[linux.git] / fs / ext4 / mmp.c

// SPDX-License-Identifier: GPL-2.0
#include <linux/fs.h>
#include <linux/random.h>
#include <linux/buffer_head.h>
#include <linux/utsname.h>
#include <linux/kthread.h>

#include "ext4.h"

/* Checksumming functions */
static __le32 ext4_mmp_csum(struct super_block *sb, struct mmp_struct *mmp)
{
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	int offset = offsetof(struct mmp_struct, mmp_checksum);
	__u32 csum;

	csum = ext4_chksum(sbi, sbi->s_csum_seed, (char *)mmp, offset);

	return cpu_to_le32(csum);
}

static int ext4_mmp_csum_verify(struct super_block *sb, struct mmp_struct *mmp)
{
	if (!ext4_has_metadata_csum(sb))
		return 1;

	return mmp->mmp_checksum == ext4_mmp_csum(sb, mmp);
}

static void ext4_mmp_csum_set(struct super_block *sb, struct mmp_struct *mmp)
{
	if (!ext4_has_metadata_csum(sb))
		return;

	mmp->mmp_checksum = ext4_mmp_csum(sb, mmp);
}

/*
 * Write the MMP block using REQ_SYNC to try to get the block on-disk
 * faster.
 */
static int write_mmp_block_thawed(struct super_block *sb,
				  struct buffer_head *bh)
{
	struct mmp_struct *mmp = (struct mmp_struct *)(bh->b_data);

	ext4_mmp_csum_set(sb, mmp);
	lock_buffer(bh);
	bh->b_end_io = end_buffer_write_sync;
	get_bh(bh);
	submit_bh(REQ_OP_WRITE | REQ_SYNC | REQ_META | REQ_PRIO, bh);
	wait_on_buffer(bh);
	if (unlikely(!buffer_uptodate(bh)))
		return -EIO;
	return 0;
}

static int write_mmp_block(struct super_block *sb, struct buffer_head *bh)
{
	int err;

	/*
	 * We protect against freezing so that we don't create dirty buffers
	 * on frozen filesystem.
	 */
	sb_start_write(sb);
	err = write_mmp_block_thawed(sb, bh);
	sb_end_write(sb);
	return err;
}

/*
 * Read the MMP block. It _must_ be read from disk and hence we clear the
 * uptodate flag on the buffer.
 */
static int read_mmp_block(struct super_block *sb, struct buffer_head **bh,
			  ext4_fsblk_t mmp_block)
{
	struct mmp_struct *mmp;
	int ret;

	if (*bh)
		clear_buffer_uptodate(*bh);

	/* This would be sb_bread(sb, mmp_block), except we need to be sure
	 * that the MD RAID device cache has been bypassed, and that the read
	 * is not blocked in the elevator. */
	if (!*bh) {
		*bh = sb_getblk(sb, mmp_block);
		if (!*bh) {
			ret = -ENOMEM;
			goto warn_exit;
		}
	}

	lock_buffer(*bh);
	ret = ext4_read_bh(*bh, REQ_META | REQ_PRIO, NULL);
	if (ret)
		goto warn_exit;

	mmp = (struct mmp_struct *)((*bh)->b_data);
	if (le32_to_cpu(mmp->mmp_magic) != EXT4_MMP_MAGIC) {
		ret = -EFSCORRUPTED;
		goto warn_exit;
	}
	if (!ext4_mmp_csum_verify(sb, mmp)) {
		ret = -EFSBADCRC;
		goto warn_exit;
	}
	return 0;
warn_exit:
	brelse(*bh);
	*bh = NULL;
	ext4_warning(sb, "Error %d while reading MMP block %llu",
		     ret, mmp_block);
	return ret;
}

/*
 * Dump as much information as possible to help the admin.
 */
void __dump_mmp_msg(struct super_block *sb, struct mmp_struct *mmp,
		    const char *function, unsigned int line, const char *msg)
{
	__ext4_warning(sb, function, line, "%s", msg);
	__ext4_warning(sb, function, line,
		       "MMP failure info: last update time: %llu, last update node: %.*s, last update device: %.*s",
		       (unsigned long long)le64_to_cpu(mmp->mmp_time),
		       (int)sizeof(mmp->mmp_nodename), mmp->mmp_nodename,
		       (int)sizeof(mmp->mmp_bdevname), mmp->mmp_bdevname);
}

