Jemalloc updated to 4.4.0.

The original jemalloc source tree was modified to:

1. Remove the configure error that prevents nested builds.
2. Insert the Redis private Jemalloc API in order to allow the
Redis fragmentation function to work.

41 files changed, 3343 insertions, 971 deletions

diff --git a/deps/jemalloc/include/jemalloc/internal/arena.h b/deps/jemalloc/include/jemalloc/internal/arena.h
index 12c617979..ce4e6029e 100644
--- a/deps/jemalloc/include/jemalloc/internal/arena.h
+++ b/deps/jemalloc/include/jemalloc/internal/arena.h
@@ -23,14 +23,29 @@
  */
 #define	LG_DIRTY_MULT_DEFAULT	3
 
+typedef enum {
+	purge_mode_ratio = 0,
+	purge_mode_decay = 1,
+
+	purge_mode_limit = 2
+} purge_mode_t;
+#define	PURGE_DEFAULT		purge_mode_ratio
+/* Default decay time in seconds. */
+#define	DECAY_TIME_DEFAULT	10
+/* Number of event ticks between time checks. */
+#define	DECAY_NTICKS_PER_UPDATE	1000
+
 typedef struct arena_runs_dirty_link_s arena_runs_dirty_link_t;
+typedef struct arena_avail_links_s arena_avail_links_t;
 typedef struct arena_run_s arena_run_t;
 typedef struct arena_chunk_map_bits_s arena_chunk_map_bits_t;
 typedef struct arena_chunk_map_misc_s arena_chunk_map_misc_t;
 typedef struct arena_chunk_s arena_chunk_t;
 typedef struct arena_bin_info_s arena_bin_info_t;
+typedef struct arena_decay_s arena_decay_t;
 typedef struct arena_bin_s arena_bin_t;
 typedef struct arena_s arena_t;
+typedef struct arena_tdata_s arena_tdata_t;
 
 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
@@ -140,13 +155,13 @@ struct arena_runs_dirty_link_s {
  */
 struct arena_chunk_map_misc_s {
 	/*
-	 * Linkage for run trees.  There are two disjoint uses:
+	 * Linkage for run heaps.  There are two disjoint uses:
 	 *
-	 * 1) arena_t's runs_avail tree.
+	 * 1) arena_t's runs_avail heaps.
 	 * 2) arena_run_t conceptually uses this linkage for in-use non-full
 	 *    runs, rather than directly embedding linkage.
 	 */
-	rb_node(arena_chunk_map_misc_t)		rb_link;
+	phn(arena_chunk_map_misc_t)		ph_link;
 
 	union {
 		/* Linkage for list of dirty runs. */
@@ -154,16 +169,15 @@ struct arena_chunk_map_misc_s {
 
 		/* Profile counters, used for large object runs. */
 		union {
-			void				*prof_tctx_pun;
-			prof_tctx_t			*prof_tctx;
+			void			*prof_tctx_pun;
+			prof_tctx_t		*prof_tctx;
 		};
 
 		/* Small region run metadata. */
 		arena_run_t			run;
 	};
 };
-typedef rb_tree(arena_chunk_map_misc_t) arena_avail_tree_t;
-typedef rb_tree(arena_chunk_map_misc_t) arena_run_tree_t;
+typedef ph(arena_chunk_map_misc_t) arena_run_heap_t;
 #endif /* JEMALLOC_ARENA_STRUCTS_A */
 
 #ifdef JEMALLOC_ARENA_STRUCTS_B
@@ -177,6 +191,14 @@ struct arena_chunk_s {
 	extent_node_t		node;
 
 	/*
+	 * True if memory could be backed by transparent huge pages.  This is
+	 * only directly relevant to Linux, since it is the only supported
+	 * platform on which jemalloc interacts with explicit transparent huge
+	 * page controls.
+	 */
+	bool			hugepage;
+
+	/*
 	 * Map of pages within chunk that keeps track of free/large/small.  The
 	 * first map_bias entries are omitted, since the chunk header does not
 	 * need to be tracked in the map.  This omission saves a header page
@@ -220,28 +242,71 @@ struct arena_chunk_s {
  */
 struct arena_bin_info_s {
 	/* Size of regions in a run for this bin's size class. */
-	size_t		reg_size;
+	size_t			reg_size;
 
 	/* Redzone size. */
-	size_t		redzone_size;
+	size_t			redzone_size;
 
 	/* Interval between regions (reg_size + (redzone_size << 1)). */
-	size_t		reg_interval;
+	size_t			reg_interval;
 
 	/* Total size of a run for this bin's size class. */
-	size_t		run_size;
+	size_t			run_size;
 
 	/* Total number of regions in a run for this bin's size class. */
-	uint32_t	nregs;
+	uint32_t		nregs;
 
 	/*
 	 * Metadata used to manipulate bitmaps for runs associated with this
 	 * bin.
 	 */
-	bitmap_info_t	bitmap_info;
+	bitmap_info_t		bitmap_info;
 
 	/* Offset of first region in a run for this bin's size class. */
-	uint32_t	reg0_offset;
+	uint32_t		reg0_offset;
+};
+
+struct arena_decay_s {
+	/*
+	 * Approximate time in seconds from the creation of a set of unused
+	 * dirty pages until an equivalent set of unused dirty pages is purged
+	 * and/or reused.
+	 */
+	ssize_t			time;
+	/* time / SMOOTHSTEP_NSTEPS. */
+	nstime_t		interval;
+	/*
+	 * Time at which the current decay interval logically started.  We do
+	 * not actually advance to a new epoch until sometime after it starts
+	 * because of scheduling and computation delays, and it is even possible
+	 * to completely skip epochs.  In all cases, during epoch advancement we
+	 * merge all relevant activity into the most recently recorded epoch.
+	 */
+	nstime_t		epoch;
+	/* Deadline randomness generator. */
+	uint64_t		jitter_state;
+	/*
+	 * Deadline for current epoch.  This is the sum of interval and per
+	 * epoch jitter which is a uniform random variable in [0..interval).
+	 * Epochs always advance by precise multiples of interval, but we
+	 * randomize the deadline to reduce the likelihood of arenas purging in
+	 * lockstep.
+	 */
+	nstime_t		deadline;
+	/*
+	 * Number of dirty pages at beginning of current epoch.  During epoch
+	 * advancement we use the delta between arena->decay.ndirty and
+	 * arena->ndirty to determine how many dirty pages, if any, were
+	 * generated.
+	 */
+	size_t			ndirty;
+	/*
+	 * Trailing log of how many unused dirty pages were generated during
+	 * each of the past SMOOTHSTEP_NSTEPS decay epochs, where the last
+	 * element is the most recent epoch.  Corresponding epoch times are
+	 * relative to epoch.
+	 */
+	size_t			backlog[SMOOTHSTEP_NSTEPS];
 };
 
 struct arena_bin_s {
@@ -251,25 +316,25 @@ struct arena_bin_s {
 	 * which may be acquired while holding one or more bin locks, but not
 	 * vise versa.
 	 */
-	malloc_mutex_t	lock;
+	malloc_mutex_t		lock;
 
 	/*
 	 * Current run being used to service allocations of this bin's size
 	 * class.
 	 */
-	arena_run_t	*runcur;
+	arena_run_t		*runcur;
 
 	/*
-	 * Tree of non-full runs.  This tree is used when looking for an
+	 * Heap of non-full runs.  This heap is used when looking for an
 	 * existing run when runcur is no longer usable.  We choose the
 	 * non-full run that is lowest in memory; this policy tends to keep
 	 * objects packed well, and it can also help reduce the number of
 	 * almost-empty chunks.
 	 */
-	arena_run_tree_t runs;
+	arena_run_heap_t	runs;
 
 	/* Bin statistics. */
-	malloc_bin_stats_t stats;
+	malloc_bin_stats_t	stats;
 };
 
 struct arena_s {
@@ -277,15 +342,23 @@ struct arena_s {
 	unsigned		ind;
 
 	/*
-	 * Number of threads currently assigned to this arena.  This field is
-	 * protected by arenas_lock.
+	 * Number of threads currently assigned to this arena, synchronized via
+	 * atomic operations.  Each thread has two distinct assignments, one for
+	 * application-serving allocation, and the other for internal metadata
+	 * allocation.  Internal metadata must not be allocated from arenas
+	 * created via the arenas.extend mallctl, because the arena.<i>.reset
+	 * mallctl indiscriminately discards all allocations for the affected
+	 * arena.
+	 *
+	 *   0: Application allocation.
+	 *   1: Internal metadata allocation.
 	 */
-	unsigned		nthreads;
+	unsigned		nthreads[2];
 
 	/*
 	 * There are three classes of arena operations from a locking
 	 * perspective:
-	 * 1) Thread assignment (modifies nthreads) is protected by arenas_lock.
+	 * 1) Thread assignment (modifies nthreads) is synchronized via atomics.
 	 * 2) Bin-related operations are protected by bin locks.
 	 * 3) Chunk- and run-related operations are protected by this mutex.
 	 */
@@ -305,10 +378,16 @@ struct arena_s {
 	 * PRNG state for cache index randomization of large allocation base
 	 * pointers.
 	 */
-	uint64_t		offset_state;
+	size_t			offset_state;
 
 	dss_prec_t		dss_prec;
 
+	/* Extant arena chunks. */
+	ql_head(extent_node_t)	achunks;
+
+	/* Extent serial number generator state. */
+	size_t			extent_sn_next;
+
 	/*
 	 * In order to avoid rapid chunk allocation/deallocation when an arena
 	 * oscillates right on the cusp of needing a new chunk, cache the most
@@ -324,7 +403,7 @@ struct arena_s {
 	/* Minimum ratio (log base 2) of nactive:ndirty. */
 	ssize_t			lg_dirty_mult;
 
-	/* True if a thread is currently executing arena_purge(). */
+	/* True if a thread is currently executing arena_purge_to_limit(). */
 	bool			purging;
 
 	/* Number of pages in active runs and huge regions. */
@@ -339,12 +418,6 @@ struct arena_s {
 	size_t			ndirty;
 
 	/*
-	 * Size/address-ordered tree of this arena's available runs.  The tree
-	 * is used for first-best-fit run allocation.
-	 */
-	arena_avail_tree_t	runs_avail;
-
-	/*
 	 * Unused dirty memory this arena manages.  Dirty memory is conceptually
 	 * tracked as an arbitrarily interleaved LRU of dirty runs and cached
 	 * chunks, but the list linkage is actually semi-duplicated in order to
@@ -375,6 +448,9 @@ struct arena_s {
 	arena_runs_dirty_link_t	runs_dirty;
 	extent_node_t		chunks_cache;
 
+	/* Decay-based purging state. */
+	arena_decay_t		decay;
+
 	/* Extant huge allocations. */
 	ql_head(extent_node_t)	huge;
 	/* Synchronizes all huge allocation/update/deallocation. */
@@ -387,9 +463,9 @@ struct arena_s {
 	 * orderings are needed, which is why there are two trees with the same
 	 * contents.
 	 */
-	extent_tree_t		chunks_szad_cached;
+	extent_tree_t		chunks_szsnad_cached;
 	extent_tree_t		chunks_ad_cached;
-	extent_tree_t		chunks_szad_retained;
+	extent_tree_t		chunks_szsnad_retained;
 	extent_tree_t		chunks_ad_retained;
 
 	malloc_mutex_t		chunks_mtx;
@@ -402,6 +478,19 @@ struct arena_s {
 
 	/* bins is used to store trees of free regions. */
 	arena_bin_t		bins[NBINS];
+
+	/*
+	 * Size-segregated address-ordered heaps of this arena's available runs,
+	 * used for first-best-fit run allocation.  Runs are quantized, i.e.
+	 * they reside in the last heap which corresponds to a size class less
+	 * than or equal to the run size.
+	 */
+	arena_run_heap_t	runs_avail[NPSIZES];
+};
+
+/* Used in conjunction with tsd for fast arena-related context lookup. */
+struct arena_tdata_s {
+	ticker_t		decay_ticker;
 };
 #endif /* JEMALLOC_ARENA_STRUCTS_B */
 
@@ -417,7 +506,10 @@ static const size_t	large_pad =
 #endif
     ;
 
+extern purge_mode_t	opt_purge;
+extern const char	*purge_mode_names[];
 extern ssize_t		opt_lg_dirty_mult;
+extern ssize_t		opt_decay_time;
 
 extern arena_bin_info_t	arena_bin_info[NBINS];
 
@@ -428,27 +520,37 @@ extern size_t		large_maxclass; /* Max large size class. */
 extern unsigned		nlclasses; /* Number of large size classes. */
 extern unsigned		nhclasses; /* Number of huge size classes. */
 
+#ifdef JEMALLOC_JET
+typedef size_t (run_quantize_t)(size_t);
+extern run_quantize_t *run_quantize_floor;
+extern run_quantize_t *run_quantize_ceil;
+#endif
 void	arena_chunk_cache_maybe_insert(arena_t *arena, extent_node_t *node,
     bool cache);
 void	arena_chunk_cache_maybe_remove(arena_t *arena, extent_node_t *node,
     bool cache);
-extent_node_t	*arena_node_alloc(arena_t *arena);
-void	arena_node_dalloc(arena_t *arena, extent_node_t *node);
-void	*arena_chunk_alloc_huge(arena_t *arena, size_t usize, size_t alignment,
-    bool *zero);
-void	arena_chunk_dalloc_huge(arena_t *arena, void *chunk, size_t usize);
-void	arena_chunk_ralloc_huge_similar(arena_t *arena, void *chunk,
-    size_t oldsize, size_t usize);
-void	arena_chunk_ralloc_huge_shrink(arena_t *arena, void *chunk,
-    size_t oldsize, size_t usize);
-bool	arena_chunk_ralloc_huge_expand(arena_t *arena, void *chunk,
-    size_t oldsize, size_t usize, bool *zero);
-ssize_t	arena_lg_dirty_mult_get(arena_t *arena);
-bool	arena_lg_dirty_mult_set(arena_t *arena, ssize_t lg_dirty_mult);
-void	arena_maybe_purge(arena_t *arena);
-void	arena_purge_all(arena_t *arena);
-void	arena_tcache_fill_small(arena_t *arena, tcache_bin_t *tbin,
-    szind_t binind, uint64_t prof_accumbytes);
+extent_node_t	*arena_node_alloc(tsdn_t *tsdn, arena_t *arena);
+void	arena_node_dalloc(tsdn_t *tsdn, arena_t *arena, extent_node_t *node);
+void	*arena_chunk_alloc_huge(tsdn_t *tsdn, arena_t *arena, size_t usize,
+    size_t alignment, size_t *sn, bool *zero);
+void	arena_chunk_dalloc_huge(tsdn_t *tsdn, arena_t *arena, void *chunk,
+    size_t usize, size_t sn);
+void	arena_chunk_ralloc_huge_similar(tsdn_t *tsdn, arena_t *arena,
+    void *chunk, size_t oldsize, size_t usize);
+void	arena_chunk_ralloc_huge_shrink(tsdn_t *tsdn, arena_t *arena,
+    void *chunk, size_t oldsize, size_t usize, size_t sn);
+bool	arena_chunk_ralloc_huge_expand(tsdn_t *tsdn, arena_t *arena,
+    void *chunk, size_t oldsize, size_t usize, bool *zero);
+ssize_t	arena_lg_dirty_mult_get(tsdn_t *tsdn, arena_t *arena);
+bool	arena_lg_dirty_mult_set(tsdn_t *tsdn, arena_t *arena,
+    ssize_t lg_dirty_mult);
+ssize_t	arena_decay_time_get(tsdn_t *tsdn, arena_t *arena);
+bool	arena_decay_time_set(tsdn_t *tsdn, arena_t *arena, ssize_t decay_time);
+void	arena_purge(tsdn_t *tsdn, arena_t *arena, bool all);
+void	arena_maybe_purge(tsdn_t *tsdn, arena_t *arena);
+void	arena_reset(tsd_t *tsd, arena_t *arena);
+void	arena_tcache_fill_small(tsdn_t *tsdn, arena_t *arena,
+    tcache_bin_t *tbin, szind_t binind, uint64_t prof_accumbytes);
 void	arena_alloc_junk_small(void *ptr, arena_bin_info_t *bin_info,
     bool zero);
 #ifdef JEMALLOC_JET
@@ -461,75 +563,100 @@ extern arena_dalloc_junk_small_t *arena_dalloc_junk_small;
 void	arena_dalloc_junk_small(void *ptr, arena_bin_info_t *bin_info);
 #endif
 void	arena_quarantine_junk_small(void *ptr, size_t usize);
-void	*arena_malloc_small(arena_t *arena, size_t size, bool zero);
-void	*arena_malloc_large(arena_t *arena, size_t size, bool zero);
-void	*arena_palloc(tsd_t *tsd, arena_t *arena, size_t usize,
+void	*arena_malloc_large(tsdn_t *tsdn, arena_t *arena, szind_t ind,
+    bool zero);
+void	*arena_malloc_hard(tsdn_t *tsdn, arena_t *arena, size_t size,
+    szind_t ind, bool zero);
+void	*arena_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize,
     size_t alignment, bool zero, tcache_t *tcache);
-void	arena_prof_promoted(const void *ptr, size_t size);
-void	arena_dalloc_bin_junked_locked(arena_t *arena, arena_chunk_t *chunk,
-    void *ptr, arena_chunk_map_bits_t *bitselm);
-void	arena_dalloc_bin(arena_t *arena, arena_chunk_t *chunk, void *ptr,
-    size_t pageind, arena_chunk_map_bits_t *bitselm);
-void	arena_dalloc_small(arena_t *arena, arena_chunk_t *chunk, void *ptr,
-    size_t pageind);
+void	arena_prof_promoted(tsdn_t *tsdn, const void *ptr, size_t size);
+void	arena_dalloc_bin_junked_locked(tsdn_t *tsdn, arena_t *arena,
+    arena_chunk_t *chunk, void *ptr, arena_chunk_map_bits_t *bitselm);
+void	arena_dalloc_bin(tsdn_t *tsdn, arena_t *arena, arena_chunk_t *chunk,
+    void *ptr, size_t pageind, arena_chunk_map_bits_t *bitselm);
+void	arena_dalloc_small(tsdn_t *tsdn, arena_t *arena, arena_chunk_t *chunk,
+    void *ptr, size_t pageind);
 #ifdef JEMALLOC_JET
 typedef void (arena_dalloc_junk_large_t)(void *, size_t);
 extern arena_dalloc_junk_large_t *arena_dalloc_junk_large;
 #else
 void	arena_dalloc_junk_large(void *ptr, size_t usize);
 #endif
-void	arena_dalloc_large_junked_locked(arena_t *arena, arena_chunk_t *chunk,
+void	arena_dalloc_large_junked_locked(tsdn_t *tsdn, arena_t *arena,
+    arena_chunk_t *chunk, void *ptr);
+void	arena_dalloc_large(tsdn_t *tsdn, arena_t *arena, arena_chunk_t *chunk,
     void *ptr);
-void	arena_dalloc_large(arena_t *arena, arena_chunk_t *chunk, void *ptr);
 #ifdef JEMALLOC_JET
 typedef void (arena_ralloc_junk_large_t)(void *, size_t, size_t);
 extern arena_ralloc_junk_large_t *arena_ralloc_junk_large;
 #endif
-bool	arena_ralloc_no_move(void *ptr, size_t oldsize, size_t size,
-    size_t extra, bool zero);
+bool	arena_ralloc_no_move(tsdn_t *tsdn, void *ptr, size_t oldsize,
+    size_t size, size_t extra, bool zero);
 void	*arena_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize,
     size_t size, size_t alignment, bool zero, tcache_t *tcache);
-dss_prec_t	arena_dss_prec_get(arena_t *arena);
-bool	arena_dss_prec_set(arena_t *arena, dss_prec_t dss_prec);
+dss_prec_t	arena_dss_prec_get(tsdn_t *tsdn, arena_t *arena);
+bool	arena_dss_prec_set(tsdn_t *tsdn, arena_t *arena, dss_prec_t dss_prec);
 ssize_t	arena_lg_dirty_mult_default_get(void);
 bool	arena_lg_dirty_mult_default_set(ssize_t lg_dirty_mult);
-void	arena_stats_merge(arena_t *arena, const char **dss,
-    ssize_t *lg_dirty_mult, size_t *nactive, size_t *ndirty,
-    arena_stats_t *astats, malloc_bin_stats_t *bstats,
-    malloc_large_stats_t *lstats, malloc_huge_stats_t *hstats);
-arena_t	*arena_new(unsigned ind);
-bool	arena_boot(void);
-void	arena_prefork(arena_t *arena);
-void	arena_postfork_parent(arena_t *arena);
-void	arena_postfork_child(arena_t *arena);
+ssize_t	arena_decay_time_default_get(void);
+bool	arena_decay_time_default_set(ssize_t decay_time);
+void	arena_basic_stats_merge(tsdn_t *tsdn, arena_t *arena,
+    unsigned *nthreads, const char **dss, ssize_t *lg_dirty_mult,
+    ssize_t *decay_time, size_t *nactive, size_t *ndirty);
+void	arena_stats_merge(tsdn_t *tsdn, arena_t *arena, unsigned *nthreads,
+    const char **dss, ssize_t *lg_dirty_mult, ssize_t *decay_time,
+    size_t *nactive, size_t *ndirty, arena_stats_t *astats,
+    malloc_bin_stats_t *bstats, malloc_large_stats_t *lstats,
+    malloc_huge_stats_t *hstats);
+unsigned	arena_nthreads_get(arena_t *arena, bool internal);
+void	arena_nthreads_inc(arena_t *arena, bool internal);
+void	arena_nthreads_dec(arena_t *arena, bool internal);
+size_t	arena_extent_sn_next(arena_t *arena);
+arena_t	*arena_new(tsdn_t *tsdn, unsigned ind);
+void	arena_boot(void);
+void	arena_prefork0(tsdn_t *tsdn, arena_t *arena);
+void	arena_prefork1(tsdn_t *tsdn, arena_t *arena);
+void	arena_prefork2(tsdn_t *tsdn, arena_t *arena);
+void	arena_prefork3(tsdn_t *tsdn, arena_t *arena);
+void	arena_postfork_parent(tsdn_t *tsdn, arena_t *arena);
+void	arena_postfork_child(tsdn_t *tsdn, arena_t *arena);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
 #ifdef JEMALLOC_H_INLINES
 
 #ifndef JEMALLOC_ENABLE_INLINE
-arena_chunk_map_bits_t	*arena_bitselm_get(arena_chunk_t *chunk,
+arena_chunk_map_bits_t	*arena_bitselm_get_mutable(arena_chunk_t *chunk,
     size_t pageind);
-arena_chunk_map_misc_t	*arena_miscelm_get(arena_chunk_t *chunk,
+const arena_chunk_map_bits_t	*arena_bitselm_get_const(
+    const arena_chunk_t *chunk, size_t pageind);
+arena_chunk_map_misc_t	*arena_miscelm_get_mutable(arena_chunk_t *chunk,
     size_t pageind);
-size_t	arena_miscelm_to_pageind(arena_chunk_map_misc_t *miscelm);
-void	*arena_miscelm_to_rpages(arena_chunk_map_misc_t *miscelm);
+const arena_chunk_map_misc_t	*arena_miscelm_get_const(
+    const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_miscelm_to_pageind(const arena_chunk_map_misc_t *miscelm);
+void	*arena_miscelm_to_rpages(const arena_chunk_map_misc_t *miscelm);
 arena_chunk_map_misc_t	*arena_rd_to_miscelm(arena_runs_dirty_link_t *rd);
 arena_chunk_map_misc_t	*arena_run_to_miscelm(arena_run_t *run);
-size_t	*arena_mapbitsp_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbitsp_read(size_t *mapbitsp);
-size_t	arena_mapbits_get(arena_chunk_t *chunk, size_t pageind);
+size_t	*arena_mapbitsp_get_mutable(arena_chunk_t *chunk, size_t pageind);
+const size_t	*arena_mapbitsp_get_const(const arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbitsp_read(const size_t *mapbitsp);
+size_t	arena_mapbits_get(const arena_chunk_t *chunk, size_t pageind);
 size_t	arena_mapbits_size_decode(size_t mapbits);
-size_t	arena_mapbits_unallocated_size_get(arena_chunk_t *chunk,
+size_t	arena_mapbits_unallocated_size_get(const arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbits_large_size_get(const arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbits_small_runind_get(const arena_chunk_t *chunk,
     size_t pageind);
-size_t	arena_mapbits_large_size_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbits_small_runind_get(arena_chunk_t *chunk, size_t pageind);
-szind_t	arena_mapbits_binind_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbits_dirty_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbits_unzeroed_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbits_decommitted_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbits_large_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbits_allocated_get(arena_chunk_t *chunk, size_t pageind);
+szind_t	arena_mapbits_binind_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_dirty_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_unzeroed_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_decommitted_get(const arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbits_large_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_allocated_get(const arena_chunk_t *chunk, size_t pageind);
 void	arena_mapbitsp_write(size_t *mapbitsp, size_t mapbits);
 size_t	arena_mapbits_size_encode(size_t size);
 void	arena_mapbits_unallocated_set(arena_chunk_t *chunk, size_t pageind,
@@ -549,27 +676,31 @@ void	arena_metadata_allocated_sub(arena_t *arena, size_t size);
 size_t	arena_metadata_allocated_get(arena_t *arena);
 bool	arena_prof_accum_impl(arena_t *arena, uint64_t accumbytes);
 bool	arena_prof_accum_locked(arena_t *arena, uint64_t accumbytes);
-bool	arena_prof_accum(arena_t *arena, uint64_t accumbytes);
+bool	arena_prof_accum(tsdn_t *tsdn, arena_t *arena, uint64_t accumbytes);
 szind_t	arena_ptr_small_binind_get(const void *ptr, size_t mapbits);
 szind_t	arena_bin_index(arena_t *arena, arena_bin_t *bin);
-unsigned	arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info,
+size_t	arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info,
     const void *ptr);
-prof_tctx_t	*arena_prof_tctx_get(const void *ptr);
-void	arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx);
-void	arena_prof_tctx_reset(const void *ptr, size_t usize,
+prof_tctx_t	*arena_prof_tctx_get(tsdn_t *tsdn, const void *ptr);
+void	arena_prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    prof_tctx_t *tctx);
+void	arena_prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize,
     const void *old_ptr, prof_tctx_t *old_tctx);
-void	*arena_malloc(tsd_t *tsd, arena_t *arena, size_t size, bool zero,
-    tcache_t *tcache);
+void	arena_decay_ticks(tsdn_t *tsdn, arena_t *arena, unsigned nticks);
+void	arena_decay_tick(tsdn_t *tsdn, arena_t *arena);
+void	*arena_malloc(tsdn_t *tsdn, arena_t *arena, size_t size, szind_t ind,
+    bool zero, tcache_t *tcache, bool slow_path);
 arena_t	*arena_aalloc(const void *ptr);
-size_t	arena_salloc(const void *ptr, bool demote);
-void	arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache);
-void	arena_sdalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache);
+size_t	arena_salloc(tsdn_t *tsdn, const void *ptr, bool demote);
+void	arena_dalloc(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool slow_path);
+void	arena_sdalloc(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path);
 #endif
 
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ARENA_C_))
 #  ifdef JEMALLOC_ARENA_INLINE_A
 JEMALLOC_ALWAYS_INLINE arena_chunk_map_bits_t *
-arena_bitselm_get(arena_chunk_t *chunk, size_t pageind)
+arena_bitselm_get_mutable(arena_chunk_t *chunk, size_t pageind)
 {
 
 	assert(pageind >= map_bias);
@@ -578,8 +709,15 @@ arena_bitselm_get(arena_chunk_t *chunk, size_t pageind)
 	return (&chunk->map_bits[pageind-map_bias]);
 }
 
+JEMALLOC_ALWAYS_INLINE const arena_chunk_map_bits_t *
+arena_bitselm_get_const(const arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (arena_bitselm_get_mutable((arena_chunk_t *)chunk, pageind));
+}
+
 JEMALLOC_ALWAYS_INLINE arena_chunk_map_misc_t *
-arena_miscelm_get(arena_chunk_t *chunk, size_t pageind)
+arena_miscelm_get_mutable(arena_chunk_t *chunk, size_t pageind)
 {
 
 	assert(pageind >= map_bias);
@@ -589,8 +727,15 @@ arena_miscelm_get(arena_chunk_t *chunk, size_t pageind)
 	    (uintptr_t)map_misc_offset) + pageind-map_bias);
 }
 
+JEMALLOC_ALWAYS_INLINE const arena_chunk_map_misc_t *
+arena_miscelm_get_const(const arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (arena_miscelm_get_mutable((arena_chunk_t *)chunk, pageind));
+}
+
 JEMALLOC_ALWAYS_INLINE size_t
-arena_miscelm_to_pageind(arena_chunk_map_misc_t *miscelm)
+arena_miscelm_to_pageind(const arena_chunk_map_misc_t *miscelm)
 {
 	arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(miscelm);
 	size_t pageind = ((uintptr_t)miscelm - ((uintptr_t)chunk +
@@ -603,7 +748,7 @@ arena_miscelm_to_pageind(arena_chunk_map_misc_t *miscelm)
 }
 
 JEMALLOC_ALWAYS_INLINE void *
-arena_miscelm_to_rpages(arena_chunk_map_misc_t *miscelm)
+arena_miscelm_to_rpages(const arena_chunk_map_misc_t *miscelm)
 {
 	arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(miscelm);
 	size_t pageind = arena_miscelm_to_pageind(miscelm);
@@ -636,24 +781,31 @@ arena_run_to_miscelm(arena_run_t *run)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t *
-arena_mapbitsp_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbitsp_get_mutable(arena_chunk_t *chunk, size_t pageind)
 {
 
-	return (&arena_bitselm_get(chunk, pageind)->bits);
+	return (&arena_bitselm_get_mutable(chunk, pageind)->bits);
+}
+
+JEMALLOC_ALWAYS_INLINE const size_t *
+arena_mapbitsp_get_const(const arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (arena_mapbitsp_get_mutable((arena_chunk_t *)chunk, pageind));
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbitsp_read(size_t *mapbitsp)
+arena_mapbitsp_read(const size_t *mapbitsp)
 {
 
 	return (*mapbitsp);
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_get(const arena_chunk_t *chunk, size_t pageind)
 {
 
-	return (arena_mapbitsp_read(arena_mapbitsp_get(chunk, pageind)));
+	return (arena_mapbitsp_read(arena_mapbitsp_get_const(chunk, pageind)));
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
@@ -673,7 +825,7 @@ arena_mapbits_size_decode(size_t mapbits)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_unallocated_size_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_unallocated_size_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -683,7 +835,7 @@ arena_mapbits_unallocated_size_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_large_size_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_large_size_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -694,7 +846,7 @@ arena_mapbits_large_size_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_small_runind_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_small_runind_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -705,7 +857,7 @@ arena_mapbits_small_runind_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE szind_t
-arena_mapbits_binind_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_binind_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 	szind_t binind;
@@ -717,7 +869,7 @@ arena_mapbits_binind_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_dirty_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_dirty_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -728,7 +880,7 @@ arena_mapbits_dirty_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_unzeroed_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_unzeroed_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -739,7 +891,7 @@ arena_mapbits_unzeroed_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_decommitted_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_decommitted_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -750,7 +902,7 @@ arena_mapbits_decommitted_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_large_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_large_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -759,7 +911,7 @@ arena_mapbits_large_get(arena_chunk_t *chunk, size_t pageind)
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_allocated_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_allocated_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 
@@ -795,7 +947,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_unallocated_set(arena_chunk_t *chunk, size_t pageind, size_t size,
     size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 
 	assert((size & PAGE_MASK) == 0);
 	assert((flags & CHUNK_MAP_FLAGS_MASK) == flags);
@@ -809,7 +961,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_unallocated_size_set(arena_chunk_t *chunk, size_t pageind,
     size_t size)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 	size_t mapbits = arena_mapbitsp_read(mapbitsp);
 
 	assert((size & PAGE_MASK) == 0);
@@ -821,7 +973,7 @@ arena_mapbits_unallocated_size_set(arena_chunk_t *chunk, size_t pageind,
 JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_internal_set(arena_chunk_t *chunk, size_t pageind, size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 
 	assert((flags & CHUNK_MAP_UNZEROED) == flags);
 	arena_mapbitsp_write(mapbitsp, flags);
@@ -831,7 +983,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_large_set(arena_chunk_t *chunk, size_t pageind, size_t size,
     size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 
 	assert((size & PAGE_MASK) == 0);
 	assert((flags & CHUNK_MAP_FLAGS_MASK) == flags);
@@ -846,7 +998,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_large_binind_set(arena_chunk_t *chunk, size_t pageind,
     szind_t binind)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 	size_t mapbits = arena_mapbitsp_read(mapbitsp);
 
 	assert(binind <= BININD_INVALID);
@@ -860,7 +1012,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_small_set(arena_chunk_t *chunk, size_t pageind, size_t runind,
     szind_t binind, size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 
 	assert(binind < BININD_INVALID);
 	assert(pageind - runind >= map_bias);
@@ -917,7 +1069,7 @@ arena_prof_accum_locked(arena_t *arena, uint64_t accumbytes)
 }
 
