Merge remote-tracking branch 'torvalds/master' into perf/core

To pick up fixes.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

1 files changed, 111 insertions, 48 deletions

diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index 485cda3eb8d7..088641ba7a8e 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -165,6 +165,10 @@ struct scrub_ctx {
 	int			readonly;
 	int			pages_per_rd_bio;
 
+	/* State of IO submission throttling affecting the associated device */
+	ktime_t			throttle_deadline;
+	u64			throttle_sent;
+
 	int			is_dev_replace;
 	u64			write_pointer;
 
@@ -605,6 +609,7 @@ static noinline_for_stack struct scrub_ctx *scrub_setup_ctx(
 	spin_lock_init(&sctx->list_lock);
 	spin_lock_init(&sctx->stat_lock);
 	init_waitqueue_head(&sctx->list_wait);
+	sctx->throttle_deadline = 0;
 
 	WARN_ON(sctx->wr_curr_bio != NULL);
 	mutex_init(&sctx->wr_lock);
@@ -626,7 +631,6 @@ nomem:
 static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root,
 				     void *warn_ctx)
 {
-	u64 isize;
 	u32 nlink;
 	int ret;
 	int i;
@@ -662,7 +666,6 @@ static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root,
 	eb = swarn->path->nodes[0];
 	inode_item = btrfs_item_ptr(eb, swarn->path->slots[0],
 					struct btrfs_inode_item);
-	isize = btrfs_inode_size(eb, inode_item);
 	nlink = btrfs_inode_nlink(eb, inode_item);
 	btrfs_release_path(swarn->path);
 
@@ -691,12 +694,12 @@ static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root,
 	 */
 	for (i = 0; i < ipath->fspath->elem_cnt; ++i)
 		btrfs_warn_in_rcu(fs_info,
-"%s at logical %llu on dev %s, physical %llu, root %llu, inode %llu, offset %llu, length %llu, links %u (path: %s)",
+"%s at logical %llu on dev %s, physical %llu, root %llu, inode %llu, offset %llu, length %u, links %u (path: %s)",
 				  swarn->errstr, swarn->logical,
 				  rcu_str_deref(swarn->dev->name),
 				  swarn->physical,
 				  root, inum, offset,
-				  min(isize - offset, (u64)PAGE_SIZE), nlink,
+				  fs_info->sectorsize, nlink,
 				  (char *)(unsigned long)ipath->fspath->val[i]);
 
 	btrfs_put_root(local_root);
@@ -885,25 +888,25 @@ static int scrub_handle_errored_block(struct scrub_block *sblock_to_check)
 	 * read all mirrors one after the other. This includes to
 	 * re-read the extent or metadata block that failed (that was
 	 * the cause that this fixup code is called) another time,
-	 * page by page this time in order to know which pages
+	 * sector by sector this time in order to know which sectors
 	 * caused I/O errors and which ones are good (for all mirrors).
 	 * It is the goal to handle the situation when more than one
 	 * mirror contains I/O errors, but the errors do not
 	 * overlap, i.e. the data can be repaired by selecting the
-	 * pages from those mirrors without I/O error on the
-	 * particular pages. One example (with blocks >= 2 * PAGE_SIZE)
-	 * would be that mirror #1 has an I/O error on the first page,
-	 * the second page is good, and mirror #2 has an I/O error on
-	 * the second page, but the first page is good.
-	 * Then the first page of the first mirror can be repaired by
-	 * taking the first page of the second mirror, and the
-	 * second page of the second mirror can be repaired by
-	 * copying the contents of the 2nd page of the 1st mirror.
-	 * One more note: if the pages of one mirror contain I/O
+	 * sectors from those mirrors without I/O error on the
+	 * particular sectors. One example (with blocks >= 2 * sectorsize)
+	 * would be that mirror #1 has an I/O error on the first sector,
+	 * the second sector is good, and mirror #2 has an I/O error on
+	 * the second sector, but the first sector is good.
+	 * Then the first sector of the first mirror can be repaired by
+	 * taking the first sector of the second mirror, and the
+	 * second sector of the second mirror can be repaired by
+	 * copying the contents of the 2nd sector of the 1st mirror.
+	 * One more note: if the sectors of one mirror contain I/O
 	 * errors, the checksum cannot be verified. In order to get
 	 * the best data for repairing, the first attempt is to find
 	 * a mirror without I/O errors and with a validated checksum.
-	 * Only if this is not possible, the pages are picked from
+	 * Only if this is not possible, the sectors are picked from
 	 * mirrors with I/O errors without considering the checksum.
 	 * If the latter is the case, at the end, the checksum of the
 	 * repaired area is verified in order to correctly maintain
@@ -1060,26 +1063,26 @@ static int scrub_handle_errored_block(struct scrub_block *sblock_to_check)
 