/*
 * kmmpd will update the MMP sequence every s_mmp_update_interval seconds
 */
static int kmmpd(void *data)
{
	struct super_block *sb = data;
	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
	struct buffer_head *bh = EXT4_SB(sb)->s_mmp_bh;
	struct mmp_struct *mmp;
	ext4_fsblk_t mmp_block;
	u32 seq = 0;
	unsigned long failed_writes = 0;
	int mmp_update_interval = le16_to_cpu(es->s_mmp_update_interval);
	unsigned mmp_check_interval;
	unsigned long last_update_time;
	unsigned long diff;
	int retval = 0;

	mmp_block = le64_to_cpu(es->s_mmp_block);
	mmp = (struct mmp_struct *)(bh->b_data);
	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
	/*
	 * Start with the higher mmp_check_interval and reduce it if
	 * the MMP block is being updated on time.
	 */
	mmp_check_interval = max(EXT4_MMP_CHECK_MULT * mmp_update_interval,
				 EXT4_MMP_MIN_CHECK_INTERVAL);
	mmp->mmp_check_interval = cpu_to_le16(mmp_check_interval);

	memcpy(mmp->mmp_nodename, init_utsname()->nodename,
	       sizeof(mmp->mmp_nodename));

	while (!kthread_should_stop() && !ext4_forced_shutdown(sb)) {
		if (!ext4_has_feature_mmp(sb)) {
			ext4_warning(sb, "kmmpd being stopped since MMP feature"
				     " has been disabled.");
			goto wait_to_exit;
		}
		if (++seq > EXT4_MMP_SEQ_MAX)
			seq = 1;

		mmp->mmp_seq = cpu_to_le32(seq);
		mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
		last_update_time = jiffies;

		retval = write_mmp_block(sb, bh);
		/*
		 * Don't spew too many error messages. Print one every
		 * (s_mmp_update_interval * 60) seconds.
		 */
		if (retval) {
			if ((failed_writes % 60) == 0) {
				ext4_error_err(sb, -retval,
					       "Error writing to MMP block");
			}
			failed_writes++;
		}

		diff = jiffies - last_update_time;
		if (diff < mmp_update_interval * HZ)
			schedule_timeout_interruptible(mmp_update_interval *
						       HZ - diff);

		/*
		 * We need to make sure that more than mmp_check_interval
		 * seconds have not passed since writing. If that has happened
		 * we need to check if the MMP block is as we left it.
		 */
		diff = jiffies - last_update_time;
		if (diff > mmp_check_interval * HZ) {
			struct buffer_head *bh_check = NULL;
			struct mmp_struct *mmp_check;

			retval = read_mmp_block(sb, &bh_check, mmp_block);
			if (retval) {
				ext4_error_err(sb, -retval,
					       "error reading MMP data: %d",
					       retval);
				goto wait_to_exit;
			}

			mmp_check = (struct mmp_struct *)(bh_check->b_data);
			if (mmp->mmp_seq != mmp_check->mmp_seq ||
			    memcmp(mmp->mmp_nodename, mmp_check->mmp_nodename,
				   sizeof(mmp->mmp_nodename))) {
				dump_mmp_msg(sb, mmp_check,
					     "Error while updating MMP info. "
					     "The filesystem seems to have been"
					     " multiply mounted.");
				ext4_error_err(sb, EBUSY, "abort");
				put_bh(bh_check);
				retval = -EBUSY;
				goto wait_to_exit;
			}
			put_bh(bh_check);
		}

		 /*
		 * Adjust the mmp_check_interval depending on how much time
		 * it took for the MMP block to be written.
		 */
		mmp_check_interval = max(min(EXT4_MMP_CHECK_MULT * diff / HZ,
					     EXT4_MMP_MAX_CHECK_INTERVAL),
					 EXT4_MMP_MIN_CHECK_INTERVAL);
		mmp->mmp_check_interval = cpu_to_le16(mmp_check_interval);
	}

