1 files changed, 247 insertions, 22 deletions

diff --git a/fs/btrfs/space-info.c b/fs/btrfs/space-info.c
index d620323d08ea..9ac94d3119e8 100644
--- a/fs/btrfs/space-info.c
+++ b/fs/btrfs/space-info.c
@@ -1,5 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 
+#include "linux/spinlock.h"
+#include <linux/minmax.h>
 #include "misc.h"
 #include "ctree.h"
 #include "space-info.h"
@@ -190,6 +192,8 @@ void btrfs_clear_space_info_full(struct btrfs_fs_info *info)
  */
 #define BTRFS_DEFAULT_ZONED_RECLAIM_THRESH			(75)
 
+#define BTRFS_UNALLOC_BLOCK_GROUP_TARGET			(10ULL)
+
 /*
  * Calculate chunk size depending on volume type (regular or zoned).
  */
@@ -232,6 +236,7 @@ static int create_space_info(struct btrfs_fs_info *info, u64 flags)
 	if (!space_info)
 		return -ENOMEM;
 
+	space_info->fs_info = info;
 	for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
 		INIT_LIST_HEAD(&space_info->block_groups[i]);
 	init_rwsem(&space_info->groups_sem);
@@ -340,11 +345,32 @@ struct btrfs_space_info *btrfs_find_space_info(struct btrfs_fs_info *info,
 	return NULL;
 }
 
+static u64 calc_effective_data_chunk_size(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_space_info *data_sinfo;
+	u64 data_chunk_size;
+
+	/*
+	 * Calculate the data_chunk_size, space_info->chunk_size is the
+	 * "optimal" chunk size based on the fs size.  However when we actually
+	 * allocate the chunk we will strip this down further, making it no
+	 * more than 10% of the disk or 1G, whichever is smaller.
+	 *
+	 * On the zoned mode, we need to use zone_size (= data_sinfo->chunk_size)
+	 * as it is.
+	 */
+	data_sinfo = btrfs_find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA);
+	if (btrfs_is_zoned(fs_info))
+		return data_sinfo->chunk_size;
+	data_chunk_size = min(data_sinfo->chunk_size,
+			      mult_perc(fs_info->fs_devices->total_rw_bytes, 10));
+	return min_t(u64, data_chunk_size, SZ_1G);
+}
+
 static u64 calc_available_free_space(struct btrfs_fs_info *fs_info,
 			  struct btrfs_space_info *space_info,
 			  enum btrfs_reserve_flush_enum flush)
 {
-	struct btrfs_space_info *data_sinfo;
 	u64 profile;
 	u64 avail;
 	u64 data_chunk_size;
@@ -368,16 +394,7 @@ static u64 calc_available_free_space(struct btrfs_fs_info *fs_info,
 	if (avail == 0)
 		return 0;
 
-	/*
-	 * Calculate the data_chunk_size, space_info->chunk_size is the
-	 * "optimal" chunk size based on the fs size.  However when we actually
-	 * allocate the chunk we will strip this down further, making it no more
-	 * than 10% of the disk or 1G, whichever is smaller.
-	 */
-	data_sinfo = btrfs_find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA);
-	data_chunk_size = min(data_sinfo->chunk_size,
-			      mult_perc(fs_info->fs_devices->total_rw_bytes, 10));
-	data_chunk_size = min_t(u64, data_chunk_size, SZ_1G);
+	data_chunk_size = calc_effective_data_chunk_size(fs_info);
 
 	/*
 	 * Since data allocations immediately use block groups as part of the
@@ -405,6 +422,17 @@ static u64 calc_available_free_space(struct btrfs_fs_info *fs_info,
 		avail >>= 3;
 	else
 		avail >>= 1;
+
+	/*
+	 * On the zoned mode, we always allocate one zone as one chunk.
+	 * Returning non-zone size alingned bytes here will result in
+	 * less pressure for the async metadata reclaim process, and it
+	 * will over-commit too much leading to ENOSPC. Align down to the
+	 * zone size to avoid that.
+	 */
+	if (btrfs_is_zoned(fs_info))
+		avail = ALIGN_DOWN(avail, fs_info->zone_size);
+
 	return avail;
 }
 
@@ -587,8 +615,6 @@ static inline u64 calc_reclaim_items_nr(const struct btrfs_fs_info *fs_info,
 	return nr;
 }
 
