/*
* Copyright (C) 2003 Sistina Software
+ * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
*
* This file is released under the LGPL.
*/
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/vmalloc.h>
+#include <linux/dm-io.h>
+#include <linux/dm-dirty-log.h>
-#include "dm-log.h"
-#include "dm-io.h"
+#include <linux/device-mapper.h>
+
+#define DM_MSG_PREFIX "dirty region log"
static LIST_HEAD(_log_types);
static DEFINE_SPINLOCK(_lock);
-int dm_register_dirty_log_type(struct dirty_log_type *type)
+static struct dm_dirty_log_type *__find_dirty_log_type(const char *name)
{
- spin_lock(&_lock);
- type->use_count = 0;
- list_add(&type->list, &_log_types);
- spin_unlock(&_lock);
+ struct dm_dirty_log_type *log_type;
- return 0;
+ list_for_each_entry(log_type, &_log_types, list)
+ if (!strcmp(name, log_type->name))
+ return log_type;
+
+ return NULL;
}
-int dm_unregister_dirty_log_type(struct dirty_log_type *type)
+static struct dm_dirty_log_type *_get_dirty_log_type(const char *name)
{
+ struct dm_dirty_log_type *log_type;
+
spin_lock(&_lock);
- if (type->use_count)
- DMWARN("Attempt to unregister a log type that is still in use");
- else
- list_del(&type->list);
+ log_type = __find_dirty_log_type(name);
+ if (log_type && !try_module_get(log_type->module))
+ log_type = NULL;
spin_unlock(&_lock);
- return 0;
+ return log_type;
}
-static struct dirty_log_type *get_type(const char *type_name)
+/*
+ * get_type
+ * @type_name
+ *
+ * Attempt to retrieve the dm_dirty_log_type by name. If not already
+ * available, attempt to load the appropriate module.
+ *
+ * Log modules are named "dm-log-" followed by the 'type_name'.
+ * Modules may contain multiple types.
+ * This function will first try the module "dm-log-<type_name>",
+ * then truncate 'type_name' on the last '-' and try again.
+ *
+ * For example, if type_name was "clustered-disk", it would search
+ * 'dm-log-clustered-disk' then 'dm-log-clustered'.
+ *
+ * Returns: dirty_log_type* on success, NULL on failure
+ */
+static struct dm_dirty_log_type *get_type(const char *type_name)
{
- struct dirty_log_type *type;
+ char *p, *type_name_dup;
+ struct dm_dirty_log_type *log_type;
+
+ if (!type_name)
+ return NULL;
+
+ log_type = _get_dirty_log_type(type_name);
+ if (log_type)
+ return log_type;
+
+ type_name_dup = kstrdup(type_name, GFP_KERNEL);
+ if (!type_name_dup) {
+ DMWARN("No memory left to attempt log module load for \"%s\"",
+ type_name);
+ return NULL;
+ }
+
+ while (request_module("dm-log-%s", type_name_dup) ||
+ !(log_type = _get_dirty_log_type(type_name))) {
+ p = strrchr(type_name_dup, '-');
+ if (!p)
+ break;
+ p[0] = '\0';
+ }
+
+ if (!log_type)
+ DMWARN("Module for logging type \"%s\" not found.", type_name);
+
+ kfree(type_name_dup);
+
+ return log_type;
+}
+
+static void put_type(struct dm_dirty_log_type *type)
+{
+ if (!type)
+ return;
spin_lock(&_lock);
- list_for_each_entry (type, &_log_types, list)
- if (!strcmp(type_name, type->name)) {
- if (!type->use_count && !try_module_get(type->module)){
- spin_unlock(&_lock);
- return NULL;
- }
- type->use_count++;
- spin_unlock(&_lock);
- return type;
- }
+ if (!__find_dirty_log_type(type->name))
+ goto out;
+ module_put(type->module);
+
+out:
spin_unlock(&_lock);
- return NULL;
}
-static void put_type(struct dirty_log_type *type)
+int dm_dirty_log_type_register(struct dm_dirty_log_type *type)
{
+ int r = 0;
+
spin_lock(&_lock);
- if (!--type->use_count)
- module_put(type->module);
+ if (!__find_dirty_log_type(type->name))
+ list_add(&type->list, &_log_types);
+ else
+ r = -EEXIST;
spin_unlock(&_lock);
+
+ return r;
}
+EXPORT_SYMBOL(dm_dirty_log_type_register);
-struct dirty_log *dm_create_dirty_log(const char *type_name, struct dm_target *ti,
