raid5-cache: check stripe finish out of order

[mirror_ubuntu-bionic-kernel.git] / drivers / md / raid5-cache.c

/*
 * Copyright (C) 2015 Shaohua Li <shli@fb.com>
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
 * more details.
 *
 */
#include <linux/kernel.h>
#include <linux/wait.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
#include <linux/raid/md_p.h>
#include <linux/crc32c.h>
#include <linux/random.h>
#include "md.h"
#include "raid5.h"

/*
 * metadata/data stored in disk with 4k size unit (a block) regardless
 * underneath hardware sector size. only works with PAGE_SIZE == 4096
 */
#define BLOCK_SECTORS (8)

/*
 * reclaim runs every 1/4 disk size or 10G reclaimable space. This can prevent
 * recovery scans a very long log
 */
#define RECLAIM_MAX_FREE_SPACE (10 * 1024 * 1024 * 2) /* sector */
#define RECLAIM_MAX_FREE_SPACE_SHIFT (2)

struct r5l_log {
	struct md_rdev *rdev;

	u32 uuid_checksum;

	sector_t device_size;		/* log device size, round to
					 * BLOCK_SECTORS */
	sector_t max_free_space;	/* reclaim run if free space is at
					 * this size */

	sector_t last_checkpoint;	/* log tail. where recovery scan
					 * starts from */
	u64 last_cp_seq;		/* log tail sequence */

	sector_t log_start;		/* log head. where new data appends */
	u64 seq;			/* log head sequence */

	struct mutex io_mutex;
	struct r5l_io_unit *current_io;	/* current io_unit accepting new data */

	spinlock_t io_list_lock;
	struct list_head running_ios;	/* io_units which are still running,
					 * and have not yet been completely
					 * written to the log */
	struct list_head io_end_ios;	/* io_units which have been completely
					 * written to the log but not yet written
					 * to the RAID */
	struct list_head flushing_ios;	/* io_units which are waiting for log
					 * cache flush */
	struct list_head flushed_ios;	/* io_units which settle down in log disk */
	struct bio flush_bio;
	struct list_head stripe_end_ios;/* io_units which have been completely
					 * written to the RAID but have not yet
					 * been considered for updating super */

	struct kmem_cache *io_kc;

	struct md_thread *reclaim_thread;
	unsigned long reclaim_target;	/* number of space that need to be
					 * reclaimed.  if it's 0, reclaim spaces
					 * used by io_units which are in
					 * IO_UNIT_STRIPE_END state (eg, reclaim
					 * dones't wait for specific io_unit
					 * switching to IO_UNIT_STRIPE_END
					 * state) */
	wait_queue_head_t iounit_wait;

	struct list_head no_space_stripes; /* pending stripes, log has no space */
	spinlock_t no_space_stripes_lock;
};

/*
 * an IO range starts from a meta data block and end at the next meta data
 * block. The io unit's the meta data block tracks data/parity followed it. io
 * unit is written to log disk with normal write, as we always flush log disk
 * first and then start move data to raid disks, there is no requirement to
 * write io unit with FLUSH/FUA
 */
struct r5l_io_unit {
	struct r5l_log *log;

	struct page *meta_page;	/* store meta block */
	int meta_offset;	/* current offset in meta_page */

	struct bio_list bios;
	atomic_t pending_io;	/* pending bios not written to log yet */
	struct bio *current_bio;/* current_bio accepting new data */

	atomic_t pending_stripe;/* how many stripes not flushed to raid */
	u64 seq;		/* seq number of the metablock */
	sector_t log_start;	/* where the io_unit starts */
	sector_t log_end;	/* where the io_unit ends */
	struct list_head log_sibling; /* log->running_ios */
	struct list_head stripe_list; /* stripes added to the io_unit */

	int state;
};

/* r5l_io_unit state */
enum r5l_io_unit_state {
	IO_UNIT_RUNNING = 0,	/* accepting new IO */
	IO_UNIT_IO_START = 1,	/* io_unit bio start writing to log,
				 * don't accepting new bio */
	IO_UNIT_IO_END = 2,	/* io_unit bio finish writing to log */
	IO_UNIT_STRIPE_END = 3,	/* stripes data finished writing to raid */
};

static sector_t r5l_ring_add(struct r5l_log *log, sector_t start, sector_t inc)
{
	start += inc;
	if (start >= log->device_size)
		start = start - log->device_size;
	return start;
}

static sector_t r5l_ring_distance(struct r5l_log *log, sector_t start,
				  sector_t end)
{
	if (end >= start)
		return end - start;
	else
		return end + log->device_size - start;
}

static bool r5l_has_free_space(struct r5l_log *log, sector_t size)
{
	sector_t used_size;

	used_size = r5l_ring_distance(log, log->last_checkpoint,
					log->log_start);

	return log->device_size > used_size + size;
}

static struct r5l_io_unit *r5l_alloc_io_unit(struct r5l_log *log)
{
	struct r5l_io_unit *io;
	/* We can't handle memory allocate failure so far */
	gfp_t gfp = GFP_NOIO | __GFP_NOFAIL;