	/*
	 * Unmount seems to be clean.
	 */
	mmp->mmp_seq = cpu_to_le32(EXT4_MMP_SEQ_CLEAN);
	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());

	retval = write_mmp_block(sb, bh);

wait_to_exit:
	while (!kthread_should_stop()) {
		set_current_state(TASK_INTERRUPTIBLE);
		if (!kthread_should_stop())
			schedule();
	}
	set_current_state(TASK_RUNNING);
	return retval;
}

void ext4_stop_mmpd(struct ext4_sb_info *sbi)
{
	if (sbi->s_mmp_tsk) {
		kthread_stop(sbi->s_mmp_tsk);
		brelse(sbi->s_mmp_bh);
		sbi->s_mmp_tsk = NULL;
	}
}

/*
 * Get a random new sequence number but make sure it is not greater than
 * EXT4_MMP_SEQ_MAX.
 */
static unsigned int mmp_new_seq(void)
{
	return get_random_u32_below(EXT4_MMP_SEQ_MAX + 1);
}

/*
 * Protect the filesystem from being mounted more than once.
 */
int ext4_multi_mount_protect(struct super_block *sb,
				    ext4_fsblk_t mmp_block)
{
	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
	struct buffer_head *bh = NULL;
	struct mmp_struct *mmp = NULL;
	u32 seq;
	unsigned int mmp_check_interval = le16_to_cpu(es->s_mmp_update_interval);
	unsigned int wait_time = 0;
	int retval;

	if (mmp_block < le32_to_cpu(es->s_first_data_block) ||
	    mmp_block >= ext4_blocks_count(es)) {
		ext4_warning(sb, "Invalid MMP block in superblock");
		retval = -EINVAL;
		goto failed;
	}

	retval = read_mmp_block(sb, &bh, mmp_block);
	if (retval)
		goto failed;

	mmp = (struct mmp_struct *)(bh->b_data);

	if (mmp_check_interval < EXT4_MMP_MIN_CHECK_INTERVAL)
		mmp_check_interval = EXT4_MMP_MIN_CHECK_INTERVAL;

	/*
	 * If check_interval in MMP block is larger, use that instead of
	 * update_interval from the superblock.
	 */
	if (le16_to_cpu(mmp->mmp_check_interval) > mmp_check_interval)
		mmp_check_interval = le16_to_cpu(mmp->mmp_check_interval);

	seq = le32_to_cpu(mmp->mmp_seq);
	if (seq == EXT4_MMP_SEQ_CLEAN)
		goto skip;

	if (seq == EXT4_MMP_SEQ_FSCK) {
		dump_mmp_msg(sb, mmp, "fsck is running on the filesystem");
		retval = -EBUSY;
		goto failed;
	}

	wait_time = min(mmp_check_interval * 2 + 1,
			mmp_check_interval + 60);

	/* Print MMP interval if more than 20 secs. */
	if (wait_time > EXT4_MMP_MIN_CHECK_INTERVAL * 4)
		ext4_warning(sb, "MMP interval %u higher than expected, please"
			     " wait.\n", wait_time * 2);

	if (schedule_timeout_interruptible(HZ * wait_time) != 0) {
		ext4_warning(sb, "MMP startup interrupted, failing mount\n");
		retval = -ETIMEDOUT;
		goto failed;
	}

	retval = read_mmp_block(sb, &bh, mmp_block);
	if (retval)
		goto failed;
	mmp = (struct mmp_struct *)(bh->b_data);
	if (seq != le32_to_cpu(mmp->mmp_seq)) {
		dump_mmp_msg(sb, mmp,
			     "Device is already active on another node.");
		retval = -EBUSY;
		goto failed;
	}

skip:
	/*
	 * write a new random sequence number.
	 */
	seq = mmp_new_seq();
	mmp->mmp_seq = cpu_to_le32(seq);

