigprocmask_wrapper)	/* 175 */
 SYSCALL(sys_rt_sigpending,sys_rt_sigpending,sys32_rt_sigpending_wrapper)
@@ -194,11 +194,11 @@
 SYSCALL(sys_getcwd,sys_getcwd,sys32_getcwd_wrapper)
 SYSCALL(sys_capget,sys_capget,sys32_capget_wrapper)
 SYSCALL(sys_capset,sys_capset,sys32_capset_wrapper)		/* 185 */
-SYSCALL(sys_sigaltstack_glue,sys_sigaltstack_glue,sys32_sigaltstack_glue)
+SYSCALL(sys_sigaltstack,sys_sigaltstack,sys32_sigaltstack)
 SYSCALL(sys_sendfile,sys_sendfile64,sys32_sendfile_wrapper)
 NI_SYSCALL							/* streams1 */
 NI_SYSCALL							/* streams2 */
-SYSCALL(sys_vfork_glue,sys_vfork_glue,sys_vfork_glue)		/* 190 */
+SYSCALL(sys_vfork,sys_vfork,sys_vfork)				/* 190 */
 SYSCALL(sys_getrlimit,sys_getrlimit,compat_sys_getrlimit_wrapper)
 SYSCALL(sys_mmap2,sys_mmap2,sys32_mmap2_wrapper)
 SYSCALL(sys_truncate64,sys_ni_syscall,sys32_truncate64_wrapper)
Index: quilt-2.6/arch/s390/kernel/sys_s390.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/sys_s390.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/sys_s390.c	2007-04-27 16:04:56.000000000 +0200
@@ -266,23 +266,3 @@
 		return -EFAULT;
 	return sys_fadvise64_64(a.fd, a.offset, a.len, a.advice);
 }
-
-/*
- * Do a system call from kernel instead of calling sys_execve so we
- * end up with proper pt_regs.
- */
-int kernel_execve(const char *filename, char *const argv[], char *const envp[])
-{
-	register const char *__arg1 asm("2") = filename;
-	register char *const*__arg2 asm("3") = argv;
-	register char *const*__arg3 asm("4") = envp;
-	register long __svcres asm("2");
-	asm volatile(
-		"svc %b1"
-		: "=d" (__svcres)
-		: "i" (__NR_execve),
-		  "0" (__arg1),
-		  "d" (__arg2),
-		  "d" (__arg3) : "memory");
-	return __svcres;
-}

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 18/38] Improved oops output.

[Martin Schwidefsky]

[-- Attachment #1: 121-oops.diff --]
[-- Type: text/plain, Size: 49321 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

This patch adds two improvements to the oops output. First it adds an
additional line after the PSW which decodes the different fields of it.
Second a disassembler is added that decodes the instructions surrounding
the faulting PSW. The output of a test oops now looks like this:

kernel BUG at init/main.c:419
illegal operation: 0001 [#1]
CPU:    0    Not tainted
Process swapper (pid: 0, task: 0000000000464968, ksp: 00000000004be000)
Krnl PSW : 0700000180000000 00000000000120b6 (rest_init+0x36/0x38)
           R:0 T:1 IO:1 EX:1 Key:0 M:0 W:0 P:0 AS:0 CC:0 PM:0 EA:3
Krnl GPRS: 0000000000000003 00000000004ba017 0000000000000022 0000000000000001
           000000000003a5f6 0000000000000000 00000000004be6a8 0000000000000000
           0000000000000000 00000000004b8200 0000000000003a50 0000000000008000
           0000000000516368 000000000033d008 00000000000120b2 00000000004bdee0
Krnl Code: 00000000000120a6: e3e0f0980024       stg     %r14,152(%r15)
           00000000000120ac: c0e500014296       brasl   %r14,3a5d8
           00000000000120b2: a7f40001           brc     15,120b4
          >00000000000120b6: 0707               bcr     0,%r7
           00000000000120b8: eb7ff0500024       stmg    %r7,%r15,80(%r15)
           00000000000120be: c0d000195825       larl    %r13,33d108
           00000000000120c4: a7f13f00           tmll    %r15,16128
           00000000000120c8: a7840001           brc     8,120ca
Call Trace:
([<00000000000120b2>] rest_init+0x32/0x38)
 [<00000000004be614>] start_kernel+0x37c/0x410
 [<0000000000012020>] _ehead+0x20/0x80

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/kernel/Makefile    |    2 
 arch/s390/kernel/dis.c       | 1278 +++++++++++++++++++++++++++++++++++++++++++
 arch/s390/kernel/traps.c     |   55 -
 include/asm-s390/processor.h |    1 
 4 files changed, 1297 insertions(+), 39 deletions(-)