	io = kmem_cache_zalloc(log->io_kc, gfp);
	io->log = log;
	io->meta_page = alloc_page(gfp | __GFP_ZERO);

	bio_list_init(&io->bios);
	INIT_LIST_HEAD(&io->log_sibling);
	INIT_LIST_HEAD(&io->stripe_list);
	io->state = IO_UNIT_RUNNING;
	return io;
}

static void r5l_free_io_unit(struct r5l_log *log, struct r5l_io_unit *io)
{
	__free_page(io->meta_page);
	kmem_cache_free(log->io_kc, io);
}

static void r5l_move_io_unit_list(struct list_head *from, struct list_head *to,
				  enum r5l_io_unit_state state)
{
	struct r5l_io_unit *io;

	while (!list_empty(from)) {
		io = list_first_entry(from, struct r5l_io_unit, log_sibling);
		/* don't change list order */
		if (io->state >= state)
			list_move_tail(&io->log_sibling, to);
		else
			break;
	}
}

/*
 * We don't want too many io_units reside in stripe_end_ios list, which will
 * waste a lot of memory. So we try to remove some. But we must keep at least 2
 * io_units. The superblock must point to a valid meta, if it's the last meta,
 * recovery can scan less
 */
static void r5l_compress_stripe_end_list(struct r5l_log *log)
{
	struct r5l_io_unit *first, *last, *io;

	first = list_first_entry(&log->stripe_end_ios,
				 struct r5l_io_unit, log_sibling);
	last = list_last_entry(&log->stripe_end_ios,
			       struct r5l_io_unit, log_sibling);
	if (first == last)
		return;
	list_del(&first->log_sibling);
	list_del(&last->log_sibling);
	while (!list_empty(&log->stripe_end_ios)) {
		io = list_first_entry(&log->stripe_end_ios,
				      struct r5l_io_unit, log_sibling);
		list_del(&io->log_sibling);
		first->log_end = io->log_end;
		r5l_free_io_unit(log, io);
	}
	list_add_tail(&first->log_sibling, &log->stripe_end_ios);
	list_add_tail(&last->log_sibling, &log->stripe_end_ios);
}

static void __r5l_set_io_unit_state(struct r5l_io_unit *io,
				    enum r5l_io_unit_state state)
{
	if (WARN_ON(io->state >= state))
		return;
	io->state = state;
}

/* XXX: totally ignores I/O errors */
static void r5l_log_endio(struct bio *bio)
{
	struct r5l_io_unit *io = bio->bi_private;
	struct r5l_log *log = io->log;
	unsigned long flags;

	bio_put(bio);

	if (!atomic_dec_and_test(&io->pending_io))
		return;

	spin_lock_irqsave(&log->io_list_lock, flags);
	__r5l_set_io_unit_state(io, IO_UNIT_IO_END);
	r5l_move_io_unit_list(&log->running_ios, &log->io_end_ios,
			IO_UNIT_IO_END);
	spin_unlock_irqrestore(&log->io_list_lock, flags);

	md_wakeup_thread(log->rdev->mddev->thread);
}

static void r5l_submit_current_io(struct r5l_log *log)
{
	struct r5l_io_unit *io = log->current_io;
	struct r5l_meta_block *block;
	struct bio *bio;
	unsigned long flags;
	u32 crc;

	if (!io)
		return;

	block = page_address(io->meta_page);
	block->meta_size = cpu_to_le32(io->meta_offset);
	crc = crc32c_le(log->uuid_checksum, block, PAGE_SIZE);
	block->checksum = cpu_to_le32(crc);

	log->current_io = NULL;
	spin_lock_irqsave(&log->io_list_lock, flags);
	__r5l_set_io_unit_state(io, IO_UNIT_IO_START);
	spin_unlock_irqrestore(&log->io_list_lock, flags);

	while ((bio = bio_list_pop(&io->bios))) {
		/* all IO must start from rdev->data_offset */
		bio->bi_iter.bi_sector += log->rdev->data_offset;
		submit_bio(WRITE, bio);
	}
}

static struct r5l_io_unit *r5l_new_meta(struct r5l_log *log)
{
	struct r5l_io_unit *io;
	struct r5l_meta_block *block;
	struct bio *bio;

	io = r5l_alloc_io_unit(log);

	block = page_address(io->meta_page);
	block->magic = cpu_to_le32(R5LOG_MAGIC);
	block->version = R5LOG_VERSION;
	block->seq = cpu_to_le64(log->seq);
	block->position = cpu_to_le64(log->log_start);

	io->log_start = log->log_start;
	io->meta_offset = sizeof(struct r5l_meta_block);
	io->seq = log->seq;

	bio = bio_kmalloc(GFP_NOIO | __GFP_NOFAIL, BIO_MAX_PAGES);
	io->current_bio = bio;
	bio->bi_rw = WRITE;
	bio->bi_bdev = log->rdev->bdev;
	bio->bi_iter.bi_sector = log->log_start;
	bio_add_page(bio, io->meta_page, PAGE_SIZE, 0);
	bio->bi_end_io = r5l_log_endio;
	bio->bi_private = io;