	/*
	 * On mount / remount we are protected against fs freezing (by s_umount
	 * semaphore) and grabbing freeze protection upsets lockdep
	 */
	retval = write_mmp_block_thawed(sb, bh);
	if (retval)
		goto failed;

	/*
	 * wait for MMP interval and check mmp_seq.
	 */
	if (schedule_timeout_interruptible(HZ * wait_time) != 0) {
		ext4_warning(sb, "MMP startup interrupted, failing mount");
		retval = -ETIMEDOUT;
		goto failed;
	}

	retval = read_mmp_block(sb, &bh, mmp_block);
	if (retval)
		goto failed;
	mmp = (struct mmp_struct *)(bh->b_data);
	if (seq != le32_to_cpu(mmp->mmp_seq)) {
		dump_mmp_msg(sb, mmp,
			     "Device is already active on another node.");
		retval = -EBUSY;
		goto failed;
	}

	EXT4_SB(sb)->s_mmp_bh = bh;

	BUILD_BUG_ON(sizeof(mmp->mmp_bdevname) < BDEVNAME_SIZE);
	snprintf(mmp->mmp_bdevname, sizeof(mmp->mmp_bdevname),
		 "%pg", bh->b_bdev);

	/*
	 * Start a kernel thread to update the MMP block periodically.
	 */
	EXT4_SB(sb)->s_mmp_tsk = kthread_run(kmmpd, sb, "kmmpd-%.*s",
					     (int)sizeof(mmp->mmp_bdevname),
					     mmp->mmp_bdevname);
	if (IS_ERR(EXT4_SB(sb)->s_mmp_tsk)) {
		EXT4_SB(sb)->s_mmp_tsk = NULL;
		ext4_warning(sb, "Unable to create kmmpd thread for %s.",
			     sb->s_id);
		retval = -ENOMEM;
		goto failed;
	}

	return 0;

