2006-02-01 11:57:12

by Heiko Carstens

[permalink] [raw]
Subject: [PATCH 1/3] s390: dasd extended error reporting module.

From: Stefan Weinhuber <[email protected]>

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.
This is a resubmit of this patch because at first submission it
didn't get included due to Christoph's ioctl changes.
Since these aren't in the -mm tree anymore this one should be
merged now.

Signed-off-by: Stefan Weinhuber <[email protected]>
Signed-off-by: Martin Schwidefsky <[email protected]>
Signed-off-by: Heiko Carstens <[email protected]>
---

drivers/s390/block/Kconfig | 14
drivers/s390/block/Makefile | 2
drivers/s390/block/dasd.c | 76 ++
drivers/s390/block/dasd_3990_erp.c | 3
drivers/s390/block/dasd_eckd.h | 1
drivers/s390/block/dasd_eer.c | 1090 +++++++++++++++++++++++++++++++++++++
drivers/s390/block/dasd_int.h | 37 +
include/asm-s390/dasd.h | 13
8 files changed, 1231 insertions(+), 5 deletions(-)

diff -urpN linux-2.6/drivers/s390/block/dasd_3990_erp.c linux-2.6-patched/drivers/s390/block/dasd_3990_erp.c
--- linux-2.6/drivers/s390/block/dasd_3990_erp.c 2006-02-01 11:03:41.000000000 +0100
+++ linux-2.6-patched/drivers/s390/block/dasd_3990_erp.c 2006-02-01 11:04:09.000000000 +0100
@@ -1108,6 +1108,9 @@ dasd_3990_handle_env_data(struct dasd_cc
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",
diff -urpN linux-2.6/drivers/s390/block/dasd.c linux-2.6-patched/drivers/s390/block/dasd.c
--- linux-2.6/drivers/s390/block/dasd.c 2006-02-01 11:03:59.000000000 +0100
+++ linux-2.6-patched/drivers/s390/block/dasd.c 2006-02-01 11:04:09.000000000 +0100
@@ -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_
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_devi
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
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 * d
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 *c
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_dri
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_onlin
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
diff -urpN linux-2.6/drivers/s390/block/dasd_eckd.h linux-2.6-patched/drivers/s390/block/dasd_eckd.h
--- linux-2.6/drivers/s390/block/dasd_eckd.h 2006-02-01 11:03:41.000000000 +0100
+++ linux-2.6-patched/drivers/s390/block/dasd_eckd.h 2006-02-01 11:04:09.000000000 +0100
@@ -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
diff -urpN linux-2.6/drivers/s390/block/dasd_eer.c linux-2.6-patched/drivers/s390/block/dasd_eer.c
--- linux-2.6/drivers/s390/block/dasd_eer.c 1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6-patched/drivers/s390/block/dasd_eer.c 2006-02-01 11:04:09.000000000 +0100
@@ -0,0 +1,1090 @@
+/*
+ * character device driver for extended error reporting
+ *
+ *
+ * Copyright (C) 2005 IBM Corporation
+ * extended error reporting for DASD ECKD devices
+ * Author(s): Stefan Weinhuber <[email protected]>
+ *
+ */
+
+#include <linux/init.h>
+#include <linux/fs.h>
+#include <linux/kernel.h>
+#include <linux/miscdevice.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/device.h>
+#include <linux/workqueue.h>
+#include <linux/poll.h>
+#include <linux/notifier.h>
+
+#include <asm/uaccess.h>
+#include <asm/semaphore.h>
+#include <asm/atomic.h>
+#include <asm/ebcdic.h>
+
+#include "dasd_int.h"
+#include "dasd_eckd.h"
+
+
+MODULE_LICENSE("GPL");
+
+MODULE_AUTHOR("Stefan Weinhuber <[email protected]>");
+MODULE_DESCRIPTION("DASD extended error reporting module");
+
+
+#ifdef PRINTK_HEADER
+#undef PRINTK_HEADER
+#endif /* PRINTK_HEADER */
+#define PRINTK_HEADER "dasd(eer):"
+
+
+
+
+
+/*****************************************************************************/
+/* the internal buffer */
+/*****************************************************************************/
+
+/*
+ * The internal buffer is meant to store obaque blobs of data, so it doesn't
+ * know of higher level concepts like triggers.
+ * It consists of a number of pages that are used as a ringbuffer. Each data
+ * blob is stored in a simple record that consists of an integer, which
+ * contains the size of the following data, and the data bytes themselfes.
+ *
+ * To allow for multiple independent readers we create one internal buffer
+ * each time the device is opened and destroy the buffer when the file is
+ * closed again.
+ *
+ * One record can be written to a buffer by using the functions
+ * - dasd_eer_start_record (one time per record to write the size to the buffer
+ * and reserve the space for the data)
+ * - dasd_eer_write_buffer (one or more times per record to write the data)
+ * The data can be written in several steps but you will have to compute
+ * the total size up front for the invocation of dasd_eer_start_record.
+ * If the ringbuffer is full, dasd_eer_start_record will remove the required