	bio_list_add(&io->bios, bio);
	atomic_inc(&io->pending_io);

	log->seq++;
	log->log_start = r5l_ring_add(log, log->log_start, BLOCK_SECTORS);
	io->log_end = log->log_start;
	/* current bio hit disk end */
	if (log->log_start == 0)
		io->current_bio = NULL;

	spin_lock_irq(&log->io_list_lock);
	list_add_tail(&io->log_sibling, &log->running_ios);
	spin_unlock_irq(&log->io_list_lock);

	return io;
}

static int r5l_get_meta(struct r5l_log *log, unsigned int payload_size)
{
	struct r5l_io_unit *io;

	io = log->current_io;
	if (io && io->meta_offset + payload_size > PAGE_SIZE)
		r5l_submit_current_io(log);
	io = log->current_io;
	if (io)
		return 0;

	log->current_io = r5l_new_meta(log);
	return 0;
}

static void r5l_append_payload_meta(struct r5l_log *log, u16 type,
				    sector_t location,
				    u32 checksum1, u32 checksum2,
				    bool checksum2_valid)
{
	struct r5l_io_unit *io = log->current_io;
	struct r5l_payload_data_parity *payload;

	payload = page_address(io->meta_page) + io->meta_offset;
	payload->header.type = cpu_to_le16(type);
	payload->header.flags = cpu_to_le16(0);
	payload->size = cpu_to_le32((1 + !!checksum2_valid) <<
				    (PAGE_SHIFT - 9));
	payload->location = cpu_to_le64(location);
	payload->checksum[0] = cpu_to_le32(checksum1);
	if (checksum2_valid)
		payload->checksum[1] = cpu_to_le32(checksum2);

	io->meta_offset += sizeof(struct r5l_payload_data_parity) +
		sizeof(__le32) * (1 + !!checksum2_valid);
}

static void r5l_append_payload_page(struct r5l_log *log, struct page *page)
{
	struct r5l_io_unit *io = log->current_io;

alloc_bio:
	if (!io->current_bio) {
		struct bio *bio;

		bio = bio_kmalloc(GFP_NOIO | __GFP_NOFAIL, BIO_MAX_PAGES);
		bio->bi_rw = WRITE;
		bio->bi_bdev = log->rdev->bdev;
		bio->bi_iter.bi_sector = log->log_start;
		bio->bi_end_io = r5l_log_endio;
		bio->bi_private = io;
		bio_list_add(&io->bios, bio);
		atomic_inc(&io->pending_io);
		io->current_bio = bio;
	}
	if (!bio_add_page(io->current_bio, page, PAGE_SIZE, 0)) {
		io->current_bio = NULL;
		goto alloc_bio;
	}
	log->log_start = r5l_ring_add(log, log->log_start,
				      BLOCK_SECTORS);
	/* current bio hit disk end */
	if (log->log_start == 0)
		io->current_bio = NULL;

	io->log_end = log->log_start;
}

static void r5l_log_stripe(struct r5l_log *log, struct stripe_head *sh,
			   int data_pages, int parity_pages)
{
	int i;
	int meta_size;
	struct r5l_io_unit *io;

	meta_size =
		((sizeof(struct r5l_payload_data_parity) + sizeof(__le32))
		 * data_pages) +
		sizeof(struct r5l_payload_data_parity) +
		sizeof(__le32) * parity_pages;

	r5l_get_meta(log, meta_size);
	io = log->current_io;

	for (i = 0; i < sh->disks; i++) {
		if (!test_bit(R5_Wantwrite, &sh->dev[i].flags))
			continue;
		if (i == sh->pd_idx || i == sh->qd_idx)
			continue;
		r5l_append_payload_meta(log, R5LOG_PAYLOAD_DATA,
					raid5_compute_blocknr(sh, i, 0),
					sh->dev[i].log_checksum, 0, false);
		r5l_append_payload_page(log, sh->dev[i].page);
	}

	if (sh->qd_idx >= 0) {
		r5l_append_payload_meta(log, R5LOG_PAYLOAD_PARITY,
					sh->sector, sh->dev[sh->pd_idx].log_checksum,
					sh->dev[sh->qd_idx].log_checksum, true);
		r5l_append_payload_page(log, sh->dev[sh->pd_idx].page);
		r5l_append_payload_page(log, sh->dev[sh->qd_idx].page);
	} else {
		r5l_append_payload_meta(log, R5LOG_PAYLOAD_PARITY,
					sh->sector, sh->dev[sh->pd_idx].log_checksum,
					0, false);
		r5l_append_payload_page(log, sh->dev[sh->pd_idx].page);
	}

	list_add_tail(&sh->log_list, &io->stripe_list);
	atomic_inc(&io->pending_stripe);
	sh->log_io = io;
}