 	/*
 	 * In case of I/O errors in the area that is supposed to be
-	 * repaired, continue by picking good copies of those pages.
-	 * Select the good pages from mirrors to rewrite bad pages from
+	 * repaired, continue by picking good copies of those sectors.
+	 * Select the good sectors from mirrors to rewrite bad sectors from
 	 * the area to fix. Afterwards verify the checksum of the block
 	 * that is supposed to be repaired. This verification step is
 	 * only done for the purpose of statistic counting and for the
 	 * final scrub report, whether errors remain.
 	 * A perfect algorithm could make use of the checksum and try
-	 * all possible combinations of pages from the different mirrors
+	 * all possible combinations of sectors from the different mirrors
 	 * until the checksum verification succeeds. For example, when
-	 * the 2nd page of mirror #1 faces I/O errors, and the 2nd page
+	 * the 2nd sector of mirror #1 faces I/O errors, and the 2nd sector
 	 * of mirror #2 is readable but the final checksum test fails,
-	 * then the 2nd page of mirror #3 could be tried, whether now
+	 * then the 2nd sector of mirror #3 could be tried, whether now
 	 * the final checksum succeeds. But this would be a rare
 	 * exception and is therefore not implemented. At least it is
 	 * avoided that the good copy is overwritten.
 	 * A more useful improvement would be to pick the sectors
 	 * without I/O error based on sector sizes (512 bytes on legacy
-	 * disks) instead of on PAGE_SIZE. Then maybe 512 byte of one
+	 * disks) instead of on sectorsize. Then maybe 512 byte of one
 	 * mirror could be repaired by taking 512 byte of a different
-	 * mirror, even if other 512 byte sectors in the same PAGE_SIZE
+	 * mirror, even if other 512 byte sectors in the same sectorsize
 	 * area are unreadable.
 	 */
 	success = 1;
@@ -1260,7 +1263,7 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
 {
 	struct scrub_ctx *sctx = original_sblock->sctx;
 	struct btrfs_fs_info *fs_info = sctx->fs_info;
-	u64 length = original_sblock->page_count * PAGE_SIZE;
+	u64 length = original_sblock->page_count * fs_info->sectorsize;
 	u64 logical = original_sblock->pagev[0]->logical;
 	u64 generation = original_sblock->pagev[0]->generation;
 	u64 flags = original_sblock->pagev[0]->flags;
@@ -1283,13 +1286,13 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
 	 */
 
 	while (length > 0) {
-		sublen = min_t(u64, length, PAGE_SIZE);
+		sublen = min_t(u64, length, fs_info->sectorsize);
 		mapped_length = sublen;
 		bbio = NULL;
 
 		/*
-		 * with a length of PAGE_SIZE, each returned stripe
-		 * represents one mirror
+		 * With a length of sectorsize, each returned stripe represents
+		 * one mirror
 		 */
 		btrfs_bio_counter_inc_blocked(fs_info);
 		ret = btrfs_map_sblock(fs_info, BTRFS_MAP_GET_READ_MIRRORS,
@@ -1480,7 +1483,7 @@ static void scrub_recheck_block(struct btrfs_fs_info *fs_info,
 		bio = btrfs_io_bio_alloc(1);
 		bio_set_dev(bio, spage->dev->bdev);
 
-		bio_add_page(bio, spage->page, PAGE_SIZE, 0);
+		bio_add_page(bio, spage->page, fs_info->sectorsize, 0);
 		bio->bi_iter.bi_sector = spage->physical >> 9;
 		bio->bi_opf = REQ_OP_READ;
 
@@ -1544,6 +1547,7 @@ static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad,
 	struct scrub_page *spage_bad = sblock_bad->pagev[page_num];
 	struct scrub_page *spage_good = sblock_good->pagev[page_num];
 	struct btrfs_fs_info *fs_info = sblock_bad->sctx->fs_info;
+	const u32 sectorsize = fs_info->sectorsize;
 
