[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(struct super_block *sb, struct buffer_head *bh)
{
	struct mmp_struct *mmp = (struct mmp_struct *)(bh->b_data);

	/*
	 * We protect against freezing so that we don't create dirty buffers
	 * on frozen filesystem.
	 */
	sb_start_write(sb);
	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);
	sb_end_write(sb);
	if (unlikely(!buffer_uptodate(bh)))
		return -EIO;

	return 0;
}

/*
 * 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 = (struct super_block *) 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);
	BUILD_BUG_ON(sizeof(mmp->mmp_bdevname) < BDEVNAME_SIZE);
	bdevname(bh->b_bdev, mmp->mmp_bdevname);

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

	while (!kthread_should_stop() && !sb_rdonly(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)
{
	u32 new_seq;

	do {
		new_seq = prandom_u32();
	} while (new_seq > EXT4_MMP_SEQ_MAX);

	return new_seq;
}

/*
 * 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");
		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");
		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");
		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.");
		goto failed;
	}

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

	retval = write_mmp_block(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");
		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.");
		goto failed;
	}

	EXT4_SB(sb)->s_mmp_bh = bh;

	/*
	 * 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),
					     bdevname(bh->b_bdev,
						      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);
		goto failed;
	}

	return 0;

failed:
	brelse(bh);
	return 1;
}
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	*/
5c359a47	42	static int write_mmp_block(struct super_block sb, struct buffer_head bh)
c5e06d10	43	{
5c359a47 DW	44	struct mmp_struct mmp = (struct mmp_struct )(bh->b_data);
5c359a47 DW	45
8e8ad8a5 JK	46	/*
	47	* We protect against freezing so that we don't create dirty buffers
	48	* on frozen filesystem.
	49	*/
	50	sb_start_write(sb);
5c359a47	51	ext4_mmp_csum_set(sb, mmp);
c5e06d10 JL	52	lock_buffer(bh);
	53	bh->b_end_io = end_buffer_write_sync;
	54	get_bh(bh);
70fd7614	55	submit_bh(REQ_OP_WRITE, REQ_SYNC \| REQ_META \| REQ_PRIO, bh);
c5e06d10	56	wait_on_buffer(bh);
8e8ad8a5	57	sb_end_write(sb);
c5e06d10	58	if (unlikely(!buffer_uptodate(bh)))
6810fad9	59	return -EIO;
c5e06d10 JL	60
	61	return 0;
	62	}
	63
	64	/*
	65	* Read the MMP block. It _must_ be read from disk and hence we clear the
	66	* uptodate flag on the buffer.
	67	*/
	68	static int read_mmp_block(struct super_block sb, struct buffer_head *bh,
	69	ext4_fsblk_t mmp_block)
	70	{
	71	struct mmp_struct *mmp;
98104468	72	int ret;
c5e06d10 JL	73
	74	if (*bh)
	75	clear_buffer_uptodate(*bh);
	76
	77	/* This would be sb_bread(sb, mmp_block), except we need to be sure
	78	* that the MD RAID device cache has been bypassed, and that the read
	79	* is not blocked in the elevator. */
98104468	80	if (!*bh) {
c5e06d10	81	*bh = sb_getblk(sb, mmp_block);
98104468 DC	82	if (!*bh) {
	83	ret = -ENOMEM;