static void r5l_wake_reclaim(struct r5l_log *log, sector_t space);
/*
 * running in raid5d, where reclaim could wait for raid5d too (when it flushes
 * data from log to raid disks), so we shouldn't wait for reclaim here
 */
int r5l_write_stripe(struct r5l_log *log, struct stripe_head *sh)
{
	int write_disks = 0;
	int data_pages, parity_pages;
	int meta_size;
	int reserve;
	int i;

	if (!log)
		return -EAGAIN;
	/* Don't support stripe batch */
	if (sh->log_io || !test_bit(R5_Wantwrite, &sh->dev[sh->pd_idx].flags) ||
	    test_bit(STRIPE_SYNCING, &sh->state)) {
		/* the stripe is written to log, we start writing it to raid */
		clear_bit(STRIPE_LOG_TRAPPED, &sh->state);
		return -EAGAIN;
	}

	for (i = 0; i < sh->disks; i++) {
		void *addr;

		if (!test_bit(R5_Wantwrite, &sh->dev[i].flags))
			continue;
		write_disks++;
		/* checksum is already calculated in last run */
		if (test_bit(STRIPE_LOG_TRAPPED, &sh->state))
			continue;
		addr = kmap_atomic(sh->dev[i].page);
		sh->dev[i].log_checksum = crc32c_le(log->uuid_checksum,
						    addr, PAGE_SIZE);
		kunmap_atomic(addr);
	}
	parity_pages = 1 + !!(sh->qd_idx >= 0);
	data_pages = write_disks - parity_pages;

	meta_size =
		((sizeof(struct r5l_payload_data_parity) + sizeof(__le32))
		 * data_pages) +
		sizeof(struct r5l_payload_data_parity) +
		sizeof(__le32) * parity_pages;
	/* Doesn't work with very big raid array */
	if (meta_size + sizeof(struct r5l_meta_block) > PAGE_SIZE)
		return -EINVAL;

	set_bit(STRIPE_LOG_TRAPPED, &sh->state);
	atomic_inc(&sh->count);

	mutex_lock(&log->io_mutex);
	/* meta + data */
	reserve = (1 + write_disks) << (PAGE_SHIFT - 9);
	if (r5l_has_free_space(log, reserve))
		r5l_log_stripe(log, sh, data_pages, parity_pages);
	else {
		spin_lock(&log->no_space_stripes_lock);
		list_add_tail(&sh->log_list, &log->no_space_stripes);
		spin_unlock(&log->no_space_stripes_lock);

		r5l_wake_reclaim(log, reserve);
	}
	mutex_unlock(&log->io_mutex);

	return 0;
}

void r5l_write_stripe_run(struct r5l_log *log)
{
	if (!log)
		return;
	mutex_lock(&log->io_mutex);
	r5l_submit_current_io(log);
	mutex_unlock(&log->io_mutex);
}

int r5l_handle_flush_request(struct r5l_log *log, struct bio *bio)
{
	if (!log)
		return -ENODEV;
	/*
	 * we flush log disk cache first, then write stripe data to raid disks.
	 * So if bio is finished, the log disk cache is flushed already. The
	 * recovery guarantees we can recovery the bio from log disk, so we
	 * don't need to flush again
	 */
	if (bio->bi_iter.bi_size == 0) {
		bio_endio(bio);
		return 0;
	}
	bio->bi_rw &= ~REQ_FLUSH;
	return -EAGAIN;
}

/* This will run after log space is reclaimed */
static void r5l_run_no_space_stripes(struct r5l_log *log)
{
	struct stripe_head *sh;

	spin_lock(&log->no_space_stripes_lock);
	while (!list_empty(&log->no_space_stripes)) {
		sh = list_first_entry(&log->no_space_stripes,
				      struct stripe_head, log_list);
		list_del_init(&sh->log_list);
		set_bit(STRIPE_HANDLE, &sh->state);
		raid5_release_stripe(sh);
	}
	spin_unlock(&log->no_space_stripes_lock);
}

static void __r5l_stripe_write_finished(struct r5l_io_unit *io)
{
	struct r5l_log *log = io->log;
	struct r5l_io_unit *last;
	sector_t reclaimable_space;
	unsigned long flags;

	spin_lock_irqsave(&log->io_list_lock, flags);
	__r5l_set_io_unit_state(io, IO_UNIT_STRIPE_END);
	/* might move 0 entry */
	r5l_move_io_unit_list(&log->flushed_ios, &log->stripe_end_ios,
			      IO_UNIT_STRIPE_END);
	if (list_empty(&log->stripe_end_ios)) {
		spin_unlock_irqrestore(&log->io_list_lock, flags);
		return;
	}

	last = list_last_entry(&log->stripe_end_ios,
			       struct r5l_io_unit, log_sibling);
	reclaimable_space = r5l_ring_distance(log, log->last_checkpoint,
					      last->log_end);
	if (reclaimable_space >= log->max_free_space)
		r5l_wake_reclaim(log, 0);

	r5l_compress_stripe_end_list(log);
	spin_unlock_irqrestore(&log->io_list_lock, flags);
	wake_up(&log->iounit_wait);
}