failed:
	brelse(bh);
	return retval;
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
c5e06d10 JL	2	#include <linux/fs.h>
	3	#include <linux/random.h>
	4	#include <linux/buffer_head.h>
	5	#include <linux/utsname.h>
	6	#include <linux/kthread.h>
	7
	8	#include "ext4.h"
	9
5c359a47	10	/* Checksumming functions */
171a7f21	11	static __le32 ext4_mmp_csum(struct super_block sb, struct mmp_struct mmp)
5c359a47 DW	12	{
	13	struct ext4_sb_info *sbi = EXT4_SB(sb);
	14	int offset = offsetof(struct mmp_struct, mmp_checksum);
	15	__u32 csum;
	16
	17	csum = ext4_chksum(sbi, sbi->s_csum_seed, (char *)mmp, offset);
	18
	19	return cpu_to_le32(csum);
	20	}
	21
c197855e	22	static int ext4_mmp_csum_verify(struct super_block sb, struct mmp_struct mmp)
5c359a47	23	{
9aa5d32b	24	if (!ext4_has_metadata_csum(sb))
5c359a47 DW	25	return 1;
	26
	27	return mmp->mmp_checksum == ext4_mmp_csum(sb, mmp);
	28	}
	29
c197855e	30	static void ext4_mmp_csum_set(struct super_block sb, struct mmp_struct mmp)
5c359a47	31	{
9aa5d32b	32	if (!ext4_has_metadata_csum(sb))
5c359a47 DW	33	return;
	34
	35	mmp->mmp_checksum = ext4_mmp_csum(sb, mmp);
	36	}
	37
c5e06d10	38	/*
70fd7614	39	* Write the MMP block using REQ_SYNC to try to get the block on-disk
c5e06d10 JL	40	* faster.
c5e06d10 JL	41	*/
949f95ff JK	42	static int write_mmp_block_thawed(struct super_block *sb,
949f95ff JK	43	struct buffer_head *bh)
c5e06d10	44	{
5c359a47 DW	45	struct mmp_struct mmp = (struct mmp_struct )(bh->b_data);
	46
	47	ext4_mmp_csum_set(sb, mmp);
c5e06d10 JL	48	lock_buffer(bh);
	49	bh->b_end_io = end_buffer_write_sync;
	50	get_bh(bh);
1420c4a5	51	submit_bh(REQ_OP_WRITE \| REQ_SYNC \| REQ_META \| REQ_PRIO, bh);
c5e06d10 JL	52	wait_on_buffer(bh);
c5e06d10 JL	53	if (unlikely(!buffer_uptodate(bh)))
6810fad9	54	return -EIO;
c5e06d10 JL	55	return 0;
	56	}
	57
949f95ff JK	58	static int write_mmp_block(struct super_block sb, struct buffer_head bh)
	59	{
	60	int err;
	61
	62	/*
	63	* We protect against freezing so that we don't create dirty buffers
	64	* on frozen filesystem.
	65	*/
	66	sb_start_write(sb);
	67	err = write_mmp_block_thawed(sb, bh);
	68	sb_end_write(sb);
	69	return err;
	70	}
	71
c5e06d10 JL	72	/*
	73	* Read the MMP block. It _must_ be read from disk and hence we clear the
	74	* uptodate flag on the buffer.
	75	*/
	76	static int read_mmp_block(struct super_block sb, struct buffer_head *bh,
	77	ext4_fsblk_t mmp_block)
	78	{
	79	struct mmp_struct *mmp;
98104468	80	int ret;
c5e06d10 JL	81
	82	if (*bh)
	83	clear_buffer_uptodate(*bh);
	84
	85	/* This would be sb_bread(sb, mmp_block), except we need to be sure
	86	* that the MD RAID device cache has been bypassed, and that the read
	87	* is not blocked in the elevator. */
98104468	88	if (!*bh) {
c5e06d10	89	*bh = sb_getblk(sb, mmp_block);
98104468 DC	90	if (!*bh) {
	91	ret = -ENOMEM;
	92	goto warn_exit;
c5e06d10 JL	93	}
c5e06d10 JL	94	}
98104468	95
98104468	96	lock_buffer(*bh);
2d069c08	97	ret = ext4_read_bh(*bh, REQ_META \| REQ_PRIO, NULL);
2d069c08	98	if (ret)
98104468	99	goto warn_exit;
2d069c08	100
c5e06d10	101	mmp = (struct mmp_struct )((bh)->b_data);
03046886	102	if (le32_to_cpu(mmp->mmp_magic) != EXT4_MMP_MAGIC) {
6a797d27	103	ret = -EFSCORRUPTED;
03046886	104	goto warn_exit;
	105	}
	106	if (!ext4_mmp_csum_verify(sb, mmp)) {
6a797d27	107	ret = -EFSBADCRC;