	84	goto warn_exit;
c5e06d10 JL	85	}
c5e06d10 JL	86	}
98104468	87
98104468	88	lock_buffer(*bh);
2d069c08	89	ret = ext4_read_bh(*bh, REQ_META \| REQ_PRIO, NULL);
2d069c08	90	if (ret)
98104468	91	goto warn_exit;
2d069c08	92
c5e06d10	93	mmp = (struct mmp_struct )((bh)->b_data);
03046886	94	if (le32_to_cpu(mmp->mmp_magic) != EXT4_MMP_MAGIC) {
6a797d27	95	ret = -EFSCORRUPTED;
03046886	96	goto warn_exit;
	97	}
	98	if (!ext4_mmp_csum_verify(sb, mmp)) {
6a797d27	99	ret = -EFSBADCRC;
03046886	100	goto warn_exit;
	101	}
	102	return 0;
98104468	103	warn_exit:
03046886	104	brelse(*bh);
03046886	105	*bh = NULL;
98104468 DC	106	ext4_warning(sb, "Error %d while reading MMP block %llu",
	107	ret, mmp_block);
	108	return ret;
c5e06d10 JL	109	}
	110
	111	/*
	112	* Dump as much information as possible to help the admin.
	113	*/
	114	void __dump_mmp_msg(struct super_block sb, struct mmp_struct mmp,
	115	const char function, unsigned int line, const char msg)
	116	{
da0b5e40	117	__ext4_warning(sb, function, line, "%s", msg);
c5e06d10	118	__ext4_warning(sb, function, line,
14c9ca05 AD	119	"MMP failure info: last update time: %llu, last update node: %.s, last update device: %.s",
	120	(unsigned long long)le64_to_cpu(mmp->mmp_time),
	121	(int)sizeof(mmp->mmp_nodename), mmp->mmp_nodename,
	122	(int)sizeof(mmp->mmp_bdevname), mmp->mmp_bdevname);
c5e06d10 JL	123	}
	124
	125	/*
	126	* kmmpd will update the MMP sequence every s_mmp_update_interval seconds
	127	*/
	128	static int kmmpd(void *data)
	129	{
618f0031	130	struct super_block sb = (struct super_block ) data;
c5e06d10	131	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
618f0031	132	struct buffer_head *bh = EXT4_SB(sb)->s_mmp_bh;
c5e06d10 JL	133	struct mmp_struct *mmp;
	134	ext4_fsblk_t mmp_block;
	135	u32 seq = 0;
	136	unsigned long failed_writes = 0;
	137	int mmp_update_interval = le16_to_cpu(es->s_mmp_update_interval);
	138	unsigned mmp_check_interval;
	139	unsigned long last_update_time;
	140	unsigned long diff;
b6654142	141	int retval = 0;
c5e06d10 JL	142
	143	mmp_block = le64_to_cpu(es->s_mmp_block);
	144	mmp = (struct mmp_struct *)(bh->b_data);
af123b37	145	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
c5e06d10 JL	146	/*
	147	* Start with the higher mmp_check_interval and reduce it if
	148	* the MMP block is being updated on time.
	149	*/
	150	mmp_check_interval = max(EXT4_MMP_CHECK_MULT * mmp_update_interval,
	151	EXT4_MMP_MIN_CHECK_INTERVAL);
	152	mmp->mmp_check_interval = cpu_to_le16(mmp_check_interval);
14c9ca05	153	BUILD_BUG_ON(sizeof(mmp->mmp_bdevname) < BDEVNAME_SIZE);
c5e06d10 JL	154	bdevname(bh->b_bdev, mmp->mmp_bdevname);
c5e06d10 JL	155
215fc6af	156	memcpy(mmp->mmp_nodename, init_utsname()->nodename,
c5e06d10 JL	157	sizeof(mmp->mmp_nodename));
c5e06d10 JL	158
61bb4a1c TT	159	while (!kthread_should_stop() && !sb_rdonly(sb)) {
	160	if (!ext4_has_feature_mmp(sb)) {
	161	ext4_warning(sb, "kmmpd being stopped since MMP feature"
	162	" has been disabled.");
	163	goto wait_to_exit;
	164	}
c5e06d10 JL	165	if (++seq > EXT4_MMP_SEQ_MAX)
	166	seq = 1;
	167
	168	mmp->mmp_seq = cpu_to_le32(seq);