Index: quilt-2.6/arch/s390/kernel/dis.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/arch/s390/kernel/dis.c	2007-04-27 16:04:56.000000000 +0200
@@ -0,0 +1,1278 @@
+/*
+ * arch/s390/kernel/dis.c
+ *
+ * Disassemble s390 instructions.
+ *
+ * Copyright IBM Corp. 2007
+ * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
+ */
+
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/ptrace.h>
+#include <linux/timer.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/kallsyms.h>
+#include <linux/reboot.h>
+#include <linux/kprobes.h>
+
+#include <asm/system.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+#include <asm/atomic.h>
+#include <asm/mathemu.h>
+#include <asm/cpcmd.h>
+#include <asm/s390_ext.h>
+#include <asm/lowcore.h>
+#include <asm/debug.h>
+#include <asm/kdebug.h>
+
+#ifndef CONFIG_64BIT
+#define ONELONG "%08lx: "
+#else /* CONFIG_64BIT */
+#define ONELONG "%016lx: "
+#endif /* CONFIG_64BIT */
+
+#define OPERAND_GPR	0x1	/* Operand printed as %rx */
+#define OPERAND_FPR	0x2	/* Operand printed as %fx */
+#define OPERAND_AR	0x4	/* Operand printed as %ax */
+#define OPERAND_CR	0x8	/* Operand printed as %cx */
+#define OPERAND_DISP	0x10	/* Operand printed as displacement */
+#define OPERAND_BASE	0x20	/* Operand printed as base register */
+#define OPERAND_INDEX	0x40	/* Operand printed as index register */
+#define OPERAND_PCREL	0x80	/* Operand printed as pc-relative symbol */
+#define OPERAND_SIGNED	0x100	/* Operand printed as signed value */
+#define OPERAND_LENGTH	0x200	/* Operand printed as length (+1) */
+
+enum {
+	UNUSED,	/* Indicates the end of the operand list */
+	R_8,	/* GPR starting at position 8 */
+	R_12,	/* GPR starting at position 12 */
+	R_16,	/* GPR starting at position 16 */
+	R_20,	/* GPR starting at position 20 */
+	R_24,	/* GPR starting at position 24 */
+	R_28,	/* GPR starting at position 28 */
+	R_32,	/* GPR starting at position 32 */
+	F_8,	/* FPR starting at position 8 */
+	F_12,	/* FPR starting at position 12 */
+	F_16,	/* FPR starting at position 16 */
+	F_20,	/* FPR starting at position 16 */
+	F_24,	/* FPR starting at position 24 */
+	F_28,	/* FPR starting at position 28 */
+	F_32,	/* FPR starting at position 32 */
+	A_8,	/* Access reg. starting at position 8 */
+	A_12,	/* Access reg. starting at position 12 */
+	A_24,	/* Access reg. starting at position 24 */
+	A_28,	/* Access reg. starting at position 28 */
+	C_8,	/* Control reg. starting at position 8 */
+	C_12,	/* Control reg. starting at position 12 */
+	B_16,	/* Base register starting at position 16 */
+	B_32,	/* Base register starting at position 32 */
+	X_12,	/* Index register starting at position 12 */
+	D_20,	/* Displacement starting at position 20 */
+	D_36,	/* Displacement starting at position 36 */
+	D20_20,	/* 20 bit displacement starting at 20 */
+	L4_8,	/* 4 bit length starting at position 8 */
+	L4_12,	/* 4 bit length starting at position 12 */
+	L8_8,	/* 8 bit length starting at position 8 */
+	U4_8,	/* 4 bit unsigned value starting at 8 */
+	U4_12,	/* 4 bit unsigned value starting at 12 */
+	U4_16,	/* 4 bit unsigned value starting at 16 */
+	U4_20,	/* 4 bit unsigned value starting at 20 */
+	U8_8,	/* 8 bit unsigned value starting at 8 */
+	U8_16,	/* 8 bit unsigned value starting at 16 */
+	I16_16,	/* 16 bit signed value starting at 16 */
+	U16_16,	/* 16 bit unsigned value starting at 16 */
+	J16_16,	/* PC relative jump offset at 16 */
+	J32_16,	/* PC relative long offset at 16 */
+	I32_16,	/* 32 bit signed value starting at 16 */
+	U32_16,	/* 32 bit unsigned value starting at 16 */
+	M_16,	/* 4 bit optional mask starting at 16 */
+	RO_28,	/* optional GPR starting at position 28 */
+};
+
+/*
+ * Enumeration of the different instruction formats.
+ * For details consult the principles of operation.
+ */
+enum {
+	INSTR_INVALID,
+	INSTR_E, INSTR_RIE_RRP, INSTR_RIL_RI, INSTR_RIL_RP, INSTR_RIL_RU,
+	INSTR_RIL_UP, INSTR_RI_RI, INSTR_RI_RP, INSTR_RI_RU, INSTR_RI_UP,
+	INSTR_RRE_00, INSTR_RRE_0R, INSTR_RRE_AA, INSTR_RRE_AR, INSTR_RRE_F0,
+	INSTR_RRE_FF, INSTR_RRE_R0, INSTR_RRE_RA, INSTR_RRE_RF, INSTR_RRE_RR,
+	INSTR_RRE_RR_OPT, INSTR_RRF_F0FF, INSTR_RRF_FUFF, INSTR_RRF_M0RR,
+	INSTR_RRF_R0RR, INSTR_RRF_RURR, INSTR_RRF_U0FF, INSTR_RRF_U0RF,
+	INSTR_RR_FF, INSTR_RR_R0, INSTR_RR_RR, INSTR_RR_U0, INSTR_RR_UR,
+	INSTR_RSE_CCRD, INSTR_RSE_RRRD, INSTR_RSE_RURD, INSTR_RSI_RRP,
+	INSTR_RSL_R0RD, INSTR_RSY_AARD, INSTR_RSY_CCRD, INSTR_RSY_RRRD,
+	INSTR_RSY_RURD, INSTR_RS_AARD, INSTR_RS_CCRD, INSTR_RS_R0RD,
+	INSTR_RS_RRRD, INSTR_RS_RURD, INSTR_RXE_FRRD, INSTR_RXE_RRRD,
+	INSTR_RXF_FRRDF, INSTR_RXY_FRRD, INSTR_RXY_RRRD, INSTR_RX_FRRD,
+	INSTR_RX_RRRD, INSTR_RX_URRD, INSTR_SIY_URD, INSTR_SI_URD,
+	INSTR_SSE_RDRD, INSTR_SSF_RRDRD, INSTR_SS_L0RDRD, INSTR_SS_LIRDRD,
+	INSTR_SS_LLRDRD, INSTR_SS_RRRDRD, INSTR_SS_RRRDRD2, INSTR_SS_RRRDRD3,
+	INSTR_S_00, INSTR_S_RD,
+};
+
+struct operand {
+	int bits;		/* The number of bits in the operand. */
+	int shift;		/* The number of bits to shift. */
+	int flags;		/* One bit syntax flags. */
+};
+
+struct insn {
+	const char name[5];
+	unsigned char opfrag;
+	unsigned char format;
+};
+
+static const struct operand operands[] =
+{
+	[UNUSED]  = { 0, 0, 0 },
+	[R_8]	 = {  4,  8, OPERAND_GPR },
+	[R_12]	 = {  4, 12, OPERAND_GPR },
+	[R_16]	 = {  4, 16, OPERAND_GPR },
+	[R_20]	 = {  4, 20, OPERAND_GPR },
+	[R_24]	 = {  4, 24, OPERAND_GPR },
+	[R_28]	 = {  4, 28, OPERAND_GPR },
+	[R_32]	 = {  4, 32, OPERAND_GPR },
+	[F_8]	 = {  4,  8, OPERAND_FPR },
+	[F_12]	 = {  4, 12, OPERAND_FPR },
+	[F_16]	 = {  4, 16, OPERAND_FPR },
+	[F_20]	 = {  4, 16, OPERAND_FPR },
+	[F_24]	 = {  4, 24, OPERAND_FPR },
+	[F_28]	 = {  4, 28, OPERAND_FPR },
+	[F_32]	 = {  4, 32, OPERAND_FPR },
+	[A_8]	 = {  4,  8, OPERAND_AR },
+	[A_12]	 = {  4, 12, OPERAND_AR },
+	[A_24]	 = {  4, 24, OPERAND_AR },
+	[A_28]	 = {  4, 28, OPERAND_AR },
+	[C_8]	 = {  4,  8, OPERAND_CR },
+	[C_12]	 = {  4, 12, OPERAND_CR },
+	[B_16]	 = {  4, 16, OPERAND_BASE | OPERAND_GPR },
+	[B_32]	 = {  4, 32, OPERAND_BASE | OPERAND_GPR },
+	[X_12]	 = {  4, 12, OPERAND_INDEX | OPERAND_GPR },
+	[D_20]	 = { 12, 20, OPERAND_DISP },
+	[D_36]	 = { 12, 36, OPERAND_DISP },
+	[D20_20] = { 20, 20, OPERAND_DISP | OPERAND_SIGNED },
+	[L4_8]	 = {  4,  8, OPERAND_LENGTH },
+	[L4_12]  = {  4, 12, OPERAND_LENGTH },
+	[L8_8]	 = {  8,  8, OPERAND_LENGTH },
+	[U4_8]	 = {  4,  8, 0 },
+	[U4_12]  = {  4, 12, 0 },
+	[U4_16]  = {  4, 16, 0 },
+	[U4_20]  = {  4, 20, 0 },
+	[U8_8]	 = {  8,  8, 0 },
+	[U8_16]  = {  8, 16, 0 },
+	[I16_16] = { 16, 16, OPERAND_SIGNED },
+	[U16_16] = { 16, 16, 0 },
+	[J16_16] = { 16, 16, OPERAND_PCREL },
+	[J32_16] = { 32, 16, OPERAND_PCREL },
+	[I32_16] = { 32, 16, OPERAND_SIGNED },
+	[U32_16] = { 32, 16, 0 },
+	[M_16]	 = {  4, 16, 0 },
+	[RO_28]  = {  4, 28, OPERAND_GPR }
+};
+
+static const unsigned char formats[][7] = {
+	[INSTR_E]	  = { 0xff, 0,0,0,0,0,0 },	       /* e.g. pr    */
+	[INSTR_RIE_RRP]	  = { 0xff, R_8,R_12,J16_16,0,0,0 },   /* e.g. brxhg */
+	[INSTR_RIL_RP]	  = { 0x0f, R_8,J32_16,0,0,0,0 },      /* e.g. brasl */
+	[INSTR_RIL_UP]	  = { 0x0f, U4_8,J32_16,0,0,0,0 },     /* e.g. brcl  */
+	[INSTR_RIL_RI]	  = { 0x0f, R_8,I32_16,0,0,0,0 },      /* e.g. afi   */
+	[INSTR_RIL_RU]	  = { 0x0f, R_8,U32_16,0,0,0,0 },      /* e.g. alfi  */
+	[INSTR_RI_RI]	  = { 0x0f, R_8,I16_16,0,0,0,0 },      /* e.g. ahi   */
+	[INSTR_RI_RP]	  = { 0x0f, R_8,J16_16,0,0,0,0 },      /* e.g. brct  */
+	[INSTR_RI_RU]	  = { 0x0f, R_8,U16_16,0,0,0,0 },      /* e.g. tml   */
+	[INSTR_RI_UP]	  = { 0x0f, U4_8,J16_16,0,0,0,0 },     /* e.g. brc   */
+	[INSTR_RRE_00]	  = { 0xff, 0,0,0,0,0,0 },	       /* e.g. palb  */
+	[INSTR_RRE_0R]	  = { 0xff, R_28,0,0,0,0,0 },	       /* e.g. tb    */
+	[INSTR_RRE_AA]	  = { 0xff, A_24,A_28,0,0,0,0 },       /* e.g. cpya  */
+	[INSTR_RRE_AR]	  = { 0xff, A_24,R_28,0,0,0,0 },       /* e.g. sar   */
+	[INSTR_RRE_F0]	  = { 0xff, F_24,0,0,0,0,0 },	       /* e.g. sqer  */
+	[INSTR_RRE_FF]	  = { 0xff, F_24,F_28,0,0,0,0 },       /* e.g. debr  */
+	[INSTR_RRE_R0]	  = { 0xff, R_24,0,0,0,0,0 },	       /* e.g. ipm   */
+	[INSTR_RRE_RA]	  = { 0xff, R_24,A_28,0,0,0,0 },       /* e.g. ear   */
+	[INSTR_RRE_RF]	  = { 0xff, R_24,F_28,0,0,0,0 },       /* e.g. cefbr */
+	[INSTR_RRE_RR]	  = { 0xff, R_24,R_28,0,0,0,0 },       /* e.g. lura  */
+	[INSTR_RRE_RR_OPT]= { 0xff, R_24,RO_28,0,0,0,0 },      /* efpc, sfpc */
+	[INSTR_RRF_F0FF]  = { 0xff, F_16,F_24,F_28,0,0,0 },    /* e.g. madbr */
+	[INSTR_RRF_FUFF]  = { 0xff, F_24,F_16,F_28,U4_20,0,0 },/* e.g. didbr */
+	[INSTR_RRF_RURR]  = { 0xff, R_24,R_28,R_16,U4_20,0,0 },/* e.g. .insn */
+	[INSTR_RRF_R0RR]  = { 0xff, R_24,R_28,R_16,0,0,0 },    /* e.g. idte  */
+	[INSTR_RRF_U0FF]  = { 0xff, F_24,U4_16,F_28,0,0,0 },   /* e.g. fixr  */
+	[INSTR_RRF_U0RF]  = { 0xff, R_24,U4_16,F_28,0,0,0 },   /* e.g. cfebr */
+	[INSTR_RRF_M0RR]  = { 0xff, R_24,R_28,M_16,0,0,0 },    /* e.g. sske  */
+	[INSTR_RR_FF]	  = { 0xff, F_8,F_12,0,0,0,0 },        /* e.g. adr   */
+	[INSTR_RR_R0]	  = { 0xff, R_8, 0,0,0,0,0 },	       /* e.g. spm   */
+	[INSTR_RR_RR]	  = { 0xff, R_8,R_12,0,0,0,0 },        /* e.g. lr    */
+	[INSTR_RR_U0]	  = { 0xff, U8_8, 0,0,0,0,0 },	       /* e.g. svc   */
+	[INSTR_RR_UR]	  = { 0xff, U4_8,R_12,0,0,0,0 },       /* e.g. bcr   */
+	[INSTR_RSE_RRRD]  = { 0xff, R_8,R_12,D_20,B_16,0,0 },  /* e.g. lmh   */
+	[INSTR_RSE_CCRD]  = { 0xff, C_8,C_12,D_20,B_16,0,0 },  /* e.g. lmh   */
+	[INSTR_RSE_RURD]  = { 0xff, R_8,U4_12,D_20,B_16,0,0 }, /* e.g. icmh  */
+	[INSTR_RSL_R0RD]  = { 0xff, R_8,D_20,B_16,0,0,0 },     /* e.g. tp    */
+	[INSTR_RSI_RRP]	  = { 0xff, R_8,R_12,J16_16,0,0,0 },   /* e.g. brxh  */
+	[INSTR_RSY_RRRD]  = { 0xff, R_8,R_12,D20_20,B_16,0,0 },/* e.g. stmy  */
+	[INSTR_RSY_RURD]  = { 0xff, R_8,U4_12,D20_20,B_16,0,0 },
+							       /* e.g. icmh  */
+	[INSTR_RSY_AARD]  = { 0xff, A_8,A_12,D20_20,B_16,0,0 },/* e.g. lamy  */
+	[INSTR_RSY_CCRD]  = { 0xff, C_8,C_12,D20_20,B_16,0,0 },/* e.g. lamy  */
+	[INSTR_RS_AARD]	  = { 0xff, A_8,A_12,D_20,B_16,0,0 },  /* e.g. lam   */
+	[INSTR_RS_CCRD]	  = { 0xff, C_8,C_12,D_20,B_16,0,0 },  /* e.g. lctl  */
+	[INSTR_RS_R0RD]	  = { 0xff, R_8,D_20,B_16,0,0,0 },     /* e.g. sll   */
+	[INSTR_RS_RRRD]	  = { 0xff, R_8,R_12,D_20,B_16,0,0 },  /* e.g. cs    */
+	[INSTR_RS_RURD]	  = { 0xff, R_8,U4_12,D_20,B_16,0,0 }, /* e.g. icm   */
+	[INSTR_RXE_FRRD]  = { 0xff, F_8,D_20,X_12,B_16,0,0 },  /* e.g. axbr  */
+	[INSTR_RXE_RRRD]  = { 0xff, R_8,D_20,X_12,B_16,0,0 },  /* e.g. lg    */
+	[INSTR_RXF_FRRDF] = { 0xff, F_32,F_8,D_20,X_12,B_16,0 },
+							       /* e.g. madb  */
+	[INSTR_RXY_RRRD]  = { 0xff, R_8,D20_20,X_12,B_16,0,0 },/* e.g. ly    */
+	[INSTR_RXY_FRRD]  = { 0xff, F_8,D20_20,X_12,B_16,0,0 },/* e.g. ley   */
+	[INSTR_RX_FRRD]	  = { 0xff, F_8,D_20,X_12,B_16,0,0 },  /* e.g. ae    */
+	[INSTR_RX_RRRD]	  = { 0xff, R_8,D_20,X_12,B_16,0,0 },  /* e.g. l     */
+	[INSTR_RX_URRD]	  = { 0x00, U4_8,D_20,X_12,B_16,0,0 }, /* e.g. bc    */
+	[INSTR_SI_URD]	  = { 0x00, D_20,B_16,U8_8,0,0,0 },    /* e.g. cli   */
+	[INSTR_SIY_URD]	  = { 0xff, D20_20,B_16,U8_8,0,0,0 },  /* e.g. tmy   */
+	[INSTR_SSE_RDRD]  = { 0xff, D_20,B_16,D_36,B_32,0,0 }, /* e.g. mvsdk */
+	[INSTR_SS_L0RDRD] = { 0xff, D_20,L8_8,B_16,D_36,B_32,0 },
+							       /* e.g. mvc   */
+	[INSTR_SS_LIRDRD] = { 0xff, D_20,L4_8,B_16,D_36,B_32,U4_12 },
+							       /* e.g. srp   */
+	[INSTR_SS_LLRDRD] = { 0xff, D_20,L4_8,B_16,D_36,L4_12,B_32 },
+							       /* e.g. pack  */
+	[INSTR_SS_RRRDRD] = { 0xff, D_20,R_8,B_16,D_36,B_32,R_12 },
+							       /* e.g. mvck  */
+	[INSTR_SS_RRRDRD2]= { 0xff, R_8,D_20,B_16,R_12,D_36,B_32 },
+							       /* e.g. plo   */
+	[INSTR_SS_RRRDRD3]= { 0xff, R_8,R_12,D_20,B_16,D_36,B_32 },
+							       /* e.g. lmd   */
+	[INSTR_S_00]	  = { 0xff, 0,0,0,0,0,0 },	       /* e.g. hsch  */
+	[INSTR_S_RD]	  = { 0xff, D_20,B_16,0,0,0,0 },       /* e.g. lpsw  */
+	[INSTR_SSF_RRDRD] = { 0x00, D_20,B_16,D_36,B_32,R_8,0 },
+							       /* e.g. mvcos */
+};
+
+static struct insn opcode[] = {
+#ifdef CONFIG_64BIT
+	{ "lmd", 0xef, INSTR_SS_RRRDRD3 },
+#endif
+	{ "spm", 0x04, INSTR_RR_R0 },
+	{ "balr", 0x05, INSTR_RR_RR },
+	{ "bctr", 0x06, INSTR_RR_RR },
+	{ "bcr", 0x07, INSTR_RR_UR },
+	{ "svc", 0x0a, INSTR_RR_U0 },
+	{ "bsm", 0x0b, INSTR_RR_RR },
+	{ "bassm", 0x0c, INSTR_RR_RR },
+	{ "basr", 0x0d, INSTR_RR_RR },
+	{ "mvcl", 0x0e, INSTR_RR_RR },
+	{ "clcl", 0x0f, INSTR_RR_RR },
+	{ "lpr", 0x10, INSTR_RR_RR },
+	{ "lnr", 0x11, INSTR_RR_RR },
+	{ "ltr", 0x12, INSTR_RR_RR },
+	{ "lcr", 0x13, INSTR_RR_RR },
+	{ "nr", 0x14, INSTR_RR_RR },
+	{ "clr", 0x15, INSTR_RR_RR },
+	{ "or", 0x16, INSTR_RR_RR },
+	{ "xr", 0x17, INSTR_RR_RR },
+	{ "lr", 0x18, INSTR_RR_RR },
+	{ "cr", 0x19, INSTR_RR_RR },
+	{ "ar", 0x1a, INSTR_RR_RR },
+	{ "sr", 0x1b, INSTR_RR_RR },
+	{ "mr", 0x1c, INSTR_RR_RR },
+	{ "dr", 0x1d, INSTR_RR_RR },
+	{ "alr", 0x1e, INSTR_RR_RR },
+	{ "slr", 0x1f, INSTR_RR_RR },
+	{ "lpdr", 0x20, INSTR_RR_FF },
+	{ "lndr", 0x21, INSTR_RR_FF },
+	{ "ltdr", 0x22, INSTR_RR_FF },
+	{ "lcdr", 0x23, INSTR_RR_FF },
+	{ "hdr", 0x24, INSTR_RR_FF },
+	{ "ldxr", 0x25, INSTR_RR_FF },
+	{ "lrdr", 0x25, INSTR_RR_FF },
+	{ "mxr", 0x26, INSTR_RR_FF },
+	{ "mxdr", 0x27, INSTR_RR_FF },
+	{ "ldr", 0x28, INSTR_RR_FF },
+	{ "cdr", 0x29, INSTR_RR_FF },
+	{ "adr", 0x2a, INSTR_RR_FF },
+	{ "sdr", 0x2b, INSTR_RR_FF },
+	{ "mdr", 0x2c, INSTR_RR_FF },
+	{ "ddr", 0x2d, INSTR_RR_FF },
+	{ "awr", 0x2e, INSTR_RR_FF },
+	{ "swr", 0x2f, INSTR_RR_FF },
+	{ "lper", 0x30, INSTR_RR_FF },
+	{ "lner", 0x31, INSTR_RR_FF },
+	{ "lter", 0x32, INSTR_RR_FF },
+	{ "lcer", 0x33, INSTR_RR_FF },
+	{ "her", 0x34, INSTR_RR_FF },
+	{ "ledr", 0x35, INSTR_RR_FF },
+	{ "lrer", 0x35, INSTR_RR_FF },
+	{ "axr", 0x36, INSTR_RR_FF },
+	{ "sxr", 0x37, INSTR_RR_FF },
+	{ "ler", 0x38, INSTR_RR_FF },
+	{ "cer", 0x39, INSTR_RR_FF },
+	{ "aer", 0x3a, INSTR_RR_FF },
+	{ "ser", 0x3b, INSTR_RR_FF },
+	{ "mder", 0x3c, INSTR_RR_FF },
+	{ "mer", 0x3c, INSTR_RR_FF },
+	{ "der", 0x3d, INSTR_RR_FF },
+	{ "aur", 0x3e, INSTR_RR_FF },
+	{ "sur", 0x3f, INSTR_RR_FF },
+	{ "sth", 0x40, INSTR_RX_RRRD },
+	{ "la", 0x41, INSTR_RX_RRRD },
+	{ "stc", 0x42, INSTR_RX_RRRD },
+	{ "ic", 0x43, INSTR_RX_RRRD },
+	{ "ex", 0x44, INSTR_RX_RRRD },
+	{ "bal", 0x45, INSTR_RX_RRRD },
+	{ "bct", 0x46, INSTR_RX_RRRD },
+	{ "bc", 0x47, INSTR_RX_URRD },
+	{ "lh", 0x48, INSTR_RX_RRRD },
+	{ "ch", 0x49, INSTR_RX_RRRD },
+	{ "ah", 0x4a, INSTR_RX_RRRD },
+	{ "sh", 0x4b, INSTR_RX_RRRD },
+	{ "mh", 0x4c, INSTR_RX_RRRD },
+	{ "bas", 0x4d, INSTR_RX_RRRD },
+	{ "cvd", 0x4e, INSTR_RX_RRRD },
+	{ "cvb", 0x4f, INSTR_RX_RRRD },
+	{ "st", 0x50, INSTR_RX_RRRD },
+	{ "lae", 0x51, INSTR_RX_RRRD },
+	{ "n", 0x54, INSTR_RX_RRRD },
+	{ "cl", 0x55, INSTR_RX_RRRD },
+	{ "o", 0x56, INSTR_RX_RRRD },
+	{ "x", 0x57, INSTR_RX_RRRD },
+	{ "l", 0x58, INSTR_RX_RRRD },
+	{ "c", 0x59, INSTR_RX_RRRD },
+	{ "a", 0x5a, INSTR_RX_RRRD },
+	{ "s", 0x5b, INSTR_RX_RRRD },
+	{ "m", 0x5c, INSTR_RX_RRRD },
+	{ "d", 0x5d, INSTR_RX_RRRD },
+	{ "al", 0x5e, INSTR_RX_RRRD },
+	{ "sl", 0x5f, INSTR_RX_RRRD },
+	{ "std", 0x60, INSTR_RX_FRRD },
+	{ "mxd", 0x67, INSTR_RX_FRRD },
+	{ "ld", 0x68, INSTR_RX_FRRD },
+	{ "cd", 0x69, INSTR_RX_FRRD },
+	{ "ad", 0x6a, INSTR_RX_FRRD },
+	{ "sd", 0x6b, INSTR_RX_FRRD },
+	{ "md", 0x6c, INSTR_RX_FRRD },
+	{ "dd", 0x6d, INSTR_RX_FRRD },
+	{ "aw", 0x6e, INSTR_RX_FRRD },
+	{ "sw", 0x6f, INSTR_RX_FRRD },
+	{ "ste", 0x70, INSTR_RX_FRRD },
+	{ "ms", 0x71, INSTR_RX_RRRD },
+	{ "le", 0x78, INSTR_RX_FRRD },
+	{ "ce", 0x79, INSTR_RX_FRRD },
+	{ "ae", 0x7a, INSTR_RX_FRRD },
+	{ "se", 0x7b, INSTR_RX_FRRD },
+	{ "mde", 0x7c, INSTR_RX_FRRD },
+	{ "me", 0x7c, INSTR_RX_FRRD },
+	{ "de", 0x7d, INSTR_RX_FRRD },
+	{ "au", 0x7e, INSTR_RX_FRRD },
+	{ "su", 0x7f, INSTR_RX_FRRD },
+	{ "ssm", 0x80, INSTR_S_RD },
+	{ "lpsw", 0x82, INSTR_S_RD },
+	{ "diag", 0x83, INSTR_RS_RRRD },
+	{ "brxh", 0x84, INSTR_RSI_RRP },
+	{ "brxle", 0x85, INSTR_RSI_RRP },
+	{ "bxh", 0x86, INSTR_RS_RRRD },
+	{ "bxle", 0x87, INSTR_RS_RRRD },
+	{ "srl", 0x88, INSTR_RS_R0RD },
+	{ "sll", 0x89, INSTR_RS_R0RD },
+	{ "sra", 0x8a, INSTR_RS_R0RD },
+	{ "sla", 0x8b, INSTR_RS_R0RD },
+	{ "srdl", 0x8c, INSTR_RS_R0RD },
+	{ "sldl", 0x8d, INSTR_RS_R0RD },
+	{ "srda", 0x8e, INSTR_RS_R0RD },
+	{ "slda", 0x8f, INSTR_RS_R0RD },
+	{ "stm", 0x90, INSTR_RS_RRRD },
+	{ "tm", 0x91, INSTR_SI_URD },
+	{ "mvi", 0x92, INSTR_SI_URD },
+	{ "ts", 0x93, INSTR_S_RD },
+	{ "ni", 0x94, INSTR_SI_URD },
+	{ "cli", 0x95, INSTR_SI_URD },
+	{ "oi", 0x96, INSTR_SI_URD },
+	{ "xi", 0x97, INSTR_SI_URD },
+	{ "lm", 0x98, INSTR_RS_RRRD },
+	{ "trace", 0x99, INSTR_RS_RRRD },
+	{ "lam", 0x9a, INSTR_RS_AARD },
+	{ "stam", 0x9b, INSTR_RS_AARD },
+	{ "mvcle", 0xa8, INSTR_RS_RRRD },
+	{ "clcle", 0xa9, INSTR_RS_RRRD },
+	{ "stnsm", 0xac, INSTR_SI_URD },
+	{ "stosm", 0xad, INSTR_SI_URD },
+	{ "sigp", 0xae, INSTR_RS_RRRD },
+	{ "mc", 0xaf, INSTR_SI_URD },
+	{ "lra", 0xb1, INSTR_RX_RRRD },
+	{ "stctl", 0xb6, INSTR_RS_CCRD },
+	{ "lctl", 0xb7, INSTR_RS_CCRD },
+	{ "cs", 0xba, INSTR_RS_RRRD },
+	{ "cds", 0xbb, INSTR_RS_RRRD },
+	{ "clm", 0xbd, INSTR_RS_RURD },
+	{ "stcm", 0xbe, INSTR_RS_RURD },
+	{ "icm", 0xbf, INSTR_RS_RURD },
+	{ "mvn", 0xd1, INSTR_SS_L0RDRD },
+	{ "mvc", 0xd2, INSTR_SS_L0RDRD },
+	{ "mvz", 0xd3, INSTR_SS_L0RDRD },
+	{ "nc", 0xd4, INSTR_SS_L0RDRD },
+	{ "clc", 0xd5, INSTR_SS_L0RDRD },
+	{ "oc", 0xd6, INSTR_SS_L0RDRD },
+	{ "xc", 0xd7, INSTR_SS_L0RDRD },
+	{ "mvck", 0xd9, INSTR_SS_RRRDRD },
+	{ "mvcp", 0xda, INSTR_SS_RRRDRD },
+	{ "mvcs", 0xdb, INSTR_SS_RRRDRD },
+	{ "tr", 0xdc, INSTR_SS_L0RDRD },
+	{ "trt", 0xdd, INSTR_SS_L0RDRD },
+	{ "ed", 0xde, INSTR_SS_L0RDRD },
+	{ "edmk", 0xdf, INSTR_SS_L0RDRD },
+	{ "pku", 0xe1, INSTR_SS_L0RDRD },
+	{ "unpku", 0xe2, INSTR_SS_L0RDRD },
+	{ "mvcin", 0xe8, INSTR_SS_L0RDRD },
+	{ "pka", 0xe9, INSTR_SS_L0RDRD },
+	{ "unpka", 0xea, INSTR_SS_L0RDRD },
+	{ "plo", 0xee, INSTR_SS_RRRDRD2 },
+	{ "srp", 0xf0, INSTR_SS_LIRDRD },
+	{ "mvo", 0xf1, INSTR_SS_LLRDRD },
+	{ "pack", 0xf2, INSTR_SS_LLRDRD },
+	{ "unpk", 0xf3, INSTR_SS_LLRDRD },
+	{ "zap", 0xf8, INSTR_SS_LLRDRD },
+	{ "cp", 0xf9, INSTR_SS_LLRDRD },
+	{ "ap", 0xfa, INSTR_SS_LLRDRD },
+	{ "sp", 0xfb, INSTR_SS_LLRDRD },
+	{ "mp", 0xfc, INSTR_SS_LLRDRD },
+	{ "dp", 0xfd, INSTR_SS_LLRDRD },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_01[] = {
+#ifdef CONFIG_64BIT
+	{ "sam64", 0x0e, INSTR_E },
+#endif
+	{ "pr", 0x01, INSTR_E },
+	{ "upt", 0x02, INSTR_E },
+	{ "sckpf", 0x07, INSTR_E },
+	{ "tam", 0x0b, INSTR_E },
+	{ "sam24", 0x0c, INSTR_E },
+	{ "sam31", 0x0d, INSTR_E },
+	{ "trap2", 0xff, INSTR_E },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_a5[] = {
+#ifdef CONFIG_64BIT
+	{ "iihh", 0x00, INSTR_RI_RU },
+	{ "iihl", 0x01, INSTR_RI_RU },
+	{ "iilh", 0x02, INSTR_RI_RU },
+	{ "iill", 0x03, INSTR_RI_RU },
+	{ "nihh", 0x04, INSTR_RI_RU },
+	{ "nihl", 0x05, INSTR_RI_RU },
+	{ "nilh", 0x06, INSTR_RI_RU },
+	{ "nill", 0x07, INSTR_RI_RU },
+	{ "oihh", 0x08, INSTR_RI_RU },
+	{ "oihl", 0x09, INSTR_RI_RU },
+	{ "oilh", 0x0a, INSTR_RI_RU },
+	{ "oill", 0x0b, INSTR_RI_RU },
+	{ "llihh", 0x0c, INSTR_RI_RU },
+	{ "llihl", 0x0d, INSTR_RI_RU },
+	{ "llilh", 0x0e, INSTR_RI_RU },
+	{ "llill", 0x0f, INSTR_RI_RU },
+#endif
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_a7[] = {
+#ifdef CONFIG_64BIT
+	{ "tmhh", 0x02, INSTR_RI_RU },
+	{ "tmhl", 0x03, INSTR_RI_RU },
+	{ "brctg", 0x07, INSTR_RI_RP },
+	{ "lghi", 0x09, INSTR_RI_RI },
+	{ "aghi", 0x0b, INSTR_RI_RI },
+	{ "mghi", 0x0d, INSTR_RI_RI },
+	{ "cghi", 0x0f, INSTR_RI_RI },
+#endif
+	{ "tmlh", 0x00, INSTR_RI_RU },
+	{ "tmll", 0x01, INSTR_RI_RU },
+	{ "brc", 0x04, INSTR_RI_UP },
+	{ "bras", 0x05, INSTR_RI_RP },
+	{ "brct", 0x06, INSTR_RI_RP },
+	{ "lhi", 0x08, INSTR_RI_RI },
+	{ "ahi", 0x0a, INSTR_RI_RI },
+	{ "mhi", 0x0c, INSTR_RI_RI },
+	{ "chi", 0x0e, INSTR_RI_RI },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_b2[] = {
+#ifdef CONFIG_64BIT
+	{ "sske", 0x2b, INSTR_RRF_M0RR },
+	{ "stckf", 0x7c, INSTR_S_RD },
+	{ "cu21", 0xa6, INSTR_RRF_M0RR },
+	{ "cuutf", 0xa6, INSTR_RRF_M0RR },
+	{ "cu12", 0xa7, INSTR_RRF_M0RR },
+	{ "cutfu", 0xa7, INSTR_RRF_M0RR },
+	{ "stfle", 0xb0, INSTR_S_RD },
+	{ "lpswe", 0xb2, INSTR_S_RD },
+#endif
+	{ "stidp", 0x02, INSTR_S_RD },
+	{ "sck", 0x04, INSTR_S_RD },
+	{ "stck", 0x05, INSTR_S_RD },
+	{ "sckc", 0x06, INSTR_S_RD },
+	{ "stckc", 0x07, INSTR_S_RD },
+	{ "spt", 0x08, INSTR_S_RD },
+	{ "stpt", 0x09, INSTR_S_RD },
+	{ "spka", 0x0a, INSTR_S_RD },
+	{ "ipk", 0x0b, INSTR_S_00 },
+	{ "ptlb", 0x0d, INSTR_S_00 },
+	{ "spx", 0x10, INSTR_S_RD },
+	{ "stpx", 0x11, INSTR_S_RD },
+	{ "stap", 0x12, INSTR_S_RD },
+	{ "sie", 0x14, INSTR_S_RD },
+	{ "pc", 0x18, INSTR_S_RD },
+	{ "sac", 0x19, INSTR_S_RD },
+	{ "cfc", 0x1a, INSTR_S_RD },
+	{ "ipte", 0x21, INSTR_RRE_RR },
+	{ "ipm", 0x22, INSTR_RRE_R0 },
+	{ "ivsk", 0x23, INSTR_RRE_RR },
+	{ "iac", 0x24, INSTR_RRE_R0 },
+	{ "ssar", 0x25, INSTR_RRE_R0 },
+	{ "epar", 0x26, INSTR_RRE_R0 },
+	{ "esar", 0x27, INSTR_RRE_R0 },
+	{ "pt", 0x28, INSTR_RRE_RR },
+	{ "iske", 0x29, INSTR_RRE_RR },
+	{ "rrbe", 0x2a, INSTR_RRE_RR },
+	{ "sske", 0x2b, INSTR_RRE_RR },
+	{ "tb", 0x2c, INSTR_RRE_0R },
+	{ "dxr", 0x2d, INSTR_RRE_F0 },
+	{ "pgin", 0x2e, INSTR_RRE_RR },
+	{ "pgout", 0x2f, INSTR_RRE_RR },
+	{ "csch", 0x30, INSTR_S_00 },
+	{ "hsch", 0x31, INSTR_S_00 },
+	{ "msch", 0x32, INSTR_S_RD },
+	{ "ssch", 0x33, INSTR_S_RD },
+	{ "stsch", 0x34, INSTR_S_RD },
+	{ "tsch", 0x35, INSTR_S_RD },
+	{ "tpi", 0x36, INSTR_S_RD },
+	{ "sal", 0x37, INSTR_S_00 },
+	{ "rsch", 0x38, INSTR_S_00 },
+	{ "stcrw", 0x39, INSTR_S_RD },
+	{ "stcps", 0x3a, INSTR_S_RD },
+	{ "rchp", 0x3b, INSTR_S_00 },
+	{ "schm", 0x3c, INSTR_S_00 },
+	{ "bakr", 0x40, INSTR_RRE_RR },
+	{ "cksm", 0x41, INSTR_RRE_RR },
+	{ "sqdr", 0x44, INSTR_RRE_F0 },
+	{ "sqer", 0x45, INSTR_RRE_F0 },
+	{ "stura", 0x46, INSTR_RRE_RR },
+	{ "msta", 0x47, INSTR_RRE_R0 },
+	{ "palb", 0x48, INSTR_RRE_00 },
+	{ "ereg", 0x49, INSTR_RRE_RR },
+	{ "esta", 0x4a, INSTR_RRE_RR },
+	{ "lura", 0x4b, INSTR_RRE_RR },
+	{ "tar", 0x4c, INSTR_RRE_AR },
+	{ "cpya", INSTR_RRE_AA },
+	{ "sar", 0x4e, INSTR_RRE_AR },
+	{ "ear", 0x4f, INSTR_RRE_RA },
+	{ "csp", 0x50, INSTR_RRE_RR },
+	{ "msr", 0x52, INSTR_RRE_RR },
+	{ "mvpg", 0x54, INSTR_RRE_RR },
+	{ "mvst", 0x55, INSTR_RRE_RR },
+	{ "cuse", 0x57, INSTR_RRE_RR },
+	{ "bsg", 0x58, INSTR_RRE_RR },
+	{ "bsa", 0x5a, INSTR_RRE_RR },
+	{ "clst", 0x5d, INSTR_RRE_RR },
+	{ "srst", 0x5e, INSTR_RRE_RR },
+	{ "cmpsc", 0x63, INSTR_RRE_RR },
+	{ "cmpsc", 0x63, INSTR_RRE_RR },
+	{ "siga", 0x74, INSTR_S_RD },
+	{ "xsch", 0x76, INSTR_S_00 },
+	{ "rp", 0x77, INSTR_S_RD },
+	{ "stcke", 0x78, INSTR_S_RD },
+	{ "sacf", 0x79, INSTR_S_RD },
+	{ "stsi", 0x7d, INSTR_S_RD },
+	{ "srnm", 0x99, INSTR_S_RD },
+	{ "stfpc", 0x9c, INSTR_S_RD },
+	{ "lfpc", 0x9d, INSTR_S_RD },
+	{ "tre", 0xa5, INSTR_RRE_RR },
+	{ "cuutf", 0xa6, INSTR_RRE_RR },
+	{ "cutfu", 0xa7, INSTR_RRE_RR },
+	{ "stfl", 0xb1, INSTR_S_RD },
+	{ "trap4", 0xff, INSTR_S_RD },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_b3[] = {
+#ifdef CONFIG_64BIT
+	{ "maylr", 0x38, INSTR_RRF_F0FF },
+	{ "mylr", 0x39, INSTR_RRF_F0FF },
+	{ "mayr", 0x3a, INSTR_RRF_F0FF },
+	{ "myr", 0x3b, INSTR_RRF_F0FF },
+	{ "mayhr", 0x3c, INSTR_RRF_F0FF },
+	{ "myhr", 0x3d, INSTR_RRF_F0FF },
+	{ "cegbr", 0xa4, INSTR_RRE_RR },
+	{ "cdgbr", 0xa5, INSTR_RRE_RR },
+	{ "cxgbr", 0xa6, INSTR_RRE_RR },
+	{ "cgebr", 0xa8, INSTR_RRF_U0RF },
+	{ "cgdbr", 0xa9, INSTR_RRF_U0RF },
+	{ "cgxbr", 0xaa, INSTR_RRF_U0RF },
+	{ "cfer", 0xb8, INSTR_RRF_U0RF },
+	{ "cfdr", 0xb9, INSTR_RRF_U0RF },
+	{ "cfxr", 0xba, INSTR_RRF_U0RF },
+	{ "cegr", 0xc4, INSTR_RRE_RR },
+	{ "cdgr", 0xc5, INSTR_RRE_RR },
+	{ "cxgr", 0xc6, INSTR_RRE_RR },
+	{ "cger", 0xc8, INSTR_RRF_U0RF },
+	{ "cgdr", 0xc9, INSTR_RRF_U0RF },
+	{ "cgxr", 0xca, INSTR_RRF_U0RF },
+#endif
+	{ "lpebr", 0x00, INSTR_RRE_FF },
+	{ "lnebr", 0x01, INSTR_RRE_FF },
+	{ "ltebr", 0x02, INSTR_RRE_FF },
+	{ "lcebr", 0x03, INSTR_RRE_FF },
+	{ "ldebr", 0x04, INSTR_RRE_FF },
+	{ "lxdbr", 0x05, INSTR_RRE_FF },
+	{ "lxebr", 0x06, INSTR_RRE_FF },
+	{ "mxdbr", 0x07, INSTR_RRE_FF },
+	{ "kebr", 0x08, INSTR_RRE_FF },
+	{ "cebr", 0x09, INSTR_RRE_FF },
+	{ "aebr", 0x0a, INSTR_RRE_FF },
+	{ "sebr", 0x0b, INSTR_RRE_FF },
+	{ "mdebr", 0x0c, INSTR_RRE_FF },
+	{ "debr", 0x0d, INSTR_RRE_FF },
+	{ "maebr", 0x0e, INSTR_RRF_F0FF },
+	{ "msebr", 0x0f, INSTR_RRF_F0FF },
+	{ "lpdbr", 0x10, INSTR_RRE_FF },
+	{ "lndbr", 0x11, INSTR_RRE_FF },
+	{ "ltdbr", 0x12, INSTR_RRE_FF },
+	{ "lcdbr", 0x13, INSTR_RRE_FF },
+	{ "sqebr", 0x14, INSTR_RRE_FF },
+	{ "sqdbr", 0x15, INSTR_RRE_FF },
+	{ "sqxbr", 0x16, INSTR_RRE_FF },
+	{ "meebr", 0x17, INSTR_RRE_FF },
+	{ "kdbr", 0x18, INSTR_RRE_FF },
+	{ "cdbr", 0x19, INSTR_RRE_FF },
+	{ "adbr", 0x1a, INSTR_RRE_FF },
+	{ "sdbr", 0x1b, INSTR_RRE_FF },
+	{ "mdbr", 0x1c, INSTR_RRE_FF },
+	{ "ddbr", 0x1d, INSTR_RRE_FF },
+	{ "madbr", 0x1e, INSTR_RRF_F0FF },
+	{ "msdbr", 0x1f, INSTR_RRF_F0FF },
+	{ "lder", 0x24, INSTR_RRE_FF },
+	{ "lxdr", 0x25, INSTR_RRE_FF },
+	{ "lxer", 0x26, INSTR_RRE_FF },
+	{ "maer", 0x2e, INSTR_RRF_F0FF },
+	{ "mser", 0x2f, INSTR_RRF_F0FF },
+	{ "sqxr", 0x36, INSTR_RRE_FF },
+	{ "meer", 0x37, INSTR_RRE_FF },
+	{ "madr", 0x3e, INSTR_RRF_F0FF },
+	{ "msdr", 0x3f, INSTR_RRF_F0FF },
+	{ "lpxbr", 0x40, INSTR_RRE_FF },
+	{ "lnxbr", 0x41, INSTR_RRE_FF },
+	{ "ltxbr", 0x42, INSTR_RRE_FF },
+	{ "lcxbr", 0x43, INSTR_RRE_FF },
+	{ "ledbr", 0x44, INSTR_RRE_FF },
+	{ "ldxbr", 0x45, INSTR_RRE_FF },
+	{ "lexbr", 0x46, INSTR_RRE_FF },
+	{ "fixbr", 0x47, INSTR_RRF_U0FF },
+	{ "kxbr", 0x48, INSTR_RRE_FF },
+	{ "cxbr", 0x49, INSTR_RRE_FF },
+	{ "axbr", 0x4a, INSTR_RRE_FF },
+	{ "sxbr", 0x4b, INSTR_RRE_FF },
+	{ "mxbr", 0x4c, INSTR_RRE_FF },
+	{ "dxbr", 0x4d, INSTR_RRE_FF },
+	{ "tbedr", 0x50, INSTR_RRF_U0FF },
+	{ "tbdr", 0x51, INSTR_RRF_U0FF },
+	{ "diebr", 0x53, INSTR_RRF_FUFF },
+	{ "fiebr", 0x57, INSTR_RRF_U0FF },
+	{ "thder", 0x58, INSTR_RRE_RR },
+	{ "thdr", 0x59, INSTR_RRE_RR },
+	{ "didbr", 0x5b, INSTR_RRF_FUFF },
+	{ "fidbr", 0x5f, INSTR_RRF_U0FF },
+	{ "lpxr", 0x60, INSTR_RRE_FF },
+	{ "lnxr", 0x61, INSTR_RRE_FF },
+	{ "ltxr", 0x62, INSTR_RRE_FF },
+	{ "lcxr", 0x63, INSTR_RRE_FF },
+	{ "lxr", 0x65, INSTR_RRE_RR },
+	{ "lexr", 0x66, INSTR_RRE_FF },
+	{ "fixr", 0x67, INSTR_RRF_U0FF },
+	{ "cxr", 0x69, INSTR_RRE_FF },
+	{ "lzer", 0x74, INSTR_RRE_R0 },
+	{ "lzdr", 0x75, INSTR_RRE_R0 },
+	{ "lzxr", 0x76, INSTR_RRE_R0 },
+	{ "fier", 0x77, INSTR_RRF_U0FF },
+	{ "fidr", 0x7f, INSTR_RRF_U0FF },
+	{ "sfpc", 0x84, INSTR_RRE_RR_OPT },
+	{ "efpc", 0x8c, INSTR_RRE_RR_OPT },
+	{ "cefbr", 0x94, INSTR_RRE_RF },
+	{ "cdfbr", 0x95, INSTR_RRE_RF },
+	{ "cxfbr", 0x96, INSTR_RRE_RF },
+	{ "cfebr", 0x98, INSTR_RRF_U0RF },
+	{ "cfdbr", 0x99, INSTR_RRF_U0RF },
+	{ "cfxbr", 0x9a, INSTR_RRF_U0RF },
+	{ "cefr", 0xb4, INSTR_RRE_RF },
+	{ "cdfr", 0xb5, INSTR_RRE_RF },
+	{ "cxfr", 0xb6, INSTR_RRE_RF },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_b9[] = {
+#ifdef CONFIG_64BIT
+	{ "lpgr", 0x00, INSTR_RRE_RR },
+	{ "lngr", 0x01, INSTR_RRE_RR },
+	{ "ltgr", 0x02, INSTR_RRE_RR },
+	{ "lcgr", 0x03, INSTR_RRE_RR },
+	{ "lgr", 0x04, INSTR_RRE_RR },
+	{ "lurag", 0x05, INSTR_RRE_RR },
+	{ "lgbr", 0x06, INSTR_RRE_RR },
+	{ "lghr", 0x07, INSTR_RRE_RR },
+	{ "agr", 0x08, INSTR_RRE_RR },
+	{ "sgr", 0x09, INSTR_RRE_RR },
+	{ "algr", 0x0a, INSTR_RRE_RR },
+	{ "slgr", 0x0b, INSTR_RRE_RR },
+	{ "msgr", 0x0c, INSTR_RRE_RR },
+	{ "dsgr", 0x0d, INSTR_RRE_RR },
+	{ "eregg", 0x0e, INSTR_RRE_RR },
+	{ "lrvgr", 0x0f, INSTR_RRE_RR },
+	{ "lpgfr", 0x10, INSTR_RRE_RR },
+	{ "lngfr", 0x11, INSTR_RRE_RR },
+	{ "ltgfr", 0x12, INSTR_RRE_RR },
+	{ "lcgfr", 0x13, INSTR_RRE_RR },
+	{ "lgfr", 0x14, INSTR_RRE_RR },
+	{ "llgfr", 0x16, INSTR_RRE_RR },
+	{ "llgtr", 0x17, INSTR_RRE_RR },
+	{ "agfr", 0x18, INSTR_RRE_RR },
+	{ "sgfr", 0x19, INSTR_RRE_RR },
+	{ "algfr", 0x1a, INSTR_RRE_RR },
+	{ "slgfr", 0x1b, INSTR_RRE_RR },
+	{ "msgfr", 0x1c, INSTR_RRE_RR },
+	{ "dsgfr", 0x1d, INSTR_RRE_RR },
+	{ "cgr", 0x20, INSTR_RRE_RR },
+	{ "clgr", 0x21, INSTR_RRE_RR },
+	{ "sturg", 0x25, INSTR_RRE_RR },
+	{ "lbr", 0x26, INSTR_RRE_RR },
+	{ "lhr", 0x27, INSTR_RRE_RR },
+	{ "cgfr", 0x30, INSTR_RRE_RR },
+	{ "clgfr", 0x31, INSTR_RRE_RR },
+	{ "bctgr", 0x46, INSTR_RRE_RR },
+	{ "ngr", 0x80, INSTR_RRE_RR },
+	{ "ogr", 0x81, INSTR_RRE_RR },
+	{ "xgr", 0x82, INSTR_RRE_RR },
+	{ "flogr", 0x83, INSTR_RRE_RR },
+	{ "llgcr", 0x84, INSTR_RRE_RR },
+	{ "llghr", 0x85, INSTR_RRE_RR },
+	{ "mlgr", 0x86, INSTR_RRE_RR },
+	{ "dlgr", 0x87, INSTR_RRE_RR },
+	{ "alcgr", 0x88, INSTR_RRE_RR },
+	{ "slbgr", 0x89, INSTR_RRE_RR },
+	{ "cspg", 0x8a, INSTR_RRE_RR },
+	{ "idte", 0x8e, INSTR_RRF_R0RR },
+	{ "llcr", 0x94, INSTR_RRE_RR },
+	{ "llhr", 0x95, INSTR_RRE_RR },
+	{ "esea", 0x9d, INSTR_RRE_R0 },
+	{ "lptea", 0xaa, INSTR_RRF_RURR },
+	{ "cu14", 0xb0, INSTR_RRF_M0RR },
+	{ "cu24", 0xb1, INSTR_RRF_M0RR },
+	{ "cu41", 0xb2, INSTR_RRF_M0RR },
+	{ "cu42", 0xb3, INSTR_RRF_M0RR },
+#endif
+	{ "kmac", 0x1e, INSTR_RRE_RR },
+	{ "lrvr", 0x1f, INSTR_RRE_RR },
+	{ "km", 0x2e, INSTR_RRE_RR },
+	{ "kmc", 0x2f, INSTR_RRE_RR },
+	{ "kimd", 0x3e, INSTR_RRE_RR },
+	{ "klmd", 0x3f, INSTR_RRE_RR },
+	{ "epsw", 0x8d, INSTR_RRE_RR },
+	{ "trtt", 0x90, INSTR_RRE_RR },
+	{ "trtt", 0x90, INSTR_RRF_M0RR },
+	{ "trto", 0x91, INSTR_RRE_RR },
+	{ "trto", 0x91, INSTR_RRF_M0RR },
+	{ "trot", 0x92, INSTR_RRE_RR },
+	{ "trot", 0x92, INSTR_RRF_M0RR },
+	{ "troo", 0x93, INSTR_RRE_RR },
+	{ "troo", 0x93, INSTR_RRF_M0RR },
+	{ "mlr", 0x96, INSTR_RRE_RR },
+	{ "dlr", 0x97, INSTR_RRE_RR },
+	{ "alcr", 0x98, INSTR_RRE_RR },
+	{ "slbr", 0x99, INSTR_RRE_RR },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_c0[] = {
+#ifdef CONFIG_64BIT
+	{ "lgfi", 0x01, INSTR_RIL_RI },
+	{ "xihf", 0x06, INSTR_RIL_RU },
+	{ "xilf", 0x07, INSTR_RIL_RU },
+	{ "iihf", 0x08, INSTR_RIL_RU },
+	{ "iilf", 0x09, INSTR_RIL_RU },
+	{ "nihf", 0x0a, INSTR_RIL_RU },
+	{ "nilf", 0x0b, INSTR_RIL_RU },
+	{ "oihf", 0x0c, INSTR_RIL_RU },
+	{ "oilf", 0x0d, INSTR_RIL_RU },
+	{ "llihf", 0x0e, INSTR_RIL_RU },
+	{ "llilf", 0x0f, INSTR_RIL_RU },
+#endif
+	{ "larl", 0x00, INSTR_RIL_RP },
+	{ "brcl", 0x04, INSTR_RIL_UP },
+	{ "brasl", 0x05, INSTR_RIL_RP },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_c2[] = {
+#ifdef CONFIG_64BIT
+	{ "slgfi", 0x04, INSTR_RIL_RU },
+	{ "slfi", 0x05, INSTR_RIL_RU },
+	{ "agfi", 0x08, INSTR_RIL_RI },
+	{ "afi", 0x09, INSTR_RIL_RI },
+	{ "algfi", 0x0a, INSTR_RIL_RU },
+	{ "alfi", 0x0b, INSTR_RIL_RU },
+	{ "cgfi", 0x0c, INSTR_RIL_RI },
+	{ "cfi", 0x0d, INSTR_RIL_RI },
+	{ "clgfi", 0x0e, INSTR_RIL_RU },
+	{ "clfi", 0x0f, INSTR_RIL_RU },
+#endif
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_c8[] = {
+#ifdef CONFIG_64BIT
+	{ "mvcos", 0x00, INSTR_SSF_RRDRD },
+#endif
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_e3[] = {
+#ifdef CONFIG_64BIT
+	{ "ltg", 0x02, INSTR_RXY_RRRD },
+	{ "lrag", 0x03, INSTR_RXY_RRRD },
+	{ "lg", 0x04, INSTR_RXY_RRRD },
+	{ "cvby", 0x06, INSTR_RXY_RRRD },
+	{ "ag", 0x08, INSTR_RXY_RRRD },
+	{ "sg", 0x09, INSTR_RXY_RRRD },
+	{ "alg", 0x0a, INSTR_RXY_RRRD },
+	{ "slg", 0x0b, INSTR_RXY_RRRD },
+	{ "msg", 0x0c, INSTR_RXY_RRRD },
+	{ "dsg", 0x0d, INSTR_RXY_RRRD },
+	{ "cvbg", 0x0e, INSTR_RXY_RRRD },
+	{ "lrvg", 0x0f, INSTR_RXY_RRRD },
+	{ "lt", 0x12, INSTR_RXY_RRRD },
+	{ "lray", 0x13, INSTR_RXY_RRRD },
+	{ "lgf", 0x14, INSTR_RXY_RRRD },
+	{ "lgh", 0x15, INSTR_RXY_RRRD },
+	{ "llgf", 0x16, INSTR_RXY_RRRD },
+	{ "llgt", 0x17, INSTR_RXY_RRRD },
+	{ "agf", 0x18, INSTR_RXY_RRRD },
+	{ "sgf", 0x19, INSTR_RXY_RRRD },
+	{ "algf", 0x1a, INSTR_RXY_RRRD },
+	{ "slgf", 0x1b, INSTR_RXY_RRRD },
+	{ "msgf", 0x1c, INSTR_RXY_RRRD },
+	{ "dsgf", 0x1d, INSTR_RXY_RRRD },
+	{ "cg", 0x20, INSTR_RXY_RRRD },
+	{ "clg", 0x21, INSTR_RXY_RRRD },
+	{ "stg", 0x24, INSTR_RXY_RRRD },
+	{ "cvdy", 0x26, INSTR_RXY_RRRD },
+	{ "cvdg", 0x2e, INSTR_RXY_RRRD },
+	{ "strvg", 0x2f, INSTR_RXY_RRRD },
+	{ "cgf", 0x30, INSTR_RXY_RRRD },
+	{ "clgf", 0x31, INSTR_RXY_RRRD },
+	{ "strvh", 0x3f, INSTR_RXY_RRRD },
+	{ "bctg", 0x46, INSTR_RXY_RRRD },
+	{ "sty", 0x50, INSTR_RXY_RRRD },
+	{ "msy", 0x51, INSTR_RXY_RRRD },
+	{ "ny", 0x54, INSTR_RXY_RRRD },
+	{ "cly", 0x55, INSTR_RXY_RRRD },
+	{ "oy", 0x56, INSTR_RXY_RRRD },
+	{ "xy", 0x57, INSTR_RXY_RRRD },
+	{ "ly", 0x58, INSTR_RXY_RRRD },
+	{ "cy", 0x59, INSTR_RXY_RRRD },
+	{ "ay", 0x5a, INSTR_RXY_RRRD },
+	{ "sy", 0x5b, INSTR_RXY_RRRD },
+	{ "aly", 0x5e, INSTR_RXY_RRRD },
+	{ "sly", 0x5f, INSTR_RXY_RRRD },
+	{ "sthy", 0x70, INSTR_RXY_RRRD },
+	{ "lay", 0x71, INSTR_RXY_RRRD },
+	{ "stcy", 0x72, INSTR_RXY_RRRD },
+	{ "icy", 0x73, INSTR_RXY_RRRD },
+	{ "lb", 0x76, INSTR_RXY_RRRD },
+	{ "lgb", 0x77, INSTR_RXY_RRRD },
+	{ "lhy", 0x78, INSTR_RXY_RRRD },
+	{ "chy", 0x79, INSTR_RXY_RRRD },
+	{ "ahy", 0x7a, INSTR_RXY_RRRD },
+	{ "shy", 0x7b, INSTR_RXY_RRRD },
+	{ "ng", 0x80, INSTR_RXY_RRRD },
+	{ "og", 0x81, INSTR_RXY_RRRD },
+	{ "xg", 0x82, INSTR_RXY_RRRD },
+	{ "mlg", 0x86, INSTR_RXY_RRRD },
+	{ "dlg", 0x87, INSTR_RXY_RRRD },
+	{ "alcg", 0x88, INSTR_RXY_RRRD },
+	{ "slbg", 0x89, INSTR_RXY_RRRD },
+	{ "stpq", 0x8e, INSTR_RXY_RRRD },
+	{ "lpq", 0x8f, INSTR_RXY_RRRD },
+	{ "llgc", 0x90, INSTR_RXY_RRRD },
+	{ "llgh", 0x91, INSTR_RXY_RRRD },
+	{ "llc", 0x94, INSTR_RXY_RRRD },
+	{ "llh", 0x95, INSTR_RXY_RRRD },
+#endif
+	{ "lrv", 0x1e, INSTR_RXY_RRRD },
+	{ "lrvh", 0x1f, INSTR_RXY_RRRD },
+	{ "strv", 0x3e, INSTR_RXY_RRRD },
+	{ "ml", 0x96, INSTR_RXY_RRRD },
+	{ "dl", 0x97, INSTR_RXY_RRRD },
+	{ "alc", 0x98, INSTR_RXY_RRRD },
+	{ "slb", 0x99, INSTR_RXY_RRRD },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_e5[] = {
+#ifdef CONFIG_64BIT
+	{ "strag", 0x02, INSTR_SSE_RDRD },
+#endif
+	{ "lasp", 0x00, INSTR_SSE_RDRD },
+	{ "tprot", 0x01, INSTR_SSE_RDRD },
+	{ "mvcsk", 0x0e, INSTR_SSE_RDRD },
+	{ "mvcdk", 0x0f, INSTR_SSE_RDRD },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_eb[] = {
+#ifdef CONFIG_64BIT
+	{ "lmg", 0x04, INSTR_RSY_RRRD },
+	{ "srag", 0x0a, INSTR_RSY_RRRD },
+	{ "slag", 0x0b, INSTR_RSY_RRRD },
+	{ "srlg", 0x0c, INSTR_RSY_RRRD },
+	{ "sllg", 0x0d, INSTR_RSY_RRRD },
+	{ "tracg", 0x0f, INSTR_RSY_RRRD },
+	{ "csy", 0x14, INSTR_RSY_RRRD },
+	{ "rllg", 0x1c, INSTR_RSY_RRRD },
+	{ "clmh", 0x20, INSTR_RSY_RURD },
+	{ "clmy", 0x21, INSTR_RSY_RURD },
+	{ "stmg", 0x24, INSTR_RSY_RRRD },
+	{ "stctg", 0x25, INSTR_RSY_CCRD },
+	{ "stmh", 0x26, INSTR_RSY_RRRD },
+	{ "stcmh", 0x2c, INSTR_RSY_RURD },
+	{ "stcmy", 0x2d, INSTR_RSY_RURD },
+	{ "lctlg", 0x2f, INSTR_RSY_CCRD },
+	{ "csg", 0x30, INSTR_RSY_RRRD },
+	{ "cdsy", 0x31, INSTR_RSY_RRRD },
+	{ "cdsg", 0x3e, INSTR_RSY_RRRD },
+	{ "bxhg", 0x44, INSTR_RSY_RRRD },
+	{ "bxleg", 0x45, INSTR_RSY_RRRD },
+	{ "tmy", 0x51, INSTR_SIY_URD },
+	{ "mviy", 0x52, INSTR_SIY_URD },
+	{ "niy", 0x54, INSTR_SIY_URD },
+	{ "cliy", 0x55, INSTR_SIY_URD },
+	{ "oiy", 0x56, INSTR_SIY_URD },
+	{ "xiy", 0x57, INSTR_SIY_URD },
+	{ "icmh", 0x80, INSTR_RSE_RURD },
+	{ "icmh", 0x80, INSTR_RSY_RURD },
+	{ "icmy", 0x81, INSTR_RSY_RURD },
+	{ "clclu", 0x8f, INSTR_RSY_RRRD },
+	{ "stmy", 0x90, INSTR_RSY_RRRD },
+	{ "lmh", 0x96, INSTR_RSY_RRRD },
+	{ "lmy", 0x98, INSTR_RSY_RRRD },
+	{ "lamy", 0x9a, INSTR_RSY_AARD },
+	{ "stamy", 0x9b, INSTR_RSY_AARD },
+#endif
+	{ "rll", 0x1d, INSTR_RSY_RRRD },
+	{ "mvclu", 0x8e, INSTR_RSY_RRRD },
+	{ "tp", 0xc0, INSTR_RSL_R0RD },
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_ec[] = {
+#ifdef CONFIG_64BIT
+	{ "brxhg", 0x44, INSTR_RIE_RRP },
+	{ "brxlg", 0x45, INSTR_RIE_RRP },
+#endif
+	{ "", 0, INSTR_INVALID }
+};
+
+static struct insn opcode_ed[] = {
+#ifdef CONFIG_64BIT
+	{ "mayl", 0x38, INSTR_RXF_FRRDF },
+	{ "myl", 0x39, INSTR_RXF_FRRDF },
+	{ "may", 0x3a, INSTR_RXF_FRRDF },
+	{ "my", 0x3b, INSTR_RXF_FRRDF },
+	{ "mayh", 0x3c, INSTR_RXF_FRRDF },
+	{ "myh", 0x3d, INSTR_RXF_FRRDF },
+	{ "ley", 0x64, INSTR_RXY_FRRD },
+	{ "ldy", 0x65, INSTR_RXY_FRRD },
+	{ "stey", 0x66, INSTR_RXY_FRRD },
+	{ "stdy", 0x67, INSTR_RXY_FRRD },
+#endif
+	{ "ldeb", 0x04, INSTR_RXE_FRRD },
+	{ "lxdb", 0x05, INSTR_RXE_FRRD },
+	{ "lxeb", 0x06, INSTR_RXE_FRRD },
+	{ "mxdb", 0x07, INSTR_RXE_FRRD },
+	{ "keb", 0x08, INSTR_RXE_FRRD },
+	{ "ceb", 0x09, INSTR_RXE_FRRD },
+	{ "aeb", 0x0a, INSTR_RXE_FRRD },
+	{ "seb", 0x0b, INSTR_RXE_FRRD },
+	{ "mdeb", 0x0c, INSTR_RXE_FRRD },
+	{ "deb", 0x0d, INSTR_RXE_FRRD },
+	{ "maeb", 0x0e, INSTR_RXF_FRRDF },
+	{ "mseb", 0x0f, INSTR_RXF_FRRDF },
+	{ "tceb", 0x10, INSTR_RXE_FRRD },
+	{ "tcdb", 0x11, INSTR_RXE_FRRD },
+	{ "tcxb", 0x12, INSTR_RXE_FRRD },
+	{ "sqeb", 0x14, INSTR_RXE_FRRD },
+	{ "sqdb", 0x15, INSTR_RXE_FRRD },
+	{ "meeb", 0x17, INSTR_RXE_FRRD },
+	{ "kdb", 0x18, INSTR_RXE_FRRD },
+	{ "cdb", 0x19, INSTR_RXE_FRRD },
+	{ "adb", 0x1a, INSTR_RXE_FRRD },
+	{ "sdb", 0x1b, INSTR_RXE_FRRD },
+	{ "mdb", 0x1c, INSTR_RXE_FRRD },
+	{ "ddb", 0x1d, INSTR_RXE_FRRD },
+	{ "madb", 0x1e, INSTR_RXF_FRRDF },
+	{ "msdb", 0x1f, INSTR_RXF_FRRDF },
+	{ "lde", 0x24, INSTR_RXE_FRRD },
+	{ "lxd", 0x25, INSTR_RXE_FRRD },
+	{ "lxe", 0x26, INSTR_RXE_FRRD },
+	{ "mae", 0x2e, INSTR_RXF_FRRDF },
+	{ "mse", 0x2f, INSTR_RXF_FRRDF },
+	{ "sqe", 0x34, INSTR_RXE_FRRD },
+	{ "mee", 0x37, INSTR_RXE_FRRD },
+	{ "mad", 0x3e, INSTR_RXF_FRRDF },
+	{ "msd", 0x3f, INSTR_RXF_FRRDF },
+	{ "", 0, INSTR_INVALID }
+};
+
+/* Extracts an operand value from an instruction.  */
+static unsigned int extract_operand(unsigned char *code,
+				    const struct operand *operand)
+{
+	unsigned int val;
+	int bits;
+
+	/* Extract fragments of the operand byte for byte.  */
+	code += operand->shift / 8;
+	bits = (operand->shift & 7) + operand->bits;
+	val = 0;
+	do {
+		val <<= 8;
+		val |= (unsigned int) *code++;
+		bits -= 8;
+	} while (bits > 0);
+	val >>= -bits;
+	val &= ((1U << (operand->bits - 1)) << 1) - 1;
+
+	/* Check for special long displacement case.  */
+	if (operand->bits == 20 && operand->shift == 20)
+		val = (val & 0xff) << 12 | (val & 0xfff00) >> 8;
+
+	/* Sign extend value if the operand is signed or pc relative.  */
+	if ((operand->flags & (OPERAND_SIGNED | OPERAND_PCREL)) &&
+	    (val & (1U << (operand->bits - 1))))
+		val |= (-1U << (operand->bits - 1)) << 1;
+
+	/* Double value if the operand is pc relative.	*/
+	if (operand->flags & OPERAND_PCREL)
+		val <<= 1;
+
+	/* Length x in an instructions has real length x + 1.  */
+	if (operand->flags & OPERAND_LENGTH)
+		val++;
+	return val;
+}
+
+static inline int insn_length(unsigned char code)
+{
+	return ((((int) code + 64) >> 7) + 1) << 1;
+}
+
+static struct insn *find_insn(unsigned char *code)
+{
+	unsigned char opfrag = code[1];
+	unsigned char opmask;
+	struct insn *table;
+
+	switch (code[0]) {
+	case 0x01:
+		table = opcode_01;
+		break;
+	case 0xa5:
+		table = opcode_a5;
+		break;
+	case 0xa7:
+		table = opcode_a7;
+		break;
+	case 0xb2:
+		table = opcode_b2;
+		break;
+	case 0xb3:
+		table = opcode_b3;
+		break;
+	case 0xb9:
+		table = opcode_b9;
+		break;
+	case 0xc0:
+		table = opcode_c0;
+		break;
+	case 0xc2:
+		table = opcode_c2;
+		break;
+	case 0xc8:
+		table = opcode_c8;
+		break;
+	case 0xe3:
+		table = opcode_e3;
+		opfrag = code[5];
+		break;
+	case 0xe5:
+		table = opcode_e5;
+		break;
+	case 0xeb:
+		table = opcode_eb;
+		opfrag = code[5];
+		break;
+	case 0xec:
+		table = opcode_ec;
+		opfrag = code[5];
+		break;
+	case 0xed:
+		table = opcode_ed;
+		opfrag = code[5];
+		break;
+	default:
+		table = opcode;
+		opfrag = code[0];
+		break;
+	}
+	while (table->format != INSTR_INVALID) {
+		opmask = formats[table->format][0];
+		if (table->opfrag == (opfrag & opmask))
+			return table;
+		table++;
+	}
+	return NULL;
+}
+
+static int print_insn(char *buffer, unsigned char *code, unsigned long addr)
+{
+	struct insn *insn;
+	const unsigned char *ops;
+	const struct operand *operand;
+	unsigned int value;
+	char separator;
+	char *ptr;
+
+	ptr = buffer;
+	insn = find_insn(code);
+	if (insn) {
+		ptr += sprintf(ptr, "%.5s\t", insn->name);
+		/* Extract the operands. */
+		separator = 0;
+		for (ops = formats[insn->format] + 1; *ops != 0; ops++) {
+			operand = operands + *ops;
+			value = extract_operand(code, operand);
+			if ((operand->flags & OPERAND_INDEX)  && value == 0)
+				continue;
+			if ((operand->flags & OPERAND_BASE) &&
+			    value == 0 && separator == '(') {
+				separator = ',';
+				continue;
+			}
+			if (separator)
+				ptr += sprintf(ptr, "%c", separator);
+			if (operand->flags & OPERAND_GPR)
+				ptr += sprintf(ptr, "%%r%i", value);
+			else if (operand->flags & OPERAND_FPR)
+				ptr += sprintf(ptr, "%%f%i", value);
+			else if (operand->flags & OPERAND_AR)
+				ptr += sprintf(ptr, "%%a%i", value);
+			else if (operand->flags & OPERAND_CR)
+				ptr += sprintf(ptr, "%%c%i", value);
+			else if (operand->flags & OPERAND_PCREL)
+				ptr += sprintf(ptr, "%lx", value + addr);
+			else if (operand->flags & OPERAND_SIGNED)
+				ptr += sprintf(ptr, "%i", value);
+			else
+				ptr += sprintf(ptr, "%u", value);
+			if (operand->flags & OPERAND_DISP)
+				separator = '(';
+			else if (operand->flags & OPERAND_BASE) {
+				ptr += sprintf(ptr, ")");
+				separator = ',';
+			} else
+				separator = ',';
+		}
+	} else
+		ptr += sprintf(ptr, "unknown");
+	return (int) (ptr - buffer);
+}
+
+void show_code(struct pt_regs *regs)
+{
+	char *mode = (regs->psw.mask & PSW_MASK_PSTATE) ? "User" : "Krnl";
+	unsigned char code[64];
+	char buffer[64], *ptr;
+	mm_segment_t old_fs;
+	unsigned long addr;
+	int start, end, opsize, hops, i;
+
+	/* Get a snapshot of the 64 bytes surrounding the fault address. */
+	old_fs = get_fs();
+	set_fs((regs->psw.mask & PSW_MASK_PSTATE) ? USER_DS : KERNEL_DS);
+	for (start = 32; start && regs->psw.addr >= 34 - start; start -= 2) {
+		addr = regs->psw.addr - 34 + start;
+		if (__copy_from_user(code + start - 2,
+				     (char __user *) addr, 2))
+			break;
+	}
+	for (end = 32; end < 64; end += 2) {
+		addr = regs->psw.addr + end - 32;
+		if (__copy_from_user(code + end,
+				     (char __user *) addr, 2))
+			break;
+	}
+	set_fs(old_fs);
+	/* Code snapshot useable ? */
+	if ((regs->psw.addr & 1) || start >= end) {
+		printk("%s Code: Bad PSW.\n", mode);
+		return;
+	}
+	/* Find a starting point for the disassembly. */
+	while (start < 32) {
+		hops = 0;
+		for (i = 0, hops = 0; start + i < 32 && hops < 3; hops++) {
+			if (!find_insn(code + start + i))
+				break;
+			i += insn_length(code[start + i]);
+		}
+		if (start + i == 32)
+			/* Looks good, sequence ends at PSW. */
+			break;
+		start += 2;
+	}
+	/* Decode the instructions. */
+	ptr = buffer;
+	ptr += sprintf(ptr, "%s Code:", mode);
+	hops = 0;
+	while (start < end && hops < 8) {
+		*ptr++ = (start == 32) ? '>' : ' ';
+		addr = regs->psw.addr + start - 32;
+		ptr += sprintf(ptr, ONELONG, addr);
+		opsize = insn_length(code[start]);
+		if (start + opsize >= end)
+			break;
+		for (i = 0; i < opsize; i++)
+			ptr += sprintf(ptr, "%02x", code[start + i]);
+		*ptr++ = '\t';
+		if (i < 6)
+			*ptr++ = '\t';
+		ptr += print_insn(ptr, code + start, addr);
+		start += opsize;
+		printk(buffer);
+		ptr = buffer;
+		ptr += sprintf(ptr, "\n          ");
+		hops++;
+	}
+	printk("\n");
+}
Index: quilt-2.6/arch/s390/kernel/Makefile
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/Makefile	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/Makefile	2007-04-27 16:04:56.000000000 +0200
@@ -6,7 +6,7 @@
 