-#define EXTENT_SIZE_PER_ITEM	SZ_256K
-
 /*
  * shrink metadata reservation for delalloc
  */
@@ -688,7 +714,7 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info,
 skip_async:
 		loops++;
 		if (wait_ordered && !trans) {
-			btrfs_wait_ordered_roots(fs_info, items, 0, (u64)-1);
+			btrfs_wait_ordered_roots(fs_info, items, NULL);
 		} else {
 			time_left = schedule_timeout_killable(1);
 			if (time_left)
@@ -807,14 +833,7 @@ static void flush_space(struct btrfs_fs_info *fs_info,
 		 * because that does not wait for a transaction to fully commit
 		 * (only for it to be unblocked, state TRANS_STATE_UNBLOCKED).
 		 */
-		trans = btrfs_attach_transaction_barrier(root);
-		if (IS_ERR(trans)) {
-			ret = PTR_ERR(trans);
-			if (ret == -ENOENT)
-				ret = 0;
-			break;
-		}
-		ret = btrfs_commit_transaction(trans);
+		ret = btrfs_commit_current_transaction(root);
 		break;
 	default:
 		ret = -ENOSPC;
@@ -1868,3 +1887,209 @@ u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo)
 
 	return free_bytes;
 }
+
+static u64 calc_pct_ratio(u64 x, u64 y)
+{
+	int err;
+
+	if (!y)
+		return 0;
+again:
+	err = check_mul_overflow(100, x, &x);
+	if (err)
+		goto lose_precision;
+	return div64_u64(x, y);
+lose_precision:
+	x >>= 10;
+	y >>= 10;
+	if (!y)
+		y = 1;
+	goto again;
+}
+
+/*
+ * A reasonable buffer for unallocated space is 10 data block_groups.
+ * If we claw this back repeatedly, we can still achieve efficient
+ * utilization when near full, and not do too much reclaim while
+ * always maintaining a solid buffer for workloads that quickly
+ * allocate and pressure the unallocated space.
+ */
+static u64 calc_unalloc_target(struct btrfs_fs_info *fs_info)
+{
+	u64 chunk_sz = calc_effective_data_chunk_size(fs_info);
+
+	return BTRFS_UNALLOC_BLOCK_GROUP_TARGET * chunk_sz;
+}
+
+/*
+ * The fundamental goal of automatic reclaim is to protect the filesystem's
+ * unallocated space and thus minimize the probability of the filesystem going
+ * read only when a metadata allocation failure causes a transaction abort.
+ *
+ * However, relocations happen into the space_info's unused space, therefore
+ * automatic reclaim must also back off as that space runs low. There is no
+ * value in doing trivial "relocations" of re-writing the same block group
+ * into a fresh one.
+ *
+ * Furthermore, we want to avoid doing too much reclaim even if there are good
+ * candidates. This is because the allocator is pretty good at filling up the
+ * holes with writes. So we want to do just enough reclaim to try and stay
+ * safe from running out of unallocated space but not be wasteful about it.
+ *
+ * Therefore, the dynamic reclaim threshold is calculated as follows:
+ * - calculate a target unallocated amount of 5 block group sized chunks
+ * - ratchet up the intensity of reclaim depending on how far we are from
+ *   that target by using a formula of unalloc / target to set the threshold.
+ *
+ * Typically with 10 block groups as the target, the discrete values this comes
+ * out to are 0, 10, 20, ... , 80, 90, and 99.
+ */
+static int calc_dynamic_reclaim_threshold(struct btrfs_space_info *space_info)
+{
+	struct btrfs_fs_info *fs_info = space_info->fs_info;
+	u64 unalloc = atomic64_read(&fs_info->free_chunk_space);
+	u64 target = calc_unalloc_target(fs_info);
+	u64 alloc = space_info->total_bytes;
+	u64 used = btrfs_space_info_used(space_info, false);
+	u64 unused = alloc - used;
+	u64 want = target > unalloc ? target - unalloc : 0;
+	u64 data_chunk_size = calc_effective_data_chunk_size(fs_info);
+
+	/* If we have no unused space, don't bother, it won't work anyway. */
+	if (unused < data_chunk_size)
+		return 0;
+
+	/* Cast to int is OK because want <= target. */