- unsigned int argc, char **argv)
+int dm_dirty_log_type_unregister(struct dm_dirty_log_type *type)
{
- struct dirty_log_type *type;
- struct dirty_log *log;
+ spin_lock(&_lock);
+
+ if (!__find_dirty_log_type(type->name)) {
+ spin_unlock(&_lock);
+ return -EINVAL;
+ }
+
+ list_del(&type->list);
+
+ spin_unlock(&_lock);
+
+ return 0;
+}
+EXPORT_SYMBOL(dm_dirty_log_type_unregister);
+
+struct dm_dirty_log *dm_dirty_log_create(const char *type_name,
+ struct dm_target *ti,
+ unsigned int argc, char **argv)
+{
+ struct dm_dirty_log_type *type;
+ struct dm_dirty_log *log;
log = kmalloc(sizeof(*log), GFP_KERNEL);
if (!log)
return log;
}
+EXPORT_SYMBOL(dm_dirty_log_create);
-void dm_destroy_dirty_log(struct dirty_log *log)
+void dm_dirty_log_destroy(struct dm_dirty_log *log)
{
log->type->dtr(log);
put_type(log->type);
kfree(log);
}
+EXPORT_SYMBOL(dm_dirty_log_destroy);
/*-----------------------------------------------------------------
* Persistent and core logs share a lot of their implementation.
FORCESYNC, /* Force a sync to happen */
} sync;
+ struct dm_io_request io_req;
+
/*
* Disk log fields
*/
+ int log_dev_failed;
struct dm_dev *log_dev;
struct log_header header;
- struct io_region header_location;
+ struct dm_io_region header_location;
struct log_header *disk_header;
};
* The touched member needs to be updated every time we access
* one of the bitsets.
*/
-static inline int log_test_bit(uint32_t *bs, unsigned bit)
+static inline int log_test_bit(uint32_t *bs, unsigned bit)
{
return ext2_test_bit(bit, (unsigned long *) bs) ? 1 : 0;
}
core->nr_regions = le64_to_cpu(disk->nr_regions);
}
+static int rw_header(struct log_c *lc, int rw)
+{
+ lc->io_req.bi_rw = rw;
+
+ return dm_io(&lc->io_req, 1, &lc->header_location, NULL);
+}
+
static int read_header(struct log_c *log)
{
int r;
- unsigned long ebits;
- r = dm_io_sync_vm(1, &log->header_location, READ,
- log->disk_header, &ebits);
+ r = rw_header(log, READ);
if (r)
return r;
return 0;
}
-static inline int write_header(struct log_c *log)
+static int _check_region_size(struct dm_target *ti, uint32_t region_size)
{
- unsigned long ebits;
+ if (region_size < 2 || region_size > ti->len)
+ return 0;
+
+ if (!is_power_of_2(region_size))
+ return 0;
- header_to_disk(&log->header, log->disk_header);
- return dm_io_sync_vm(1, &log->header_location, WRITE,
- log->disk_header, &ebits);
+ return 1;
}
/*----------------------------------------------------------------
* argv contains region_size followed optionally by [no]sync
*--------------------------------------------------------------*/
#define BYTE_SHIFT 3
-static int create_log_context(struct dirty_log *log, struct dm_target *ti,
+static int create_log_context(struct dm_dirty_log *log, struct dm_target *ti,
unsigned int argc, char **argv,
struct dm_dev *dev)
{
uint32_t region_size;
unsigned int region_count;
size_t bitset_size, buf_size;
+ int r;
if (argc < 1 || argc > 2) {
- DMWARN("wrong number of arguments to mirror log");
+ DMWARN("wrong number of arguments to dirty region log");
return -EINVAL;
}
else if (!strcmp(argv[1], "nosync"))
sync = NOSYNC;
else {
- DMWARN("unrecognised sync argument to mirror log: %s",
- argv[1]);
+ DMWARN("unrecognised sync argument to "
+ "dirty region log: %s", argv[1]);
return -EINVAL;
}
}
- if (sscanf(argv[0], "%u", ®ion_size) != 1) {
- DMWARN("invalid region size string");
+ if (sscanf(argv[0], "%u", ®ion_size) != 1 ||
+ !_check_region_size(ti, region_size)) {
+ DMWARN("invalid region size %s", argv[0]);
return -EINVAL;
}
lc->disk_header = NULL;
} else {
lc->log_dev = dev;
+ lc->log_dev_failed = 0;
lc->header_location.bdev = lc->log_dev->bdev;
lc->header_location.sector = 0;
/*
* Buffer holds both header and bitset.
