[PATCH] s390: dasd extended error reporting module

The DASD extended error reporting is a facility that allows to get detailed
information about certain problems in the DASD I/O.  This information can be
used to implement fail-over applications that can recover these problems.
Signed-off-by: Stefan Weinhuber <wein@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

drivers/s390/block/Kconfig

@@ -55,13 +55,21 @@ config DASD_DIAG
 	  Disks under VM.  If you are not running under VM or unsure what it is,
 	  say "N".
 
+config DASD_EER
+	tristate "Extended error reporting (EER)"
+	depends on DASD
+	help
+	  This driver provides a character device interface to the
+          DASD extended error reporting. This is only needed if you want to
+          use applications written for the EER facility.
+
 config DASD_CMB
 	tristate "Compatibility interface for DASD channel measurement blocks"
 	depends on DASD
 	help
-	  This driver provides an additional interface to the channel measurement
-	  facility, which is normally accessed though sysfs, with a set of
-	  ioctl functions specific to the dasd driver.
+	  This driver provides an additional interface to the channel
+          measurement facility, which is normally accessed though sysfs, with
+          a set of ioctl functions specific to the dasd driver.
 	  This is only needed if you want to use applications written for
 	  linux-2.4 dasd channel measurement facility interface.
 

drivers/s390/block/Makefile

@@ -5,6 +5,7 @@
 dasd_eckd_mod-objs := dasd_eckd.o dasd_3990_erp.o dasd_9343_erp.o
 dasd_fba_mod-objs  := dasd_fba.o dasd_3370_erp.o dasd_9336_erp.o
 dasd_diag_mod-objs := dasd_diag.o
+dasd_eer_mod-objs := dasd_eer.o
 dasd_mod-objs      := dasd.o dasd_ioctl.o dasd_proc.o dasd_devmap.o \
 			dasd_genhd.o dasd_erp.o
 
@@ -13,5 +14,6 @@ obj-$(CONFIG_DASD_DIAG) += dasd_diag_mod.o
 obj-$(CONFIG_DASD_ECKD) += dasd_eckd_mod.o
 obj-$(CONFIG_DASD_FBA)  += dasd_fba_mod.o
 obj-$(CONFIG_DASD_CMB)  += dasd_cmb.o
+obj-$(CONFIG_DASD_EER)  += dasd_eer.o
 obj-$(CONFIG_BLK_DEV_XPRAM) += xpram.o
 obj-$(CONFIG_DCSSBLK) += dcssblk.o

drivers/s390/block/dasd.c

@@ -18,6 +18,7 @@
 #include <linux/slab.h>
 #include <linux/buffer_head.h>
 #include <linux/hdreg.h>
+#include <linux/notifier.h>
 
 #include <asm/ccwdev.h>
 #include <asm/ebcdic.h>
@@ -57,6 +58,7 @@ static void dasd_int_handler(struct ccw_device *, unsigned long, struct irb *);
 static void dasd_flush_ccw_queue(struct dasd_device *, int);
 static void dasd_tasklet(struct dasd_device *);
 static void do_kick_device(void *data);
+static void dasd_disable_eer(struct dasd_device *device);
 
 /*
  * SECTION: Operations on the device structure.
@@ -151,6 +153,8 @@ dasd_state_new_to_known(struct dasd_device *device)
 static inline void
 dasd_state_known_to_new(struct dasd_device * device)
 {
+	/* disable extended error reporting for this device */
+	dasd_disable_eer(device);
 	/* Forget the discipline information. */
 	device->discipline = NULL;
 	device->state = DASD_STATE_NEW;
@@ -867,6 +871,9 @@ dasd_handle_state_change_pending(struct dasd_device *device)
 	struct dasd_ccw_req *cqr;
 	struct list_head *l, *n;
 
+	/* first of all call extended error reporting */
+	dasd_write_eer_trigger(DASD_EER_STATECHANGE, device, NULL);
+
 	device->stopped &= ~DASD_STOPPED_PENDING;
 
