Fix and disable blocks statistics during scrub

Block statistics calculation during scrub I/O issue in case of sorted
scrub accounted ditto blocks several times.  Embedded blocks on other
side were not accounted at all.  This change moves the accounting from
issue to scan stage, that fixes both problems and also allows to avoid
pool-wide locking and the lock contention it created.

Since this statistics is quite specific and is not even exposed now
anywhere, disable its calculation by default to not waste CPU time.

Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: Alexander Motin <[email protected]>
Sponsored-By: iXsystems, Inc.
Closes openzfs#13579

include/sys/dsl_pool.h

@@ -82,7 +82,6 @@ typedef struct zfs_blkstat {
 
 typedef struct zfs_all_blkstats {
 	zfs_blkstat_t	zab_type[DN_MAX_LEVELS + 1][DMU_OT_TOTAL + 1];
-	kmutex_t	zab_lock;
 } zfs_all_blkstats_t;
 
 

module/zfs/dsl_pool.c

@@ -418,10 +418,8 @@ dsl_pool_close(dsl_pool_t *dp)
 	cv_destroy(&dp->dp_spaceavail_cv);
 	taskq_destroy(dp->dp_unlinked_drain_taskq);
 	taskq_destroy(dp->dp_zrele_taskq);
-	if (dp->dp_blkstats != NULL) {
-		mutex_destroy(&dp->dp_blkstats->zab_lock);
+	if (dp->dp_blkstats != NULL)
 		vmem_free(dp->dp_blkstats, sizeof (zfs_all_blkstats_t));
-	}
 	kmem_free(dp, sizeof (dsl_pool_t));
 }
 

module/zfs/dsl_scan.c

@@ -129,6 +129,7 @@ static void scan_ds_queue_sync(dsl_scan_t *scn, dmu_tx_t *tx);
 static uint64_t dsl_scan_count_data_disks(vdev_t *vd);
 
 extern int zfs_vdev_async_write_active_min_dirty_percent;
+static int zfs_scan_blkstats = 0;
 
 /*
  * By default zfs will check to ensure it is not over the hard memory
@@ -788,13 +789,19 @@ dsl_scan_setup_sync(void *arg, dmu_tx_t *tx)
 
 	/* back to the generic stuff */
 
-	if (dp->dp_blkstats == NULL) {
-		dp->dp_blkstats =
-		    vmem_alloc(sizeof (zfs_all_blkstats_t), KM_SLEEP);
-		mutex_init(&dp->dp_blkstats->zab_lock, NULL,
-		    MUTEX_DEFAULT, NULL);
+	if (zfs_scan_blkstats) {
+		if (dp->dp_blkstats == NULL) {
+			dp->dp_blkstats =
+			    vmem_alloc(sizeof (zfs_all_blkstats_t), KM_SLEEP);
+		}
+		memset(&dp->dp_blkstats->zab_type, 0,
+		    sizeof (dp->dp_blkstats->zab_type));
+	} else {
+		if (dp->dp_blkstats) {
+			vmem_free(dp->dp_blkstats, sizeof (zfs_all_blkstats_t));
+			dp->dp_blkstats = NULL;
+		}
 	}
-	bzero(&dp->dp_blkstats->zab_type, sizeof (dp->dp_blkstats->zab_type));
 
 	if (spa_version(spa) < SPA_VERSION_DSL_SCRUB)
 		ot = DMU_OT_ZAP_OTHER;
@@ -3779,10 +3786,8 @@ dsl_scan_sync(dsl_pool_t *dp, dmu_tx_t *tx)
 }
 
 static void
-count_block(dsl_scan_t *scn, zfs_all_blkstats_t *zab, const blkptr_t *bp)
+count_block_issued(spa_t *spa, const blkptr_t *bp, boolean_t all)
 {
-	int i;
-
 	/*
 	 * Don't count embedded bp's, since we already did the work of
 	 * scanning these when we scanned the containing block.
@@ -3797,28 +3802,21 @@ count_block(dsl_scan_t *scn, zfs_all_blkstats_t *zab, const blkptr_t *bp)
 	 * zio code will only try the first one unless there is an issue.
 	 * Therefore, we should only count the first DVA for these IOs.
 	 */
-	if (scn->scn_is_sorted) {
-		atomic_add_64(&scn->scn_dp->dp_spa->spa_scan_pass_issued,
-		    DVA_GET_ASIZE(&bp->blk_dva[0]));
-	} else {
-		spa_t *spa = scn->scn_dp->dp_spa;
-
-		for (i = 0; i < BP_GET_NDVAS(bp); i++) {
-			atomic_add_64(&spa->spa_scan_pass_issued,
-			    DVA_GET_ASIZE(&bp->blk_dva[i]));
-		}
-	}
+	atomic_add_64(&spa->spa_scan_pass_issued,
+	    all ? BP_GET_ASIZE(bp) : DVA_GET_ASIZE(&bp->blk_dva[0]));
+}
 
