dm-log.c

来自「linux 内核源代码」· C语言 代码 · 共 740 行 · 第 1/2 页

/*
 * Copyright (C) 2003 Sistina Software
 *
 * This file is released under the LGPL.
 */

#include <linux/init.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/vmalloc.h>

#include "dm-log.h"
#include "dm-io.h"

#define DM_MSG_PREFIX "mirror log"

static LIST_HEAD(_log_types);
static DEFINE_SPINLOCK(_lock);

int dm_register_dirty_log_type(struct dirty_log_type *type)
{
	spin_lock(&_lock);
	type->use_count = 0;
	list_add(&type->list, &_log_types);
	spin_unlock(&_lock);

	return 0;
}

int dm_unregister_dirty_log_type(struct dirty_log_type *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);

	spin_unlock(&_lock);

	return 0;
}

static struct dirty_log_type *get_type(const char *type_name)
{
	struct dirty_log_type *type;

	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;
		}

	spin_unlock(&_lock);
	return NULL;
}

static void put_type(struct dirty_log_type *type)
{
	spin_lock(&_lock);
	if (!--type->use_count)
		module_put(type->module);
	spin_unlock(&_lock);
}

struct dirty_log *dm_create_dirty_log(const char *type_name, struct dm_target *ti,
				      unsigned int argc, char **argv)
{
	struct dirty_log_type *type;
	struct dirty_log *log;

	log = kmalloc(sizeof(*log), GFP_KERNEL);
	if (!log)
		return NULL;

	type = get_type(type_name);
	if (!type) {
		kfree(log);
		return NULL;
	}

	log->type = type;
	if (type->ctr(log, ti, argc, argv)) {
		kfree(log);
		put_type(type);
		return NULL;
	}

	return log;
}

void dm_destroy_dirty_log(struct dirty_log *log)
{
	log->type->dtr(log);
	put_type(log->type);
	kfree(log);
}

/*-----------------------------------------------------------------
 * Persistent and core logs share a lot of their implementation.
 * FIXME: need a reload method to be called from a resume
 *---------------------------------------------------------------*/
/*
 * Magic for persistent mirrors: "MiRr"
 */
#define MIRROR_MAGIC 0x4D695272

/*
 * The on-disk version of the metadata.
 */
#define MIRROR_DISK_VERSION 2
#define LOG_OFFSET 2

struct log_header {
	uint32_t magic;

	/*
	 * Simple, incrementing version. no backward
	 * compatibility.
	 */
	uint32_t version;
	sector_t nr_regions;
};

struct log_c {
	struct dm_target *ti;
	int touched;
	uint32_t region_size;
	unsigned int region_count;
	region_t sync_count;

	unsigned bitset_uint32_count;
	uint32_t *clean_bits;
	uint32_t *sync_bits;
	uint32_t *recovering_bits;	/* FIXME: this seems excessive */

	int sync_search;

	/* Resync flag */
	enum sync {
		DEFAULTSYNC,	/* Synchronize if necessary */
		NOSYNC,		/* Devices known to be already in sync */
		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 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)
{
	return ext2_test_bit(bit, (unsigned long *) bs) ? 1 : 0;
}

static inline void log_set_bit(struct log_c *l,
			       uint32_t *bs, unsigned bit)
{
	ext2_set_bit(bit, (unsigned long *) bs);
	l->touched = 1;
}

static inline void log_clear_bit(struct log_c *l,
				 uint32_t *bs, unsigned bit)
{
	ext2_clear_bit(bit, (unsigned long *) bs);
	l->touched = 1;
}