 	BUG_ON(spage_bad->page == NULL);
 	BUG_ON(spage_good->page == NULL);
@@ -1563,8 +1567,8 @@ static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad,
 		bio->bi_iter.bi_sector = spage_bad->physical >> 9;
 		bio->bi_opf = REQ_OP_WRITE;
 
-		ret = bio_add_page(bio, spage_good->page, PAGE_SIZE, 0);
-		if (PAGE_SIZE != ret) {
+		ret = bio_add_page(bio, spage_good->page, sectorsize, 0);
+		if (ret != sectorsize) {
 			bio_put(bio);
 			return -EIO;
 		}
@@ -1642,6 +1646,7 @@ static int scrub_add_page_to_wr_bio(struct scrub_ctx *sctx,
 {
 	struct scrub_bio *sbio;
 	int ret;
+	const u32 sectorsize = sctx->fs_info->sectorsize;
 
 	mutex_lock(&sctx->wr_lock);
 again:
@@ -1681,16 +1686,16 @@ again:
 		bio->bi_iter.bi_sector = sbio->physical >> 9;
 		bio->bi_opf = REQ_OP_WRITE;
 		sbio->status = 0;
-	} else if (sbio->physical + sbio->page_count * PAGE_SIZE !=
+	} else if (sbio->physical + sbio->page_count * sectorsize !=
 		   spage->physical_for_dev_replace ||
-		   sbio->logical + sbio->page_count * PAGE_SIZE !=
+		   sbio->logical + sbio->page_count * sectorsize !=
 		   spage->logical) {
 		scrub_wr_submit(sctx);
 		goto again;
 	}
 
-	ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0);
-	if (ret != PAGE_SIZE) {
+	ret = bio_add_page(sbio->bio, spage->page, sectorsize, 0);
+	if (ret != sectorsize) {
 		if (sbio->page_count < 1) {
 			bio_put(sbio->bio);
 			sbio->bio = NULL;
@@ -1729,7 +1734,8 @@ static void scrub_wr_submit(struct scrub_ctx *sctx)
 	btrfsic_submit_bio(sbio->bio);
 
 	if (btrfs_is_zoned(sctx->fs_info))
-		sctx->write_pointer = sbio->physical + sbio->page_count * PAGE_SIZE;
+		sctx->write_pointer = sbio->physical + sbio->page_count *
+			sctx->fs_info->sectorsize;
 }
 
 static void scrub_wr_bio_end_io(struct bio *bio)
@@ -1988,6 +1994,65 @@ static void scrub_page_put(struct scrub_page *spage)
 	}
 }
 
+/*
+ * Throttling of IO submission, bandwidth-limit based, the timeslice is 1
+ * second.  Limit can be set via /sys/fs/UUID/devinfo/devid/scrub_speed_max.
+ */
+static void scrub_throttle(struct scrub_ctx *sctx)
+{
+	const int time_slice = 1000;
+	struct scrub_bio *sbio;
+	struct btrfs_device *device;
+	s64 delta;
+	ktime_t now;
+	u32 div;
+	u64 bwlimit;
+
+	sbio = sctx->bios[sctx->curr];
+	device = sbio->dev;
+	bwlimit = READ_ONCE(device->scrub_speed_max);
+	if (bwlimit == 0)
+		return;
+
+	/*
+	 * Slice is divided into intervals when the IO is submitted, adjust by
+	 * bwlimit and maximum of 64 intervals.
+	 */
+	div = max_t(u32, 1, (u32)(bwlimit / (16 * 1024 * 1024)));
+	div = min_t(u32, 64, div);
+
+	/* Start new epoch, set deadline */
+	now = ktime_get();
+	if (sctx->throttle_deadline == 0) {
+		sctx->throttle_deadline = ktime_add_ms(now, time_slice / div);
+		sctx->throttle_sent = 0;
+	}
+
+	/* Still in the time to send? */
+	if (ktime_before(now, sctx->throttle_deadline)) {
+		/* If current bio is within the limit, send it */
+		sctx->throttle_sent += sbio->bio->bi_iter.bi_size;
+		if (sctx->throttle_sent <= div_u64(bwlimit, div))
+			return;
+
+		/* We're over the limit, sleep until the rest of the slice */
+		delta = ktime_ms_delta(sctx->throttle_deadline, now);
+	} else {
+		/* New request after deadline, start new epoch */
+		delta = 0;
+	}
+
+	if (delta) {
+		long timeout;
+
+		timeout = div_u64(delta * HZ, 1000);
+		schedule_timeout_interruptible(timeout);
+	}
+
+	/* Next call will start the deadline period */
+	sctx->throttle_deadline = 0;
+}
+
 static void scrub_submit(struct scrub_ctx *sctx)
 {
 	struct scrub_bio *sbio;
@@ -1995,6 +2060,8 @@ static void scrub_submit(struct scrub_ctx *sctx)
 	if (sctx->curr == -1)
 		return;
 