03046886	108	goto warn_exit;
	109	}
	110	return 0;
98104468	111	warn_exit:
03046886	112	brelse(*bh);
03046886	113	*bh = NULL;
98104468 DC	114	ext4_warning(sb, "Error %d while reading MMP block %llu",
	115	ret, mmp_block);
	116	return ret;
c5e06d10 JL	117	}
	118
	119	/*
	120	* Dump as much information as possible to help the admin.
	121	*/
	122	void __dump_mmp_msg(struct super_block sb, struct mmp_struct mmp,
	123	const char function, unsigned int line, const char msg)
	124	{
da0b5e40	125	__ext4_warning(sb, function, line, "%s", msg);
c5e06d10	126	__ext4_warning(sb, function, line,
14c9ca05 AD	127	"MMP failure info: last update time: %llu, last update node: %.s, last update device: %.s",
	128	(unsigned long long)le64_to_cpu(mmp->mmp_time),
	129	(int)sizeof(mmp->mmp_nodename), mmp->mmp_nodename,
	130	(int)sizeof(mmp->mmp_bdevname), mmp->mmp_bdevname);
c5e06d10 JL	131	}
	132
	133	/*
	134	* kmmpd will update the MMP sequence every s_mmp_update_interval seconds
	135	*/
	136	static int kmmpd(void *data)
	137	{
c30365b9	138	struct super_block *sb = data;
c5e06d10	139	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
618f0031	140	struct buffer_head *bh = EXT4_SB(sb)->s_mmp_bh;
c5e06d10 JL	141	struct mmp_struct *mmp;
	142	ext4_fsblk_t mmp_block;
	143	u32 seq = 0;
	144	unsigned long failed_writes = 0;
	145	int mmp_update_interval = le16_to_cpu(es->s_mmp_update_interval);
	146	unsigned mmp_check_interval;
	147	unsigned long last_update_time;
	148	unsigned long diff;
b6654142	149	int retval = 0;
c5e06d10 JL	150
	151	mmp_block = le64_to_cpu(es->s_mmp_block);
	152	mmp = (struct mmp_struct *)(bh->b_data);
af123b37	153	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
c5e06d10 JL	154	/*
	155	* Start with the higher mmp_check_interval and reduce it if
	156	* the MMP block is being updated on time.
	157	*/
	158	mmp_check_interval = max(EXT4_MMP_CHECK_MULT * mmp_update_interval,
	159	EXT4_MMP_MIN_CHECK_INTERVAL);
	160	mmp->mmp_check_interval = cpu_to_le16(mmp_check_interval);
c5e06d10	161
215fc6af	162	memcpy(mmp->mmp_nodename, init_utsname()->nodename,
c5e06d10 JL	163	sizeof(mmp->mmp_nodename));
c5e06d10 JL	164
1e1566b9	165	while (!kthread_should_stop() && !ext4_forced_shutdown(sb)) {
61bb4a1c TT	166	if (!ext4_has_feature_mmp(sb)) {
	167	ext4_warning(sb, "kmmpd being stopped since MMP feature"
	168	" has been disabled.");
	169	goto wait_to_exit;
	170	}
c5e06d10 JL	171	if (++seq > EXT4_MMP_SEQ_MAX)
	172	seq = 1;
	173
	174	mmp->mmp_seq = cpu_to_le32(seq);
af123b37	175	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
c5e06d10 JL	176	last_update_time = jiffies;
c5e06d10 JL	177
5c359a47	178	retval = write_mmp_block(sb, bh);
c5e06d10 JL	179	/*
	180	* Don't spew too many error messages. Print one every
	181	* (s_mmp_update_interval * 60) seconds.
	182	*/
bdfc230f	183	if (retval) {
878520ac	184	if ((failed_writes % 60) == 0) {
54d3adbc TT	185	ext4_error_err(sb, -retval,
54d3adbc TT	186	"Error writing to MMP block");
878520ac	187	}
c5e06d10 JL	188	failed_writes++;
	189	}
	190
c5e06d10 JL	191	diff = jiffies - last_update_time;
	192	if (diff < mmp_update_interval * HZ)
	193	schedule_timeout_interruptible(mmp_update_interval *
	194	HZ - diff);
	195
	196	/*
	197	* We need to make sure that more than mmp_check_interval