 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 irq.o ipl.o
+	    semaphore.o s390_ext.o debug.o irq.o ipl.o dis.o
 
 obj-y	+= $(if $(CONFIG_64BIT),entry64.o,entry.o)
 obj-y	+= $(if $(CONFIG_64BIT),reipl64.o,reipl.o)
Index: quilt-2.6/arch/s390/kernel/traps.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/traps.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/traps.c	2007-04-27 16:04:56.000000000 +0200
@@ -188,18 +188,31 @@
 
 EXPORT_SYMBOL(dump_stack);
 
+static inline int mask_bits(struct pt_regs *regs, unsigned long bits)
+{
+	return (regs->psw.mask & bits) / ((~bits + 1) & bits);
+}
+
 void show_registers(struct pt_regs *regs)
 {
-	mm_segment_t old_fs;
 	char *mode;
-	int i;
 
 	mode = (regs->psw.mask & PSW_MASK_PSTATE) ? "User" : "Krnl";
 	printk("%s PSW : %p %p",
 	       mode, (void *) regs->psw.mask,
 	       (void *) regs->psw.addr);
 	print_symbol(" (%s)\n", regs->psw.addr & PSW_ADDR_INSN);
-	printk("%s GPRS: " FOURLONG, mode,
+	printk("           R:%x T:%x IO:%x EX:%x Key:%x M:%x W:%x "
+	       "P:%x AS:%x CC:%x PM:%x", mask_bits(regs, PSW_MASK_PER),
+	       mask_bits(regs, PSW_MASK_DAT), mask_bits(regs, PSW_MASK_IO),
+	       mask_bits(regs, PSW_MASK_EXT), mask_bits(regs, PSW_MASK_KEY),
+	       mask_bits(regs, PSW_MASK_MCHECK), mask_bits(regs, PSW_MASK_WAIT),
+	       mask_bits(regs, PSW_MASK_PSTATE), mask_bits(regs, PSW_MASK_ASC),
+	       mask_bits(regs, PSW_MASK_CC), mask_bits(regs, PSW_MASK_PM));
+#ifdef CONFIG_64BIT
+	printk(" EA:%x", mask_bits(regs, PSW_BASE_BITS));
+#endif
+	printk("\n%s GPRS: " FOURLONG, mode,
 	       regs->gprs[0], regs->gprs[1], regs->gprs[2], regs->gprs[3]);
 	printk("           " FOURLONG,
 	       regs->gprs[4], regs->gprs[5], regs->gprs[6], regs->gprs[7]);
@@ -208,41 +221,7 @@
 	printk("           " FOURLONG,
 	       regs->gprs[12], regs->gprs[13], regs->gprs[14], regs->gprs[15]);
 
-#if 0
-	/* FIXME: this isn't needed any more but it changes the ksymoops
-	 * input. To remove or not to remove ... */
-	save_access_regs(regs->acrs);
-	printk("%s ACRS: %08x %08x %08x %08x\n", mode,
-	       regs->acrs[0], regs->acrs[1], regs->acrs[2], regs->acrs[3]);
-	printk("           %08x %08x %08x %08x\n",
-	       regs->acrs[4], regs->acrs[5], regs->acrs[6], regs->acrs[7]);
-	printk("           %08x %08x %08x %08x\n",
-	       regs->acrs[8], regs->acrs[9], regs->acrs[10], regs->acrs[11]);
-	printk("           %08x %08x %08x %08x\n",
-	       regs->acrs[12], regs->acrs[13], regs->acrs[14], regs->acrs[15]);
-#endif
-
-	/*
-	 * Print the first 20 byte of the instruction stream at the
-	 * time of the fault.
-	 */
-	old_fs = get_fs();
-	if (regs->psw.mask & PSW_MASK_PSTATE)
-		set_fs(USER_DS);
-	else
-		set_fs(KERNEL_DS);
-	printk("%s Code: ", mode);
-	for (i = 0; i < 20; i++) {
-		unsigned char c;
-		if (__get_user(c, (char __user *)(regs->psw.addr + i))) {
-			printk(" Bad PSW.");
-			break;
-		}
-		printk("%02x ", c);
-	}
-	set_fs(old_fs);
-
-	printk("\n");
+	show_code(regs);
 }	
 
 /* This is called from fs/proc/array.c */
Index: quilt-2.6/include/asm-s390/processor.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/processor.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/processor.h	2007-04-27 16:04:56.000000000 +0200
@@ -196,6 +196,7 @@
 extern char *task_show_regs(struct task_struct *task, char *buffer);
 
 extern void show_registers(struct pt_regs *regs);
+extern void show_code(struct pt_regs *regs);
 extern void show_trace(struct task_struct *task, unsigned long *sp);
 
 unsigned long get_wchan(struct task_struct *p);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 19/38] Use generic bug.

[Martin Schwidefsky]

[-- Attachment #1: 122-bug-output.diff --]
[-- Type: text/plain, Size: 7001 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

Generic bug implementation for s390. Will increase the value of the
console output on BUG() statements since registers r0-r5,r14 will
not be clobbered by a printk() call that was previously done before
the illegal instruction of BUG() was hit.
Also implements an architecture specific WARN_ON(). Output of that
could be increased but requires common code change.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/Kconfig              |    5 ++
 arch/s390/defconfig            |    2 +
 arch/s390/kernel/module.c      |    4 +-
 arch/s390/kernel/traps.c       |   17 ++++++++--
 arch/s390/kernel/vmlinux.lds.S |   10 +++++
 include/asm-s390/bug.h         |   69 +++++++++++++++++++++++++++++++++--------
 6 files changed, 89 insertions(+), 18 deletions(-)

Index: quilt-2.6/arch/s390/defconfig
===================================================================
--- quilt-2.6.orig/arch/s390/defconfig	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/defconfig	2007-04-27 16:04:56.000000000 +0200
@@ -12,6 +12,7 @@
 # CONFIG_ARCH_HAS_ILOG2_U64 is not set
 CONFIG_GENERIC_HWEIGHT=y
 CONFIG_GENERIC_TIME=y
+CONFIG_GENERIC_BUG=y
 CONFIG_NO_IOMEM=y
 CONFIG_S390=y
 CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
@@ -705,6 +706,7 @@
 CONFIG_DEBUG_SPINLOCK_SLEEP=y
 # CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
 # CONFIG_DEBUG_KOBJECT is not set
+CONFIG_DEBUG_BUGVERBOSE=y
 # CONFIG_DEBUG_INFO is not set
 # CONFIG_DEBUG_VM is not set
 # CONFIG_DEBUG_LIST is not set
Index: quilt-2.6/arch/s390/Kconfig
===================================================================
--- quilt-2.6.orig/arch/s390/Kconfig	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/Kconfig	2007-04-27 16:04:56.000000000 +0200
@@ -41,6 +41,11 @@
 config GENERIC_TIME
 	def_bool y
 
+config GENERIC_BUG
+	bool
+	depends on BUG
+	default y
+
 config NO_IOMEM
 	def_bool y
 
Index: quilt-2.6/arch/s390/kernel/module.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/module.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/module.c	2007-04-27 16:04:56.000000000 +0200
@@ -31,6 +31,7 @@
 #include <linux/string.h>
 #include <linux/kernel.h>
 #include <linux/moduleloader.h>
+#include <linux/bug.h>
 
 #if 0
 #define DEBUGP printk
@@ -398,9 +399,10 @@
 		    struct module *me)
 {
 	vfree(me->arch.syminfo);
-	return 0;
+	return module_bug_finalize(hdr, sechdrs, me);
 }
 
 void module_arch_cleanup(struct module *mod)
 {
+	module_bug_cleanup(mod);
 }
Index: quilt-2.6/arch/s390/kernel/traps.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/traps.c	2007-04-27 16:04:56.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/traps.c	2007-04-27 16:04:56.000000000 +0200
@@ -30,7 +30,7 @@
 #include <linux/kallsyms.h>
 #include <linux/reboot.h>
 #include <linux/kprobes.h>
-
+#include <linux/bug.h>
 #include <asm/system.h>
 #include <asm/uaccess.h>
 #include <asm/io.h>
@@ -297,6 +297,11 @@
 #endif
 }
 
+int is_valid_bugaddr(unsigned long addr)
+{
+	return 1;
+}
+
 static void __kprobes inline do_trap(long interruption_code, int signr,
 					char *str, struct pt_regs *regs,
 					siginfo_t *info)
@@ -323,8 +328,14 @@
                 fixup = search_exception_tables(regs->psw.addr & PSW_ADDR_INSN);
                 if (fixup)
                         regs->psw.addr = fixup->fixup | PSW_ADDR_AMODE;
-                else
-                        die(str, regs, interruption_code);
+		else {
+			enum bug_trap_type btt;
+
+			btt = report_bug(regs->psw.addr & PSW_ADDR_INSN);
+			if (btt == BUG_TRAP_TYPE_WARN)
+				return;
+			die(str, regs, interruption_code);
+		}
         }
 }
 
Index: quilt-2.6/arch/s390/kernel/vmlinux.lds.S
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/vmlinux.lds.S	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/vmlinux.lds.S	2007-04-27 16:04:56.000000000 +0200
@@ -45,6 +45,8 @@
   __ex_table : { *(__ex_table) }
   __stop___ex_table = .;
 
+  BUG_TABLE
+
   .data : {			/* Data */
 	*(.data)
 	CONSTRUCTORS
@@ -77,6 +79,12 @@
 	*(.init.text)
 	_einittext = .;
   }
+  /*
+   * .exit.text is discarded at runtime, not link time,
+   * to deal with references from __bug_table
+   */
+  .exit.text :	 { *(.exit.text) }
+
   .init.data : { *(.init.data) }
   . = ALIGN(256);
   __setup_start = .;
@@ -116,7 +124,7 @@
 
   /* Sections to be discarded */
   /DISCARD/ : {
-	*(.exit.text) *(.exit.data) *(.exitcall.exit)
+	*(.exit.data) *(.exitcall.exit)
 	}
 
   /* Stabs debugging sections.  */
Index: quilt-2.6/include/asm-s390/bug.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/bug.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/bug.h	2007-04-27 16:04:56.000000000 +0200
@@ -1,27 +1,70 @@
-#ifndef _S390_BUG_H
-#define _S390_BUG_H
+#ifndef _ASM_S390_BUG_H
+#define _ASM_S390_BUG_H
 
 #include <linux/kernel.h>
 
 #ifdef CONFIG_BUG
 
-static inline __attribute__((noreturn)) void __do_illegal_op(void)
-{
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 3)
-	__builtin_trap();
+#ifdef CONFIG_64BIT
+#define S390_LONG ".quad"
 #else
-	asm volatile(".long 0");
+#define S390_LONG ".long"
 #endif
-}
 
-#define BUG() do { \
-	printk("kernel BUG at %s:%d!\n", __FILE__, __LINE__); \
-	__do_illegal_op(); \
+#ifdef CONFIG_DEBUG_BUGVERBOSE
+
+#define __EMIT_BUG(x) do {					\
+	asm volatile(						\
+		"0:	j	0b+2\n"				\
+		"1:\n"						\
+		".section .rodata.str,\"aMS\",@progbits,1\n"	\
+		"2:	.asciz	\""__FILE__"\"\n"		\
+		".previous\n"					\
+		".section __bug_table,\"a\"\n"			\
+		"3:\t"	S390_LONG "\t1b,2b\n"			\
+		"	.short	%0,%1\n"			\
+		"	.org	3b+%2\n"			\
+		".previous\n"					\
+		: : "i" (__LINE__),				\
+		    "i" (x),					\
+		    "i" (sizeof(struct bug_entry)));		\
 } while (0)
 
+#else /* CONFIG_DEBUG_BUGVERBOSE */
+
+#define __EMIT_BUG(x) do {				\
+	asm volatile(					\
+		"0:	j	0b+2\n"			\
+		"1:\n"					\
+		".section __bug_table,\"a\"\n"		\
+		"2:\t"	S390_LONG "\t1b\n"		\
+		"	.short	%0\n"			\
+		"	.org	2b+%1\n"		\
+		".previous\n"				\
+		: : "i" (x),				\
+		    "i" (sizeof(struct bug_entry)));	\
+} while (0)
+
+#endif /* CONFIG_DEBUG_BUGVERBOSE */
+
+#define BUG()	__EMIT_BUG(0)
+
+#define WARN_ON(x) ({					\
+	typeof(x) __ret_warn_on = (x);			\
+	if (__builtin_constant_p(__ret_warn_on)) {	\
+		if (__ret_warn_on)			\
+			__EMIT_BUG(BUGFLAG_WARNING);	\
+	} else {					\
+		if (unlikely(__ret_warn_on))		\
+			__EMIT_BUG(BUGFLAG_WARNING);	\
+	}						\
+	unlikely(__ret_warn_on);			\
+})
+
 #define HAVE_ARCH_BUG
-#endif
+#define HAVE_ARCH_WARN_ON
+#endif /* CONFIG_BUG */
 
 #include <asm-generic/bug.h>
 
-#endif
+#endif /* _ASM_S390_BUG_H */

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 20/38] Minor fault path optimization.

[Martin Schwidefsky]

[-- Attachment #1: 123-minor-fault.diff --]
[-- Type: text/plain, Size: 10626 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

The minor fault path has grown a lot in terms of cycles. In particular
the kprobes hook is very costly. Optimize the path to save a couple of
cycles. If kprobes is enabled more than 300 cycles can be avoided if 
kprobes_running() is false.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/mm/fault.c |  259 +++++++++++++++++++++++++++------------------------
 1 files changed, 141 insertions(+), 118 deletions(-)

Index: quilt-2.6/arch/s390/mm/fault.c
===================================================================
--- quilt-2.6.orig/arch/s390/mm/fault.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/mm/fault.c	2007-04-27 16:04:57.000000000 +0200
@@ -63,21 +63,25 @@
 	return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb);
 }
 
-static inline int notify_page_fault(enum die_val val, const char *str,
-			struct pt_regs *regs, long err, int trap, int sig)
+static int __kprobes __notify_page_fault(struct pt_regs *regs, long err)
 {
-	struct die_args args = {
-		.regs = regs,
-		.str = str,
-		.err = err,
-		.trapnr = trap,
-		.signr = sig
-	};
-	return atomic_notifier_call_chain(&notify_page_fault_chain, val, &args);
+	struct die_args args = { .str = "page fault",
+				 .trapnr = 14,
+				 .signr = SIGSEGV };
+	args.regs = regs;
+	args.err = err;
+	return atomic_notifier_call_chain(&notify_page_fault_chain,
+					  DIE_PAGE_FAULT, &args);
+}
+
+static inline int notify_page_fault(struct pt_regs *regs, long err)
+{
+	if (unlikely(kprobe_running()))
+		return __notify_page_fault(regs, err);
+	return NOTIFY_DONE;
 }
 #else
-static inline int notify_page_fault(enum die_val val, const char *str,
-			struct pt_regs *regs, long err, int trap, int sig)
+static inline int notify_page_fault(struct pt_regs *regs, long err)
 {
 	return NOTIFY_DONE;
 }
@@ -170,6 +174,89 @@
 	force_sig_info(SIGSEGV, &si, current);
 }
 
+static void do_no_context(struct pt_regs *regs, unsigned long error_code,
+			  unsigned long address)
+{
+	const struct exception_table_entry *fixup;
+
+	/* Are we prepared to handle this kernel fault?  */
+	fixup = search_exception_tables(regs->psw.addr & __FIXUP_MASK);
+	if (fixup) {
+		regs->psw.addr = fixup->fixup | PSW_ADDR_AMODE;
+		return;
+	}
+
+	/*
+	 * Oops. The kernel tried to access some bad page. We'll have to
+	 * terminate things with extreme prejudice.
+	 */
+	if (check_space(current) == 0)
+		printk(KERN_ALERT "Unable to handle kernel pointer dereference"
+		       " at virtual kernel address %p\n", (void *)address);
+	else
+		printk(KERN_ALERT "Unable to handle kernel paging request"
+		       " at virtual user address %p\n", (void *)address);
+
+	die("Oops", regs, error_code);
+	do_exit(SIGKILL);
+}
+
+static void do_low_address(struct pt_regs *regs, unsigned long error_code)
+{
+	/* Low-address protection hit in kernel mode means
+	   NULL pointer write access in kernel mode.  */
+	if (regs->psw.mask & PSW_MASK_PSTATE) {
+		/* Low-address protection hit in user mode 'cannot happen'. */
+		die ("Low-address protection", regs, error_code);
+		do_exit(SIGKILL);
+	}
+
+	do_no_context(regs, error_code, 0);
+}
+
+/*
+ * We ran out of memory, or some other thing happened to us that made
+ * us unable to handle the page fault gracefully.
+ */
+static int do_out_of_memory(struct pt_regs *regs, unsigned long error_code,
+			    unsigned long address)
+{
+	struct task_struct *tsk = current;
+	struct mm_struct *mm = tsk->mm;
+
+	up_read(&mm->mmap_sem);
+	if (is_init(tsk)) {
+		yield();
+		down_read(&mm->mmap_sem);
+		return 1;
+	}
+	printk("VM: killing process %s\n", tsk->comm);
+	if (regs->psw.mask & PSW_MASK_PSTATE)
+		do_exit(SIGKILL);
+	do_no_context(regs, error_code, address);
+	return 0;
+}
+
+static void do_sigbus(struct pt_regs *regs, unsigned long error_code,
+		      unsigned long address)
+{
+	struct task_struct *tsk = current;
+	struct mm_struct *mm = tsk->mm;
+
+	up_read(&mm->mmap_sem);
+	/*
+	 * Send a sigbus, regardless of whether we were in kernel
+	 * or user mode.
+	 */
+	tsk->thread.prot_addr = address;
+	tsk->thread.trap_no = error_code;
+	force_sig(SIGBUS, tsk);
+
+	/* Kernel mode? Handle exceptions or die */
+	if (!(regs->psw.mask & PSW_MASK_PSTATE))
+		do_no_context(regs, error_code, address);
+}
+
 #ifdef CONFIG_S390_EXEC_PROTECT
 extern long sys_sigreturn(struct pt_regs *regs);
 extern long sys_rt_sigreturn(struct pt_regs *regs);
@@ -253,49 +340,23 @@
  *   3b       Region third trans.  ->  Not present       (nullification)
  */
 static inline void
-do_exception(struct pt_regs *regs, unsigned long error_code, int is_protection)
+do_exception(struct pt_regs *regs, unsigned long error_code, int write)
 {
-        struct task_struct *tsk;
-        struct mm_struct *mm;
-        struct vm_area_struct * vma;
-        unsigned long address;
-	const struct exception_table_entry *fixup;
-	int si_code;
+	struct task_struct *tsk;
+	struct mm_struct *mm;
+	struct vm_area_struct *vma;
+	unsigned long address;
 	int space;
+	int si_code;
 
-        tsk = current;
-        mm = tsk->mm;
-	
-	if (notify_page_fault(DIE_PAGE_FAULT, "page fault", regs, error_code, 14,
-					SIGSEGV) == NOTIFY_STOP)
+	if (notify_page_fault(regs, error_code) == NOTIFY_STOP)
 		return;
 
-	/* 
-         * Check for low-address protection.  This needs to be treated
-	 * as a special case because the translation exception code 
-	 * field is not guaranteed to contain valid data in this case.
-	 */
-	if (is_protection && !(S390_lowcore.trans_exc_code & 4)) {
-
-		/* Low-address protection hit in kernel mode means 
-		   NULL pointer write access in kernel mode.  */
- 		if (!(regs->psw.mask & PSW_MASK_PSTATE)) {
-			address = 0;
-			space = 0;
-			goto no_context;
-		}
-
-		/* Low-address protection hit in user mode 'cannot happen'.  */
-		die ("Low-address protection", regs, error_code);
-        	do_exit(SIGKILL);
-	}
+	tsk = current;
+	mm = tsk->mm;
 
-        /* 
-         * get the failing address 
-         * more specific the segment and page table portion of 
-         * the address 
-         */
-        address = S390_lowcore.trans_exc_code & __FAIL_ADDR_MASK;
+	/* get the failing address and the affected space */
+	address = S390_lowcore.trans_exc_code & __FAIL_ADDR_MASK;
 	space = check_space(tsk);
 
 	/*
@@ -313,7 +374,7 @@
 	 */
 	local_irq_enable();
 
-        down_read(&mm->mmap_sem);
+	down_read(&mm->mmap_sem);
 
 	si_code = SEGV_MAPERR;
 	vma = find_vma(mm, address);
@@ -330,19 +391,19 @@
 			return;
 #endif
 
-        if (vma->vm_start <= address) 
-                goto good_area;
-        if (!(vma->vm_flags & VM_GROWSDOWN))
-                goto bad_area;
-        if (expand_stack(vma, address))
-                goto bad_area;
+	if (vma->vm_start <= address)
+		goto good_area;
+	if (!(vma->vm_flags & VM_GROWSDOWN))
+		goto bad_area;
+	if (expand_stack(vma, address))
+		goto bad_area;
 /*
  * Ok, we have a good vm_area for this memory access, so
  * we can handle it..
  */
 good_area:
 	si_code = SEGV_ACCERR;
-	if (!is_protection) {
+	if (!write) {
 		/* page not present, check vm flags */
 		if (!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)))
 			goto bad_area;
@@ -357,7 +418,7 @@
 	 * make sure we exit gracefully rather than endlessly redo
 	 * the fault.
 	 */
-	switch (handle_mm_fault(mm, vma, address, is_protection)) {
+	switch (handle_mm_fault(mm, vma, address, write)) {
 	case VM_FAULT_MINOR:
 		tsk->min_flt++;
 		break;
@@ -365,9 +426,12 @@
 		tsk->maj_flt++;
 		break;
 	case VM_FAULT_SIGBUS:
-		goto do_sigbus;
+		do_sigbus(regs, error_code, address);
+		return;
 	case VM_FAULT_OOM:
-		goto out_of_memory;
+		if (do_out_of_memory(regs, error_code, address))
+			goto survive;
+		return;
 	default:
 		BUG();
 	}
@@ -385,75 +449,34 @@
  * Fix it, but check if it's kernel or user first..
  */
 bad_area:
-        up_read(&mm->mmap_sem);
+	up_read(&mm->mmap_sem);
 
-        /* User mode accesses just cause a SIGSEGV */
-        if (regs->psw.mask & PSW_MASK_PSTATE) {
-                tsk->thread.prot_addr = address;
-                tsk->thread.trap_no = error_code;
+	/* User mode accesses just cause a SIGSEGV */
+	if (regs->psw.mask & PSW_MASK_PSTATE) {
+		tsk->thread.prot_addr = address;
+		tsk->thread.trap_no = error_code;
 		do_sigsegv(regs, error_code, si_code, address);
-                return;
+		return;
 	}
 
 no_context:
-        /* Are we prepared to handle this kernel fault?  */
-	fixup = search_exception_tables(regs->psw.addr & __FIXUP_MASK);
-	if (fixup) {
-		regs->psw.addr = fixup->fixup | PSW_ADDR_AMODE;
-                return;
-        }
-
-/*
- * Oops. The kernel tried to access some bad page. We'll have to
- * terminate things with extreme prejudice.
- */
-	if (space == 0)
-                printk(KERN_ALERT "Unable to handle kernel pointer dereference"
-        	       " at virtual kernel address %p\n", (void *)address);
-        else
-                printk(KERN_ALERT "Unable to handle kernel paging request"
-		       " at virtual user address %p\n", (void *)address);
-
-        die("Oops", regs, error_code);
-        do_exit(SIGKILL);
-
-
-/*
- * We ran out of memory, or some other thing happened to us that made
- * us unable to handle the page fault gracefully.
-*/
-out_of_memory:
-	up_read(&mm->mmap_sem);
-	if (is_init(tsk)) {
-		yield();
-		down_read(&mm->mmap_sem);
-		goto survive;
-	}
-	printk("VM: killing process %s\n", tsk->comm);
-	if (regs->psw.mask & PSW_MASK_PSTATE)
-		do_exit(SIGKILL);
-	goto no_context;
-
-do_sigbus:
-	up_read(&mm->mmap_sem);
-
-	/*
-	 * Send a sigbus, regardless of whether we were in kernel
-	 * or user mode.
-	 */
-        tsk->thread.prot_addr = address;
-        tsk->thread.trap_no = error_code;
-	force_sig(SIGBUS, tsk);
-
-	/* Kernel mode? Handle exceptions or die */
-	if (!(regs->psw.mask & PSW_MASK_PSTATE))
-		goto no_context;
+	do_no_context(regs, error_code, address);
 }
 
 void __kprobes do_protection_exception(struct pt_regs *regs,
 				       unsigned long error_code)
 {
+	/* Protection exception is supressing, decrement psw address. */
 	regs->psw.addr -= (error_code >> 16);
+	/*
+	 * Check for low-address protection.  This needs to be treated
+	 * as a special case because the translation exception code
+	 * field is not guaranteed to contain valid data in this case.
+	 */
+	if (unlikely(!(S390_lowcore.trans_exc_code & 4))) {
+		do_low_address(regs, error_code);
+		return;
+	}
 	do_exception(regs, 4, 1);
 }
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 21/38] No execute support cleanup.

[Martin Schwidefsky]

[-- Attachment #1: 124-nx-cleanup.diff --]
[-- Type: text/plain, Size: 3579 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Simplify the signal_return function that checks for the two special
system calls sigreturn and rt_sigreturn. No need to do a page table
walk, a call to copy_from_user while disabled page faults will work
as well.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/mm/fault.c |   72 ++++++++++++++-------------------------------------
 1 files changed, 21 insertions(+), 51 deletions(-)

Index: quilt-2.6/arch/s390/mm/fault.c
===================================================================
--- quilt-2.6.orig/arch/s390/mm/fault.c	2007-04-27 16:04:57.000000000 +0200
+++ quilt-2.6/arch/s390/mm/fault.c	2007-04-27 16:04:57.000000000 +0200
@@ -26,9 +26,9 @@
 #include <linux/module.h>
 #include <linux/hardirq.h>
 #include <linux/kprobes.h>
+#include <linux/uaccess.h>
 
 #include <asm/system.h>
-#include <asm/uaccess.h>
 #include <asm/pgtable.h>
 #include <asm/kdebug.h>
 #include <asm/s390_ext.h>
@@ -263,68 +263,38 @@
 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)
+static int signal_return(struct mm_struct *mm, struct pt_regs *regs,
+			 unsigned long address, unsigned long error_code)
 {
+	u16 instruction;
+	int rc, compat;
+
+	pagefault_disable();
+	rc = __get_user(instruction, (u16 __user *) regs->psw.addr);
+	pagefault_enable();
+	if (rc)
+		return -EFAULT;
+
 	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);
+	compat = test_tsk_thread_flag(current, TIF_31BIT);
+	if (compat && instruction == 0x0a77)
+		sys32_sigreturn(regs);
+	else if (compat && instruction == 0x0aad)
+		sys32_rt_sigreturn(regs);
 	else
+#endif
+	if (instruction == 0x0a77)
 		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 if (instruction == 0x0aad)
+		sys_rt_sigreturn(regs);
 	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 */
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 22/38] Get rid of console setup functions.

[Martin Schwidefsky]

[-- Attachment #1: 125-console-setup.diff --]
[-- Type: text/plain, Size: 2123 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

We get this:

Section mismatch: reference to .init.text:con3270_consetup from .data
		  between 'con3270' (at offset 0x45c8) and 'con3270_fn'
Section mismatch: reference to .init.text:con3215_consetup from .data
		  between 'con3215' (at offset 0x4678) and
		  'raw3215_ccw_driver'

Since there is no difference between a non present console setup
function and one that returns only 0 remove them.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 drivers/s390/char/con3215.c |    7 -------
 drivers/s390/char/con3270.c |    7 -------
 2 files changed, 14 deletions(-)

Index: quilt-2.6/drivers/s390/char/con3215.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/con3215.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/con3215.c	2007-04-27 16:04:58.000000000 +0200
@@ -813,12 +813,6 @@
 	spin_unlock_irqrestore(get_ccwdev_lock(raw->cdev), flags);
 }
 
-static int __init 
-con3215_consetup(struct console *co, char *options)
-{
-	return 0;
-}
-
 /*
  *  The console structure for the 3215 console
  */
@@ -827,7 +821,6 @@
 	.write	 = con3215_write,
 	.device	 = con3215_device,
 	.unblank = con3215_unblank,
-	.setup	 = con3215_consetup,
 	.flags	 = CON_PRINTBUFFER,
 };
 
Index: quilt-2.6/drivers/s390/char/con3270.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/con3270.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/con3270.c	2007-04-27 16:04:58.000000000 +0200
@@ -555,12 +555,6 @@
 	spin_unlock_irqrestore(&cp->view.lock, flags);
 }
 
-static int __init 
-con3270_consetup(struct console *co, char *options)
-{
-	return 0;
-}
-
 /*
  *  The console structure for the 3270 console
  */
@@ -569,7 +563,6 @@
 	.write	 = con3270_write,
 	.device	 = con3270_device,
 	.unblank = con3270_unblank,
-	.setup	 = con3270_consetup,
 	.flags	 = CON_PRINTBUFFER,
 };
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 23/38] Improved kernel stack overflow checking.

[Martin Schwidefsky]

[-- Attachment #1: 126-stack-overflow.diff --]
[-- Type: text/plain, Size: 1234 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Recent cvs versions of gcc have support for an improved stack overflow
checking that calculates the size of the guard size for each function.
If the compiler accepts -mstack-size without -mstack-guard then the
new stack check is available. We always want to use the new stack
checker.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/Makefile |    2 ++
 1 files changed, 2 insertions(+)

Index: quilt-2.6/arch/s390/Makefile
===================================================================
--- quilt-2.6.orig/arch/s390/Makefile	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/Makefile	2007-04-27 16:04:58.000000000 +0200
@@ -67,8 +67,10 @@
 
 ifeq ($(call cc-option-yn,-mstack-size=8192 -mstack-guard=128),y)
 cflags-$(CONFIG_CHECK_STACK) += -mstack-size=$(STACK_SIZE)
+ifneq ($(call cc-option-yn,-mstack-size=8192),y)
 cflags-$(CONFIG_CHECK_STACK) += -mstack-guard=$(CONFIG_STACK_GUARD)
 endif
+endif
 
 ifeq ($(call cc-option-yn,-mwarn-dynamicstack),y)
 cflags-$(CONFIG_WARN_STACK) += -mwarn-dynamicstack

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 24/38] dasd: Add sysfs attribute status and generate uevents.