         /* restart all 'running' IO on queue */
@@ -1086,6 +1093,19 @@ restart:
 			}
 			goto restart;
 		}
+
+		/* first of all call extended error reporting */
+		if (device->eer && cqr->status == DASD_CQR_FAILED) {
+			dasd_write_eer_trigger(DASD_EER_FATALERROR,
+					       device, cqr);
+
+			/* restart request  */
+			cqr->status = DASD_CQR_QUEUED;
+			cqr->retries = 255;
+			device->stopped |= DASD_STOPPED_QUIESCE;
+			goto restart;
+		}
+
 		/* Process finished ERP request. */
 		if (cqr->refers) {
 			__dasd_process_erp(device, cqr);
@@ -1223,7 +1243,8 @@ __dasd_start_head(struct dasd_device * device)
 	cqr = list_entry(device->ccw_queue.next, struct dasd_ccw_req, list);
         /* check FAILFAST */
 	if (device->stopped & ~DASD_STOPPED_PENDING &&
-	    test_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags)) {
+	    test_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags) &&
+	    (!device->eer)) {
 		cqr->status = DASD_CQR_FAILED;
 		dasd_schedule_bh(device);
 	}
@@ -1965,6 +1986,9 @@ dasd_generic_notify(struct ccw_device *cdev, int event)
 	switch (event) {
 	case CIO_GONE:
 	case CIO_NO_PATH:
+		/* first of all call extended error reporting */
+		dasd_write_eer_trigger(DASD_EER_NOPATH, device, NULL);
+
 		if (device->state < DASD_STATE_BASIC)
 			break;
 		/* Device is active. We want to keep it. */
@@ -2022,6 +2046,51 @@ dasd_generic_auto_online (struct ccw_driver *dasd_discipline_driver)
 	put_driver(drv);
 }
 
+/*
+ * notifications for extended error reports
+ */
+static struct notifier_block *dasd_eer_chain;
+
+int
+dasd_register_eer_notifier(struct notifier_block *nb)
+{
+	return notifier_chain_register(&dasd_eer_chain, nb);
+}
+
+int
+dasd_unregister_eer_notifier(struct notifier_block *nb)
+{
+	return notifier_chain_unregister(&dasd_eer_chain, nb);
+}
+
+/*
+ * Notify the registered error reporting module of a problem
+ */
+void
+dasd_write_eer_trigger(unsigned int id, struct dasd_device *device,
+		       struct dasd_ccw_req *cqr)
+{
+	if (device->eer) {
+		struct dasd_eer_trigger temp;
+		temp.id = id;
+		temp.device = device;
+		temp.cqr = cqr;
+		notifier_call_chain(&dasd_eer_chain, DASD_EER_TRIGGER,
+				    (void *)&temp);
+	}
+}
+
+/*
+ * Tell the registered error reporting module to disable error reporting for
+ * a given device and to cleanup any private data structures on that device.
+ */
+static void
+dasd_disable_eer(struct dasd_device *device)
+{
+	notifier_call_chain(&dasd_eer_chain, DASD_EER_DISABLE, (void *)device);
+}
+
+
 static int __init
 dasd_init(void)
 {
@@ -2103,6 +2172,11 @@ EXPORT_SYMBOL_GPL(dasd_generic_set_online);
 EXPORT_SYMBOL_GPL(dasd_generic_set_offline);
 EXPORT_SYMBOL_GPL(dasd_generic_auto_online);
 
+EXPORT_SYMBOL(dasd_register_eer_notifier);
+EXPORT_SYMBOL(dasd_unregister_eer_notifier);
+EXPORT_SYMBOL(dasd_write_eer_trigger);
+
+
 /*
  * Overrides for Emacs so that we follow Linus's tabbing style.
  * Emacs will notice this stuff at the end of the file and automatically

drivers/s390/block/dasd_3990_erp.c

@@ -1108,6 +1108,9 @@ dasd_3990_handle_env_data(struct dasd_ccw_req * erp, char *sense)
 		case 0x0B:
 			DEV_MESSAGE(KERN_WARNING, device, "%s",
 				    "FORMAT F - Volume is suspended duplex");
+			/* call extended error reporting (EER) */
+			dasd_write_eer_trigger(DASD_EER_PPRCSUSPEND, device,
+					       erp->refers);
 			break;
 		case 0x0C:
 			DEV_MESSAGE(KERN_WARNING, device, "%s",

drivers/s390/block/dasd_eckd.h

@@ -29,6 +29,7 @@
 #define DASD_ECKD_CCW_PSF		 0x27
 #define DASD_ECKD_CCW_RSSD		 0x3e
 #define DASD_ECKD_CCW_LOCATE_RECORD	 0x47
+#define DASD_ECKD_CCW_SNSS               0x54
 #define DASD_ECKD_CCW_DEFINE_EXTENT	 0x63
 #define DASD_ECKD_CCW_WRITE_MT		 0x85
 #define DASD_ECKD_CCW_READ_MT		 0x86

drivers/s390/block/dasd_int.h

@@ -275,6 +275,34 @@ struct dasd_discipline {
 
 extern struct dasd_discipline *dasd_diag_discipline_pointer;
 