 JEMALLOC_INLINE bool
-arena_prof_accum(arena_t *arena, uint64_t accumbytes)
+arena_prof_accum(tsdn_t *tsdn, arena_t *arena, uint64_t accumbytes)
 {
 
 	cassert(config_prof);
@@ -928,9 +1080,9 @@ arena_prof_accum(arena_t *arena, uint64_t accumbytes)
 	{
 		bool ret;
 
-		malloc_mutex_lock(&arena->lock);
+		malloc_mutex_lock(tsdn, &arena->lock);
 		ret = arena_prof_accum_impl(arena, accumbytes);
-		malloc_mutex_unlock(&arena->lock);
+		malloc_mutex_unlock(tsdn, &arena->lock);
 		return (ret);
 	}
 }
@@ -948,12 +1100,12 @@ arena_ptr_small_binind_get(const void *ptr, size_t mapbits)
 		size_t pageind;
 		size_t actual_mapbits;
 		size_t rpages_ind;
-		arena_run_t *run;
+		const arena_run_t *run;
 		arena_bin_t *bin;
 		szind_t run_binind, actual_binind;
 		arena_bin_info_t *bin_info;
-		arena_chunk_map_misc_t *miscelm;
-		void *rpages;
+		const arena_chunk_map_misc_t *miscelm;
+		const void *rpages;
 
 		assert(binind != BININD_INVALID);
 		assert(binind < NBINS);
@@ -966,11 +1118,11 @@ arena_ptr_small_binind_get(const void *ptr, size_t mapbits)
 		assert(arena_mapbits_allocated_get(chunk, pageind) != 0);
 		rpages_ind = pageind - arena_mapbits_small_runind_get(chunk,
 		    pageind);
-		miscelm = arena_miscelm_get(chunk, rpages_ind);
+		miscelm = arena_miscelm_get_const(chunk, rpages_ind);
 		run = &miscelm->run;
 		run_binind = run->binind;
 		bin = &arena->bins[run_binind];
-		actual_binind = bin - arena->bins;
+		actual_binind = (szind_t)(bin - arena->bins);
 		assert(run_binind == actual_binind);
 		bin_info = &arena_bin_info[actual_binind];
 		rpages = arena_miscelm_to_rpages(miscelm);
@@ -987,16 +1139,15 @@ arena_ptr_small_binind_get(const void *ptr, size_t mapbits)
 JEMALLOC_INLINE szind_t
 arena_bin_index(arena_t *arena, arena_bin_t *bin)
 {
-	szind_t binind = bin - arena->bins;
+	szind_t binind = (szind_t)(bin - arena->bins);
 	assert(binind < NBINS);
 	return (binind);
 }
 
-JEMALLOC_INLINE unsigned
+JEMALLOC_INLINE size_t
 arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
 {
-	unsigned shift, diff, regind;
-	size_t interval;
+	size_t diff, interval, shift, regind;
 	arena_chunk_map_misc_t *miscelm = arena_run_to_miscelm(run);
 	void *rpages = arena_miscelm_to_rpages(miscelm);
 
@@ -1011,12 +1162,12 @@ arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
 	 * Avoid doing division with a variable divisor if possible.  Using
 	 * actual division here can reduce allocator throughput by over 20%!
 	 */
-	diff = (unsigned)((uintptr_t)ptr - (uintptr_t)rpages -
+	diff = (size_t)((uintptr_t)ptr - (uintptr_t)rpages -
 	    bin_info->reg0_offset);
 
 	/* Rescale (factor powers of 2 out of the numerator and denominator). */
 	interval = bin_info->reg_interval;
-	shift = jemalloc_ffs(interval) - 1;
+	shift = ffs_zu(interval) - 1;
 	diff >>= shift;
 	interval >>= shift;
 
@@ -1038,9 +1189,9 @@ arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
 		 * divide by 0, and 1 and 2 are both powers of two, which are
 		 * handled above.
 		 */
-#define	SIZE_INV_SHIFT	((sizeof(unsigned) << 3) - LG_RUN_MAXREGS)
-#define	SIZE_INV(s)	(((1U << SIZE_INV_SHIFT) / (s)) + 1)
-		static const unsigned interval_invs[] = {
+#define	SIZE_INV_SHIFT	((sizeof(size_t) << 3) - LG_RUN_MAXREGS)
+#define	SIZE_INV(s)	(((ZU(1) << SIZE_INV_SHIFT) / (s)) + 1)
+		static const size_t interval_invs[] = {
 		    SIZE_INV(3),
 		    SIZE_INV(4), SIZE_INV(5), SIZE_INV(6), SIZE_INV(7),
 		    SIZE_INV(8), SIZE_INV(9), SIZE_INV(10), SIZE_INV(11),
@@ -1051,8 +1202,8 @@ arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
 		    SIZE_INV(28), SIZE_INV(29), SIZE_INV(30), SIZE_INV(31)
 		};
 
-		if (likely(interval <= ((sizeof(interval_invs) /
-		    sizeof(unsigned)) + 2))) {
+		if (likely(interval <= ((sizeof(interval_invs) / sizeof(size_t))
+		    + 2))) {
 			regind = (diff * interval_invs[interval - 3]) >>
 			    SIZE_INV_SHIFT;
 		} else
@@ -1067,7 +1218,7 @@ arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
 }
 
 JEMALLOC_INLINE prof_tctx_t *
-arena_prof_tctx_get(const void *ptr)
+arena_prof_tctx_get(tsdn_t *tsdn, const void *ptr)
 {
 	prof_tctx_t *ret;
 	arena_chunk_t *chunk;
@@ -1083,18 +1234,19 @@ arena_prof_tctx_get(const void *ptr)
 		if (likely((mapbits & CHUNK_MAP_LARGE) == 0))
 			ret = (prof_tctx_t *)(uintptr_t)1U;
 		else {
-			arena_chunk_map_misc_t *elm = arena_miscelm_get(chunk,
-			    pageind);
+			arena_chunk_map_misc_t *elm =
+			    arena_miscelm_get_mutable(chunk, pageind);
 			ret = atomic_read_p(&elm->prof_tctx_pun);
 		}
 	} else
-		ret = huge_prof_tctx_get(ptr);
+		ret = huge_prof_tctx_get(tsdn, ptr);
 
 	return (ret);
 }
 
 JEMALLOC_INLINE void
-arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)
+arena_prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    prof_tctx_t *tctx)
 {
 	arena_chunk_t *chunk;
 
@@ -1113,7 +1265,7 @@ arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)
 
 			assert(arena_mapbits_large_get(chunk, pageind) != 0);
 
-			elm = arena_miscelm_get(chunk, pageind);
+			elm = arena_miscelm_get_mutable(chunk, pageind);
 			atomic_write_p(&elm->prof_tctx_pun, tctx);
 		} else {
 			/*
@@ -1125,12 +1277,12 @@ arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)
 			assert(arena_mapbits_large_get(chunk, pageind) == 0);
 		}
 	} else
-		huge_prof_tctx_set(ptr, tctx);
+		huge_prof_tctx_set(tsdn, ptr, tctx);
 }
 
 JEMALLOC_INLINE void
-arena_prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
-    prof_tctx_t *old_tctx)
+arena_prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize,
+    const void *old_ptr, prof_tctx_t *old_tctx)
 {
 
 	cassert(config_prof);
@@ -1149,43 +1301,59 @@ arena_prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
 			    0);
 			assert(arena_mapbits_large_get(chunk, pageind) != 0);
 
-			elm = arena_miscelm_get(chunk, pageind);
+			elm = arena_miscelm_get_mutable(chunk, pageind);
 			atomic_write_p(&elm->prof_tctx_pun,
 			    (prof_tctx_t *)(uintptr_t)1U);
 		} else
-			huge_prof_tctx_reset(ptr);
+			huge_prof_tctx_reset(tsdn, ptr);
 	}
 }
 
+JEMALLOC_ALWAYS_INLINE void
+arena_decay_ticks(tsdn_t *tsdn, arena_t *arena, unsigned nticks)
+{
+	tsd_t *tsd;
+	ticker_t *decay_ticker;
+
+	if (unlikely(tsdn_null(tsdn)))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	decay_ticker = decay_ticker_get(tsd, arena->ind);
+	if (unlikely(decay_ticker == NULL))
+		return;
+	if (unlikely(ticker_ticks(decay_ticker, nticks)))
+		arena_purge(tsdn, arena, false);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_decay_tick(tsdn_t *tsdn, arena_t *arena)
+{
+
+	arena_decay_ticks(tsdn, arena, 1);
+}
+
 JEMALLOC_ALWAYS_INLINE void *
-arena_malloc(tsd_t *tsd, arena_t *arena, size_t size, bool zero,
-    tcache_t *tcache)
+arena_malloc(tsdn_t *tsdn, arena_t *arena, size_t size, szind_t ind, bool zero,
+    tcache_t *tcache, bool slow_path)
 {
 
+	assert(!tsdn_null(tsdn) || tcache == NULL);
 	assert(size != 0);
 
-	arena = arena_choose(tsd, arena);
-	if (unlikely(arena == NULL))
-		return (NULL);
+	if (likely(tcache != NULL)) {
+		if (likely(size <= SMALL_MAXCLASS)) {
+			return (tcache_alloc_small(tsdn_tsd(tsdn), arena,
+			    tcache, size, ind, zero, slow_path));
+		}
+		if (likely(size <= tcache_maxclass)) {
+			return (tcache_alloc_large(tsdn_tsd(tsdn), arena,
+			    tcache, size, ind, zero, slow_path));
+		}
+		/* (size > tcache_maxclass) case falls through. */
+		assert(size > tcache_maxclass);
+	}
 
-	if (likely(size <= SMALL_MAXCLASS)) {
-		if (likely(tcache != NULL)) {
-			return (tcache_alloc_small(tsd, arena, tcache, size,
-			    zero));
-		} else
-			return (arena_malloc_small(arena, size, zero));
-	} else if (likely(size <= large_maxclass)) {
-		/*
-		 * Initialize tcache after checking size in order to avoid
-		 * infinite recursion during tcache initialization.
-		 */
-		if (likely(tcache != NULL) && size <= tcache_maxclass) {
-			return (tcache_alloc_large(tsd, arena, tcache, size,
-			    zero));
-		} else
-			return (arena_malloc_large(arena, size, zero));
-	} else
-		return (huge_malloc(tsd, arena, size, zero, tcache));
+	return (arena_malloc_hard(tsdn, arena, size, ind, zero));
 }
 
 JEMALLOC_ALWAYS_INLINE arena_t *
@@ -1202,7 +1370,7 @@ arena_aalloc(const void *ptr)
 
 /* Return the size of the allocation pointed to by ptr. */
 JEMALLOC_ALWAYS_INLINE size_t
-arena_salloc(const void *ptr, bool demote)
+arena_salloc(tsdn_t *tsdn, const void *ptr, bool demote)
 {
 	size_t ret;
 	arena_chunk_t *chunk;
@@ -1245,17 +1413,18 @@ arena_salloc(const void *ptr, bool demote)
 			ret = index2size(binind);
 		}
 	} else
-		ret = huge_salloc(ptr);
+		ret = huge_salloc(tsdn, ptr);
 
 	return (ret);
 }
 
 JEMALLOC_ALWAYS_INLINE void
-arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache)
+arena_dalloc(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool slow_path)
 {
 	arena_chunk_t *chunk;
 	size_t pageind, mapbits;
 
+	assert(!tsdn_null(tsdn) || tcache == NULL);
 	assert(ptr != NULL);
 
 	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
@@ -1268,10 +1437,12 @@ arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache)
 			if (likely(tcache != NULL)) {
 				szind_t binind = arena_ptr_small_binind_get(ptr,
 				    mapbits);
-				tcache_dalloc_small(tsd, tcache, ptr, binind);
+				tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr,
+				    binind, slow_path);
 			} else {
-				arena_dalloc_small(extent_node_arena_get(
-				    &chunk->node), chunk, ptr, pageind);
+				arena_dalloc_small(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr, pageind);
 			}
 		} else {
 			size_t size = arena_mapbits_large_size_get(chunk,
@@ -1282,28 +1453,33 @@ arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache)
 
 			if (likely(tcache != NULL) && size - large_pad <=
 			    tcache_maxclass) {
-				tcache_dalloc_large(tsd, tcache, ptr, size -
-				    large_pad);
+				tcache_dalloc_large(tsdn_tsd(tsdn), tcache, ptr,
+				    size - large_pad, slow_path);
 			} else {
-				arena_dalloc_large(extent_node_arena_get(
-				    &chunk->node), chunk, ptr);
+				arena_dalloc_large(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr);
 			}
 		}
 	} else
-		huge_dalloc(tsd, ptr, tcache);
+		huge_dalloc(tsdn, ptr);
 }
 
 JEMALLOC_ALWAYS_INLINE void
-arena_sdalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
+arena_sdalloc(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path)
 {
 	arena_chunk_t *chunk;
 
+	assert(!tsdn_null(tsdn) || tcache == NULL);
+
 	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
 	if (likely(chunk != ptr)) {
 		if (config_prof && opt_prof) {
 			size_t pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >>
 			    LG_PAGE;
-			assert(arena_mapbits_allocated_get(chunk, pageind) != 0);
+			assert(arena_mapbits_allocated_get(chunk, pageind) !=
+			    0);
 			if (arena_mapbits_large_get(chunk, pageind) != 0) {
 				/*
 				 * Make sure to use promoted size, not request
@@ -1313,32 +1489,36 @@ arena_sdalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
 				    pageind) - large_pad;
 			}
 		}
-		assert(s2u(size) == s2u(arena_salloc(ptr, false)));
+		assert(s2u(size) == s2u(arena_salloc(tsdn, ptr, false)));
 
 		if (likely(size <= SMALL_MAXCLASS)) {
 			/* Small allocation. */
 			if (likely(tcache != NULL)) {
 				szind_t binind = size2index(size);
-				tcache_dalloc_small(tsd, tcache, ptr, binind);
+				tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr,
+				    binind, slow_path);
 			} else {
 				size_t pageind = ((uintptr_t)ptr -
 				    (uintptr_t)chunk) >> LG_PAGE;
-				arena_dalloc_small(extent_node_arena_get(
-				    &chunk->node), chunk, ptr, pageind);
+				arena_dalloc_small(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr, pageind);
 			}
 		} else {
 			assert(config_cache_oblivious || ((uintptr_t)ptr &
 			    PAGE_MASK) == 0);
 
-			if (likely(tcache != NULL) && size <= tcache_maxclass)
-				tcache_dalloc_large(tsd, tcache, ptr, size);
-			else {
-				arena_dalloc_large(extent_node_arena_get(
-				    &chunk->node), chunk, ptr);
+			if (likely(tcache != NULL) && size <= tcache_maxclass) {
+				tcache_dalloc_large(tsdn_tsd(tsdn), tcache, ptr,
+				    size, slow_path);
+			} else {
+				arena_dalloc_large(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr);
 			}
 		}
 	} else
-		huge_dalloc(tsd, ptr, tcache);
+		huge_dalloc(tsdn, ptr);
 }
 #  endif /* JEMALLOC_ARENA_INLINE_B */
 #endif
diff --git a/deps/jemalloc/include/jemalloc/internal/assert.h b/deps/jemalloc/include/jemalloc/internal/assert.h
new file mode 100644
index 000000000..6f8f7eb93
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/assert.h
@@ -0,0 +1,45 @@
+/*
+ * Define a custom assert() in order to reduce the chances of deadlock during
+ * assertion failure.
+ */
+#ifndef assert
+#define	assert(e) do {							\
+	if (unlikely(config_debug && !(e))) {				\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Failed assertion: \"%s\"\n",	\
+		    __FILE__, __LINE__, #e);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#ifndef not_reached
+#define	not_reached() do {						\
+	if (config_debug) {						\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Unreachable code reached\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+	unreachable();							\
+} while (0)
+#endif
+
+#ifndef not_implemented
+#define	not_implemented() do {						\
+	if (config_debug) {						\
+		malloc_printf("<jemalloc>: %s:%d: Not implemented\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#ifndef assert_not_implemented
+#define	assert_not_implemented(e) do {					\
+	if (unlikely(config_debug && !(e)))				\
+		not_implemented();					\
+} while (0)
+#endif
+
+
diff --git a/deps/jemalloc/include/jemalloc/internal/atomic.h b/deps/jemalloc/include/jemalloc/internal/atomic.h
index a9aad35d1..3f15ea149 100644
--- a/deps/jemalloc/include/jemalloc/internal/atomic.h
+++ b/deps/jemalloc/include/jemalloc/internal/atomic.h
@@ -28,8 +28,8 @@
  * callers.
  *
  *   <t> atomic_read_<t>(<t> *p) { return (*p); }
- *   <t> atomic_add_<t>(<t> *p, <t> x) { return (*p + x); }
- *   <t> atomic_sub_<t>(<t> *p, <t> x) { return (*p - x); }
+ *   <t> atomic_add_<t>(<t> *p, <t> x) { return (*p += x); }
+ *   <t> atomic_sub_<t>(<t> *p, <t> x) { return (*p -= x); }
  *   bool atomic_cas_<t>(<t> *p, <t> c, <t> s)
  *   {
  *     if (*p != c)
diff --git a/deps/jemalloc/include/jemalloc/internal/base.h b/deps/jemalloc/include/jemalloc/internal/base.h
index 39e46ee44..d6b81e162 100644
--- a/deps/jemalloc/include/jemalloc/internal/base.h
+++ b/deps/jemalloc/include/jemalloc/internal/base.h
@@ -9,12 +9,13 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS
 
-void	*base_alloc(size_t size);
-void	base_stats_get(size_t *allocated, size_t *resident, size_t *mapped);
+void	*base_alloc(tsdn_t *tsdn, size_t size);
+void	base_stats_get(tsdn_t *tsdn, size_t *allocated, size_t *resident,
+    size_t *mapped);
 bool	base_boot(void);
-void	base_prefork(void);
-void	base_postfork_parent(void);
-void	base_postfork_child(void);
+void	base_prefork(tsdn_t *tsdn);
+void	base_postfork_parent(tsdn_t *tsdn);
+void	base_postfork_child(tsdn_t *tsdn);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/bitmap.h b/deps/jemalloc/include/jemalloc/internal/bitmap.h
index fcc6005c7..36f38b59c 100644
--- a/deps/jemalloc/include/jemalloc/internal/bitmap.h
+++ b/deps/jemalloc/include/jemalloc/internal/bitmap.h
@@ -15,6 +15,15 @@ typedef unsigned long bitmap_t;
 #define	BITMAP_GROUP_NBITS		(ZU(1) << LG_BITMAP_GROUP_NBITS)
 #define	BITMAP_GROUP_NBITS_MASK		(BITMAP_GROUP_NBITS-1)
 
+/*
+ * Do some analysis on how big the bitmap is before we use a tree.  For a brute
+ * force linear search, if we would have to call ffs_lu() more than 2^3 times,
+ * use a tree instead.
+ */
+#if LG_BITMAP_MAXBITS - LG_BITMAP_GROUP_NBITS > 3
+#  define USE_TREE
+#endif
+
 /* Number of groups required to store a given number of bits. */
 #define	BITMAP_BITS2GROUPS(nbits)					\
     ((nbits + BITMAP_GROUP_NBITS_MASK) >> LG_BITMAP_GROUP_NBITS)
@@ -48,6 +57,8 @@ typedef unsigned long bitmap_t;
 /*
  * Maximum number of groups required to support LG_BITMAP_MAXBITS.
  */
+#ifdef USE_TREE
+
 #if LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS
 #  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_1_LEVEL(BITMAP_MAXBITS)
 #elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 2
@@ -65,6 +76,12 @@ typedef unsigned long bitmap_t;
     (LG_BITMAP_MAXBITS / LG_SIZEOF_BITMAP)				\
     + !!(LG_BITMAP_MAXBITS % LG_SIZEOF_BITMAP)
 
+#else /* USE_TREE */
+
+#define	BITMAP_GROUPS_MAX BITMAP_BITS2GROUPS(BITMAP_MAXBITS)
+
+#endif /* USE_TREE */
+
 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
 #ifdef JEMALLOC_H_STRUCTS
@@ -78,6 +95,7 @@ struct bitmap_info_s {
 	/* Logical number of bits in bitmap (stored at bottom level). */
 	size_t nbits;
 
+#ifdef USE_TREE
 	/* Number of levels necessary for nbits. */
 	unsigned nlevels;
 
@@ -86,6 +104,10 @@ struct bitmap_info_s {
 	 * bottom to top (e.g. the bottom level is stored in levels[0]).
 	 */
 	bitmap_level_t levels[BITMAP_MAX_LEVELS+1];
+#else /* USE_TREE */
+	/* Number of groups necessary for nbits. */
+	size_t ngroups;
+#endif /* USE_TREE */
 };
 
 #endif /* JEMALLOC_H_STRUCTS */
@@ -93,9 +115,8 @@ struct bitmap_info_s {
 #ifdef JEMALLOC_H_EXTERNS
 
 void	bitmap_info_init(bitmap_info_t *binfo, size_t nbits);
-size_t	bitmap_info_ngroups(const bitmap_info_t *binfo);
-size_t	bitmap_size(size_t nbits);
 void	bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo);
+size_t	bitmap_size(const bitmap_info_t *binfo);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
@@ -113,10 +134,20 @@ void	bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit);
 JEMALLOC_INLINE bool
 bitmap_full(bitmap_t *bitmap, const bitmap_info_t *binfo)
 {
-	unsigned rgoff = binfo->levels[binfo->nlevels].group_offset - 1;
+#ifdef USE_TREE
+	size_t rgoff = binfo->levels[binfo->nlevels].group_offset - 1;
 	bitmap_t rg = bitmap[rgoff];
 	/* The bitmap is full iff the root group is 0. */
 	return (rg == 0);
+#else
+	size_t i;
+
+	for (i = 0; i < binfo->ngroups; i++) {
+		if (bitmap[i] != 0)
+			return (false);
+	}
+	return (true);
+#endif
 }
 
 JEMALLOC_INLINE bool
@@ -128,7 +159,7 @@ bitmap_get(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
 	assert(bit < binfo->nbits);
 	goff = bit >> LG_BITMAP_GROUP_NBITS;
 	g = bitmap[goff];
-	return (!(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK))));
+	return (!(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK))));
 }
 
 JEMALLOC_INLINE void
@@ -143,10 +174,11 @@ bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
 	goff = bit >> LG_BITMAP_GROUP_NBITS;
 	gp = &bitmap[goff];
 	g = *gp;
-	assert(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)));
-	g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+	assert(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+	g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
 	*gp = g;
 	assert(bitmap_get(bitmap, binfo, bit));
+#ifdef USE_TREE
 	/* Propagate group state transitions up the tree. */
 	if (g == 0) {
 		unsigned i;
@@ -155,13 +187,14 @@ bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
 			goff = bit >> LG_BITMAP_GROUP_NBITS;
 			gp = &bitmap[binfo->levels[i].group_offset + goff];
 			g = *gp;
-			assert(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)));
-			g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+			assert(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+			g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
 			*gp = g;
 			if (g != 0)
 				break;
 		}
 	}
+#endif
 }
 
 /* sfu: set first unset. */
@@ -174,15 +207,24 @@ bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo)
 
 	assert(!bitmap_full(bitmap, binfo));
 
+#ifdef USE_TREE
 	i = binfo->nlevels - 1;
 	g = bitmap[binfo->levels[i].group_offset];
-	bit = jemalloc_ffsl(g) - 1;
+	bit = ffs_lu(g) - 1;
 	while (i > 0) {
 		i--;
 		g = bitmap[binfo->levels[i].group_offset + bit];
-		bit = (bit << LG_BITMAP_GROUP_NBITS) + (jemalloc_ffsl(g) - 1);
+		bit = (bit << LG_BITMAP_GROUP_NBITS) + (ffs_lu(g) - 1);
 	}
-
+#else
+	i = 0;
+	g = bitmap[0];
+	while ((bit = ffs_lu(g)) == 0) {
+		i++;
+		g = bitmap[i];
+	}
+	bit = (i << LG_BITMAP_GROUP_NBITS) + (bit - 1);
+#endif
 	bitmap_set(bitmap, binfo, bit);
 	return (bit);
 }
@@ -193,7 +235,7 @@ bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
 	size_t goff;
 	bitmap_t *gp;
 	bitmap_t g;
-	bool propagate;
+	UNUSED bool propagate;
 
 	assert(bit < binfo->nbits);
 	assert(bitmap_get(bitmap, binfo, bit));
@@ -201,10 +243,11 @@ bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
 	gp = &bitmap[goff];
 	g = *gp;
 	propagate = (g == 0);
-	assert((g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK))) == 0);
-	g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+	assert((g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK))) == 0);
+	g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
 	*gp = g;
 	assert(!bitmap_get(bitmap, binfo, bit));
+#ifdef USE_TREE
 	/* Propagate group state transitions up the tree. */
 	if (propagate) {
 		unsigned i;
@@ -214,14 +257,15 @@ bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
 			gp = &bitmap[binfo->levels[i].group_offset + goff];
 			g = *gp;
 			propagate = (g == 0);
-			assert((g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)))
+			assert((g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)))
 			    == 0);
-			g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+			g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
 			*gp = g;
 			if (!propagate)
 				break;
 		}
 	}
+#endif /* USE_TREE */
 }
 
 #endif
diff --git a/deps/jemalloc/include/jemalloc/internal/chunk.h b/deps/jemalloc/include/jemalloc/internal/chunk.h
index 5d1938353..50b9904b0 100644
--- a/deps/jemalloc/include/jemalloc/internal/chunk.h
+++ b/deps/jemalloc/include/jemalloc/internal/chunk.h
@@ -48,32 +48,30 @@ extern size_t		chunk_npages;
 
 extern const chunk_hooks_t	chunk_hooks_default;
 
-chunk_hooks_t	chunk_hooks_get(arena_t *arena);
-chunk_hooks_t	chunk_hooks_set(arena_t *arena,
+chunk_hooks_t	chunk_hooks_get(tsdn_t *tsdn, arena_t *arena);
+chunk_hooks_t	chunk_hooks_set(tsdn_t *tsdn, arena_t *arena,
     const chunk_hooks_t *chunk_hooks);
 
-bool	chunk_register(const void *chunk, const extent_node_t *node);
+bool	chunk_register(tsdn_t *tsdn, const void *chunk,
+    const extent_node_t *node);
 void	chunk_deregister(const void *chunk, const extent_node_t *node);
 void	*chunk_alloc_base(size_t size);
-void	*chunk_alloc_cache(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *new_addr, size_t size, size_t alignment, bool *zero,
-    bool dalloc_node);
-void	*chunk_alloc_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *new_addr, size_t size, size_t alignment, bool *zero, bool *commit);
-void	chunk_dalloc_cache(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *chunk, size_t size, bool committed);
-void	chunk_dalloc_arena(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *chunk, size_t size, bool zeroed, bool committed);
-void	chunk_dalloc_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *chunk, size_t size, bool committed);
-bool	chunk_purge_arena(arena_t *arena, void *chunk, size_t offset,
+void	*chunk_alloc_cache(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *new_addr, size_t size, size_t alignment,
+    size_t *sn, bool *zero, bool *commit, bool dalloc_node);
+void	*chunk_alloc_wrapper(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *new_addr, size_t size, size_t alignment,
+    size_t *sn, bool *zero, bool *commit);
+void	chunk_dalloc_cache(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *chunk, size_t size, size_t sn,
+    bool committed);
+void	chunk_dalloc_wrapper(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *chunk, size_t size, size_t sn,
+    bool zeroed, bool committed);
+bool	chunk_purge_wrapper(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *chunk, size_t size, size_t offset,
     size_t length);
-bool	chunk_purge_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *chunk, size_t size, size_t offset, size_t length);
 bool	chunk_boot(void);
-void	chunk_prefork(void);
-void	chunk_postfork_parent(void);
-void	chunk_postfork_child(void);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/chunk_dss.h b/deps/jemalloc/include/jemalloc/internal/chunk_dss.h
index 388f46be0..da8511ba0 100644
--- a/deps/jemalloc/include/jemalloc/internal/chunk_dss.h
+++ b/deps/jemalloc/include/jemalloc/internal/chunk_dss.h
@@ -23,13 +23,11 @@ extern const char *dss_prec_names[];
 
 dss_prec_t	chunk_dss_prec_get(void);
 bool	chunk_dss_prec_set(dss_prec_t dss_prec);
-void	*chunk_alloc_dss(arena_t *arena, void *new_addr, size_t size,
-    size_t alignment, bool *zero, bool *commit);
+void	*chunk_alloc_dss(tsdn_t *tsdn, arena_t *arena, void *new_addr,
+    size_t size, size_t alignment, bool *zero, bool *commit);
 bool	chunk_in_dss(void *chunk);
-bool	chunk_dss_boot(void);
-void	chunk_dss_prefork(void);
-void	chunk_dss_postfork_parent(void);
-void	chunk_dss_postfork_child(void);
+bool	chunk_dss_mergeable(void *chunk_a, void *chunk_b);
+void	chunk_dss_boot(void);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/chunk_mmap.h b/deps/jemalloc/include/jemalloc/internal/chunk_mmap.h
index 7d8014c58..6f2d0ac2e 100644
--- a/deps/jemalloc/include/jemalloc/internal/chunk_mmap.h
+++ b/deps/jemalloc/include/jemalloc/internal/chunk_mmap.h
@@ -9,8 +9,8 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS
 
-void	*chunk_alloc_mmap(size_t size, size_t alignment, bool *zero,
-    bool *commit);
+void	*chunk_alloc_mmap(void *new_addr, size_t size, size_t alignment,
+    bool *zero, bool *commit);
 bool	chunk_dalloc_mmap(void *chunk, size_t size);
 
 #endif /* JEMALLOC_H_EXTERNS */
diff --git a/deps/jemalloc/include/jemalloc/internal/ckh.h b/deps/jemalloc/include/jemalloc/internal/ckh.h
index 75c1c979f..f75ad90b7 100644
--- a/deps/jemalloc/include/jemalloc/internal/ckh.h
+++ b/deps/jemalloc/include/jemalloc/internal/ckh.h
@@ -40,9 +40,7 @@ struct ckh_s {
 #endif
 
 	/* Used for pseudo-random number generation. */
-#define	CKH_A		1103515241
-#define	CKH_C		12347
-	uint32_t	prng_state;
+	uint64_t	prng_state;
 
 	/* Total number of items. */
 	size_t		count;
@@ -74,7 +72,7 @@ bool	ckh_iter(ckh_t *ckh, size_t *tabind, void **key, void **data);
 bool	ckh_insert(tsd_t *tsd, ckh_t *ckh, const void *key, const void *data);
 bool	ckh_remove(tsd_t *tsd, ckh_t *ckh, const void *searchkey, void **key,
     void **data);
-bool	ckh_search(ckh_t *ckh, const void *seachkey, void **key, void **data);
+bool	ckh_search(ckh_t *ckh, const void *searchkey, void **key, void **data);
 void	ckh_string_hash(const void *key, size_t r_hash[2]);
 bool	ckh_string_keycomp(const void *k1, const void *k2);
 void	ckh_pointer_hash(const void *key, size_t r_hash[2]);
diff --git a/deps/jemalloc/include/jemalloc/internal/ctl.h b/deps/jemalloc/include/jemalloc/internal/ctl.h
index 751c14b5b..af0f6d7c5 100644
--- a/deps/jemalloc/include/jemalloc/internal/ctl.h
+++ b/deps/jemalloc/include/jemalloc/internal/ctl.h
@@ -21,13 +21,14 @@ struct ctl_named_node_s {
 	/* If (nchildren == 0), this is a terminal node. */
 	unsigned		nchildren;
 	const			ctl_node_t *children;
-	int			(*ctl)(const size_t *, size_t, void *, size_t *,
-	    void *, size_t);
+	int			(*ctl)(tsd_t *, const size_t *, size_t, void *,
+	    size_t *, void *, size_t);
 };
 
 struct ctl_indexed_node_s {
 	struct ctl_node_s	node;
-	const ctl_named_node_t	*(*index)(const size_t *, size_t, size_t);
+	const ctl_named_node_t	*(*index)(tsdn_t *, const size_t *, size_t,
+	    size_t);
 };
 
 struct ctl_arena_stats_s {
@@ -35,8 +36,12 @@ struct ctl_arena_stats_s {
 	unsigned		nthreads;
 	const char		*dss;
 	ssize_t			lg_dirty_mult;
+	ssize_t			decay_time;
 	size_t			pactive;
 	size_t			pdirty;
+
+	/* The remainder are only populated if config_stats is true. */
+
 	arena_stats_t		astats;
 
 	/* Aggregate stats for small size classes, based on bin stats. */
@@ -56,6 +61,7 @@ struct ctl_stats_s {
 	size_t			metadata;
 	size_t			resident;
 	size_t			mapped;
+	size_t			retained;
 	unsigned		narenas;
 	ctl_arena_stats_t	*arenas;	/* (narenas + 1) elements. */
 };
@@ -64,16 +70,17 @@ struct ctl_stats_s {
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS
 