void r5l_stripe_write_finished(struct stripe_head *sh)
{
	struct r5l_io_unit *io;

	io = sh->log_io;
	sh->log_io = NULL;

	if (io && atomic_dec_and_test(&io->pending_stripe))
		__r5l_stripe_write_finished(io);
}

static void r5l_log_flush_endio(struct bio *bio)
{
	struct r5l_log *log = container_of(bio, struct r5l_log,
		flush_bio);
	unsigned long flags;
	struct r5l_io_unit *io;
	struct stripe_head *sh;

	spin_lock_irqsave(&log->io_list_lock, flags);
	list_for_each_entry(io, &log->flushing_ios, log_sibling) {
		while (!list_empty(&io->stripe_list)) {
			sh = list_first_entry(&io->stripe_list,
				struct stripe_head, log_list);
			list_del_init(&sh->log_list);
			set_bit(STRIPE_HANDLE, &sh->state);
			raid5_release_stripe(sh);
		}
	}
	list_splice_tail_init(&log->flushing_ios, &log->flushed_ios);
	spin_unlock_irqrestore(&log->io_list_lock, flags);
}

/*
 * Starting dispatch IO to raid.
 * io_unit(meta) consists of a log. There is one situation we want to avoid. A
 * broken meta in the middle of a log causes recovery can't find meta at the
 * head of log. If operations require meta at the head persistent in log, we
 * must make sure meta before it persistent in log too. A case is:
 *
 * stripe data/parity is in log, we start write stripe to raid disks. stripe
 * data/parity must be persistent in log before we do the write to raid disks.
 *
 * The solution is we restrictly maintain io_unit list order. In this case, we
 * only write stripes of an io_unit to raid disks till the io_unit is the first
 * one whose data/parity is in log.
 */
void r5l_flush_stripe_to_raid(struct r5l_log *log)
{
	bool do_flush;
	if (!log)
		return;

	spin_lock_irq(&log->io_list_lock);
	/* flush bio is running */
	if (!list_empty(&log->flushing_ios)) {
		spin_unlock_irq(&log->io_list_lock);
		return;
	}
	list_splice_tail_init(&log->io_end_ios, &log->flushing_ios);
	do_flush = !list_empty(&log->flushing_ios);
	spin_unlock_irq(&log->io_list_lock);

	if (!do_flush)
		return;
	bio_reset(&log->flush_bio);
	log->flush_bio.bi_bdev = log->rdev->bdev;
	log->flush_bio.bi_end_io = r5l_log_flush_endio;
	submit_bio(WRITE_FLUSH, &log->flush_bio);
}

static void r5l_kick_io_unit(struct r5l_log *log)
{
	md_wakeup_thread(log->rdev->mddev->thread);
	wait_event_lock_irq(log->iounit_wait, !list_empty(&log->stripe_end_ios),
			    log->io_list_lock);
}

static void r5l_write_super(struct r5l_log *log, sector_t cp);
static void r5l_do_reclaim(struct r5l_log *log)
{
	struct r5l_io_unit *io, *last;
	LIST_HEAD(list);
	sector_t free = 0;
	sector_t reclaim_target = xchg(&log->reclaim_target, 0);

	spin_lock_irq(&log->io_list_lock);
	/*
	 * move proper io_unit to reclaim list. We should not change the order.
	 * reclaimable/unreclaimable io_unit can be mixed in the list, we
	 * shouldn't reuse space of an unreclaimable io_unit
	 */
	while (1) {
		struct list_head *target_list = NULL;

		while (!list_empty(&log->stripe_end_ios)) {
			io = list_first_entry(&log->stripe_end_ios,
					      struct r5l_io_unit, log_sibling);
			list_move_tail(&io->log_sibling, &list);
			free += r5l_ring_distance(log, io->log_start,
						  io->log_end);
		}

		if (free >= reclaim_target ||
		    (list_empty(&log->running_ios) &&
		     list_empty(&log->io_end_ios) &&
		     list_empty(&log->flushing_ios) &&
		     list_empty(&log->flushed_ios)))
			break;

		/* Below waiting mostly happens when we shutdown the raid */
		if (!list_empty(&log->flushed_ios))
			target_list = &log->flushed_ios;
		else if (!list_empty(&log->flushing_ios))
			target_list = &log->flushing_ios;
		else if (!list_empty(&log->io_end_ios))
			target_list = &log->io_end_ios;
		else if (!list_empty(&log->running_ios))
			target_list = &log->running_ios;

		r5l_kick_io_unit(log);
	}
	spin_unlock_irq(&log->io_list_lock);

	if (list_empty(&list))
		return;