[Martin Schwidefsky]

[-- Attachment #1: 127-dasd-status.diff --]
[-- Type: text/plain, Size: 2976 bytes --]

From: Horst Hummel <horst.hummel@de.ibm.com>

This patch adds a sysfs-attribute 'status' to make the DASD device-status
accessible from user-space. In addition, the DASD driver generates an
uevent(CHANGE) for the ccw-device on each device-status change.
This enables user-space applications (e.g. udev) to do related processing.

Signed-off-by: Horst Hummel <horst.hummel@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/block/dasd.c        |    3 ++
 drivers/s390/block/dasd_devmap.c |   41 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 44 insertions(+)

Index: quilt-2.6/drivers/s390/block/dasd.c
===================================================================
--- quilt-2.6.orig/drivers/s390/block/dasd.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/block/dasd.c	2007-04-27 16:04:58.000000000 +0200
@@ -398,6 +398,9 @@
 
 	if (device->state == device->target)
 		wake_up(&dasd_init_waitq);
+
+	/* let user-space know that the device status changed */
+	kobject_uevent(&device->cdev->dev.kobj, KOBJ_CHANGE);
 }
 
 /*
Index: quilt-2.6/drivers/s390/block/dasd_devmap.c
===================================================================
--- quilt-2.6.orig/drivers/s390/block/dasd_devmap.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/block/dasd_devmap.c	2007-04-27 16:04:58.000000000 +0200
@@ -829,6 +829,46 @@
 static DEVICE_ATTR(discipline, 0444, dasd_discipline_show, NULL);
 
 static ssize_t
+dasd_device_status_show(struct device *dev, struct device_attribute *attr,
+		     char *buf)
+{
+	struct dasd_device *device;
+	ssize_t len;
+
+	device = dasd_device_from_cdev(to_ccwdev(dev));
+	if (!IS_ERR(device)) {
+		switch (device->state) {
+		case DASD_STATE_NEW:
+			len = snprintf(buf, PAGE_SIZE, "new\n");
+			break;
+		case DASD_STATE_KNOWN:
+			len = snprintf(buf, PAGE_SIZE, "detected\n");
+			break;
+		case DASD_STATE_BASIC:
+			len = snprintf(buf, PAGE_SIZE, "basic\n");
+			break;
+		case DASD_STATE_UNFMT:
+			len = snprintf(buf, PAGE_SIZE, "unformatted\n");
+			break;
+		case DASD_STATE_READY:
+			len = snprintf(buf, PAGE_SIZE, "ready\n");
+			break;
+		case DASD_STATE_ONLINE:
+			len = snprintf(buf, PAGE_SIZE, "online\n");
+			break;
+		default:
+			len = snprintf(buf, PAGE_SIZE, "no stat\n");
+			break;
+		}
+		dasd_put_device(device);
+	} else
+		len = snprintf(buf, PAGE_SIZE, "unknown\n");
+	return len;
+}
+
+static DEVICE_ATTR(status, 0444, dasd_device_status_show, NULL);
+
+static ssize_t
 dasd_alias_show(struct device *dev, struct device_attribute *attr, char *buf)
 {
 	struct dasd_devmap *devmap;
@@ -939,6 +979,7 @@
 static struct attribute * dasd_attrs[] = {
 	&dev_attr_readonly.attr,
 	&dev_attr_discipline.attr,
+	&dev_attr_status.attr,
 	&dev_attr_alias.attr,
 	&dev_attr_vendor.attr,
 	&dev_attr_uid.attr,

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 25/38] dasd: Add ipldev parameter.

[Martin Schwidefsky]

[-- Attachment #1: 128-dasd-ipldev.diff --]
[-- Type: text/plain, Size: 1683 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Specifying 'ipldev' in the dasd= kernel parameter will automatically
activate the boot device for use by the dasd driver.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/block/dasd_devmap.c |   17 +++++++++++++++++
 1 files changed, 17 insertions(+)

Index: quilt-2.6/drivers/s390/block/dasd_devmap.c
===================================================================
--- quilt-2.6.orig/drivers/s390/block/dasd_devmap.c	2007-04-27 16:04:58.000000000 +0200
+++ quilt-2.6/drivers/s390/block/dasd_devmap.c	2007-04-27 16:04:58.000000000 +0200
@@ -19,6 +19,7 @@
 
 #include <asm/debug.h>
 #include <asm/uaccess.h>
+#include <asm/ipl.h>
 
 /* This is ugly... */
 #define PRINTK_HEADER "dasd_devmap:"
@@ -133,6 +134,8 @@
 __setup ("dasd=", dasd_call_setup);
 #endif	/* #ifndef MODULE */
 
+#define	DASD_IPLDEV	"ipldev"
+
 /*
  * Read a device busid/devno from a string.
  */
@@ -141,6 +144,20 @@
 {
 	int val, old_style;
 
+	/* Interpret ipldev busid */
+	if (strncmp(DASD_IPLDEV, *str, strlen(DASD_IPLDEV)) == 0) {
+		if (ipl_info.type != IPL_TYPE_CCW) {
+			MESSAGE(KERN_ERR, "%s", "ipl device is not a ccw "
+				"device");
+			return -EINVAL;
+		}
+		*id0 = 0;
+		*id1 = ipl_info.data.ccw.dev_id.ssid;
+		*devno = ipl_info.data.ccw.dev_id.devno;
+		*str += strlen(DASD_IPLDEV);
+
+		return 0;
+	}
 	/* check for leading '0x' */
 	old_style = 0;
 	if ((*str)[0] == '0' && (*str)[1] == 'x') {

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 26/38] zfcpdump support.

[Martin Schwidefsky]

[-- Attachment #1: 129-zfcp-dump.diff --]
[-- Type: text/plain, Size: 55769 bytes --]

From: Michael Holzheu <holzheu@de.ibm.com>

s390 machines provide hardware support for creating Linux dumps on SCSI
disks. For creating a dump a special purpose dump Linux is used. The first
32 MB of memory are saved by the hardware before the dump Linux is
booted. Via an SCLP interface, the saved memory can be accessed from
Linux. This patch exports memory and registers of the crashed Linux to
userspace via a debugfs file. For more information refer to
Documentation/s390/zfcpdump.txt, which is included in this patch.

Signed-off-by: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 Documentation/s390/zfcpdump.txt |   87 +++++
 arch/s390/Kconfig               |    8 
 arch/s390/Makefile              |    3 
 arch/s390/defconfig             |    1 
 arch/s390/kernel/head64.S       |   72 ++++
 arch/s390/kernel/ipl.c          |  232 +++++++++-----
 arch/s390/kernel/setup.c        |   38 ++
 arch/s390/kernel/smp.c          |   62 +++
 drivers/s390/char/Makefile      |    3 
 drivers/s390/char/sclp.h        |    2 
 drivers/s390/char/sclp_sdias.c  |  255 +++++++++++++++
 drivers/s390/char/zcore.c       |  651 ++++++++++++++++++++++++++++++++++++++++
 include/asm-s390/ipl.h          |   32 +
 include/asm-s390/lowcore.h      |   46 ++
 include/asm-s390/sclp.h         |    2 
 include/asm-s390/setup.h        |    2 
 include/asm-s390/smp.h          |    1 
 17 files changed, 1411 insertions(+), 86 deletions(-)

Index: quilt-2.6/arch/s390/defconfig
===================================================================
--- quilt-2.6.orig/arch/s390/defconfig	2007-04-27 16:04:56.000000000 +0200
+++ quilt-2.6/arch/s390/defconfig	2007-04-27 16:04:59.000000000 +0200
@@ -167,6 +167,7 @@
 CONFIG_NO_IDLE_HZ_INIT=y
 CONFIG_S390_HYPFS_FS=y
 CONFIG_KEXEC=y
+# CONFIG_ZFCPDUMP is not set
 
 #
 # Networking
Index: quilt-2.6/arch/s390/Kconfig
===================================================================
--- quilt-2.6.orig/arch/s390/Kconfig	2007-04-27 16:04:56.000000000 +0200
+++ quilt-2.6/arch/s390/Kconfig	2007-04-27 16:04:59.000000000 +0200
@@ -519,6 +519,14 @@
 	  current kernel, and to start another kernel.  It is like a reboot
 	  but is independent of hardware/microcode support.
 
+config ZFCPDUMP
+	tristate "zfcpdump support"
+	select SMP
+	default n
+	help
+	  Select this option if you want to build an zfcpdump enabled kernel.
+	  Refer to "Documentation/s390/zfcpdump.txt" for more details on this.
+
 endmenu
 
 source "net/Kconfig"
Index: quilt-2.6/arch/s390/kernel/head64.S
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/head64.S	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/head64.S	2007-04-27 16:04:59.000000000 +0200
@@ -39,7 +39,69 @@
 	basr	%r13,0			# get base
 .LPG1:	sll	%r13,1			# remove high order bit
 	srl	%r13,1
-	lhi	%r1,1			# mode 1 = esame
+
+#ifdef CONFIG_ZFCPDUMP
+
+	# check if we have been ipled using zfcp dump:
+
+	tm	0xb9,0x01		# test if subchannel is enabled
+	jno	.nodump			# subchannel disabled
+	l	%r1,0xb8
+	la	%r5,.Lipl_schib-.LPG1(%r13)
+	stsch	0(%r5)			# get schib of subchannel
+	jne	.nodump			# schib not available
+	tm	5(%r5),0x01		# devno valid?
+	jno	.nodump
+	tm	4(%r5),0x80		# qdio capable device?
+	jno	.nodump
+	l	%r2,20(%r0)		# address of ipl parameter block
+	lhi	%r3,0
+	ic	%r3,0x148(%r2)		# get opt field
+	chi	%r3,0x20		# load with dump?
+	jne	.nodump
+
+	# store all prefix registers in case of load with dump:
+
+	la	%r7,0			# base register for 0 page
+	la	%r8,0			# first cpu
+	l	%r11,.Lpref_arr_ptr-.LPG1(%r13)	# address of prefix array
+	ahi	%r11,4			# skip boot cpu
+	lr	%r12,%r11
+	ahi	%r12,(CONFIG_NR_CPUS*4)	# end of prefix array
+	stap	.Lcurrent_cpu+2-.LPG1(%r13)	# store current cpu addr
+1:
+	cl	%r8,.Lcurrent_cpu-.LPG1(%r13)	# is ipl cpu ?
+	je	4f				# if yes get next cpu
+2:
+	lr	%r9,%r7
+	sigp	%r9,%r8,0x9		# stop & store status of cpu
+	brc	8,3f			# accepted
+	brc	4,4f			# status stored: next cpu
+	brc	2,2b			# busy: 	 try again
+	brc	1,4f			# not op:	 next cpu
+3:
+	mvc	0(4,%r11),264(%r7)	# copy prefix register to prefix array
+	ahi	%r11,4			# next element in prefix array
+	clr	%r11,%r12
+	je	5f			# no more space in prefix array
+4:
+	ahi	%r8,1				# next cpu (r8 += 1)
+	cl	%r8,.Llast_cpu-.LPG1(%r13)	# is last possible cpu ?
+	jl	1b				# jump if not last cpu
+5:
+	lhi	%r1,2			# mode 2 = esame (dump)
+	j	6f
+	.align 4
+.Lipl_schib:
+	.rept 13
+	.long 0
+	.endr
+.nodump:
+	lhi	%r1,1			# mode 1 = esame (normal ipl)
+6:
+#else
+	lhi	%r1,1			# mode 1 = esame (normal ipl)
+#endif /* CONFIG_ZFCPDUMP */
 	mvi	__LC_AR_MODE_ID,1	# set esame flag
 	slr	%r0,%r0 		# set cpuid to zero
 	sigp	%r1,%r0,0x12		# switch to esame mode
@@ -149,6 +211,14 @@
 .L4malign:.quad 0xffffffffffc00000
 .Lscan2g:.quad	0x80000000 + 0x20000 - 8	# 2GB + 128K - 8
 .Lnop:	.long	0x07000700
+#ifdef CONFIG_ZFCPDUMP
+.Lcurrent_cpu:
+	.long 0x0
+.Llast_cpu:
+	.long 0x0000ffff
+.Lpref_arr_ptr:
+	.long zfcpdump_prefix_array
+#endif /* CONFIG_ZFCPDUMP */
 .Lparmaddr:
 	.quad	PARMAREA
 	.align	64
Index: quilt-2.6/arch/s390/kernel/ipl.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/ipl.c	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/ipl.c	2007-04-27 16:04:59.000000000 +0200
@@ -29,36 +29,21 @@
 #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"
-
-/*
- * Must be in data section since the bss section
- * is not cleared when these are accessed.
- */
-static u16 ipl_devno __attribute__((__section__(".data"))) = 0;
-u32 ipl_flags __attribute__((__section__(".data"))) = 0;
+#define IPL_UNKNOWN_STR		"unknown"
+#define IPL_CCW_STR		"ccw"
+#define IPL_FCP_STR		"fcp"
+#define IPL_FCP_DUMP_STR	"fcp_dump"
+#define IPL_NSS_STR		"nss"
 
 static char *ipl_type_str(enum ipl_type type)
 {
 	switch (type) {
-	case IPL_TYPE_NONE:
-		return IPL_NONE_STR;
 	case IPL_TYPE_CCW:
 		return IPL_CCW_STR;
 	case IPL_TYPE_FCP:
 		return IPL_FCP_STR;
+	case IPL_TYPE_FCP_DUMP:
+		return IPL_FCP_DUMP_STR;
 	case IPL_TYPE_NSS:
 		return IPL_NSS_STR;
 	case IPL_TYPE_UNKNOWN:
@@ -67,15 +52,55 @@
 	}
 }
 
+enum dump_type {
+	DUMP_TYPE_NONE	= 1,
+	DUMP_TYPE_CCW	= 2,
+	DUMP_TYPE_FCP	= 4,
+};
+
+#define DUMP_NONE_STR	 "none"
+#define DUMP_CCW_STR	 "ccw"
+#define DUMP_FCP_STR	 "fcp"
+
+static char *dump_type_str(enum dump_type type)
+{
+	switch (type) {
+	case DUMP_TYPE_NONE:
+		return DUMP_NONE_STR;
+	case DUMP_TYPE_CCW:
+		return DUMP_CCW_STR;
+	case DUMP_TYPE_FCP:
+		return DUMP_FCP_STR;
+	default:
+		return NULL;
+	}
+}
+
+/*
+ * Must be in data section since the bss section
+ * is not cleared when these are accessed.
+ */
+static u16 ipl_devno __attribute__((__section__(".data"))) = 0;
+u32 ipl_flags __attribute__((__section__(".data"))) = 0;
+
 enum ipl_method {
-	IPL_METHOD_NONE,
-	IPL_METHOD_CCW_CIO,
-	IPL_METHOD_CCW_DIAG,
-	IPL_METHOD_CCW_VM,
-	IPL_METHOD_FCP_RO_DIAG,
-	IPL_METHOD_FCP_RW_DIAG,
-	IPL_METHOD_FCP_RO_VM,
-	IPL_METHOD_NSS,
+	REIPL_METHOD_CCW_CIO,
+	REIPL_METHOD_CCW_DIAG,
+	REIPL_METHOD_CCW_VM,
+	REIPL_METHOD_FCP_RO_DIAG,
+	REIPL_METHOD_FCP_RW_DIAG,
+	REIPL_METHOD_FCP_RO_VM,
+	REIPL_METHOD_FCP_DUMP,
+	REIPL_METHOD_NSS,
+	REIPL_METHOD_DEFAULT,
+};
+
+enum dump_method {
+	DUMP_METHOD_NONE,
+	DUMP_METHOD_CCW_CIO,
+	DUMP_METHOD_CCW_DIAG,
+	DUMP_METHOD_CCW_VM,
+	DUMP_METHOD_FCP_DIAG,
 };
 
 enum shutdown_action {
@@ -107,15 +132,15 @@
 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 enum ipl_method reipl_method = REIPL_METHOD_DEFAULT;
 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;
+static int dump_capabilities = DUMP_TYPE_NONE;
+static enum dump_type dump_type = DUMP_TYPE_NONE;
+static enum dump_method dump_method = DUMP_METHOD_NONE;
 static struct ipl_parameter_block *dump_block_fcp;
 static struct ipl_parameter_block *dump_block_ccw;
 
@@ -134,6 +159,7 @@
 		: "d" (subcode) : "cc", "memory");
 	return _rc;
 }
+EXPORT_SYMBOL_GPL(diag308);
 
 /* SYSFS */
 
@@ -197,7 +223,7 @@
  * ipl section
  */
 
-static enum ipl_type ipl_get_type(void)
+static __init enum ipl_type get_ipl_type(void)
 {
 	struct ipl_parameter_block *ipl = IPL_PARMBLOCK_START;
 
@@ -211,12 +237,44 @@
 		return IPL_TYPE_UNKNOWN;
 	if (ipl->hdr.pbt != DIAG308_IPL_TYPE_FCP)
 		return IPL_TYPE_UNKNOWN;
+	if (ipl->ipl_info.fcp.opt == DIAG308_IPL_OPT_DUMP)
+		return IPL_TYPE_FCP_DUMP;
 	return IPL_TYPE_FCP;
 }
 
+void __init setup_ipl_info(void)
+{
+	ipl_info.type = get_ipl_type();
+	switch (ipl_info.type) {
+	case IPL_TYPE_CCW:
+		ipl_info.data.ccw.dev_id.devno = ipl_devno;
+		ipl_info.data.ccw.dev_id.ssid = 0;
+		break;
+	case IPL_TYPE_FCP:
+	case IPL_TYPE_FCP_DUMP:
+		ipl_info.data.fcp.dev_id.devno =
+			IPL_PARMBLOCK_START->ipl_info.fcp.devno;
+		ipl_info.data.fcp.dev_id.ssid = 0;
+		ipl_info.data.fcp.wwpn = IPL_PARMBLOCK_START->ipl_info.fcp.wwpn;
+		ipl_info.data.fcp.lun = IPL_PARMBLOCK_START->ipl_info.fcp.lun;
+		break;
+	case IPL_TYPE_NSS:
+		strncpy(ipl_info.data.nss.name, kernel_nss_name,
+			sizeof(ipl_info.data.nss.name));
+		break;
+	case IPL_TYPE_UNKNOWN:
+	default:
+		/* We have no info to copy */
+		break;
+	}
+}
+
+struct ipl_info ipl_info;
+EXPORT_SYMBOL_GPL(ipl_info);
+
 static ssize_t ipl_type_show(struct subsystem *subsys, char *page)
 {
-	return sprintf(page, "%s\n", ipl_type_str(ipl_get_type()));
+	return sprintf(page, "%s\n", ipl_type_str(ipl_info.type));
 }
 
 static struct subsys_attribute sys_ipl_type_attr = __ATTR_RO(ipl_type);
@@ -225,10 +283,11 @@
 {
 	struct ipl_parameter_block *ipl = IPL_PARMBLOCK_START;
 
-	switch (ipl_get_type()) {
+	switch (ipl_info.type) {
 	case IPL_TYPE_CCW:
 		return sprintf(page, "0.0.%04x\n", ipl_devno);
 	case IPL_TYPE_FCP:
+	case IPL_TYPE_FCP_DUMP:
 		return sprintf(page, "0.0.%04x\n", ipl->ipl_info.fcp.devno);
 	default:
 		return 0;
@@ -485,23 +544,29 @@
 	switch(type) {
 	case IPL_TYPE_CCW:
 		if (MACHINE_IS_VM)
-			reipl_method = IPL_METHOD_CCW_VM;
+			reipl_method = REIPL_METHOD_CCW_VM;
 		else
-			reipl_method = IPL_METHOD_CCW_CIO;
+			reipl_method = REIPL_METHOD_CCW_CIO;
 		break;
 	case IPL_TYPE_FCP:
 		if (diag308_set_works)
-			reipl_method = IPL_METHOD_FCP_RW_DIAG;
+			reipl_method = REIPL_METHOD_FCP_RW_DIAG;
 		else if (MACHINE_IS_VM)
-			reipl_method = IPL_METHOD_FCP_RO_VM;
+			reipl_method = REIPL_METHOD_FCP_RO_VM;
 		else
-			reipl_method = IPL_METHOD_FCP_RO_DIAG;
+			reipl_method = REIPL_METHOD_FCP_RO_DIAG;
+		break;
+	case IPL_TYPE_FCP_DUMP:
+		reipl_method = REIPL_METHOD_FCP_DUMP;
 		break;
 	case IPL_TYPE_NSS:
-		reipl_method = IPL_METHOD_NSS;
+		reipl_method = REIPL_METHOD_NSS;
+		break;
+	case IPL_TYPE_UNKNOWN:
+		reipl_method = REIPL_METHOD_DEFAULT;
 		break;
 	default:
-		reipl_method = IPL_METHOD_NONE;
+		BUG();
 	}
 	reipl_type = type;
 	return 0;
@@ -579,22 +644,22 @@
 
 /* dump type */
 
-static int dump_set_type(enum ipl_type type)
+static int dump_set_type(enum dump_type type)
 {
 	if (!(dump_capabilities & type))
 		return -EINVAL;
 	switch(type) {
-	case IPL_TYPE_CCW:
+	case DUMP_TYPE_CCW:
 		if (MACHINE_IS_VM)
-			dump_method = IPL_METHOD_CCW_VM;
+			dump_method = DUMP_METHOD_CCW_VM;
 		else
-			dump_method = IPL_METHOD_CCW_CIO;
+			dump_method = DUMP_METHOD_CCW_CIO;
 		break;
-	case IPL_TYPE_FCP:
-		dump_method = IPL_METHOD_FCP_RW_DIAG;
+	case DUMP_TYPE_FCP:
+		dump_method = DUMP_METHOD_FCP_DIAG;
 		break;
 	default:
-		dump_method = IPL_METHOD_NONE;
+		dump_method = DUMP_METHOD_NONE;
 	}
 	dump_type = type;
 	return 0;
@@ -602,7 +667,7 @@
 
 static ssize_t dump_type_show(struct subsystem *subsys, char *page)
 {
-	return sprintf(page, "%s\n", ipl_type_str(dump_type));
+	return sprintf(page, "%s\n", dump_type_str(dump_type));
 }
 
 static ssize_t dump_type_store(struct subsystem *subsys, const char *buf,
@@ -610,12 +675,12 @@
 {
 	int rc = -EINVAL;
 
-	if (strncmp(buf, IPL_NONE_STR, strlen(IPL_NONE_STR)) == 0)
-		rc = dump_set_type(IPL_TYPE_NONE);
-	else if (strncmp(buf, IPL_CCW_STR, strlen(IPL_CCW_STR)) == 0)
-		rc = dump_set_type(IPL_TYPE_CCW);
-	else if (strncmp(buf, IPL_FCP_STR, strlen(IPL_FCP_STR)) == 0)
-		rc = dump_set_type(IPL_TYPE_FCP);
+	if (strncmp(buf, DUMP_NONE_STR, strlen(DUMP_NONE_STR)) == 0)
+		rc = dump_set_type(DUMP_TYPE_NONE);
+	else if (strncmp(buf, DUMP_CCW_STR, strlen(DUMP_CCW_STR)) == 0)
+		rc = dump_set_type(DUMP_TYPE_CCW);
+	else if (strncmp(buf, DUMP_FCP_STR, strlen(DUMP_FCP_STR)) == 0)
+		rc = dump_set_type(DUMP_TYPE_FCP);
 	return (rc != 0) ? rc : len;
 }
 
@@ -664,14 +729,14 @@
 	char loadparm[LOADPARM_LEN + 1];
 
 	switch (reipl_method) {
-	case IPL_METHOD_CCW_CIO:
+	case REIPL_METHOD_CCW_CIO:
 		devid.devno = reipl_block_ccw->ipl_info.ccw.devno;
-		if (ipl_get_type() == IPL_TYPE_CCW && devid.devno == ipl_devno)
+		if (ipl_info.type == IPL_TYPE_CCW && devid.devno == ipl_devno)
 			diag308(DIAG308_IPL, NULL);
 		devid.ssid  = 0;
 		reipl_ccw_dev(&devid);
 		break;
-	case IPL_METHOD_CCW_VM:
+	case REIPL_METHOD_CCW_VM:
 		reipl_get_ascii_loadparm(loadparm);
 		if (strlen(loadparm) == 0)
 			sprintf(buf, "IPL %X",
@@ -681,30 +746,32 @@
 				reipl_block_ccw->ipl_info.ccw.devno, loadparm);
 		__cpcmd(buf, NULL, 0, NULL);
 		break;
-	case IPL_METHOD_CCW_DIAG:
+	case REIPL_METHOD_CCW_DIAG:
 		diag308(DIAG308_SET, reipl_block_ccw);
 		diag308(DIAG308_IPL, NULL);
 		break;
-	case IPL_METHOD_FCP_RW_DIAG:
+	case REIPL_METHOD_FCP_RW_DIAG:
 		diag308(DIAG308_SET, reipl_block_fcp);
 		diag308(DIAG308_IPL, NULL);
 		break;
-	case IPL_METHOD_FCP_RO_DIAG:
+	case REIPL_METHOD_FCP_RO_DIAG:
 		diag308(DIAG308_IPL, NULL);
 		break;
-	case IPL_METHOD_FCP_RO_VM:
+	case REIPL_METHOD_FCP_RO_VM:
 		__cpcmd("IPL", NULL, 0, NULL);
 		break;
-	case IPL_METHOD_NSS:
+	case REIPL_METHOD_NSS:
 		sprintf(buf, "IPL %s", reipl_nss_name);
 		__cpcmd(buf, NULL, 0, NULL);
 		break;
-	case IPL_METHOD_NONE:
-	default:
+	case REIPL_METHOD_DEFAULT:
 		if (MACHINE_IS_VM)
 			__cpcmd("IPL", NULL, 0, NULL);
 		diag308(DIAG308_IPL, NULL);
 		break;
+	case REIPL_METHOD_FCP_DUMP:
+	default:
+		break;
 	}
 	signal_processor(smp_processor_id(), sigp_stop_and_store_status);
 }
@@ -715,28 +782,28 @@
 	static char buf[100];
 
 	switch (dump_method) {
-	case IPL_METHOD_CCW_CIO:
+	case DUMP_METHOD_CCW_CIO:
 		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:
+	case DUMP_METHOD_CCW_VM:
 		smp_send_stop();
 		sprintf(buf, "STORE STATUS");
 		__cpcmd(buf, NULL, 0, NULL);
 		sprintf(buf, "IPL %X", dump_block_ccw->ipl_info.ccw.devno);
 		__cpcmd(buf, NULL, 0, NULL);
 		break;
-	case IPL_METHOD_CCW_DIAG:
+	case DUMP_METHOD_CCW_DIAG:
 		diag308(DIAG308_SET, dump_block_ccw);
 		diag308(DIAG308_DUMP, NULL);
 		break;
-	case IPL_METHOD_FCP_RW_DIAG:
+	case DUMP_METHOD_FCP_DIAG:
 		diag308(DIAG308_SET, dump_block_fcp);
 		diag308(DIAG308_DUMP, NULL);
 		break;
-	case IPL_METHOD_NONE:
+	case DUMP_METHOD_NONE:
 	default:
 		return;
 	}
@@ -777,12 +844,13 @@
 	rc = firmware_register(&ipl_subsys);
 	if (rc)
 		return rc;
-	switch (ipl_get_type()) {
+	switch (ipl_info.type) {
 	case IPL_TYPE_CCW:
 		rc = sysfs_create_group(&ipl_subsys.kset.kobj,
 					&ipl_ccw_attr_group);
 		break;
 	case IPL_TYPE_FCP:
+	case IPL_TYPE_FCP_DUMP:
 		rc = ipl_register_fcp_files();
 		break;
 	case IPL_TYPE_NSS:
@@ -852,7 +920,7 @@
 	/* 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)
+	if (ipl_info.type == IPL_TYPE_CCW)
 		reipl_block_ccw->ipl_info.ccw.devno = ipl_devno;
 	reipl_capabilities |= IPL_TYPE_CCW;
 	return 0;
@@ -862,9 +930,9 @@
 {
 	int rc;
 
-	if ((!diag308_set_works) && (ipl_get_type() != IPL_TYPE_FCP))
+	if ((!diag308_set_works) && (ipl_info.type != IPL_TYPE_FCP))
 		return 0;
-	if ((!diag308_set_works) && (ipl_get_type() == IPL_TYPE_FCP))
+	if ((!diag308_set_works) && (ipl_info.type == IPL_TYPE_FCP))
 		make_attrs_ro(reipl_fcp_attrs);
 
 	reipl_block_fcp = (void *) get_zeroed_page(GFP_KERNEL);
@@ -875,7 +943,7 @@
 		free_page((unsigned long)reipl_block_fcp);
 		return rc;
 	}
-	if (ipl_get_type() == IPL_TYPE_FCP) {
+	if (ipl_info.type == IPL_TYPE_FCP) {
 		memcpy(reipl_block_fcp, IPL_PARMBLOCK_START, PAGE_SIZE);
 	} else {
 		reipl_block_fcp->hdr.len = IPL_PARM_BLK_FCP_LEN;
@@ -909,7 +977,7 @@
 	rc = reipl_nss_init();
 	if (rc)
 		return rc;
-	rc = reipl_set_type(ipl_get_type());
+	rc = reipl_set_type(ipl_info.type);
 	if (rc)
 		return rc;
 	return 0;
@@ -931,7 +999,7 @@
 	dump_block_ccw->hdr.version = IPL_PARM_BLOCK_VERSION;
 	dump_block_ccw->hdr.blk0_len = IPL_PARM_BLK0_CCW_LEN;
 	dump_block_ccw->hdr.pbt = DIAG308_IPL_TYPE_CCW;
-	dump_capabilities |= IPL_TYPE_CCW;
+	dump_capabilities |= DUMP_TYPE_CCW;
 	return 0;
 }
 
@@ -956,7 +1024,7 @@
 	dump_block_fcp->hdr.blk0_len = IPL_PARM_BLK0_FCP_LEN;
 	dump_block_fcp->hdr.pbt = DIAG308_IPL_TYPE_FCP;
 	dump_block_fcp->ipl_info.fcp.opt = DIAG308_IPL_OPT_DUMP;
-	dump_capabilities |= IPL_TYPE_FCP;
+	dump_capabilities |= DUMP_TYPE_FCP;
 	return 0;
 }
 
@@ -995,7 +1063,7 @@
 	rc = dump_fcp_init();
 	if (rc)
 		return rc;
-	dump_set_type(IPL_TYPE_NONE);
+	dump_set_type(DUMP_TYPE_NONE);
 	return 0;
 }
 
Index: quilt-2.6/arch/s390/kernel/setup.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/setup.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/setup.c	2007-04-27 16:04:59.000000000 +0200
@@ -285,6 +285,26 @@
 	}
 }
 
+#if defined(CONFIG_ZFCPDUMP) || defined(CONFIG_ZFCPDUMP_MODULE)
+static void __init setup_zfcpdump(unsigned int console_devno)
+{
+	static char str[64];
+
+	if (ipl_info.type != IPL_TYPE_FCP_DUMP)
+		return;
+	if (console_devno != -1)
+		sprintf(str, "cio_ignore=all,!0.0.%04x,!0.0.%04x",
+			ipl_info.data.fcp.dev_id.devno, console_devno);
+	else
+		sprintf(str, "cio_ignore=all,!0.0.%04x",
+			ipl_info.data.fcp.dev_id.devno);
+	strcat(COMMAND_LINE, str);
+	console_loglevel = 2;
+}
+#else
+static inline void setup_zfcpdump(unsigned int console_devno) {}
+#endif /* CONFIG_ZFCPDUMP */
+
 #ifdef CONFIG_SMP
 void (*_machine_restart)(char *command) = machine_restart_smp;
 void (*_machine_halt)(void) = machine_halt_smp;
@@ -586,13 +606,20 @@
 	}
 }
 
+unsigned long real_memory_size;
+EXPORT_SYMBOL_GPL(real_memory_size);
+
 static void __init setup_memory_end(void)
 {
-	unsigned long real_size, memory_size;
+	unsigned long memory_size;
 	unsigned long max_mem, max_phys;
 	int i;
 
-	memory_size = real_size = 0;
+#if defined(CONFIG_ZFCPDUMP) || defined(CONFIG_ZFCPDUMP_MODULE)
+	if (ipl_info.type == IPL_TYPE_FCP_DUMP)
+		memory_end = ZFCPDUMP_HSA_SIZE;
+#endif
+	memory_size = 0;
 	max_phys = VMALLOC_END_INIT - VMALLOC_MIN_SIZE;
 	memory_end &= PAGE_MASK;
 
@@ -601,7 +628,8 @@
 	for (i = 0; i < MEMORY_CHUNKS; i++) {
 		struct mem_chunk *chunk = &memory_chunk[i];
 
-		real_size = max(real_size, chunk->addr + chunk->size);
+		real_memory_size = max(real_memory_size,
+				       chunk->addr + chunk->size);
 		if (chunk->addr >= max_mem) {
 			memset(chunk, 0, sizeof(*chunk));
 			continue;
@@ -765,6 +793,7 @@
 
 	parse_early_param();
 
+	setup_ipl_info();
 	setup_memory_end();
 	setup_addressing_mode();
 	setup_memory();
@@ -782,6 +811,9 @@
 
         /* Setup default console */
 	conmode_default();
+
+	/* Setup zfcpdump support */
+	setup_zfcpdump(console_devno);
 }
 
 void print_cpu_info(struct cpuinfo_S390 *cpuinfo)
Index: quilt-2.6/arch/s390/kernel/smp.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/smp.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/smp.c	2007-04-27 16:04:59.000000000 +0200
@@ -31,6 +31,7 @@
 #include <linux/interrupt.h>
 #include <linux/cpu.h>
 #include <linux/timex.h>
+#include <linux/bootmem.h>
 #include <asm/ipl.h>
 #include <asm/setup.h>
 #include <asm/sigp.h>
@@ -40,6 +41,7 @@
 #include <asm/cpcmd.h>
 #include <asm/tlbflush.h>
 #include <asm/timer.h>
+#include <asm/lowcore.h>
 
 extern volatile int __cpu_logical_map[];
 
@@ -395,6 +397,65 @@
 	on_each_cpu(smp_ctl_bit_callback, &parms, 0, 1);
 }
 
+#if defined(CONFIG_ZFCPDUMP) || defined(CONFIG_ZFCPDUMP_MODULE)
+
+/*
+ * zfcpdump_prefix_array holds prefix registers for the following scenario:
+ * 64 bit zfcpdump kernel and 31 bit kernel which is to be dumped. We have to
+ * save its prefix registers, since they get lost, when switching from 31 bit
+ * to 64 bit.
+ */
+unsigned int zfcpdump_prefix_array[NR_CPUS + 1] \
+	__attribute__((__section__(".data")));
+
+static void __init smp_get_save_areas(void)
+{
+	unsigned int cpu, cpu_num, rc;
+	__u16 boot_cpu_addr;
+
+	if (ipl_info.type != IPL_TYPE_FCP_DUMP)
+		return;
+	boot_cpu_addr = S390_lowcore.cpu_data.cpu_addr;
+	cpu_num = 1;
+	for (cpu = 0; cpu <= 65535; cpu++) {
+		if ((u16) cpu == boot_cpu_addr)
+			continue;
+		__cpu_logical_map[1] = (__u16) cpu;
+		if (signal_processor(1, sigp_sense) == sigp_not_operational)
+			continue;
+		if (cpu_num >= NR_CPUS) {
+			printk("WARNING: Registers for cpu %i are not "
+			       "saved, since dump kernel was compiled with"
+			       "NR_CPUS=%i!\n", cpu_num, NR_CPUS);
+			continue;
+		}
+		zfcpdump_save_areas[cpu_num] =
+			alloc_bootmem(sizeof(union save_area));
+		while (1) {
+			rc = signal_processor(1, sigp_stop_and_store_status);
+			if (rc != sigp_busy)
+				break;
+			cpu_relax();
+		}
+		memcpy(zfcpdump_save_areas[cpu_num],
+		       (void *)(unsigned long) store_prefix() +
+		       SAVE_AREA_BASE, SAVE_AREA_SIZE);
+#ifdef __s390x__
+		/* copy original prefix register */
+		zfcpdump_save_areas[cpu_num]->s390x.pref_reg =
+			zfcpdump_prefix_array[cpu_num];
+#endif
+		cpu_num++;
+	}
+}
+
+union save_area *zfcpdump_save_areas[NR_CPUS + 1];
+EXPORT_SYMBOL_GPL(zfcpdump_save_areas);
+
+#else
+#define smp_get_save_areas() do { } while (0)
+#endif
+
 /*
  * Lets check how many CPUs we have.
  */
@@ -589,6 +650,7 @@
 {
 	unsigned int phy_cpus, pos_cpus, cpu;
 