+ * number of old records.
+ *
+ * A record is typically read in two steps, first read the integer that
+ * specifies the size of the following data, then read the data.
+ * Both can be done by
+ * - dasd_eer_read_buffer
+ *
+ * For all mentioned functions you need to get the bufferlock first and keep it
+ * until a complete record is written or read.
+ */
+
+
+/*
+ * Alle information necessary to keep track of an internal buffer is kept in
+ * a struct eerbuffer. The buffer specific to a file pointer is strored in
+ * the private_data field of that file. To be able to write data to all
+ * existing buffers, each buffer is also added to the bufferlist.
+ * If the user doesn't want to read a complete record in one go, we have to
+ * keep track of the rest of the record. residual stores the number of bytes
+ * that are still to deliver. If the rest of the record is invalidated between
+ * two reads then residual will be set to -1 so that the next read will fail.
+ * All entries in the eerbuffer structure are protected with the bufferlock.
+ * To avoid races between writing to a buffer on the one side and creating
+ * and destroying buffers on the other side, the bufferlock must also be used
+ * to protect the bufferlist.
+ */
+
+struct eerbuffer {
+ struct list_head list;
+ char **buffer;
+ int buffersize;
+ int buffer_page_count;
+ int head;
+ int tail;
+ int residual;
+};
+
+LIST_HEAD(bufferlist);
+
+static spinlock_t bufferlock = SPIN_LOCK_UNLOCKED;
+
+DECLARE_WAIT_QUEUE_HEAD(dasd_eer_read_wait_queue);
+
+/*
+ * How many free bytes are available on the buffer.
+ * needs to be called with bufferlock held
+ */
+static int
+dasd_eer_get_free_bytes(struct eerbuffer *eerb)
+{
+ if (eerb->head < eerb->tail) {
+ return eerb->tail - eerb->head - 1;
+ } else
+ return eerb->buffersize - eerb->head + eerb->tail -1;
+}
+
+/*
+ * How many bytes of buffer space are used.
+ * needs to be called with bufferlock held
+ */
+static int
+dasd_eer_get_filled_bytes(struct eerbuffer *eerb)
+{
+
+ if (eerb->head >= eerb->tail) {
+ return eerb->head - eerb->tail;
+ } else
+ return eerb->buffersize - eerb->tail + eerb->head;
+}
+
+/*
+ * The dasd_eer_write_buffer function just copies count bytes of data
+ * to the buffer. Make sure to call dasd_eer_start_record first, to
+ * make sure that enough free space is available.
+ * needs to be called with bufferlock held
+ */
+static void
+dasd_eer_write_buffer(struct eerbuffer *eerb, int count, char *data)
+{
+
+ unsigned long headindex,localhead;
+ unsigned long rest, len;
+ char *nextdata;
+
+ nextdata = data;
+ rest = count;
+ while (rest > 0) {
+ headindex = eerb->head / PAGE_SIZE;
+ localhead = eerb->head % PAGE_SIZE;
+ len = min(rest, (PAGE_SIZE - localhead));
+ memcpy(eerb->buffer[headindex]+localhead, nextdata, len);
+ nextdata += len;
+ rest -= len;
+ eerb->head += len;
+ if ( eerb->head == eerb->buffersize )
+ eerb->head = 0; /* wrap around */
+ if (eerb->head > eerb->buffersize) {
+ MESSAGE(KERN_ERR, "%s", "runaway buffer head.");
+ BUG();
+ }
+ }
+}
+
+/*
+ * needs to be called with bufferlock held
+ */
+static int
+dasd_eer_read_buffer(struct eerbuffer *eerb, int count, char *data)
+{
+
+ unsigned long tailindex,localtail;
+ unsigned long rest, len, finalcount;
+ char *nextdata;
+
+ finalcount = min(count, dasd_eer_get_filled_bytes(eerb));
+ nextdata = data;
+ rest = finalcount;
+ while (rest > 0) {
+ tailindex = eerb->tail / PAGE_SIZE;
+ localtail = eerb->tail % PAGE_SIZE;
+ len = min(rest, (PAGE_SIZE - localtail));
+ memcpy(nextdata, eerb->buffer[tailindex]+localtail, len);
+ nextdata += len;
+ rest -= len;
+ eerb->tail += len;
+ if ( eerb->tail == eerb->buffersize )
+ eerb->tail = 0; /* wrap around */
+ if (eerb->tail > eerb->buffersize) {
+ MESSAGE(KERN_ERR, "%s", "runaway buffer tail.");
+ BUG();
+ }
+ }
+ return finalcount;
+}
+
+/*
+ * Whenever you want to write a blob of data to the internal buffer you
+ * have to start by using this function first. It will write the number
+ * of bytes that will be written to the buffer. If necessary it will remove
+ * old records to make room for the new one.
+ * needs to be called with bufferlock held
+ */
+static int
+dasd_eer_start_record(struct eerbuffer *eerb, int count)
+{
+ int tailcount;
+ if (count + sizeof(count) > eerb->buffersize)
+ return -ENOMEM;
+ while (dasd_eer_get_free_bytes(eerb) < count + sizeof(count)) {
+ if (eerb->residual > 0) {
+ eerb->tail += eerb->residual;
+ if (eerb->tail >= eerb->buffersize)
+ eerb->tail -= eerb->buffersize;
+ eerb->residual = -1;
+ }
+ dasd_eer_read_buffer(eerb, sizeof(tailcount),
+ (char*)(&tailcount));
+ eerb->tail += tailcount;
+ if (eerb->tail >= eerb->buffersize)
+ eerb->tail -= eerb->buffersize;
+ }
+ dasd_eer_write_buffer(eerb, sizeof(count), (char*)(&count));
+
+ return 0;