+
+/*
+ * Notification numbers for extended error reporting notifications:
+ * The DASD_EER_DISABLE notification is sent before a dasd_device (and it's
+ * eer pointer) is freed. The error reporting module needs to do all necessary
+ * cleanup steps.
+ * The DASD_EER_TRIGGER notification sends the actual error reports (triggers).
+ */
+#define DASD_EER_DISABLE 0
+#define DASD_EER_TRIGGER 1
+
+/* Trigger IDs for extended error reporting DASD_EER_TRIGGER notification */
+#define DASD_EER_FATALERROR  1
+#define DASD_EER_NOPATH      2
+#define DASD_EER_STATECHANGE 3
+#define DASD_EER_PPRCSUSPEND 4
+
+/*
+ * The dasd_eer_trigger structure contains all data that we need to send
+ * along with an DASD_EER_TRIGGER notification.
+ */
+struct dasd_eer_trigger {
+	unsigned int id;
+	struct dasd_device *device;
+	struct dasd_ccw_req *cqr;
+};
+
+
 struct dasd_device {
 	/* Block device stuff. */
 	struct gendisk *gdp;
@@ -288,6 +316,9 @@ struct dasd_device {
 	unsigned long flags;		/* per device flags */
 	unsigned short features;        /* copy of devmap-features (read-only!) */
 
+	/* extended error reporting stuff (eer) */
+	void *eer;
+
 	/* Device discipline stuff. */
 	struct dasd_discipline *discipline;
 	char *private;
@@ -488,6 +519,12 @@ int dasd_generic_set_online(struct ccw_device *, struct dasd_discipline *);
 int dasd_generic_set_offline (struct ccw_device *cdev);
 int dasd_generic_notify(struct ccw_device *, int);
 void dasd_generic_auto_online (struct ccw_driver *);
+int dasd_register_eer_notifier(struct notifier_block *);
+int dasd_unregister_eer_notifier(struct notifier_block *);
+void dasd_write_eer_trigger(unsigned int , struct dasd_device *,
+			struct dasd_ccw_req *);
+
+
 
 /* externals in dasd_devmap.c */
 extern int dasd_max_devindex;

include/asm-s390/dasd.h

@@ -204,7 +204,8 @@ typedef struct attrib_data_t {
  *
  * Here ist how the ioctl-nr should be used:
  *    0 -   31   DASD driver itself
- *   32 -  239   still open
+ *   32 -  229   still open
+ *  230 -  239   DASD extended error reporting
  *  240 -  255   reserved for EMC 
  *******************************************************************************/
 
@@ -236,12 +237,22 @@ typedef struct attrib_data_t {
 #define BIODASDPSRD    _IOR(DASD_IOCTL_LETTER,4,dasd_rssd_perf_stats_t)
 /* Get Attributes (cache operations) */
 #define BIODASDGATTR   _IOR(DASD_IOCTL_LETTER,5,attrib_data_t) 
+/* retrieve extended error-reporting value */
+#define BIODASDEERGET  _IOR(DASD_IOCTL_LETTER,6,int)
 
 
 /* #define BIODASDFORMAT  _IOW(IOCTL_LETTER,0,format_data_t) , deprecated */
 #define BIODASDFMT     _IOW(DASD_IOCTL_LETTER,1,format_data_t) 
 /* Set Attributes (cache operations) */
 #define BIODASDSATTR   _IOW(DASD_IOCTL_LETTER,2,attrib_data_t) 
+/* retrieve extended error-reporting value */
+#define BIODASDEERSET  _IOW(DASD_IOCTL_LETTER,3,int)
+
+
+/* remove all records from the eer buffer */
+#define DASD_EER_PURGE       _IO(DASD_IOCTL_LETTER,230)
+/* set the number of pages that are used for the internal eer buffer */
+#define DASD_EER_SETBUFSIZE  _IOW(DASD_IOCTL_LETTER,230,int)
 
 
 #endif				/* DASD_H */