+	smp_get_save_areas();
 	phy_cpus = smp_count_cpus();
 	pos_cpus = min(phy_cpus + additional_cpus, (unsigned int) NR_CPUS);
 
Index: quilt-2.6/arch/s390/Makefile
===================================================================
--- quilt-2.6.orig/arch/s390/Makefile	2007-04-27 16:04:58.000000000 +0200
+++ quilt-2.6/arch/s390/Makefile	2007-04-27 16:04:59.000000000 +0200
@@ -105,6 +105,9 @@
 image: vmlinux
 	$(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
 
+zfcpdump:
+	$(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
+
 archclean:
 	$(Q)$(MAKE) $(clean)=$(boot)
 
Index: quilt-2.6/Documentation/s390/zfcpdump.txt
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/Documentation/s390/zfcpdump.txt	2007-04-27 16:04:59.000000000 +0200
@@ -0,0 +1,87 @@
+s390 SCSI dump tool (zfcpdump)
+
+System z machines (z900 or higher) provide hardware support for creating system
+dumps on SCSI disks. The dump process is initiated by booting a dump tool, which
+has to create a dump of the current (probably crashed) Linux image. In order to
+not overwrite memory of the crashed Linux with data of the dump tool, the
+hardware saves some memory plus the register sets of the boot cpu before the
+dump tool is loaded. There exists an SCLP hardware interface to obtain the saved
+memory afterwards. Currently 32 MB are saved.
+
+This zfcpdump implementation consists of a Linux dump kernel together with
+a userspace dump tool, which are loaded together into the saved memory region
+below 32 MB. zfcpdump is installed on a SCSI disk using zipl (as contained in
+the s390-tools package) to make the device bootable. The operator of a Linux
+system can then trigger a SCSI dump by booting the SCSI disk, where zfcpdump
+resides on.
+
+The kernel part of zfcpdump is implemented as a debugfs file under "zcore/mem",
+which exports memory and registers of the crashed Linux in an s390
+standalone dump format. It can be used in the same way as e.g. /dev/mem. The
+dump format defines a 4K header followed by plain uncompressed memory. The
+register sets are stored in the prefix pages of the respective cpus. To build a
+dump enabled kernel with the zcore driver, the kernel config option
+CONFIG_ZFCPDUMP has to be set. When reading from "zcore/mem", the part of
+memory, which has been saved by hardware is read by the driver via the SCLP
+hardware interface. The second part is just copied from the non overwritten real
+memory.
+
+The userspace application of zfcpdump can reside e.g. in an intitramfs or an
+initrd. It reads from zcore/mem and writes the system dump to a file on a
+SCSI disk.
+
+To build a zfcpdump kernel use the following settings in your kernel
+configuration:
+ * CONFIG_ZFCPDUMP=y
+ * Enable ZFCP driver
+ * Enable SCSI driver
+ * Enable ext2 and ext3 filesystems
+ * Disable as many features as possible to keep the kernel small.
+   E.g. network support is not needed at all.
+
+To use the zfcpdump userspace application in an initramfs you have to do the
+following:
+
+ * Copy the zfcpdump executable somewhere into your Linux tree.
+   E.g. to "arch/s390/boot/zfcpdump. If you do not want to include
+   shared libraries, compile the tool with the "-static" gcc option.
+ * If you want to include e2fsck, add it to your source tree, too. The zfcpdump
+   application attempts to start /sbin/e2fsck from the ramdisk.
+ * Use an initramfs config file like the following:
+
+   dir /dev 755 0 0
+   nod /dev/console 644 0 0 c 5 1
+   nod /dev/null 644 0 0 c 1 3
+   nod /dev/sda1 644 0 0 b 8 1
+   nod /dev/sda2 644 0 0 b 8 2
+   nod /dev/sda3 644 0 0 b 8 3
+   nod /dev/sda4 644 0 0 b 8 4
+   nod /dev/sda5 644 0 0 b 8 5
+   nod /dev/sda6 644 0 0 b 8 6
+   nod /dev/sda7 644 0 0 b 8 7
+   nod /dev/sda8 644 0 0 b 8 8
+   nod /dev/sda9 644 0 0 b 8 9
+   nod /dev/sda10 644 0 0 b 8 10
+   nod /dev/sda11 644 0 0 b 8 11
+   nod /dev/sda12 644 0 0 b 8 12
+   nod /dev/sda13 644 0 0 b 8 13
+   nod /dev/sda14 644 0 0 b 8 14
+   nod /dev/sda15 644 0 0 b 8 15
+   file /init arch/s390/boot/zfcpdump 755 0 0
+   file /sbin/e2fsck arch/s390/boot/e2fsck 755 0 0
+   dir /proc 755 0 0
+   dir /sys 755 0 0
+   dir /mnt 755 0 0
+   dir /sbin 755 0 0
+
+ * Issue "make image" to build the zfcpdump image with initramfs.
+
+In a Linux distribution the zfcpdump enabled kernel image must be copied to
+/usr/share/zfcpdump/zfcpdump.image, where the s390 zipl tool is looking for the
+dump kernel when preparing a SCSI dump disk.
+
+If you use a ramdisk copy it to "/usr/share/zfcpdump/zfcpdump.rd".
+
+For more information on how to use zfcpdump refer to the s390 'Using the Dump
+Tools book', which is available from
+http://www.ibm.com/developerworks/linux/linux390.
Index: quilt-2.6/drivers/s390/char/Makefile
===================================================================
--- quilt-2.6.orig/drivers/s390/char/Makefile	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/drivers/s390/char/Makefile	2007-04-27 16:04:59.000000000 +0200
@@ -29,3 +29,6 @@
 obj-$(CONFIG_S390_TAPE_3590) += tape_3590.o
 obj-$(CONFIG_MONREADER) += monreader.o
 obj-$(CONFIG_MONWRITER) += monwriter.o
+
+zcore_mod-objs := sclp_sdias.o zcore.o
+obj-$(CONFIG_ZFCPDUMP) += zcore_mod.o
Index: quilt-2.6/drivers/s390/char/sclp.h
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp.h	2007-04-27 16:04:59.000000000 +0200
@@ -27,6 +27,7 @@
 #define EvTyp_CntlProgIdent	0x0B
 #define EvTyp_SigQuiesce	0x1D
 #define EvTyp_VT220Msg		0x1A
+#define EvTyp_SDIAS		0x1C
 
 #define EvTyp_OpCmd_Mask	0x80000000
 #define EvTyp_Msg_Mask		0x40000000
@@ -36,6 +37,7 @@
 #define EvTyp_CtlProgIdent_Mask	0x00200000
 #define EvTyp_SigQuiesce_Mask	0x00000008
 #define EvTyp_VT220Msg_Mask	0x00000040
+#define EvTyp_SDIAS_Mask	0x00000010
 
 #define GnrlMsgFlgs_DOM		0x8000
 #define GnrlMsgFlgs_SndAlrm	0x4000
Index: quilt-2.6/drivers/s390/char/sclp_sdias.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/char/sclp_sdias.c	2007-04-27 16:04:59.000000000 +0200
@@ -0,0 +1,255 @@
+/*
+ * Sclp "store data in absolut storage"
+ *
+ * Copyright IBM Corp. 2003,2007
+ * Author(s): Michael Holzheu
+ */
+
+#include <linux/sched.h>
+#include <asm/sclp.h>
+#include <asm/debug.h>
+#include <asm/ipl.h>
+#include "sclp.h"
+#include "sclp_rw.h"
+
+#define TRACE(x...) debug_sprintf_event(sdias_dbf, 1, x)
+#define ERROR_MSG(x...) printk ( KERN_ALERT "SDIAS: " x )
+
+#define SDIAS_RETRIES 300
+#define SDIAS_SLEEP_TICKS 50
+
+#define EQ_STORE_DATA	0x0
+#define EQ_SIZE		0x1
+#define DI_FCP_DUMP	0x0
+#define ASA_SIZE_32	0x0
+#define ASA_SIZE_64	0x1
+#define EVSTATE_ALL_STORED	0x0
+#define EVSTATE_NO_DATA		0x3
+#define EVSTATE_PART_STORED	0x10
+
+static struct debug_info *sdias_dbf;
+
+static struct sclp_register sclp_sdias_register = {
+	.send_mask = EvTyp_SDIAS_Mask,
+};
+
+struct sdias_evbuf {
+	struct	evbuf_header hdr;
+	u8	event_qual;
+	u8	data_id;
+	u64	reserved2;
+	u32	event_id;
+	u16	reserved3;
+	u8	asa_size;
+	u8	event_status;
+	u32	reserved4;
+	u32	blk_cnt;
+	u64	asa;
+	u32	reserved5;
+	u32	fbn;
+	u32	reserved6;
+	u32	lbn;
+	u16	reserved7;
+	u16	dbs;
+} __attribute__((packed));
+
+struct sdias_sccb {
+	struct sccb_header  hdr;
+	struct sdias_evbuf  evbuf;
+} __attribute__((packed));
+
+static struct sdias_sccb sccb __attribute__((aligned(4096)));
+
+static int sclp_req_done;
+static wait_queue_head_t sdias_wq;
+static DEFINE_MUTEX(sdias_mutex);
+
+static void sdias_callback(struct sclp_req *request, void *data)
+{
+	struct sdias_sccb *sccb;
+
+	sccb = (struct sdias_sccb *) request->sccb;
+	sclp_req_done = 1;
+	wake_up(&sdias_wq); /* Inform caller, that request is complete */
+	TRACE("callback done\n");
+}
+
+static int sdias_sclp_send(struct sclp_req *req)
+{
+	int retries;
+	int rc;
+
+	for (retries = SDIAS_RETRIES; retries; retries--) {
+		sclp_req_done = 0;
+		TRACE("add request\n");
+		rc = sclp_add_request(req);
+		if (rc) {
+			/* not initiated, wait some time and retry */
+			set_current_state(TASK_INTERRUPTIBLE);
+			TRACE("add request failed: rc = %i\n",rc);
+			schedule_timeout(SDIAS_SLEEP_TICKS);
+			continue;
+		}
+		/* initiated, wait for completion of service call */
+		wait_event(sdias_wq, (sclp_req_done == 1));
+		if (req->status == SCLP_REQ_FAILED) {
+			TRACE("sclp request failed\n");
+			rc = -EIO;
+			continue;
+		}
+		TRACE("request done\n");
+		break;
+	}
+	return rc;
+}
+
+/*
+ * Get number of blocks (4K) available in the HSA
+ */
+int sclp_sdias_blk_count(void)
+{
+	struct sclp_req request;
+	int rc;
+
+	mutex_lock(&sdias_mutex);
+
+	memset(&sccb, 0, sizeof(sccb));
+	memset(&request, 0, sizeof(request));
+
+	sccb.hdr.length = sizeof(sccb);
+	sccb.evbuf.hdr.length = sizeof(struct sdias_evbuf);
+	sccb.evbuf.hdr.type = EvTyp_SDIAS;
+	sccb.evbuf.event_qual = EQ_SIZE;
+	sccb.evbuf.data_id = DI_FCP_DUMP;
+	sccb.evbuf.event_id = 4712;
+	sccb.evbuf.dbs = 1;
+
+	request.sccb = &sccb;
+	request.command = SCLP_CMDW_WRITE_EVENT_DATA;
+	request.status = SCLP_REQ_FILLED;
+	request.callback = sdias_callback;
+
+	rc = sdias_sclp_send(&request);
+	if (rc) {
+		ERROR_MSG("sclp_send failed for get_nr_blocks\n");
+		goto out;
+	}
+	if (sccb.hdr.response_code != 0x0020) {
+		TRACE("send failed: %x\n", sccb.hdr.response_code);
+		rc = -EIO;
+		goto out;
+	}
+
+	switch (sccb.evbuf.event_status) {
+		case 0:
+			rc = sccb.evbuf.blk_cnt;
+			break;
+		default:
+			ERROR_MSG("SCLP error: %x\n", sccb.evbuf.event_status);
+			rc = -EIO;
+			goto out;
+	}
+	TRACE("%i blocks\n", rc);
+out:
+	mutex_unlock(&sdias_mutex);
+	return rc;
+}
+
+/*
+ * Copy from HSA to absolute storage (not reentrant):
+ *
+ * @dest     : Address of buffer where data should be copied
+ * @start_blk: Start Block (beginning with 1)
+ * @nr_blks  : Number of 4K blocks to copy
+ *
+ * Return Value: 0 : Requested 'number' of blocks of data copied
+ *		 <0: ERROR - negative event status
+ */
+int sclp_sdias_copy(void *dest, int start_blk, int nr_blks)
+{
+	struct sclp_req request;
+	int rc;
+
+	mutex_lock(&sdias_mutex);
+
+	memset(&sccb, 0, sizeof(sccb));
+	memset(&request, 0, sizeof(request));
+
+	sccb.hdr.length = sizeof(sccb);
+	sccb.evbuf.hdr.length = sizeof(struct sdias_evbuf);
+	sccb.evbuf.hdr.type = EvTyp_SDIAS;
+	sccb.evbuf.hdr.flags = 0;
+	sccb.evbuf.event_qual = EQ_STORE_DATA;
+	sccb.evbuf.data_id = DI_FCP_DUMP;
+	sccb.evbuf.event_id = 4712;
+#ifdef __s390x__
+	sccb.evbuf.asa_size = ASA_SIZE_64;
+#else
+	sccb.evbuf.asa_size = ASA_SIZE_32;
+#endif
+	sccb.evbuf.event_status = 0;
+	sccb.evbuf.blk_cnt = nr_blks;
+	sccb.evbuf.asa = (unsigned long)dest;
+	sccb.evbuf.fbn = start_blk;
+	sccb.evbuf.lbn = 0;
+	sccb.evbuf.dbs = 1;
+
+	request.sccb	 = &sccb;
+	request.command  = SCLP_CMDW_WRITE_EVENT_DATA;
+	request.status	 = SCLP_REQ_FILLED;
+	request.callback = sdias_callback;
+
+	rc = sdias_sclp_send(&request);
+	if (rc) {
+		ERROR_MSG("sclp_send failed: %x\n", rc);
+		goto out;
+	}
+	if (sccb.hdr.response_code != 0x0020) {
+		TRACE("copy failed: %x\n", sccb.hdr.response_code);
+		rc = -EIO;
+		goto out;
+	}
+
+	switch (sccb.evbuf.event_status) {
+		case EVSTATE_ALL_STORED:
+			TRACE("all stored\n");
+		case EVSTATE_PART_STORED:
+			TRACE("part stored: %i\n", sccb.evbuf.blk_cnt);
+			break;
+		case EVSTATE_NO_DATA:
+			TRACE("no data\n");
+		default:
+			ERROR_MSG("Error from SCLP while copying hsa. "
+				  "Event status = %x\n",
+				sccb.evbuf.event_status);
+			rc = -EIO;
+	}
+out:
+	mutex_unlock(&sdias_mutex);
+	return rc;
+}
+
+int __init sdias_init(void)
+{
+	int rc;
+
+	if (ipl_info.type != IPL_TYPE_FCP_DUMP)
+		return 0;
+	sdias_dbf = debug_register("dump_sdias", 4, 1, 4 * sizeof(long));
+	debug_register_view(sdias_dbf, &debug_sprintf_view);
+	debug_set_level(sdias_dbf, 6);
+	rc = sclp_register(&sclp_sdias_register);
+	if (rc) {
+		ERROR_MSG("sclp register failed\n");
+		return rc;
+	}
+	init_waitqueue_head(&sdias_wq);
+	TRACE("init done\n");
+	return 0;
+}
+
+void __exit sdias_exit(void)
+{
+	debug_unregister(sdias_dbf);
+	sclp_unregister(&sclp_sdias_register);
+}
Index: quilt-2.6/drivers/s390/char/zcore.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/char/zcore.c	2007-04-27 16:04:59.000000000 +0200
@@ -0,0 +1,651 @@
+/*
+ * zcore module to export memory content and register sets for creating system
+ * dumps on SCSI disks (zfcpdump). The "zcore/mem" debugfs file shows the same
+ * dump format as s390 standalone dumps.
+ *
+ * For more information please refer to Documentation/s390/zfcpdump.txt
+ *
+ * Copyright IBM Corp. 2003,2007
+ * Author(s): Michael Holzheu
+ */
+
+#include <linux/init.h>
+#include <linux/miscdevice.h>
+#include <linux/utsname.h>
+#include <linux/debugfs.h>
+#include <asm/ipl.h>
+#include <asm/sclp.h>
+#include <asm/setup.h>
+#include <asm/sigp.h>
+#include <asm/uaccess.h>
+#include <asm/debug.h>
+#include <asm/processor.h>
+#include <asm/irqflags.h>
+
+#define TRACE(x...) debug_sprintf_event(zcore_dbf, 1, x)
+#define MSG(x...) printk( KERN_ALERT x )
+#define ERROR_MSG(x...) printk ( KERN_ALERT "DUMP: " x )
+
+#define TO_USER		0
+#define TO_KERNEL	1
+
+enum arch_id {
+	ARCH_S390	= 0,
+	ARCH_S390X	= 1,
+};
+
+/* dump system info */
+
+struct sys_info {
+	enum arch_id	arch;
+	unsigned long	sa_base;
+	u32		sa_size;
+	int		cpu_map[NR_CPUS];
+	unsigned long	mem_size;
+	union save_area	lc_mask;
+};
+
+static struct sys_info sys_info;
+static struct debug_info *zcore_dbf;
+static int hsa_available;
+static struct dentry *zcore_dir;
+static struct dentry *zcore_file;
+
+/*
+ * Copy memory from HSA to kernel or user memory (not reentrant):
+ *
+ * @dest:  Kernel or user buffer where memory should be copied to
+ * @src:   Start address within HSA where data should be copied
+ * @count: Size of buffer, which should be copied
+ * @mode:  Either TO_KERNEL or TO_USER
+ */
+static int memcpy_hsa(void *dest, unsigned long src, size_t count, int mode)
+{
+	int offs, blk_num;
+	static char buf[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
+
+	if (count == 0)
+		return 0;
+
+	/* copy first block */
+	offs = 0;
+	if ((src % PAGE_SIZE) != 0) {
+		blk_num = src / PAGE_SIZE + 2;
+		if (sclp_sdias_copy(buf, blk_num, 1)) {
+			TRACE("sclp_sdias_copy() failed\n");
+			return -EIO;
+		}
+		offs = min((PAGE_SIZE - (src % PAGE_SIZE)), count);
+		if (mode == TO_USER) {
+			if (copy_to_user((__force __user void*) dest,
+					 buf + (src % PAGE_SIZE), offs))
+				return -EFAULT;
+		} else
+			memcpy(dest, buf + (src % PAGE_SIZE), offs);
+	}
+	if (offs == count)
+		goto out;
+
+	/* copy middle */
+	for (; (offs + PAGE_SIZE) <= count; offs += PAGE_SIZE) {
+		blk_num = (src + offs) / PAGE_SIZE + 2;
+		if (sclp_sdias_copy(buf, blk_num, 1)) {
+			TRACE("sclp_sdias_copy() failed\n");
+			return -EIO;
+		}
+		if (mode == TO_USER) {
+			if (copy_to_user((__force __user void*) dest + offs,
+					 buf, PAGE_SIZE))
+				return -EFAULT;
+		} else
+			memcpy(dest + offs, buf, PAGE_SIZE);
+	}
+	if (offs == count)
+		goto out;
+
+	/* copy last block */
+	blk_num = (src + offs) / PAGE_SIZE + 2;
+	if (sclp_sdias_copy(buf, blk_num, 1)) {
+		TRACE("sclp_sdias_copy() failed\n");
+		return -EIO;
+	}
+	if (mode == TO_USER) {
+		if (copy_to_user((__force __user void*) dest + offs, buf,
+				 PAGE_SIZE))
+			return -EFAULT;
+	} else
+		memcpy(dest + offs, buf, count - offs);
+out:
+	return 0;
+}
+
+static int memcpy_hsa_user(void __user *dest, unsigned long src, size_t count)
+{
+	return memcpy_hsa((void __force *) dest, src, count, TO_USER);
+}
+
+static int memcpy_hsa_kernel(void *dest, unsigned long src, size_t count)
+{
+	return memcpy_hsa(dest, src, count, TO_KERNEL);
+}
+
+static int memcpy_real(void *dest, unsigned long src, size_t count)
+{
+	unsigned long flags;
+	int rc = -EFAULT;
+	register unsigned long _dest asm("2") = (unsigned long) dest;
+	register unsigned long _len1 asm("3") = (unsigned long) count;
+	register unsigned long _src  asm("4") = src;
+	register unsigned long _len2 asm("5") = (unsigned long) count;
+
+	if (count == 0)
+		return 0;
+	flags = __raw_local_irq_stnsm(0xf8); /* switch to real mode */
+	asm volatile (
+		"0:	mvcle	%1,%2,0x0\n"
+		"1:	jo	0b\n"
+		"	lhi	%0,0x0\n"
+		"2:\n"
+		EX_TABLE(1b,2b)
+		: "+d" (rc)
+		: "d" (_dest), "d" (_src), "d" (_len1), "d" (_len2)
+		: "cc", "memory");
+	__raw_local_irq_ssm(flags);
+
+	return rc;
+}
+
+static int memcpy_real_user(__user void *dest, unsigned long src, size_t count)
+{
+	static char buf[4096];
+	int offs = 0, size;
+
+	while (offs < count) {
+		size = min(sizeof(buf), count - offs);
+		if (memcpy_real(buf, src + offs, size))
+			return -EFAULT;
+		if (copy_to_user(dest + offs, buf, size))
+			return -EFAULT;
+		offs += size;
+	}
+	return 0;
+}
+
+#ifdef __s390x__
+/*
+ * Convert s390x (64 bit) cpu info to s390 (32 bit) cpu info
+ */
+static void __init s390x_to_s390_regs(union save_area *out, union save_area *in,
+				      int cpu)
+{
+	int i;
+
+	for (i = 0; i < 16; i++) {
+		out->s390.gp_regs[i] = in->s390x.gp_regs[i] & 0x00000000ffffffff;
+		out->s390.acc_regs[i] = in->s390x.acc_regs[i];
+		out->s390.ctrl_regs[i] =
+			in->s390x.ctrl_regs[i] & 0x00000000ffffffff;
+	}
+	/* locore for 31 bit has only space for fpregs 0,2,4,6 */
+	out->s390.fp_regs[0] = in->s390x.fp_regs[0];
+	out->s390.fp_regs[1] = in->s390x.fp_regs[2];
+	out->s390.fp_regs[2] = in->s390x.fp_regs[4];
+	out->s390.fp_regs[3] = in->s390x.fp_regs[6];
+	memcpy(&(out->s390.psw[0]), &(in->s390x.psw[0]), 4);
+	out->s390.psw[1] |= 0x8; /* set bit 12 */
+	memcpy(&(out->s390.psw[4]),&(in->s390x.psw[12]), 4);
+	out->s390.psw[4] |= 0x80; /* set (31bit) addressing bit */
+	out->s390.pref_reg = in->s390x.pref_reg;
+	out->s390.timer = in->s390x.timer;
+	out->s390.clk_cmp = in->s390x.clk_cmp;
+}
+
+static void __init s390x_to_s390_save_areas(void)
+{
+	int i = 1;
+	static union save_area tmp;
+
+	while (zfcpdump_save_areas[i]) {
+		s390x_to_s390_regs(&tmp, zfcpdump_save_areas[i], i);
+		memcpy(zfcpdump_save_areas[i], &tmp, sizeof(tmp));
+		i++;
+	}
+}
+
+#endif /* __s390x__ */
+
+static int __init init_cpu_info(enum arch_id arch)
+{
+	union save_area *sa;
+
+	/* get info for boot cpu from lowcore, stored in the HSA */
+
+	sa = kmalloc(sizeof(*sa), GFP_KERNEL);
+	if (!sa) {
+		ERROR_MSG("kmalloc failed: %s: %i\n",__FUNCTION__, __LINE__);
+		return -ENOMEM;
+	}
+	if (memcpy_hsa_kernel(sa, sys_info.sa_base, sys_info.sa_size) < 0) {
+		ERROR_MSG("could not copy from HSA\n");
+		kfree(sa);
+		return -EIO;
+	}
+	zfcpdump_save_areas[0] = sa;
+
+#ifdef __s390x__
+	/* convert s390x regs to s390, if we are dumping an s390 Linux */
+
+	if (arch == ARCH_S390)
+		s390x_to_s390_save_areas();
+#endif
+
+	return 0;
+}
+
+static DEFINE_MUTEX(zcore_mutex);
+
+#define DUMP_VERSION	0x3
+#define DUMP_MAGIC	0xa8190173618f23fdULL
+#define DUMP_ARCH_S390X	2
+#define DUMP_ARCH_S390	1
+#define HEADER_SIZE	4096
+
+/* dump header dumped according to s390 crash dump format */
+
+struct zcore_header {
+	u64 magic;
+	u32 version;
+	u32 header_size;
+	u32 dump_level;
+	u32 page_size;
+	u64 mem_size;
+	u64 mem_start;
+	u64 mem_end;
+	u32 num_pages;
+	u32 pad1;
+	u64 tod;
+	cpuid_t cpu_id;
+	u32 arch_id;
+	u32 build_arch;
+	char pad2[4016];
+} __attribute__((packed,__aligned__(16)));
+
+static struct zcore_header zcore_header = {
+	.magic		= DUMP_MAGIC,
+	.version	= DUMP_VERSION,
+	.header_size	= 4096,
+	.dump_level	= 0,
+	.page_size	= PAGE_SIZE,
+	.mem_start	= 0,
+#ifdef __s390x__
+	.build_arch	= DUMP_ARCH_S390X,
+#else
+	.build_arch	= DUMP_ARCH_S390,
+#endif
+};
+
+/*
+ * Copy lowcore info to buffer. Use map in order to copy only register parts.
+ *
+ * @buf:    User buffer
+ * @sa:     Pointer to save area
+ * @sa_off: Offset in save area to copy
+ * @len:    Number of bytes to copy
+ */
+static int copy_lc(void __user *buf, void *sa, int sa_off, int len)
+{
+	int i;
+	char *lc_mask = (char*)&sys_info.lc_mask;
+
+	for (i = 0; i < len; i++) {
+		if (!lc_mask[i + sa_off])
+			continue;
+		if (copy_to_user(buf + i, sa + sa_off + i, 1))
+			return -EFAULT;
+	}
+	return 0;
+}
+
+/*
+ * Copy lowcores info to memory, if necessary
+ *
+ * @buf:   User buffer
+ * @addr:  Start address of buffer in dump memory
+ * @count: Size of buffer
+ */
+static int zcore_add_lc(char __user *buf, unsigned long start, size_t count)
+{
+	unsigned long end;
+	int i = 0;
+
+	if (count == 0)
+		return 0;
+
+	end = start + count;
+	while (zfcpdump_save_areas[i]) {
+		unsigned long cp_start, cp_end; /* copy range */
+		unsigned long sa_start, sa_end; /* save area range */
+		unsigned long prefix;
+		unsigned long sa_off, len, buf_off;
+
+		if (sys_info.arch == ARCH_S390)
+			prefix = zfcpdump_save_areas[i]->s390.pref_reg;
+		else
+			prefix = zfcpdump_save_areas[i]->s390x.pref_reg;
+
+		sa_start = prefix + sys_info.sa_base;
+		sa_end = prefix + sys_info.sa_base + sys_info.sa_size;
+
+		if ((end < sa_start) || (start > sa_end))
+			goto next;
+		cp_start = max(start, sa_start);
+		cp_end = min(end, sa_end);
+
+		buf_off = cp_start - start;
+		sa_off = cp_start - sa_start;
+		len = cp_end - cp_start;
+
+		TRACE("copy_lc for: %lx\n", start);
+		if (copy_lc(buf + buf_off, zfcpdump_save_areas[i], sa_off, len))
+			return -EFAULT;
+next:
+		i++;
+	}
+	return 0;
+}
+
+/*
+ * Read routine for zcore character device
+ * First 4K are dump header
+ * Next 32MB are HSA Memory
+ * Rest is read from absolute Memory
+ */
+static ssize_t zcore_read(struct file *file, char __user *buf, size_t count,
+			  loff_t *ppos)
+{
+	unsigned long mem_start; /* Start address in memory */
+	size_t mem_offs;	 /* Offset in dump memory */
+	size_t hdr_count;	 /* Size of header part of output buffer */
+	size_t size;
+	int rc;
+
+	mutex_lock(&zcore_mutex);
+
+	if (*ppos > (sys_info.mem_size + HEADER_SIZE)) {
+		rc = -EINVAL;
+		goto fail;
+	}
+
+	count = min(count, (size_t) (sys_info.mem_size + HEADER_SIZE - *ppos));
+
+	/* Copy dump header */
+	if (*ppos < HEADER_SIZE) {
+		size = min(count, (size_t) (HEADER_SIZE - *ppos));
+		if (copy_to_user(buf, &zcore_header + *ppos, size)) {
+			rc = -EFAULT;
+			goto fail;
+		}
+		hdr_count = size;
+		mem_start = 0;
+	} else {
+		hdr_count = 0;
+		mem_start = *ppos - HEADER_SIZE;
+	}
+
+	mem_offs = 0;
+
+	/* Copy from HSA data */
+	if (*ppos < (ZFCPDUMP_HSA_SIZE + HEADER_SIZE)) {
+		size = min((count - hdr_count), (size_t) (ZFCPDUMP_HSA_SIZE
+			   - mem_start));
+		rc = memcpy_hsa_user(buf + hdr_count, mem_start, size);
+		if (rc)
+			goto fail;
+
+		mem_offs += size;
+	}
+
+	/* Copy from real mem */
+	size = count - mem_offs - hdr_count;
+	rc = memcpy_real_user(buf + hdr_count + mem_offs, mem_start + mem_offs,
+			      size);
+	if (rc)
+		goto fail;
+
+	/*
+	 * Since s390 dump analysis tools like lcrash or crash
+	 * expect register sets in the prefix pages of the cpus,
+	 * we copy them into the read buffer, if necessary.
+	 * buf + hdr_count: Start of memory part of output buffer
+	 * mem_start: Start memory address to copy from
+	 * count - hdr_count: Size of memory area to copy
+	 */
+	if (zcore_add_lc(buf + hdr_count, mem_start, count - hdr_count)) {
+		rc = -EFAULT;
+		goto fail;
+	}
+	*ppos += count;
+fail:
+	mutex_unlock(&zcore_mutex);
+	return (rc < 0) ? rc : count;
+}
+
+static int zcore_open(struct inode *inode, struct file *filp)
+{
+	if (!hsa_available)
+		return -ENODATA;
+	else
+		return capable(CAP_SYS_RAWIO) ? 0 : -EPERM;
+}
+
+static int zcore_release(struct inode *inode, struct file *filep)
+{
+	diag308(DIAG308_REL_HSA, NULL);
+	hsa_available = 0;
+	return 0;
+}
+
+static loff_t zcore_lseek(struct file *file, loff_t offset, int orig)
+{
+	loff_t rc;
+
+	mutex_lock(&zcore_mutex);
+	switch (orig) {
+	case 0:
+		file->f_pos = offset;
+		rc = file->f_pos;
+		break;
+	case 1:
+		file->f_pos += offset;
+		rc = file->f_pos;
+		break;
+	default:
+		rc = -EINVAL;
+	}
+	mutex_unlock(&zcore_mutex);
+	return rc;
+}
+
+static struct file_operations zcore_fops = {
+	.owner		= THIS_MODULE,
+	.llseek		= zcore_lseek,
+	.read		= zcore_read,
+	.open		= zcore_open,
+	.release	= zcore_release,
+};
+
+
+static void __init set_s390_lc_mask(union save_area *map)
+{
+	memset(&map->s390.ext_save, 0xff, sizeof(map->s390.ext_save));
+	memset(&map->s390.timer, 0xff, sizeof(map->s390.timer));
+	memset(&map->s390.clk_cmp, 0xff, sizeof(map->s390.clk_cmp));
+	memset(&map->s390.psw, 0xff, sizeof(map->s390.psw));
+	memset(&map->s390.pref_reg, 0xff, sizeof(map->s390.pref_reg));
+	memset(&map->s390.acc_regs, 0xff, sizeof(map->s390.acc_regs));
+	memset(&map->s390.fp_regs, 0xff, sizeof(map->s390.fp_regs));
+	memset(&map->s390.gp_regs, 0xff, sizeof(map->s390.gp_regs));
+	memset(&map->s390.ctrl_regs, 0xff, sizeof(map->s390.ctrl_regs));
+}
+
+static void __init set_s390x_lc_mask(union save_area *map)
+{
+	memset(&map->s390x.fp_regs, 0xff, sizeof(map->s390x.fp_regs));
+	memset(&map->s390x.gp_regs, 0xff, sizeof(map->s390x.gp_regs));
+	memset(&map->s390x.psw, 0xff, sizeof(map->s390x.psw));
+	memset(&map->s390x.pref_reg, 0xff, sizeof(map->s390x.pref_reg));
+	memset(&map->s390x.fp_ctrl_reg, 0xff, sizeof(map->s390x.fp_ctrl_reg));
+	memset(&map->s390x.tod_reg, 0xff, sizeof(map->s390x.tod_reg));
+	memset(&map->s390x.timer, 0xff, sizeof(map->s390x.timer));
+	memset(&map->s390x.clk_cmp, 0xff, sizeof(map->s390x.clk_cmp));
+	memset(&map->s390x.acc_regs, 0xff, sizeof(map->s390x.acc_regs));
+	memset(&map->s390x.ctrl_regs, 0xff, sizeof(map->s390x.ctrl_regs));
+}
+
+/*
+ * Initialize dump globals for a given architecture
+ */
+static int __init sys_info_init(enum arch_id arch)
+{
+	switch (arch) {
+	case ARCH_S390X:
+		MSG("DETECTED 'S390X (64 bit) OS'\n");
+		sys_info.sa_base = SAVE_AREA_BASE_S390X;
+		sys_info.sa_size = sizeof(struct save_area_s390x);
+		set_s390x_lc_mask(&sys_info.lc_mask);
+		break;
+	case ARCH_S390:
+		MSG("DETECTED 'S390 (32 bit) OS'\n");
+		sys_info.sa_base = SAVE_AREA_BASE_S390;
+		sys_info.sa_size = sizeof(struct save_area_s390);
+		set_s390_lc_mask(&sys_info.lc_mask);
+		break;
+	default:
+		ERROR_MSG("unknown architecture 0x%x.\n",arch);
+		return -EINVAL;
+	}
+	sys_info.arch = arch;
+	if (init_cpu_info(arch)) {
+		ERROR_MSG("get cpu info failed\n");
+		return -ENOMEM;
+	}
+	sys_info.mem_size = real_memory_size;
+
+	return 0;
+}
+
+static int __init check_sdias(void)
+{
+	int rc, act_hsa_size;
+
+	rc = sclp_sdias_blk_count();
+	if (rc < 0) {
+		ERROR_MSG("Could not determine HSA size\n");
+		return rc;
+	}
+	act_hsa_size = (rc - 1) * PAGE_SIZE;
+	if (act_hsa_size < ZFCPDUMP_HSA_SIZE) {
+		ERROR_MSG("HSA size too small: %i\n", act_hsa_size);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static void __init zcore_header_init(int arch, struct zcore_header *hdr)
+{
+	if (arch == ARCH_S390X)
+		hdr->arch_id = DUMP_ARCH_S390X;
+	else
+		hdr->arch_id = DUMP_ARCH_S390;
+	hdr->mem_size = sys_info.mem_size;
+	hdr->mem_end = sys_info.mem_size;
+	hdr->num_pages = sys_info.mem_size / PAGE_SIZE;
+	hdr->tod = get_clock();
+	get_cpu_id(&hdr->cpu_id);
+}
+
+extern int sdias_init(void);
+
+static int __init zcore_init(void)
+{
+	unsigned char arch;
+	int rc;
+
+	if (ipl_info.type != IPL_TYPE_FCP_DUMP)
+		return -ENODATA;
+
+	zcore_dbf = debug_register("zcore", 4, 1, 4 * sizeof(long));
+	debug_register_view(zcore_dbf, &debug_sprintf_view);
+	debug_set_level(zcore_dbf, 6);
+
+	TRACE("devno:  %x\n", ipl_info.data.fcp.dev_id.devno);
+	TRACE("wwpn:   %llx\n", (unsigned long long) ipl_info.data.fcp.wwpn);
+	TRACE("lun:    %llx\n", (unsigned long long) ipl_info.data.fcp.lun);
+
+	rc = sdias_init();
+	if (rc)
+		goto fail;
+
+	rc = check_sdias();
+	if (rc) {
+		ERROR_MSG("Dump initialization failed\n");
+		goto fail;
+	}
+
+	rc = memcpy_hsa_kernel(&arch, __LC_AR_MODE_ID, 1);
+	if (rc) {
+		ERROR_MSG("sdial memcpy for arch id failed\n");
+		goto fail;
+	}
+
+#ifndef __s390x__
+	if (arch == ARCH_S390X) {
+		ERROR_MSG("32 bit dumper can't dump 64 bit system!\n");
+		rc = -EINVAL;
+		goto fail;
+	}
+#endif
+
+	rc = sys_info_init(arch);
+	if (rc) {
+		ERROR_MSG("arch init failed\n");
+		goto fail;
+	}
+
+	zcore_header_init(arch, &zcore_header);
+
+	zcore_dir = debugfs_create_dir("zcore" , NULL);
+	if (!zcore_dir) {
+		rc = -ENOMEM;
+		goto fail;
+	}
+	zcore_file = debugfs_create_file("mem", S_IRUSR, zcore_dir, NULL,
+					 &zcore_fops);
+	if (!zcore_file) {
+		debugfs_remove(zcore_dir);
+		rc = -ENOMEM;
+		goto fail;
+	}
+	hsa_available = 1;
+	return 0;
+
+fail:
+	diag308(DIAG308_REL_HSA, NULL);
+	return rc;
+}
+
+extern void sdias_exit(void);
+
+static void __exit zcore_exit(void)
+{
+	debug_unregister(zcore_dbf);
+	sdias_exit();
+	diag308(DIAG308_REL_HSA, NULL);
+}
+
+MODULE_AUTHOR("Copyright IBM Corp. 2003,2007");
+MODULE_DESCRIPTION("zcore module for zfcpdump support");
+MODULE_LICENSE("GPL");
+
+subsys_initcall(zcore_init);
+module_exit(zcore_exit);
Index: quilt-2.6/include/asm-s390/ipl.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/ipl.h	2007-04-27 16:04:52.000000000 +0200
+++ quilt-2.6/include/asm-s390/ipl.h	2007-04-27 16:04:59.000000000 +0200
@@ -8,6 +8,8 @@
 #define _ASM_S390_IPL_H
 