*/
- buf_size = dm_round_up((LOG_OFFSET << SECTOR_SHIFT) +
- bitset_size, ti->limits.hardsect_size);
+ buf_size =
+ dm_round_up((LOG_OFFSET << SECTOR_SHIFT) + bitset_size,
+ bdev_logical_block_size(lc->header_location.
+ bdev));
+
+ if (buf_size > i_size_read(dev->bdev->bd_inode)) {
+ DMWARN("log device %s too small: need %llu bytes",
+ dev->name, (unsigned long long)buf_size);
+ kfree(lc);
+ return -EINVAL;
+ }
+
lc->header_location.count = buf_size >> SECTOR_SHIFT;
+ lc->io_req.mem.type = DM_IO_VMA;
+ lc->io_req.notify.fn = NULL;
+ lc->io_req.client = dm_io_client_create(dm_div_up(buf_size,
+ PAGE_SIZE));
+ if (IS_ERR(lc->io_req.client)) {
+ r = PTR_ERR(lc->io_req.client);
+ DMWARN("couldn't allocate disk io client");
+ kfree(lc);
+ return -ENOMEM;
+ }
+
lc->disk_header = vmalloc(buf_size);
if (!lc->disk_header) {
DMWARN("couldn't allocate disk log buffer");
+ dm_io_client_destroy(lc->io_req.client);
kfree(lc);
return -ENOMEM;
}
+ lc->io_req.mem.ptr.vma = lc->disk_header;
lc->clean_bits = (void *)lc->disk_header +
(LOG_OFFSET << SECTOR_SHIFT);
}
DMWARN("couldn't allocate sync bitset");
if (!dev)
vfree(lc->clean_bits);
+ else
+ dm_io_client_destroy(lc->io_req.client);
vfree(lc->disk_header);
kfree(lc);
return -ENOMEM;
vfree(lc->sync_bits);
if (!dev)
vfree(lc->clean_bits);
+ else
+ dm_io_client_destroy(lc->io_req.client);
vfree(lc->disk_header);
kfree(lc);
return -ENOMEM;
return 0;
}
-static int core_ctr(struct dirty_log *log, struct dm_target *ti,
+static int core_ctr(struct dm_dirty_log *log, struct dm_target *ti,
unsigned int argc, char **argv)
{
return create_log_context(log, ti, argc, argv, NULL);
kfree(lc);
}
-static void core_dtr(struct dirty_log *log)
+static void core_dtr(struct dm_dirty_log *log)
{
struct log_c *lc = (struct log_c *) log->context;
*
* argv contains log_device region_size followed optionally by [no]sync
*--------------------------------------------------------------*/
-static int disk_ctr(struct dirty_log *log, struct dm_target *ti,
+static int disk_ctr(struct dm_dirty_log *log, struct dm_target *ti,
unsigned int argc, char **argv)
{
int r;
struct dm_dev *dev;
if (argc < 2 || argc > 3) {
- DMWARN("wrong number of arguments to disk mirror log");
+ DMWARN("wrong number of arguments to disk dirty region log");
return -EINVAL;
}
return 0;
}
-static void disk_dtr(struct dirty_log *log)
+static void disk_dtr(struct dm_dirty_log *log)
{
struct log_c *lc = (struct log_c *) log->context;
dm_put_device(lc->ti, lc->log_dev);
vfree(lc->disk_header);
+ dm_io_client_destroy(lc->io_req.client);
destroy_log_context(lc);
}
return count;
}
-static int disk_resume(struct dirty_log *log)
+static void fail_log_device(struct log_c *lc)
+{
+ if (lc->log_dev_failed)
+ return;
+
+ lc->log_dev_failed = 1;
+ dm_table_event(lc->ti->table);
+}
+
+static int disk_resume(struct dm_dirty_log *log)
{
int r;
unsigned i;
/* read the disk header */
r = read_header(lc);
- if (r)
- return r;
+ if (r) {
+ DMWARN("%s: Failed to read header on dirty region log device",
+ lc->log_dev->name);
+ fail_log_device(lc);
+ /*
+ * If the log device cannot be read, we must assume
+ * all regions are out-of-sync. If we simply return
+ * here, the state will be uninitialized and could
+ * lead us to return 'in-sync' status for regions
+ * that are actually 'out-of-sync'.