+	return calc_pct_ratio(want, target);
+}
+
+int btrfs_calc_reclaim_threshold(struct btrfs_space_info *space_info)
+{
+	lockdep_assert_held(&space_info->lock);
+
+	if (READ_ONCE(space_info->dynamic_reclaim))
+		return calc_dynamic_reclaim_threshold(space_info);
+	return READ_ONCE(space_info->bg_reclaim_threshold);
+}
+
+/*
+ * Under "urgent" reclaim, we will reclaim even fresh block groups that have
+ * recently seen successful allocations, as we are desperate to reclaim
+ * whatever we can to avoid ENOSPC in a transaction leading to a readonly fs.
+ */
+static bool is_reclaim_urgent(struct btrfs_space_info *space_info)
+{
+	struct btrfs_fs_info *fs_info = space_info->fs_info;
+	u64 unalloc = atomic64_read(&fs_info->free_chunk_space);
+	u64 data_chunk_size = calc_effective_data_chunk_size(fs_info);
+
+	return unalloc < data_chunk_size;
+}
+
+static int do_reclaim_sweep(struct btrfs_fs_info *fs_info,
+			    struct btrfs_space_info *space_info, int raid)
+{
+	struct btrfs_block_group *bg;
+	int thresh_pct;
+	bool try_again = true;
+	bool urgent;
+
+	spin_lock(&space_info->lock);
+	urgent = is_reclaim_urgent(space_info);
+	thresh_pct = btrfs_calc_reclaim_threshold(space_info);
+	spin_unlock(&space_info->lock);
+
+	down_read(&space_info->groups_sem);
+again:
+	list_for_each_entry(bg, &space_info->block_groups[raid], list) {
+		u64 thresh;
+		bool reclaim = false;
+
+		btrfs_get_block_group(bg);
+		spin_lock(&bg->lock);
+		thresh = mult_perc(bg->length, thresh_pct);
+		if (bg->used < thresh && bg->reclaim_mark) {
+			try_again = false;
+			reclaim = true;
+		}
+		bg->reclaim_mark++;
+		spin_unlock(&bg->lock);
+		if (reclaim)
+			btrfs_mark_bg_to_reclaim(bg);
+		btrfs_put_block_group(bg);
+	}
+
+	/*
+	 * In situations where we are very motivated to reclaim (low unalloc)
+	 * use two passes to make the reclaim mark check best effort.
+	 *
+	 * If we have any staler groups, we don't touch the fresher ones, but if we
+	 * really need a block group, do take a fresh one.
+	 */
+	if (try_again && urgent) {
+		try_again = false;
+		goto again;
+	}
+
+	up_read(&space_info->groups_sem);
+	return 0;
+}
+
+void btrfs_space_info_update_reclaimable(struct btrfs_space_info *space_info, s64 bytes)
+{
+	u64 chunk_sz = calc_effective_data_chunk_size(space_info->fs_info);
+
+	lockdep_assert_held(&space_info->lock);
+	space_info->reclaimable_bytes += bytes;
+
+	if (space_info->reclaimable_bytes >= chunk_sz)
+		btrfs_set_periodic_reclaim_ready(space_info, true);
+}
+
+void btrfs_set_periodic_reclaim_ready(struct btrfs_space_info *space_info, bool ready)
+{
+	lockdep_assert_held(&space_info->lock);
+	if (!READ_ONCE(space_info->periodic_reclaim))
+		return;
+	if (ready != space_info->periodic_reclaim_ready) {
+		space_info->periodic_reclaim_ready = ready;
+		if (!ready)
+			space_info->reclaimable_bytes = 0;
+	}
+}
+
+bool btrfs_should_periodic_reclaim(struct btrfs_space_info *space_info)
+{
+	bool ret;
+
+	if (space_info->flags & BTRFS_BLOCK_GROUP_SYSTEM)
+		return false;
+	if (!READ_ONCE(space_info->periodic_reclaim))
+		return false;
+
+	spin_lock(&space_info->lock);
+	ret = space_info->periodic_reclaim_ready;
+	btrfs_set_periodic_reclaim_ready(space_info, false);
+	spin_unlock(&space_info->lock);
+
+	return ret;
+}
+
+int btrfs_reclaim_sweep(struct btrfs_fs_info *fs_info)
+{
+	int ret;
+	int raid;
+	struct btrfs_space_info *space_info;
+
+	list_for_each_entry(space_info, &fs_info->space_info, list) {
+		if (!btrfs_should_periodic_reclaim(space_info))
+			continue;
+		for (raid = 0; raid < BTRFS_NR_RAID_TYPES; raid++) {
+			ret = do_reclaim_sweep(fs_info, space_info, raid);
+			if (ret)
+				return ret;
+		}
+	}
+
+	return ret;
+}