+};
+
+/*
+ * release pages that are not used anymore
+ */
+static void
+dasd_eer_free_buffer_pages(char **buf, int no_pages)
+{
+ int i;
+
+ for (i = 0; i < no_pages; ++i) {
+ free_page((unsigned long)buf[i]);
+ }
+}
+
+/*
+ * allocate a new set of memory pages
+ */
+static int
+dasd_eer_allocate_buffer_pages(char **buf, int no_pages)
+{
+ int i;
+
+ for (i = 0; i < no_pages; ++i) {
+ buf[i] = (char *) get_zeroed_page(GFP_KERNEL);
+ if (!buf[i]) {
+ dasd_eer_free_buffer_pages(buf, i);
+ return -ENOMEM;
+ }
+ }
+ return 0;
+}
+
+/*
+ * empty the buffer by resetting head and tail
+ * In case there is a half read data blob in the buffer, we set residual
+ * to -1 to indicate that the remainder of the blob is lost.
+ */
+static void
+dasd_eer_purge_buffer(struct eerbuffer *eerb)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&bufferlock, flags);
+ if (eerb->residual > 0)
+ eerb->residual = -1;
+ eerb->tail=0;
+ eerb->head=0;
+ spin_unlock_irqrestore(&bufferlock, flags);
+}
+
+/*
+ * set the size of the buffer, newsize is the new number of pages to be used
+ * we don't try to copy any data back an forth, so any resize will also purge
+ * the buffer
+ */
+static int
+dasd_eer_resize_buffer(struct eerbuffer *eerb, int newsize)
+{
+ int i, oldcount, reuse;
+ char **new;
+ char **old;
+ unsigned long flags;
+
+ if (newsize < 1)
+ return -EINVAL;
+ if (eerb->buffer_page_count == newsize) {
+ /* documented behaviour is that any successfull invocation
+ * will purge all records */
+ dasd_eer_purge_buffer(eerb);
+ return 0;
+ }
+ new = kmalloc(newsize*sizeof(char*), GFP_KERNEL);
+ if (!new)
+ return -ENOMEM;
+
+ reuse=min(eerb->buffer_page_count, newsize);
+ for (i = 0; i < reuse; ++i) {
+ new[i] = eerb->buffer[i];
+ }
+ if (eerb->buffer_page_count < newsize) {
+ if (dasd_eer_allocate_buffer_pages(
+ &new[eerb->buffer_page_count],
+ newsize - eerb->buffer_page_count)) {
+ kfree(new);
+ return -ENOMEM;
+ }
+ }
+
+ spin_lock_irqsave(&bufferlock, flags);
+ old = eerb->buffer;
+ eerb->buffer = new;
+ if (eerb->residual > 0)
+ eerb->residual = -1;
+ eerb->tail = 0;
+ eerb->head = 0;
+ oldcount = eerb->buffer_page_count;
+ eerb->buffer_page_count = newsize;
+ spin_unlock_irqrestore(&bufferlock, flags);
+
+ if (oldcount > newsize) {
+ for (i = newsize; i < oldcount; ++i) {
+ free_page((unsigned long)old[i]);
+ }
+ }
+ kfree(old);
+
+ return 0;
+}
+
+
+/*****************************************************************************/
+/* The extended error reporting functionality */
+/*****************************************************************************/
+
+/*
+ * When a DASD device driver wants to report an error, it calls the
+ * function dasd_eer_write_trigger (via a notifier mechanism) and gives the
+ * respective trigger ID as parameter.
+ * Currently there are four kinds of triggers:
+ *
+ * DASD_EER_FATALERROR: all kinds of unrecoverable I/O problems
+ * DASD_EER_PPRCSUSPEND: PPRC was suspended
+ * DASD_EER_NOPATH: There is no path to the device left.
+ * DASD_EER_STATECHANGE: The state of the device has changed.
+ *
+ * For the first three triggers all required information can be supplied by
+ * the caller. For these triggers a record is written by the function
+ * dasd_eer_write_standard_trigger.
+ *
+ * When dasd_eer_write_trigger is called to write a DASD_EER_STATECHANGE
+ * trigger, we have to gather the necessary sense data first. We cannot queue
+ * the necessary SNSS (sense subsystem status) request immediatly, since we
+ * are likely to run in a deadlock situation. Instead, we schedule a
+ * work_struct that calls the function dasd_eer_sense_subsystem_status to
+ * create and start an SNSS request asynchronously.
+ *
+ * To avoid memory allocations at runtime, the necessary memory is allocated
+ * when the extended error reporting is enabled for a device (by
+ * dasd_eer_probe). There is one private eer data structure for each eer
+ * enabled DASD device. It contains memory for the work_struct, one SNSS cqr
+ * and a flags field that is used to coordinate the use of the cqr. The call
+ * to write a state change trigger can come in at any time, so we have one flag
+ * CQR_IN_USE that protects the cqr itself. When this flag indicates that the
+ * cqr is currently in use, dasd_eer_sense_subsystem_status cannot start a
+ * second request but sets the SNSS_REQUESTED flag instead.
+ *
+ * When the request is finished, the callback function dasd_eer_SNSS_cb
+ * is called. This function will invoke the function
+ * dasd_eer_write_SNSS_trigger to finally write the trigger. It will also
+ * check the SNSS_REQUESTED flag and if it is set it will call
+ * dasd_eer_sense_subsystem_status again.
+ *
+ * To avoid race conditions during the handling of the lock, the flags must
+ * be protected by the snsslock.