+ */
+ lc->header.nr_regions = 0;
+ }
- /* set or clear any new bits */
+ /* set or clear any new bits -- device has grown */
if (lc->sync == NOSYNC)
for (i = lc->header.nr_regions; i < lc->region_count; i++)
/* FIXME: amazingly inefficient */
/* FIXME: amazingly inefficient */
log_clear_bit(lc, lc->clean_bits, i);
+ /* clear any old bits -- device has shrunk */
+ for (i = lc->region_count; i % (sizeof(*lc->clean_bits) << BYTE_SHIFT); i++)
+ log_clear_bit(lc, lc->clean_bits, i);
+
/* copy clean across to sync */
memcpy(lc->sync_bits, lc->clean_bits, size);
lc->sync_count = count_bits32(lc->clean_bits, lc->bitset_uint32_count);
+ lc->sync_search = 0;
/* set the correct number of regions in the header */
lc->header.nr_regions = lc->region_count;
+ header_to_disk(&lc->header, lc->disk_header);
+
/* write the new header */
- return write_header(lc);
+ r = rw_header(lc, WRITE);
+ if (r) {
+ DMWARN("%s: Failed to write header on dirty region log device",
+ lc->log_dev->name);
+ fail_log_device(lc);
+ }
+
+ return r;
}
-static uint32_t core_get_region_size(struct dirty_log *log)
+static uint32_t core_get_region_size(struct dm_dirty_log *log)
{
struct log_c *lc = (struct log_c *) log->context;
return lc->region_size;
}
-static int core_is_clean(struct dirty_log *log, region_t region)
+static int core_resume(struct dm_dirty_log *log)
+{
+ struct log_c *lc = (struct log_c *) log->context;
+ lc->sync_search = 0;
+ return 0;
+}
+
+static int core_is_clean(struct dm_dirty_log *log, region_t region)
{
struct log_c *lc = (struct log_c *) log->context;
return log_test_bit(lc->clean_bits, region);
}
-static int core_in_sync(struct dirty_log *log, region_t region, int block)
+static int core_in_sync(struct dm_dirty_log *log, region_t region, int block)
{
struct log_c *lc = (struct log_c *) log->context;
return log_test_bit(lc->sync_bits, region);
}
-static int core_flush(struct dirty_log *log)
+static int core_flush(struct dm_dirty_log *log)
{
/* no op */
return 0;
}
-static int disk_flush(struct dirty_log *log)
+static int disk_flush(struct dm_dirty_log *log)
{
int r;
struct log_c *lc = (struct log_c *) log->context;
if (!lc->touched)
return 0;
- r = write_header(lc);
- if (!r)
+ r = rw_header(lc, WRITE);
+ if (r)
+ fail_log_device(lc);
+ else
lc->touched = 0;
return r;
}
-static void core_mark_region(struct dirty_log *log, region_t region)
+static void core_mark_region(struct dm_dirty_log *log, region_t region)
{
struct log_c *lc = (struct log_c *) log->context;
log_clear_bit(lc, lc->clean_bits, region);
}
-static void core_clear_region(struct dirty_log *log, region_t region)
+static void core_clear_region(struct dm_dirty_log *log, region_t region)
{
struct log_c *lc = (struct log_c *) log->context;
log_set_bit(lc, lc->clean_bits, region);
}
-static int core_get_resync_work(struct dirty_log *log, region_t *region)
+static int core_get_resync_work(struct dm_dirty_log *log, region_t *region)
{
struct log_c *lc = (struct log_c *) log->context;
return 1;
}
-static void core_complete_resync_work(struct dirty_log *log, region_t region,
- int success)
+static void core_set_region_sync(struct dm_dirty_log *log, region_t region,
+ int in_sync)
{
struct log_c *lc = (struct log_c *) log->context;
log_clear_bit(lc, lc->recovering_bits, region);
- if (success) {
+ if (in_sync) {
log_set_bit(lc, lc->sync_bits, region);
lc->sync_count++;
- }
+ } else if (log_test_bit(lc->sync_bits, region)) {
+ lc->sync_count--;
+ log_clear_bit(lc, lc->sync_bits, region);