 #include <asm/types.h>
+#include <asm/cio.h>
+#include <asm/setup.h>
 
 #define IPL_PARMBLOCK_ORIGIN	0x2000
 
@@ -79,6 +81,7 @@
 extern u32 ipl_flags;
 
 extern u32 dump_prefix_page;
+
 extern void do_reipl(void);
 extern void ipl_save_parameters(void);
 
@@ -88,6 +91,35 @@
 	IPL_NSS_VALID		= 4,
 };
 
+enum ipl_type {
+	IPL_TYPE_UNKNOWN	= 1,
+	IPL_TYPE_CCW		= 2,
+	IPL_TYPE_FCP		= 4,
+	IPL_TYPE_FCP_DUMP	= 8,
+	IPL_TYPE_NSS		= 16,
+};
+
+struct ipl_info
+{
+	enum ipl_type type;
+	union {
+		struct {
+			struct ccw_dev_id dev_id;
+		} ccw;
+		struct {
+			struct ccw_dev_id dev_id;
+			u64 wwpn;
+			u64 lun;
+		} fcp;
+		struct {
+			char name[NSS_NAME_SIZE + 1];
+		} nss;
+	} data;
+};
+
+extern struct ipl_info ipl_info;
+extern void setup_ipl_info(void);
+
 /*
  * DIAG 308 support
  */
Index: quilt-2.6/include/asm-s390/lowcore.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/lowcore.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/lowcore.h	2007-04-27 16:04:59.000000000 +0200
@@ -147,6 +147,52 @@
 void mcck_int_handler(void);
 void io_int_handler(void);
 
+struct save_area_s390 {
+	u32	ext_save;
+	u64	timer;
+	u64	clk_cmp;
+	u8	pad1[24];
+	u8	psw[8];
+	u32	pref_reg;
+	u8	pad2[20];
+	u32	acc_regs[16];
+	u64	fp_regs[4];
+	u32	gp_regs[16];
+	u32	ctrl_regs[16];
+}  __attribute__((packed));
+
+struct save_area_s390x {
+	u64	fp_regs[16];
+	u64	gp_regs[16];
+	u8	psw[16];
+	u8	pad1[8];
+	u32	pref_reg;
+	u32	fp_ctrl_reg;
+	u8	pad2[4];
+	u32	tod_reg;
+	u64	timer;
+	u64	clk_cmp;
+	u8	pad3[8];
+	u32	acc_regs[16];
+	u64	ctrl_regs[16];
+}  __attribute__((packed));
+
+union save_area {
+	struct save_area_s390	s390;
+	struct save_area_s390x	s390x;
+};
+
+#define SAVE_AREA_BASE_S390	0xd4
+#define SAVE_AREA_BASE_S390X	0x1200
+
+#ifndef __s390x__
+#define SAVE_AREA_SIZE sizeof(struct save_area_s390)
+#define SAVE_AREA_BASE SAVE_AREA_BASE_S390
+#else
+#define SAVE_AREA_SIZE sizeof(struct save_area_s390x)
+#define SAVE_AREA_BASE SAVE_AREA_BASE_S390X
+#endif
+
 struct _lowcore
 {
 #ifndef __s390x__
Index: quilt-2.6/include/asm-s390/sclp.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/sclp.h	2007-04-27 16:04:54.000000000 +0200
+++ quilt-2.6/include/asm-s390/sclp.h	2007-04-27 16:04:59.000000000 +0200
@@ -44,6 +44,8 @@
 
 extern struct sclp_readinfo_sccb s390_readinfo_sccb;
 extern void sclp_readinfo_early(void);
+extern int sclp_sdias_blk_count(void);
+extern int sclp_sdias_copy(void *dest, int blk_num, int nr_blks);
 extern int sclp_chp_configure(struct chp_id chpid);
 extern int sclp_chp_deconfigure(struct chp_id chpid);
 extern int sclp_chp_read_info(struct sclp_chp_info *info);
Index: quilt-2.6/include/asm-s390/setup.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/setup.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/setup.h	2007-04-27 16:04:59.000000000 +0200
@@ -40,6 +40,7 @@
 };
 
 extern struct mem_chunk memory_chunk[];
+extern unsigned long real_memory_size;
 
 #ifdef CONFIG_S390_SWITCH_AMODE
 extern unsigned int switch_amode;
@@ -77,6 +78,7 @@
 #endif /* __s390x__ */
 
 #define MACHINE_HAS_SCLP	(!MACHINE_IS_P390)
+#define ZFCPDUMP_HSA_SIZE	(32UL<<20)
 
 /*
  * Console mode. Override with conmode=
Index: quilt-2.6/include/asm-s390/smp.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/smp.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/smp.h	2007-04-27 16:04:59.000000000 +0200
@@ -119,4 +119,5 @@
 #define smp_setup_cpu_possible_map()	do { } while (0)
 #endif
 
+extern union save_area *zfcpdump_save_areas[NR_CPUS + 1];
 #endif

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 27/38] ctc: kmalloc->kzalloc/casting cleanups.

[Martin Schwidefsky]

[-- Attachment #1: 130-ctc-kmalloc.diff --]
[-- Type: text/plain, Size: 3110 bytes --]

From: "Ahmed S. Darwish" <darwish.07@gmail.com>

A patch for the CTC / ESCON network driver.  Switch from kmalloc to kzalloc
when appropriate, remove some unnecessary kmalloc casts too.  Since I have no
s390 machine, I didn't compile it but I examined it carefully.

Signed-off-by: "Ahmed S. Darwish" <darwish.07@gmail.com>
Cc: Frank Pavlic <fpavlic@de.ibm.com>
Cc: Ursula Braun <braunu@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/net/ctcmain.c |   23 +++++++++--------------
 1 files changed, 9 insertions(+), 14 deletions(-)

Index: quilt-2.6/drivers/s390/net/ctcmain.c
===================================================================
--- quilt-2.6.orig/drivers/s390/net/ctcmain.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/net/ctcmain.c	2007-04-27 16:04:59.000000000 +0200
@@ -1638,21 +1638,19 @@
 	struct channel *ch;
 
 	DBF_TEXT(trace, 2, __FUNCTION__);
-	if ((ch =
-	     (struct channel *) kmalloc(sizeof (struct channel),
-					GFP_KERNEL)) == NULL) {
+	ch = kzalloc(sizeof(struct channel), GFP_KERNEL);
+	if (!ch) {
 		ctc_pr_warn("ctc: Out of memory in add_channel\n");
 		return -1;
 	}
-	memset(ch, 0, sizeof (struct channel));
-	if ((ch->ccw = kmalloc(8*sizeof(struct ccw1),
-					       GFP_KERNEL | GFP_DMA)) == NULL) {
+	/* assure all flags and counters are reset */
+	ch->ccw = kzalloc(8 * sizeof(struct ccw1), GFP_KERNEL | GFP_DMA);
+	if (!ch->ccw) {
 		kfree(ch);
 		ctc_pr_warn("ctc: Out of memory in add_channel\n");
 		return -1;
 	}
 
-	memset(ch->ccw, 0, 8*sizeof(struct ccw1));	// assure all flags and counters are reset
 
 	/**
 	 * "static" ccws are used in the following way:
@@ -1692,15 +1690,14 @@
 		return -1;
 	}
 	fsm_newstate(ch->fsm, CH_STATE_IDLE);
-	if ((ch->irb = kmalloc(sizeof (struct irb),
-					      GFP_KERNEL)) == NULL) {
+	ch->irb = kzalloc(sizeof(struct irb), GFP_KERNEL);
+	if (!ch->irb) {
 		ctc_pr_warn("ctc: Out of memory in add_channel\n");
 		kfree_fsm(ch->fsm);
 		kfree(ch->ccw);
 		kfree(ch);
 		return -1;
 	}
-	memset(ch->irb, 0, sizeof (struct irb));
 	while (*c && less_than((*c)->id, ch->id))
 		c = &(*c)->next;
 	if (*c && (!strncmp((*c)->id, ch->id, CTC_ID_SIZE))) {
@@ -2745,14 +2742,13 @@
 	if (!get_device(&cgdev->dev))
 		return -ENODEV;
 
-	priv = kmalloc(sizeof (struct ctc_priv), GFP_KERNEL);
+	priv = kzalloc(sizeof(struct ctc_priv), GFP_KERNEL);
 	if (!priv) {
 		ctc_pr_err("%s: Out of memory\n", __func__);
 		put_device(&cgdev->dev);
 		return -ENOMEM;
 	}
 
-	memset(priv, 0, sizeof (struct ctc_priv));
 	rc = ctc_add_files(&cgdev->dev);
 	if (rc) {
 		kfree(priv);
@@ -2793,10 +2789,9 @@
 	DBF_TEXT(setup, 3, __FUNCTION__);
 
 	if (alloc_device) {
-		dev = kmalloc(sizeof (struct net_device), GFP_KERNEL);
+		dev = kzalloc(sizeof(struct net_device), GFP_KERNEL);
 		if (!dev)
 			return NULL;
-		memset(dev, 0, sizeof (struct net_device));
 	}
 
 	dev->priv = privptr;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 28/38] sclp: initialize early.

[Martin Schwidefsky]

[-- Attachment #1: 131-sclp-init.diff --]
[-- Type: text/plain, Size: 1101 bytes --]

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

Add explicit sclp initialization for those sclp users that do not
register with the interface.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/char/sclp.c |    8 ++++++++
 1 files changed, 8 insertions(+)

Index: quilt-2.6/drivers/s390/char/sclp.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp.c	2007-04-27 16:04:59.000000000 +0200
@@ -15,6 +15,7 @@
 #include <linux/timer.h>
 #include <linux/reboot.h>
 #include <linux/jiffies.h>
+#include <linux/init.h>
 #include <asm/types.h>
 #include <asm/s390_ext.h>
 
@@ -930,3 +931,10 @@
 	sclp_init_mask(1);
 	return 0;
 }
+
+static __init int sclp_initcall(void)
+{
+	return sclp_init();
+}
+
+arch_initcall(sclp_initcall);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 29/38] vmlogrdr: stop IUCV connection in vmlogrdr_release.

[Martin Schwidefsky]

[-- Attachment #1: 132-vmlogrdr-iucv.diff --]
[-- Type: text/plain, Size: 993 bytes --]

From: Ursula Braun <braunu@de.ibm.com>

Reopen of /dev/account failed. The IUCV path has to be terminated
in vmlogrdr_release.

Signed-off-by: Ursula Braun <braunu@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/char/vmlogrdr.c |    3 +++
 1 files changed, 3 insertions(+)

Index: quilt-2.6/drivers/s390/char/vmlogrdr.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/vmlogrdr.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/vmlogrdr.c	2007-04-27 16:04:59.000000000 +0200
@@ -385,6 +385,9 @@
 
 	struct vmlogrdr_priv_t * logptr = filp->private_data;
 
+	iucv_path_sever(logptr->path, NULL);
+	kfree(logptr->path);
+	logptr->path = NULL;
 	if (logptr->autorecording) {
 		ret = vmlogrdr_recording(logptr,0,logptr->autopurge);
 		if (ret)

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 30/38] sclp: fix coding style.

[Martin Schwidefsky]

[-- Attachment #1: 133-sclp-coding.diff --]
[-- Type: text/plain, Size: 11226 bytes --]

From: Stefan Haberland <stefan.haberland@de.ibm.com>

Use only capital letters for defines.

Cc: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Stefan Haberland <stefan.haberland@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/char/sclp.c         |    2 -
 drivers/s390/char/sclp.h         |   72 +++++++++++++++++++--------------------
 drivers/s390/char/sclp_cpi.c     |    4 +-
 drivers/s390/char/sclp_quiesce.c |    2 -
 drivers/s390/char/sclp_rw.c      |   16 ++++----
 drivers/s390/char/sclp_sdias.c   |    6 +--
 drivers/s390/char/sclp_tty.c     |    6 +--
 drivers/s390/char/sclp_vt220.c   |    8 ++--
 8 files changed, 58 insertions(+), 58 deletions(-)

Index: quilt-2.6/drivers/s390/char/sclp.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp.c	2007-04-27 16:04:59.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp.c	2007-04-27 16:05:00.000000000 +0200
@@ -511,7 +511,7 @@
 }
 
 static struct sclp_register sclp_state_change_event = {
-	.receive_mask = EvTyp_StateChange_Mask,
+	.receive_mask = EVTYP_STATECHANGE_MASK,
 	.receiver_fn = sclp_state_change_cb
 };
 
Index: quilt-2.6/drivers/s390/char/sclp_cpi.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp_cpi.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp_cpi.c	2007-04-27 16:05:00.000000000 +0200
@@ -46,7 +46,7 @@
 /* Event type structure for write message and write priority message */
 static struct sclp_register sclp_cpi_event =
 {
-	.send_mask = EvTyp_CtlProgIdent_Mask
+	.send_mask = EVTYP_CTLPROGIDENT_MASK
 };
 
 MODULE_LICENSE("GPL");
@@ -201,7 +201,7 @@
 		       "console.\n");
 		return -EINVAL;
 	}
-	if (!(sclp_cpi_event.sclp_send_mask & EvTyp_CtlProgIdent_Mask)) {
+	if (!(sclp_cpi_event.sclp_send_mask & EVTYP_CTLPROGIDENT_MASK)) {
 		printk(KERN_WARNING "cpi: no control program identification "
 		       "support\n");
 		sclp_unregister(&sclp_cpi_event);
Index: quilt-2.6/drivers/s390/char/sclp.h
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp.h	2007-04-27 16:04:59.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp.h	2007-04-27 16:05:00.000000000 +0200
@@ -19,35 +19,35 @@
 #define MAX_KMEM_PAGES (sizeof(unsigned long) << 3)
 #define MAX_CONSOLE_PAGES	4
 
-#define EvTyp_OpCmd		0x01
-#define EvTyp_Msg		0x02
-#define EvTyp_StateChange	0x08
-#define EvTyp_PMsgCmd		0x09
-#define EvTyp_CntlProgOpCmd	0x20
-#define EvTyp_CntlProgIdent	0x0B
-#define EvTyp_SigQuiesce	0x1D
-#define EvTyp_VT220Msg		0x1A
-#define EvTyp_SDIAS		0x1C
-
-#define EvTyp_OpCmd_Mask	0x80000000
-#define EvTyp_Msg_Mask		0x40000000
-#define EvTyp_StateChange_Mask	0x01000000
-#define EvTyp_PMsgCmd_Mask	0x00800000
-#define EvTyp_CtlProgOpCmd_Mask	0x00000001
-#define EvTyp_CtlProgIdent_Mask	0x00200000
-#define EvTyp_SigQuiesce_Mask	0x00000008
-#define EvTyp_VT220Msg_Mask	0x00000040
-#define EvTyp_SDIAS_Mask	0x00000010
-
-#define GnrlMsgFlgs_DOM		0x8000
-#define GnrlMsgFlgs_SndAlrm	0x4000
-#define GnrlMsgFlgs_HoldMsg	0x2000
-
-#define LnTpFlgs_CntlText	0x8000
-#define LnTpFlgs_LabelText	0x4000
-#define LnTpFlgs_DataText	0x2000
-#define LnTpFlgs_EndText	0x1000
-#define LnTpFlgs_PromptText	0x0800
+#define EVTYP_OPCMD		0x01
+#define EVTYP_MSG		0x02
+#define EVTYP_STATECHANGE	0x08
+#define EVTYP_PMSGCMD		0x09
+#define EVTYP_CNTLPROGOPCMD	0x20
+#define EVTYP_CNTLPROGIDENT	0x0B
+#define EVTYP_SIGQUIESCE	0x1D
+#define EVTYP_VT220MSG		0x1A
+#define EVTYP_SDIAS		0x1C
+
+#define EVTYP_OPCMD_MASK	0x80000000
+#define EVTYP_MSG_MASK		0x40000000
+#define EVTYP_STATECHANGE_MASK	0x01000000
+#define EVTYP_PMSGCMD_MASK	0x00800000
+#define EVTYP_CTLPROGOPCMD_MASK	0x00000001
+#define EVTYP_CTLPROGIDENT_MASK	0x00200000
+#define EVTYP_SIGQUIESCE_MASK	0x00000008
+#define EVTYP_VT220MSG_MASK	0x00000040
+#define EVTYP_SDIAS_MASK	0x00000010
+
+#define GNRLMSGFLGS_DOM		0x8000
+#define GNRLMSGFLGS_SNDALRM	0x4000
+#define GNRLMSGFLGS_HOLDMSG	0x2000
+
+#define LNTPFLGS_CNTLTEXT	0x8000
+#define LNTPFLGS_LABELTEXT	0x4000
+#define LNTPFLGS_DATATEXT	0x2000
+#define LNTPFLGS_ENDTEXT	0x1000
+#define LNTPFLGS_PROMPTTEXT	0x0800
 
 typedef unsigned int sclp_cmdw_t;
 
@@ -58,15 +58,15 @@
 #define SCLP_CMDW_READ_SCP_INFO_FORCED	0x00120001
 
 #define GDS_ID_MDSMU		0x1310
-#define GDS_ID_MDSRouteInfo	0x1311
-#define GDS_ID_AgUnWrkCorr	0x1549
-#define GDS_ID_SNACondReport	0x1532
+#define GDS_ID_MDSROUTEINFO	0x1311
+#define GDS_ID_AGUNWRKCORR	0x1549
+#define GDS_ID_SNACONDREPORT	0x1532
 #define GDS_ID_CPMSU		0x1212
-#define GDS_ID_RoutTargInstr	0x154D
-#define GDS_ID_OpReq		0x8070
-#define GDS_ID_TextCmd		0x1320
+#define GDS_ID_ROUTTARGINSTR	0x154D
+#define GDS_ID_OPREQ		0x8070
+#define GDS_ID_TEXTCMD		0x1320
 
-#define GDS_KEY_SelfDefTextMsg	0x31
+#define GDS_KEY_SELFDEFTEXTMSG	0x31
 
 typedef u32 sccb_mask_t;	/* ATTENTION: assumes 32bit mask !!! */
 
Index: quilt-2.6/drivers/s390/char/sclp_quiesce.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp_quiesce.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp_quiesce.c	2007-04-27 16:05:00.000000000 +0200
@@ -43,7 +43,7 @@
 }
 
 static struct sclp_register sclp_quiesce_event = {
-	.receive_mask = EvTyp_SigQuiesce_Mask,
+	.receive_mask = EVTYP_SIGQUIESCE_MASK,
 	.receiver_fn = sclp_quiesce_handler
 };
 
Index: quilt-2.6/drivers/s390/char/sclp_rw.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp_rw.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp_rw.c	2007-04-27 16:05:00.000000000 +0200
@@ -30,7 +30,7 @@
 
 /* Event type structure for write message and write priority message */
 static struct sclp_register sclp_rw_event = {
-	.send_mask = EvTyp_Msg_Mask | EvTyp_PMsgCmd_Mask
+	.send_mask = EVTYP_MSG_MASK | EVTYP_PMSGCMD_MASK
 };
 
 /*
@@ -64,7 +64,7 @@
 	memset(sccb, 0, sizeof(struct write_sccb));
 	sccb->header.length = sizeof(struct write_sccb);
 	sccb->msg_buf.header.length = sizeof(struct msg_buf);
-	sccb->msg_buf.header.type = EvTyp_Msg;
+	sccb->msg_buf.header.type = EVTYP_MSG;
 	sccb->msg_buf.mdb.header.length = sizeof(struct mdb);
 	sccb->msg_buf.mdb.header.type = 1;
 	sccb->msg_buf.mdb.header.tag = 0xD4C4C240;	/* ebcdic "MDB " */
@@ -114,7 +114,7 @@
 	memset(mto, 0, sizeof(struct mto));
 	mto->length = sizeof(struct mto);
 	mto->type = 4;	/* message text object */
-	mto->line_type_flags = LnTpFlgs_EndText; /* end text */
+	mto->line_type_flags = LNTPFLGS_ENDTEXT; /* end text */
 
 	/* set pointer to first byte after struct mto. */
 	buffer->current_line = (char *) (mto + 1);
@@ -215,7 +215,7 @@
 		case '\a':	/* bell, one for several times	*/
 			/* set SCLP sound alarm bit in General Object */
 			buffer->sccb->msg_buf.mdb.go.general_msg_flags |=
-				GnrlMsgFlgs_SndAlrm;
+				GNRLMSGFLGS_SNDALRM;
 			break;
 		case '\t':	/* horizontal tabulator	 */
 			/* check if new mto needs to be created */
@@ -452,12 +452,12 @@
 		return -EIO;
 
 	sccb = buffer->sccb;
-	if (sclp_rw_event.sclp_send_mask & EvTyp_Msg_Mask)
+	if (sclp_rw_event.sclp_send_mask & EVTYP_MSG_MASK)
 		/* Use normal write message */
-		sccb->msg_buf.header.type = EvTyp_Msg;
-	else if (sclp_rw_event.sclp_send_mask & EvTyp_PMsgCmd_Mask)
+		sccb->msg_buf.header.type = EVTYP_MSG;
+	else if (sclp_rw_event.sclp_send_mask & EVTYP_PMSGCMD_MASK)
 		/* Use write priority message */
-		sccb->msg_buf.header.type = EvTyp_PMsgCmd;
+		sccb->msg_buf.header.type = EVTYP_PMSGCMD;
 	else
 		return -ENOSYS;
 	buffer->request.command = SCLP_CMDW_WRITE_EVENT_DATA;
Index: quilt-2.6/drivers/s390/char/sclp_sdias.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp_sdias.c	2007-04-27 16:04:59.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp_sdias.c	2007-04-27 16:05:00.000000000 +0200
@@ -30,7 +30,7 @@
 static struct debug_info *sdias_dbf;
 
 static struct sclp_register sclp_sdias_register = {
-	.send_mask = EvTyp_SDIAS_Mask,
+	.send_mask = EVTYP_SDIAS_MASK,
 };
 
 struct sdias_evbuf {
@@ -118,7 +118,7 @@
 
 	sccb.hdr.length = sizeof(sccb);
 	sccb.evbuf.hdr.length = sizeof(struct sdias_evbuf);
-	sccb.evbuf.hdr.type = EvTyp_SDIAS;
+	sccb.evbuf.hdr.type = EVTYP_SDIAS;
 	sccb.evbuf.event_qual = EQ_SIZE;
 	sccb.evbuf.data_id = DI_FCP_DUMP;
 	sccb.evbuf.event_id = 4712;
@@ -177,7 +177,7 @@
 
 	sccb.hdr.length = sizeof(sccb);
 	sccb.evbuf.hdr.length = sizeof(struct sdias_evbuf);
-	sccb.evbuf.hdr.type = EvTyp_SDIAS;
+	sccb.evbuf.hdr.type = EVTYP_SDIAS;
 	sccb.evbuf.hdr.flags = 0;
 	sccb.evbuf.event_qual = EQ_STORE_DATA;
 	sccb.evbuf.data_id = DI_FCP_DUMP;
Index: quilt-2.6/drivers/s390/char/sclp_tty.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp_tty.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp_tty.c	2007-04-27 16:05:00.000000000 +0200
@@ -648,7 +648,7 @@
 	subvec = start;
 	while (subvec < end) {
 		subvec = find_gds_subvector(subvec, end,
-					    GDS_KEY_SelfDefTextMsg);
+					    GDS_KEY_SELFDEFTEXTMSG);
 		if (!subvec)
 			break;
 		sclp_eval_selfdeftextmsg((struct gds_subvector *)(subvec + 1),
@@ -664,7 +664,7 @@
 
 	vec = start;
 	while (vec < end) {
-		vec = find_gds_vector(vec, end, GDS_ID_TextCmd);
+		vec = find_gds_vector(vec, end, GDS_ID_TEXTCMD);
 		if (!vec)
 			break;
 		sclp_eval_textcmd((struct gds_subvector *)(vec + 1),
@@ -703,7 +703,7 @@
 
 static struct sclp_register sclp_input_event =
 {
-	.receive_mask = EvTyp_OpCmd_Mask | EvTyp_PMsgCmd_Mask,
+	.receive_mask = EVTYP_OPCMD_MASK | EVTYP_PMSGCMD_MASK,
 	.state_change_fn = sclp_tty_state_change,
 	.receiver_fn = sclp_tty_receiver
 };
Index: quilt-2.6/drivers/s390/char/sclp_vt220.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp_vt220.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp_vt220.c	2007-04-27 16:05:00.000000000 +0200
@@ -99,8 +99,8 @@
 
 /* Registration structure for our interest in SCLP event buffers */
 static struct sclp_register sclp_vt220_register = {
-	.send_mask		= EvTyp_VT220Msg_Mask,
-	.receive_mask		= EvTyp_VT220Msg_Mask,
+	.send_mask		= EVTYP_VT220MSG_MASK,
+	.receive_mask		= EVTYP_VT220MSG_MASK,
 	.state_change_fn	= NULL,
 	.receiver_fn		= sclp_vt220_receiver_fn
 };
@@ -202,7 +202,7 @@
 static int
 __sclp_vt220_emit(struct sclp_vt220_request *request)
 {
-	if (!(sclp_vt220_register.sclp_send_mask & EvTyp_VT220Msg_Mask)) {
+	if (!(sclp_vt220_register.sclp_send_mask & EVTYP_VT220MSG_MASK)) {
 		request->sclp_req.status = SCLP_REQ_FAILED;
 		return -EIO;
 	}
@@ -284,7 +284,7 @@
 	sccb->header.length = sizeof(struct sclp_vt220_sccb);
 	sccb->header.function_code = SCLP_NORMAL_WRITE;
 	sccb->header.response_code = 0x0000;
-	sccb->evbuf.type = EvTyp_VT220Msg;
+	sccb->evbuf.type = EVTYP_VT220MSG;
 	sccb->evbuf.length = sizeof(struct evbuf_header);
 
 	return request;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 31/38] crypto: cleanup.

[Martin Schwidefsky]

[-- Attachment #1: 134-crypto.diff --]
[-- Type: text/plain, Size: 11857 bytes --]

From: Jan Glauber <jan.glauber@de.ibm.com>

Cleanup code and remove obsolete documentation.

Signed-off-by: Jan Glauber <jan.glauber@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 Documentation/s390/crypto/crypto-API.txt |   79 -------------------
 arch/s390/crypto/sha1_s390.c             |  127 +++++++++++++++----------------
 arch/s390/crypto/sha256_s390.c           |   38 +++------
 3 files changed, 80 insertions(+), 164 deletions(-)

Index: quilt-2.6/arch/s390/crypto/sha1_s390.c
===================================================================
--- quilt-2.6.orig/arch/s390/crypto/sha1_s390.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/crypto/sha1_s390.c	2007-04-27 16:05:00.000000000 +0200
@@ -25,99 +25,100 @@
  */
 #include <linux/init.h>
 #include <linux/module.h>
-#include <linux/mm.h>
 #include <linux/crypto.h>
-#include <asm/scatterlist.h>
-#include <asm/byteorder.h>
+
 #include "crypt_s390.h"
 
 #define SHA1_DIGEST_SIZE	20
 #define SHA1_BLOCK_SIZE		64
 
-struct crypt_s390_sha1_ctx {
-	u64 count;
+struct s390_sha1_ctx {
+	u64 count;		/* message length */
 	u32 state[5];
-	u32 buf_len;
-	u8 buffer[2 * SHA1_BLOCK_SIZE];
+	u8 buf[2 * SHA1_BLOCK_SIZE];
 };
 
 static void sha1_init(struct crypto_tfm *tfm)
 {
-	struct crypt_s390_sha1_ctx *ctx = crypto_tfm_ctx(tfm);
-
-	ctx->state[0] = 0x67452301;
-	ctx->state[1] =	0xEFCDAB89;
-	ctx->state[2] =	0x98BADCFE;
-	ctx->state[3] = 0x10325476;
-	ctx->state[4] =	0xC3D2E1F0;
+	struct s390_sha1_ctx *sctx = crypto_tfm_ctx(tfm);
 
-	ctx->count = 0;
-	ctx->buf_len = 0;
+	sctx->state[0] = 0x67452301;
+	sctx->state[1] = 0xEFCDAB89;
+	sctx->state[2] = 0x98BADCFE;
+	sctx->state[3] = 0x10325476;
+	sctx->state[4] = 0xC3D2E1F0;
+	sctx->count = 0;
 }
 
 static void sha1_update(struct crypto_tfm *tfm, const u8 *data,
 			unsigned int len)
 {
-	struct crypt_s390_sha1_ctx *sctx;
-	long imd_len;
-
-	sctx = crypto_tfm_ctx(tfm);
-	sctx->count += len * 8; /* message bit length */
-
-	/* anything in buffer yet? -> must be completed */
-	if (sctx->buf_len && (sctx->buf_len + len) >= SHA1_BLOCK_SIZE) {
-		/* complete full block and hash */
-		memcpy(sctx->buffer + sctx->buf_len, data,
-		       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;
-		len -= SHA1_BLOCK_SIZE - sctx->buf_len;
-		sctx->buf_len = 0;
+	struct s390_sha1_ctx *sctx = crypto_tfm_ctx(tfm);
+	unsigned int index;
+	int ret;
+
+	/* how much is already in the buffer? */
+	index = sctx->count & 0x3f;
+
+	sctx->count += len;
+
+	if (index + len < SHA1_BLOCK_SIZE)
+		goto store;
+
+	/* process one stored block */
+	if (index) {
+		memcpy(sctx->buf + index, data, SHA1_BLOCK_SIZE - index);
+		ret = crypt_s390_kimd(KIMD_SHA_1, sctx->state, sctx->buf,
+				      SHA1_BLOCK_SIZE);
+		BUG_ON(ret != SHA1_BLOCK_SIZE);
+		data += SHA1_BLOCK_SIZE - index;
+		len -= SHA1_BLOCK_SIZE - index;
 	}
 
-	/* rest of data contains full blocks? */
-	imd_len = len & ~0x3ful;
-	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) {
-		memcpy(sctx->buffer + sctx->buf_len , data, len);
-		sctx->buf_len += len;
+	/* process as many blocks as possible */
+	if (len >= SHA1_BLOCK_SIZE) {
+		ret = crypt_s390_kimd(KIMD_SHA_1, sctx->state, data,
+				      len & ~(SHA1_BLOCK_SIZE - 1));
+		BUG_ON(ret != (len & ~(SHA1_BLOCK_SIZE - 1)));
+		data += ret;
+		len -= ret;
 	}
-}
 
+store:
+	/* anything left? */
+	if (len)
+		memcpy(sctx->buf + index , data, len);
+}
 
-static void pad_message(struct crypt_s390_sha1_ctx* sctx)
+/* Add padding and return the message digest. */
+static void sha1_final(struct crypto_tfm *tfm, u8 *out)
 {
-	int index;
+	struct s390_sha1_ctx *sctx = crypto_tfm_ctx(tfm);
+	u64 bits;
+	unsigned int index, end;
+	int ret;
+
+	/* must perform manual padding */
+	index = sctx->count & 0x3f;
+	end =  (index < 56) ? SHA1_BLOCK_SIZE : (2 * SHA1_BLOCK_SIZE);
 
-	index = sctx->buf_len;
-	sctx->buf_len = (sctx->buf_len < 56) ?
-			 SHA1_BLOCK_SIZE:2 * SHA1_BLOCK_SIZE;
 	/* start pad with 1 */
-	sctx->buffer[index] = 0x80;
+	sctx->buf[index] = 0x80;
+
 	/* pad with zeros */
 	index++;
-	memset(sctx->buffer + index, 0x00, sctx->buf_len - index);
-	/* append length */
-	memcpy(sctx->buffer + sctx->buf_len - 8, &sctx->count,
-	       sizeof sctx->count);
-}
+	memset(sctx->buf + index, 0x00, end - index - 8);
 
-/* Add padding and return the message digest. */
-static void sha1_final(struct crypto_tfm *tfm, u8 *out)
-{
-	struct crypt_s390_sha1_ctx *sctx = crypto_tfm_ctx(tfm);
+	/* append message length */
+	bits = sctx->count * 8;
+	memcpy(sctx->buf + end - 8, &bits, sizeof(bits));
+
+	ret = crypt_s390_kimd(KIMD_SHA_1, sctx->state, sctx->buf, end);
+	BUG_ON(ret != end);
 
-	/* must perform manual padding */
-	pad_message(sctx);
-	crypt_s390_kimd(KIMD_SHA_1, sctx->state, sctx->buffer, sctx->buf_len);
 	/* copy digest to out */
 	memcpy(out, sctx->state, SHA1_DIGEST_SIZE);
+
 	/* wipe context */
 	memset(sctx, 0, sizeof *sctx);
 }
@@ -128,7 +129,7 @@
 	.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_ctxsize	=	sizeof(struct s390_sha1_ctx),
 	.cra_module	=	THIS_MODULE,
 	.cra_list	=	LIST_HEAD_INIT(alg.cra_list),
 	.cra_u		=	{ .digest = {
Index: quilt-2.6/arch/s390/crypto/sha256_s390.c
===================================================================
--- quilt-2.6.orig/arch/s390/crypto/sha256_s390.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/crypto/sha256_s390.c	2007-04-27 16:05:00.000000000 +0200
@@ -26,7 +26,7 @@
 #define SHA256_BLOCK_SIZE	64
 
 struct s390_sha256_ctx {
-	u64 count;
+	u64 count;		/* message length */
 	u32 state[8];
 	u8 buf[2 * SHA256_BLOCK_SIZE];
 };
@@ -54,10 +54,9 @@
 	int ret;
 
 	/* how much is already in the buffer? */
-	index = sctx->count / 8 & 0x3f;
+	index = sctx->count & 0x3f;
 
-	/* update message bit length */
-	sctx->count += len * 8;
+	sctx->count += len;
 
 	if ((index + len) < SHA256_BLOCK_SIZE)
 		goto store;
@@ -87,12 +86,17 @@
 		memcpy(sctx->buf + index , data, len);
 }
 
-static void pad_message(struct s390_sha256_ctx* sctx)
+/* Add padding and return the message digest */
+static void sha256_final(struct crypto_tfm *tfm, u8 *out)
 {
-	int index, end;
+	struct s390_sha256_ctx *sctx = crypto_tfm_ctx(tfm);
+	u64 bits;
+	unsigned int index, end;
+	int ret;
 
-	index = sctx->count / 8 & 0x3f;
-	end = index < 56 ? SHA256_BLOCK_SIZE : 2 * SHA256_BLOCK_SIZE;
+	/* must perform manual padding */
+	index = sctx->count & 0x3f;
+	end = (index < 56) ? SHA256_BLOCK_SIZE : (2 * SHA256_BLOCK_SIZE);
 
 	/* start pad with 1 */
 	sctx->buf[index] = 0x80;
@@ -102,21 +106,11 @@
 	memset(sctx->buf + index, 0x00, end - index - 8);
 
 	/* append message length */
-	memcpy(sctx->buf + end - 8, &sctx->count, sizeof sctx->count);
-
-	sctx->count = end * 8;
-}
-
-/* Add padding and return the message digest */
-static void sha256_final(struct crypto_tfm *tfm, u8 *out)
-{
-	struct s390_sha256_ctx *sctx = crypto_tfm_ctx(tfm);
-
-	/* must perform manual padding */
-	pad_message(sctx);
+	bits = sctx->count * 8;
+	memcpy(sctx->buf + end - 8, &bits, sizeof(bits));
 
-	crypt_s390_kimd(KIMD_SHA_256, sctx->state, sctx->buf,
-			sctx->count / 8);
+	ret = crypt_s390_kimd(KIMD_SHA_256, sctx->state, sctx->buf, end);
+	BUG_ON(ret != end);
 
 	/* copy digest to out */
 	memcpy(out, sctx->state, SHA256_DIGEST_SIZE);
Index: quilt-2.6/Documentation/s390/crypto/crypto-API.txt
===================================================================
--- quilt-2.6.orig/Documentation/s390/crypto/crypto-API.txt	2007-04-27 16:01:49.000000000 +0200
+++ /dev/null	1970-01-01 00:00:00.000000000 +0000
@@ -1,83 +0,0 @@
-crypto-API support for z990 Message Security Assist (MSA) instructions
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-AUTHOR:	Thomas Spatzier (tspat@de.ibm.com)
-
-
-1. Introduction crypto-API
-~~~~~~~~~~~~~~~~~~~~~~~~~~
-See Documentation/crypto/api-intro.txt for an introduction/description of the
-kernel crypto API.
-According to api-intro.txt support for z990 crypto instructions has been added
-in the algorithm api layer of the crypto API. Several files containing z990
-optimized implementations of crypto algorithms are placed in the
-arch/s390/crypto directory.
-
-
-2. Probing for availability of MSA
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-It should be possible to use Kernels with the z990 crypto implementations both
-on machines with MSA available and on those without MSA (pre z990 or z990
-without MSA). Therefore a simple probing mechanism has been implemented:
-In the init function of each crypto module the availability of MSA and of the
-respective crypto algorithm in particular will be tested. If the algorithm is
-available the module will load and register its algorithm with the crypto API.
-
-If the respective crypto algorithm is not available, the init function will
-return -ENOSYS. In that case a fallback to the standard software implementation
-of the crypto algorithm must be taken ( -> the standard crypto modules are
-also built when compiling the kernel).
-
-
-3. Ensuring z990 crypto module preference
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-If z990 crypto instructions are available the optimized modules should be
-preferred instead of standard modules.
-
-3.1. compiled-in modules
-~~~~~~~~~~~~~~~~~~~~~~~~
-For compiled-in modules it has to be ensured that the z990 modules are linked
-before the standard crypto modules. Then, on system startup the init functions
-of z990 crypto modules will be called first and query for availability of z990
-crypto instructions. If instruction is available, the z990 module will register
-its crypto algorithm implementation -> the load of the standard module will fail
-since the algorithm is already registered.
-If z990 crypto instruction is not available the load of the z990 module will
-fail -> the standard module will load and register its algorithm.
-
-3.2. dynamic modules
-~~~~~~~~~~~~~~~~~~~~
-A system administrator has to take care of giving preference to z990 crypto
-modules. If MSA is available appropriate lines have to be added to
-/etc/modprobe.conf.
-
-Example:	z990 crypto instruction for SHA1 algorithm is available
-
-		add the following line to /etc/modprobe.conf (assuming the
-		z990 crypto modules for SHA1 is called sha1_z990):
-
-		alias sha1 sha1_z990
-
-		-> when the sha1 algorithm is requested through the crypto API
-		(which has a module autoloader) the z990 module will be loaded.
-
-TBD:	a userspace module probing mechanism
-	something like 'probe sha1 sha1_z990 sha1' in modprobe.conf
-	-> try module sha1_z990, if it fails to load standard module sha1
-	the 'probe' statement is currently not supported in modprobe.conf
-
-
-4. Currently implemented z990 crypto algorithms
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-The following crypto algorithms with z990 MSA support are currently implemented.
-The name of each algorithm under which it is registered in crypto API and the
-name of the respective module is given in square brackets.
-
-- SHA1 Digest Algorithm [sha1 -> sha1_z990]
-- DES Encrypt/Decrypt Algorithm (64bit key) [des -> des_z990]
-- Triple DES Encrypt/Decrypt Algorithm (128bit key) [des3_ede128 -> des_z990]
-- Triple DES Encrypt/Decrypt Algorithm (192bit key) [des3_ede -> des_z990]
-
-In order to load, for example, the sha1_z990 module when the sha1 algorithm is
-requested (see 3.2.) add 'alias sha1 sha1_z990' to /etc/modprobe.conf.
-

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 32/38] vtime: cleanup per_cpu usage.