-int	ctl_byname(const char *name, void *oldp, size_t *oldlenp, void *newp,
-    size_t newlen);
-int	ctl_nametomib(const char *name, size_t *mibp, size_t *miblenp);
-
-int	ctl_bymib(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp,
+int	ctl_byname(tsd_t *tsd, const char *name, void *oldp, size_t *oldlenp,
     void *newp, size_t newlen);
+int	ctl_nametomib(tsdn_t *tsdn, const char *name, size_t *mibp,
+    size_t *miblenp);
+
+int	ctl_bymib(tsd_t *tsd, const size_t *mib, size_t miblen, void *oldp,
+    size_t *oldlenp, void *newp, size_t newlen);
 bool	ctl_boot(void);
-void	ctl_prefork(void);
-void	ctl_postfork_parent(void);
-void	ctl_postfork_child(void);
+void	ctl_prefork(tsdn_t *tsdn);
+void	ctl_postfork_parent(tsdn_t *tsdn);
+void	ctl_postfork_child(tsdn_t *tsdn);
 
 #define	xmallctl(name, oldp, oldlenp, newp, newlen) do {		\
 	if (je_mallctl(name, oldp, oldlenp, newp, newlen)		\
diff --git a/deps/jemalloc/include/jemalloc/internal/extent.h b/deps/jemalloc/include/jemalloc/internal/extent.h
index 386d50ef4..168ffe643 100644
--- a/deps/jemalloc/include/jemalloc/internal/extent.h
+++ b/deps/jemalloc/include/jemalloc/internal/extent.h
@@ -19,6 +19,20 @@ struct extent_node_s {
 	size_t			en_size;
 
 	/*
+	 * Serial number (potentially non-unique).
+	 *
+	 * In principle serial numbers can wrap around on 32-bit systems if
+	 * JEMALLOC_MUNMAP is defined, but as long as comparison functions fall
+	 * back on address comparison for equal serial numbers, stable (if
+	 * imperfect) ordering is maintained.
+	 *
+	 * Serial numbers may not be unique even in the absence of wrap-around,
+	 * e.g. when splitting an extent and assigning the same serial number to
+	 * both resulting adjacent extents.
+	 */
+	size_t			en_sn;
+
+	/*
 	 * The zeroed flag is used by chunk recycling code to track whether
 	 * memory is zero-filled.
 	 */
@@ -45,10 +59,10 @@ struct extent_node_s {
 	qr(extent_node_t)	cc_link;
 
 	union {
-		/* Linkage for the size/address-ordered tree. */
-		rb_node(extent_node_t)	szad_link;
+		/* Linkage for the size/sn/address-ordered tree. */
+		rb_node(extent_node_t)	szsnad_link;
 
-		/* Linkage for arena's huge and node_cache lists. */
+		/* Linkage for arena's achunks, huge, and node_cache lists. */
 		ql_elm(extent_node_t)	ql_link;
 	};
 
@@ -61,7 +75,7 @@ typedef rb_tree(extent_node_t) extent_tree_t;
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS
 
-rb_proto(, extent_tree_szad_, extent_tree_t, extent_node_t)
+rb_proto(, extent_tree_szsnad_, extent_tree_t, extent_node_t)
 
 rb_proto(, extent_tree_ad_, extent_tree_t, extent_node_t)
 
@@ -73,6 +87,7 @@ rb_proto(, extent_tree_ad_, extent_tree_t, extent_node_t)
 arena_t	*extent_node_arena_get(const extent_node_t *node);
 void	*extent_node_addr_get(const extent_node_t *node);
 size_t	extent_node_size_get(const extent_node_t *node);
+size_t	extent_node_sn_get(const extent_node_t *node);
 bool	extent_node_zeroed_get(const extent_node_t *node);
 bool	extent_node_committed_get(const extent_node_t *node);
 bool	extent_node_achunk_get(const extent_node_t *node);
@@ -80,12 +95,13 @@ prof_tctx_t	*extent_node_prof_tctx_get(const extent_node_t *node);
 void	extent_node_arena_set(extent_node_t *node, arena_t *arena);
 void	extent_node_addr_set(extent_node_t *node, void *addr);
 void	extent_node_size_set(extent_node_t *node, size_t size);
+void	extent_node_sn_set(extent_node_t *node, size_t sn);
 void	extent_node_zeroed_set(extent_node_t *node, bool zeroed);
 void	extent_node_committed_set(extent_node_t *node, bool committed);
 void	extent_node_achunk_set(extent_node_t *node, bool achunk);
 void	extent_node_prof_tctx_set(extent_node_t *node, prof_tctx_t *tctx);
 void	extent_node_init(extent_node_t *node, arena_t *arena, void *addr,
-    size_t size, bool zeroed, bool committed);
+    size_t size, size_t sn, bool zeroed, bool committed);
 void	extent_node_dirty_linkage_init(extent_node_t *node);
 void	extent_node_dirty_insert(extent_node_t *node,
     arena_runs_dirty_link_t *runs_dirty, extent_node_t *chunks_dirty);
@@ -114,6 +130,13 @@ extent_node_size_get(const extent_node_t *node)
 	return (node->en_size);
 }
 
+JEMALLOC_INLINE size_t
+extent_node_sn_get(const extent_node_t *node)
+{
+
+	return (node->en_sn);
+}
+
 JEMALLOC_INLINE bool
 extent_node_zeroed_get(const extent_node_t *node)
 {
@@ -165,6 +188,13 @@ extent_node_size_set(extent_node_t *node, size_t size)
 }
 
 JEMALLOC_INLINE void
+extent_node_sn_set(extent_node_t *node, size_t sn)
+{
+
+	node->en_sn = sn;
+}
+
+JEMALLOC_INLINE void
 extent_node_zeroed_set(extent_node_t *node, bool zeroed)
 {
 
@@ -194,12 +224,13 @@ extent_node_prof_tctx_set(extent_node_t *node, prof_tctx_t *tctx)
 
 JEMALLOC_INLINE void
 extent_node_init(extent_node_t *node, arena_t *arena, void *addr, size_t size,
-    bool zeroed, bool committed)
+    size_t sn, bool zeroed, bool committed)
 {
 
 	extent_node_arena_set(node, arena);
 	extent_node_addr_set(node, addr);
 	extent_node_size_set(node, size);
+	extent_node_sn_set(node, sn);
 	extent_node_zeroed_set(node, zeroed);
 	extent_node_committed_set(node, committed);
 	extent_node_achunk_set(node, false);
diff --git a/deps/jemalloc/include/jemalloc/internal/hash.h b/deps/jemalloc/include/jemalloc/internal/hash.h
index bcead337a..1ff2d9a05 100644
--- a/deps/jemalloc/include/jemalloc/internal/hash.h
+++ b/deps/jemalloc/include/jemalloc/internal/hash.h
@@ -1,6 +1,6 @@
 /*
  * The following hash function is based on MurmurHash3, placed into the public
- * domain by Austin Appleby.  See http://code.google.com/p/smhasher/ for
+ * domain by Austin Appleby.  See https://github.com/aappleby/smhasher for
  * details.
  */
 /******************************************************************************/
@@ -49,6 +49,14 @@ JEMALLOC_INLINE uint32_t
 hash_get_block_32(const uint32_t *p, int i)
 {
 
+	/* Handle unaligned read. */
+	if (unlikely((uintptr_t)p & (sizeof(uint32_t)-1)) != 0) {
+		uint32_t ret;
+
+		memcpy(&ret, (uint8_t *)(p + i), sizeof(uint32_t));
+		return (ret);
+	}
+
 	return (p[i]);
 }
 
@@ -56,6 +64,14 @@ JEMALLOC_INLINE uint64_t
 hash_get_block_64(const uint64_t *p, int i)
 {
 
+	/* Handle unaligned read. */
+	if (unlikely((uintptr_t)p & (sizeof(uint64_t)-1)) != 0) {
+		uint64_t ret;
+
+		memcpy(&ret, (uint8_t *)(p + i), sizeof(uint64_t));
+		return (ret);
+	}
+
 	return (p[i]);
 }
 
@@ -321,13 +337,18 @@ hash_x64_128(const void *key, const int len, const uint32_t seed,
 JEMALLOC_INLINE void
 hash(const void *key, size_t len, const uint32_t seed, size_t r_hash[2])
 {
+
+	assert(len <= INT_MAX); /* Unfortunate implementation limitation. */
+
 #if (LG_SIZEOF_PTR == 3 && !defined(JEMALLOC_BIG_ENDIAN))
-	hash_x64_128(key, len, seed, (uint64_t *)r_hash);
+	hash_x64_128(key, (int)len, seed, (uint64_t *)r_hash);
 #else
-	uint64_t hashes[2];
-	hash_x86_128(key, len, seed, hashes);
-	r_hash[0] = (size_t)hashes[0];
-	r_hash[1] = (size_t)hashes[1];
+	{
+		uint64_t hashes[2];
+		hash_x86_128(key, (int)len, seed, hashes);
+		r_hash[0] = (size_t)hashes[0];
+		r_hash[1] = (size_t)hashes[1];
+	}
 #endif
 }
 #endif
diff --git a/deps/jemalloc/include/jemalloc/internal/huge.h b/deps/jemalloc/include/jemalloc/internal/huge.h
index ece7af980..22184d9bb 100644
--- a/deps/jemalloc/include/jemalloc/internal/huge.h
+++ b/deps/jemalloc/include/jemalloc/internal/huge.h
@@ -9,24 +9,23 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS
 
-void	*huge_malloc(tsd_t *tsd, arena_t *arena, size_t size, bool zero,
-    tcache_t *tcache);
-void	*huge_palloc(tsd_t *tsd, arena_t *arena, size_t size, size_t alignment,
-    bool zero, tcache_t *tcache);
-bool	huge_ralloc_no_move(void *ptr, size_t oldsize, size_t usize_min,
-    size_t usize_max, bool zero);
+void	*huge_malloc(tsdn_t *tsdn, arena_t *arena, size_t usize, bool zero);
+void	*huge_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize,
+    size_t alignment, bool zero);
+bool	huge_ralloc_no_move(tsdn_t *tsdn, void *ptr, size_t oldsize,
+    size_t usize_min, size_t usize_max, bool zero);
 void	*huge_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize,
     size_t usize, size_t alignment, bool zero, tcache_t *tcache);
 #ifdef JEMALLOC_JET
 typedef void (huge_dalloc_junk_t)(void *, size_t);
 extern huge_dalloc_junk_t *huge_dalloc_junk;
 #endif
-void	huge_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache);
+void	huge_dalloc(tsdn_t *tsdn, void *ptr);
 arena_t	*huge_aalloc(const void *ptr);
-size_t	huge_salloc(const void *ptr);
-prof_tctx_t	*huge_prof_tctx_get(const void *ptr);
-void	huge_prof_tctx_set(const void *ptr, prof_tctx_t *tctx);
-void	huge_prof_tctx_reset(const void *ptr);
+size_t	huge_salloc(tsdn_t *tsdn, const void *ptr);
+prof_tctx_t	*huge_prof_tctx_get(tsdn_t *tsdn, const void *ptr);
+void	huge_prof_tctx_set(tsdn_t *tsdn, const void *ptr, prof_tctx_t *tctx);
+void	huge_prof_tctx_reset(tsdn_t *tsdn, const void *ptr);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in b/deps/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in
index 8536a3eda..e7ace7d8c 100644
--- a/deps/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in
+++ b/deps/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in
@@ -49,6 +49,7 @@ static const bool config_lazy_lock =
     false
 #endif
     ;
+static const char * const config_malloc_conf = JEMALLOC_CONFIG_MALLOC_CONF;
 static const bool config_prof =
 #ifdef JEMALLOC_PROF
     true
@@ -160,7 +161,10 @@ static const bool config_cache_oblivious =
 #include <malloc/malloc.h>
 #endif
 
+#include "jemalloc/internal/ph.h"
+#ifndef __PGI
 #define	RB_COMPACT
+#endif
 #include "jemalloc/internal/rb.h"
 #include "jemalloc/internal/qr.h"
 #include "jemalloc/internal/ql.h"
@@ -183,6 +187,9 @@ static const bool config_cache_oblivious =
 
 #include "jemalloc/internal/jemalloc_internal_macros.h"
 
+/* Page size index type. */
+typedef unsigned pszind_t;
+
 /* Size class index type. */
 typedef unsigned szind_t;
 
@@ -232,7 +239,7 @@ typedef unsigned szind_t;
 #  ifdef __alpha__
 #    define LG_QUANTUM		4
 #  endif
-#  if (defined(__sparc64__) || defined(__sparcv9))
+#  if (defined(__sparc64__) || defined(__sparcv9) || defined(__sparc_v9__))
 #    define LG_QUANTUM		4
 #  endif
 #  if (defined(__amd64__) || defined(__x86_64__) || defined(_M_X64))
@@ -256,6 +263,9 @@ typedef unsigned szind_t;
 #  ifdef __powerpc__
 #    define LG_QUANTUM		4
 #  endif
+#  ifdef __riscv__
+#    define LG_QUANTUM		4
+#  endif
 #  ifdef __s390__
 #    define LG_QUANTUM		4
 #  endif
@@ -317,13 +327,17 @@ typedef unsigned szind_t;
 #define	PAGE		((size_t)(1U << LG_PAGE))
 #define	PAGE_MASK	((size_t)(PAGE - 1))
 
+/* Return the page base address for the page containing address a. */
+#define	PAGE_ADDR2BASE(a)						\
+	((void *)((uintptr_t)(a) & ~PAGE_MASK))
+
 /* Return the smallest pagesize multiple that is >= s. */
 #define	PAGE_CEILING(s)							\
 	(((s) + PAGE_MASK) & ~PAGE_MASK)
 
 /* Return the nearest aligned address at or below a. */
 #define	ALIGNMENT_ADDR2BASE(a, alignment)				\
-	((void *)((uintptr_t)(a) & (-(alignment))))
+	((void *)((uintptr_t)(a) & ((~(alignment)) + 1)))
 
 /* Return the offset between a and the nearest aligned address at or below a. */
 #define	ALIGNMENT_ADDR2OFFSET(a, alignment)				\
@@ -331,7 +345,7 @@ typedef unsigned szind_t;
 
 /* Return the smallest alignment multiple that is >= s. */
 #define	ALIGNMENT_CEILING(s, alignment)					\
-	(((s) + (alignment - 1)) & (-(alignment)))
+	(((s) + (alignment - 1)) & ((~(alignment)) + 1))
 
 /* Declare a variable-length array. */
 #if __STDC_VERSION__ < 199901L
@@ -351,14 +365,19 @@ typedef unsigned szind_t;
 #  define VARIABLE_ARRAY(type, name, count) type name[(count)]
 #endif
 
+#include "jemalloc/internal/nstime.h"
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
+#include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
 #include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/witness.h"
 #include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/tsd.h"
 #include "jemalloc/internal/mb.h"
@@ -379,14 +398,19 @@ typedef unsigned szind_t;
 /******************************************************************************/
 #define	JEMALLOC_H_STRUCTS
 
+#include "jemalloc/internal/nstime.h"
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
+#include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
 #include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/witness.h"
 #include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/mb.h"
 #include "jemalloc/internal/bitmap.h"
@@ -422,14 +446,28 @@ extern bool	opt_redzone;
 extern bool	opt_utrace;
 extern bool	opt_xmalloc;
 extern bool	opt_zero;
-extern size_t	opt_narenas;
+extern unsigned	opt_narenas;
 
 extern bool	in_valgrind;
 
 /* Number of CPUs. */
-extern unsigned		ncpus;
+extern unsigned	ncpus;
+
+/* Number of arenas used for automatic multiplexing of threads and arenas. */
+extern unsigned	narenas_auto;
 
 /*
+ * Arenas that are used to service external requests.  Not all elements of the
+ * arenas array are necessarily used; arenas are created lazily as needed.
+ */
+extern arena_t	**arenas;
+
+/*
+ * pind2sz_tab encodes the same information as could be computed by
+ * pind2sz_compute().
+ */
+extern size_t const	pind2sz_tab[NPSIZES];
+/*
  * index2size_tab encodes the same information as could be computed (at
  * unacceptable cost in some code paths) by index2size_compute().
  */
@@ -447,31 +485,35 @@ void	a0dalloc(void *ptr);
 void	*bootstrap_malloc(size_t size);
 void	*bootstrap_calloc(size_t num, size_t size);
 void	bootstrap_free(void *ptr);
-arena_t	*arenas_extend(unsigned ind);
-arena_t	*arena_init(unsigned ind);
 unsigned	narenas_total_get(void);
-arena_t	*arena_get_hard(tsd_t *tsd, unsigned ind, bool init_if_missing);
-arena_t	*arena_choose_hard(tsd_t *tsd);
+arena_t	*arena_init(tsdn_t *tsdn, unsigned ind);
+arena_tdata_t	*arena_tdata_get_hard(tsd_t *tsd, unsigned ind);
+arena_t	*arena_choose_hard(tsd_t *tsd, bool internal);
 void	arena_migrate(tsd_t *tsd, unsigned oldind, unsigned newind);
-unsigned	arena_nbound(unsigned ind);
 void	thread_allocated_cleanup(tsd_t *tsd);
 void	thread_deallocated_cleanup(tsd_t *tsd);
+void	iarena_cleanup(tsd_t *tsd);
 void	arena_cleanup(tsd_t *tsd);
-void	arenas_cache_cleanup(tsd_t *tsd);
-void	narenas_cache_cleanup(tsd_t *tsd);
-void	arenas_cache_bypass_cleanup(tsd_t *tsd);
+void	arenas_tdata_cleanup(tsd_t *tsd);
+void	narenas_tdata_cleanup(tsd_t *tsd);
+void	arenas_tdata_bypass_cleanup(tsd_t *tsd);
 void	jemalloc_prefork(void);
 void	jemalloc_postfork_parent(void);
 void	jemalloc_postfork_child(void);
 
+#include "jemalloc/internal/nstime.h"
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
+#include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
 #include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/witness.h"
 #include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/mb.h"
 #include "jemalloc/internal/bitmap.h"
@@ -492,16 +534,21 @@ void	jemalloc_postfork_child(void);
 /******************************************************************************/
 #define	JEMALLOC_H_INLINES
 
+#include "jemalloc/internal/nstime.h"
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
+#include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
 #include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
-#include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/tsd.h"
+#include "jemalloc/internal/witness.h"
+#include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/mb.h"
 #include "jemalloc/internal/extent.h"
 #include "jemalloc/internal/base.h"
@@ -511,6 +558,11 @@ void	jemalloc_postfork_child(void);
 #include "jemalloc/internal/huge.h"
 
 #ifndef JEMALLOC_ENABLE_INLINE
+pszind_t	psz2ind(size_t psz);
+size_t	pind2sz_compute(pszind_t pind);
+size_t	pind2sz_lookup(pszind_t pind);
+size_t	pind2sz(pszind_t pind);
+size_t	psz2u(size_t psz);
 szind_t	size2index_compute(size_t size);
 szind_t	size2index_lookup(size_t size);
 szind_t	size2index(size_t size);
@@ -521,39 +573,121 @@ size_t	s2u_compute(size_t size);
 size_t	s2u_lookup(size_t size);
 size_t	s2u(size_t size);
 size_t	sa2u(size_t size, size_t alignment);
+arena_t	*arena_choose_impl(tsd_t *tsd, arena_t *arena, bool internal);
 arena_t	*arena_choose(tsd_t *tsd, arena_t *arena);
-arena_t	*arena_get(tsd_t *tsd, unsigned ind, bool init_if_missing,
+arena_t	*arena_ichoose(tsd_t *tsd, arena_t *arena);
+arena_tdata_t	*arena_tdata_get(tsd_t *tsd, unsigned ind,
     bool refresh_if_missing);
+arena_t	*arena_get(tsdn_t *tsdn, unsigned ind, bool init_if_missing);
+ticker_t	*decay_ticker_get(tsd_t *tsd, unsigned ind);
 #endif
 
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_C_))
+JEMALLOC_INLINE pszind_t
+psz2ind(size_t psz)
+{
+
+	if (unlikely(psz > HUGE_MAXCLASS))
+		return (NPSIZES);
+	{
+		pszind_t x = lg_floor((psz<<1)-1);
+		pszind_t shift = (x < LG_SIZE_CLASS_GROUP + LG_PAGE) ? 0 : x -
+		    (LG_SIZE_CLASS_GROUP + LG_PAGE);
+		pszind_t grp = shift << LG_SIZE_CLASS_GROUP;
+
+		pszind_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_PAGE + 1) ?
+		    LG_PAGE : x - LG_SIZE_CLASS_GROUP - 1;
+
+		size_t delta_inverse_mask = ZI(-1) << lg_delta;
+		pszind_t mod = ((((psz-1) & delta_inverse_mask) >> lg_delta)) &
+		    ((ZU(1) << LG_SIZE_CLASS_GROUP) - 1);
+
+		pszind_t ind = grp + mod;
+		return (ind);
+	}
+}
+
+JEMALLOC_INLINE size_t
+pind2sz_compute(pszind_t pind)
+{
+
+	{
+		size_t grp = pind >> LG_SIZE_CLASS_GROUP;
+		size_t mod = pind & ((ZU(1) << LG_SIZE_CLASS_GROUP) - 1);
+
+		size_t grp_size_mask = ~((!!grp)-1);
+		size_t grp_size = ((ZU(1) << (LG_PAGE +
+		    (LG_SIZE_CLASS_GROUP-1))) << grp) & grp_size_mask;
+
+		size_t shift = (grp == 0) ? 1 : grp;
+		size_t lg_delta = shift + (LG_PAGE-1);
+		size_t mod_size = (mod+1) << lg_delta;
+
+		size_t sz = grp_size + mod_size;
+		return (sz);
+	}
+}
+
+JEMALLOC_INLINE size_t
+pind2sz_lookup(pszind_t pind)
+{
+	size_t ret = (size_t)pind2sz_tab[pind];
+	assert(ret == pind2sz_compute(pind));
+	return (ret);
+}
+
+JEMALLOC_INLINE size_t
+pind2sz(pszind_t pind)
+{
+
+	assert(pind < NPSIZES);
+	return (pind2sz_lookup(pind));
+}
+
+JEMALLOC_INLINE size_t
+psz2u(size_t psz)
+{
+
+	if (unlikely(psz > HUGE_MAXCLASS))
+		return (0);
+	{
+		size_t x = lg_floor((psz<<1)-1);
+		size_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_PAGE + 1) ?
+		    LG_PAGE : x - LG_SIZE_CLASS_GROUP - 1;
+		size_t delta = ZU(1) << lg_delta;
+		size_t delta_mask = delta - 1;
+		size_t usize = (psz + delta_mask) & ~delta_mask;
+		return (usize);
+	}
+}
+
 JEMALLOC_INLINE szind_t
 size2index_compute(size_t size)
 {
 
+	if (unlikely(size > HUGE_MAXCLASS))
+		return (NSIZES);
 #if (NTBINS != 0)
 	if (size <= (ZU(1) << LG_TINY_MAXCLASS)) {
-		size_t lg_tmin = LG_TINY_MAXCLASS - NTBINS + 1;
-		size_t lg_ceil = lg_floor(pow2_ceil(size));
+		szind_t lg_tmin = LG_TINY_MAXCLASS - NTBINS + 1;
+		szind_t lg_ceil = lg_floor(pow2_ceil_zu(size));
 		return (lg_ceil < lg_tmin ? 0 : lg_ceil - lg_tmin);
 	}
 #endif
 	{
-		size_t x = unlikely(ZI(size) < 0) ? ((size<<1) ?
-		    (ZU(1)<<(LG_SIZEOF_PTR+3)) : ((ZU(1)<<(LG_SIZEOF_PTR+3))-1))
-		    : lg_floor((size<<1)-1);
-		size_t shift = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM) ? 0 :
+		szind_t x = lg_floor((size<<1)-1);
+		szind_t shift = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM) ? 0 :
 		    x - (LG_SIZE_CLASS_GROUP + LG_QUANTUM);
-		size_t grp = shift << LG_SIZE_CLASS_GROUP;
+		szind_t grp = shift << LG_SIZE_CLASS_GROUP;
 
-		size_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM + 1)
+		szind_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM + 1)
 		    ? LG_QUANTUM : x - LG_SIZE_CLASS_GROUP - 1;
 
 		size_t delta_inverse_mask = ZI(-1) << lg_delta;
-		size_t mod = ((((size-1) & delta_inverse_mask) >> lg_delta)) &
+		szind_t mod = ((((size-1) & delta_inverse_mask) >> lg_delta)) &
 		    ((ZU(1) << LG_SIZE_CLASS_GROUP) - 1);
 
-		size_t index = NTBINS + grp + mod;
+		szind_t index = NTBINS + grp + mod;
 		return (index);
 	}
 }
@@ -564,8 +698,7 @@ size2index_lookup(size_t size)
 
 	assert(size <= LOOKUP_MAXCLASS);
 	{
-		size_t ret = ((size_t)(size2index_tab[(size-1) >>
-		    LG_TINY_MIN]));
+		szind_t ret = (size2index_tab[(size-1) >> LG_TINY_MIN]);
 		assert(ret == size2index_compute(size));
 		return (ret);
 	}
@@ -628,18 +761,18 @@ JEMALLOC_ALWAYS_INLINE size_t
 s2u_compute(size_t size)
 {
 
+	if (unlikely(size > HUGE_MAXCLASS))
+		return (0);
 #if (NTBINS > 0)
 	if (size <= (ZU(1) << LG_TINY_MAXCLASS)) {
 		size_t lg_tmin = LG_TINY_MAXCLASS - NTBINS + 1;
-		size_t lg_ceil = lg_floor(pow2_ceil(size));
+		size_t lg_ceil = lg_floor(pow2_ceil_zu(size));
 		return (lg_ceil < lg_tmin ? (ZU(1) << lg_tmin) :
 		    (ZU(1) << lg_ceil));
 	}
 #endif
 	{
-		size_t x = unlikely(ZI(size) < 0) ? ((size<<1) ?
-		    (ZU(1)<<(LG_SIZEOF_PTR+3)) : ((ZU(1)<<(LG_SIZEOF_PTR+3))-1))
-		    : lg_floor((size<<1)-1);
+		size_t x = lg_floor((size<<1)-1);
 		size_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM + 1)
 		    ?  LG_QUANTUM : x - LG_SIZE_CLASS_GROUP - 1;
 		size_t delta = ZU(1) << lg_delta;
@@ -723,17 +856,16 @@ sa2u(size_t size, size_t alignment)
 			return (usize);
 	}
 
-	/* Huge size class.  Beware of size_t overflow. */
+	/* Huge size class.  Beware of overflow. */
+
+	if (unlikely(alignment > HUGE_MAXCLASS))
+		return (0);
 
 	/*
 	 * We can't achieve subchunk alignment, so round up alignment to the
 	 * minimum that can actually be supported.
 	 */
 	alignment = CHUNK_CEILING(alignment);
-	if (alignment == 0) {
-		/* size_t overflow. */
-		return (0);
-	}
 
 	/* Make sure result is a huge size class. */
 	if (size <= chunksize)
@@ -759,45 +891,84 @@ sa2u(size_t size, size_t alignment)
 
 /* Choose an arena based on a per-thread value. */
 JEMALLOC_INLINE arena_t *
-arena_choose(tsd_t *tsd, arena_t *arena)
+arena_choose_impl(tsd_t *tsd, arena_t *arena, bool internal)
 {
 	arena_t *ret;
 
 	if (arena != NULL)
 		return (arena);
 
-	if (unlikely((ret = tsd_arena_get(tsd)) == NULL))
-		ret = arena_choose_hard(tsd);
+	ret = internal ? tsd_iarena_get(tsd) : tsd_arena_get(tsd);
+	if (unlikely(ret == NULL))
+		ret = arena_choose_hard(tsd, internal);
 
 	return (ret);
 }
 
 JEMALLOC_INLINE arena_t *
-arena_get(tsd_t *tsd, unsigned ind, bool init_if_missing,
-    bool refresh_if_missing)
+arena_choose(tsd_t *tsd, arena_t *arena)
 {
-	arena_t *arena;
-	arena_t **arenas_cache = tsd_arenas_cache_get(tsd);
 
-	/* init_if_missing requires refresh_if_missing. */
-	assert(!init_if_missing || refresh_if_missing);
+	return (arena_choose_impl(tsd, arena, false));
+}
+
+JEMALLOC_INLINE arena_t *
+arena_ichoose(tsd_t *tsd, arena_t *arena)
+{
 
-	if (unlikely(arenas_cache == NULL)) {
-		/* arenas_cache hasn't been initialized yet. */
-		return (arena_get_hard(tsd, ind, init_if_missing));
+	return (arena_choose_impl(tsd, arena, true));
+}
+
+JEMALLOC_INLINE arena_tdata_t *
+arena_tdata_get(tsd_t *tsd, unsigned ind, bool refresh_if_missing)
+{
+	arena_tdata_t *tdata;
+	arena_tdata_t *arenas_tdata = tsd_arenas_tdata_get(tsd);
+
+	if (unlikely(arenas_tdata == NULL)) {
+		/* arenas_tdata hasn't been initialized yet. */
+		return (arena_tdata_get_hard(tsd, ind));
 	}
-	if (unlikely(ind >= tsd_narenas_cache_get(tsd))) {
+	if (unlikely(ind >= tsd_narenas_tdata_get(tsd))) {
 		/*
-		 * ind is invalid, cache is old (too small), or arena to be
+		 * ind is invalid, cache is old (too small), or tdata to be
 		 * initialized.
 		 */
-		return (refresh_if_missing ? arena_get_hard(tsd, ind,
-		    init_if_missing) : NULL);
+		return (refresh_if_missing ? arena_tdata_get_hard(tsd, ind) :
+		    NULL);
 	}
-	arena = arenas_cache[ind];
-	if (likely(arena != NULL) || !refresh_if_missing)
-		return (arena);
-	return (arena_get_hard(tsd, ind, init_if_missing));
+
+	tdata = &arenas_tdata[ind];
+	if (likely(tdata != NULL) || !refresh_if_missing)
+		return (tdata);
+	return (arena_tdata_get_hard(tsd, ind));
+}
+
+JEMALLOC_INLINE arena_t *
+arena_get(tsdn_t *tsdn, unsigned ind, bool init_if_missing)
+{
+	arena_t *ret;
+
+	assert(ind <= MALLOCX_ARENA_MAX);
+
+	ret = arenas[ind];
+	if (unlikely(ret == NULL)) {
+		ret = atomic_read_p((void *)&arenas[ind]);
+		if (init_if_missing && unlikely(ret == NULL))
+			ret = arena_init(tsdn, ind);
+	}
+	return (ret);
+}
+
+JEMALLOC_INLINE ticker_t *
+decay_ticker_get(tsd_t *tsd, unsigned ind)
+{
+	arena_tdata_t *tdata;
+
+	tdata = arena_tdata_get(tsd, ind, true);
+	if (unlikely(tdata == NULL))
+		return (NULL);
+	return (&tdata->decay_ticker);
 }
 #endif
 
@@ -818,27 +989,27 @@ arena_get(tsd_t *tsd, unsigned ind, bool init_if_missing,
 
 #ifndef JEMALLOC_ENABLE_INLINE
 arena_t	*iaalloc(const void *ptr);
-size_t	isalloc(const void *ptr, bool demote);
-void	*iallocztm(tsd_t *tsd, size_t size, bool zero, tcache_t *tcache,
-    bool is_metadata, arena_t *arena);
-void	*imalloct(tsd_t *tsd, size_t size, tcache_t *tcache, arena_t *arena);
-void	*imalloc(tsd_t *tsd, size_t size);
-void	*icalloct(tsd_t *tsd, size_t size, tcache_t *tcache, arena_t *arena);
-void	*icalloc(tsd_t *tsd, size_t size);
-void	*ipallocztm(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+size_t	isalloc(tsdn_t *tsdn, const void *ptr, bool demote);
+void	*iallocztm(tsdn_t *tsdn, size_t size, szind_t ind, bool zero,
+    tcache_t *tcache, bool is_metadata, arena_t *arena, bool slow_path);
+void	*ialloc(tsd_t *tsd, size_t size, szind_t ind, bool zero,
+    bool slow_path);
+void	*ipallocztm(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
     tcache_t *tcache, bool is_metadata, arena_t *arena);
-void	*ipalloct(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+void	*ipalloct(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
     tcache_t *tcache, arena_t *arena);
 void	*ipalloc(tsd_t *tsd, size_t usize, size_t alignment, bool zero);
-size_t	ivsalloc(const void *ptr, bool demote);
+size_t	ivsalloc(tsdn_t *tsdn, const void *ptr, bool demote);
 size_t	u2rz(size_t usize);
-size_t	p2rz(const void *ptr);
-void	idalloctm(tsd_t *tsd, void *ptr, tcache_t *tcache, bool is_metadata);
-void	idalloct(tsd_t *tsd, void *ptr, tcache_t *tcache);
+size_t	p2rz(tsdn_t *tsdn, const void *ptr);
+void	idalloctm(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool is_metadata,
+    bool slow_path);
 void	idalloc(tsd_t *tsd, void *ptr);
-void	iqalloc(tsd_t *tsd, void *ptr, tcache_t *tcache);
-void	isdalloct(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache);
-void	isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache);
+void	iqalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path);
+void	isdalloct(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path);
+void	isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path);
 void	*iralloct_realign(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
     size_t extra, size_t alignment, bool zero, tcache_t *tcache,
     arena_t *arena);
@@ -846,8 +1017,8 @@ void	*iralloct(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
     size_t alignment, bool zero, tcache_t *tcache, arena_t *arena);
 void	*iralloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
     size_t alignment, bool zero);
-bool	ixalloc(void *ptr, size_t oldsize, size_t size, size_t extra,
-    size_t alignment, bool zero);
+bool	ixalloc(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
+    size_t extra, size_t alignment, bool zero);
 #endif
 