	198	* seconds have not passed since writing. If that has happened
	199	* we need to check if the MMP block is as we left it.
	200	*/
	201	diff = jiffies - last_update_time;
	202	if (diff > mmp_check_interval * HZ) {
	203	struct buffer_head *bh_check = NULL;
	204	struct mmp_struct *mmp_check;
	205
	206	retval = read_mmp_block(sb, &bh_check, mmp_block);
	207	if (retval) {
54d3adbc TT	208	ext4_error_err(sb, -retval,
	209	"error reading MMP data: %d",
	210	retval);
61bb4a1c	211	goto wait_to_exit;
c5e06d10 JL	212	}
	213
	214	mmp_check = (struct mmp_struct *)(bh_check->b_data);
	215	if (mmp->mmp_seq != mmp_check->mmp_seq \|\|
	216	memcmp(mmp->mmp_nodename, mmp_check->mmp_nodename,
	217	sizeof(mmp->mmp_nodename))) {
	218	dump_mmp_msg(sb, mmp_check,
	219	"Error while updating MMP info. "
	220	"The filesystem seems to have been"
	221	" multiply mounted.");
54d3adbc	222	ext4_error_err(sb, EBUSY, "abort");
03046886	223	put_bh(bh_check);
03046886	224	retval = -EBUSY;
61bb4a1c	225	goto wait_to_exit;
c5e06d10 JL	226	}
	227	put_bh(bh_check);
	228	}
	229
	230	/*
	231	* Adjust the mmp_check_interval depending on how much time
	232	* it took for the MMP block to be written.
	233	*/
	234	mmp_check_interval = max(min(EXT4_MMP_CHECK_MULT * diff / HZ,
	235	EXT4_MMP_MAX_CHECK_INTERVAL),
	236	EXT4_MMP_MIN_CHECK_INTERVAL);
	237	mmp->mmp_check_interval = cpu_to_le16(mmp_check_interval);
	238	}
	239
	240	/*
	241	* Unmount seems to be clean.
	242	*/
	243	mmp->mmp_seq = cpu_to_le32(EXT4_MMP_SEQ_CLEAN);
af123b37	244	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
c5e06d10	245
5c359a47	246	retval = write_mmp_block(sb, bh);
c5e06d10	247
61bb4a1c TT	248	wait_to_exit:
	249	while (!kthread_should_stop()) {
	250	set_current_state(TASK_INTERRUPTIBLE);
	251	if (!kthread_should_stop())
	252	schedule();
	253	}
	254	set_current_state(TASK_RUNNING);
c5e06d10 JL	255	return retval;
	256	}
	257
618f0031 PS	258	void ext4_stop_mmpd(struct ext4_sb_info *sbi)
	259	{
	260	if (sbi->s_mmp_tsk) {
	261	kthread_stop(sbi->s_mmp_tsk);
	262	brelse(sbi->s_mmp_bh);
	263	sbi->s_mmp_tsk = NULL;
	264	}
	265	}
	266
c5e06d10 JL	267	/*
	268	* Get a random new sequence number but make sure it is not greater than
	269	* EXT4_MMP_SEQ_MAX.
	270	*/
	271	static unsigned int mmp_new_seq(void)
	272	{
d247aabd	273	return get_random_u32_below(EXT4_MMP_SEQ_MAX + 1);
c5e06d10 JL	274	}
	275
	276	/*
	277	* Protect the filesystem from being mounted more than once.
	278	*/
	279	int ext4_multi_mount_protect(struct super_block *sb,
	280	ext4_fsblk_t mmp_block)
	281	{
	282	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
	283	struct buffer_head *bh = NULL;
	284	struct mmp_struct *mmp = NULL;
c5e06d10 JL	285	u32 seq;
	286	unsigned int mmp_check_interval = le16_to_cpu(es->s_mmp_update_interval);
	287	unsigned int wait_time = 0;
	288	int retval;
	289
	290	if (mmp_block < le32_to_cpu(es->s_first_data_block) \|\|
	291	mmp_block >= ext4_blocks_count(es)) {
	292	ext4_warning(sb, "Invalid MMP block in superblock");
3b50d501	293	retval = -EINVAL;
c5e06d10 JL	294	goto failed;
	295	}
	296
	297	retval = read_mmp_block(sb, &bh, mmp_block);
	298	if (retval)
	299	goto failed;
	300
	301	mmp = (struct mmp_struct *)(bh->b_data);
	302
	303	if (mmp_check_interval < EXT4_MMP_MIN_CHECK_INTERVAL)
	304	mmp_check_interval = EXT4_MMP_MIN_CHECK_INTERVAL;