[Martin Schwidefsky]

[-- Attachment #1: 135-vtime-percpu.diff --]
[-- Type: text/plain, Size: 2728 bytes --]

From: Jan Glauber <jan.glauber@de.ibm.com>

Replace per_cpu(... , smp_processor_id()) with __get_cpu_var()

Signed-off-by: Jan Glauber <jan.glauber@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 arch/s390/kernel/vtime.c |   16 +++++++---------
 1 files changed, 7 insertions(+), 9 deletions(-)

Index: quilt-2.6/arch/s390/kernel/vtime.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/vtime.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/vtime.c	2007-04-27 16:05:00.000000000 +0200
@@ -128,7 +128,7 @@
 	S390_lowcore.last_update_timer = expires;
 
 	/* store expire time for this CPU timer */
-	per_cpu(virt_cpu_timer, smp_processor_id()).to_expire = expires;
+	__get_cpu_var(virt_cpu_timer).to_expire = expires;
 }
 #else
 static inline void set_vtimer(__u64 expires)
@@ -137,7 +137,7 @@
 	asm volatile ("SPT %0" : : "m" (S390_lowcore.last_update_timer));
 
 	/* store expire time for this CPU timer */
-	per_cpu(virt_cpu_timer, smp_processor_id()).to_expire = expires;
+	__get_cpu_var(virt_cpu_timer).to_expire = expires;
 }
 #endif
 
@@ -145,7 +145,7 @@
 {
 	struct vtimer_queue *vt_list;
 
-	vt_list = &per_cpu(virt_cpu_timer, smp_processor_id());
+	vt_list = &__get_cpu_var(virt_cpu_timer);
 
 	/* CPU timer interrupt is pending, don't reprogramm it */
 	if (vt_list->idle & 1LL<<63)
@@ -159,7 +159,7 @@
 {
 	struct vtimer_queue *vt_list;
 
-	vt_list = &per_cpu(virt_cpu_timer, smp_processor_id());
+	vt_list = &__get_cpu_var(virt_cpu_timer);
 
 	/* nothing to do */
 	if (list_empty(&vt_list->list)) {
@@ -219,7 +219,7 @@
 	if (list_empty(cb_list))
 		return;
 
-	vt_list = &per_cpu(virt_cpu_timer, smp_processor_id());
+	vt_list = &__get_cpu_var(virt_cpu_timer);
 
 	list_for_each_entry_safe(event, tmp, cb_list, entry) {
 		fn = event->function;
@@ -244,7 +244,6 @@
  */
 static void do_cpu_timer_interrupt(__u16 error_code)
 {
-	int cpu;
 	__u64 next, delta;
 	struct vtimer_queue *vt_list;
 	struct vtimer_list *event, *tmp;
@@ -253,8 +252,7 @@
 	struct list_head cb_list;
 
 	INIT_LIST_HEAD(&cb_list);
-	cpu = smp_processor_id();
-	vt_list = &per_cpu(virt_cpu_timer, cpu);
+	vt_list = &__get_cpu_var(virt_cpu_timer);
 
 	/* walk timer list, fire all expired events */
 	spin_lock(&vt_list->lock);
@@ -534,7 +532,7 @@
 	/* enable cpu timer interrupts */
 	__ctl_set_bit(0,10);
 
-	vt_list = &per_cpu(virt_cpu_timer, smp_processor_id());
+	vt_list = &__get_cpu_var(virt_cpu_timer);
 	INIT_LIST_HEAD(&vt_list->list);
 	spin_lock_init(&vt_list->lock);
 	vt_list->to_expire = 0;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 33/38] Processor degradation notification.

[Martin Schwidefsky]

[-- Attachment #1: 136-cpu-degradation.diff --]
[-- Type: text/plain, Size: 7044 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

Generate uevents for all cpus if cpu capability changes. This can
happen e.g. because the cpus are overheating. The cpu capability can
be read via /sys/devices/system/cpu/cpuN/capability.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/kernel/smp.c          |   48 ++++++++++++++++++++++---
 drivers/s390/char/Makefile      |    2 -
 drivers/s390/char/sclp.h        |    2 +
 drivers/s390/char/sclp_config.c |   75 ++++++++++++++++++++++++++++++++++++++++
 drivers/s390/sysinfo.c          |   18 +++++++++
 include/asm-s390/processor.h    |    1 
 6 files changed, 140 insertions(+), 6 deletions(-)

Index: quilt-2.6/arch/s390/kernel/smp.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/smp.c	2007-04-27 16:04:59.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/smp.c	2007-04-27 16:05:00.000000000 +0200
@@ -821,19 +821,57 @@
 
 static DEFINE_PER_CPU(struct cpu, cpu_devices);
 
+static ssize_t show_capability(struct sys_device *dev, char *buf)
+{
+	unsigned int capability;
+	int rc;
+
+	rc = get_cpu_capability(&capability);
+	if (rc)
+		return rc;
+	return sprintf(buf, "%u\n", capability);
+}
+static SYSDEV_ATTR(capability, 0444, show_capability, NULL);
+
+static int __cpuinit smp_cpu_notify(struct notifier_block *self,
+				    unsigned long action, void *hcpu)
+{
+	unsigned int cpu = (unsigned int)(long)hcpu;
+	struct cpu *c = &per_cpu(cpu_devices, cpu);
+	struct sys_device *s = &c->sysdev;
+
+	switch (action) {
+	case CPU_ONLINE:
+		if (sysdev_create_file(s, &attr_capability))
+			return NOTIFY_BAD;
+		break;
+	case CPU_DEAD:
+		sysdev_remove_file(s, &attr_capability);
+		break;
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block __cpuinitdata smp_cpu_nb = {
+	.notifier_call	= smp_cpu_notify,
+};
+
 static int __init topology_init(void)
 {
 	int cpu;
-	int ret;
+
+	register_cpu_notifier(&smp_cpu_nb);
 
 	for_each_possible_cpu(cpu) {
 		struct cpu *c = &per_cpu(cpu_devices, cpu);
+		struct sys_device *s = &c->sysdev;
 
 		c->hotpluggable = 1;
-		ret = register_cpu(c, cpu);
-		if (ret)
-			printk(KERN_WARNING "topology_init: register_cpu %d "
-			       "failed (%d)\n", cpu, ret);
+		register_cpu(c, cpu);
+		if (!cpu_online(cpu))
+			continue;
+		s = &c->sysdev;
+		sysdev_create_file(s, &attr_capability);
 	}
 	return 0;
 }
Index: quilt-2.6/drivers/s390/char/Makefile
===================================================================
--- quilt-2.6.orig/drivers/s390/char/Makefile	2007-04-27 16:04:59.000000000 +0200
+++ quilt-2.6/drivers/s390/char/Makefile	2007-04-27 16:05:00.000000000 +0200
@@ -3,7 +3,7 @@
 #
 
 obj-y += ctrlchar.o keyboard.o defkeymap.o sclp.o sclp_rw.o sclp_quiesce.o \
-	 sclp_info.o sclp_chp.o
+	 sclp_info.o sclp_config.o sclp_chp.o
 
 obj-$(CONFIG_TN3270) += raw3270.o
 obj-$(CONFIG_TN3270_CONSOLE) += con3270.o
Index: quilt-2.6/drivers/s390/char/sclp_config.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ quilt-2.6/drivers/s390/char/sclp_config.c	2007-04-27 16:05:00.000000000 +0200
@@ -0,0 +1,75 @@
+/*
+ *  drivers/s390/char/sclp_config.c
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ */
+
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/cpu.h>
+#include <linux/sysdev.h>
+#include <linux/workqueue.h>
+#include "sclp.h"
+
+#define TAG	"sclp_config: "
+
+struct conf_mgm_data {
+	u8 reserved;
+	u8 ev_qualifier;
+} __attribute__((packed));
+
+#define EV_QUAL_CAP_CHANGE	3
+
+static struct work_struct sclp_cpu_capability_work;
+
+static void sclp_cpu_capability_notify(struct work_struct *work)
+{
+	int cpu;
+	struct sys_device *sysdev;
+
+	printk(KERN_WARNING TAG "cpu capability changed.\n");
+	lock_cpu_hotplug();
+	for_each_online_cpu(cpu) {
+		sysdev = get_cpu_sysdev(cpu);
+		kobject_uevent(&sysdev->kobj, KOBJ_CHANGE);
+	}
+	unlock_cpu_hotplug();
+}
+
+static void sclp_conf_receiver_fn(struct evbuf_header *evbuf)
+{
+	struct conf_mgm_data *cdata;
+
+	cdata = (struct conf_mgm_data *)(evbuf + 1);
+	if (cdata->ev_qualifier == EV_QUAL_CAP_CHANGE)
+		schedule_work(&sclp_cpu_capability_work);
+}
+
+static struct sclp_register sclp_conf_register =
+{
+	.receive_mask = EVTYP_CONFMGMDATA_MASK,
+	.receiver_fn  = sclp_conf_receiver_fn,
+};
+
+static int __init sclp_conf_init(void)
+{
+	int rc;
+
+	INIT_WORK(&sclp_cpu_capability_work, sclp_cpu_capability_notify);
+
+	rc = sclp_register(&sclp_conf_register);
+	if (rc) {
+		printk(KERN_ERR TAG "failed to register (%d).\n", rc);
+		return rc;
+	}
+
+	if (!(sclp_conf_register.sclp_receive_mask & EVTYP_CONFMGMDATA_MASK)) {
+		printk(KERN_WARNING TAG "no configuration management.\n");
+		sclp_unregister(&sclp_conf_register);
+		rc = -ENOSYS;
+	}
+	return rc;
+}
+
+__initcall(sclp_conf_init);
Index: quilt-2.6/drivers/s390/char/sclp.h
===================================================================
--- quilt-2.6.orig/drivers/s390/char/sclp.h	2007-04-27 16:05:00.000000000 +0200
+++ quilt-2.6/drivers/s390/char/sclp.h	2007-04-27 16:05:00.000000000 +0200
@@ -27,6 +27,7 @@
 #define EVTYP_CNTLPROGIDENT	0x0B
 #define EVTYP_SIGQUIESCE	0x1D
 #define EVTYP_VT220MSG		0x1A
+#define EVTYP_CONFMGMDATA	0x04
 #define EVTYP_SDIAS		0x1C
 
 #define EVTYP_OPCMD_MASK	0x80000000
@@ -37,6 +38,7 @@
 #define EVTYP_CTLPROGIDENT_MASK	0x00200000
 #define EVTYP_SIGQUIESCE_MASK	0x00000008
 #define EVTYP_VT220MSG_MASK	0x00000040
+#define EVTYP_CONFMGMDATA_MASK	0x10000000
 #define EVTYP_SDIAS_MASK	0x00000010
 
 #define GNRLMSGFLGS_DOM		0x8000
Index: quilt-2.6/drivers/s390/sysinfo.c
===================================================================
--- quilt-2.6.orig/drivers/s390/sysinfo.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/sysinfo.c	2007-04-27 16:05:00.000000000 +0200
@@ -357,6 +357,24 @@
 
 __initcall(create_proc_sysinfo);
 
+int get_cpu_capability(unsigned int *capability)
+{
+	struct sysinfo_1_2_2 *info;
+	int rc;
+
+	info = (void *) get_zeroed_page(GFP_KERNEL);
+	if (!info)
+		return -ENOMEM;
+	rc = stsi(info, 1, 2, 2);
+	if (rc == -ENOSYS)
+		goto out;
+	rc = 0;
+	*capability = info->capability;
+out:
+	free_page((unsigned long) info);
+	return rc;
+}
+
 /*
  * CPU capability might have changed. Therefore recalculate loops_per_jiffy.
  */
Index: quilt-2.6/include/asm-s390/processor.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/processor.h	2007-04-27 16:04:56.000000000 +0200
+++ quilt-2.6/include/asm-s390/processor.h	2007-04-27 16:05:00.000000000 +0200
@@ -57,6 +57,7 @@
 
 extern void s390_adjust_jiffies(void);
 extern void print_cpu_info(struct cpuinfo_S390 *);
+extern int get_cpu_capability(unsigned int *);
 
 /* Lazy FPU handling on uni-processor */
 extern struct task_struct *last_task_used_math;

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 34/38] split page_test_and_clear_dirty.

[Martin Schwidefsky]

[-- Attachment #1: 137-test-clear-dirty.diff --]
[-- Type: text/plain, Size: 4300 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

The page_test_and_clear_dirty primitive really consists of two
operations, page_test_dirty and the page_clear_dirty. The combination
of the two is not an atomic operation, so it makes more sense to have
two separate operations instead of one.
In addition to the improved readability of the s390 version of
SetPageUptodate, it now avoids the page_test_dirty operation which is
an insert-storage-key-extended (iske) instruction which is an expensive
operation.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 include/asm-generic/pgtable.h |   11 +++++++++--
 include/asm-s390/pgtable.h    |   15 ++++++++-------
 include/linux/page-flags.h    |    2 +-
 mm/rmap.c                     |    8 ++++++--
 4 files changed, 24 insertions(+), 12 deletions(-)

Index: quilt-2.6/include/asm-generic/pgtable.h
===================================================================
--- quilt-2.6.orig/include/asm-generic/pgtable.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-generic/pgtable.h	2007-04-27 16:05:01.000000000 +0200
@@ -139,8 +139,15 @@
 #define pte_same(A,B)	(pte_val(A) == pte_val(B))
 #endif
 
-#ifndef __HAVE_ARCH_PAGE_TEST_AND_CLEAR_DIRTY
-#define page_test_and_clear_dirty(page) (0)
+#ifndef __HAVE_ARCH_PAGE_TEST_DIRTY
+#define page_test_dirty(page)		(0)
+#endif
+
+#ifndef __HAVE_ARCH_PAGE_CLEAR_DIRTY
+#define page_clear_dirty(page)		do { } while (0)
+#endif
+
+#ifndef __HAVE_ARCH_PAGE_TEST_DIRTY
 #define pte_maybe_dirty(pte)		pte_dirty(pte)
 #else
 #define pte_maybe_dirty(pte)		(1)
Index: quilt-2.6/include/asm-s390/pgtable.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/pgtable.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/asm-s390/pgtable.h	2007-04-27 16:05:01.000000000 +0200
@@ -753,14 +753,14 @@
  * should therefore only be called if it is not mapped in any
  * address space.
  */
-static inline int page_test_and_clear_dirty(struct page *page)
+static inline int page_test_dirty(struct page *page)
 {
-	unsigned long physpage = page_to_phys(page);
-	int skey = page_get_storage_key(physpage);
+	return (page_get_storage_key(page_to_phys(page)) & _PAGE_CHANGED) != 0;
+}
 
-	if (skey & _PAGE_CHANGED)
-		page_set_storage_key(physpage, skey & ~_PAGE_CHANGED);
-	return skey & _PAGE_CHANGED;
+static inline void page_clear_dirty(struct page *page)
+{
+	page_set_storage_key(page_to_phys(page), PAGE_DEFAULT_KEY);
 }
 
 /*
@@ -953,7 +953,8 @@
 #define __HAVE_ARCH_PTEP_CLEAR_FLUSH
 #define __HAVE_ARCH_PTEP_SET_WRPROTECT
 #define __HAVE_ARCH_PTE_SAME
-#define __HAVE_ARCH_PAGE_TEST_AND_CLEAR_DIRTY
+#define __HAVE_ARCH_PAGE_TEST_DIRTY
+#define __HAVE_ARCH_PAGE_CLEAR_DIRTY
 #define __HAVE_ARCH_PAGE_TEST_AND_CLEAR_YOUNG
 #include <asm-generic/pgtable.h>
 
Index: quilt-2.6/include/linux/page-flags.h
===================================================================
--- quilt-2.6.orig/include/linux/page-flags.h	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/include/linux/page-flags.h	2007-04-27 16:05:01.000000000 +0200
@@ -133,7 +133,7 @@
 static inline void SetPageUptodate(struct page *page)
 {
 	if (!test_and_set_bit(PG_uptodate, &page->flags))
-		page_test_and_clear_dirty(page);
+		page_clear_dirty(page);
 }
 #else
 #define SetPageUptodate(page)	set_bit(PG_uptodate, &(page)->flags)
Index: quilt-2.6/mm/rmap.c
===================================================================
--- quilt-2.6.orig/mm/rmap.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/mm/rmap.c	2007-04-27 16:05:01.000000000 +0200
@@ -498,8 +498,10 @@
 		struct address_space *mapping = page_mapping(page);
 		if (mapping)
 			ret = page_mkclean_file(mapping, page);
-		if (page_test_and_clear_dirty(page))
+		if (page_test_dirty(page)) {
+			page_clear_dirty(page);
 			ret = 1;
+		}
 	}
 
 	return ret;
@@ -605,8 +607,10 @@
 		 * Leaving it set also helps swapoff to reinstate ptes
 		 * faster for those pages still in swapcache.
 		 */
-		if (page_test_and_clear_dirty(page))
+		if (page_test_dirty(page)) {
+			page_clear_dirty(page);
 			set_page_dirty(page);
+		}
 		__dec_zone_page_state(page,
 				PageAnon(page) ? NR_ANON_PAGES : NR_FILE_MAPPED);
 	}

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 35/38] Switch etr from tasklet to workqueue.

[Martin Schwidefsky]

[-- Attachment #1: 138-etr-work.diff --]
[-- Type: text/plain, Size: 4331 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

The clock synchronization of the ETR code requires an smp_call_function
to synchronize all cpus. Calling smp_call_function from a tasklet is
illegal. Replace the tasklet with a job on the global workqueue.
ETR work is rare and can be postponed to a be done by a kernel thread.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 arch/s390/kernel/time.c |   34 +++++++++++++++++-----------------
 1 files changed, 17 insertions(+), 17 deletions(-)

Index: quilt-2.6/arch/s390/kernel/time.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/time.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/time.c	2007-04-27 16:05:01.000000000 +0200
@@ -280,7 +280,6 @@
 }
 
 static void etr_reset(void);
-static void etr_init(void);
 static void etr_ext_handler(__u16);
 
 /*
@@ -355,7 +354,6 @@
 #ifdef CONFIG_VIRT_TIMER
 	vtime_init();
 #endif
-	etr_init();
 }
 
 /*
@@ -426,11 +424,11 @@
 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);
+static void etr_work_fn(struct work_struct *work);
+static DECLARE_WORK(etr_work, etr_work_fn);
 
 /*
  * The etr get_clock function. It will write the current clock value
@@ -507,29 +505,31 @@
 	}
 }
 
-static void etr_init(void)
+static int __init etr_init(void)
 {
 	struct etr_aib aib;
 
 	if (test_bit(ETR_FLAG_ENOSYS, &etr_flags))
-		return;
+		return 0;
 	/* 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);
+		schedule_work(&etr_work);
 	}
 	if (etr_port1_online) {
 		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
-		tasklet_hi_schedule(&etr_tasklet);
+		schedule_work(&etr_work);
 	}
+	return 0;
 }
 
+arch_initcall(etr_init);
+
 /*
  * Two sorts of ETR machine checks. The architecture reads:
  * "When a machine-check niterruption occurs and if a switch-to-local or
@@ -549,7 +549,7 @@
 		return;
 	etr_disable_sync_clock(NULL);
 	set_bit(ETR_EVENT_SWITCH_LOCAL, &etr_events);
-	tasklet_hi_schedule(&etr_tasklet);
+	schedule_work(&etr_work);
 }
 
 /*
@@ -564,7 +564,7 @@
 		return;
 	etr_disable_sync_clock(NULL);
 	set_bit(ETR_EVENT_SYNC_CHECK, &etr_events);
-	tasklet_hi_schedule(&etr_tasklet);
+	schedule_work(&etr_work);
 }
 
 /*
@@ -591,13 +591,13 @@
 		 * Both ports are not up-to-date now.
 		 */
 		set_bit(ETR_EVENT_PORT_ALERT, &etr_events);
-	tasklet_hi_schedule(&etr_tasklet);
+	schedule_work(&etr_work);
 }
 
 static void etr_timeout(unsigned long dummy)
 {
 	set_bit(ETR_EVENT_UPDATE, &etr_events);
-	tasklet_hi_schedule(&etr_tasklet);
+	schedule_work(&etr_work);
 }
 
 /*
@@ -927,7 +927,7 @@
 	if (!eacr.e0 && !eacr.e1)
 		return eacr;
 
-	/* Update port0 or port1 with aib stored in etr_tasklet_fn. */
+	/* Update port0 or port1 with aib stored in etr_work_fn. */
 	if (aib->esw.q == 0) {
 		/* Information for port 0 stored. */
 		if (eacr.p0 && !etr_port0_uptodate) {
@@ -1007,7 +1007,7 @@
  * 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)
+static void etr_work_fn(struct work_struct *work)
 {
 	unsigned long long now;
 	struct etr_eacr eacr;
@@ -1220,13 +1220,13 @@
 			return count;	/* Nothing to do. */
 		etr_port0_online = value;
 		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
-		tasklet_hi_schedule(&etr_tasklet);
+		schedule_work(&etr_work);
 	} 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);
+		schedule_work(&etr_work);
 	}
 	return count;
 }

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 36/38] Remove debugging junk.

[Martin Schwidefsky]

[-- Attachment #1: 139-appldata.diff --]
[-- Type: text/plain, Size: 1889 bytes --]

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

arch/s390/appldata/appldata_base.c has some confusing debugging code left
over to allow compiling it as a module. In practice, it cannot be configured
as module and there is no need to keep that code.

Signed-off-by: Gerald Schaefer <geraldsc@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 arch/s390/appldata/appldata_base.c |   38 -------------------------------------
 1 files changed, 38 deletions(-)

Index: quilt-2.6/arch/s390/appldata/appldata_base.c
===================================================================
--- quilt-2.6.orig/arch/s390/appldata/appldata_base.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/arch/s390/appldata/appldata_base.c	2007-04-27 16:05:01.000000000 +0200
@@ -668,45 +668,7 @@
 EXPORT_SYMBOL_GPL(appldata_unregister_ops);
 EXPORT_SYMBOL_GPL(appldata_diag);
 
-#ifdef MODULE
-/*
- * Kernel symbols needed by appldata_mem and appldata_os modules.
- * However, if this file is compiled as a module (for testing only), these
- * symbols are not exported. In this case, we define them locally and export
- * those.
- */
-void si_swapinfo(struct sysinfo *val)
-{
-	val->freeswap = -1ul;
-	val->totalswap = -1ul;
-}
-
-unsigned long avenrun[3] = {-1 - FIXED_1/200, -1 - FIXED_1/200,
-				-1 - FIXED_1/200};
-int nr_threads = -1;
-
-void get_full_page_state(struct page_state *ps)
-{
-	memset(ps, -1, sizeof(struct page_state));
-}
-
-unsigned long nr_running(void)
-{
-	return -1;
-}
-
-unsigned long nr_iowait(void)
-{
-	return -1;
-}
-
-/*unsigned long nr_context_switches(void)
-{
-	return -1;
-}*/
-#endif /* MODULE */
 EXPORT_SYMBOL_GPL(si_swapinfo);
 EXPORT_SYMBOL_GPL(nr_threads);
 EXPORT_SYMBOL_GPL(nr_running);
 EXPORT_SYMBOL_GPL(nr_iowait);
-//EXPORT_SYMBOL_GPL(nr_context_switches);

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 37/38] Clean up smp code in preparation for some larger changes.

[Martin Schwidefsky]

[-- Attachment #1: 140-smp-cleanup.diff --]
[-- Type: text/plain, Size: 16579 bytes --]

From: Heiko Carstens <heiko.carstens@de.ibm.com>

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---

 arch/s390/kernel/smp.c |  259 ++++++++++++++++++-------------------------------
 include/asm-s390/smp.h |    5 
 2 files changed, 99 insertions(+), 165 deletions(-)

Index: quilt-2.6/arch/s390/kernel/smp.c
===================================================================
--- quilt-2.6.orig/arch/s390/kernel/smp.c	2007-04-27 16:05:00.000000000 +0200
+++ quilt-2.6/arch/s390/kernel/smp.c	2007-04-27 16:05:02.000000000 +0200
@@ -1,12 +1,12 @@
 /*
  *  arch/s390/kernel/smp.c
  *
- *    Copyright (C) IBM Corp. 1999,2006
+ *    Copyright IBM Corp. 1999,2007
  *    Author(s): Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
- *               Martin Schwidefsky (schwidefsky@de.ibm.com)
- *               Heiko Carstens (heiko.carstens@de.ibm.com)
+ *		 Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *		 Heiko Carstens (heiko.carstens@de.ibm.com)
  *
- *  based on other smp stuff by 
+ *  based on other smp stuff by
  *    (c) 1995 Alan Cox, CymruNET Ltd  <alan@cymru.net>
  *    (c) 1998 Ingo Molnar
  *
@@ -43,16 +43,17 @@
 #include <asm/timer.h>
 #include <asm/lowcore.h>
 
-extern volatile int __cpu_logical_map[];
-
 /*
  * An array with a pointer the lowcore of every CPU.
  */
-
 struct _lowcore *lowcore_ptr[NR_CPUS];
+EXPORT_SYMBOL(lowcore_ptr);
 
 cpumask_t cpu_online_map = CPU_MASK_NONE;
+EXPORT_SYMBOL(cpu_online_map);
+
 cpumask_t cpu_possible_map = CPU_MASK_NONE;
+EXPORT_SYMBOL(cpu_possible_map);
 
 static struct task_struct *current_set[NR_CPUS];
 
@@ -72,7 +73,7 @@
 	int wait;
 };
 
-static struct call_data_struct * call_data;
+static struct call_data_struct *call_data;
 
 /*
  * 'Call function' interrupt callback
@@ -152,8 +153,8 @@
  *
  * Run a function on all other CPUs.
  *
- * You must not call this function with disabled interrupts or from a
- * hardware interrupt handler. You may call it from a bottom half.
+ * You must not call this function with disabled interrupts, from a
+ * hardware interrupt handler or from a bottom half.
  */
 int smp_call_function(void (*func) (void *info), void *info, int nonatomic,
 		      int wait)
@@ -179,11 +180,11 @@
  *
  * Run a function on one processor.
  *
- * You must not call this function with disabled interrupts or from a
- * hardware interrupt handler. You may call it from a bottom half.
+ * You must not call this function with disabled interrupts, from a
+ * hardware interrupt handler or from a bottom half.
  */
 int smp_call_function_on(void (*func) (void *info), void *info, int nonatomic,
-			  int wait, int cpu)
+			 int wait, int cpu)
 {
 	cpumask_t map = CPU_MASK_NONE;
 
@@ -197,9 +198,9 @@
 
 static void do_send_stop(void)
 {
-        int cpu, rc;
+	int cpu, rc;
 
-        /* stop all processors */
+	/* stop all processors */
 	for_each_online_cpu(cpu) {
 		if (cpu == smp_processor_id())
 			continue;
@@ -211,9 +212,9 @@
 
 static void do_store_status(void)
 {
-        int cpu, rc;
+	int cpu, rc;
 
-        /* store status of all processors in their lowcores (real 0) */
+	/* store status of all processors in their lowcores (real 0) */
 	for_each_online_cpu(cpu) {
 		if (cpu == smp_processor_id())
 			continue;
@@ -221,8 +222,8 @@
 			rc = signal_processor_p(
 				(__u32)(unsigned long) lowcore_ptr[cpu], cpu,
 				sigp_store_status_at_address);
-		} while(rc == sigp_busy);
-        }
+		} while (rc == sigp_busy);
+	}
 }
 
 static void do_wait_for_stop(void)
@@ -233,7 +234,7 @@
 	for_each_online_cpu(cpu) {
 		if (cpu == smp_processor_id())
 			continue;
-		while(!smp_cpu_not_running(cpu))
+		while (!smp_cpu_not_running(cpu))
 			cpu_relax();
 	}
 }
@@ -247,7 +248,7 @@
 	/* Disable all interrupts/machine checks */
 	__load_psw_mask(psw_kernel_bits & ~PSW_MASK_MCHECK);
 
-        /* write magic number to zero page (absolute 0) */
+	/* write magic number to zero page (absolute 0) */
 	lowcore_ptr[smp_processor_id()]->panic_magic = __PANIC_MAGIC;
 
 	/* stop other processors. */
@@ -263,8 +264,7 @@
 /*
  * Reboot, halt and power_off routines for SMP.
  */
-
-void machine_restart_smp(char * __unused) 
+void machine_restart_smp(char *__unused)
 {
 	smp_send_stop();
 	do_reipl();
@@ -295,17 +295,17 @@
 
 static void do_ext_call_interrupt(__u16 code)
 {
-        unsigned long bits;
+	unsigned long bits;
 
-        /*
-         * handle bit signal external calls
-         *
-         * For the ec_schedule signal we have to do nothing. All the work
-         * is done automatically when we return from the interrupt.
-         */
+	/*
+	 * handle bit signal external calls
+	 *
+	 * For the ec_schedule signal we have to do nothing. All the work
+	 * is done automatically when we return from the interrupt.
+	 */
 	bits = xchg(&S390_lowcore.ext_call_fast, 0);
 
-	if (test_bit(ec_call_function, &bits)) 
+	if (test_bit(ec_call_function, &bits))
 		do_call_function();
 }
 
@@ -315,11 +315,11 @@
  */
 static void smp_ext_bitcall(int cpu, ec_bit_sig sig)
 {
-        /*
-         * Set signaling bit in lowcore of target cpu and kick it
-         */
+	/*
+	 * Set signaling bit in lowcore of target cpu and kick it
+	 */
 	set_bit(sig, (unsigned long *) &lowcore_ptr[cpu]->ext_call_fast);
-	while(signal_processor(cpu, sigp_emergency_signal) == sigp_busy)
+	while (signal_processor(cpu, sigp_emergency_signal) == sigp_busy)
 		udelay(10);
 }
 
@@ -334,7 +334,7 @@
 
 void smp_ptlb_all(void)
 {
-        on_each_cpu(smp_ptlb_callback, NULL, 0, 1);
+	on_each_cpu(smp_ptlb_callback, NULL, 0, 1);
 }
 EXPORT_SYMBOL(smp_ptlb_all);
 #endif /* ! CONFIG_64BIT */
@@ -346,7 +346,7 @@
  */
 void smp_send_reschedule(int cpu)
 {
-        smp_ext_bitcall(cpu, ec_schedule);
+	smp_ext_bitcall(cpu, ec_schedule);
 }
 
 /*
@@ -360,11 +360,12 @@
 /*
  * callback for setting/clearing control bits
  */
-static 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;
-	
+
 	__ctl_store(cregs, 0, 15);
 	for (i = 0; i <= 15; i++)
 		cregs[i] = (cregs[i] & pp->andvals[i]) | pp->orvals[i];
@@ -383,6 +384,7 @@
 	parms.orvals[cr] = 1 << bit;
 	on_each_cpu(smp_ctl_bit_callback, &parms, 0, 1);
 }
+EXPORT_SYMBOL(smp_ctl_set_bit);
 
 /*
  * Clear a bit in a control register of all cpus
@@ -396,6 +398,7 @@
 	parms.andvals[cr] = ~(1L << bit);
 	on_each_cpu(smp_ctl_bit_callback, &parms, 0, 1);
 }
+EXPORT_SYMBOL(smp_ctl_clear_bit);
 
 #if defined(CONFIG_ZFCPDUMP) || defined(CONFIG_ZFCPDUMP_MODULE)
 