	/* super always point to last valid meta */
	last = list_last_entry(&list, struct r5l_io_unit, log_sibling);
	/*
	 * write_super will flush cache of each raid disk. We must write super
	 * here, because the log area might be reused soon and we don't want to
	 * confuse recovery
	 */
	r5l_write_super(log, last->log_start);

	mutex_lock(&log->io_mutex);
	log->last_checkpoint = last->log_start;
	log->last_cp_seq = last->seq;
	mutex_unlock(&log->io_mutex);
	r5l_run_no_space_stripes(log);

	while (!list_empty(&list)) {
		io = list_first_entry(&list, struct r5l_io_unit, log_sibling);
		list_del(&io->log_sibling);
		r5l_free_io_unit(log, io);
	}
}

static void r5l_reclaim_thread(struct md_thread *thread)
{
	struct mddev *mddev = thread->mddev;
	struct r5conf *conf = mddev->private;
	struct r5l_log *log = conf->log;

	if (!log)
		return;
	r5l_do_reclaim(log);
}

static void r5l_wake_reclaim(struct r5l_log *log, sector_t space)
{
	unsigned long target;
	unsigned long new = (unsigned long)space; /* overflow in theory */

	do {
		target = log->reclaim_target;
		if (new < target)
			return;
	} while (cmpxchg(&log->reclaim_target, target, new) != target);
	md_wakeup_thread(log->reclaim_thread);
}

struct r5l_recovery_ctx {
	struct page *meta_page;		/* current meta */
	sector_t meta_total_blocks;	/* total size of current meta and data */
	sector_t pos;			/* recovery position */
	u64 seq;			/* recovery position seq */
};

static int r5l_read_meta_block(struct r5l_log *log,
			       struct r5l_recovery_ctx *ctx)
{
	struct page *page = ctx->meta_page;
	struct r5l_meta_block *mb;
	u32 crc, stored_crc;

	if (!sync_page_io(log->rdev, ctx->pos, PAGE_SIZE, page, READ, false))
		return -EIO;

	mb = page_address(page);
	stored_crc = le32_to_cpu(mb->checksum);
	mb->checksum = 0;

	if (le32_to_cpu(mb->magic) != R5LOG_MAGIC ||
	    le64_to_cpu(mb->seq) != ctx->seq ||
	    mb->version != R5LOG_VERSION ||
	    le64_to_cpu(mb->position) != ctx->pos)
		return -EINVAL;

	crc = crc32c_le(log->uuid_checksum, mb, PAGE_SIZE);
	if (stored_crc != crc)
		return -EINVAL;

	if (le32_to_cpu(mb->meta_size) > PAGE_SIZE)
		return -EINVAL;

	ctx->meta_total_blocks = BLOCK_SECTORS;

	return 0;
}

static int r5l_recovery_flush_one_stripe(struct r5l_log *log,
					 struct r5l_recovery_ctx *ctx,
					 sector_t stripe_sect,
					 int *offset, sector_t *log_offset)
{
	struct r5conf *conf = log->rdev->mddev->private;
	struct stripe_head *sh;
	struct r5l_payload_data_parity *payload;
	int disk_index;

	sh = raid5_get_active_stripe(conf, stripe_sect, 0, 0, 0);
	while (1) {
		payload = page_address(ctx->meta_page) + *offset;

		if (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_DATA) {
			raid5_compute_sector(conf,
					     le64_to_cpu(payload->location), 0,
					     &disk_index, sh);

			sync_page_io(log->rdev, *log_offset, PAGE_SIZE,
				     sh->dev[disk_index].page, READ, false);
			sh->dev[disk_index].log_checksum =
				le32_to_cpu(payload->checksum[0]);
			set_bit(R5_Wantwrite, &sh->dev[disk_index].flags);
			ctx->meta_total_blocks += BLOCK_SECTORS;
		} else {
			disk_index = sh->pd_idx;
			sync_page_io(log->rdev, *log_offset, PAGE_SIZE,
				     sh->dev[disk_index].page, READ, false);
			sh->dev[disk_index].log_checksum =
				le32_to_cpu(payload->checksum[0]);
			set_bit(R5_Wantwrite, &sh->dev[disk_index].flags);

			if (sh->qd_idx >= 0) {
				disk_index = sh->qd_idx;
				sync_page_io(log->rdev,
					     r5l_ring_add(log, *log_offset, BLOCK_SECTORS),
					     PAGE_SIZE, sh->dev[disk_index].page,
					     READ, false);
				sh->dev[disk_index].log_checksum =
					le32_to_cpu(payload->checksum[1]);
				set_bit(R5_Wantwrite,
					&sh->dev[disk_index].flags);
			}
			ctx->meta_total_blocks += BLOCK_SECTORS * conf->max_degraded;
		}

		*log_offset = r5l_ring_add(log, *log_offset,
					   le32_to_cpu(payload->size));
		*offset += sizeof(struct r5l_payload_data_parity) +
			sizeof(__le32) *
			(le32_to_cpu(payload->size) >> (PAGE_SHIFT - 9));
		if (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_PARITY)
			break;
	}

	for (disk_index = 0; disk_index < sh->disks; disk_index++) {
		void *addr;
		u32 checksum;

		if (!test_bit(R5_Wantwrite, &sh->dev[disk_index].flags))
			continue;
		addr = kmap_atomic(sh->dev[disk_index].page);
		checksum = crc32c_le(log->uuid_checksum, addr, PAGE_SIZE);
		kunmap_atomic(addr);
		if (checksum != sh->dev[disk_index].log_checksum)
			goto error;
	}

	for (disk_index = 0; disk_index < sh->disks; disk_index++) {
		struct md_rdev *rdev, *rrdev;

		if (!test_and_clear_bit(R5_Wantwrite,
					&sh->dev[disk_index].flags))
			continue;