+ */
+
+struct dasd_eer_private {
+ struct dasd_ccw_req *cqr;
+ unsigned long flags;
+ struct work_struct worker;
+};
+
+static void dasd_eer_destroy(struct dasd_device *device,
+ struct dasd_eer_private *eer);
+static int
+dasd_eer_write_trigger(struct dasd_eer_trigger *trigger);
+static void dasd_eer_sense_subsystem_status(void *data);
+static int dasd_eer_notify(struct notifier_block *self,
+ unsigned long action, void *data);
+
+struct workqueue_struct *dasd_eer_workqueue;
+
+#define SNSS_DATA_SIZE 44
+static spinlock_t snsslock = SPIN_LOCK_UNLOCKED;
+
+#define DASD_EER_BUSID_SIZE 10
+struct dasd_eer_header {
+ __u32 total_size;
+ __u32 trigger;
+ __u64 tv_sec;
+ __u64 tv_usec;
+ char busid[DASD_EER_BUSID_SIZE];
+} __attribute__ ((packed));
+
+static struct notifier_block dasd_eer_nb = {
+ .notifier_call = dasd_eer_notify,
+};
+
+/*
+ * flags for use with dasd_eer_private
+ */
+#define CQR_IN_USE 0
+#define SNSS_REQUESTED 1
+
+/*
+ * This function checks if extended error reporting is available for a given
+ * dasd_device. If yes, then it creates and returns a struct dasd_eer,
+ * otherwise it returns an -EPERM error pointer.
+ */
+struct dasd_eer_private *
+dasd_eer_probe(struct dasd_device *device)
+{
+ struct dasd_eer_private *private;
+
+ if (!(device && device->discipline
+ && !strcmp(device->discipline->name, "ECKD"))) {
+ return ERR_PTR(-EPERM);
+ }
+ /* allocate the private data structure */
+ private = (struct dasd_eer_private *)kmalloc(
+ sizeof(struct dasd_eer_private), GFP_KERNEL);
+ if (!private) {
+ return ERR_PTR(-ENOMEM);
+ }
+ INIT_WORK(&private->worker, dasd_eer_sense_subsystem_status,
+ (void *)device);
+ private->cqr = dasd_kmalloc_request("ECKD",
+ 1 /* SNSS */ ,
+ SNSS_DATA_SIZE ,
+ device);
+ if (!private->cqr) {
+ kfree(private);
+ return ERR_PTR(-ENOMEM);
+ }
+ private->flags = 0;
+ return private;
+};
+
+/*
+ * If our private SNSS request is queued, remove it from the
+ * dasd ccw queue so we can free the requests memory.
+ */
+static void
+dasd_eer_dequeue_SNSS_request(struct dasd_device *device,
+ struct dasd_eer_private *eer)
+{
+ struct list_head *lst, *nxt;
+ struct dasd_ccw_req *cqr, *erpcqr;
+ dasd_erp_fn_t erp_fn;
+
+ spin_lock_irq(get_ccwdev_lock(device->cdev));
+ list_for_each_safe(lst, nxt, &device->ccw_queue) {
+ cqr = list_entry(lst, struct dasd_ccw_req, list);
+ /* we are looking for two kinds or requests */
+ /* first kind: our SNSS request: */
+ if (cqr == eer->cqr) {
+ if (cqr->status == DASD_CQR_IN_IO)
+ device->discipline->term_IO(cqr);
+ list_del(&cqr->list);
+ break;
+ }
+ /* second kind: ERP requests for our SNSS request */
+ if (cqr->refers) {
+ /* If this erp request chain ends in our cqr, then */
+ /* cal the erp_postaction to clean it up */
+ erpcqr = cqr;
+ while (erpcqr->refers) {
+ erpcqr = erpcqr->refers;
+ }
+ if (erpcqr == eer->cqr) {
+ erp_fn = device->discipline->erp_postaction(
+ cqr);
+ erp_fn(cqr);
+ }
+ continue;
+ }
+ }
+ spin_unlock_irq(get_ccwdev_lock(device->cdev));
+}
+
+/*
+ * This function dismantles a struct dasd_eer that was created by
+ * dasd_eer_probe. Since we want to free our private data structure,
+ * we must make sure that the memory is not in use anymore.
+ * We have to flush the work queue and remove a possible SNSS request
+ * from the dasd queue.
+ */
+static void
+dasd_eer_destroy(struct dasd_device *device, struct dasd_eer_private *eer)
+{
+ flush_workqueue(dasd_eer_workqueue);
+ dasd_eer_dequeue_SNSS_request(device, eer);
+ dasd_kfree_request(eer->cqr, device);
+ kfree(eer);
+};
+
+/*
+ * enable the extended error reporting for a particular device
+ */
+static int
+dasd_eer_enable_on_device(struct dasd_device *device)
+{
+ void *eer;
+ if (!device)
+ return -ENODEV;
+ if (device->eer)
+ return 0;
+ if (!try_module_get(THIS_MODULE)) {
+ return -EINVAL;
+ }
+ eer = (void *)dasd_eer_probe(device);
+ if (IS_ERR(eer)) {
+ module_put(THIS_MODULE);
+ return PTR_ERR(eer);
+ }
+ device->eer = eer;
+ return 0;
+}
+
+/*
+ * enable the extended error reporting for a particular device
+ */
+static int
+dasd_eer_disable_on_device(struct dasd_device *device)
+{
+ struct dasd_eer_private *eer = device->eer;
+
+ if (!device)
+ return -ENODEV;
+ if (!device->eer)
+ return 0;
+ device->eer = NULL;
+ dasd_eer_destroy(device,eer);
+ module_put(THIS_MODULE);
+
+ return 0;
+}
+
+/*
+ * Set extended error reporting (eer)
+ * Note: This will be registered as a DASD ioctl, to be called on DASD devices.
+ */
+static int
+dasd_ioctl_set_eer(struct block_device *bdev, int no, long args)
+{
+ struct dasd_device *device;
+ int intval;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EACCES;
+ if (bdev != bdev->bd_contains)
+ /* Error-reporting is not allowed for partitions */
+ return -EINVAL;
+ if (get_user(intval, (int __user *) args))