af123b37	169	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
c5e06d10 JL	170	last_update_time = jiffies;
c5e06d10 JL	171
5c359a47	172	retval = write_mmp_block(sb, bh);
c5e06d10 JL	173	/*
	174	* Don't spew too many error messages. Print one every
	175	* (s_mmp_update_interval * 60) seconds.
	176	*/
bdfc230f	177	if (retval) {
878520ac	178	if ((failed_writes % 60) == 0) {
54d3adbc TT	179	ext4_error_err(sb, -retval,
54d3adbc TT	180	"Error writing to MMP block");
878520ac	181	}
c5e06d10 JL	182	failed_writes++;
	183	}
	184
c5e06d10 JL	185	diff = jiffies - last_update_time;
	186	if (diff < mmp_update_interval * HZ)
	187	schedule_timeout_interruptible(mmp_update_interval *
	188	HZ - diff);
	189
	190	/*
	191	* We need to make sure that more than mmp_check_interval
	192	* seconds have not passed since writing. If that has happened
	193	* we need to check if the MMP block is as we left it.
	194	*/
	195	diff = jiffies - last_update_time;
	196	if (diff > mmp_check_interval * HZ) {
	197	struct buffer_head *bh_check = NULL;
	198	struct mmp_struct *mmp_check;
	199
	200	retval = read_mmp_block(sb, &bh_check, mmp_block);
	201	if (retval) {
54d3adbc TT	202	ext4_error_err(sb, -retval,
	203	"error reading MMP data: %d",
	204	retval);
61bb4a1c	205	goto wait_to_exit;
c5e06d10 JL	206	}
	207
	208	mmp_check = (struct mmp_struct *)(bh_check->b_data);
	209	if (mmp->mmp_seq != mmp_check->mmp_seq \|\|
	210	memcmp(mmp->mmp_nodename, mmp_check->mmp_nodename,
	211	sizeof(mmp->mmp_nodename))) {
	212	dump_mmp_msg(sb, mmp_check,
	213	"Error while updating MMP info. "
	214	"The filesystem seems to have been"
	215	" multiply mounted.");
54d3adbc	216	ext4_error_err(sb, EBUSY, "abort");
03046886	217	put_bh(bh_check);
03046886	218	retval = -EBUSY;
61bb4a1c	219	goto wait_to_exit;
c5e06d10 JL	220	}
	221	put_bh(bh_check);
	222	}
	223
	224	/*
	225	* Adjust the mmp_check_interval depending on how much time
	226	* it took for the MMP block to be written.
	227	*/
	228	mmp_check_interval = max(min(EXT4_MMP_CHECK_MULT * diff / HZ,
	229	EXT4_MMP_MAX_CHECK_INTERVAL),
	230	EXT4_MMP_MIN_CHECK_INTERVAL);
	231	mmp->mmp_check_interval = cpu_to_le16(mmp_check_interval);
	232	}
	233
	234	/*
	235	* Unmount seems to be clean.
	236	*/
	237	mmp->mmp_seq = cpu_to_le32(EXT4_MMP_SEQ_CLEAN);
af123b37	238	mmp->mmp_time = cpu_to_le64(ktime_get_real_seconds());
c5e06d10	239
5c359a47	240	retval = write_mmp_block(sb, bh);
c5e06d10	241
61bb4a1c TT	242	wait_to_exit:
	243	while (!kthread_should_stop()) {
	244	set_current_state(TASK_INTERRUPTIBLE);
	245	if (!kthread_should_stop())
	246	schedule();
	247	}
	248	set_current_state(TASK_RUNNING);
c5e06d10 JL	249	return retval;
	250	}
	251
618f0031 PS	252	void ext4_stop_mmpd(struct ext4_sb_info *sbi)
	253	{
	254	if (sbi->s_mmp_tsk) {
	255	kthread_stop(sbi->s_mmp_tsk);
	256	brelse(sbi->s_mmp_bh);
	257	sbi->s_mmp_tsk = NULL;
	258	}
	259	}
	260
c5e06d10 JL	261	/*
	262	* Get a random new sequence number but make sure it is not greater than
	263	* EXT4_MMP_SEQ_MAX.
	264	*/
	265	static unsigned int mmp_new_seq(void)
	266	{
	267	u32 new_seq;
	268
	269	do {
dd1f723b	270	new_seq = prandom_u32();