		/* in case device is broken */
		rdev = rcu_dereference(conf->disks[disk_index].rdev);
		if (rdev)
			sync_page_io(rdev, stripe_sect, PAGE_SIZE,
				     sh->dev[disk_index].page, WRITE, false);
		rrdev = rcu_dereference(conf->disks[disk_index].replacement);
		if (rrdev)
			sync_page_io(rrdev, stripe_sect, PAGE_SIZE,
				     sh->dev[disk_index].page, WRITE, false);
	}
	raid5_release_stripe(sh);
	return 0;

error:
	for (disk_index = 0; disk_index < sh->disks; disk_index++)
		sh->dev[disk_index].flags = 0;
	raid5_release_stripe(sh);
	return -EINVAL;
}

static int r5l_recovery_flush_one_meta(struct r5l_log *log,
				       struct r5l_recovery_ctx *ctx)
{
	struct r5conf *conf = log->rdev->mddev->private;
	struct r5l_payload_data_parity *payload;
	struct r5l_meta_block *mb;
	int offset;
	sector_t log_offset;
	sector_t stripe_sector;

	mb = page_address(ctx->meta_page);
	offset = sizeof(struct r5l_meta_block);
	log_offset = r5l_ring_add(log, ctx->pos, BLOCK_SECTORS);

	while (offset < le32_to_cpu(mb->meta_size)) {
		int dd;

		payload = (void *)mb + offset;
		stripe_sector = raid5_compute_sector(conf,
						     le64_to_cpu(payload->location), 0, &dd, NULL);
		if (r5l_recovery_flush_one_stripe(log, ctx, stripe_sector,
						  &offset, &log_offset))
			return -EINVAL;
	}
	return 0;
}

/* copy data/parity from log to raid disks */
static void r5l_recovery_flush_log(struct r5l_log *log,
				   struct r5l_recovery_ctx *ctx)
{
	while (1) {
		if (r5l_read_meta_block(log, ctx))
			return;
		if (r5l_recovery_flush_one_meta(log, ctx))
			return;
		ctx->seq++;
		ctx->pos = r5l_ring_add(log, ctx->pos, ctx->meta_total_blocks);
	}
}

static int r5l_log_write_empty_meta_block(struct r5l_log *log, sector_t pos,
					  u64 seq)
{
	struct page *page;
	struct r5l_meta_block *mb;
	u32 crc;

	page = alloc_page(GFP_KERNEL | __GFP_ZERO);
	if (!page)
		return -ENOMEM;
	mb = page_address(page);
	mb->magic = cpu_to_le32(R5LOG_MAGIC);
	mb->version = R5LOG_VERSION;
	mb->meta_size = cpu_to_le32(sizeof(struct r5l_meta_block));
	mb->seq = cpu_to_le64(seq);
	mb->position = cpu_to_le64(pos);
	crc = crc32c_le(log->uuid_checksum, mb, PAGE_SIZE);
	mb->checksum = cpu_to_le32(crc);

	if (!sync_page_io(log->rdev, pos, PAGE_SIZE, page, WRITE_FUA, false)) {
		__free_page(page);
		return -EIO;
	}
	__free_page(page);
	return 0;
}

static int r5l_recovery_log(struct r5l_log *log)
{
	struct r5l_recovery_ctx ctx;

	ctx.pos = log->last_checkpoint;
	ctx.seq = log->last_cp_seq;
	ctx.meta_page = alloc_page(GFP_KERNEL);
	if (!ctx.meta_page)
		return -ENOMEM;

	r5l_recovery_flush_log(log, &ctx);
	__free_page(ctx.meta_page);

	/*
	 * we did a recovery. Now ctx.pos points to an invalid meta block. New
	 * log will start here. but we can't let superblock point to last valid
	 * meta block. The log might looks like:
	 * | meta 1| meta 2| meta 3|
	 * meta 1 is valid, meta 2 is invalid. meta 3 could be valid. If
	 * superblock points to meta 1, we write a new valid meta 2n.  if crash
	 * happens again, new recovery will start from meta 1. Since meta 2n is
	 * valid now, recovery will think meta 3 is valid, which is wrong.
	 * The solution is we create a new meta in meta2 with its seq == meta
	 * 1's seq + 10 and let superblock points to meta2. The same recovery will
	 * not think meta 3 is a valid meta, because its seq doesn't match
	 */
	if (ctx.seq > log->last_cp_seq + 1) {
		int ret;