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_C_))
@@ -862,100 +1033,85 @@ iaalloc(const void *ptr)
 
 /*
  * Typical usage:
+ *   tsdn_t *tsdn = [...]
  *   void *ptr = [...]
- *   size_t sz = isalloc(ptr, config_prof);
+ *   size_t sz = isalloc(tsdn, ptr, config_prof);
  */
 JEMALLOC_ALWAYS_INLINE size_t
-isalloc(const void *ptr, bool demote)
+isalloc(tsdn_t *tsdn, const void *ptr, bool demote)
 {
 
 	assert(ptr != NULL);
 	/* Demotion only makes sense if config_prof is true. */
 	assert(config_prof || !demote);
 
-	return (arena_salloc(ptr, demote));
+	return (arena_salloc(tsdn, ptr, demote));
 }
 
 JEMALLOC_ALWAYS_INLINE void *
-iallocztm(tsd_t *tsd, size_t size, bool zero, tcache_t *tcache, bool is_metadata,
-    arena_t *arena)
+iallocztm(tsdn_t *tsdn, size_t size, szind_t ind, bool zero, tcache_t *tcache,
+    bool is_metadata, arena_t *arena, bool slow_path)
 {
 	void *ret;
 
 	assert(size != 0);
+	assert(!is_metadata || tcache == NULL);
+	assert(!is_metadata || arena == NULL || arena->ind < narenas_auto);
 
-	ret = arena_malloc(tsd, arena, size, zero, tcache);
+	ret = arena_malloc(tsdn, arena, size, ind, zero, tcache, slow_path);
 	if (config_stats && is_metadata && likely(ret != NULL)) {
-		arena_metadata_allocated_add(iaalloc(ret), isalloc(ret,
-		    config_prof));
+		arena_metadata_allocated_add(iaalloc(ret),
+		    isalloc(tsdn, ret, config_prof));
 	}
 	return (ret);
 }
 
 JEMALLOC_ALWAYS_INLINE void *
-imalloct(tsd_t *tsd, size_t size, tcache_t *tcache, arena_t *arena)
+ialloc(tsd_t *tsd, size_t size, szind_t ind, bool zero, bool slow_path)
 {
 
-	return (iallocztm(tsd, size, false, tcache, false, arena));
+	return (iallocztm(tsd_tsdn(tsd), size, ind, zero, tcache_get(tsd, true),
+	    false, NULL, slow_path));
 }
 
 JEMALLOC_ALWAYS_INLINE void *
-imalloc(tsd_t *tsd, size_t size)
-{
-
-	return (iallocztm(tsd, size, false, tcache_get(tsd, true), false, NULL));
-}
-
-JEMALLOC_ALWAYS_INLINE void *
-icalloct(tsd_t *tsd, size_t size, tcache_t *tcache, arena_t *arena)
-{
-
-	return (iallocztm(tsd, size, true, tcache, false, arena));
-}
-
-JEMALLOC_ALWAYS_INLINE void *
-icalloc(tsd_t *tsd, size_t size)
-{
-
-	return (iallocztm(tsd, size, true, tcache_get(tsd, true), false, NULL));
-}
-
-JEMALLOC_ALWAYS_INLINE void *
-ipallocztm(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+ipallocztm(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
     tcache_t *tcache, bool is_metadata, arena_t *arena)
 {
 	void *ret;
 
 	assert(usize != 0);
 	assert(usize == sa2u(usize, alignment));
+	assert(!is_metadata || tcache == NULL);
+	assert(!is_metadata || arena == NULL || arena->ind < narenas_auto);
 
-	ret = arena_palloc(tsd, arena, usize, alignment, zero, tcache);
+	ret = arena_palloc(tsdn, arena, usize, alignment, zero, tcache);
 	assert(ALIGNMENT_ADDR2BASE(ret, alignment) == ret);
 	if (config_stats && is_metadata && likely(ret != NULL)) {
-		arena_metadata_allocated_add(iaalloc(ret), isalloc(ret,
+		arena_metadata_allocated_add(iaalloc(ret), isalloc(tsdn, ret,
 		    config_prof));
 	}
 	return (ret);
 }
 
 JEMALLOC_ALWAYS_INLINE void *
-ipalloct(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+ipalloct(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
     tcache_t *tcache, arena_t *arena)
 {
 
-	return (ipallocztm(tsd, usize, alignment, zero, tcache, false, arena));
+	return (ipallocztm(tsdn, usize, alignment, zero, tcache, false, arena));
 }
 
 JEMALLOC_ALWAYS_INLINE void *
 ipalloc(tsd_t *tsd, size_t usize, size_t alignment, bool zero)
 {
 
-	return (ipallocztm(tsd, usize, alignment, zero, tcache_get(tsd,
-	    NULL), false, NULL));
+	return (ipallocztm(tsd_tsdn(tsd), usize, alignment, zero,
+	    tcache_get(tsd, true), false, NULL));
 }
 
 JEMALLOC_ALWAYS_INLINE size_t
-ivsalloc(const void *ptr, bool demote)
+ivsalloc(tsdn_t *tsdn, const void *ptr, bool demote)
 {
 	extent_node_t *node;
 
@@ -967,7 +1123,7 @@ ivsalloc(const void *ptr, bool demote)
 	assert(extent_node_addr_get(node) == ptr ||
 	    extent_node_achunk_get(node));
 
-	return (isalloc(ptr, demote));
+	return (isalloc(tsdn, ptr, demote));
 }
 
 JEMALLOC_INLINE size_t
@@ -985,65 +1141,62 @@ u2rz(size_t usize)
 }
 
 JEMALLOC_INLINE size_t
-p2rz(const void *ptr)
+p2rz(tsdn_t *tsdn, const void *ptr)
 {
-	size_t usize = isalloc(ptr, false);
+	size_t usize = isalloc(tsdn, ptr, false);
 
 	return (u2rz(usize));
 }
 
 JEMALLOC_ALWAYS_INLINE void
-idalloctm(tsd_t *tsd, void *ptr, tcache_t *tcache, bool is_metadata)
+idalloctm(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool is_metadata,
+    bool slow_path)
 {
 
 	assert(ptr != NULL);
+	assert(!is_metadata || tcache == NULL);
+	assert(!is_metadata || iaalloc(ptr)->ind < narenas_auto);
 	if (config_stats && is_metadata) {
-		arena_metadata_allocated_sub(iaalloc(ptr), isalloc(ptr,
+		arena_metadata_allocated_sub(iaalloc(ptr), isalloc(tsdn, ptr,
 		    config_prof));
 	}
 
-	arena_dalloc(tsd, ptr, tcache);
-}
-
-JEMALLOC_ALWAYS_INLINE void
-idalloct(tsd_t *tsd, void *ptr, tcache_t *tcache)
-{
-
-	idalloctm(tsd, ptr, tcache, false);
+	arena_dalloc(tsdn, ptr, tcache, slow_path);
 }
 
 JEMALLOC_ALWAYS_INLINE void
 idalloc(tsd_t *tsd, void *ptr)
 {
 
-	idalloctm(tsd, ptr, tcache_get(tsd, false), false);
+	idalloctm(tsd_tsdn(tsd), ptr, tcache_get(tsd, false), false, true);
 }
 
 JEMALLOC_ALWAYS_INLINE void
-iqalloc(tsd_t *tsd, void *ptr, tcache_t *tcache)
+iqalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path)
 {
 
-	if (config_fill && unlikely(opt_quarantine))
+	if (slow_path && config_fill && unlikely(opt_quarantine))
 		quarantine(tsd, ptr);
 	else
-		idalloctm(tsd, ptr, tcache, false);
+		idalloctm(tsd_tsdn(tsd), ptr, tcache, false, slow_path);
 }
 
 JEMALLOC_ALWAYS_INLINE void
-isdalloct(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
+isdalloct(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path)
 {
 
-	arena_sdalloc(tsd, ptr, size, tcache);
+	arena_sdalloc(tsdn, ptr, size, tcache, slow_path);
 }
 
 JEMALLOC_ALWAYS_INLINE void
-isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
+isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache, bool slow_path)
 {
 
-	if (config_fill && unlikely(opt_quarantine))
+	if (slow_path && config_fill && unlikely(opt_quarantine))
 		quarantine(tsd, ptr);
 	else
-		isdalloct(tsd, ptr, size, tcache);
+		isdalloct(tsd_tsdn(tsd), ptr, size, tcache, slow_path);
 }
 
 JEMALLOC_ALWAYS_INLINE void *
@@ -1054,17 +1207,18 @@ iralloct_realign(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
 	size_t usize, copysize;
 
 	usize = sa2u(size + extra, alignment);
-	if (usize == 0)
+	if (unlikely(usize == 0 || usize > HUGE_MAXCLASS))
 		return (NULL);
-	p = ipalloct(tsd, usize, alignment, zero, tcache, arena);
+	p = ipalloct(tsd_tsdn(tsd), usize, alignment, zero, tcache, arena);
 	if (p == NULL) {
 		if (extra == 0)
 			return (NULL);
 		/* Try again, without extra this time. */
 		usize = sa2u(size, alignment);
-		if (usize == 0)
+		if (unlikely(usize == 0 || usize > HUGE_MAXCLASS))
 			return (NULL);
-		p = ipalloct(tsd, usize, alignment, zero, tcache, arena);
+		p = ipalloct(tsd_tsdn(tsd), usize, alignment, zero, tcache,
+		    arena);
 		if (p == NULL)
 			return (NULL);
 	}
@@ -1074,7 +1228,7 @@ iralloct_realign(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
 	 */
 	copysize = (size < oldsize) ? size : oldsize;
 	memcpy(p, ptr, copysize);
-	isqalloc(tsd, ptr, oldsize, tcache);
+	isqalloc(tsd, ptr, oldsize, tcache, true);
 	return (p);
 }
 
@@ -1110,8 +1264,8 @@ iralloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size, size_t alignment,
 }
 
 JEMALLOC_ALWAYS_INLINE bool
-ixalloc(void *ptr, size_t oldsize, size_t size, size_t extra, size_t alignment,
-    bool zero)
+ixalloc(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t extra,
+    size_t alignment, bool zero)
 {
 
 	assert(ptr != NULL);
@@ -1123,7 +1277,7 @@ ixalloc(void *ptr, size_t oldsize, size_t size, size_t extra, size_t alignment,
 		return (true);
 	}
 
-	return (arena_ralloc_no_move(ptr, oldsize, size, extra, zero));
+	return (arena_ralloc_no_move(tsdn, ptr, oldsize, size, extra, zero));
 }
 #endif
 
diff --git a/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_decls.h b/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_decls.h
index a601d6ebb..c907d9109 100644
--- a/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_decls.h
+++ b/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_decls.h
@@ -17,7 +17,18 @@
 #    include <sys/uio.h>
 #  endif
 #  include <pthread.h>
+#  ifdef JEMALLOC_OS_UNFAIR_LOCK
+#    include <os/lock.h>
+#  endif
+#  ifdef JEMALLOC_GLIBC_MALLOC_HOOK
+#    include <sched.h>
+#  endif
 #  include <errno.h>
+#  include <sys/time.h>
+#  include <time.h>
+#  ifdef JEMALLOC_HAVE_MACH_ABSOLUTE_TIME
+#    include <mach/mach_time.h>
+#  endif
 #endif
 #include <sys/types.h>
 
diff --git a/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_defs.h.in b/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_defs.h.in
index b0f8caaf8..def4ba550 100644
--- a/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_defs.h.in
+++ b/deps/jemalloc/include/jemalloc/internal/jemalloc_internal_defs.h.in
@@ -56,9 +56,9 @@
 #undef JEMALLOC_HAVE_BUILTIN_CLZ
 
 /*
- * Defined if madvise(2) is available.
+ * Defined if os_unfair_lock_*() functions are available, as provided by Darwin.
  */
-#undef JEMALLOC_HAVE_MADVISE
+#undef JEMALLOC_OS_UNFAIR_LOCK
 
 /*
  * Defined if OSSpin*() functions are available, as provided by Darwin, and
@@ -66,6 +66,9 @@
  */
 #undef JEMALLOC_OSSPIN
 
+/* Defined if syscall(2) is usable. */
+#undef JEMALLOC_USE_SYSCALL
+
 /*
  * Defined if secure_getenv(3) is available.
  */
@@ -76,6 +79,24 @@
  */
 #undef JEMALLOC_HAVE_ISSETUGID
 
+/* Defined if pthread_atfork(3) is available. */
+#undef JEMALLOC_HAVE_PTHREAD_ATFORK
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC_COARSE, ...) is available.
+ */
+#undef JEMALLOC_HAVE_CLOCK_MONOTONIC_COARSE
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC, ...) is available.
+ */
+#undef JEMALLOC_HAVE_CLOCK_MONOTONIC
+
+/*
+ * Defined if mach_absolute_time() is available.
+ */
+#undef JEMALLOC_HAVE_MACH_ABSOLUTE_TIME
+
 /*
  * Defined if _malloc_thread_cleanup() exists.  At least in the case of
  * FreeBSD, pthread_key_create() allocates, which if used during malloc
@@ -189,9 +210,16 @@
 #undef JEMALLOC_TLS
 
 /*
- * ffs()/ffsl() functions to use for bitmapping.  Don't use these directly;
- * instead, use jemalloc_ffs() or jemalloc_ffsl() from util.h.
+ * Used to mark unreachable code to quiet "end of non-void" compiler warnings.
+ * Don't use this directly; instead use unreachable() from util.h
  */
+#undef JEMALLOC_INTERNAL_UNREACHABLE
+
+/*
+ * ffs*() functions to use for bitmapping.  Don't use these directly; instead,
+ * use ffs_*() from util.h.
+ */
+#undef JEMALLOC_INTERNAL_FFSLL
 #undef JEMALLOC_INTERNAL_FFSL
 #undef JEMALLOC_INTERNAL_FFS
 
@@ -214,17 +242,34 @@
 #undef JEMALLOC_ZONE_VERSION
 
 /*
+ * Methods for determining whether the OS overcommits.
+ * JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY: Linux's
+ *                                         /proc/sys/vm.overcommit_memory file.
+ * JEMALLOC_SYSCTL_VM_OVERCOMMIT: FreeBSD's vm.overcommit sysctl.
+ */
+#undef JEMALLOC_SYSCTL_VM_OVERCOMMIT
+#undef JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY
+
+/* Defined if madvise(2) is available. */
+#undef JEMALLOC_HAVE_MADVISE
+
+/*
  * Methods for purging unused pages differ between operating systems.
  *
- *   madvise(..., MADV_DONTNEED) : On Linux, this immediately discards pages,
- *                                 such that new pages will be demand-zeroed if
- *                                 the address region is later touched.
- *   madvise(..., MADV_FREE) : On FreeBSD and Darwin, this marks pages as being
- *                             unused, such that they will be discarded rather
- *                             than swapped out.
+ *   madvise(..., MADV_FREE) : This marks pages as being unused, such that they
+ *                             will be discarded rather than swapped out.
+ *   madvise(..., MADV_DONTNEED) : This immediately discards pages, such that
+ *                                 new pages will be demand-zeroed if the
+ *                                 address region is later touched.
  */
-#undef JEMALLOC_PURGE_MADVISE_DONTNEED
 #undef JEMALLOC_PURGE_MADVISE_FREE
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED
+
+/*
+ * Defined if transparent huge pages are supported via the MADV_[NO]HUGEPAGE
+ * arguments to madvise(2).
+ */
+#undef JEMALLOC_THP
 
 /* Define if operating system has alloca.h header. */
 #undef JEMALLOC_HAS_ALLOCA_H
@@ -241,6 +286,9 @@
 /* sizeof(long) == 2^LG_SIZEOF_LONG. */
 #undef LG_SIZEOF_LONG
 
+/* sizeof(long long) == 2^LG_SIZEOF_LONG_LONG. */
+#undef LG_SIZEOF_LONG_LONG
+
 /* sizeof(intmax_t) == 2^LG_SIZEOF_INTMAX_T. */
 #undef LG_SIZEOF_INTMAX_T
 
@@ -259,4 +307,7 @@
  */
 #undef JEMALLOC_EXPORT
 
+/* config.malloc_conf options string. */
+#undef JEMALLOC_CONFIG_MALLOC_CONF
+
 #endif /* JEMALLOC_INTERNAL_DEFS_H_ */
diff --git a/deps/jemalloc/include/jemalloc/internal/mb.h b/deps/jemalloc/include/jemalloc/internal/mb.h
index 3cfa78729..5384728fd 100644
--- a/deps/jemalloc/include/jemalloc/internal/mb.h
+++ b/deps/jemalloc/include/jemalloc/internal/mb.h
@@ -42,7 +42,7 @@ mb_write(void)
 	    : /* Inputs. */
 	    : "memory" /* Clobbers. */
 	    );
-#else
+#  else
 	/*
 	 * This is hopefully enough to keep the compiler from reordering
 	 * instructions around this one.
@@ -52,7 +52,7 @@ mb_write(void)
 	    : /* Inputs. */
 	    : "memory" /* Clobbers. */
 	    );
-#endif
+#  endif
 }
 #elif (defined(__amd64__) || defined(__x86_64__))
 JEMALLOC_INLINE void
@@ -104,9 +104,9 @@ mb_write(void)
 {
 	malloc_mutex_t mtx;
 
-	malloc_mutex_init(&mtx);
-	malloc_mutex_lock(&mtx);
-	malloc_mutex_unlock(&mtx);
+	malloc_mutex_init(&mtx, "mb", WITNESS_RANK_OMIT);
+	malloc_mutex_lock(TSDN_NULL, &mtx);
+	malloc_mutex_unlock(TSDN_NULL, &mtx);
 }
 #endif
 #endif
diff --git a/deps/jemalloc/include/jemalloc/internal/mutex.h b/deps/jemalloc/include/jemalloc/internal/mutex.h
index f051f2917..b442d2d4e 100644
--- a/deps/jemalloc/include/jemalloc/internal/mutex.h
+++ b/deps/jemalloc/include/jemalloc/internal/mutex.h
@@ -5,18 +5,25 @@ typedef struct malloc_mutex_s malloc_mutex_t;
 
 #ifdef _WIN32
 #  define MALLOC_MUTEX_INITIALIZER
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+#  define MALLOC_MUTEX_INITIALIZER					\
+     {OS_UNFAIR_LOCK_INIT, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #elif (defined(JEMALLOC_OSSPIN))
-#  define MALLOC_MUTEX_INITIALIZER {0}
+#  define MALLOC_MUTEX_INITIALIZER {0, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #elif (defined(JEMALLOC_MUTEX_INIT_CB))
-#  define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER, NULL}
+#  define MALLOC_MUTEX_INITIALIZER					\
+    {PTHREAD_MUTEX_INITIALIZER, NULL, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #else
 #  if (defined(JEMALLOC_HAVE_PTHREAD_MUTEX_ADAPTIVE_NP) &&		\
        defined(PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP))
 #    define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_ADAPTIVE_NP
-#    define MALLOC_MUTEX_INITIALIZER {PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP}
+#    define MALLOC_MUTEX_INITIALIZER					\
+       {PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP,				\
+        WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #  else
 #    define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_DEFAULT
-#    define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER}
+#    define MALLOC_MUTEX_INITIALIZER					\
+       {PTHREAD_MUTEX_INITIALIZER, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #  endif
 #endif
 
@@ -31,6 +38,8 @@ struct malloc_mutex_s {
 #  else
 	CRITICAL_SECTION	lock;
 #  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+	os_unfair_lock		lock;
 #elif (defined(JEMALLOC_OSSPIN))
 	OSSpinLock		lock;
 #elif (defined(JEMALLOC_MUTEX_INIT_CB))
@@ -39,6 +48,7 @@ struct malloc_mutex_s {
 #else
 	pthread_mutex_t		lock;
 #endif
+	witness_t		witness;
 };
 
 #endif /* JEMALLOC_H_STRUCTS */
@@ -52,52 +62,62 @@ extern bool isthreaded;
 #  define isthreaded true
 #endif
 
-bool	malloc_mutex_init(malloc_mutex_t *mutex);
-void	malloc_mutex_prefork(malloc_mutex_t *mutex);
-void	malloc_mutex_postfork_parent(malloc_mutex_t *mutex);
-void	malloc_mutex_postfork_child(malloc_mutex_t *mutex);
-bool	mutex_boot(void);
+bool	malloc_mutex_init(malloc_mutex_t *mutex, const char *name,
+    witness_rank_t rank);
+void	malloc_mutex_prefork(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_postfork_parent(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_postfork_child(tsdn_t *tsdn, malloc_mutex_t *mutex);
+bool	malloc_mutex_boot(void);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
 #ifdef JEMALLOC_H_INLINES
 
 #ifndef JEMALLOC_ENABLE_INLINE
-void	malloc_mutex_lock(malloc_mutex_t *mutex);
-void	malloc_mutex_unlock(malloc_mutex_t *mutex);
+void	malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_unlock(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_assert_owner(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_assert_not_owner(tsdn_t *tsdn, malloc_mutex_t *mutex);
 #endif
 
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
 JEMALLOC_INLINE void
-malloc_mutex_lock(malloc_mutex_t *mutex)
+malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex)
 {
 
 	if (isthreaded) {
+		witness_assert_not_owner(tsdn, &mutex->witness);
 #ifdef _WIN32
 #  if _WIN32_WINNT >= 0x0600
 		AcquireSRWLockExclusive(&mutex->lock);
 #  else
 		EnterCriticalSection(&mutex->lock);
 #  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+		os_unfair_lock_lock(&mutex->lock);
 #elif (defined(JEMALLOC_OSSPIN))
 		OSSpinLockLock(&mutex->lock);
 #else
 		pthread_mutex_lock(&mutex->lock);
 #endif
+		witness_lock(tsdn, &mutex->witness);
 	}
 }
 
 JEMALLOC_INLINE void
-malloc_mutex_unlock(malloc_mutex_t *mutex)
+malloc_mutex_unlock(tsdn_t *tsdn, malloc_mutex_t *mutex)
 {
 
 	if (isthreaded) {
+		witness_unlock(tsdn, &mutex->witness);
 #ifdef _WIN32
 #  if _WIN32_WINNT >= 0x0600
 		ReleaseSRWLockExclusive(&mutex->lock);
 #  else
 		LeaveCriticalSection(&mutex->lock);
 #  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+		os_unfair_lock_unlock(&mutex->lock);
 #elif (defined(JEMALLOC_OSSPIN))
 		OSSpinLockUnlock(&mutex->lock);
 #else
@@ -105,6 +125,22 @@ malloc_mutex_unlock(malloc_mutex_t *mutex)
 #endif
 	}
 }
+
+JEMALLOC_INLINE void
+malloc_mutex_assert_owner(tsdn_t *tsdn, malloc_mutex_t *mutex)
+{
+
+	if (isthreaded)
+		witness_assert_owner(tsdn, &mutex->witness);
+}
+
+JEMALLOC_INLINE void
+malloc_mutex_assert_not_owner(tsdn_t *tsdn, malloc_mutex_t *mutex)
+{
+
+	if (isthreaded)
+		witness_assert_not_owner(tsdn, &mutex->witness);
+}
 #endif
 
 #endif /* JEMALLOC_H_INLINES */
diff --git a/deps/jemalloc/include/jemalloc/internal/nstime.h b/deps/jemalloc/include/jemalloc/internal/nstime.h
new file mode 100644
index 000000000..93b27dc80
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/nstime.h
@@ -0,0 +1,48 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct nstime_s nstime_t;
+
+/* Maximum supported number of seconds (~584 years). */
+#define	NSTIME_SEC_MAX	KQU(18446744072)
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct nstime_s {
+	uint64_t	ns;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+void	nstime_init(nstime_t *time, uint64_t ns);
+void	nstime_init2(nstime_t *time, uint64_t sec, uint64_t nsec);
+uint64_t	nstime_ns(const nstime_t *time);
+uint64_t	nstime_sec(const nstime_t *time);
+uint64_t	nstime_nsec(const nstime_t *time);
+void	nstime_copy(nstime_t *time, const nstime_t *source);
+int	nstime_compare(const nstime_t *a, const nstime_t *b);
+void	nstime_add(nstime_t *time, const nstime_t *addend);
+void	nstime_subtract(nstime_t *time, const nstime_t *subtrahend);
+void	nstime_imultiply(nstime_t *time, uint64_t multiplier);
+void	nstime_idivide(nstime_t *time, uint64_t divisor);
+uint64_t	nstime_divide(const nstime_t *time, const nstime_t *divisor);
+#ifdef JEMALLOC_JET
+typedef bool (nstime_monotonic_t)(void);
+extern nstime_monotonic_t *nstime_monotonic;
+typedef bool (nstime_update_t)(nstime_t *);
+extern nstime_update_t *nstime_update;
+#else
+bool	nstime_monotonic(void);
+bool	nstime_update(nstime_t *time);
+#endif
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/pages.h b/deps/jemalloc/include/jemalloc/internal/pages.h
index da7eb9686..4ae9f156a 100644
--- a/deps/jemalloc/include/jemalloc/internal/pages.h
+++ b/deps/jemalloc/include/jemalloc/internal/pages.h
@@ -9,13 +9,16 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS
 
