diff --git a/module/zfs/metaslab.c b/module/zfs/metaslab.c index 536611a06a6c..c5c7fe51f786 100644 --- a/module/zfs/metaslab.c +++ b/module/zfs/metaslab.c @@ -289,8 +289,8 @@ int zfs_metaslab_mem_limit = 25; */ unsigned long zfs_metaslab_max_size_cache_sec = 3600; /* 1 hour */ -static uint64_t metaslab_weight(metaslab_t *); -static void metaslab_set_fragmentation(metaslab_t *); +static uint64_t metaslab_weight(metaslab_t *, boolean_t); +static void metaslab_set_fragmentation(metaslab_t *, boolean_t); static void metaslab_free_impl(vdev_t *, uint64_t, uint64_t, boolean_t); static void metaslab_check_free_impl(vdev_t *, uint64_t, uint64_t); @@ -1855,13 +1855,19 @@ metaslab_verify_weight_and_frag(metaslab_t *msp) msp->ms_fragmentation = 0; /* - * This function is used for verification purposes. Regardless of - * whether metaslab_weight() thinks this metaslab should be active or - * not, we want to ensure that the actual weight (and therefore the - * value of ms_weight) would be the same if it was to be recalculated - * at this point. + * This function is used for verification purposes and thus should + * not introduce any side-effects/mutations on the system's state. + * + * Regardless of whether metaslab_weight() thinks this metaslab + * should be active or not, we want to ensure that the actual weight + * (and therefore the value of ms_weight) would be the same if it + * was to be recalculated at this point. + * + * In addition we set the nodirty flag so metaslab_weight() does + * not dirty the metaslab for future TXGs (e.g. when trying to + * force condensing to upgrade the metaslab spacemaps). */ - msp->ms_weight = metaslab_weight(msp) | was_active; + msp->ms_weight = metaslab_weight(msp, B_TRUE) | was_active; VERIFY3U(max_segsize, ==, msp->ms_max_size); @@ -2345,7 +2351,7 @@ metaslab_init(metaslab_group_t *mg, uint64_t id, uint64_t object, ms->ms_trim = range_tree_create(NULL, NULL); metaslab_group_add(mg, ms); - metaslab_set_fragmentation(ms); + metaslab_set_fragmentation(ms, B_FALSE); /* * If we're opening an existing pool (txg == 0) or creating @@ -2520,7 +2526,7 @@ int zfs_frag_table[FRAGMENTATION_TABLE_SIZE] = { * value should be in the range [0, 100]. */ static void -metaslab_set_fragmentation(metaslab_t *msp) +metaslab_set_fragmentation(metaslab_t *msp, boolean_t nodirty) { spa_t *spa = msp->ms_group->mg_vd->vdev_spa; uint64_t fragmentation = 0; @@ -2555,9 +2561,11 @@ metaslab_set_fragmentation(metaslab_t *msp) * be shutting down the pool. We don't want to dirty * any data past this point so skip setting the condense * flag. We can retry this action the next time the pool - * is imported. + * is imported. We also skip marking this metaslab for + * condensing if the caller has explicitly set nodirty. */ - if (spa_writeable(spa) && txg < spa_final_dirty_txg(spa)) { + if (!nodirty && + spa_writeable(spa) && txg < spa_final_dirty_txg(spa)) { msp->ms_condense_wanted = B_TRUE; vdev_dirty(vd, VDD_METASLAB, msp, txg + 1); zfs_dbgmsg("txg %llu, requesting force condense: " @@ -2854,8 +2862,9 @@ metaslab_should_allocate(metaslab_t *msp, uint64_t asize, boolean_t try_hard) return (should_allocate); } + static uint64_t -metaslab_weight(metaslab_t *msp) +metaslab_weight(metaslab_t *msp, boolean_t nodirty) { vdev_t *vd = msp->ms_group->mg_vd; spa_t *spa = vd->vdev_spa; @@ -2863,7 +2872,7 @@ metaslab_weight(metaslab_t *msp) ASSERT(MUTEX_HELD(&msp->ms_lock)); - metaslab_set_fragmentation(msp); + metaslab_set_fragmentation(msp, nodirty); /* * Update the maximum size. If the metaslab is loaded, this will @@ -2904,7 +2913,7 @@ metaslab_recalculate_weight_and_sort(metaslab_t *msp) /* note: we preserve the mask (e.g. indication of primary, etc..) */ uint64_t was_active = msp->ms_weight & METASLAB_ACTIVE_MASK; metaslab_group_sort(msp->ms_group, msp, - metaslab_weight(msp) | was_active); + metaslab_weight(msp, B_FALSE) | was_active); } static int