c5e06d10 JL	271	} while (new_seq > EXT4_MMP_SEQ_MAX);
	272
	273	return new_seq;
	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");
	293	goto failed;
	294	}
	295
	296	retval = read_mmp_block(sb, &bh, mmp_block);
	297	if (retval)
	298	goto failed;
	299
	300	mmp = (struct mmp_struct *)(bh->b_data);
	301
	302	if (mmp_check_interval < EXT4_MMP_MIN_CHECK_INTERVAL)
	303	mmp_check_interval = EXT4_MMP_MIN_CHECK_INTERVAL;
	304
	305	/*
	306	* If check_interval in MMP block is larger, use that instead of
	307	* update_interval from the superblock.
	308	*/
85d21650 SN	309	if (le16_to_cpu(mmp->mmp_check_interval) > mmp_check_interval)
85d21650 SN	310	mmp_check_interval = le16_to_cpu(mmp->mmp_check_interval);
c5e06d10 JL	311
	312	seq = le32_to_cpu(mmp->mmp_seq);
	313	if (seq == EXT4_MMP_SEQ_CLEAN)
	314	goto skip;
	315
	316	if (seq == EXT4_MMP_SEQ_FSCK) {
	317	dump_mmp_msg(sb, mmp, "fsck is running on the filesystem");
	318	goto failed;
	319	}
	320
	321	wait_time = min(mmp_check_interval * 2 + 1,
	322	mmp_check_interval + 60);
	323
	324	/* Print MMP interval if more than 20 secs. */
	325	if (wait_time > EXT4_MMP_MIN_CHECK_INTERVAL * 4)
	326	ext4_warning(sb, "MMP interval %u higher than expected, please"
	327	" wait.\n", wait_time * 2);
	328
	329	if (schedule_timeout_interruptible(HZ * wait_time) != 0) {
	330	ext4_warning(sb, "MMP startup interrupted, failing mount\n");
	331	goto failed;
	332	}
	333
	334	retval = read_mmp_block(sb, &bh, mmp_block);
	335	if (retval)
	336	goto failed;
	337	mmp = (struct mmp_struct *)(bh->b_data);
	338	if (seq != le32_to_cpu(mmp->mmp_seq)) {
	339	dump_mmp_msg(sb, mmp,
	340	"Device is already active on another node.");
	341	goto failed;
	342	}
	343
	344	skip:
	345	/*
	346	* write a new random sequence number.
	347	*/
f6f96fdb DW	348	seq = mmp_new_seq();
f6f96fdb DW	349	mmp->mmp_seq = cpu_to_le32(seq);
c5e06d10	350
5c359a47	351	retval = write_mmp_block(sb, bh);
c5e06d10 JL	352	if (retval)
	353	goto failed;
	354
	355	/*
	356	* wait for MMP interval and check mmp_seq.
	357	*/
	358	if (schedule_timeout_interruptible(HZ * wait_time) != 0) {
8d2ae1cb	359	ext4_warning(sb, "MMP startup interrupted, failing mount");
c5e06d10 JL	360	goto failed;
	361	}
	362
	363	retval = read_mmp_block(sb, &bh, mmp_block);
	364	if (retval)
	365	goto failed;
	366	mmp = (struct mmp_struct *)(bh->b_data);
	367	if (seq != le32_to_cpu(mmp->mmp_seq)) {
	368	dump_mmp_msg(sb, mmp,
	369	"Device is already active on another node.");
	370	goto failed;
	371	}
	372
618f0031	373	EXT4_SB(sb)->s_mmp_bh = bh;
c5e06d10 JL	374
	375	/*
	376	* Start a kernel thread to update the MMP block periodically.
	377	*/
618f0031	378	EXT4_SB(sb)->s_mmp_tsk = kthread_run(kmmpd, sb, "kmmpd-%.*s",
14c9ca05	379	(int)sizeof(mmp->mmp_bdevname),
c5e06d10 JL	380	bdevname(bh->b_bdev,
	381	mmp->mmp_bdevname));
	382	if (IS_ERR(EXT4_SB(sb)->s_mmp_tsk)) {
	383	EXT4_SB(sb)->s_mmp_tsk = NULL;
c5e06d10 JL	384	ext4_warning(sb, "Unable to create kmmpd thread for %s.",
	385	sb->s_id);
	386	goto failed;
	387	}
	388
	389	return 0;
	390
	391	failed:
	392	brelse(bh);
	393	return 1;
	394	}