+ return -EFAULT;
+ device = bdev->bd_disk->private_data;
+ if (device == NULL)
+ return -ENODEV;
+
+ intval = (intval != 0);
+ DEV_MESSAGE (KERN_DEBUG, device,
+ "set eer on device to %d", intval);
+ if (intval)
+ return dasd_eer_enable_on_device(device);
+ else
+ return dasd_eer_disable_on_device(device);
+}
+
+/*
+ * Get value of extended error reporting.
+ * Note: This will be registered as a DASD ioctl, to be called on DASD devices.
+ */
+static int
+dasd_ioctl_get_eer(struct block_device *bdev, int no, long args)
+{
+ struct dasd_device *device;
+
+ device = bdev->bd_disk->private_data;
+ if (device == NULL)
+ return -ENODEV;
+ return put_user((device->eer != NULL), (int __user *) args);
+}
+
+/*
+ * The following function can be used for those triggers that have
+ * all necessary data available when the function is called.
+ * If the parameter cqr is not NULL, the chain of requests will be searched
+ * for valid sense data, and all valid sense data sets will be added to
+ * the triggers data.
+ */
+static int
+dasd_eer_write_standard_trigger(int trigger, struct dasd_device *device,
+ struct dasd_ccw_req *cqr)
+{
+ struct dasd_ccw_req *temp_cqr;
+ int data_size;
+ struct timeval tv;
+ struct dasd_eer_header header;
+ unsigned long flags;
+ struct eerbuffer *eerb;
+
+ /* go through cqr chain and count the valid sense data sets */
+ temp_cqr = cqr;
+ data_size = 0;
+ while (temp_cqr) {
+ if (temp_cqr->irb.esw.esw0.erw.cons)
+ data_size += 32;
+ temp_cqr = temp_cqr->refers;
+ }
+
+ header.total_size = sizeof(header) + data_size + 4; /* "EOR" */
+ header.trigger = trigger;
+ do_gettimeofday(&tv);
+ header.tv_sec = tv.tv_sec;
+ header.tv_usec = tv.tv_usec;
+ strncpy(header.busid, device->cdev->dev.bus_id, DASD_EER_BUSID_SIZE);
+
+ spin_lock_irqsave(&bufferlock, flags);
+ list_for_each_entry(eerb, &bufferlist, list) {
+ dasd_eer_start_record(eerb, header.total_size);
+ dasd_eer_write_buffer(eerb, sizeof(header), (char*)(&header));
+ temp_cqr = cqr;
+ while (temp_cqr) {
+ if (temp_cqr->irb.esw.esw0.erw.cons)
+ dasd_eer_write_buffer(eerb, 32, cqr->irb.ecw);
+ temp_cqr = temp_cqr->refers;
+ }
+ dasd_eer_write_buffer(eerb, 4,"EOR");
+ }
+ spin_unlock_irqrestore(&bufferlock, flags);
+
+ wake_up_interruptible(&dasd_eer_read_wait_queue);
+
+ return 0;
+}
+
+/*
+ * This function writes a DASD_EER_STATECHANGE trigger.
+ */
+static void
+dasd_eer_write_SNSS_trigger(struct dasd_device *device,
+ struct dasd_ccw_req *cqr)
+{
+ int data_size;
+ int snss_rc;
+ struct timeval tv;
+ struct dasd_eer_header header;
+ unsigned long flags;
+ struct eerbuffer *eerb;
+
+ snss_rc = (cqr->status == DASD_CQR_FAILED) ? -EIO : 0;
+ if (snss_rc)
+ data_size = 0;
+ else
+ data_size = SNSS_DATA_SIZE;
+
+ header.total_size = sizeof(header) + data_size + 4; /* "EOR" */
+ header.trigger = DASD_EER_STATECHANGE;
+ do_gettimeofday(&tv);
+ header.tv_sec = tv.tv_sec;
+ header.tv_usec = tv.tv_usec;
+ strncpy(header.busid, device->cdev->dev.bus_id, DASD_EER_BUSID_SIZE);
+
+ spin_lock_irqsave(&bufferlock, flags);
+ list_for_each_entry(eerb, &bufferlist, list) {
+ dasd_eer_start_record(eerb, header.total_size);
+ dasd_eer_write_buffer(eerb, sizeof(header),(char*)(&header));
+ if (!snss_rc)
+ dasd_eer_write_buffer(eerb, SNSS_DATA_SIZE, cqr->data);
+ dasd_eer_write_buffer(eerb, 4,"EOR");
+ }
+ spin_unlock_irqrestore(&bufferlock, flags);
+
+ wake_up_interruptible(&dasd_eer_read_wait_queue);
+}
+
+/*
+ * callback function for use with SNSS request
+ */
+static void
+dasd_eer_SNSS_cb(struct dasd_ccw_req *cqr, void *data)
+{
+ struct dasd_device *device;
+ struct dasd_eer_private *private;
+ unsigned long irqflags;
+
+ device = (struct dasd_device *)data;
+ private = (struct dasd_eer_private *)device->eer;
+ dasd_eer_write_SNSS_trigger(device, cqr);
+ spin_lock_irqsave(&snsslock, irqflags);
+ if(!test_and_clear_bit(SNSS_REQUESTED, &private->flags)) {
+ clear_bit(CQR_IN_USE, &private->flags);
+ spin_unlock_irqrestore(&snsslock, irqflags);
+ return;
+ };
+ clear_bit(CQR_IN_USE, &private->flags);
+ spin_unlock_irqrestore(&snsslock, irqflags);
+ dasd_eer_sense_subsystem_status(device);
+ return;
+}
+
+/*
+ * clean a used cqr before using it again
+ */
+static void
+dasd_eer_clean_SNSS_request(struct dasd_ccw_req *cqr)
+{
+ struct ccw1 *cpaddr = cqr->cpaddr;
+ void *data = cqr->data;
+
+ memset(cqr, 0, sizeof(struct dasd_ccw_req));
+ memset(cpaddr, 0, sizeof(struct ccw1));
+ memset(data, 0, SNSS_DATA_SIZE);
+ cqr->cpaddr = cpaddr;
+ cqr->data = data;
+ strncpy((char *) &cqr->magic, "ECKD", 4);
+ ASCEBC((char *) &cqr->magic, 4);
+ set_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags);
+}
+
+/*
+ * build and start an SNSS request
+ * This function is called from a work queue so we have to
+ * pass the dasd_device pointer as a void pointer.
+ */
+static void