		ret = r5l_log_write_empty_meta_block(log, ctx.pos, ctx.seq + 10);
		if (ret)
			return ret;
		log->seq = ctx.seq + 11;
		log->log_start = r5l_ring_add(log, ctx.pos, BLOCK_SECTORS);
		r5l_write_super(log, ctx.pos);
	} else {
		log->log_start = ctx.pos;
		log->seq = ctx.seq;
	}
	return 0;
}

static void r5l_write_super(struct r5l_log *log, sector_t cp)
{
	struct mddev *mddev = log->rdev->mddev;

	log->rdev->journal_tail = cp;
	set_bit(MD_CHANGE_DEVS, &mddev->flags);
}

static int r5l_load_log(struct r5l_log *log)
{
	struct md_rdev *rdev = log->rdev;
	struct page *page;
	struct r5l_meta_block *mb;
	sector_t cp = log->rdev->journal_tail;
	u32 stored_crc, expected_crc;
	bool create_super = false;
	int ret;

	/* Make sure it's valid */
	if (cp >= rdev->sectors || round_down(cp, BLOCK_SECTORS) != cp)
		cp = 0;
	page = alloc_page(GFP_KERNEL);
	if (!page)
		return -ENOMEM;

	if (!sync_page_io(rdev, cp, PAGE_SIZE, page, READ, false)) {
		ret = -EIO;
		goto ioerr;
	}
	mb = page_address(page);

	if (le32_to_cpu(mb->magic) != R5LOG_MAGIC ||
	    mb->version != R5LOG_VERSION) {
		create_super = true;
		goto create;
	}
	stored_crc = le32_to_cpu(mb->checksum);
	mb->checksum = 0;
	expected_crc = crc32c_le(log->uuid_checksum, mb, PAGE_SIZE);
	if (stored_crc != expected_crc) {
		create_super = true;
		goto create;
	}
	if (le64_to_cpu(mb->position) != cp) {
		create_super = true;
		goto create;
	}
create:
	if (create_super) {
		log->last_cp_seq = prandom_u32();
		cp = 0;
		/*
		 * Make sure super points to correct address. Log might have
		 * data very soon. If super hasn't correct log tail address,
		 * recovery can't find the log
		 */
		r5l_write_super(log, cp);
	} else
		log->last_cp_seq = le64_to_cpu(mb->seq);

	log->device_size = round_down(rdev->sectors, BLOCK_SECTORS);
	log->max_free_space = log->device_size >> RECLAIM_MAX_FREE_SPACE_SHIFT;
	if (log->max_free_space > RECLAIM_MAX_FREE_SPACE)
		log->max_free_space = RECLAIM_MAX_FREE_SPACE;
	log->last_checkpoint = cp;

	__free_page(page);

	return r5l_recovery_log(log);
ioerr:
	__free_page(page);
	return ret;
}

int r5l_init_log(struct r5conf *conf, struct md_rdev *rdev)
{
	struct r5l_log *log;

	if (PAGE_SIZE != 4096)
		return -EINVAL;
	log = kzalloc(sizeof(*log), GFP_KERNEL);
	if (!log)
		return -ENOMEM;
	log->rdev = rdev;

	log->uuid_checksum = crc32c_le(~0, rdev->mddev->uuid,
				       sizeof(rdev->mddev->uuid));

	mutex_init(&log->io_mutex);

	spin_lock_init(&log->io_list_lock);
	INIT_LIST_HEAD(&log->running_ios);
	INIT_LIST_HEAD(&log->io_end_ios);
	INIT_LIST_HEAD(&log->stripe_end_ios);
	INIT_LIST_HEAD(&log->flushing_ios);
	INIT_LIST_HEAD(&log->flushed_ios);
	bio_init(&log->flush_bio);

	log->io_kc = KMEM_CACHE(r5l_io_unit, 0);
	if (!log->io_kc)
		goto io_kc;

	log->reclaim_thread = md_register_thread(r5l_reclaim_thread,
						 log->rdev->mddev, "reclaim");
	if (!log->reclaim_thread)
		goto reclaim_thread;
	init_waitqueue_head(&log->iounit_wait);

	INIT_LIST_HEAD(&log->no_space_stripes);
	spin_lock_init(&log->no_space_stripes_lock);

	if (r5l_load_log(log))
		goto error;

	conf->log = log;
	return 0;
error:
	md_unregister_thread(&log->reclaim_thread);
reclaim_thread:
	kmem_cache_destroy(log->io_kc);
io_kc:
	kfree(log);
	return -EINVAL;
}

void r5l_exit_log(struct r5l_log *log)
{
	/*
	 * at this point all stripes are finished, so io_unit is at least in
	 * STRIPE_END state
	 */
	r5l_wake_reclaim(log, -1L);
	md_unregister_thread(&log->reclaim_thread);
	r5l_do_reclaim(log);
	/*
	 * force a super update, r5l_do_reclaim might updated the super.
	 * mddev->thread is already stopped
	 */
	md_update_sb(log->rdev->mddev, 1);

	kmem_cache_destroy(log->io_kc);
	kfree(log);
}