	305
	306	/*
	307	* If check_interval in MMP block is larger, use that instead of
	308	* update_interval from the superblock.
	309	*/
85d21650 SN	310	if (le16_to_cpu(mmp->mmp_check_interval) > mmp_check_interval)
85d21650 SN	311	mmp_check_interval = le16_to_cpu(mmp->mmp_check_interval);
c5e06d10 JL	312
	313	seq = le32_to_cpu(mmp->mmp_seq);
	314	if (seq == EXT4_MMP_SEQ_CLEAN)
	315	goto skip;
	316
	317	if (seq == EXT4_MMP_SEQ_FSCK) {
	318	dump_mmp_msg(sb, mmp, "fsck is running on the filesystem");
3b50d501	319	retval = -EBUSY;
c5e06d10 JL	320	goto failed;
	321	}
	322
	323	wait_time = min(mmp_check_interval * 2 + 1,
	324	mmp_check_interval + 60);
	325
	326	/* Print MMP interval if more than 20 secs. */
	327	if (wait_time > EXT4_MMP_MIN_CHECK_INTERVAL * 4)
	328	ext4_warning(sb, "MMP interval %u higher than expected, please"
	329	" wait.\n", wait_time * 2);
	330
	331	if (schedule_timeout_interruptible(HZ * wait_time) != 0) {
	332	ext4_warning(sb, "MMP startup interrupted, failing mount\n");
3b50d501	333	retval = -ETIMEDOUT;
c5e06d10 JL	334	goto failed;
	335	}
	336
	337	retval = read_mmp_block(sb, &bh, mmp_block);
	338	if (retval)
	339	goto failed;
	340	mmp = (struct mmp_struct *)(bh->b_data);
	341	if (seq != le32_to_cpu(mmp->mmp_seq)) {
	342	dump_mmp_msg(sb, mmp,
	343	"Device is already active on another node.");
3b50d501	344	retval = -EBUSY;
c5e06d10 JL	345	goto failed;
	346	}
	347
	348	skip:
	349	/*
	350	* write a new random sequence number.
	351	*/
f6f96fdb DW	352	seq = mmp_new_seq();
f6f96fdb DW	353	mmp->mmp_seq = cpu_to_le32(seq);
c5e06d10	354
949f95ff JK	355	/*
	356	* On mount / remount we are protected against fs freezing (by s_umount
	357	* semaphore) and grabbing freeze protection upsets lockdep
	358	*/
	359	retval = write_mmp_block_thawed(sb, bh);
c5e06d10 JL	360	if (retval)
	361	goto failed;
	362
	363	/*
	364	* wait for MMP interval and check mmp_seq.
	365	*/
	366	if (schedule_timeout_interruptible(HZ * wait_time) != 0) {
8d2ae1cb	367	ext4_warning(sb, "MMP startup interrupted, failing mount");
3b50d501	368	retval = -ETIMEDOUT;
c5e06d10 JL	369	goto failed;
	370	}
	371
	372	retval = read_mmp_block(sb, &bh, mmp_block);
	373	if (retval)
	374	goto failed;
	375	mmp = (struct mmp_struct *)(bh->b_data);
	376	if (seq != le32_to_cpu(mmp->mmp_seq)) {
	377	dump_mmp_msg(sb, mmp,
	378	"Device is already active on another node.");
3b50d501	379	retval = -EBUSY;
c5e06d10 JL	380	goto failed;
	381	}
	382
618f0031	383	EXT4_SB(sb)->s_mmp_bh = bh;
c5e06d10	384
c5b045b9	385	BUILD_BUG_ON(sizeof(mmp->mmp_bdevname) < BDEVNAME_SIZE);
900d156b CH	386	snprintf(mmp->mmp_bdevname, sizeof(mmp->mmp_bdevname),
900d156b CH	387	"%pg", bh->b_bdev);
c5b045b9	388
c5e06d10 JL	389	/*
	390	* Start a kernel thread to update the MMP block periodically.
	391	*/
618f0031	392	EXT4_SB(sb)->s_mmp_tsk = kthread_run(kmmpd, sb, "kmmpd-%.*s",
14c9ca05	393	(int)sizeof(mmp->mmp_bdevname),
c5b045b9	394	mmp->mmp_bdevname);
c5e06d10 JL	395	if (IS_ERR(EXT4_SB(sb)->s_mmp_tsk)) {
c5e06d10 JL	396	EXT4_SB(sb)->s_mmp_tsk = NULL;
c5e06d10 JL	397	ext4_warning(sb, "Unable to create kmmpd thread for %s.",
c5e06d10 JL	398	sb->s_id);
3b50d501	399	retval = -ENOMEM;
c5e06d10 JL	400	goto failed;
	401	}
	402
	403	return 0;
	404
	405	failed:
	406	brelse(bh);
3b50d501	407	return retval;
c5e06d10	408	}