+static void
+count_block(zfs_all_blkstats_t *zab, const blkptr_t *bp)
+{
 	/*
 	 * If we resume after a reboot, zab will be NULL; don't record
 	 * incomplete stats in that case.
 	 */
 	if (zab == NULL)
 		return;
 
-	mutex_enter(&zab->zab_lock);
-
-	for (i = 0; i < 4; i++) {
+	for (int i = 0; i < 4; i++) {
 		int l = (i < 2) ? BP_GET_LEVEL(bp) : DN_MAX_LEVELS;
 		int t = (i & 1) ? BP_GET_TYPE(bp) : DMU_OT_TOTAL;
 
@@ -3853,8 +3851,6 @@ count_block(dsl_scan_t *scn, zfs_all_blkstats_t *zab, const blkptr_t *bp)
 			break;
 		}
 	}
-
-	mutex_exit(&zab->zab_lock);
 }
 
 static void
@@ -3952,10 +3948,10 @@ dsl_scan_scrub_cb(dsl_pool_t *dp,
 	boolean_t needs_io = B_FALSE;
 	int zio_flags = ZIO_FLAG_SCAN_THREAD | ZIO_FLAG_RAW | ZIO_FLAG_CANFAIL;
 
-
+	count_block(dp->dp_blkstats, bp);
 	if (phys_birth <= scn->scn_phys.scn_min_txg ||
 	    phys_birth >= scn->scn_phys.scn_max_txg) {
-		count_block(scn, dp->dp_blkstats, bp);
+		count_block_issued(spa, bp, B_TRUE);
 		return (0);
 	}
 
@@ -3996,7 +3992,7 @@ dsl_scan_scrub_cb(dsl_pool_t *dp,
 	if (needs_io && !zfs_no_scrub_io) {
 		dsl_scan_enqueue(dp, bp, zio_flags, zb);
 	} else {
-		count_block(scn, dp->dp_blkstats, bp);
+		count_block_issued(spa, bp, B_TRUE);
 	}
 
 	/* do not relocate this block */
@@ -4070,7 +4066,7 @@ scan_exec_io(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags,
 	}
 
 	ASSERT(pio != NULL);
-	count_block(scn, dp->dp_blkstats, bp);
+	count_block_issued(spa, bp, queue == NULL);
 	zio_nowait(zio_read(pio, spa, bp, data, size, dsl_scan_scrub_done,
 	    queue, ZIO_PRIORITY_SCRUB, zio_flags, zb));
 }
@@ -4355,7 +4351,7 @@ dsl_scan_freed_dva(spa_t *spa, const blkptr_t *bp, int dva_i)
 
 		/* count the block as though we issued it */
 		sio2bp(sio, &tmpbp);
-		count_block(scn, dp->dp_blkstats, &tmpbp);
+		count_block_issued(spa, &tmpbp, B_FALSE);
 
 		sio_free(sio);
 	}
@@ -4447,6 +4443,9 @@ ZFS_MODULE_PARAM(zfs, zfs_, max_async_dedup_frees, ULONG, ZMOD_RW,
 ZFS_MODULE_PARAM(zfs, zfs_, free_bpobj_enabled, INT, ZMOD_RW,
 	"Enable processing of the free_bpobj");
 
+ZFS_MODULE_PARAM(zfs, zfs_, scan_blkstats, INT, ZMOD_RW,
+	"Enable block statistics calculation during scrub");
+
 ZFS_MODULE_PARAM(zfs, zfs_, scan_mem_lim_fact, INT, ZMOD_RW,
 	"Fraction of RAM for scan hard limit");