+dasd_eer_sense_subsystem_status(void *data)
+{
+ struct dasd_device *device;
+ struct dasd_eer_private *private;
+ struct dasd_ccw_req *cqr;
+ struct ccw1 *ccw;
+ unsigned long irqflags;
+
+ device = (struct dasd_device *)data;
+ private = (struct dasd_eer_private *)device->eer;
+ if (!private) /* device not eer enabled any more */
+ return;
+ cqr = private->cqr;
+ spin_lock_irqsave(&snsslock, irqflags);
+ if(test_and_set_bit(CQR_IN_USE, &private->flags)) {
+ set_bit(SNSS_REQUESTED, &private->flags);
+ spin_unlock_irqrestore(&snsslock, irqflags);
+ return;
+ };
+ spin_unlock_irqrestore(&snsslock, irqflags);
+ dasd_eer_clean_SNSS_request(cqr);
+ cqr->device = device;
+ cqr->retries = 255;
+ cqr->expires = 10 * HZ;
+
+ ccw = cqr->cpaddr;
+ ccw->cmd_code = DASD_ECKD_CCW_SNSS;
+ ccw->count = SNSS_DATA_SIZE;
+ ccw->flags = 0;
+ ccw->cda = (__u32)(addr_t)cqr->data;
+
+ cqr->buildclk = get_clock();
+ cqr->status = DASD_CQR_FILLED;
+ cqr->callback = dasd_eer_SNSS_cb;
+ cqr->callback_data = (void *)device;
+ dasd_add_request_head(cqr);
+
+ return;
+}
+
+/*
+ * This function is called for all triggers. It calls the appropriate
+ * function that writes the actual trigger records.
+ */
+static int
+dasd_eer_write_trigger(struct dasd_eer_trigger *trigger)
+{
+ int rc;
+ struct dasd_eer_private *private = trigger->device->eer;
+
+ switch (trigger->id) {
+ case DASD_EER_FATALERROR:
+ case DASD_EER_PPRCSUSPEND:
+ rc = dasd_eer_write_standard_trigger(
+ trigger->id, trigger->device, trigger->cqr);
+ break;
+ case DASD_EER_NOPATH:
+ rc = dasd_eer_write_standard_trigger(
+ trigger->id, trigger->device, NULL);
+ break;
+ case DASD_EER_STATECHANGE:
+ if (queue_work(dasd_eer_workqueue, &private->worker)) {
+ rc=0;
+ } else {
+ /* If the work_struct was already queued, it can't
+ * be queued again. But this is OK since we don't
+ * need to have it queued twice.
+ */
+ rc = -EBUSY;
+ }
+ break;
+ default: /* unknown trigger, so we write it without any sense data */
+ rc = dasd_eer_write_standard_trigger(
+ trigger->id, trigger->device, NULL);
+ break;
+ }
+ return rc;
+}
+
+/*
+ * This function is registered with the dasd device driver and gets called
+ * for all dasd eer notifications.
+ */
+static int dasd_eer_notify(struct notifier_block *self,
+ unsigned long action, void *data)
+{
+ switch (action) {
+ case DASD_EER_DISABLE:
+ dasd_eer_disable_on_device((struct dasd_device *)data);
+ break;
+ case DASD_EER_TRIGGER:
+ dasd_eer_write_trigger((struct dasd_eer_trigger *)data);
+ break;
+ }
+ return NOTIFY_OK;
+}
+
+
+/*****************************************************************************/
+/* the device operations */
+/*****************************************************************************/
+
+/*
+ * On the one side we need a lock to access our internal buffer, on the
+ * other side a copy_to_user can sleep. So we need to copy the data we have
+ * to transfer in a readbuffer, which is protected by the readbuffer_mutex.
+ */
+static char readbuffer[PAGE_SIZE];
+DECLARE_MUTEX(readbuffer_mutex);
+
+
+static int
+dasd_eer_open(struct inode *inp, struct file *filp)
+{
+ struct eerbuffer *eerb;
+ unsigned long flags;
+
+ eerb = kmalloc(sizeof(struct eerbuffer), GFP_KERNEL);
+ eerb->head = 0;
+ eerb->tail = 0;
+ eerb->residual = 0;
+ eerb->buffer_page_count = 1;
+ eerb->buffersize = eerb->buffer_page_count * PAGE_SIZE;
+ eerb->buffer = kmalloc(eerb->buffer_page_count*sizeof(char*),
+ GFP_KERNEL);
+ if (!eerb->buffer)
+ return -ENOMEM;
+ if (dasd_eer_allocate_buffer_pages(eerb->buffer,
+ eerb->buffer_page_count)) {
+ kfree(eerb->buffer);
+ return -ENOMEM;
+ }
+ filp->private_data = eerb;
+ spin_lock_irqsave(&bufferlock, flags);
+ list_add(&eerb->list, &bufferlist);
+ spin_unlock_irqrestore(&bufferlock, flags);
+
+ return nonseekable_open(inp,filp);
+}
+
+static int
+dasd_eer_close(struct inode *inp, struct file *filp)
+{
+ struct eerbuffer *eerb;
+ unsigned long flags;
+
+ eerb = (struct eerbuffer *)filp->private_data;
+ spin_lock_irqsave(&bufferlock, flags);
+ list_del(&eerb->list);
+ spin_unlock_irqrestore(&bufferlock, flags);
+ dasd_eer_free_buffer_pages(eerb->buffer, eerb->buffer_page_count);
+ kfree(eerb->buffer);
+ kfree(eerb);
+
+ return 0;
+}
+
+static long
+dasd_eer_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
+{
+ int intval;
+ struct eerbuffer *eerb;
+
+ eerb = (struct eerbuffer *)filp->private_data;
+ switch (cmd) {
+ case DASD_EER_PURGE:
+ dasd_eer_purge_buffer(eerb);
+ return 0;
+ case DASD_EER_SETBUFSIZE:
+ if (get_user(intval, (int __user *)arg))
+ return -EFAULT;
+ return dasd_eer_resize_buffer(eerb, intval);
+ default:
+ return -ENOIOCTLCMD;
+ }
+}
+
+static ssize_t
+dasd_eer_read(struct file *filp, char __user *buf, size_t count, loff_t *ppos)
+{
+ int tc,rc;
+ int tailcount,effective_count;