+	scrub_throttle(sctx);
+
 	sbio = sctx->bios[sctx->curr];
 	sctx->curr = -1;
 	scrub_pending_bio_inc(sctx);
@@ -2006,6 +2073,7 @@ static int scrub_add_page_to_rd_bio(struct scrub_ctx *sctx,
 {
 	struct scrub_block *sblock = spage->sblock;
 	struct scrub_bio *sbio;
+	const u32 sectorsize = sctx->fs_info->sectorsize;
 	int ret;
 
 again:
@@ -2044,9 +2112,9 @@ again:
 		bio->bi_iter.bi_sector = sbio->physical >> 9;
 		bio->bi_opf = REQ_OP_READ;
 		sbio->status = 0;
-	} else if (sbio->physical + sbio->page_count * PAGE_SIZE !=
+	} else if (sbio->physical + sbio->page_count * sectorsize !=
 		   spage->physical ||
-		   sbio->logical + sbio->page_count * PAGE_SIZE !=
+		   sbio->logical + sbio->page_count * sectorsize !=
 		   spage->logical ||
 		   sbio->dev != spage->dev) {
 		scrub_submit(sctx);
@@ -2054,8 +2122,8 @@ again:
 	}
 
 	sbio->pagev[sbio->page_count] = spage;
-	ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0);
-	if (ret != PAGE_SIZE) {
+	ret = bio_add_page(sbio->bio, spage->page, sectorsize, 0);
+	if (ret != sectorsize) {
 		if (sbio->page_count < 1) {
 			bio_put(sbio->bio);
 			sbio->bio = NULL;
@@ -2398,7 +2466,7 @@ static void scrub_block_complete(struct scrub_block *sblock)
 	if (sblock->sparity && corrupted && !sblock->data_corrected) {
 		u64 start = sblock->pagev[0]->logical;
 		u64 end = sblock->pagev[sblock->page_count - 1]->logical +
-			  PAGE_SIZE;
+			  sblock->sctx->fs_info->sectorsize;
 
 		ASSERT(end - start <= U32_MAX);
 		scrub_parity_mark_sectors_error(sblock->sparity,
@@ -2418,7 +2486,7 @@ static void drop_csum_range(struct scrub_ctx *sctx, struct btrfs_ordered_sum *su
  * the csum into @csum.
  *
  * The search source is sctx->csum_list, which is a pre-populated list
- * storing bytenr ordered csum ranges.  We're reponsible to cleanup any range
+ * storing bytenr ordered csum ranges.  We're responsible to cleanup any range
  * that is before @logical.
  *
  * Return 0 if there is no csum for the range.
@@ -3138,28 +3206,23 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 	physical = map->stripes[num].physical;
 	offset = 0;
 	nstripes = div64_u64(length, map->stripe_len);
+	mirror_num = 1;
+	increment = map->stripe_len;
 	if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
 		offset = map->stripe_len * num;
 		increment = map->stripe_len * map->num_stripes;
-		mirror_num = 1;
 	} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
 		int factor = map->num_stripes / map->sub_stripes;
 		offset = map->stripe_len * (num / map->sub_stripes);
 		increment = map->stripe_len * factor;
 		mirror_num = num % map->sub_stripes + 1;
 	} else if (map->type & BTRFS_BLOCK_GROUP_RAID1_MASK) {
-		increment = map->stripe_len;
 		mirror_num = num % map->num_stripes + 1;
 	} else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
-		increment = map->stripe_len;
 		mirror_num = num % map->num_stripes + 1;
 	} else if (map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) {
 		get_raid56_logic_offset(physical, num, map, &offset, NULL);
 		increment = map->stripe_len * nr_data_stripes(map);
-		mirror_num = 1;
-	} else {
-		increment = map->stripe_len;
-		mirror_num = 1;
 	}
 
 	path = btrfs_alloc_path();