-void	*pages_map(void *addr, size_t size);
+void	*pages_map(void *addr, size_t size, bool *commit);
 void	pages_unmap(void *addr, size_t size);
 void	*pages_trim(void *addr, size_t alloc_size, size_t leadsize,
-    size_t size);
+    size_t size, bool *commit);
 bool	pages_commit(void *addr, size_t size);
 bool	pages_decommit(void *addr, size_t size);
 bool	pages_purge(void *addr, size_t size);
+bool	pages_huge(void *addr, size_t size);
+bool	pages_nohuge(void *addr, size_t size);
+void	pages_boot(void);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/ph.h b/deps/jemalloc/include/jemalloc/internal/ph.h
new file mode 100644
index 000000000..4f91c333f
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/ph.h
@@ -0,0 +1,345 @@
+/*
+ * A Pairing Heap implementation.
+ *
+ * "The Pairing Heap: A New Form of Self-Adjusting Heap"
+ * https://www.cs.cmu.edu/~sleator/papers/pairing-heaps.pdf
+ *
+ * With auxiliary twopass list, described in a follow on paper.
+ *
+ * "Pairing Heaps: Experiments and Analysis"
+ * http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.106.2988&rep=rep1&type=pdf
+ *
+ *******************************************************************************
+ */
+
+#ifndef PH_H_
+#define	PH_H_
+
+/* Node structure. */
+#define	phn(a_type)							\
+struct {								\
+	a_type	*phn_prev;						\
+	a_type	*phn_next;						\
+	a_type	*phn_lchild;						\
+}
+
+/* Root structure. */
+#define	ph(a_type)							\
+struct {								\
+	a_type	*ph_root;						\
+}
+
+/* Internal utility macros. */
+#define	phn_lchild_get(a_type, a_field, a_phn)				\
+	(a_phn->a_field.phn_lchild)
+#define	phn_lchild_set(a_type, a_field, a_phn, a_lchild) do {		\
+	a_phn->a_field.phn_lchild = a_lchild;				\
+} while (0)
+
+#define	phn_next_get(a_type, a_field, a_phn)				\
+	(a_phn->a_field.phn_next)
+#define	phn_prev_set(a_type, a_field, a_phn, a_prev) do {		\
+	a_phn->a_field.phn_prev = a_prev;				\
+} while (0)
+
+#define	phn_prev_get(a_type, a_field, a_phn)				\
+	(a_phn->a_field.phn_prev)
+#define	phn_next_set(a_type, a_field, a_phn, a_next) do {		\
+	a_phn->a_field.phn_next = a_next;				\
+} while (0)
+
+#define	phn_merge_ordered(a_type, a_field, a_phn0, a_phn1, a_cmp) do {	\
+	a_type *phn0child;						\
+									\
+	assert(a_phn0 != NULL);						\
+	assert(a_phn1 != NULL);						\
+	assert(a_cmp(a_phn0, a_phn1) <= 0);				\
+									\
+	phn_prev_set(a_type, a_field, a_phn1, a_phn0);			\
+	phn0child = phn_lchild_get(a_type, a_field, a_phn0);		\
+	phn_next_set(a_type, a_field, a_phn1, phn0child);		\
+	if (phn0child != NULL)						\
+		phn_prev_set(a_type, a_field, phn0child, a_phn1);	\
+	phn_lchild_set(a_type, a_field, a_phn0, a_phn1);		\
+} while (0)
+
+#define	phn_merge(a_type, a_field, a_phn0, a_phn1, a_cmp, r_phn) do {	\
+	if (a_phn0 == NULL)						\
+		r_phn = a_phn1;						\
+	else if (a_phn1 == NULL)					\
+		r_phn = a_phn0;						\
+	else if (a_cmp(a_phn0, a_phn1) < 0) {				\
+		phn_merge_ordered(a_type, a_field, a_phn0, a_phn1,	\
+		    a_cmp);						\
+		r_phn = a_phn0;						\
+	} else {							\
+		phn_merge_ordered(a_type, a_field, a_phn1, a_phn0,	\
+		    a_cmp);						\
+		r_phn = a_phn1;						\
+	}								\
+} while (0)
+
+#define	ph_merge_siblings(a_type, a_field, a_phn, a_cmp, r_phn) do {	\
+	a_type *head = NULL;						\
+	a_type *tail = NULL;						\
+	a_type *phn0 = a_phn;						\
+	a_type *phn1 = phn_next_get(a_type, a_field, phn0);		\
+									\
+	/*								\
+	 * Multipass merge, wherein the first two elements of a FIFO	\
+	 * are repeatedly merged, and each result is appended to the	\
+	 * singly linked FIFO, until the FIFO contains only a single	\
+	 * element.  We start with a sibling list but no reference to	\
+	 * its tail, so we do a single pass over the sibling list to	\
+	 * populate the FIFO.						\
+	 */								\
+	if (phn1 != NULL) {						\
+		a_type *phnrest = phn_next_get(a_type, a_field, phn1);	\
+		if (phnrest != NULL)					\
+			phn_prev_set(a_type, a_field, phnrest, NULL);	\
+		phn_prev_set(a_type, a_field, phn0, NULL);		\
+		phn_next_set(a_type, a_field, phn0, NULL);		\
+		phn_prev_set(a_type, a_field, phn1, NULL);		\
+		phn_next_set(a_type, a_field, phn1, NULL);		\
+		phn_merge(a_type, a_field, phn0, phn1, a_cmp, phn0);	\
+		head = tail = phn0;					\
+		phn0 = phnrest;						\
+		while (phn0 != NULL) {					\
+			phn1 = phn_next_get(a_type, a_field, phn0);	\
+			if (phn1 != NULL) {				\
+				phnrest = phn_next_get(a_type, a_field,	\
+				    phn1);				\
+				if (phnrest != NULL) {			\
+					phn_prev_set(a_type, a_field,	\
+					    phnrest, NULL);		\
+				}					\
+				phn_prev_set(a_type, a_field, phn0,	\
+				    NULL);				\
+				phn_next_set(a_type, a_field, phn0,	\
+				    NULL);				\
+				phn_prev_set(a_type, a_field, phn1,	\
+				    NULL);				\
+				phn_next_set(a_type, a_field, phn1,	\
+				    NULL);				\
+				phn_merge(a_type, a_field, phn0, phn1,	\
+				    a_cmp, phn0);			\
+				phn_next_set(a_type, a_field, tail,	\
+				    phn0);				\
+				tail = phn0;				\
+				phn0 = phnrest;				\
+			} else {					\
+				phn_next_set(a_type, a_field, tail,	\
+				    phn0);				\
+				tail = phn0;				\
+				phn0 = NULL;				\
+			}						\
+		}							\
+		phn0 = head;						\
+		phn1 = phn_next_get(a_type, a_field, phn0);		\
+		if (phn1 != NULL) {					\
+			while (true) {					\
+				head = phn_next_get(a_type, a_field,	\
+				    phn1);				\
+				assert(phn_prev_get(a_type, a_field,	\
+				    phn0) == NULL);			\
+				phn_next_set(a_type, a_field, phn0,	\
+				    NULL);				\
+				assert(phn_prev_get(a_type, a_field,	\
+				    phn1) == NULL);			\
+				phn_next_set(a_type, a_field, phn1,	\
+				    NULL);				\
+				phn_merge(a_type, a_field, phn0, phn1,	\
+				    a_cmp, phn0);			\
+				if (head == NULL)			\
+					break;				\
+				phn_next_set(a_type, a_field, tail,	\
+				    phn0);				\
+				tail = phn0;				\
+				phn0 = head;				\
+				phn1 = phn_next_get(a_type, a_field,	\
+				    phn0);				\
+			}						\
+		}							\
+	}								\
+	r_phn = phn0;							\
+} while (0)
+
+#define	ph_merge_aux(a_type, a_field, a_ph, a_cmp) do {			\
+	a_type *phn = phn_next_get(a_type, a_field, a_ph->ph_root);	\
+	if (phn != NULL) {						\
+		phn_prev_set(a_type, a_field, a_ph->ph_root, NULL);	\
+		phn_next_set(a_type, a_field, a_ph->ph_root, NULL);	\
+		phn_prev_set(a_type, a_field, phn, NULL);		\
+		ph_merge_siblings(a_type, a_field, phn, a_cmp, phn);	\
+		assert(phn_next_get(a_type, a_field, phn) == NULL);	\
+		phn_merge(a_type, a_field, a_ph->ph_root, phn, a_cmp,	\
+		    a_ph->ph_root);					\
+	}								\
+} while (0)
+
+#define	ph_merge_children(a_type, a_field, a_phn, a_cmp, r_phn) do {	\
+	a_type *lchild = phn_lchild_get(a_type, a_field, a_phn);	\
+	if (lchild == NULL)						\
+		r_phn = NULL;						\
+	else {								\
+		ph_merge_siblings(a_type, a_field, lchild, a_cmp,	\
+		    r_phn);						\
+	}								\
+} while (0)
+
+/*
+ * The ph_proto() macro generates function prototypes that correspond to the
+ * functions generated by an equivalently parameterized call to ph_gen().
+ */
+#define	ph_proto(a_attr, a_prefix, a_ph_type, a_type)			\
+a_attr void	a_prefix##new(a_ph_type *ph);				\
+a_attr bool	a_prefix##empty(a_ph_type *ph);				\
+a_attr a_type	*a_prefix##first(a_ph_type *ph);			\
+a_attr void	a_prefix##insert(a_ph_type *ph, a_type *phn);		\
+a_attr a_type	*a_prefix##remove_first(a_ph_type *ph);			\
+a_attr void	a_prefix##remove(a_ph_type *ph, a_type *phn);
+
+/*
+ * The ph_gen() macro generates a type-specific pairing heap implementation,
+ * based on the above cpp macros.
+ */
+#define	ph_gen(a_attr, a_prefix, a_ph_type, a_type, a_field, a_cmp)	\
+a_attr void								\
+a_prefix##new(a_ph_type *ph)						\
+{									\
+									\
+	memset(ph, 0, sizeof(ph(a_type)));				\
+}									\
+a_attr bool								\
+a_prefix##empty(a_ph_type *ph)						\
+{									\
+									\
+	return (ph->ph_root == NULL);					\
+}									\
+a_attr a_type *								\
+a_prefix##first(a_ph_type *ph)						\
+{									\
+									\
+	if (ph->ph_root == NULL)					\
+		return (NULL);						\
+	ph_merge_aux(a_type, a_field, ph, a_cmp);			\
+	return (ph->ph_root);						\
+}									\
+a_attr void								\
+a_prefix##insert(a_ph_type *ph, a_type *phn)				\
+{									\
+									\
+	memset(&phn->a_field, 0, sizeof(phn(a_type)));			\
+									\
+	/*								\
+	 * Treat the root as an aux list during insertion, and lazily	\
+	 * merge during a_prefix##remove_first().  For elements that	\
+	 * are inserted, then removed via a_prefix##remove() before the	\
+	 * aux list is ever processed, this makes insert/remove		\
+	 * constant-time, whereas eager merging would make insert	\
+	 * O(log n).							\
+	 */								\
+	if (ph->ph_root == NULL)					\
+		ph->ph_root = phn;					\
+	else {								\
+		phn_next_set(a_type, a_field, phn, phn_next_get(a_type,	\
+		    a_field, ph->ph_root));				\
+		if (phn_next_get(a_type, a_field, ph->ph_root) !=	\
+		    NULL) {						\
+			phn_prev_set(a_type, a_field,			\
+			    phn_next_get(a_type, a_field, ph->ph_root),	\
+			    phn);					\
+		}							\
+		phn_prev_set(a_type, a_field, phn, ph->ph_root);	\
+		phn_next_set(a_type, a_field, ph->ph_root, phn);	\
+	}								\
+}									\
+a_attr a_type *								\
+a_prefix##remove_first(a_ph_type *ph)					\
+{									\
+	a_type *ret;							\
+									\
+	if (ph->ph_root == NULL)					\
+		return (NULL);						\
+	ph_merge_aux(a_type, a_field, ph, a_cmp);			\
+									\
+	ret = ph->ph_root;						\
+									\
+	ph_merge_children(a_type, a_field, ph->ph_root, a_cmp,		\
+	    ph->ph_root);						\
+									\
+	return (ret);							\
+}									\
+a_attr void								\
+a_prefix##remove(a_ph_type *ph, a_type *phn)				\
+{									\
+	a_type *replace, *parent;					\
+									\
+	/*								\
+	 * We can delete from aux list without merging it, but we need	\
+	 * to merge if we are dealing with the root node.		\
+	 */								\
+	if (ph->ph_root == phn) {					\
+		ph_merge_aux(a_type, a_field, ph, a_cmp);		\
+		if (ph->ph_root == phn) {				\
+			ph_merge_children(a_type, a_field, ph->ph_root,	\
+			    a_cmp, ph->ph_root);			\
+			return;						\
+		}							\
+	}								\
+									\
+	/* Get parent (if phn is leftmost child) before mutating. */	\
+	if ((parent = phn_prev_get(a_type, a_field, phn)) != NULL) {	\
+		if (phn_lchild_get(a_type, a_field, parent) != phn)	\
+			parent = NULL;					\
+	}								\
+	/* Find a possible replacement node, and link to parent. */	\
+	ph_merge_children(a_type, a_field, phn, a_cmp, replace);	\
+	/* Set next/prev for sibling linked list. */			\
+	if (replace != NULL) {						\
+		if (parent != NULL) {					\
+			phn_prev_set(a_type, a_field, replace, parent);	\
+			phn_lchild_set(a_type, a_field, parent,		\
+			    replace);					\
+		} else {						\
+			phn_prev_set(a_type, a_field, replace,		\
+			    phn_prev_get(a_type, a_field, phn));	\
+			if (phn_prev_get(a_type, a_field, phn) !=	\
+			    NULL) {					\
+				phn_next_set(a_type, a_field,		\
+				    phn_prev_get(a_type, a_field, phn),	\
+				    replace);				\
+			}						\
+		}							\
+		phn_next_set(a_type, a_field, replace,			\
+		    phn_next_get(a_type, a_field, phn));		\
+		if (phn_next_get(a_type, a_field, phn) != NULL) {	\
+			phn_prev_set(a_type, a_field,			\
+			    phn_next_get(a_type, a_field, phn),		\
+			    replace);					\
+		}							\
+	} else {							\
+		if (parent != NULL) {					\
+			a_type *next = phn_next_get(a_type, a_field,	\
+			    phn);					\
+			phn_lchild_set(a_type, a_field, parent, next);	\
+			if (next != NULL) {				\
+				phn_prev_set(a_type, a_field, next,	\
+				    parent);				\
+			}						\
+		} else {						\
+			assert(phn_prev_get(a_type, a_field, phn) !=	\
+			    NULL);					\
+			phn_next_set(a_type, a_field,			\
+			    phn_prev_get(a_type, a_field, phn),		\
+			    phn_next_get(a_type, a_field, phn));	\
+		}							\
+		if (phn_next_get(a_type, a_field, phn) != NULL) {	\
+			phn_prev_set(a_type, a_field,			\
+			    phn_next_get(a_type, a_field, phn),		\
+			    phn_prev_get(a_type, a_field, phn));	\
+		}							\
+	}								\
+}
+
+#endif /* PH_H_ */
diff --git a/deps/jemalloc/include/jemalloc/internal/private_symbols.txt b/deps/jemalloc/include/jemalloc/internal/private_symbols.txt
index a90021aa6..c1c6c4090 100644
--- a/deps/jemalloc/include/jemalloc/internal/private_symbols.txt
+++ b/deps/jemalloc/include/jemalloc/internal/private_symbols.txt
@@ -3,12 +3,15 @@ a0get
 a0malloc
 arena_aalloc
 arena_alloc_junk_small
+arena_basic_stats_merge
 arena_bin_index
 arena_bin_info
-arena_bitselm_get
+arena_bitselm_get_const
+arena_bitselm_get_mutable
 arena_boot
 arena_choose
 arena_choose_hard
+arena_choose_impl
 arena_chunk_alloc_huge
 arena_chunk_cache_maybe_insert
 arena_chunk_cache_maybe_remove
@@ -25,18 +28,25 @@ arena_dalloc_junk_small
 arena_dalloc_large
 arena_dalloc_large_junked_locked
 arena_dalloc_small
+arena_decay_tick
+arena_decay_ticks
+arena_decay_time_default_get
+arena_decay_time_default_set
+arena_decay_time_get
+arena_decay_time_set
 arena_dss_prec_get
 arena_dss_prec_set
+arena_extent_sn_next
 arena_get
-arena_get_hard
+arena_ichoose
 arena_init
 arena_lg_dirty_mult_default_get
 arena_lg_dirty_mult_default_set
 arena_lg_dirty_mult_get
 arena_lg_dirty_mult_set
 arena_malloc
+arena_malloc_hard
 arena_malloc_large
-arena_malloc_small
 arena_mapbits_allocated_get
 arena_mapbits_binind_get
 arena_mapbits_decommitted_get
@@ -47,9 +57,6 @@ arena_mapbits_large_binind_set
 arena_mapbits_large_get
 arena_mapbits_large_set
 arena_mapbits_large_size_get
-arena_mapbitsp_get
-arena_mapbitsp_read
-arena_mapbitsp_write
 arena_mapbits_size_decode
 arena_mapbits_size_encode
 arena_mapbits_small_runind_get
@@ -58,23 +65,33 @@ arena_mapbits_unallocated_set
 arena_mapbits_unallocated_size_get
 arena_mapbits_unallocated_size_set
 arena_mapbits_unzeroed_get
+arena_mapbitsp_get_const
+arena_mapbitsp_get_mutable
+arena_mapbitsp_read
+arena_mapbitsp_write
 arena_maxrun
 arena_maybe_purge
 arena_metadata_allocated_add
 arena_metadata_allocated_get
 arena_metadata_allocated_sub
 arena_migrate
-arena_miscelm_get
+arena_miscelm_get_const
+arena_miscelm_get_mutable
 arena_miscelm_to_pageind
 arena_miscelm_to_rpages
-arena_nbound
 arena_new
 arena_node_alloc
 arena_node_dalloc
+arena_nthreads_dec
+arena_nthreads_get
+arena_nthreads_inc
 arena_palloc
 arena_postfork_child
 arena_postfork_parent
-arena_prefork
+arena_prefork0
+arena_prefork1
+arena_prefork2
+arena_prefork3
 arena_prof_accum
 arena_prof_accum_impl
 arena_prof_accum_locked
@@ -83,21 +100,25 @@ arena_prof_tctx_get
 arena_prof_tctx_reset
 arena_prof_tctx_set
 arena_ptr_small_binind_get
-arena_purge_all
+arena_purge
 arena_quarantine_junk_small
 arena_ralloc
 arena_ralloc_junk_large
 arena_ralloc_no_move
 arena_rd_to_miscelm
 arena_redzone_corruption
+arena_reset
 arena_run_regind
 arena_run_to_miscelm
 arena_salloc
-arenas_cache_bypass_cleanup
-arenas_cache_cleanup
 arena_sdalloc
 arena_stats_merge
 arena_tcache_fill_small
+arena_tdata_get
+arena_tdata_get_hard
+arenas
+arenas_tdata_bypass_cleanup
+arenas_tdata_cleanup
 atomic_add_p
 atomic_add_u
 atomic_add_uint32
@@ -113,6 +134,11 @@ atomic_sub_u
 atomic_sub_uint32
 atomic_sub_uint64
 atomic_sub_z
+atomic_write_p
+atomic_write_u
+atomic_write_uint32
+atomic_write_uint64
+atomic_write_z
 base_alloc
 base_boot
 base_postfork_child
@@ -122,7 +148,6 @@ base_stats_get
 bitmap_full
 bitmap_get
 bitmap_info_init
-bitmap_info_ngroups
 bitmap_init
 bitmap_set
 bitmap_sfu
@@ -139,32 +164,25 @@ chunk_alloc_dss
 chunk_alloc_mmap
 chunk_alloc_wrapper
 chunk_boot
-chunk_dalloc_arena
 chunk_dalloc_cache
 chunk_dalloc_mmap
 chunk_dalloc_wrapper
 chunk_deregister
 chunk_dss_boot
-chunk_dss_postfork_child
-chunk_dss_postfork_parent
+chunk_dss_mergeable
 chunk_dss_prec_get
 chunk_dss_prec_set
-chunk_dss_prefork
 chunk_hooks_default
 chunk_hooks_get
 chunk_hooks_set
 chunk_in_dss
 chunk_lookup
 chunk_npages
-chunk_postfork_child
-chunk_postfork_parent
-chunk_prefork
-chunk_purge_arena
 chunk_purge_wrapper
 chunk_register
+chunks_rtree
 chunksize
 chunksize_mask
-chunks_rtree
 ckh_count
 ckh_delete
 ckh_insert
@@ -183,6 +201,7 @@ ctl_nametomib
 ctl_postfork_child
 ctl_postfork_parent
 ctl_prefork
+decay_ticker_get
 dss_prec_names
 extent_node_achunk_get
 extent_node_achunk_set
@@ -190,6 +209,8 @@ extent_node_addr_get
 extent_node_addr_set
 extent_node_arena_get
 extent_node_arena_set
+extent_node_committed_get
+extent_node_committed_set
 extent_node_dirty_insert
 extent_node_dirty_linkage_init
 extent_node_dirty_remove
@@ -198,8 +219,12 @@ extent_node_prof_tctx_get
 extent_node_prof_tctx_set
 extent_node_size_get
 extent_node_size_set
+extent_node_sn_get
+extent_node_sn_set
 extent_node_zeroed_get
 extent_node_zeroed_set
+extent_tree_ad_destroy
+extent_tree_ad_destroy_recurse
 extent_tree_ad_empty
 extent_tree_ad_first
 extent_tree_ad_insert
@@ -217,23 +242,31 @@ extent_tree_ad_reverse_iter
 extent_tree_ad_reverse_iter_recurse
 extent_tree_ad_reverse_iter_start
 extent_tree_ad_search
-extent_tree_szad_empty
-extent_tree_szad_first
-extent_tree_szad_insert
-extent_tree_szad_iter
-extent_tree_szad_iter_recurse
-extent_tree_szad_iter_start
-extent_tree_szad_last
-extent_tree_szad_new
-extent_tree_szad_next
-extent_tree_szad_nsearch
-extent_tree_szad_prev
-extent_tree_szad_psearch
-extent_tree_szad_remove
-extent_tree_szad_reverse_iter
-extent_tree_szad_reverse_iter_recurse
-extent_tree_szad_reverse_iter_start
-extent_tree_szad_search
+extent_tree_szsnad_destroy
+extent_tree_szsnad_destroy_recurse
+extent_tree_szsnad_empty
+extent_tree_szsnad_first
+extent_tree_szsnad_insert
+extent_tree_szsnad_iter
+extent_tree_szsnad_iter_recurse
+extent_tree_szsnad_iter_start
+extent_tree_szsnad_last
+extent_tree_szsnad_new
+extent_tree_szsnad_next
+extent_tree_szsnad_nsearch
+extent_tree_szsnad_prev
+extent_tree_szsnad_psearch
+extent_tree_szsnad_remove
+extent_tree_szsnad_reverse_iter
+extent_tree_szsnad_reverse_iter_recurse
+extent_tree_szsnad_reverse_iter_start
+extent_tree_szsnad_search
+ffs_llu
+ffs_lu
+ffs_u
+ffs_u32
+ffs_u64
+ffs_zu
 get_errno
 hash
 hash_fmix_32
@@ -257,19 +290,16 @@ huge_ralloc
 huge_ralloc_no_move
 huge_salloc
 iaalloc
+ialloc
 iallocztm
-icalloc
-icalloct
+iarena_cleanup
 idalloc
-idalloct
 idalloctm
-imalloc
-imalloct
+in_valgrind
 index2size
 index2size_compute
 index2size_lookup
 index2size_tab
-in_valgrind
 ipalloc
 ipalloct
 ipallocztm
@@ -288,7 +318,11 @@ jemalloc_postfork_parent
 jemalloc_prefork
 large_maxclass
 lg_floor
+lg_prof_sample
 malloc_cprintf
+malloc_mutex_assert_not_owner
+malloc_mutex_assert_owner
+malloc_mutex_boot
 malloc_mutex_init
 malloc_mutex_lock
 malloc_mutex_postfork_child
@@ -310,12 +344,29 @@ malloc_write
 map_bias
 map_misc_offset
 mb_write
-mutex_boot
-narenas_cache_cleanup
+narenas_auto
+narenas_tdata_cleanup
 narenas_total_get
 ncpus
 nhbins
+nhclasses
+nlclasses
+nstime_add
+nstime_compare
+nstime_copy
+nstime_divide
+nstime_idivide
+nstime_imultiply
+nstime_init
+nstime_init2
+nstime_monotonic
+nstime_ns
+nstime_nsec
+nstime_sec
+nstime_subtract
+nstime_update
 opt_abort
+opt_decay_time
 opt_dss
 opt_junk
 opt_junk_alloc
@@ -334,6 +385,7 @@ opt_prof_gdump
 opt_prof_leak
 opt_prof_prefix
 opt_prof_thread_active_init
+opt_purge
 opt_quarantine
 opt_redzone
 opt_stats_print
@@ -342,13 +394,32 @@ opt_utrace
 opt_xmalloc
 opt_zero
 p2rz
+pages_boot
 pages_commit
 pages_decommit
+pages_huge
 pages_map
+pages_nohuge
 pages_purge
 pages_trim
 pages_unmap
-pow2_ceil
+pind2sz
+pind2sz_compute
+pind2sz_lookup
+pind2sz_tab
+pow2_ceil_u32
+pow2_ceil_u64
+pow2_ceil_zu
+prng_lg_range_u32
+prng_lg_range_u64
+prng_lg_range_zu
+prng_range_u32
+prng_range_u64
+prng_range_zu
+prng_state_next_u32
+prng_state_next_u64
+prng_state_next_zu
+prof_active
 prof_active_get
 prof_active_get_unlocked
 prof_active_set
@@ -358,6 +429,7 @@ prof_backtrace
 prof_boot0
 prof_boot1
 prof_boot2
+prof_bt_count
 prof_dump_header
 prof_dump_open
 prof_free
@@ -375,7 +447,8 @@ prof_malloc_sample_object
 prof_mdump
 prof_postfork_child
 prof_postfork_parent
-prof_prefork
+prof_prefork0
+prof_prefork1
 prof_realloc
 prof_reset
 prof_sample_accum_update
@@ -384,6 +457,7 @@ prof_tctx_get
 prof_tctx_reset
 prof_tctx_set
 prof_tdata_cleanup
+prof_tdata_count
 prof_tdata_get
 prof_tdata_init
 prof_tdata_reinit
@@ -393,11 +467,13 @@ prof_thread_active_init_set
 prof_thread_active_set
 prof_thread_name_get
 prof_thread_name_set
+psz2ind
+psz2u
+purge_mode_names
 quarantine
 quarantine_alloc_hook
 quarantine_alloc_hook_work
 quarantine_cleanup
-register_zone
 rtree_child_read
 rtree_child_read_hard
 rtree_child_tryread
@@ -413,6 +489,8 @@ rtree_subtree_read_hard
 rtree_subtree_tryread
 rtree_val_read
 rtree_val_write
+run_quantize_ceil
+run_quantize_floor
 s2u
 s2u_compute
 s2u_lookup
@@ -422,6 +500,8 @@ size2index
 size2index_compute
 size2index_lookup
 size2index_tab
+spin_adaptive
+spin_init
 stats_cactive
 stats_cactive_add
 stats_cactive_get
@@ -431,8 +511,6 @@ tcache_alloc_easy
 tcache_alloc_large
 tcache_alloc_small
 tcache_alloc_small_hard
-tcache_arena_associate
-tcache_arena_dissociate
 tcache_arena_reassociate
 tcache_bin_flush_large
 tcache_bin_flush_small
@@ -451,49 +529,103 @@ tcache_flush
 tcache_get
 tcache_get_hard
 tcache_maxclass
-tcaches
 tcache_salloc
+tcache_stats_merge
+tcaches
 tcaches_create
 tcaches_destroy
 tcaches_flush
 tcaches_get
-tcache_stats_merge
 thread_allocated_cleanup
 thread_deallocated_cleanup
+ticker_copy
+ticker_init
+ticker_read
+ticker_tick
+ticker_ticks
 tsd_arena_get
 tsd_arena_set
+tsd_arenap_get
+tsd_arenas_tdata_bypass_get
+tsd_arenas_tdata_bypass_set
+tsd_arenas_tdata_bypassp_get
+tsd_arenas_tdata_get
+tsd_arenas_tdata_set
+tsd_arenas_tdatap_get
 tsd_boot
 tsd_boot0
 tsd_boot1
 tsd_booted
+tsd_booted_get
 tsd_cleanup
 tsd_cleanup_wrapper
 tsd_fetch
+tsd_fetch_impl
 tsd_get
-tsd_wrapper_get
-tsd_wrapper_set
+tsd_get_allocates
+tsd_iarena_get
+tsd_iarena_set
+tsd_iarenap_get
 tsd_initialized
 tsd_init_check_recursion
 tsd_init_finish
 tsd_init_head
+tsd_narenas_tdata_get
+tsd_narenas_tdata_set
+tsd_narenas_tdatap_get
+tsd_wrapper_get
+tsd_wrapper_set
 tsd_nominal
+tsd_prof_tdata_get
+tsd_prof_tdata_set
+tsd_prof_tdatap_get
 tsd_quarantine_get
 tsd_quarantine_set
+tsd_quarantinep_get
 tsd_set
 tsd_tcache_enabled_get
 tsd_tcache_enabled_set
+tsd_tcache_enabledp_get
 tsd_tcache_get
 tsd_tcache_set
-tsd_tls
-tsd_tsd
-tsd_prof_tdata_get
-tsd_prof_tdata_set
+tsd_tcachep_get
 tsd_thread_allocated_get
 tsd_thread_allocated_set
+tsd_thread_allocatedp_get
 tsd_thread_deallocated_get
 tsd_thread_deallocated_set
+tsd_thread_deallocatedp_get
+tsd_tls
+tsd_tsd
+tsd_tsdn
+tsd_witness_fork_get
+tsd_witness_fork_set
+tsd_witness_forkp_get
+tsd_witnesses_get
+tsd_witnesses_set
+tsd_witnessesp_get
+tsdn_fetch
+tsdn_null
+tsdn_tsd
 u2rz
 valgrind_freelike_block
 valgrind_make_mem_defined
 valgrind_make_mem_noaccess
 valgrind_make_mem_undefined
+witness_assert_lockless
+witness_assert_not_owner
+witness_assert_owner
+witness_fork_cleanup
+witness_init
+witness_lock
+witness_lock_error
+witness_lockless_error
+witness_not_owner_error
+witness_owner
+witness_owner_error
+witness_postfork_child
+witness_postfork_parent
+witness_prefork
+witness_unlock
+witnesses_cleanup
+zone_register
diff --git a/deps/jemalloc/include/jemalloc/internal/prng.h b/deps/jemalloc/include/jemalloc/internal/prng.h
index 216d0ef47..c2bda19c6 100644
--- a/deps/jemalloc/include/jemalloc/internal/prng.h
+++ b/deps/jemalloc/include/jemalloc/internal/prng.h
@@ -18,31 +18,13 @@
  * proportional to bit position.  For example, the lowest bit has a cycle of 2,
  * the next has a cycle of 4, etc.  For this reason, we prefer to use the upper
  * bits.
- *
- * Macro parameters:
- *   uint32_t r          : Result.
- *   unsigned lg_range   : (0..32], number of least significant bits to return.
- *   uint32_t state      : Seed value.
- *   const uint32_t a, c : See above discussion.
  */
-#define	prng32(r, lg_range, state, a, c) do {				\
-	assert((lg_range) > 0);						\
-	assert((lg_range) <= 32);					\
-									\
-	r = (state * (a)) + (c);					\
-	state = r;							\
-	r >>= (32 - (lg_range));					\
-} while (false)
-
-/* Same as prng32(), but 64 bits of pseudo-randomness, using uint64_t. */
-#define	prng64(r, lg_range, state, a, c) do {				\
-	assert((lg_range) > 0);						\
-	assert((lg_range) <= 64);					\
-									\
-	r = (state * (a)) + (c);					\
-	state = r;							\
-	r >>= (64 - (lg_range));					\
-} while (false)
+
+#define	PRNG_A_32	UINT32_C(1103515241)
+#define	PRNG_C_32	UINT32_C(12347)
+
+#define	PRNG_A_64	UINT64_C(6364136223846793005)
+#define	PRNG_C_64	UINT64_C(1442695040888963407)
 
 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
@@ -56,5 +38,170 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_INLINES
 
+#ifndef JEMALLOC_ENABLE_INLINE
+uint32_t	prng_state_next_u32(uint32_t state);
+uint64_t	prng_state_next_u64(uint64_t state);
+size_t	prng_state_next_zu(size_t state);
+
+uint32_t	prng_lg_range_u32(uint32_t *state, unsigned lg_range,
+    bool atomic);
+uint64_t	prng_lg_range_u64(uint64_t *state, unsigned lg_range);
+size_t	prng_lg_range_zu(size_t *state, unsigned lg_range, bool atomic);
+
+uint32_t	prng_range_u32(uint32_t *state, uint32_t range, bool atomic);
+uint64_t	prng_range_u64(uint64_t *state, uint64_t range);
+size_t	prng_range_zu(size_t *state, size_t range, bool atomic);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_PRNG_C_))
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_state_next_u32(uint32_t state)
+{
+
+	return ((state * PRNG_A_32) + PRNG_C_32);
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_state_next_u64(uint64_t state)
+{
+
+	return ((state * PRNG_A_64) + PRNG_C_64);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_state_next_zu(size_t state)
+{
+
+#if LG_SIZEOF_PTR == 2
+	return ((state * PRNG_A_32) + PRNG_C_32);
+#elif LG_SIZEOF_PTR == 3
+	return ((state * PRNG_A_64) + PRNG_C_64);
+#else
+#error Unsupported pointer size
+#endif
+}
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_lg_range_u32(uint32_t *state, unsigned lg_range, bool atomic)
+{
+	uint32_t ret, state1;
+
+	assert(lg_range > 0);
+	assert(lg_range <= 32);
+
+	if (atomic) {
+		uint32_t state0;
+
+		do {
+			state0 = atomic_read_uint32(state);
+			state1 = prng_state_next_u32(state0);
+		} while (atomic_cas_uint32(state, state0, state1));
+	} else {
+		state1 = prng_state_next_u32(*state);
+		*state = state1;
+	}
+	ret = state1 >> (32 - lg_range);
+
+	return (ret);
+}
+
+/* 64-bit atomic operations cannot be supported on all relevant platforms. */
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_lg_range_u64(uint64_t *state, unsigned lg_range)
+{
+	uint64_t ret, state1;
+
+	assert(lg_range > 0);
+	assert(lg_range <= 64);
+
+	state1 = prng_state_next_u64(*state);
+	*state = state1;
+	ret = state1 >> (64 - lg_range);
+
+	return (ret);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_lg_range_zu(size_t *state, unsigned lg_range, bool atomic)
+{
+	size_t ret, state1;
+
+	assert(lg_range > 0);
+	assert(lg_range <= ZU(1) << (3 + LG_SIZEOF_PTR));
+
+	if (atomic) {
+		size_t state0;
+
+		do {
+			state0 = atomic_read_z(state);
+			state1 = prng_state_next_zu(state0);
+		} while (atomic_cas_z(state, state0, state1));
+	} else {
+		state1 = prng_state_next_zu(*state);
+		*state = state1;
+	}
+	ret = state1 >> ((ZU(1) << (3 + LG_SIZEOF_PTR)) - lg_range);
+
+	return (ret);
+}
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_range_u32(uint32_t *state, uint32_t range, bool atomic)
+{
+	uint32_t ret;
+	unsigned lg_range;
+
+	assert(range > 1);
+
+	/* Compute the ceiling of lg(range). */
+	lg_range = ffs_u32(pow2_ceil_u32(range)) - 1;
+
+	/* Generate a result in [0..range) via repeated trial. */
+	do {
+		ret = prng_lg_range_u32(state, lg_range, atomic);
+	} while (ret >= range);
+
+	return (ret);
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_range_u64(uint64_t *state, uint64_t range)
+{
+	uint64_t ret;
+	unsigned lg_range;
+
+	assert(range > 1);
+
+	/* Compute the ceiling of lg(range). */
+	lg_range = ffs_u64(pow2_ceil_u64(range)) - 1;
+
+	/* Generate a result in [0..range) via repeated trial. */
+	do {
+		ret = prng_lg_range_u64(state, lg_range);
+	} while (ret >= range);
+
+	return (ret);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_range_zu(size_t *state, size_t range, bool atomic)
+{
+	size_t ret;
+	unsigned lg_range;
+
+	assert(range > 1);
+
+	/* Compute the ceiling of lg(range). */
+	lg_range = ffs_u64(pow2_ceil_u64(range)) - 1;
+
+	/* Generate a result in [0..range) via repeated trial. */
+	do {
+		ret = prng_lg_range_zu(state, lg_range, atomic);
+	} while (ret >= range);
+
+	return (ret);
+}
+#endif
+
 #endif /* JEMALLOC_H_INLINES */
 /******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/prof.h b/deps/jemalloc/include/jemalloc/internal/prof.h
index e5198c3e8..8293b71ed 100644
--- a/deps/jemalloc/include/jemalloc/internal/prof.h
+++ b/deps/jemalloc/include/jemalloc/internal/prof.h
@@ -281,7 +281,7 @@ extern uint64_t	prof_interval;
 extern size_t	lg_prof_sample;
 
 void	prof_alloc_rollback(tsd_t *tsd, prof_tctx_t *tctx, bool updated);
-void	prof_malloc_sample_object(const void *ptr, size_t usize,
+void	prof_malloc_sample_object(tsdn_t *tsdn, const void *ptr, size_t usize,
     prof_tctx_t *tctx);
 void	prof_free_sampled_object(tsd_t *tsd, size_t usize, prof_tctx_t *tctx);
 void	bt_init(prof_bt_t *bt, void **vec);
@@ -293,32 +293,33 @@ size_t	prof_bt_count(void);
 const prof_cnt_t *prof_cnt_all(void);
 typedef int (prof_dump_open_t)(bool, const char *);
 extern prof_dump_open_t *prof_dump_open;
-typedef bool (prof_dump_header_t)(bool, const prof_cnt_t *);
+typedef bool (prof_dump_header_t)(tsdn_t *, bool, const prof_cnt_t *);
 extern prof_dump_header_t *prof_dump_header;
 #endif
-void	prof_idump(void);
-bool	prof_mdump(const char *filename);
-void	prof_gdump(void);
+void	prof_idump(tsdn_t *tsdn);
+bool	prof_mdump(tsd_t *tsd, const char *filename);
+void	prof_gdump(tsdn_t *tsdn);
 prof_tdata_t	*prof_tdata_init(tsd_t *tsd);
 prof_tdata_t	*prof_tdata_reinit(tsd_t *tsd, prof_tdata_t *tdata);
 void	prof_reset(tsd_t *tsd, size_t lg_sample);
 void	prof_tdata_cleanup(tsd_t *tsd);
-const char	*prof_thread_name_get(void);
-bool	prof_active_get(void);
-bool	prof_active_set(bool active);
+bool	prof_active_get(tsdn_t *tsdn);
+bool	prof_active_set(tsdn_t *tsdn, bool active);
+const char	*prof_thread_name_get(tsd_t *tsd);
 int	prof_thread_name_set(tsd_t *tsd, const char *thread_name);
-bool	prof_thread_active_get(void);
-bool	prof_thread_active_set(bool active);
-bool	prof_thread_active_init_get(void);
-bool	prof_thread_active_init_set(bool active_init);
-bool	prof_gdump_get(void);
-bool	prof_gdump_set(bool active);
+bool	prof_thread_active_get(tsd_t *tsd);
+bool	prof_thread_active_set(tsd_t *tsd, bool active);
+bool	prof_thread_active_init_get(tsdn_t *tsdn);
+bool	prof_thread_active_init_set(tsdn_t *tsdn, bool active_init);
+bool	prof_gdump_get(tsdn_t *tsdn);
+bool	prof_gdump_set(tsdn_t *tsdn, bool active);
 void	prof_boot0(void);
 void	prof_boot1(void);
-bool	prof_boot2(void);
-void	prof_prefork(void);
-void	prof_postfork_parent(void);
-void	prof_postfork_child(void);
+bool	prof_boot2(tsd_t *tsd);
+void	prof_prefork0(tsdn_t *tsdn);
+void	prof_prefork1(tsdn_t *tsdn);
+void	prof_postfork_parent(tsdn_t *tsdn);
+void	prof_postfork_child(tsdn_t *tsdn);
 void	prof_sample_threshold_update(prof_tdata_t *tdata);
 