+ }
}
-static region_t core_get_sync_count(struct dirty_log *log)
+static region_t core_get_sync_count(struct dm_dirty_log *log)
{
struct log_c *lc = (struct log_c *) log->context;
if (lc->sync != DEFAULTSYNC) \
DMEMIT("%ssync ", lc->sync == NOSYNC ? "no" : "")
-static int core_status(struct dirty_log *log, status_type_t status,
+static int core_status(struct dm_dirty_log *log, status_type_t status,
char *result, unsigned int maxlen)
{
int sz = 0;
switch(status) {
case STATUSTYPE_INFO:
+ DMEMIT("1 %s", log->type->name);
break;
case STATUSTYPE_TABLE:
return sz;
}
-static int disk_status(struct dirty_log *log, status_type_t status,
+static int disk_status(struct dm_dirty_log *log, status_type_t status,
char *result, unsigned int maxlen)
{
int sz = 0;
- char buffer[16];
struct log_c *lc = log->context;
switch(status) {
case STATUSTYPE_INFO:
+ DMEMIT("3 %s %s %c", log->type->name, lc->log_dev->name,
+ lc->log_dev_failed ? 'D' : 'A');
break;
case STATUSTYPE_TABLE:
- format_dev_t(buffer, lc->log_dev->bdev->bd_dev);
DMEMIT("%s %u %s %u ", log->type->name,
- lc->sync == DEFAULTSYNC ? 2 : 3, buffer,
+ lc->sync == DEFAULTSYNC ? 2 : 3, lc->log_dev->name,
lc->region_size);
DMEMIT_SYNC;
}
return sz;
}
-static struct dirty_log_type _core_type = {
+static struct dm_dirty_log_type _core_type = {
.name = "core",
.module = THIS_MODULE,
.ctr = core_ctr,
.dtr = core_dtr,
+ .resume = core_resume,
.get_region_size = core_get_region_size,
.is_clean = core_is_clean,
.in_sync = core_in_sync,
.mark_region = core_mark_region,
.clear_region = core_clear_region,
.get_resync_work = core_get_resync_work,
- .complete_resync_work = core_complete_resync_work,
+ .set_region_sync = core_set_region_sync,
.get_sync_count = core_get_sync_count,
.status = core_status,
};
-static struct dirty_log_type _disk_type = {
+static struct dm_dirty_log_type _disk_type = {
.name = "disk",
.module = THIS_MODULE,
.ctr = disk_ctr,
.dtr = disk_dtr,
- .suspend = disk_flush,
+ .postsuspend = disk_flush,
.resume = disk_resume,
.get_region_size = core_get_region_size,
.is_clean = core_is_clean,
.mark_region = core_mark_region,
.clear_region = core_clear_region,
.get_resync_work = core_get_resync_work,
- .complete_resync_work = core_complete_resync_work,
+ .set_region_sync = core_set_region_sync,
.get_sync_count = core_get_sync_count,
.status = disk_status,
};
-int __init dm_dirty_log_init(void)
+static int __init dm_dirty_log_init(void)
{
int r;
- r = dm_register_dirty_log_type(&_core_type);
+ r = dm_dirty_log_type_register(&_core_type);
if (r)
DMWARN("couldn't register core log");
- r = dm_register_dirty_log_type(&_disk_type);
+ r = dm_dirty_log_type_register(&_disk_type);
if (r) {
DMWARN("couldn't register disk type");
- dm_unregister_dirty_log_type(&_core_type);
+ dm_dirty_log_type_unregister(&_core_type);
}
return r;
}
-void dm_dirty_log_exit(void)
+static void __exit dm_dirty_log_exit(void)
{
- dm_unregister_dirty_log_type(&_disk_type);
- dm_unregister_dirty_log_type(&_core_type);
+ dm_dirty_log_type_unregister(&_disk_type);
+ dm_dirty_log_type_unregister(&_core_type);
}
-EXPORT_SYMBOL(dm_register_dirty_log_type);
-EXPORT_SYMBOL(dm_unregister_dirty_log_type);
-EXPORT_SYMBOL(dm_create_dirty_log);
-EXPORT_SYMBOL(dm_destroy_dirty_log);
+module_init(dm_dirty_log_init);
+module_exit(dm_dirty_log_exit);
+
+MODULE_DESCRIPTION(DM_NAME " dirty region log");
+MODULE_AUTHOR("Joe Thornber, Heinz Mauelshagen <dm-devel@redhat.com>");
+MODULE_LICENSE("GPL");