/*----------------------------------------------------------------
 * Header IO
 *--------------------------------------------------------------*/
static void header_to_disk(struct log_header *core, struct log_header *disk)
{
	disk->magic = cpu_to_le32(core->magic);
	disk->version = cpu_to_le32(core->version);
	disk->nr_regions = cpu_to_le64(core->nr_regions);
}

static void header_from_disk(struct log_header *core, struct log_header *disk)
{
	core->magic = le32_to_cpu(disk->magic);
	core->version = le32_to_cpu(disk->version);
	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;
	lc->io_req.mem.ptr.vma = lc->disk_header;
	lc->io_req.notify.fn = NULL;

	return dm_io(&lc->io_req, 1, &lc->header_location, NULL);
}

static int read_header(struct log_c *log)
{
	int r;

	r = rw_header(log, READ);
	if (r)
		return r;

	header_from_disk(&log->header, log->disk_header);

	/* New log required? */
	if (log->sync != DEFAULTSYNC || log->header.magic != MIRROR_MAGIC) {
		log->header.magic = MIRROR_MAGIC;
		log->header.version = MIRROR_DISK_VERSION;
		log->header.nr_regions = 0;
	}

#ifdef __LITTLE_ENDIAN
	if (log->header.version == 1)
		log->header.version = 2;
#endif

	if (log->header.version != MIRROR_DISK_VERSION) {
		DMWARN("incompatible disk log version");
		return -EINVAL;
	}

	return 0;
}

static inline int write_header(struct log_c *log)
{
	header_to_disk(&log->header, log->disk_header);
	return rw_header(log, WRITE);
}

/*----------------------------------------------------------------
 * core log constructor/destructor
 *
 * 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,
			      unsigned int argc, char **argv,
			      struct dm_dev *dev)
{
	enum sync sync = DEFAULTSYNC;

	struct log_c *lc;
	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");
		return -EINVAL;
	}

	if (argc > 1) {
		if (!strcmp(argv[1], "sync"))
			sync = FORCESYNC;
		else if (!strcmp(argv[1], "nosync"))
			sync = NOSYNC;
		else {
			DMWARN("unrecognised sync argument to mirror log: %s",
			       argv[1]);
			return -EINVAL;
		}
	}

	if (sscanf(argv[0], "%u", &region_size) != 1) {
		DMWARN("invalid region size string");
		return -EINVAL;
	}

	region_count = dm_sector_div_up(ti->len, region_size);

	lc = kmalloc(sizeof(*lc), GFP_KERNEL);
	if (!lc) {
		DMWARN("couldn't allocate core log");
		return -ENOMEM;
	}

	lc->ti = ti;
	lc->touched = 0;
	lc->region_size = region_size;
	lc->region_count = region_count;
	lc->sync = sync;

	/*
	 * Work out how many "unsigned long"s we need to hold the bitset.
	 */
	bitset_size = dm_round_up(region_count,
				  sizeof(*lc->clean_bits) << BYTE_SHIFT);
	bitset_size >>= BYTE_SHIFT;

	lc->bitset_uint32_count = bitset_size / sizeof(*lc->clean_bits);

	/*
	 * Disk log?
	 */
	if (!dev) {
		lc->clean_bits = vmalloc(bitset_size);
		if (!lc->clean_bits) {
			DMWARN("couldn't allocate clean bitset");
			kfree(lc);
			return -ENOMEM;
		}
		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);
		lc->header_location.count = buf_size >> SECTOR_SHIFT;
		lc->io_req.mem.type = DM_IO_VMA;
		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");
			kfree(lc);
			return -ENOMEM;
		}

		lc->clean_bits = (void *)lc->disk_header +
				 (LOG_OFFSET << SECTOR_SHIFT);
	}

	memset(lc->clean_bits, -1, bitset_size);

	lc->sync_bits = vmalloc(bitset_size);
	if (!lc->sync_bits) {
		DMWARN("couldn't allocate sync bitset");
		if (!dev)
			vfree(lc->clean_bits);
		vfree(lc->disk_header);
		kfree(lc);
		return -ENOMEM;
	}
	memset(lc->sync_bits, (sync == NOSYNC) ? -1 : 0, bitset_size);
	lc->sync_count = (sync == NOSYNC) ? region_count : 0;