 #endif /* JEMALLOC_H_EXTERNS */
@@ -329,17 +330,17 @@ void	prof_sample_threshold_update(prof_tdata_t *tdata);
 bool	prof_active_get_unlocked(void);
 bool	prof_gdump_get_unlocked(void);
 prof_tdata_t	*prof_tdata_get(tsd_t *tsd, bool create);
+prof_tctx_t	*prof_tctx_get(tsdn_t *tsdn, const void *ptr);
+void	prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    prof_tctx_t *tctx);
+void	prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize,
+    const void *old_ptr, prof_tctx_t *tctx);
 bool	prof_sample_accum_update(tsd_t *tsd, size_t usize, bool commit,
     prof_tdata_t **tdata_out);
 prof_tctx_t	*prof_alloc_prep(tsd_t *tsd, size_t usize, bool prof_active,
     bool update);
-prof_tctx_t	*prof_tctx_get(const void *ptr);
-void	prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx);
-void	prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
-    prof_tctx_t *tctx);
-void	prof_malloc_sample_object(const void *ptr, size_t usize,
+void	prof_malloc(tsdn_t *tsdn, const void *ptr, size_t usize,
     prof_tctx_t *tctx);
-void	prof_malloc(const void *ptr, size_t usize, prof_tctx_t *tctx);
 void	prof_realloc(tsd_t *tsd, const void *ptr, size_t usize,
     prof_tctx_t *tctx, bool prof_active, bool updated, const void *old_ptr,
     size_t old_usize, prof_tctx_t *old_tctx);
@@ -397,34 +398,34 @@ prof_tdata_get(tsd_t *tsd, bool create)
 }
 
 JEMALLOC_ALWAYS_INLINE prof_tctx_t *
-prof_tctx_get(const void *ptr)
+prof_tctx_get(tsdn_t *tsdn, const void *ptr)
 {
 
 	cassert(config_prof);
 	assert(ptr != NULL);
 
-	return (arena_prof_tctx_get(ptr));
+	return (arena_prof_tctx_get(tsdn, ptr));
 }
 
 JEMALLOC_ALWAYS_INLINE void
-prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)
+prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize, prof_tctx_t *tctx)
 {
 
 	cassert(config_prof);
 	assert(ptr != NULL);
 
-	arena_prof_tctx_set(ptr, usize, tctx);
+	arena_prof_tctx_set(tsdn, ptr, usize, tctx);
 }
 
 JEMALLOC_ALWAYS_INLINE void
-prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
+prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize, const void *old_ptr,
     prof_tctx_t *old_tctx)
 {
 
 	cassert(config_prof);
 	assert(ptr != NULL);
 
-	arena_prof_tctx_reset(ptr, usize, old_ptr, old_tctx);
+	arena_prof_tctx_reset(tsdn, ptr, usize, old_ptr, old_tctx);
 }
 
 JEMALLOC_ALWAYS_INLINE bool
@@ -436,16 +437,16 @@ prof_sample_accum_update(tsd_t *tsd, size_t usize, bool update,
 	cassert(config_prof);
 
 	tdata = prof_tdata_get(tsd, true);
-	if ((uintptr_t)tdata <= (uintptr_t)PROF_TDATA_STATE_MAX)
+	if (unlikely((uintptr_t)tdata <= (uintptr_t)PROF_TDATA_STATE_MAX))
 		tdata = NULL;
 
 	if (tdata_out != NULL)
 		*tdata_out = tdata;
 
-	if (tdata == NULL)
+	if (unlikely(tdata == NULL))
 		return (true);
 
-	if (tdata->bytes_until_sample >= usize) {
+	if (likely(tdata->bytes_until_sample >= usize)) {
 		if (update)
 			tdata->bytes_until_sample -= usize;
 		return (true);
@@ -479,17 +480,17 @@ prof_alloc_prep(tsd_t *tsd, size_t usize, bool prof_active, bool update)
 }
 
 JEMALLOC_ALWAYS_INLINE void
-prof_malloc(const void *ptr, size_t usize, prof_tctx_t *tctx)
+prof_malloc(tsdn_t *tsdn, const void *ptr, size_t usize, prof_tctx_t *tctx)
 {
 
 	cassert(config_prof);
 	assert(ptr != NULL);
-	assert(usize == isalloc(ptr, true));
+	assert(usize == isalloc(tsdn, ptr, true));
 
 	if (unlikely((uintptr_t)tctx > (uintptr_t)1U))
-		prof_malloc_sample_object(ptr, usize, tctx);
+		prof_malloc_sample_object(tsdn, ptr, usize, tctx);
 	else
-		prof_tctx_set(ptr, usize, (prof_tctx_t *)(uintptr_t)1U);
+		prof_tctx_set(tsdn, ptr, usize, (prof_tctx_t *)(uintptr_t)1U);
 }
 
 JEMALLOC_ALWAYS_INLINE void
@@ -503,7 +504,7 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 	assert(ptr != NULL || (uintptr_t)tctx <= (uintptr_t)1U);
 
 	if (prof_active && !updated && ptr != NULL) {
-		assert(usize == isalloc(ptr, true));
+		assert(usize == isalloc(tsd_tsdn(tsd), ptr, true));
 		if (prof_sample_accum_update(tsd, usize, true, NULL)) {
 			/*
 			 * Don't sample.  The usize passed to prof_alloc_prep()
@@ -512,6 +513,7 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 			 * though its actual usize was insufficient to cross the
 			 * sample threshold.
 			 */
+			prof_alloc_rollback(tsd, tctx, true);
 			tctx = (prof_tctx_t *)(uintptr_t)1U;
 		}
 	}
@@ -520,9 +522,9 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 	old_sampled = ((uintptr_t)old_tctx > (uintptr_t)1U);
 
 	if (unlikely(sampled))
-		prof_malloc_sample_object(ptr, usize, tctx);
+		prof_malloc_sample_object(tsd_tsdn(tsd), ptr, usize, tctx);
 	else
-		prof_tctx_reset(ptr, usize, old_ptr, old_tctx);
+		prof_tctx_reset(tsd_tsdn(tsd), ptr, usize, old_ptr, old_tctx);
 
 	if (unlikely(old_sampled))
 		prof_free_sampled_object(tsd, old_usize, old_tctx);
@@ -531,10 +533,10 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 JEMALLOC_ALWAYS_INLINE void
 prof_free(tsd_t *tsd, const void *ptr, size_t usize)
 {
-	prof_tctx_t *tctx = prof_tctx_get(ptr);
+	prof_tctx_t *tctx = prof_tctx_get(tsd_tsdn(tsd), ptr);
 
 	cassert(config_prof);
-	assert(usize == isalloc(ptr, true));
+	assert(usize == isalloc(tsd_tsdn(tsd), ptr, true));
 
 	if (unlikely((uintptr_t)tctx > (uintptr_t)1U))
 		prof_free_sampled_object(tsd, usize, tctx);
diff --git a/deps/jemalloc/include/jemalloc/internal/rb.h b/deps/jemalloc/include/jemalloc/internal/rb.h
index 2ca8e5933..3770342f8 100644
--- a/deps/jemalloc/include/jemalloc/internal/rb.h
+++ b/deps/jemalloc/include/jemalloc/internal/rb.h
@@ -42,7 +42,6 @@ struct {								\
 #define	rb_tree(a_type)							\
 struct {								\
     a_type *rbt_root;							\
-    a_type rbt_nil;							\
 }
 
 /* Left accessors. */
@@ -79,6 +78,15 @@ struct {								\
     (a_node)->a_field.rbn_right_red = (a_type *) (((intptr_t)		\
       (a_node)->a_field.rbn_right_red) & ((ssize_t)-2));		\
 } while (0)
+
+/* Node initializer. */
+#define	rbt_node_new(a_type, a_field, a_rbt, a_node) do {		\
+    /* Bookkeeping bit cannot be used by node pointer. */		\
+    assert(((uintptr_t)(a_node) & 0x1) == 0);				\
+    rbtn_left_set(a_type, a_field, (a_node), NULL);	\
+    rbtn_right_set(a_type, a_field, (a_node), NULL);	\
+    rbtn_red_set(a_type, a_field, (a_node));				\
+} while (0)
 #else
 /* Right accessors. */
 #define	rbtn_right_get(a_type, a_field, a_node)				\
@@ -99,28 +107,26 @@ struct {								\
 #define	rbtn_black_set(a_type, a_field, a_node) do {			\
     (a_node)->a_field.rbn_red = false;					\
 } while (0)
-#endif
 
 /* Node initializer. */
 #define	rbt_node_new(a_type, a_field, a_rbt, a_node) do {		\
-    rbtn_left_set(a_type, a_field, (a_node), &(a_rbt)->rbt_nil);	\
-    rbtn_right_set(a_type, a_field, (a_node), &(a_rbt)->rbt_nil);	\
+    rbtn_left_set(a_type, a_field, (a_node), NULL);	\
+    rbtn_right_set(a_type, a_field, (a_node), NULL);	\
     rbtn_red_set(a_type, a_field, (a_node));				\
 } while (0)
+#endif
 
 /* Tree initializer. */
 #define	rb_new(a_type, a_field, a_rbt) do {				\
-    (a_rbt)->rbt_root = &(a_rbt)->rbt_nil;				\
-    rbt_node_new(a_type, a_field, a_rbt, &(a_rbt)->rbt_nil);		\
-    rbtn_black_set(a_type, a_field, &(a_rbt)->rbt_nil);			\
+    (a_rbt)->rbt_root = NULL;						\
 } while (0)
 
 /* Internal utility macros. */
 #define	rbtn_first(a_type, a_field, a_rbt, a_root, r_node) do {		\
     (r_node) = (a_root);						\
-    if ((r_node) != &(a_rbt)->rbt_nil) {				\
+    if ((r_node) != NULL) {						\
 	for (;								\
-	  rbtn_left_get(a_type, a_field, (r_node)) != &(a_rbt)->rbt_nil;\
+	  rbtn_left_get(a_type, a_field, (r_node)) != NULL;		\
 	  (r_node) = rbtn_left_get(a_type, a_field, (r_node))) {	\
 	}								\
     }									\
@@ -128,10 +134,9 @@ struct {								\
 
 #define	rbtn_last(a_type, a_field, a_rbt, a_root, r_node) do {		\
     (r_node) = (a_root);						\
-    if ((r_node) != &(a_rbt)->rbt_nil) {				\
-	for (; rbtn_right_get(a_type, a_field, (r_node)) !=		\
-	  &(a_rbt)->rbt_nil; (r_node) = rbtn_right_get(a_type, a_field,	\
-	  (r_node))) {							\
+    if ((r_node) != NULL) {						\
+	for (; rbtn_right_get(a_type, a_field, (r_node)) != NULL;	\
+	  (r_node) = rbtn_right_get(a_type, a_field, (r_node))) {	\
 	}								\
     }									\
 } while (0)
@@ -169,11 +174,11 @@ a_prefix##next(a_rbt_type *rbtree, a_type *node);			\
 a_attr a_type *								\
 a_prefix##prev(a_rbt_type *rbtree, a_type *node);			\
 a_attr a_type *								\
-a_prefix##search(a_rbt_type *rbtree, a_type *key);			\
+a_prefix##search(a_rbt_type *rbtree, const a_type *key);		\
 a_attr a_type *								\
-a_prefix##nsearch(a_rbt_type *rbtree, a_type *key);			\
+a_prefix##nsearch(a_rbt_type *rbtree, const a_type *key);		\
 a_attr a_type *								\
-a_prefix##psearch(a_rbt_type *rbtree, a_type *key);			\
+a_prefix##psearch(a_rbt_type *rbtree, const a_type *key);		\
 a_attr void								\
 a_prefix##insert(a_rbt_type *rbtree, a_type *node);			\
 a_attr void								\
@@ -183,7 +188,10 @@ a_prefix##iter(a_rbt_type *rbtree, a_type *start, a_type *(*cb)(	\
   a_rbt_type *, a_type *, void *), void *arg);				\
 a_attr a_type *								\
 a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
-  a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg);
+  a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg);		\
+a_attr void								\
+a_prefix##destroy(a_rbt_type *rbtree, void (*cb)(a_type *, void *),	\
+  void *arg);
 
 /*
  * The rb_gen() macro generates a type-specific red-black tree implementation,
@@ -254,7 +262,7 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
  *            last/first.
  *
  *   static ex_node_t *
- *   ex_search(ex_t *tree, ex_node_t *key);
+ *   ex_search(ex_t *tree, const ex_node_t *key);
  *       Description: Search for node that matches key.
  *       Args:
  *         tree: Pointer to an initialized red-black tree object.
@@ -262,9 +270,9 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
  *       Ret: Node in tree that matches key, or NULL if no match.
  *
  *   static ex_node_t *
- *   ex_nsearch(ex_t *tree, ex_node_t *key);
+ *   ex_nsearch(ex_t *tree, const ex_node_t *key);
  *   static ex_node_t *
- *   ex_psearch(ex_t *tree, ex_node_t *key);
+ *   ex_psearch(ex_t *tree, const ex_node_t *key);
  *       Description: Search for node that matches key.  If no match is found,
  *                    return what would be key's successor/predecessor, were
  *                    key in tree.
@@ -312,6 +320,20 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
  *         arg  : Opaque pointer passed to cb().
  *       Ret: NULL if iteration completed, or the non-NULL callback return value
  *            that caused termination of the iteration.
+ *
+ *   static void
+ *   ex_destroy(ex_t *tree, void (*cb)(ex_node_t *, void *), void *arg);
+ *       Description: Iterate over the tree with post-order traversal, remove
+ *                    each node, and run the callback if non-null.  This is
+ *                    used for destroying a tree without paying the cost to
+ *                    rebalance it.  The tree must not be otherwise altered
+ *                    during traversal.
+ *       Args:
+ *         tree: Pointer to an initialized red-black tree object.
+ *         cb  : Callback function, which, if non-null, is called for each node
+ *               during iteration.  There is no way to stop iteration once it
+ *               has begun.
+ *         arg : Opaque pointer passed to cb().
  */
 #define	rb_gen(a_attr, a_prefix, a_rbt_type, a_type, a_field, a_cmp)	\
 a_attr void								\
@@ -320,36 +342,30 @@ a_prefix##new(a_rbt_type *rbtree) {					\
 }									\
 a_attr bool								\
 a_prefix##empty(a_rbt_type *rbtree) {					\
-    return (rbtree->rbt_root == &rbtree->rbt_nil);			\
+    return (rbtree->rbt_root == NULL);					\
 }									\
 a_attr a_type *								\
 a_prefix##first(a_rbt_type *rbtree) {					\
     a_type *ret;							\
     rbtn_first(a_type, a_field, rbtree, rbtree->rbt_root, ret);		\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
-    }									\
     return (ret);							\
 }									\
 a_attr a_type *								\
 a_prefix##last(a_rbt_type *rbtree) {					\
     a_type *ret;							\
     rbtn_last(a_type, a_field, rbtree, rbtree->rbt_root, ret);		\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
-    }									\
     return (ret);							\
 }									\
 a_attr a_type *								\
 a_prefix##next(a_rbt_type *rbtree, a_type *node) {			\
     a_type *ret;							\
-    if (rbtn_right_get(a_type, a_field, node) != &rbtree->rbt_nil) {	\
+    if (rbtn_right_get(a_type, a_field, node) != NULL) {		\
 	rbtn_first(a_type, a_field, rbtree, rbtn_right_get(a_type,	\
 	  a_field, node), ret);						\
     } else {								\
 	a_type *tnode = rbtree->rbt_root;				\
-	assert(tnode != &rbtree->rbt_nil);				\
-	ret = &rbtree->rbt_nil;						\
+	assert(tnode != NULL);						\
+	ret = NULL;							\
 	while (true) {							\
 	    int cmp = (a_cmp)(node, tnode);				\
 	    if (cmp < 0) {						\
@@ -360,24 +376,21 @@ a_prefix##next(a_rbt_type *rbtree, a_type *node) {			\
 	    } else {							\
 		break;							\
 	    }								\
-	    assert(tnode != &rbtree->rbt_nil);				\
+	    assert(tnode != NULL);					\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
     return (ret);							\
 }									\
 a_attr a_type *								\
 a_prefix##prev(a_rbt_type *rbtree, a_type *node) {			\
     a_type *ret;							\
-    if (rbtn_left_get(a_type, a_field, node) != &rbtree->rbt_nil) {	\
+    if (rbtn_left_get(a_type, a_field, node) != NULL) {			\
 	rbtn_last(a_type, a_field, rbtree, rbtn_left_get(a_type,	\
 	  a_field, node), ret);						\
     } else {								\
 	a_type *tnode = rbtree->rbt_root;				\
-	assert(tnode != &rbtree->rbt_nil);				\
-	ret = &rbtree->rbt_nil;						\
+	assert(tnode != NULL);						\
+	ret = NULL;							\
 	while (true) {							\
 	    int cmp = (a_cmp)(node, tnode);				\
 	    if (cmp < 0) {						\
@@ -388,20 +401,17 @@ a_prefix##prev(a_rbt_type *rbtree, a_type *node) {			\
 	    } else {							\
 		break;							\
 	    }								\
-	    assert(tnode != &rbtree->rbt_nil);				\
+	    assert(tnode != NULL);					\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
     return (ret);							\
 }									\
 a_attr a_type *								\
-a_prefix##search(a_rbt_type *rbtree, a_type *key) {			\
+a_prefix##search(a_rbt_type *rbtree, const a_type *key) {		\
     a_type *ret;							\
     int cmp;								\
     ret = rbtree->rbt_root;						\
-    while (ret != &rbtree->rbt_nil					\
+    while (ret != NULL							\
       && (cmp = (a_cmp)(key, ret)) != 0) {				\
 	if (cmp < 0) {							\
 	    ret = rbtn_left_get(a_type, a_field, ret);			\
@@ -409,17 +419,14 @@ a_prefix##search(a_rbt_type *rbtree, a_type *key) {			\
 	    ret = rbtn_right_get(a_type, a_field, ret);			\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
     return (ret);							\
 }									\
 a_attr a_type *								\
-a_prefix##nsearch(a_rbt_type *rbtree, a_type *key) {			\
+a_prefix##nsearch(a_rbt_type *rbtree, const a_type *key) {		\
     a_type *ret;							\
     a_type *tnode = rbtree->rbt_root;					\
-    ret = &rbtree->rbt_nil;						\
-    while (tnode != &rbtree->rbt_nil) {					\
+    ret = NULL;								\
+    while (tnode != NULL) {						\
 	int cmp = (a_cmp)(key, tnode);					\
 	if (cmp < 0) {							\
 	    ret = tnode;						\
@@ -431,17 +438,14 @@ a_prefix##nsearch(a_rbt_type *rbtree, a_type *key) {			\
 	    break;							\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
     return (ret);							\
 }									\
 a_attr a_type *								\
-a_prefix##psearch(a_rbt_type *rbtree, a_type *key) {			\
+a_prefix##psearch(a_rbt_type *rbtree, const a_type *key) {		\
     a_type *ret;							\
     a_type *tnode = rbtree->rbt_root;					\
-    ret = &rbtree->rbt_nil;						\
-    while (tnode != &rbtree->rbt_nil) {					\
+    ret = NULL;								\
+    while (tnode != NULL) {						\
 	int cmp = (a_cmp)(key, tnode);					\
 	if (cmp < 0) {							\
 	    tnode = rbtn_left_get(a_type, a_field, tnode);		\
@@ -453,9 +457,6 @@ a_prefix##psearch(a_rbt_type *rbtree, a_type *key) {			\
 	    break;							\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
     return (ret);							\
 }									\
 a_attr void								\
@@ -467,7 +468,7 @@ a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
     rbt_node_new(a_type, a_field, rbtree, node);			\
     /* Wind. */								\
     path->node = rbtree->rbt_root;					\
-    for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) {	\
+    for (pathp = path; pathp->node != NULL; pathp++) {			\
 	int cmp = pathp->cmp = a_cmp(node, pathp->node);		\
 	assert(cmp != 0);						\
 	if (cmp < 0) {							\
@@ -487,7 +488,8 @@ a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
 	    rbtn_left_set(a_type, a_field, cnode, left);		\
 	    if (rbtn_red_get(a_type, a_field, left)) {			\
 		a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
-		if (rbtn_red_get(a_type, a_field, leftleft)) {		\
+		if (leftleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  leftleft)) {						\
 		    /* Fix up 4-node. */				\
 		    a_type *tnode;					\
 		    rbtn_black_set(a_type, a_field, leftleft);		\
@@ -502,7 +504,8 @@ a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
 	    rbtn_right_set(a_type, a_field, cnode, right);		\
 	    if (rbtn_red_get(a_type, a_field, right)) {			\
 		a_type *left = rbtn_left_get(a_type, a_field, cnode);	\
-		if (rbtn_red_get(a_type, a_field, left)) {		\
+		if (left != NULL && rbtn_red_get(a_type, a_field,	\
+		  left)) {						\
 		    /* Split 4-node. */					\
 		    rbtn_black_set(a_type, a_field, left);		\
 		    rbtn_black_set(a_type, a_field, right);		\
@@ -535,7 +538,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
     /* Wind. */								\
     nodep = NULL; /* Silence compiler warning. */			\
     path->node = rbtree->rbt_root;					\
-    for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) {	\
+    for (pathp = path; pathp->node != NULL; pathp++) {			\
 	int cmp = pathp->cmp = a_cmp(node, pathp->node);		\
 	if (cmp < 0) {							\
 	    pathp[1].node = rbtn_left_get(a_type, a_field,		\
@@ -547,7 +550,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 	        /* Find node's successor, in preparation for swap. */	\
 		pathp->cmp = 1;						\
 		nodep = pathp;						\
-		for (pathp++; pathp->node != &rbtree->rbt_nil;		\
+		for (pathp++; pathp->node != NULL;			\
 		  pathp++) {						\
 		    pathp->cmp = -1;					\
 		    pathp[1].node = rbtn_left_get(a_type, a_field,	\
@@ -590,7 +593,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 	}								\
     } else {								\
 	a_type *left = rbtn_left_get(a_type, a_field, node);		\
-	if (left != &rbtree->rbt_nil) {					\
+	if (left != NULL) {						\
 	    /* node has no successor, but it has a left child.        */\
 	    /* Splice node out, without losing the left child.        */\
 	    assert(!rbtn_red_get(a_type, a_field, node));		\
@@ -610,33 +613,32 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 	    return;							\
 	} else if (pathp == path) {					\
 	    /* The tree only contained one node. */			\
-	    rbtree->rbt_root = &rbtree->rbt_nil;			\
+	    rbtree->rbt_root = NULL;					\
 	    return;							\
 	}								\
     }									\
     if (rbtn_red_get(a_type, a_field, pathp->node)) {			\
 	/* Prune red node, which requires no fixup. */			\
 	assert(pathp[-1].cmp < 0);					\
-	rbtn_left_set(a_type, a_field, pathp[-1].node,			\
-	  &rbtree->rbt_nil);						\
+	rbtn_left_set(a_type, a_field, pathp[-1].node, NULL);		\
 	return;								\
     }									\
     /* The node to be pruned is black, so unwind until balance is     */\
     /* restored.                                                      */\
-    pathp->node = &rbtree->rbt_nil;					\
+    pathp->node = NULL;							\
     for (pathp--; (uintptr_t)pathp >= (uintptr_t)path; pathp--) {	\
 	assert(pathp->cmp != 0);					\
 	if (pathp->cmp < 0) {						\
 	    rbtn_left_set(a_type, a_field, pathp->node,			\
 	      pathp[1].node);						\
-	    assert(!rbtn_red_get(a_type, a_field, pathp[1].node));	\
 	    if (rbtn_red_get(a_type, a_field, pathp->node)) {		\
 		a_type *right = rbtn_right_get(a_type, a_field,		\
 		  pathp->node);						\
 		a_type *rightleft = rbtn_left_get(a_type, a_field,	\
 		  right);						\
 		a_type *tnode;						\
-		if (rbtn_red_get(a_type, a_field, rightleft)) {		\
+		if (rightleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  rightleft)) {						\
 		    /* In the following diagrams, ||, //, and \\      */\
 		    /* indicate the path to the removed node.         */\
 		    /*                                                */\
@@ -679,7 +681,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		  pathp->node);						\
 		a_type *rightleft = rbtn_left_get(a_type, a_field,	\
 		  right);						\
-		if (rbtn_red_get(a_type, a_field, rightleft)) {		\
+		if (rightleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  rightleft)) {						\
 		    /*      ||                                        */\
 		    /*    pathp(b)                                    */\
 		    /*  //        \                                   */\
@@ -733,7 +736,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		  left);						\
 		a_type *leftrightleft = rbtn_left_get(a_type, a_field,	\
 		  leftright);						\
-		if (rbtn_red_get(a_type, a_field, leftrightleft)) {	\
+		if (leftrightleft != NULL && rbtn_red_get(a_type,	\
+		  a_field, leftrightleft)) {				\
 		    /*      ||                                        */\
 		    /*    pathp(b)                                    */\
 		    /*   /        \\                                  */\
@@ -759,7 +763,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		    /*   (b)                                          */\
 		    /*   /                                            */\
 		    /* (b)                                            */\
-		    assert(leftright != &rbtree->rbt_nil);		\
+		    assert(leftright != NULL);				\
 		    rbtn_red_set(a_type, a_field, leftright);		\
 		    rbtn_rotate_right(a_type, a_field, pathp->node,	\
 		      tnode);						\
@@ -782,7 +786,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		return;							\
 	    } else if (rbtn_red_get(a_type, a_field, pathp->node)) {	\
 		a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
-		if (rbtn_red_get(a_type, a_field, leftleft)) {		\
+		if (leftleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  leftleft)) {						\
 		    /*        ||                                      */\
 		    /*      pathp(r)                                  */\
 		    /*     /        \\                                */\
@@ -820,7 +825,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		}							\
 	    } else {							\
 		a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
-		if (rbtn_red_get(a_type, a_field, leftleft)) {		\
+		if (leftleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  leftleft)) {						\
 		    /*               ||                               */\
 		    /*             pathp(b)                           */\
 		    /*            /        \\                         */\
@@ -866,13 +872,13 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 a_attr a_type *								\
 a_prefix##iter_recurse(a_rbt_type *rbtree, a_type *node,		\
   a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) {		\
-    if (node == &rbtree->rbt_nil) {					\
-	return (&rbtree->rbt_nil);					\
+    if (node == NULL) {							\
+	return (NULL);							\
     } else {								\
 	a_type *ret;							\
 	if ((ret = a_prefix##iter_recurse(rbtree, rbtn_left_get(a_type,	\
-	  a_field, node), cb, arg)) != &rbtree->rbt_nil			\
-	  || (ret = cb(rbtree, node, arg)) != NULL) {			\
+	  a_field, node), cb, arg)) != NULL || (ret = cb(rbtree, node,	\
+	  arg)) != NULL) {						\
 	    return (ret);						\
 	}								\
 	return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
@@ -886,8 +892,8 @@ a_prefix##iter_start(a_rbt_type *rbtree, a_type *start, a_type *node,	\
     if (cmp < 0) {							\
 	a_type *ret;							\
 	if ((ret = a_prefix##iter_start(rbtree, start,			\
-	  rbtn_left_get(a_type, a_field, node), cb, arg)) !=		\
-	  &rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) {	\
+	  rbtn_left_get(a_type, a_field, node), cb, arg)) != NULL ||	\
+	  (ret = cb(rbtree, node, arg)) != NULL) {			\
 	    return (ret);						\
 	}								\
 	return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
@@ -914,21 +920,18 @@ a_prefix##iter(a_rbt_type *rbtree, a_type *start, a_type *(*cb)(	\
     } else {								\
 	ret = a_prefix##iter_recurse(rbtree, rbtree->rbt_root, cb, arg);\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
-    }									\
     return (ret);							\
 }									\
 a_attr a_type *								\
 a_prefix##reverse_iter_recurse(a_rbt_type *rbtree, a_type *node,	\
   a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) {		\
-    if (node == &rbtree->rbt_nil) {					\
-	return (&rbtree->rbt_nil);					\
+    if (node == NULL) {							\
+	return (NULL);							\
     } else {								\
 	a_type *ret;							\
 	if ((ret = a_prefix##reverse_iter_recurse(rbtree,		\
-	  rbtn_right_get(a_type, a_field, node), cb, arg)) !=		\
-	  &rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) {	\
+	  rbtn_right_get(a_type, a_field, node), cb, arg)) != NULL ||	\
+	  (ret = cb(rbtree, node, arg)) != NULL) {			\
 	    return (ret);						\
 	}								\
 	return (a_prefix##reverse_iter_recurse(rbtree,			\
@@ -943,8 +946,8 @@ a_prefix##reverse_iter_start(a_rbt_type *rbtree, a_type *start,		\
     if (cmp > 0) {							\
 	a_type *ret;							\
 	if ((ret = a_prefix##reverse_iter_start(rbtree, start,		\
-	  rbtn_right_get(a_type, a_field, node), cb, arg)) !=		\
-	  &rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) {	\
+	  rbtn_right_get(a_type, a_field, node), cb, arg)) != NULL ||	\
+	  (ret = cb(rbtree, node, arg)) != NULL) {			\
 	    return (ret);						\
 	}								\
 	return (a_prefix##reverse_iter_recurse(rbtree,			\
@@ -972,10 +975,29 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
 	ret = a_prefix##reverse_iter_recurse(rbtree, rbtree->rbt_root,	\
 	  cb, arg);							\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
-    }									\
     return (ret);							\
+}									\
+a_attr void								\
+a_prefix##destroy_recurse(a_rbt_type *rbtree, a_type *node, void (*cb)(	\
+  a_type *, void *), void *arg) {					\
+    if (node == NULL) {							\
+	return;								\
+    }									\
+    a_prefix##destroy_recurse(rbtree, rbtn_left_get(a_type, a_field,	\
+      node), cb, arg);							\
+    rbtn_left_set(a_type, a_field, (node), NULL);			\
+    a_prefix##destroy_recurse(rbtree, rbtn_right_get(a_type, a_field,	\
+      node), cb, arg);							\
+    rbtn_right_set(a_type, a_field, (node), NULL);			\
+    if (cb) {								\
+	cb(node, arg);							\
+    }									\
+}									\
+a_attr void								\
+a_prefix##destroy(a_rbt_type *rbtree, void (*cb)(a_type *, void *),	\
+  void *arg) {								\
+    a_prefix##destroy_recurse(rbtree, rbtree->rbt_root, cb, arg);	\
+    rbtree->rbt_root = NULL;						\
 }
 
 #endif /* RB_H_ */
diff --git a/deps/jemalloc/include/jemalloc/internal/rtree.h b/deps/jemalloc/include/jemalloc/internal/rtree.h
index 28ae9d1dd..8d0c584da 100644
--- a/deps/jemalloc/include/jemalloc/internal/rtree.h
+++ b/deps/jemalloc/include/jemalloc/internal/rtree.h
@@ -15,9 +15,10 @@ typedef struct rtree_s rtree_t;
  * machine address width.
  */
 #define	LG_RTREE_BITS_PER_LEVEL	4
-#define	RTREE_BITS_PER_LEVEL	(ZU(1) << LG_RTREE_BITS_PER_LEVEL)
+#define	RTREE_BITS_PER_LEVEL	(1U << LG_RTREE_BITS_PER_LEVEL)
+/* Maximum rtree height. */
 #define	RTREE_HEIGHT_MAX						\
-    ((ZU(1) << (LG_SIZEOF_PTR+3)) / RTREE_BITS_PER_LEVEL)
+    ((1U << (LG_SIZEOF_PTR+3)) / RTREE_BITS_PER_LEVEL)
 
 /* Used for two-stage lock-free node initialization. */
 #define	RTREE_NODE_INITIALIZING	((rtree_node_elm_t *)0x1)
@@ -111,22 +112,25 @@ unsigned	rtree_start_level(rtree_t *rtree, uintptr_t key);
 uintptr_t	rtree_subkey(rtree_t *rtree, uintptr_t key, unsigned level);
 
 bool	rtree_node_valid(rtree_node_elm_t *node);
-rtree_node_elm_t	*rtree_child_tryread(rtree_node_elm_t *elm);
+rtree_node_elm_t	*rtree_child_tryread(rtree_node_elm_t *elm,
+    bool dependent);
 rtree_node_elm_t	*rtree_child_read(rtree_t *rtree, rtree_node_elm_t *elm,
-    unsigned level);
+    unsigned level, bool dependent);
 extent_node_t	*rtree_val_read(rtree_t *rtree, rtree_node_elm_t *elm,
     bool dependent);
 void	rtree_val_write(rtree_t *rtree, rtree_node_elm_t *elm,
     const extent_node_t *val);
-rtree_node_elm_t	*rtree_subtree_tryread(rtree_t *rtree, unsigned level);
-rtree_node_elm_t	*rtree_subtree_read(rtree_t *rtree, unsigned level);
+rtree_node_elm_t	*rtree_subtree_tryread(rtree_t *rtree, unsigned level,
+    bool dependent);
+rtree_node_elm_t	*rtree_subtree_read(rtree_t *rtree, unsigned level,
+    bool dependent);
 
 extent_node_t	*rtree_get(rtree_t *rtree, uintptr_t key, bool dependent);
 bool	rtree_set(rtree_t *rtree, uintptr_t key, const extent_node_t *val);
 #endif
 
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_RTREE_C_))
-JEMALLOC_INLINE unsigned
+JEMALLOC_ALWAYS_INLINE unsigned
 rtree_start_level(rtree_t *rtree, uintptr_t key)
 {
 	unsigned start_level;
@@ -140,7 +144,7 @@ rtree_start_level(rtree_t *rtree, uintptr_t key)
 	return (start_level);
 }
 