+ unsigned long flags;
+ struct eerbuffer *eerb;
+
+ eerb = (struct eerbuffer *)filp->private_data;
+ if(down_interruptible(&readbuffer_mutex))
+ return -ERESTARTSYS;
+
+ spin_lock_irqsave(&bufferlock, flags);
+
+ if (eerb->residual < 0) { /* the remainder of this record */
+ /* has been deleted */
+ eerb->residual = 0;
+ spin_unlock_irqrestore(&bufferlock, flags);
+ up(&readbuffer_mutex);
+ return -EIO;
+ } else if (eerb->residual > 0) {
+ /* OK we still have a second half of a record to deliver */
+ effective_count = min(eerb->residual, (int)count);
+ eerb->residual -= effective_count;
+ } else {
+ tc = 0;
+ while (!tc) {
+ tc = dasd_eer_read_buffer(eerb,
+ sizeof(tailcount), (char*)(&tailcount));
+ if (!tc) {
+ /* no data available */
+ spin_unlock_irqrestore(&bufferlock, flags);
+ up(&readbuffer_mutex);
+ if (filp->f_flags & O_NONBLOCK)
+ return -EAGAIN;
+ rc = wait_event_interruptible(
+ dasd_eer_read_wait_queue,
+ eerb->head != eerb->tail);
+ if (rc) {
+ return rc;
+ }
+ if(down_interruptible(&readbuffer_mutex))
+ return -ERESTARTSYS;
+ spin_lock_irqsave(&bufferlock, flags);
+ }
+ }
+ WARN_ON(tc != sizeof(tailcount));
+ effective_count = min(tailcount,(int)count);
+ eerb->residual = tailcount - effective_count;
+ }
+
+ tc = dasd_eer_read_buffer(eerb, effective_count, readbuffer);
+ WARN_ON(tc != effective_count);
+
+ spin_unlock_irqrestore(&bufferlock, flags);
+
+ if (copy_to_user(buf, readbuffer, effective_count)) {
+ up(&readbuffer_mutex);
+ return -EFAULT;
+ }
+
+ up(&readbuffer_mutex);
+ return effective_count;
+}
+
+static unsigned int
+dasd_eer_poll (struct file *filp, poll_table *ptable)
+{
+ unsigned int mask;
+ unsigned long flags;
+ struct eerbuffer *eerb;
+
+ eerb = (struct eerbuffer *)filp->private_data;
+ poll_wait(filp, &dasd_eer_read_wait_queue, ptable);
+ spin_lock_irqsave(&bufferlock, flags);
+ if (eerb->head != eerb->tail)
+ mask = POLLIN | POLLRDNORM ;
+ else
+ mask = 0;
+ spin_unlock_irqrestore(&bufferlock, flags);
+ return mask;
+}
+
+static struct file_operations dasd_eer_fops = {
+ .open = &dasd_eer_open,
+ .release = &dasd_eer_close,
+ .unlocked_ioctl = &dasd_eer_ioctl,
+ .compat_ioctl = &dasd_eer_ioctl,
+ .read = &dasd_eer_read,
+ .poll = &dasd_eer_poll,
+ .owner = THIS_MODULE,
+};
+
+static struct miscdevice dasd_eer_dev = {
+ .minor = MISC_DYNAMIC_MINOR,
+ .name = "dasd_eer",
+ .fops = &dasd_eer_fops,
+};
+
+
+/*****************************************************************************/
+/* Init and exit */
+/*****************************************************************************/
+
+static int
+__init dasd_eer_init(void)
+{
+ int rc;
+
+ dasd_eer_workqueue = create_singlethread_workqueue("dasd_eer");
+ if (!dasd_eer_workqueue) {
+ MESSAGE(KERN_ERR , "%s", "dasd_eer_init could not "
+ "create workqueue \n");
+ rc = -ENOMEM;
+ goto out;
+ }
+
+ rc = dasd_register_eer_notifier(&dasd_eer_nb);
+ if (rc) {
+ MESSAGE(KERN_ERR, "%s", "dasd_eer_init could not "
+ "register error reporting");
+ goto queue;
+ }
+
+ dasd_ioctl_no_register(THIS_MODULE, BIODASDEERSET, dasd_ioctl_set_eer);
+ dasd_ioctl_no_register(THIS_MODULE, BIODASDEERGET, dasd_ioctl_get_eer);
+
+ /* we don't need our own character device,
+ * so we just register as misc device */
+ rc = misc_register(&dasd_eer_dev);
+ if (rc) {
+ MESSAGE(KERN_ERR, "%s", "dasd_eer_init could not "
+ "register misc device");
+ goto unregister;
+ }
+
+ return 0;
+
+unregister:
+ dasd_unregister_eer_notifier(&dasd_eer_nb);
+ dasd_ioctl_no_unregister(THIS_MODULE, BIODASDEERSET,
+ dasd_ioctl_set_eer);
+ dasd_ioctl_no_unregister(THIS_MODULE, BIODASDEERGET,
+ dasd_ioctl_get_eer);
+queue:
+ destroy_workqueue(dasd_eer_workqueue);
+out:
+ return rc;
+
+}
+module_init(dasd_eer_init);
+
+static void
+__exit dasd_eer_exit(void)
+{
+ dasd_unregister_eer_notifier(&dasd_eer_nb);
+ dasd_ioctl_no_unregister(THIS_MODULE, BIODASDEERSET,
+ dasd_ioctl_set_eer);
+ dasd_ioctl_no_unregister(THIS_MODULE, BIODASDEERGET,
+ dasd_ioctl_get_eer);
+ destroy_workqueue(dasd_eer_workqueue);
+
+ WARN_ON(misc_deregister(&dasd_eer_dev) != 0);
+}
+module_exit(dasd_eer_exit);
diff -urpN linux-2.6/drivers/s390/block/dasd_int.h linux-2.6-patched/drivers/s390/block/dasd_int.h
--- linux-2.6/drivers/s390/block/dasd_int.h 2006-02-01 11:03:41.000000000 +0100
+++ linux-2.6-patched/drivers/s390/block/dasd_int.h 2006-02-01 11:04:09.000000000 +0100
@@ -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_d
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;
diff -urpN linux-2.6/drivers/s390/block/Kconfig linux-2.6-patched/drivers/s390/block/Kconfig
--- linux-2.6/drivers/s390/block/Kconfig 2006-02-01 11:02:58.000000000 +0100
+++ linux-2.6-patched/drivers/s390/block/Kconfig 2006-02-01 11:04:09.000000000 +0100
@@ -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.