@@ -460,8 +463,7 @@
  * Lets check how many CPUs we have.
  */
 
-static unsigned int
-__init smp_count_cpus(void)
+static unsigned int __init smp_count_cpus(void)
 {
 	unsigned int cpu, num_cpus;
 	__u16 boot_cpu_addr;
@@ -477,31 +479,30 @@
 		if ((__u16) cpu == boot_cpu_addr)
 			continue;
 		__cpu_logical_map[1] = (__u16) cpu;
-		if (signal_processor(1, sigp_sense) ==
-		    sigp_not_operational)
+		if (signal_processor(1, sigp_sense) == sigp_not_operational)
 			continue;
 		num_cpus++;
 	}
 
-	printk("Detected %d CPU's\n",(int) num_cpus);
+	printk("Detected %d CPU's\n", (int) num_cpus);
 	printk("Boot cpu address %2X\n", boot_cpu_addr);
 
 	return num_cpus;
 }
 
 /*
- *      Activate a secondary processor.
+ *	Activate a secondary processor.
  */
 int __devinit start_secondary(void *cpuvoid)
 {
-        /* Setup the cpu */
-        cpu_init();
+	/* Setup the cpu */
+	cpu_init();
 	preempt_disable();
 	/* Enable TOD clock interrupts on the secondary cpu. */
-        init_cpu_timer();
+	init_cpu_timer();
 #ifdef CONFIG_VIRT_TIMER
 	/* Enable cpu timer interrupts on the secondary cpu. */
-        init_cpu_vtimer();
+	init_cpu_vtimer();
 #endif
 	/* Enable pfault pseudo page faults on this cpu. */
 	pfault_init();
@@ -510,11 +511,11 @@
 	cpu_set(smp_processor_id(), cpu_online_map);
 	/* Switch on interrupts */
 	local_irq_enable();
-        /* Print info about this processor */
-        print_cpu_info(&S390_lowcore.cpu_data);
-        /* cpu_idle will call schedule for us */
-        cpu_idle();
-        return 0;
+	/* Print info about this processor */
+	print_cpu_info(&S390_lowcore.cpu_data);
+	/* cpu_idle will call schedule for us */
+	cpu_idle();
+	return 0;
 }
 
 static void __init smp_create_idle(unsigned int cpu)
@@ -531,56 +532,13 @@
 	current_set[cpu] = p;
 }
 
-/* Reserving and releasing of CPUs */
-
-static DEFINE_SPINLOCK(smp_reserve_lock);
-static int smp_cpu_reserved[NR_CPUS];
-
-int
-smp_get_cpu(cpumask_t cpu_mask)
-{
-	unsigned long flags;
-	int cpu;
-
-	spin_lock_irqsave(&smp_reserve_lock, flags);
-	/* Try to find an already reserved cpu. */
-	for_each_cpu_mask(cpu, cpu_mask) {
-		if (smp_cpu_reserved[cpu] != 0) {
-			smp_cpu_reserved[cpu]++;
-			/* Found one. */
-			goto out;
-		}
-	}
-	/* Reserve a new cpu from cpu_mask. */
-	for_each_cpu_mask(cpu, cpu_mask) {
-		if (cpu_online(cpu)) {
-			smp_cpu_reserved[cpu]++;
-			goto out;
-		}
-	}
-	cpu = -ENODEV;
-out:
-	spin_unlock_irqrestore(&smp_reserve_lock, flags);
-	return cpu;
-}
-
-void
-smp_put_cpu(int cpu)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&smp_reserve_lock, flags);
-	smp_cpu_reserved[cpu]--;
-	spin_unlock_irqrestore(&smp_reserve_lock, flags);
-}
-
-static int
-cpu_stopped(int cpu)
+static int cpu_stopped(int cpu)
 {
 	__u32 status;
 
 	/* Check for stopped state */
-	if (signal_processor_ps(&status, 0, cpu, sigp_sense) == sigp_status_stored) {
+	if (signal_processor_ps(&status, 0, cpu, sigp_sense) ==
+	    sigp_status_stored) {
 		if (status & 0x40)
 			return 1;
 	}
@@ -589,14 +547,13 @@
 
 /* Upping and downing of CPUs */
 
-int
-__cpu_up(unsigned int cpu)
+int __cpu_up(unsigned int cpu)
 {
 	struct task_struct *idle;
-        struct _lowcore    *cpu_lowcore;
+	struct _lowcore *cpu_lowcore;
 	struct stack_frame *sf;
-        sigp_ccode          ccode;
-	int                 curr_cpu;
+	sigp_ccode ccode;
+	int curr_cpu;
 
 	for (curr_cpu = 0; curr_cpu <= 65535; curr_cpu++) {
 		__cpu_logical_map[cpu] = (__u16) curr_cpu;
@@ -609,7 +566,7 @@
 
 	ccode = signal_processor_p((__u32)(unsigned long)(lowcore_ptr[cpu]),
 				   cpu, sigp_set_prefix);
-	if (ccode){
+	if (ccode) {
 		printk("sigp_set_prefix failed for cpu %d "
 		       "with condition code %d\n",
 		       (int) cpu, (int) ccode);
@@ -617,9 +574,9 @@
 	}
 
 	idle = current_set[cpu];
-        cpu_lowcore = lowcore_ptr[cpu];
+	cpu_lowcore = lowcore_ptr[cpu];
 	cpu_lowcore->kernel_stack = (unsigned long)
-		task_stack_page(idle) + (THREAD_SIZE);
+		task_stack_page(idle) + THREAD_SIZE;
 	sf = (struct stack_frame *) (cpu_lowcore->kernel_stack
 				     - sizeof(struct pt_regs)
 				     - sizeof(struct stack_frame));
@@ -631,11 +588,11 @@
 		"	stam	0,15,0(%0)"
 		: : "a" (&cpu_lowcore->access_regs_save_area) : "memory");
 	cpu_lowcore->percpu_offset = __per_cpu_offset[cpu];
-        cpu_lowcore->current_task = (unsigned long) idle;
-        cpu_lowcore->cpu_data.cpu_nr = cpu;
+	cpu_lowcore->current_task = (unsigned long) idle;
+	cpu_lowcore->cpu_data.cpu_nr = cpu;
 	eieio();
 
-	while (signal_processor(cpu,sigp_restart) == sigp_busy)
+	while (signal_processor(cpu, sigp_restart) == sigp_busy)
 		udelay(10);
 
 	while (!cpu_online(cpu))
@@ -682,18 +639,11 @@
 }
 early_param("possible_cpus", setup_possible_cpus);
 
-int
-__cpu_disable(void)
+int __cpu_disable(void)
 {
-	unsigned long flags;
 	struct ec_creg_mask_parms cr_parms;
 	int cpu = smp_processor_id();
 
-	spin_lock_irqsave(&smp_reserve_lock, flags);
-	if (smp_cpu_reserved[cpu] != 0) {
-		spin_unlock_irqrestore(&smp_reserve_lock, flags);
-		return -EBUSY;
-	}
 	cpu_clear(cpu, cpu_online_map);
 
 	/* Disable pfault pseudo page faults on this cpu. */
@@ -704,24 +654,23 @@
 
 	/* disable all external interrupts */
 	cr_parms.orvals[0] = 0;
-	cr_parms.andvals[0] = ~(1<<15 | 1<<14 | 1<<13 | 1<<12 |
-				1<<11 | 1<<10 | 1<< 6 | 1<< 4);
+	cr_parms.andvals[0] = ~(1 << 15 | 1 << 14 | 1 << 13 | 1 << 12 |
+				1 << 11 | 1 << 10 | 1 <<  6 | 1 <<  4);
 	/* disable all I/O interrupts */
 	cr_parms.orvals[6] = 0;
-	cr_parms.andvals[6] = ~(1<<31 | 1<<30 | 1<<29 | 1<<28 |
-				1<<27 | 1<<26 | 1<<25 | 1<<24);
+	cr_parms.andvals[6] = ~(1 << 31 | 1 << 30 | 1 << 29 | 1 << 28 |
+				1 << 27 | 1 << 26 | 1 << 25 | 1 << 24);
 	/* disable most machine checks */
 	cr_parms.orvals[14] = 0;
-	cr_parms.andvals[14] = ~(1<<28 | 1<<27 | 1<<26 | 1<<25 | 1<<24);
+	cr_parms.andvals[14] = ~(1 << 28 | 1 << 27 | 1 << 26 |
+				 1 << 25 | 1 << 24);
 
 	smp_ctl_bit_callback(&cr_parms);
 
-	spin_unlock_irqrestore(&smp_reserve_lock, flags);
 	return 0;
 }
 
-void
-__cpu_die(unsigned int cpu)
+void __cpu_die(unsigned int cpu)
 {
 	/* Wait until target cpu is down */
 	while (!smp_cpu_not_running(cpu))
@@ -729,13 +678,12 @@
 	printk("Processor %d spun down\n", cpu);
 }
 
-void
-cpu_die(void)
+void cpu_die(void)
 {
 	idle_task_exit();
 	signal_processor(smp_processor_id(), sigp_stop);
 	BUG();
-	for(;;);
+	for (;;);
 }
 
 #endif /* CONFIG_HOTPLUG_CPU */
@@ -748,36 +696,36 @@
 {
 	unsigned long stack;
 	unsigned int cpu;
-        int i;
+	int i;
 
-        /* request the 0x1201 emergency signal external interrupt */
-        if (register_external_interrupt(0x1201, do_ext_call_interrupt) != 0)
-                panic("Couldn't request external interrupt 0x1201");
-        memset(lowcore_ptr,0,sizeof(lowcore_ptr));  
-        /*
-         *  Initialize prefix pages and stacks for all possible cpus
-         */
+	/* request the 0x1201 emergency signal external interrupt */
+	if (register_external_interrupt(0x1201, do_ext_call_interrupt) != 0)
+		panic("Couldn't request external interrupt 0x1201");
+	memset(lowcore_ptr, 0, sizeof(lowcore_ptr));
+	/*
+	 *  Initialize prefix pages and stacks for all possible cpus
+	 */
 	print_cpu_info(&S390_lowcore.cpu_data);
 
-        for_each_possible_cpu(i) {
+	for_each_possible_cpu(i) {
 		lowcore_ptr[i] = (struct _lowcore *)
-			__get_free_pages(GFP_KERNEL|GFP_DMA, 
-					sizeof(void*) == 8 ? 1 : 0);
-		stack = __get_free_pages(GFP_KERNEL,ASYNC_ORDER);
-		if (lowcore_ptr[i] == NULL || stack == 0ULL)
+			__get_free_pages(GFP_KERNEL | GFP_DMA,
+					 sizeof(void*) == 8 ? 1 : 0);
+		stack = __get_free_pages(GFP_KERNEL, ASYNC_ORDER);
+		if (!lowcore_ptr[i] || !stack)
 			panic("smp_boot_cpus failed to allocate memory\n");
 
 		*(lowcore_ptr[i]) = S390_lowcore;
-		lowcore_ptr[i]->async_stack = stack + (ASYNC_SIZE);
-		stack = __get_free_pages(GFP_KERNEL,0);
-		if (stack == 0ULL)
+		lowcore_ptr[i]->async_stack = stack + ASYNC_SIZE;
+		stack = __get_free_pages(GFP_KERNEL, 0);
+		if (!stack)
 			panic("smp_boot_cpus failed to allocate memory\n");
-		lowcore_ptr[i]->panic_stack = stack + (PAGE_SIZE);
+		lowcore_ptr[i]->panic_stack = stack + PAGE_SIZE;
 #ifndef CONFIG_64BIT
 		if (MACHINE_HAS_IEEE) {
 			lowcore_ptr[i]->extended_save_area_addr =
-				(__u32) __get_free_pages(GFP_KERNEL,0);
-			if (lowcore_ptr[i]->extended_save_area_addr == 0)
+				(__u32) __get_free_pages(GFP_KERNEL, 0);
+			if (!lowcore_ptr[i]->extended_save_area_addr)
 				panic("smp_boot_cpus failed to "
 				      "allocate memory\n");
 		}
@@ -816,7 +764,7 @@
  */
 int setup_profiling_timer(unsigned int multiplier)
 {
-        return 0;
+	return 0;
 }
 
 static DEFINE_PER_CPU(struct cpu, cpu_devices);
@@ -853,7 +801,7 @@
 }
 
 static struct notifier_block __cpuinitdata smp_cpu_nb = {
-	.notifier_call	= smp_cpu_notify,
+	.notifier_call = smp_cpu_notify,
 };
 
 static int __init topology_init(void)
@@ -875,13 +823,4 @@
 	}
 	return 0;
 }
-
 subsys_initcall(topology_init);
-
-EXPORT_SYMBOL(cpu_online_map);
-EXPORT_SYMBOL(cpu_possible_map);
-EXPORT_SYMBOL(lowcore_ptr);
-EXPORT_SYMBOL(smp_ctl_set_bit);
-EXPORT_SYMBOL(smp_ctl_clear_bit);
-EXPORT_SYMBOL(smp_get_cpu);
-EXPORT_SYMBOL(smp_put_cpu);
Index: quilt-2.6/include/asm-s390/smp.h
===================================================================
--- quilt-2.6.orig/include/asm-s390/smp.h	2007-04-27 16:04:59.000000000 +0200
+++ quilt-2.6/include/asm-s390/smp.h	2007-04-27 16:05:02.000000000 +0200
@@ -54,9 +54,6 @@
 
 #define raw_smp_processor_id()	(S390_lowcore.cpu_data.cpu_nr)
 
-extern int smp_get_cpu(cpumask_t cpu_map);
-extern void smp_put_cpu(int cpu);
-
 static inline __u16 hard_smp_processor_id(void)
 {
         __u16 cpu_address;
@@ -114,8 +111,6 @@
 }
 
 #define smp_cpu_not_running(cpu)	1
-#define smp_get_cpu(cpu) ({ 0; })
-#define smp_put_cpu(cpu) ({ 0; })
 #define smp_setup_cpu_possible_map()	do { } while (0)
 #endif
 

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

[patch 38/38] SPIN_LOCK_UNLOCKED cleanup in drivers/s390

[Martin Schwidefsky]

[-- Attachment #1: 141-spinlock.diff --]
[-- Type: text/plain, Size: 2095 bytes --]

From: Milind Arun Choudhary <milindchoudhary@gmail.com>

SPIN_LOCK_UNLOCKED cleanup,use __SPIN_LOCK_UNLOCKED instead.

Signed-off-by: Milind Arun Choudhary <milindchoudhary@gmail.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---

 drivers/s390/char/vmlogrdr.c |    6 +++---
 drivers/s390/cio/cmf.c       |    2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)

Index: quilt-2.6/drivers/s390/char/vmlogrdr.c
===================================================================
--- quilt-2.6.orig/drivers/s390/char/vmlogrdr.c	2007-04-27 16:04:59.000000000 +0200
+++ quilt-2.6/drivers/s390/char/vmlogrdr.c	2007-04-27 16:05:02.000000000 +0200
@@ -125,7 +125,7 @@
 	  .recording_name = "EREP",
 	  .minor_num      = 0,
 	  .buffer_free    = 1,
-	  .priv_lock      = SPIN_LOCK_UNLOCKED,
+	  .priv_lock	  = __SPIN_LOCK_UNLOCKED(sys_ser[0].priv_lock),
 	  .autorecording  = 1,
 	  .autopurge      = 1,
 	},
@@ -134,7 +134,7 @@
 	  .recording_name = "ACCOUNT",
 	  .minor_num      = 1,
 	  .buffer_free    = 1,
-	  .priv_lock      = SPIN_LOCK_UNLOCKED,
+	  .priv_lock	  = __SPIN_LOCK_UNLOCKED(sys_ser[1].priv_lock),
 	  .autorecording  = 1,
 	  .autopurge      = 1,
 	},
@@ -143,7 +143,7 @@
 	  .recording_name = "SYMPTOM",
 	  .minor_num      = 2,
 	  .buffer_free    = 1,
-	  .priv_lock      = SPIN_LOCK_UNLOCKED,
+	  .priv_lock	  = __SPIN_LOCK_UNLOCKED(sys_ser[2].priv_lock),
 	  .autorecording  = 1,
 	  .autopurge      = 1,
 	}
Index: quilt-2.6/drivers/s390/cio/cmf.c
===================================================================
--- quilt-2.6.orig/drivers/s390/cio/cmf.c	2007-04-27 16:01:49.000000000 +0200
+++ quilt-2.6/drivers/s390/cio/cmf.c	2007-04-27 16:05:02.000000000 +0200
@@ -476,7 +476,7 @@
 };
 
 static struct cmb_area cmb_area = {
-	.lock = SPIN_LOCK_UNLOCKED,
+	.lock = __SPIN_LOCK_UNLOCKED(cmb_area.lock),
 	.list = LIST_HEAD_INIT(cmb_area.list),
 	.num_channels  = 1024,
 };

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [patch 18/38] Improved oops output.

[Christoph Hellwig]

On Fri, Apr 27, 2007 at 04:05:21PM +0200, Martin Schwidefsky wrote:
> From: Martin Schwidefsky <schwidefsky@de.ibm.com>
> 
> This patch adds two improvements to the oops output. First it adds an
> additional line after the PSW which decodes the different fields of it.
> Second a disassembler is added that decodes the instructions surrounding
> the faulting PSW. The output of a test oops now looks like this:

So we're growing in-kernel disassemblers now? :)

I wonder whether this is really worth it, but at least your handwritten
one is much nicer than the copy & paste from gdb in the kdb patches.

Re: [patch 18/38] Improved oops output.

[Martin Schwidefsky]

On Fri, 2007-04-27 at 15:25 +0100, Christoph Hellwig wrote:
> On Fri, Apr 27, 2007 at 04:05:21PM +0200, Martin Schwidefsky wrote:
> > From: Martin Schwidefsky <schwidefsky@de.ibm.com>
> > 
> > This patch adds two improvements to the oops output. First it adds an
> > additional line after the PSW which decodes the different fields of it.
> > Second a disassembler is added that decodes the instructions surrounding
> > the faulting PSW. The output of a test oops now looks like this:
> 
> So we're growing in-kernel disassemblers now? :)
> 
> I wonder whether this is really worth it, but at least your handwritten
> one is much nicer than the copy & paste from gdb in the kdb patches.

The disassembler uses about 8K of memory and it really helps in
debugging. For me it is worth the cost.

-- 
blue skies,
  Martin.

"Reality continues to ruin my life." - Calvin.

Re: [patch 18/38] Improved oops output.

[Chuck Ebbert]

Martin Schwidefsky wrote:
>> So we're growing in-kernel disassemblers now? :)
>>
>> I wonder whether this is really worth it, but at least your handwritten
>> one is much nicer than the copy & paste from gdb in the kdb patches.
> 
> The disassembler uses about 8K of memory and it really helps in
> debugging. For me it is worth the cost.
> 

No ksymoops for your architecture?

Re: [patch 18/38] Improved oops output.

[Martin Schwidefsky]

On Fri, 2007-04-27 at 12:01 -0400, Chuck Ebbert wrote:
> Martin Schwidefsky wrote:
> >> So we're growing in-kernel disassemblers now? :)
> >>
> >> I wonder whether this is really worth it, but at least your handwritten
> >> one is much nicer than the copy & paste from gdb in the kdb patches.
> > 
> > The disassembler uses about 8K of memory and it really helps in
> > debugging. For me it is worth the cost.
> > 
> 
> No ksymoops for your architecture?

There is ksymoops for s390 and it should be able to disassemble the code
as well. But you need all sorts of files to feed into ksymoops to get
something out of it. At least I often get bug reports with the console
output attached and it simply saves time if I can see the disassembly
without an extra step.

-- 
blue skies,
  Martin.

"Reality continues to ruin my life." - Calvin.

Re: [patch 18/38] Improved oops output.

[Chuck Ebbert]

Martin Schwidefsky wrote:
>> No ksymoops for your architecture?
> 
> There is ksymoops for s390 and it should be able to disassemble the code
> as well. But you need all sorts of files to feed into ksymoops to get
> something out of it. At least I often get bug reports with the console
> output attached and it simply saves time if I can see the disassembly
> without an extra step.
> 

The only file you need on i386 is the oops itself. ksymoops will scream
about missing files but will still work for disassembly.

Re: [patch 18/38] Improved oops output.

[Andi Kleen]

Martin Schwidefsky <schwidefsky@de.ibm.com> writes:

> Second a disassembler is added that decodes the instructions surrounding
> the faulting PSW. 

Is that really needed? I use a simple wrapper around as/objdump for this
and it works great.

(ftp://ftp.firstfloor.org/pub/ak/shell/decodecode) 

-Andi

Re: [patch 18/38] Improved oops output.

[Martin Schwidefsky]

On Fri, 2007-04-27 at 20:14 +0200, Andi Kleen wrote:
> Martin Schwidefsky <schwidefsky@de.ibm.com> writes:
> 
> > Second a disassembler is added that decodes the instructions surrounding
> > the faulting PSW. 
> 
> Is that really needed? I use a simple wrapper around as/objdump for this
> and it works great.
> 
> (ftp://ftp.firstfloor.org/pub/ak/shell/decodecode) 

It is not absolutly needed but it is damn convenient. Usually I can
immediatly see the reason for the oops. And the size is only 8KB which
is next to nothing.

-- 
blue skies,
  Martin.

"Reality continues to ruin my life." - Calvin.

Re: [patch 18/38] Improved oops output.

[Chuck Ebbert]

Martin Schwidefsky wrote:
> On Fri, 2007-04-27 at 20:14 +0200, Andi Kleen wrote:
>> Martin Schwidefsky <schwidefsky@de.ibm.com> writes:
>>
>>> Second a disassembler is added that decodes the instructions surrounding
>>> the faulting PSW. 
>> Is that really needed? I use a simple wrapper around as/objdump for this
>> and it works great.
>>
>> (ftp://ftp.firstfloor.org/pub/ak/shell/decodecode) 
> 
> It is not absolutly needed but it is damn convenient. Usually I can
> immediatly see the reason for the oops. And the size is only 8KB which
> is next to nothing.
> 

Yeah, well your instruction set is really simple anyway. :)

I wonder how big the x86 disassembler is...

Re: [patch 20/38] Minor fault path optimization.

[Paul Mackerras]

Martin Schwidefsky writes:

> The minor fault path has grown a lot in terms of cycles. In particular
> the kprobes hook is very costly. Optimize the path to save a couple of
> cycles. If kprobes is enabled more than 300 cycles can be avoided if 
> kprobes_running() is false.

There's no good reason to use a notifier for page faults, since
there's only one external piece of code that wants to know about
them...

Regards,
Paul.

Re: [patch 20/38] Minor fault path optimization.

[Christoph Hellwig]

On Sat, Apr 28, 2007 at 02:49:55PM +1000, Paul Mackerras wrote:
> Martin Schwidefsky writes:
> 
> > The minor fault path has grown a lot in terms of cycles. In particular
> > the kprobes hook is very costly. Optimize the path to save a couple of
> > cycles. If kprobes is enabled more than 300 cycles can be avoided if 
> > kprobes_running() is false.
> 
> There's no good reason to use a notifier for page faults, since
> there's only one external piece of code that wants to know about
> them...

Well, here's my powerpc patch to do the direct call that I sent out
a few weeks ago.  It not just speed up the pagefault path a lot, 
but also is a major code cleanup.  Andi and Anton didn't like it
because they have ambition for murky debugger code to possibly
use it one day..

Index: linux-2.6.20/arch/powerpc/mm/fault.c
===================================================================
--- linux-2.6.20.orig/arch/powerpc/mm/fault.c	2007-03-19 14:50:41.000000000 +0100
+++ linux-2.6.20/arch/powerpc/mm/fault.c	2007-03-19 15:00:00.000000000 +0100
@@ -39,37 +39,26 @@
 #include <asm/kdebug.h>
 #include <asm/siginfo.h>
 
-#ifdef CONFIG_KPROBES
-ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
 
-/* Hook to register for page fault notifications */
-int register_page_fault_notifier(struct notifier_block *nb)
+#ifdef CONFIG_KPROBES
+static inline int notify_page_fault(struct pt_regs *regs)
 {
-	return atomic_notifier_chain_register(&notify_page_fault_chain, nb);
-}
+	int ret = 0;
 
-int unregister_page_fault_notifier(struct notifier_block *nb)
-{
-	return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb);
-}
+	/* kprobe_running() needs smp_processor_id() */
+	if (!user_mode(regs)) {
+		preempt_disable();
+		if (kprobe_running() && kprobe_fault_handler(regs, 11))
+			ret = 1;
+		preempt_enable();
+	}
 
-static inline int notify_page_fault(enum die_val val, const char *str,
-			struct pt_regs *regs, long err, int trap, int sig)
-{
-	struct die_args args = {
-		.regs = regs,
-		.str = str,
-		.err = err,
-		.trapnr = trap,
-		.signr = sig
-	};
-	return atomic_notifier_call_chain(&notify_page_fault_chain, val, &args);
+	return ret;
 }
 #else
-static inline int notify_page_fault(enum die_val val, const char *str,
-			struct pt_regs *regs, long err, int trap, int sig)
+static inline int notify_page_fault(struct pt_regs *regs)
 {
-	return NOTIFY_DONE;
+	return 0;
 }
 #endif
 
@@ -175,8 +164,7 @@ int __kprobes do_page_fault(struct pt_re
 	is_write = error_code & ESR_DST;
 #endif /* CONFIG_4xx || CONFIG_BOOKE */
 
-	if (notify_page_fault(DIE_PAGE_FAULT, "page_fault", regs, error_code,
-				11, SIGSEGV) == NOTIFY_STOP)
+	if (notify_page_fault(regs))
 		return 0;
 
 	if (trap == 0x300) {
Index: linux-2.6.20/arch/powerpc/kernel/kprobes.c
===================================================================
--- linux-2.6.20.orig/arch/powerpc/kernel/kprobes.c	2007-03-19 14:53:01.000000000 +0100
+++ linux-2.6.20/arch/powerpc/kernel/kprobes.c	2007-03-19 15:03:41.000000000 +0100
@@ -376,7 +376,7 @@ out:
 	return 1;
 }
 
-static int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
+int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
 {
 	struct kprobe *cur = kprobe_running();
 	struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
@@ -461,14 +461,6 @@ int __kprobes kprobe_exceptions_notify(s
 		if (post_kprobe_handler(args->regs))
 			ret = NOTIFY_STOP;
 		break;
-	case DIE_PAGE_FAULT:
-		/* kprobe_running() needs smp_processor_id() */
-		preempt_disable();
-		if (kprobe_running() &&
-		    kprobe_fault_handler(args->regs, args->trapnr))
-			ret = NOTIFY_STOP;
-		preempt_enable();
-		break;
 	default:
 		break;
 	}
Index: linux-2.6.20/include/asm-powerpc/kdebug.h
===================================================================
--- linux-2.6.20.orig/include/asm-powerpc/kdebug.h	2007-03-19 14:55:13.000000000 +0100
+++ linux-2.6.20/include/asm-powerpc/kdebug.h	2007-03-19 15:03:27.000000000 +0100
@@ -18,8 +18,20 @@ struct die_args {
 
 extern int register_die_notifier(struct notifier_block *);
 extern int unregister_die_notifier(struct notifier_block *);
-extern int register_page_fault_notifier(struct notifier_block *);
-extern int unregister_page_fault_notifier(struct notifier_block *);
+
+/*
+ * These are only here because kprobes.c wants them to implement a
+ * blatant layer violation.  Will hopefully go away soon once all
+ * architectures are updated.
+ */
+static inline int register_page_fault_notifier(struct notifier_block *nb)
+{
+	return 0;
+}
+static inline int unregister_page_fault_notifier(struct notifier_block *nb)
+{
+	return 0;
+}
 extern struct atomic_notifier_head powerpc_die_chain;
 
 /* Grossly misnamed. */
@@ -29,7 +41,6 @@ enum die_val {
 	DIE_DABR_MATCH,
 	DIE_BPT,
 	DIE_SSTEP,
-	DIE_PAGE_FAULT,
 };
 
 static inline int notify_die(enum die_val val,char *str,struct pt_regs *regs,long err,int trap, int sig)
Index: linux-2.6.20/include/asm-powerpc/kprobes.h
===================================================================
--- linux-2.6.20.orig/include/asm-powerpc/kprobes.h	2007-03-19 15:01:14.000000000 +0100
+++ linux-2.6.20/include/asm-powerpc/kprobes.h	2007-03-19 15:01:55.000000000 +0100
@@ -100,5 +100,6 @@ struct kprobe_ctlblk {
 
 extern int kprobe_exceptions_notify(struct notifier_block *self,
 					unsigned long val, void *data);
+extern int kprobe_fault_handler(struct pt_regs *regs, int trapnr);
 #endif /* __KERNEL__ */
 #endif	/* _ASM_POWERPC_KPROBES_H */

Re: [patch 20/38] Minor fault path optimization.

[Paul Mackerras]

Christoph Hellwig writes:

> Well, here's my powerpc patch to do the direct call that I sent out
> a few weeks ago.  It not just speed up the pagefault path a lot, 
> but also is a major code cleanup.  Andi and Anton didn't like it
> because they have ambition for murky debugger code to possibly
> use it one day..

Looks good.  Could you resend it to me with a proper description and a
sign-off?

Thanks,
Paul.

[PATCH] powerpc: minor fault path optimization

[Christoph Hellwig]

Call the kprobes pagefault handler directly instead of going through
the complex notifier chain.


Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: linux-2.6/arch/powerpc/mm/fault.c
===================================================================
--- linux-2.6.orig/arch/powerpc/mm/fault.c	2007-03-23 20:07:28.000000000 +0100
+++ linux-2.6/arch/powerpc/mm/fault.c	2007-04-30 11:33:36.000000000 +0200
@@ -39,37 +39,26 @@
 #include <asm/kdebug.h>
 #include <asm/siginfo.h>
 
-#ifdef CONFIG_KPROBES
-ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
 
-/* Hook to register for page fault notifications */
-int register_page_fault_notifier(struct notifier_block *nb)
+#ifdef CONFIG_KPROBES
+static inline int notify_page_fault(struct pt_regs *regs)
 {
-	return atomic_notifier_chain_register(&notify_page_fault_chain, nb);
-}
+	int ret = 0;
 
-int unregister_page_fault_notifier(struct notifier_block *nb)
-{
-	return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb);
-}
+	/* kprobe_running() needs smp_processor_id() */
+	if (!user_mode(regs)) {
+		preempt_disable();
+		if (kprobe_running() && kprobe_fault_handler(regs, 11))
+			ret = 1;
+		preempt_enable();
+	}
 
-static inline int notify_page_fault(enum die_val val, const char *str,
-			struct pt_regs *regs, long err, int trap, int sig)
-{
-	struct die_args args = {
-		.regs = regs,
-		.str = str,
-		.err = err,
-		.trapnr = trap,
-		.signr = sig
-	};
-	return atomic_notifier_call_chain(&notify_page_fault_chain, val, &args);
+	return ret;
 }
 #else
-static inline int notify_page_fault(enum die_val val, const char *str,
-			struct pt_regs *regs, long err, int trap, int sig)
+static inline int notify_page_fault(struct pt_regs *regs)
 {
-	return NOTIFY_DONE;
+	return 0;
 }
 #endif
 
@@ -175,8 +164,7 @@ int __kprobes do_page_fault(struct pt_re
 	is_write = error_code & ESR_DST;
 #endif /* CONFIG_4xx || CONFIG_BOOKE */
 
-	if (notify_page_fault(DIE_PAGE_FAULT, "page_fault", regs, error_code,
-				11, SIGSEGV) == NOTIFY_STOP)
+	if (notify_page_fault(regs))
 		return 0;
 
 	if (trap == 0x300) {
Index: linux-2.6/arch/powerpc/kernel/kprobes.c
===================================================================
--- linux-2.6.orig/arch/powerpc/kernel/kprobes.c	2007-04-02 20:18:18.000000000 +0200
+++ linux-2.6/arch/powerpc/kernel/kprobes.c	2007-04-30 11:33:36.000000000 +0200
@@ -376,7 +376,7 @@ out:
 	return 1;
 }
 
-static int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
+int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
 {
 	struct kprobe *cur = kprobe_running();
 	struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
@@ -461,14 +461,6 @@ int __kprobes kprobe_exceptions_notify(s
 		if (post_kprobe_handler(args->regs))
 			ret = NOTIFY_STOP;
 		break;
-	case DIE_PAGE_FAULT:
-		/* kprobe_running() needs smp_processor_id() */
-		preempt_disable();
-		if (kprobe_running() &&
-		    kprobe_fault_handler(args->regs, args->trapnr))
-			ret = NOTIFY_STOP;
-		preempt_enable();
-		break;
 	default:
 		break;
 	}
Index: linux-2.6/include/asm-powerpc/kdebug.h
===================================================================
--- linux-2.6.orig/include/asm-powerpc/kdebug.h	2007-03-23 20:07:29.000000000 +0100
+++ linux-2.6/include/asm-powerpc/kdebug.h	2007-04-30 11:51:04.000000000 +0200
@@ -18,8 +18,20 @@ struct die_args {
 
 extern int register_die_notifier(struct notifier_block *);
 extern int unregister_die_notifier(struct notifier_block *);
-extern int register_page_fault_notifier(struct notifier_block *);
-extern int unregister_page_fault_notifier(struct notifier_block *);
+
+/*
+ * These are only here because kprobes.c wants them to implement a
+ * blatant layering violation.  Will hopefully go away soon once all
+ * architectures are updated.
+ */
+static inline int register_page_fault_notifier(struct notifier_block *nb)
+{
+	return 0;
+}
+static inline int unregister_page_fault_notifier(struct notifier_block *nb)
+{
+	return 0;
+}
 extern struct atomic_notifier_head powerpc_die_chain;
 
 /* Grossly misnamed. */
@@ -29,7 +41,6 @@ enum die_val {
 	DIE_DABR_MATCH,
 	DIE_BPT,
 	DIE_SSTEP,
-	DIE_PAGE_FAULT,
 };
 
 static inline int notify_die(enum die_val val,char *str,struct pt_regs *regs,long err,int trap, int sig)
Index: linux-2.6/include/asm-powerpc/kprobes.h
===================================================================
--- linux-2.6.orig/include/asm-powerpc/kprobes.h	2007-02-11 08:17:32.000000000 +0100
+++ linux-2.6/include/asm-powerpc/kprobes.h	2007-04-30 11:33:36.000000000 +0200
@@ -105,5 +105,6 @@ struct kprobe_ctlblk {
 
 extern int kprobe_exceptions_notify(struct notifier_block *self,
 					unsigned long val, void *data);
+extern int kprobe_fault_handler(struct pt_regs *regs, int trapnr);
 #endif /* __KERNEL__ */
 #endif	/* _ASM_POWERPC_KPROBES_H */

Re: [patch 20/38] Minor fault path optimization.

[Martin Schwidefsky]

On Sat, 2007-04-28 at 09:34 +0100, Christoph Hellwig wrote:
> Well, here's my powerpc patch to do the direct call that I sent out
> a few weeks ago.  It not just speed up the pagefault path a lot, 
> but also is a major code cleanup.  Andi and Anton didn't like it
> because they have ambition for murky debugger code to possibly
> use it one day..

I thought about using a direct call for s390 as well. The advantage of a
direct call is that it avoids the overhead of a notifier call even if
kprobes is running. The disadvantage is that there cannot be a second
consumer of the fault notifications. I choose the middle way: avoid the
overhead if kprobes is not active and use the full notifier if kprobes
is running. It should not hurt the performance too badly since
kprobe_running is only true if the cpu gets a page fault while executing
a kprobe.

-- 
blue skies,
  Martin.

"Reality continues to ruin my life." - Calvin.

Re: [patch 20/38] Minor fault path optimization.

[Christoph Hellwig]

On Wed, May 02, 2007 at 12:45:06PM +0200, Martin Schwidefsky wrote:
> I thought about using a direct call for s390 as well. The advantage of a
> direct call is that it avoids the overhead of a notifier call even if
> kprobes is running. The disadvantage is that there cannot be a second
> consumer of the fault notifications. I choose the middle way: avoid the
> overhead if kprobes is not active and use the full notifier if kprobes
> is running. It should not hurt the performance too badly since
> kprobe_running is only true if the cpu gets a page fault while executing
> a kprobe.

There is not second caller :)  Also it wouldn't work because you check
for kprobe_running() which wouldn't be true for it.  It also means
you go through the whole notifier chain mechanism for every userspace
page fault when kprobes are running which is the least of them.

So let's no do odd preparations for the future but optimize for
the current case and add things if needed in the future.  Note that
I plan to send similar patches for all kprobes supporting architectures
so that we have common code again.  Eventually notify_page_fault should
move into include/linux/kprobes.h



> 
> --kprobe_running 
> blue skies,
>   Martin.
> 
> "Reality continues to ruin my life." - Calvin.
> 
> 
---end quoted text---