-JEMALLOC_INLINE uintptr_t
+JEMALLOC_ALWAYS_INLINE uintptr_t
 rtree_subkey(rtree_t *rtree, uintptr_t key, unsigned level)
 {
 
@@ -149,37 +153,40 @@ rtree_subkey(rtree_t *rtree, uintptr_t key, unsigned level)
 	    rtree->levels[level].bits) - 1));
 }
 
-JEMALLOC_INLINE bool
+JEMALLOC_ALWAYS_INLINE bool
 rtree_node_valid(rtree_node_elm_t *node)
 {
 
 	return ((uintptr_t)node > (uintptr_t)RTREE_NODE_INITIALIZING);
 }
 
-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_child_tryread(rtree_node_elm_t *elm)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_child_tryread(rtree_node_elm_t *elm, bool dependent)
 {
 	rtree_node_elm_t *child;
 
 	/* Double-checked read (first read may be stale. */
 	child = elm->child;
-	if (!rtree_node_valid(child))
+	if (!dependent && !rtree_node_valid(child))
 		child = atomic_read_p(&elm->pun);
+	assert(!dependent || child != NULL);
 	return (child);
 }
 
-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_child_read(rtree_t *rtree, rtree_node_elm_t *elm, unsigned level)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_child_read(rtree_t *rtree, rtree_node_elm_t *elm, unsigned level,
+    bool dependent)
 {
 	rtree_node_elm_t *child;
 
-	child = rtree_child_tryread(elm);
-	if (unlikely(!rtree_node_valid(child)))
+	child = rtree_child_tryread(elm, dependent);
+	if (!dependent && unlikely(!rtree_node_valid(child)))
 		child = rtree_child_read_hard(rtree, elm, level);
+	assert(!dependent || child != NULL);
 	return (child);
 }
 
-JEMALLOC_INLINE extent_node_t *
+JEMALLOC_ALWAYS_INLINE extent_node_t *
 rtree_val_read(rtree_t *rtree, rtree_node_elm_t *elm, bool dependent)
 {
 
@@ -208,54 +215,119 @@ rtree_val_write(rtree_t *rtree, rtree_node_elm_t *elm, const extent_node_t *val)
 	atomic_write_p(&elm->pun, val);
 }
 
-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_subtree_tryread(rtree_t *rtree, unsigned level)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_subtree_tryread(rtree_t *rtree, unsigned level, bool dependent)
 {
 	rtree_node_elm_t *subtree;
 
 	/* Double-checked read (first read may be stale. */
 	subtree = rtree->levels[level].subtree;
-	if (!rtree_node_valid(subtree))
+	if (!dependent && unlikely(!rtree_node_valid(subtree)))
 		subtree = atomic_read_p(&rtree->levels[level].subtree_pun);
+	assert(!dependent || subtree != NULL);
 	return (subtree);
 }
 
-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_subtree_read(rtree_t *rtree, unsigned level)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_subtree_read(rtree_t *rtree, unsigned level, bool dependent)
 {
 	rtree_node_elm_t *subtree;
 
-	subtree = rtree_subtree_tryread(rtree, level);
-	if (unlikely(!rtree_node_valid(subtree)))
+	subtree = rtree_subtree_tryread(rtree, level, dependent);
+	if (!dependent && unlikely(!rtree_node_valid(subtree)))
 		subtree = rtree_subtree_read_hard(rtree, level);
+	assert(!dependent || subtree != NULL);
 	return (subtree);
 }
 
-JEMALLOC_INLINE extent_node_t *
+JEMALLOC_ALWAYS_INLINE extent_node_t *
 rtree_get(rtree_t *rtree, uintptr_t key, bool dependent)
 {
 	uintptr_t subkey;
-	unsigned i, start_level;
-	rtree_node_elm_t *node, *child;
+	unsigned start_level;
+	rtree_node_elm_t *node;
 
 	start_level = rtree_start_level(rtree, key);
 
-	for (i = start_level, node = rtree_subtree_tryread(rtree, start_level);
-	    /**/; i++, node = child) {
-		if (!dependent && unlikely(!rtree_node_valid(node)))
-			return (NULL);
-		subkey = rtree_subkey(rtree, key, i);
-		if (i == rtree->height - 1) {
-			/*
-			 * node is a leaf, so it contains values rather than
-			 * child pointers.
-			 */
-			return (rtree_val_read(rtree, &node[subkey],
-			    dependent));
-		}
-		assert(i < rtree->height - 1);
-		child = rtree_child_tryread(&node[subkey]);
+	node = rtree_subtree_tryread(rtree, start_level, dependent);
+#define	RTREE_GET_BIAS	(RTREE_HEIGHT_MAX - rtree->height)
+	switch (start_level + RTREE_GET_BIAS) {
+#define	RTREE_GET_SUBTREE(level)					\
+	case level:							\
+		assert(level < (RTREE_HEIGHT_MAX-1));			\
+		if (!dependent && unlikely(!rtree_node_valid(node)))	\
+			return (NULL);					\
+		subkey = rtree_subkey(rtree, key, level -		\
+		    RTREE_GET_BIAS);					\
+		node = rtree_child_tryread(&node[subkey], dependent);	\
+		/* Fall through. */
+#define	RTREE_GET_LEAF(level)						\
+	case level:							\
+		assert(level == (RTREE_HEIGHT_MAX-1));			\
+		if (!dependent && unlikely(!rtree_node_valid(node)))	\
+			return (NULL);					\
+		subkey = rtree_subkey(rtree, key, level -		\
+		    RTREE_GET_BIAS);					\
+		/*							\
+		 * node is a leaf, so it contains values rather than	\
+		 * child pointers.					\
+		 */							\
+		return (rtree_val_read(rtree, &node[subkey],		\
+		    dependent));
+#if RTREE_HEIGHT_MAX > 1
+	RTREE_GET_SUBTREE(0)
+#endif
+#if RTREE_HEIGHT_MAX > 2
+	RTREE_GET_SUBTREE(1)
+#endif
+#if RTREE_HEIGHT_MAX > 3
+	RTREE_GET_SUBTREE(2)
+#endif
+#if RTREE_HEIGHT_MAX > 4
+	RTREE_GET_SUBTREE(3)
+#endif
+#if RTREE_HEIGHT_MAX > 5
+	RTREE_GET_SUBTREE(4)
+#endif
+#if RTREE_HEIGHT_MAX > 6
+	RTREE_GET_SUBTREE(5)
+#endif
+#if RTREE_HEIGHT_MAX > 7
+	RTREE_GET_SUBTREE(6)
+#endif
+#if RTREE_HEIGHT_MAX > 8
+	RTREE_GET_SUBTREE(7)
+#endif
+#if RTREE_HEIGHT_MAX > 9
+	RTREE_GET_SUBTREE(8)
+#endif
+#if RTREE_HEIGHT_MAX > 10
+	RTREE_GET_SUBTREE(9)
+#endif
+#if RTREE_HEIGHT_MAX > 11
+	RTREE_GET_SUBTREE(10)
+#endif
+#if RTREE_HEIGHT_MAX > 12
+	RTREE_GET_SUBTREE(11)
+#endif
+#if RTREE_HEIGHT_MAX > 13
+	RTREE_GET_SUBTREE(12)
+#endif
+#if RTREE_HEIGHT_MAX > 14
+	RTREE_GET_SUBTREE(13)
+#endif
+#if RTREE_HEIGHT_MAX > 15
+	RTREE_GET_SUBTREE(14)
+#endif
+#if RTREE_HEIGHT_MAX > 16
+#  error Unsupported RTREE_HEIGHT_MAX
+#endif
+	RTREE_GET_LEAF(RTREE_HEIGHT_MAX-1)
+#undef RTREE_GET_SUBTREE
+#undef RTREE_GET_LEAF
+	default: not_reached();
 	}
+#undef RTREE_GET_BIAS
 	not_reached();
 }
 
@@ -268,7 +340,7 @@ rtree_set(rtree_t *rtree, uintptr_t key, const extent_node_t *val)
 
 	start_level = rtree_start_level(rtree, key);
 
-	node = rtree_subtree_read(rtree, start_level);
+	node = rtree_subtree_read(rtree, start_level, false);
 	if (node == NULL)
 		return (true);
 	for (i = start_level; /**/; i++, node = child) {
@@ -282,7 +354,7 @@ rtree_set(rtree_t *rtree, uintptr_t key, const extent_node_t *val)
 			return (false);
 		}
 		assert(i + 1 < rtree->height);
-		child = rtree_child_read(rtree, &node[subkey], i);
+		child = rtree_child_read(rtree, &node[subkey], i, false);
 		if (child == NULL)
 			return (true);
 	}
diff --git a/deps/jemalloc/include/jemalloc/internal/size_classes.sh b/deps/jemalloc/include/jemalloc/internal/size_classes.sh
index fc82036d3..f6fbce4ef 100755
--- a/deps/jemalloc/include/jemalloc/internal/size_classes.sh
+++ b/deps/jemalloc/include/jemalloc/internal/size_classes.sh
@@ -48,6 +48,21 @@ size_class() {
   lg_p=$5
   lg_kmax=$6
 
+  if [ ${lg_delta} -ge ${lg_p} ] ; then
+    psz="yes"
+  else
+    pow2 ${lg_p}; p=${pow2_result}
+    pow2 ${lg_grp}; grp=${pow2_result}
+    pow2 ${lg_delta}; delta=${pow2_result}
+    sz=$((${grp} + ${delta} * ${ndelta}))
+    npgs=$((${sz} / ${p}))
+    if [ ${sz} -eq $((${npgs} * ${p})) ] ; then
+      psz="yes"
+    else
+      psz="no"
+    fi
+  fi
+
   lg ${ndelta}; lg_ndelta=${lg_result}; pow2 ${lg_ndelta}
   if [ ${pow2_result} -lt ${ndelta} ] ; then
     rem="yes"
@@ -74,14 +89,15 @@ size_class() {
   else
     lg_delta_lookup="no"
   fi
-  printf '    SC(%3d, %6d, %8d, %6d, %3s, %2s) \\\n' ${index} ${lg_grp} ${lg_delta} ${ndelta} ${bin} ${lg_delta_lookup}
+  printf '    SC(%3d, %6d, %8d, %6d, %3s, %3s, %2s) \\\n' ${index} ${lg_grp} ${lg_delta} ${ndelta} ${psz} ${bin} ${lg_delta_lookup}
   # Defined upon return:
-  # - lg_delta_lookup (${lg_delta} or "no")
+  # - psz ("yes" or "no")
   # - bin ("yes" or "no")
+  # - lg_delta_lookup (${lg_delta} or "no")
 }
 
 sep_line() {
-  echo "                                               \\"
+  echo "                                                         \\"
 }
 
 size_classes() {
@@ -95,12 +111,13 @@ size_classes() {
   pow2 ${lg_g}; g=${pow2_result}
 
   echo "#define	SIZE_CLASSES \\"
-  echo "  /* index, lg_grp, lg_delta, ndelta, bin, lg_delta_lookup */ \\"
+  echo "  /* index, lg_grp, lg_delta, ndelta, psz, bin, lg_delta_lookup */ \\"
 
   ntbins=0
   nlbins=0
   lg_tiny_maxclass='"NA"'
   nbins=0
+  npsizes=0
 
   # Tiny size classes.
   ndelta=0
@@ -112,6 +129,9 @@ size_classes() {
     if [ ${lg_delta_lookup} != "no" ] ; then
       nlbins=$((${index} + 1))
     fi
+    if [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
     if [ ${bin} != "no" ] ; then
       nbins=$((${index} + 1))
     fi
@@ -133,19 +153,25 @@ size_classes() {
     index=$((${index} + 1))
     lg_grp=$((${lg_grp} + 1))
     lg_delta=$((${lg_delta} + 1))
+    if [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
   fi
   while [ ${ndelta} -lt ${g} ] ; do
     size_class ${index} ${lg_grp} ${lg_delta} ${ndelta} ${lg_p} ${lg_kmax}
     index=$((${index} + 1))
     ndelta=$((${ndelta} + 1))
+    if [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
   done
 
   # All remaining groups.
   lg_grp=$((${lg_grp} + ${lg_g}))
-  while [ ${lg_grp} -lt ${ptr_bits} ] ; do
+  while [ ${lg_grp} -lt $((${ptr_bits} - 1)) ] ; do
     sep_line
     ndelta=1
-    if [ ${lg_grp} -eq $((${ptr_bits} - 1)) ] ; then
+    if [ ${lg_grp} -eq $((${ptr_bits} - 2)) ] ; then
       ndelta_limit=$((${g} - 1))
     else
       ndelta_limit=${g}
@@ -157,6 +183,9 @@ size_classes() {
         # Final written value is correct:
         lookup_maxclass="((((size_t)1) << ${lg_grp}) + (((size_t)${ndelta}) << ${lg_delta}))"
       fi
+      if [ ${psz} = "yes" ] ; then
+        npsizes=$((${npsizes} + 1))
+      fi
       if [ ${bin} != "no" ] ; then
         nbins=$((${index} + 1))
         # Final written value is correct:
@@ -183,6 +212,7 @@ size_classes() {
   # - nlbins
   # - nbins
   # - nsizes
+  # - npsizes
   # - lg_tiny_maxclass
   # - lookup_maxclass
   # - small_maxclass
@@ -200,13 +230,13 @@ cat <<EOF
  * be defined prior to inclusion, and it in turn defines:
  *
  *   LG_SIZE_CLASS_GROUP: Lg of size class count for each size doubling.
- *   SIZE_CLASSES: Complete table of
- *                 SC(index, lg_grp, lg_delta, ndelta, bin, lg_delta_lookup)
- *                 tuples.
+ *   SIZE_CLASSES: Complete table of SC(index, lg_grp, lg_delta, ndelta, psz,
+ *                 bin, lg_delta_lookup) tuples.
  *     index: Size class index.
  *     lg_grp: Lg group base size (no deltas added).
  *     lg_delta: Lg delta to previous size class.
  *     ndelta: Delta multiplier.  size == 1<<lg_grp + ndelta<<lg_delta
+ *     psz: 'yes' if a multiple of the page size, 'no' otherwise.
  *     bin: 'yes' if a small bin size class, 'no' otherwise.
  *     lg_delta_lookup: Same as lg_delta if a lookup table size class, 'no'
  *                      otherwise.
@@ -214,6 +244,7 @@ cat <<EOF
  *   NLBINS: Number of bins supported by the lookup table.
  *   NBINS: Number of small size class bins.
  *   NSIZES: Number of size classes.
+ *   NPSIZES: Number of size classes that are a multiple of (1U << LG_PAGE).
  *   LG_TINY_MAXCLASS: Lg of maximum tiny size class.
  *   LOOKUP_MAXCLASS: Maximum size class included in lookup table.
  *   SMALL_MAXCLASS: Maximum small size class.
@@ -238,6 +269,7 @@ for lg_z in ${lg_zarr} ; do
         echo "#define	NLBINS			${nlbins}"
         echo "#define	NBINS			${nbins}"
         echo "#define	NSIZES			${nsizes}"
+        echo "#define	NPSIZES			${npsizes}"
         echo "#define	LG_TINY_MAXCLASS	${lg_tiny_maxclass}"
         echo "#define	LOOKUP_MAXCLASS		${lookup_maxclass}"
         echo "#define	SMALL_MAXCLASS		${small_maxclass}"
diff --git a/deps/jemalloc/include/jemalloc/internal/smoothstep.h b/deps/jemalloc/include/jemalloc/internal/smoothstep.h
new file mode 100644
index 000000000..c5333ccad
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/smoothstep.h
@@ -0,0 +1,246 @@
+/*
+ * This file was generated by the following command:
+ *   sh smoothstep.sh smoother 200 24 3 15
+ */
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/*
+ * This header defines a precomputed table based on the smoothstep family of
+ * sigmoidal curves (https://en.wikipedia.org/wiki/Smoothstep) that grow from 0
+ * to 1 in 0 <= x <= 1.  The table is stored as integer fixed point values so
+ * that floating point math can be avoided.
+ *
+ *                      3     2
+ *   smoothstep(x) = -2x  + 3x
+ *
+ *                       5      4      3
+ *   smootherstep(x) = 6x  - 15x  + 10x
+ *
+ *                          7      6      5      4
+ *   smootheststep(x) = -20x  + 70x  - 84x  + 35x
+ */
+
+#define	SMOOTHSTEP_VARIANT	"smoother"
+#define	SMOOTHSTEP_NSTEPS	200
+#define	SMOOTHSTEP_BFP		24
+#define	SMOOTHSTEP \
+ /* STEP(step, h,                            x,     y) */ \
+    STEP(   1, UINT64_C(0x0000000000000014), 0.005, 0.000001240643750) \
+    STEP(   2, UINT64_C(0x00000000000000a5), 0.010, 0.000009850600000) \
+    STEP(   3, UINT64_C(0x0000000000000229), 0.015, 0.000032995181250) \
+    STEP(   4, UINT64_C(0x0000000000000516), 0.020, 0.000077619200000) \
+    STEP(   5, UINT64_C(0x00000000000009dc), 0.025, 0.000150449218750) \
+    STEP(   6, UINT64_C(0x00000000000010e8), 0.030, 0.000257995800000) \
+    STEP(   7, UINT64_C(0x0000000000001aa4), 0.035, 0.000406555756250) \
+    STEP(   8, UINT64_C(0x0000000000002777), 0.040, 0.000602214400000) \
+    STEP(   9, UINT64_C(0x00000000000037c2), 0.045, 0.000850847793750) \
+    STEP(  10, UINT64_C(0x0000000000004be6), 0.050, 0.001158125000000) \
+    STEP(  11, UINT64_C(0x000000000000643c), 0.055, 0.001529510331250) \
+    STEP(  12, UINT64_C(0x000000000000811f), 0.060, 0.001970265600000) \
+    STEP(  13, UINT64_C(0x000000000000a2e2), 0.065, 0.002485452368750) \
+    STEP(  14, UINT64_C(0x000000000000c9d8), 0.070, 0.003079934200000) \
+    STEP(  15, UINT64_C(0x000000000000f64f), 0.075, 0.003758378906250) \
+    STEP(  16, UINT64_C(0x0000000000012891), 0.080, 0.004525260800000) \
+    STEP(  17, UINT64_C(0x00000000000160e7), 0.085, 0.005384862943750) \
+    STEP(  18, UINT64_C(0x0000000000019f95), 0.090, 0.006341279400000) \
+    STEP(  19, UINT64_C(0x000000000001e4dc), 0.095, 0.007398417481250) \
+    STEP(  20, UINT64_C(0x00000000000230fc), 0.100, 0.008560000000000) \
+    STEP(  21, UINT64_C(0x0000000000028430), 0.105, 0.009829567518750) \
+    STEP(  22, UINT64_C(0x000000000002deb0), 0.110, 0.011210480600000) \
+    STEP(  23, UINT64_C(0x00000000000340b1), 0.115, 0.012705922056250) \
+    STEP(  24, UINT64_C(0x000000000003aa67), 0.120, 0.014318899200000) \
+    STEP(  25, UINT64_C(0x0000000000041c00), 0.125, 0.016052246093750) \
+    STEP(  26, UINT64_C(0x00000000000495a8), 0.130, 0.017908625800000) \
+    STEP(  27, UINT64_C(0x000000000005178b), 0.135, 0.019890532631250) \
+    STEP(  28, UINT64_C(0x000000000005a1cf), 0.140, 0.022000294400000) \
+    STEP(  29, UINT64_C(0x0000000000063498), 0.145, 0.024240074668750) \
+    STEP(  30, UINT64_C(0x000000000006d009), 0.150, 0.026611875000000) \
+    STEP(  31, UINT64_C(0x000000000007743f), 0.155, 0.029117537206250) \
+    STEP(  32, UINT64_C(0x0000000000082157), 0.160, 0.031758745600000) \
+    STEP(  33, UINT64_C(0x000000000008d76b), 0.165, 0.034537029243750) \
+    STEP(  34, UINT64_C(0x0000000000099691), 0.170, 0.037453764200000) \
+    STEP(  35, UINT64_C(0x00000000000a5edf), 0.175, 0.040510175781250) \
+    STEP(  36, UINT64_C(0x00000000000b3067), 0.180, 0.043707340800000) \
+    STEP(  37, UINT64_C(0x00000000000c0b38), 0.185, 0.047046189818750) \
+    STEP(  38, UINT64_C(0x00000000000cef5e), 0.190, 0.050527509400000) \
+    STEP(  39, UINT64_C(0x00000000000ddce6), 0.195, 0.054151944356250) \
+    STEP(  40, UINT64_C(0x00000000000ed3d8), 0.200, 0.057920000000000) \
+    STEP(  41, UINT64_C(0x00000000000fd439), 0.205, 0.061832044393750) \
+    STEP(  42, UINT64_C(0x000000000010de0e), 0.210, 0.065888310600000) \
+    STEP(  43, UINT64_C(0x000000000011f158), 0.215, 0.070088898931250) \
+    STEP(  44, UINT64_C(0x0000000000130e17), 0.220, 0.074433779200000) \
+    STEP(  45, UINT64_C(0x0000000000143448), 0.225, 0.078922792968750) \
+    STEP(  46, UINT64_C(0x00000000001563e7), 0.230, 0.083555655800000) \
+    STEP(  47, UINT64_C(0x0000000000169cec), 0.235, 0.088331959506250) \
+    STEP(  48, UINT64_C(0x000000000017df4f), 0.240, 0.093251174400000) \
+    STEP(  49, UINT64_C(0x0000000000192b04), 0.245, 0.098312651543750) \
+    STEP(  50, UINT64_C(0x00000000001a8000), 0.250, 0.103515625000000) \
+    STEP(  51, UINT64_C(0x00000000001bde32), 0.255, 0.108859214081250) \
+    STEP(  52, UINT64_C(0x00000000001d458b), 0.260, 0.114342425600000) \
+    STEP(  53, UINT64_C(0x00000000001eb5f8), 0.265, 0.119964156118750) \
+    STEP(  54, UINT64_C(0x0000000000202f65), 0.270, 0.125723194200000) \
+    STEP(  55, UINT64_C(0x000000000021b1bb), 0.275, 0.131618222656250) \
+    STEP(  56, UINT64_C(0x0000000000233ce3), 0.280, 0.137647820800000) \
+    STEP(  57, UINT64_C(0x000000000024d0c3), 0.285, 0.143810466693750) \
+    STEP(  58, UINT64_C(0x0000000000266d40), 0.290, 0.150104539400000) \
+    STEP(  59, UINT64_C(0x000000000028123d), 0.295, 0.156528321231250) \
+    STEP(  60, UINT64_C(0x000000000029bf9c), 0.300, 0.163080000000000) \
+    STEP(  61, UINT64_C(0x00000000002b753d), 0.305, 0.169757671268750) \
+    STEP(  62, UINT64_C(0x00000000002d32fe), 0.310, 0.176559340600000) \
+    STEP(  63, UINT64_C(0x00000000002ef8bc), 0.315, 0.183482925806250) \
+    STEP(  64, UINT64_C(0x000000000030c654), 0.320, 0.190526259200000) \
+    STEP(  65, UINT64_C(0x0000000000329b9f), 0.325, 0.197687089843750) \
+    STEP(  66, UINT64_C(0x0000000000347875), 0.330, 0.204963085800000) \
+    STEP(  67, UINT64_C(0x0000000000365cb0), 0.335, 0.212351836381250) \
+    STEP(  68, UINT64_C(0x0000000000384825), 0.340, 0.219850854400000) \
+    STEP(  69, UINT64_C(0x00000000003a3aa8), 0.345, 0.227457578418750) \
+    STEP(  70, UINT64_C(0x00000000003c340f), 0.350, 0.235169375000000) \
+    STEP(  71, UINT64_C(0x00000000003e342b), 0.355, 0.242983540956250) \
+    STEP(  72, UINT64_C(0x0000000000403ace), 0.360, 0.250897305600000) \
+    STEP(  73, UINT64_C(0x00000000004247c8), 0.365, 0.258907832993750) \
+    STEP(  74, UINT64_C(0x0000000000445ae9), 0.370, 0.267012224200000) \
+    STEP(  75, UINT64_C(0x0000000000467400), 0.375, 0.275207519531250) \
+    STEP(  76, UINT64_C(0x00000000004892d8), 0.380, 0.283490700800000) \
+    STEP(  77, UINT64_C(0x00000000004ab740), 0.385, 0.291858693568750) \
+    STEP(  78, UINT64_C(0x00000000004ce102), 0.390, 0.300308369400000) \
+    STEP(  79, UINT64_C(0x00000000004f0fe9), 0.395, 0.308836548106250) \
+    STEP(  80, UINT64_C(0x00000000005143bf), 0.400, 0.317440000000000) \
+    STEP(  81, UINT64_C(0x0000000000537c4d), 0.405, 0.326115448143750) \
+    STEP(  82, UINT64_C(0x000000000055b95b), 0.410, 0.334859570600000) \
+    STEP(  83, UINT64_C(0x000000000057fab1), 0.415, 0.343669002681250) \
+    STEP(  84, UINT64_C(0x00000000005a4015), 0.420, 0.352540339200000) \
+    STEP(  85, UINT64_C(0x00000000005c894e), 0.425, 0.361470136718750) \
+    STEP(  86, UINT64_C(0x00000000005ed622), 0.430, 0.370454915800000) \
+    STEP(  87, UINT64_C(0x0000000000612655), 0.435, 0.379491163256250) \
+    STEP(  88, UINT64_C(0x00000000006379ac), 0.440, 0.388575334400000) \
+    STEP(  89, UINT64_C(0x000000000065cfeb), 0.445, 0.397703855293750) \
+    STEP(  90, UINT64_C(0x00000000006828d6), 0.450, 0.406873125000000) \
+    STEP(  91, UINT64_C(0x00000000006a842f), 0.455, 0.416079517831250) \
+    STEP(  92, UINT64_C(0x00000000006ce1bb), 0.460, 0.425319385600000) \
+    STEP(  93, UINT64_C(0x00000000006f413a), 0.465, 0.434589059868750) \
+    STEP(  94, UINT64_C(0x000000000071a270), 0.470, 0.443884854200000) \
+    STEP(  95, UINT64_C(0x000000000074051d), 0.475, 0.453203066406250) \
+    STEP(  96, UINT64_C(0x0000000000766905), 0.480, 0.462539980800000) \
+    STEP(  97, UINT64_C(0x000000000078cde7), 0.485, 0.471891870443750) \
+    STEP(  98, UINT64_C(0x00000000007b3387), 0.490, 0.481254999400000) \
+    STEP(  99, UINT64_C(0x00000000007d99a4), 0.495, 0.490625624981250) \
+    STEP( 100, UINT64_C(0x0000000000800000), 0.500, 0.500000000000000) \
+    STEP( 101, UINT64_C(0x000000000082665b), 0.505, 0.509374375018750) \
+    STEP( 102, UINT64_C(0x000000000084cc78), 0.510, 0.518745000600000) \
+    STEP( 103, UINT64_C(0x0000000000873218), 0.515, 0.528108129556250) \
+    STEP( 104, UINT64_C(0x00000000008996fa), 0.520, 0.537460019200000) \
+    STEP( 105, UINT64_C(0x00000000008bfae2), 0.525, 0.546796933593750) \
+    STEP( 106, UINT64_C(0x00000000008e5d8f), 0.530, 0.556115145800000) \
+    STEP( 107, UINT64_C(0x000000000090bec5), 0.535, 0.565410940131250) \
+    STEP( 108, UINT64_C(0x0000000000931e44), 0.540, 0.574680614400000) \
+    STEP( 109, UINT64_C(0x0000000000957bd0), 0.545, 0.583920482168750) \
+    STEP( 110, UINT64_C(0x000000000097d729), 0.550, 0.593126875000000) \
+    STEP( 111, UINT64_C(0x00000000009a3014), 0.555, 0.602296144706250) \
+    STEP( 112, UINT64_C(0x00000000009c8653), 0.560, 0.611424665600000) \
+    STEP( 113, UINT64_C(0x00000000009ed9aa), 0.565, 0.620508836743750) \
+    STEP( 114, UINT64_C(0x0000000000a129dd), 0.570, 0.629545084200000) \
+    STEP( 115, UINT64_C(0x0000000000a376b1), 0.575, 0.638529863281250) \
+    STEP( 116, UINT64_C(0x0000000000a5bfea), 0.580, 0.647459660800000) \
+    STEP( 117, UINT64_C(0x0000000000a8054e), 0.585, 0.656330997318750) \
+    STEP( 118, UINT64_C(0x0000000000aa46a4), 0.590, 0.665140429400000) \
+    STEP( 119, UINT64_C(0x0000000000ac83b2), 0.595, 0.673884551856250) \
+    STEP( 120, UINT64_C(0x0000000000aebc40), 0.600, 0.682560000000000) \
+    STEP( 121, UINT64_C(0x0000000000b0f016), 0.605, 0.691163451893750) \
+    STEP( 122, UINT64_C(0x0000000000b31efd), 0.610, 0.699691630600000) \
+    STEP( 123, UINT64_C(0x0000000000b548bf), 0.615, 0.708141306431250) \
+    STEP( 124, UINT64_C(0x0000000000b76d27), 0.620, 0.716509299200000) \
+    STEP( 125, UINT64_C(0x0000000000b98c00), 0.625, 0.724792480468750) \
+    STEP( 126, UINT64_C(0x0000000000bba516), 0.630, 0.732987775800000) \
+    STEP( 127, UINT64_C(0x0000000000bdb837), 0.635, 0.741092167006250) \
+    STEP( 128, UINT64_C(0x0000000000bfc531), 0.640, 0.749102694400000) \
+    STEP( 129, UINT64_C(0x0000000000c1cbd4), 0.645, 0.757016459043750) \
+    STEP( 130, UINT64_C(0x0000000000c3cbf0), 0.650, 0.764830625000000) \
+    STEP( 131, UINT64_C(0x0000000000c5c557), 0.655, 0.772542421581250) \
+    STEP( 132, UINT64_C(0x0000000000c7b7da), 0.660, 0.780149145600000) \
+    STEP( 133, UINT64_C(0x0000000000c9a34f), 0.665, 0.787648163618750) \
+    STEP( 134, UINT64_C(0x0000000000cb878a), 0.670, 0.795036914200000) \
+    STEP( 135, UINT64_C(0x0000000000cd6460), 0.675, 0.802312910156250) \
+    STEP( 136, UINT64_C(0x0000000000cf39ab), 0.680, 0.809473740800000) \
+    STEP( 137, UINT64_C(0x0000000000d10743), 0.685, 0.816517074193750) \
+    STEP( 138, UINT64_C(0x0000000000d2cd01), 0.690, 0.823440659400000) \
+    STEP( 139, UINT64_C(0x0000000000d48ac2), 0.695, 0.830242328731250) \
+    STEP( 140, UINT64_C(0x0000000000d64063), 0.700, 0.836920000000000) \
+    STEP( 141, UINT64_C(0x0000000000d7edc2), 0.705, 0.843471678768750) \
+    STEP( 142, UINT64_C(0x0000000000d992bf), 0.710, 0.849895460600000) \
+    STEP( 143, UINT64_C(0x0000000000db2f3c), 0.715, 0.856189533306250) \
+    STEP( 144, UINT64_C(0x0000000000dcc31c), 0.720, 0.862352179200000) \
+    STEP( 145, UINT64_C(0x0000000000de4e44), 0.725, 0.868381777343750) \
+    STEP( 146, UINT64_C(0x0000000000dfd09a), 0.730, 0.874276805800000) \
+    STEP( 147, UINT64_C(0x0000000000e14a07), 0.735, 0.880035843881250) \
+    STEP( 148, UINT64_C(0x0000000000e2ba74), 0.740, 0.885657574400000) \
+    STEP( 149, UINT64_C(0x0000000000e421cd), 0.745, 0.891140785918750) \
+    STEP( 150, UINT64_C(0x0000000000e58000), 0.750, 0.896484375000000) \
+    STEP( 151, UINT64_C(0x0000000000e6d4fb), 0.755, 0.901687348456250) \
+    STEP( 152, UINT64_C(0x0000000000e820b0), 0.760, 0.906748825600000) \
+    STEP( 153, UINT64_C(0x0000000000e96313), 0.765, 0.911668040493750) \
+    STEP( 154, UINT64_C(0x0000000000ea9c18), 0.770, 0.916444344200000) \
+    STEP( 155, UINT64_C(0x0000000000ebcbb7), 0.775, 0.921077207031250) \
+    STEP( 156, UINT64_C(0x0000000000ecf1e8), 0.780, 0.925566220800000) \
+    STEP( 157, UINT64_C(0x0000000000ee0ea7), 0.785, 0.929911101068750) \
+    STEP( 158, UINT64_C(0x0000000000ef21f1), 0.790, 0.934111689400000) \
+    STEP( 159, UINT64_C(0x0000000000f02bc6), 0.795, 0.938167955606250) \
+    STEP( 160, UINT64_C(0x0000000000f12c27), 0.800, 0.942080000000000) \
+    STEP( 161, UINT64_C(0x0000000000f22319), 0.805, 0.945848055643750) \
+    STEP( 162, UINT64_C(0x0000000000f310a1), 0.810, 0.949472490600000) \
+    STEP( 163, UINT64_C(0x0000000000f3f4c7), 0.815, 0.952953810181250) \
+    STEP( 164, UINT64_C(0x0000000000f4cf98), 0.820, 0.956292659200000) \
+    STEP( 165, UINT64_C(0x0000000000f5a120), 0.825, 0.959489824218750) \
+    STEP( 166, UINT64_C(0x0000000000f6696e), 0.830, 0.962546235800000) \
+    STEP( 167, UINT64_C(0x0000000000f72894), 0.835, 0.965462970756250) \
+    STEP( 168, UINT64_C(0x0000000000f7dea8), 0.840, 0.968241254400000) \
+    STEP( 169, UINT64_C(0x0000000000f88bc0), 0.845, 0.970882462793750) \
+    STEP( 170, UINT64_C(0x0000000000f92ff6), 0.850, 0.973388125000000) \
+    STEP( 171, UINT64_C(0x0000000000f9cb67), 0.855, 0.975759925331250) \
+    STEP( 172, UINT64_C(0x0000000000fa5e30), 0.860, 0.977999705600000) \
+    STEP( 173, UINT64_C(0x0000000000fae874), 0.865, 0.980109467368750) \
+    STEP( 174, UINT64_C(0x0000000000fb6a57), 0.870, 0.982091374200000) \
+    STEP( 175, UINT64_C(0x0000000000fbe400), 0.875, 0.983947753906250) \
+    STEP( 176, UINT64_C(0x0000000000fc5598), 0.880, 0.985681100800000) \
+    STEP( 177, UINT64_C(0x0000000000fcbf4e), 0.885, 0.987294077943750) \
+    STEP( 178, UINT64_C(0x0000000000fd214f), 0.890, 0.988789519400000) \
+    STEP( 179, UINT64_C(0x0000000000fd7bcf), 0.895, 0.990170432481250) \
+    STEP( 180, UINT64_C(0x0000000000fdcf03), 0.900, 0.991440000000000) \
+    STEP( 181, UINT64_C(0x0000000000fe1b23), 0.905, 0.992601582518750) \
+    STEP( 182, UINT64_C(0x0000000000fe606a), 0.910, 0.993658720600000) \
+    STEP( 183, UINT64_C(0x0000000000fe9f18), 0.915, 0.994615137056250) \
+    STEP( 184, UINT64_C(0x0000000000fed76e), 0.920, 0.995474739200000) \
+    STEP( 185, UINT64_C(0x0000000000ff09b0), 0.925, 0.996241621093750) \
+    STEP( 186, UINT64_C(0x0000000000ff3627), 0.930, 0.996920065800000) \
+    STEP( 187, UINT64_C(0x0000000000ff5d1d), 0.935, 0.997514547631250) \
+    STEP( 188, UINT64_C(0x0000000000ff7ee0), 0.940, 0.998029734400000) \
+    STEP( 189, UINT64_C(0x0000000000ff9bc3), 0.945, 0.998470489668750) \
+    STEP( 190, UINT64_C(0x0000000000ffb419), 0.950, 0.998841875000000) \
+    STEP( 191, UINT64_C(0x0000000000ffc83d), 0.955, 0.999149152206250) \
+    STEP( 192, UINT64_C(0x0000000000ffd888), 0.960, 0.999397785600000) \
+    STEP( 193, UINT64_C(0x0000000000ffe55b), 0.965, 0.999593444243750) \
+    STEP( 194, UINT64_C(0x0000000000ffef17), 0.970, 0.999742004200000) \
+    STEP( 195, UINT64_C(0x0000000000fff623), 0.975, 0.999849550781250) \
+    STEP( 196, UINT64_C(0x0000000000fffae9), 0.980, 0.999922380800000) \
+    STEP( 197, UINT64_C(0x0000000000fffdd6), 0.985, 0.999967004818750) \
+    STEP( 198, UINT64_C(0x0000000000ffff5a), 0.990, 0.999990149400000) \
+    STEP( 199, UINT64_C(0x0000000000ffffeb), 0.995, 0.999998759356250) \
+    STEP( 200, UINT64_C(0x0000000001000000), 1.000, 1.000000000000000) \
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/smoothstep.sh b/deps/jemalloc/include/jemalloc/internal/smoothstep.sh
new file mode 100755
index 000000000..8124693f7
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/smoothstep.sh
@@ -0,0 +1,115 @@
+#!/bin/sh
+#
+# Generate a discrete lookup table for a sigmoid function in the smoothstep
+# family (https://en.wikipedia.org/wiki/Smoothstep), where the lookup table
+# entries correspond to x in [1/nsteps, 2/nsteps, ..., nsteps/nsteps].  Encode
+# the entries using a binary fixed point representation.
+#
+# Usage: smoothstep.sh <variant> <nsteps> <bfp> <xprec> <yprec>
+#
+#        <variant> is in {smooth, smoother, smoothest}.
+#        <nsteps> must be greater than zero.
+#        <bfp> must be in [0..62]; reasonable values are roughly [10..30].
+#        <xprec> is x decimal precision.
+#        <yprec> is y decimal precision.
+
+#set -x
+
+cmd="sh smoothstep.sh $*"
+variant=$1
+nsteps=$2
+bfp=$3
+xprec=$4
+yprec=$5
+
+case "${variant}" in
+  smooth)
+    ;;
+  smoother)
+    ;;
+  smoothest)
+    ;;
+  *)
+    echo "Unsupported variant"
+    exit 1
+    ;;
+esac
+
+smooth() {
+  step=$1
+  y=`echo ${yprec} k ${step} ${nsteps} / sx _2 lx 3 ^ '*' 3 lx 2 ^ '*' + p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  h=`echo ${yprec} k 2 ${bfp} ^ ${y} '*' p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g' | tr '.' ' ' | awk '{print $1}' `
+}
+
+smoother() {
+  step=$1
+  y=`echo ${yprec} k ${step} ${nsteps} / sx 6 lx 5 ^ '*' _15 lx 4 ^ '*' + 10 lx 3 ^ '*' + p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  h=`echo ${yprec} k 2 ${bfp} ^ ${y} '*' p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g' | tr '.' ' ' | awk '{print $1}' `
+}
+
+smoothest() {
+  step=$1
+  y=`echo ${yprec} k ${step} ${nsteps} / sx _20 lx 7 ^ '*' 70 lx 6 ^ '*' + _84 lx 5 ^ '*' + 35 lx 4 ^ '*' + p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  h=`echo ${yprec} k 2 ${bfp} ^ ${y} '*' p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g' | tr '.' ' ' | awk '{print $1}' `
+}
+
+cat <<EOF
+/*
+ * This file was generated by the following command:
+ *   $cmd
+ */
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/*
+ * This header defines a precomputed table based on the smoothstep family of
+ * sigmoidal curves (https://en.wikipedia.org/wiki/Smoothstep) that grow from 0
+ * to 1 in 0 <= x <= 1.  The table is stored as integer fixed point values so
+ * that floating point math can be avoided.
+ *
+ *                      3     2
+ *   smoothstep(x) = -2x  + 3x
+ *
+ *                       5      4      3
+ *   smootherstep(x) = 6x  - 15x  + 10x
+ *
+ *                          7      6      5      4
+ *   smootheststep(x) = -20x  + 70x  - 84x  + 35x
+ */
+
+#define	SMOOTHSTEP_VARIANT	"${variant}"
+#define	SMOOTHSTEP_NSTEPS	${nsteps}
+#define	SMOOTHSTEP_BFP		${bfp}
+#define	SMOOTHSTEP \\
+ /* STEP(step, h,                            x,     y) */ \\
+EOF
+
+s=1
+while [ $s -le $nsteps ] ; do
+  $variant ${s}
+  x=`echo ${xprec} k ${s} ${nsteps} / p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  printf '    STEP(%4d, UINT64_C(0x%016x), %s, %s) \\\n' ${s} ${h} ${x} ${y}
+
+  s=$((s+1))
+done
+echo
+
+cat <<EOF
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+EOF
diff --git a/deps/jemalloc/include/jemalloc/internal/spin.h b/deps/jemalloc/include/jemalloc/internal/spin.h
new file mode 100644
index 000000000..9ef5ceb92
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/spin.h
@@ -0,0 +1,51 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct spin_s spin_t;
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct spin_s {
+	unsigned iteration;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	spin_init(spin_t *spin);
+void	spin_adaptive(spin_t *spin);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_SPIN_C_))
+JEMALLOC_INLINE void
+spin_init(spin_t *spin)
+{
+
+	spin->iteration = 0;
+}
+
+JEMALLOC_INLINE void
+spin_adaptive(spin_t *spin)
+{
+	volatile uint64_t i;
+
+	for (i = 0; i < (KQU(1) << spin->iteration); i++)
+		CPU_SPINWAIT;
+
+	if (spin->iteration < 63)
+		spin->iteration++;
+}
+
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+
diff --git a/deps/jemalloc/include/jemalloc/internal/stats.h b/deps/jemalloc/include/jemalloc/internal/stats.h
index c91dba99d..04e7dae14 100644
--- a/deps/jemalloc/include/jemalloc/internal/stats.h
+++ b/deps/jemalloc/include/jemalloc/internal/stats.h
@@ -103,6 +103,14 @@ struct arena_stats_s {
 	size_t		mapped;
 