diff -urpN linux-2.6/drivers/s390/block/Makefile linux-2.6-patched/drivers/s390/block/Makefile
--- linux-2.6/drivers/s390/block/Makefile 2006-01-03 04:21:10.000000000 +0100
+++ linux-2.6-patched/drivers/s390/block/Makefile 2006-02-01 11:04:09.000000000 +0100
@@ -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-$(CONFIG_DASD_DIAG) += dasd_diag_mod
obj-$(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
diff -urpN linux-2.6/include/asm-s390/dasd.h linux-2.6-patched/include/asm-s390/dasd.h
--- linux-2.6/include/asm-s390/dasd.h 2006-02-01 11:03:41.000000000 +0100
+++ linux-2.6-patched/include/asm-s390/dasd.h 2006-02-01 11:04:09.000000000 +0100
@@ -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 */


2006-02-04 13:14:35

by Christoph Hellwig

[permalink] [raw]
Subject: Re: [PATCH 1/3] s390: dasd extended error reporting module.

On Wed, Feb 01, 2006 at 12:56:49PM +0100, Heiko Carstens wrote:
> From: Stefan Weinhuber <[email protected]>
>
> 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.
> This is a resubmit of this patch because at first submission it
> didn't get included due to Christoph's ioctl changes.
> Since these aren't in the -mm tree anymore this one should be
> merged now.

NACK again. (although andrew unfortunately pushed this to Linus
already, it would be good if this went out again before 2.6.16).

If you really want this functionality in a separate module you need
to call try_module_get on the module structure of the eer module
everytime before calling into it. You need to register a structure
ala

struct dasd_eer_ops {
struct module *owner;
void (*disable_on_device)(struct dasd_device *);
void (*write_trigger)(struct dasd_eer_trigger *);
};

and then have wrappers in the core dasd code ala

eer_disable_on_device(struct dasd_device *dev)
{
if (try_module_get(eer_ops->owner)) {
eer_ops->write_trigger(dev);
module_put(eer_ops->owner);
}
}

the module_get/put on THIS_MODULE in the patch are completely broken.

the ioctl registration is also totally not acceptable. firtly this
interface is absolutely suitable for sysfs, although that might only work
if it's in the main module. But even if you go for ioctls it should be
on your separate char device and not use the broken dynamic ioctl
registration which _must_ not be used in any new code.

Besides that there's tons of the usual style issues on the code, please
use mutexes, please make lots of variables that could be static, etc.

2006-02-06 11:30:43

by Jan Blunck

[permalink] [raw]
Subject: Re: [PATCH 1/3] s390: dasd extended error reporting module.

On Wed, Feb 01, Heiko Carstens wrote:

> From: Stefan Weinhuber <[email protected]>
>
> 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.
> This is a resubmit of this patch because at first submission it
> didn't get included due to Christoph's ioctl changes.
> Since these aren't in the -mm tree anymore this one should be
> merged now.

Why don't you use the sysfs for this purpose? This new character device
interface seems very odd to me. Why don't you introduce new attributes to the
dasd device for that purpose and make online pollable for failovers?

Or use dm-netlink to report the extended errors via multipath to the user
space.

Regards,
Jan

--
Jan Blunck [email protected]
SuSE LINUX AG - A Novell company
Maxfeldstr. 5 +49-911-74053-608
D-90409 N?rnberg http://www.suse.de

2006-02-06 18:15:50

by Stefan Weinhuber

[permalink] [raw]
Subject: Re: [PATCH 1/3] s390: dasd extended error reporting module.

Jan Blunck <[email protected]> wrote on 06.02.2006 12:30:46:

> Why don't you use the sysfs for this purpose? This new character device
> interface seems very odd to me. Why don't you introduce new attributes
to the
> dasd device for that purpose and make online pollable for failovers?

We need to transfer binary data records, and in case of an error
situation probably quite a number of them. Maybe sysfs could be bent to
that purpose but to me this seems even more awkward then a character
device.

> Or use dm-netlink to report the extended errors via multipath to the
user
> space.

Our interface is supposed to work under memory constrained situations
e.g. when the connection to the storage device failed. With our character
device we can constrain memory allocations to uncritical situations:
- when error reporting is enabled
- when the device is opened
- when the internal buffer size is adjusted
The error reporting itself is done without any additional allocations.

Best Regards / Mit freundlichen Gr??en

Stefan Weinhuber

-------------------------------------------------------------------
IBM Deutschland Entwicklung GmbH
Linux for zSeries Development & Services

2006-02-06 18:23:48

by Stefan Weinhuber

[permalink] [raw]
Subject: Re: [PATCH 1/3] s390: dasd extended error reporting module.

Christoph Hellwig <[email protected]> wrote on 04.02.2006 14:14:19:
>
> NACK again. (although andrew unfortunately pushed this to Linus
> already, it would be good if this went out again before 2.6.16).
>
> If you really want this functionality in a separate module you need
> to call try_module_get on the module structure of the eer module
> everytime before calling into it. You need to register a structure
> ala
>
> struct dasd_eer_ops {
> struct module *owner;
> void (*disable_on_device)(struct dasd_device *);
> void (*write_trigger)(struct dasd_eer_trigger *);
> };
>
> and then have wrappers in the core dasd code ala
>
> eer_disable_on_device(struct dasd_device *dev)
> {
> if (try_module_get(eer_ops->owner)) {
> eer_ops->write_trigger(dev);
> module_put(eer_ops->owner);
> }
> }
>
> the module_get/put on THIS_MODULE in the patch are completely broken.

For the dynamically registered ioctls there is exactly such a
concept in place. So a try_module_get is done _before_ the
dasd_ioctl_set_eer function is called. That function again calls
dasd_eer_enable_on_device which does the try_module_get you complain
about. So, for each device on which the error reporing is enabled, a
try_module_get is done, and it should be safe for those devices to
call the respective notifier functions later on.

>
> the ioctl registration is also totally not acceptable. firtly this
> interface is absolutely suitable for sysfs, although that might only
work
> if it's in the main module. But even if you go for ioctls it should be
> on your separate char device and not use the broken dynamic ioctl
> registration which _must_ not be used in any new code.

I know that you don't like that dynamic ioctl interface, but I would
like to leave that discussion to Horst.
But besides that, the ioctl in question is needed on the DASD device,
since we want to enable or disable error reporting per DASD device.

Best Regards / Mit freundlichen Gr??en

Stefan Weinhuber

-------------------------------------------------------------------
IBM Deutschland Entwicklung GmbH
Linux for zSeries Development & Services