 	/*
+	 * Number of bytes currently retained as a side effect of munmap() being
+	 * disabled/bypassed.  Retained bytes are technically mapped (though
+	 * always decommitted or purged), but they are excluded from the mapped
+	 * statistic (above).
+	 */
+	size_t		retained;
+
+	/*
 	 * Total number of purge sweeps, total number of madvise calls made,
 	 * and total pages purged in order to keep dirty unused memory under
 	 * control.
@@ -168,6 +176,9 @@ JEMALLOC_INLINE void
 stats_cactive_add(size_t size)
 {
 
+	assert(size > 0);
+	assert((size & chunksize_mask) == 0);
+
 	atomic_add_z(&stats_cactive, size);
 }
 
@@ -175,6 +186,9 @@ JEMALLOC_INLINE void
 stats_cactive_sub(size_t size)
 {
 
+	assert(size > 0);
+	assert((size & chunksize_mask) == 0);
+
 	atomic_sub_z(&stats_cactive, size);
 }
 #endif
diff --git a/deps/jemalloc/include/jemalloc/internal/tcache.h b/deps/jemalloc/include/jemalloc/internal/tcache.h
index 5079cd266..01ba062de 100644
--- a/deps/jemalloc/include/jemalloc/internal/tcache.h
+++ b/deps/jemalloc/include/jemalloc/internal/tcache.h
@@ -70,13 +70,20 @@ struct tcache_bin_s {
 	int		low_water;	/* Min # cached since last GC. */
 	unsigned	lg_fill_div;	/* Fill (ncached_max >> lg_fill_div). */
 	unsigned	ncached;	/* # of cached objects. */
+	/*
+	 * To make use of adjacent cacheline prefetch, the items in the avail
+	 * stack goes to higher address for newer allocations.  avail points
+	 * just above the available space, which means that
+	 * avail[-ncached, ... -1] are available items and the lowest item will
+	 * be allocated first.
+	 */
 	void		**avail;	/* Stack of available objects. */
 };
 
 struct tcache_s {
 	ql_elm(tcache_t) link;		/* Used for aggregating stats. */
 	uint64_t	prof_accumbytes;/* Cleared after arena_prof_accum(). */
-	unsigned	ev_cnt;		/* Event count since incremental GC. */
+	ticker_t	gc_ticker;	/* Drives incremental GC. */
 	szind_t		next_gc_bin;	/* Next bin to GC. */
 	tcache_bin_t	tbins[1];	/* Dynamically sized. */
 	/*
@@ -108,7 +115,7 @@ extern tcache_bin_info_t	*tcache_bin_info;
  * Number of tcache bins.  There are NBINS small-object bins, plus 0 or more
  * large-object bins.
  */
-extern size_t	nhbins;
+extern unsigned	nhbins;
 
 /* Maximum cached size class. */
 extern size_t	tcache_maxclass;
@@ -123,27 +130,25 @@ extern size_t	tcache_maxclass;
  */
 extern tcaches_t	*tcaches;
 
-size_t	tcache_salloc(const void *ptr);
+size_t	tcache_salloc(tsdn_t *tsdn, const void *ptr);
 void	tcache_event_hard(tsd_t *tsd, tcache_t *tcache);
-void	*tcache_alloc_small_hard(tsd_t *tsd, arena_t *arena, tcache_t *tcache,
-    tcache_bin_t *tbin, szind_t binind);
+void	*tcache_alloc_small_hard(tsdn_t *tsdn, arena_t *arena, tcache_t *tcache,
+    tcache_bin_t *tbin, szind_t binind, bool *tcache_success);
 void	tcache_bin_flush_small(tsd_t *tsd, tcache_t *tcache, tcache_bin_t *tbin,
     szind_t binind, unsigned rem);
 void	tcache_bin_flush_large(tsd_t *tsd, tcache_bin_t *tbin, szind_t binind,
     unsigned rem, tcache_t *tcache);
-void	tcache_arena_associate(tcache_t *tcache, arena_t *arena);
-void	tcache_arena_reassociate(tcache_t *tcache, arena_t *oldarena,
-    arena_t *newarena);
-void	tcache_arena_dissociate(tcache_t *tcache, arena_t *arena);
+void	tcache_arena_reassociate(tsdn_t *tsdn, tcache_t *tcache,
+    arena_t *oldarena, arena_t *newarena);
 tcache_t *tcache_get_hard(tsd_t *tsd);
-tcache_t *tcache_create(tsd_t *tsd, arena_t *arena);
+tcache_t *tcache_create(tsdn_t *tsdn, arena_t *arena);
 void	tcache_cleanup(tsd_t *tsd);
 void	tcache_enabled_cleanup(tsd_t *tsd);
-void	tcache_stats_merge(tcache_t *tcache, arena_t *arena);
+void	tcache_stats_merge(tsdn_t *tsdn, tcache_t *tcache, arena_t *arena);
 bool	tcaches_create(tsd_t *tsd, unsigned *r_ind);
 void	tcaches_flush(tsd_t *tsd, unsigned ind);
 void	tcaches_destroy(tsd_t *tsd, unsigned ind);
-bool	tcache_boot(void);
+bool	tcache_boot(tsdn_t *tsdn);
 
 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
@@ -155,15 +160,15 @@ void	tcache_flush(void);
 bool	tcache_enabled_get(void);
 tcache_t *tcache_get(tsd_t *tsd, bool create);
 void	tcache_enabled_set(bool enabled);
-void	*tcache_alloc_easy(tcache_bin_t *tbin);
+void	*tcache_alloc_easy(tcache_bin_t *tbin, bool *tcache_success);
 void	*tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache,
-    size_t size, bool zero);
+    size_t size, szind_t ind, bool zero, bool slow_path);
 void	*tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache,
-    size_t size, bool zero);
+    size_t size, szind_t ind, bool zero, bool slow_path);
 void	tcache_dalloc_small(tsd_t *tsd, tcache_t *tcache, void *ptr,
-    szind_t binind);
+    szind_t binind, bool slow_path);
 void	tcache_dalloc_large(tsd_t *tsd, tcache_t *tcache, void *ptr,
-    size_t size);
+    size_t size, bool slow_path);
 tcache_t	*tcaches_get(tsd_t *tsd, unsigned ind);
 #endif
 
@@ -240,51 +245,74 @@ tcache_event(tsd_t *tsd, tcache_t *tcache)
 	if (TCACHE_GC_INCR == 0)
 		return;
 
-	tcache->ev_cnt++;
-	assert(tcache->ev_cnt <= TCACHE_GC_INCR);
-	if (unlikely(tcache->ev_cnt == TCACHE_GC_INCR))
+	if (unlikely(ticker_tick(&tcache->gc_ticker)))
 		tcache_event_hard(tsd, tcache);
 }
 
 JEMALLOC_ALWAYS_INLINE void *
-tcache_alloc_easy(tcache_bin_t *tbin)
+tcache_alloc_easy(tcache_bin_t *tbin, bool *tcache_success)
 {
 	void *ret;
 
 	if (unlikely(tbin->ncached == 0)) {
 		tbin->low_water = -1;
+		*tcache_success = false;
 		return (NULL);
 	}
+	/*
+	 * tcache_success (instead of ret) should be checked upon the return of
+	 * this function.  We avoid checking (ret == NULL) because there is
+	 * never a null stored on the avail stack (which is unknown to the
+	 * compiler), and eagerly checking ret would cause pipeline stall
+	 * (waiting for the cacheline).
+	 */
+	*tcache_success = true;
+	ret = *(tbin->avail - tbin->ncached);
 	tbin->ncached--;
+
 	if (unlikely((int)tbin->ncached < tbin->low_water))
 		tbin->low_water = tbin->ncached;
-	ret = tbin->avail[tbin->ncached];
+
 	return (ret);
 }
 
 JEMALLOC_ALWAYS_INLINE void *
 tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
-    bool zero)
+    szind_t binind, bool zero, bool slow_path)
 {
 	void *ret;
-	szind_t binind;
-	size_t usize;
 	tcache_bin_t *tbin;
+	bool tcache_success;
+	size_t usize JEMALLOC_CC_SILENCE_INIT(0);
 
-	binind = size2index(size);
 	assert(binind < NBINS);
 	tbin = &tcache->tbins[binind];
-	usize = index2size(binind);
-	ret = tcache_alloc_easy(tbin);
-	if (unlikely(ret == NULL)) {
-		ret = tcache_alloc_small_hard(tsd, arena, tcache, tbin, binind);
-		if (ret == NULL)
+	ret = tcache_alloc_easy(tbin, &tcache_success);
+	assert(tcache_success == (ret != NULL));
+	if (unlikely(!tcache_success)) {
+		bool tcache_hard_success;
+		arena = arena_choose(tsd, arena);
+		if (unlikely(arena == NULL))
+			return (NULL);
+
+		ret = tcache_alloc_small_hard(tsd_tsdn(tsd), arena, tcache,
+		    tbin, binind, &tcache_hard_success);
+		if (tcache_hard_success == false)
 			return (NULL);
 	}
-	assert(tcache_salloc(ret) == usize);
+
+	assert(ret);
+	/*
+	 * Only compute usize if required.  The checks in the following if
+	 * statement are all static.
+	 */
+	if (config_prof || (slow_path && config_fill) || unlikely(zero)) {
+		usize = index2size(binind);
+		assert(tcache_salloc(tsd_tsdn(tsd), ret) == usize);
+	}
 
 	if (likely(!zero)) {
-		if (config_fill) {
+		if (slow_path && config_fill) {
 			if (unlikely(opt_junk_alloc)) {
 				arena_alloc_junk_small(ret,
 				    &arena_bin_info[binind], false);
@@ -292,7 +320,7 @@ tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 				memset(ret, 0, usize);
 		}
 	} else {
-		if (config_fill && unlikely(opt_junk_alloc)) {
+		if (slow_path && config_fill && unlikely(opt_junk_alloc)) {
 			arena_alloc_junk_small(ret, &arena_bin_info[binind],
 			    true);
 		}
@@ -309,28 +337,38 @@ tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 
 JEMALLOC_ALWAYS_INLINE void *
 tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
-    bool zero)
+    szind_t binind, bool zero, bool slow_path)
 {
 	void *ret;
-	szind_t binind;
-	size_t usize;
 	tcache_bin_t *tbin;
+	bool tcache_success;
 
-	binind = size2index(size);
-	usize = index2size(binind);
-	assert(usize <= tcache_maxclass);
 	assert(binind < nhbins);
 	tbin = &tcache->tbins[binind];
-	ret = tcache_alloc_easy(tbin);
-	if (unlikely(ret == NULL)) {
+	ret = tcache_alloc_easy(tbin, &tcache_success);
+	assert(tcache_success == (ret != NULL));
+	if (unlikely(!tcache_success)) {
 		/*
 		 * Only allocate one large object at a time, because it's quite
 		 * expensive to create one and not use it.
 		 */
-		ret = arena_malloc_large(arena, usize, zero);
+		arena = arena_choose(tsd, arena);
+		if (unlikely(arena == NULL))
+			return (NULL);
+
+		ret = arena_malloc_large(tsd_tsdn(tsd), arena, binind, zero);
 		if (ret == NULL)
 			return (NULL);
 	} else {
+		size_t usize JEMALLOC_CC_SILENCE_INIT(0);
+
+		/* Only compute usize on demand */
+		if (config_prof || (slow_path && config_fill) ||
+		    unlikely(zero)) {
+			usize = index2size(binind);
+			assert(usize <= tcache_maxclass);
+		}
+
 		if (config_prof && usize == LARGE_MINCLASS) {
 			arena_chunk_t *chunk =
 			    (arena_chunk_t *)CHUNK_ADDR2BASE(ret);
@@ -340,10 +378,11 @@ tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 			    BININD_INVALID);
 		}
 		if (likely(!zero)) {
-			if (config_fill) {
-				if (unlikely(opt_junk_alloc))
-					memset(ret, 0xa5, usize);
-				else if (unlikely(opt_zero))
+			if (slow_path && config_fill) {
+				if (unlikely(opt_junk_alloc)) {
+					memset(ret, JEMALLOC_ALLOC_JUNK,
+					    usize);
+				} else if (unlikely(opt_zero))
 					memset(ret, 0, usize);
 			}
 		} else
@@ -360,14 +399,15 @@ tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 }
 
 JEMALLOC_ALWAYS_INLINE void
-tcache_dalloc_small(tsd_t *tsd, tcache_t *tcache, void *ptr, szind_t binind)
+tcache_dalloc_small(tsd_t *tsd, tcache_t *tcache, void *ptr, szind_t binind,
+    bool slow_path)
 {
 	tcache_bin_t *tbin;
 	tcache_bin_info_t *tbin_info;
 
-	assert(tcache_salloc(ptr) <= SMALL_MAXCLASS);
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) <= SMALL_MAXCLASS);
 
-	if (config_fill && unlikely(opt_junk_free))
+	if (slow_path && config_fill && unlikely(opt_junk_free))
 		arena_dalloc_junk_small(ptr, &arena_bin_info[binind]);
 
 	tbin = &tcache->tbins[binind];
@@ -377,26 +417,27 @@ tcache_dalloc_small(tsd_t *tsd, tcache_t *tcache, void *ptr, szind_t binind)
 		    (tbin_info->ncached_max >> 1));
 	}
 	assert(tbin->ncached < tbin_info->ncached_max);
-	tbin->avail[tbin->ncached] = ptr;
 	tbin->ncached++;
+	*(tbin->avail - tbin->ncached) = ptr;
 
 	tcache_event(tsd, tcache);
 }
 
 JEMALLOC_ALWAYS_INLINE void
-tcache_dalloc_large(tsd_t *tsd, tcache_t *tcache, void *ptr, size_t size)
+tcache_dalloc_large(tsd_t *tsd, tcache_t *tcache, void *ptr, size_t size,
+    bool slow_path)
 {
 	szind_t binind;
 	tcache_bin_t *tbin;
 	tcache_bin_info_t *tbin_info;
 
 	assert((size & PAGE_MASK) == 0);
-	assert(tcache_salloc(ptr) > SMALL_MAXCLASS);
-	assert(tcache_salloc(ptr) <= tcache_maxclass);
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) > SMALL_MAXCLASS);
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) <= tcache_maxclass);
 
 	binind = size2index(size);
 
-	if (config_fill && unlikely(opt_junk_free))
+	if (slow_path && config_fill && unlikely(opt_junk_free))
 		arena_dalloc_junk_large(ptr, size);
 
 	tbin = &tcache->tbins[binind];
@@ -406,8 +447,8 @@ tcache_dalloc_large(tsd_t *tsd, tcache_t *tcache, void *ptr, size_t size)
 		    (tbin_info->ncached_max >> 1), tcache);
 	}
 	assert(tbin->ncached < tbin_info->ncached_max);
-	tbin->avail[tbin->ncached] = ptr;
 	tbin->ncached++;
+	*(tbin->avail - tbin->ncached) = ptr;
 
 	tcache_event(tsd, tcache);
 }
@@ -416,8 +457,10 @@ JEMALLOC_ALWAYS_INLINE tcache_t *
 tcaches_get(tsd_t *tsd, unsigned ind)
 {
 	tcaches_t *elm = &tcaches[ind];
-	if (unlikely(elm->tcache == NULL))
-		elm->tcache = tcache_create(tsd, arena_choose(tsd, NULL));
+	if (unlikely(elm->tcache == NULL)) {
+		elm->tcache = tcache_create(tsd_tsdn(tsd), arena_choose(tsd,
+		    NULL));
+	}
 	return (elm->tcache);
 }
 #endif
diff --git a/deps/jemalloc/include/jemalloc/internal/ticker.h b/deps/jemalloc/include/jemalloc/internal/ticker.h
new file mode 100644
index 000000000..4696e56d2
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/ticker.h
@@ -0,0 +1,75 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct ticker_s ticker_t;
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct ticker_s {
+	int32_t	tick;
+	int32_t	nticks;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	ticker_init(ticker_t *ticker, int32_t nticks);
+void	ticker_copy(ticker_t *ticker, const ticker_t *other);
+int32_t	ticker_read(const ticker_t *ticker);
+bool	ticker_ticks(ticker_t *ticker, int32_t nticks);
+bool	ticker_tick(ticker_t *ticker);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_TICKER_C_))
+JEMALLOC_INLINE void
+ticker_init(ticker_t *ticker, int32_t nticks)
+{
+
+	ticker->tick = nticks;
+	ticker->nticks = nticks;
+}
+
+JEMALLOC_INLINE void
+ticker_copy(ticker_t *ticker, const ticker_t *other)
+{
+
+	*ticker = *other;
+}
+
+JEMALLOC_INLINE int32_t
+ticker_read(const ticker_t *ticker)
+{
+
+	return (ticker->tick);
+}
+
+JEMALLOC_INLINE bool
+ticker_ticks(ticker_t *ticker, int32_t nticks)
+{
+
+	if (unlikely(ticker->tick < nticks)) {
+		ticker->tick = ticker->nticks;
+		return (true);
+	}
+	ticker->tick -= nticks;
+	return(false);
+}
+
+JEMALLOC_INLINE bool
+ticker_tick(ticker_t *ticker)
+{
+
+	return (ticker_ticks(ticker, 1));
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/internal/tsd.h b/deps/jemalloc/include/jemalloc/internal/tsd.h
index eed7aa013..9055acafd 100644
--- a/deps/jemalloc/include/jemalloc/internal/tsd.h
+++ b/deps/jemalloc/include/jemalloc/internal/tsd.h
@@ -13,6 +13,9 @@ typedef struct tsd_init_head_s tsd_init_head_t;
 #endif
 
 typedef struct tsd_s tsd_t;
+typedef struct tsdn_s tsdn_t;
+
+#define	TSDN_NULL	((tsdn_t *)0)
 
 typedef enum {
 	tsd_state_uninitialized,
@@ -44,7 +47,8 @@ typedef enum {
  * The result is a set of generated functions, e.g.:
  *
  *   bool example_tsd_boot(void) {...}
- *   example_t *example_tsd_get() {...}
+ *   bool example_tsd_booted_get(void) {...}
+ *   example_t *example_tsd_get(bool init) {...}
  *   void example_tsd_set(example_t *val) {...}
  *
  * Note that all of the functions deal in terms of (a_type *) rather than
@@ -98,8 +102,10 @@ a_attr void								\
 a_name##tsd_boot1(void);						\
 a_attr bool								\
 a_name##tsd_boot(void);							\
+a_attr bool								\
+a_name##tsd_booted_get(void);						\
 a_attr a_type *								\
-a_name##tsd_get(void);							\
+a_name##tsd_get(bool init);						\
 a_attr void								\
 a_name##tsd_set(a_type *val);
 
@@ -201,9 +207,21 @@ a_name##tsd_boot(void)							\
 									\
 	return (a_name##tsd_boot0());					\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (false);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 									\
 	assert(a_name##tsd_booted);					\
@@ -246,9 +264,21 @@ a_name##tsd_boot(void)							\
 									\
 	return (a_name##tsd_boot0());					\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (false);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 									\
 	assert(a_name##tsd_booted);					\
@@ -307,14 +337,14 @@ a_name##tsd_wrapper_set(a_name##tsd_wrapper_t *wrapper)			\
 	}								\
 }									\
 a_attr a_name##tsd_wrapper_t *						\
-a_name##tsd_wrapper_get(void)						\
+a_name##tsd_wrapper_get(bool init)					\
 {									\
 	DWORD error = GetLastError();					\
 	a_name##tsd_wrapper_t *wrapper = (a_name##tsd_wrapper_t *)	\
 	    TlsGetValue(a_name##tsd_tsd);				\
 	SetLastError(error);						\
 									\
-	if (unlikely(wrapper == NULL)) {				\
+	if (init && unlikely(wrapper == NULL)) {			\
 		wrapper = (a_name##tsd_wrapper_t *)			\
 		    malloc_tsd_malloc(sizeof(a_name##tsd_wrapper_t));	\
 		if (wrapper == NULL) {					\
@@ -368,14 +398,28 @@ a_name##tsd_boot(void)							\
 	a_name##tsd_boot1();						\
 	return (false);							\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (true);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(init);			\
+	if (a_name##tsd_get_allocates() && !init && wrapper == NULL)	\
+		return (NULL);						\
 	return (&wrapper->val);						\
 }									\
 a_attr void								\
@@ -384,7 +428,7 @@ a_name##tsd_set(a_type *val)						\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(true);			\
 	wrapper->val = *(val);						\
 	if (a_cleanup != malloc_tsd_no_cleanup)				\
 		wrapper->initialized = true;				\
@@ -428,12 +472,12 @@ a_name##tsd_wrapper_set(a_name##tsd_wrapper_t *wrapper)			\
 	}								\
 }									\
 a_attr a_name##tsd_wrapper_t *						\
-a_name##tsd_wrapper_get(void)						\
+a_name##tsd_wrapper_get(bool init)					\
 {									\
 	a_name##tsd_wrapper_t *wrapper = (a_name##tsd_wrapper_t *)	\
 	    pthread_getspecific(a_name##tsd_tsd);			\
 									\
-	if (unlikely(wrapper == NULL)) {				\
+	if (init && unlikely(wrapper == NULL)) {			\
 		tsd_init_block_t block;					\
 		wrapper = tsd_init_check_recursion(			\
 		    &a_name##tsd_init_head, &block);			\
@@ -490,14 +534,28 @@ a_name##tsd_boot(void)							\
 	a_name##tsd_boot1();						\
 	return (false);							\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (true);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(init);			\
+	if (a_name##tsd_get_allocates() && !init && wrapper == NULL)	\
+		return (NULL);						\
 	return (&wrapper->val);						\
 }									\
 a_attr void								\
@@ -506,7 +564,7 @@ a_name##tsd_set(a_type *val)						\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(true);			\
 	wrapper->val = *(val);						\
 	if (a_cleanup != malloc_tsd_no_cleanup)				\
 		wrapper->initialized = true;				\
@@ -536,12 +594,15 @@ struct tsd_init_head_s {
     O(thread_allocated,		uint64_t)				\
     O(thread_deallocated,	uint64_t)				\
     O(prof_tdata,		prof_tdata_t *)				\
+    O(iarena,			arena_t *)				\
     O(arena,			arena_t *)				\
-    O(arenas_cache,		arena_t **)				\
-    O(narenas_cache,		unsigned)				\
-    O(arenas_cache_bypass,	bool)					\
+    O(arenas_tdata,		arena_tdata_t *)			\
+    O(narenas_tdata,		unsigned)				\
+    O(arenas_tdata_bypass,	bool)					\
     O(tcache_enabled,		tcache_enabled_t)			\
     O(quarantine,		quarantine_t *)				\
+    O(witnesses,		witness_list_t)				\
+    O(witness_fork,		bool)					\
 
 #define	TSD_INITIALIZER {						\
     tsd_state_uninitialized,						\
@@ -551,10 +612,13 @@ struct tsd_init_head_s {
     NULL,								\
     NULL,								\
     NULL,								\
+    NULL,								\
     0,									\
     false,								\
     tcache_enabled_default,						\
-    NULL								\
+    NULL,								\
+    ql_head_initializer(witnesses),					\
+    false								\
 }
 
 struct tsd_s {
@@ -565,6 +629,15 @@ MALLOC_TSD
 #undef O
 };
 
+/*
+ * Wrapper around tsd_t that makes it possible to avoid implicit conversion
+ * between tsd_t and tsdn_t, where tsdn_t is "nullable" and has to be
+ * explicitly converted to tsd_t, which is non-nullable.
+ */
+struct tsdn_s {
+	tsd_t	tsd;
+};
+
 static const tsd_t tsd_initializer = TSD_INITIALIZER;
 
 malloc_tsd_types(, tsd_t)
@@ -577,7 +650,7 @@ void	*malloc_tsd_malloc(size_t size);
 void	malloc_tsd_dalloc(void *wrapper);
 void	malloc_tsd_no_cleanup(void *arg);
 void	malloc_tsd_cleanup_register(bool (*f)(void));
-bool	malloc_tsd_boot0(void);
+tsd_t	*malloc_tsd_boot0(void);
 void	malloc_tsd_boot1(void);
 #if (!defined(JEMALLOC_MALLOC_THREAD_CLEANUP) && !defined(JEMALLOC_TLS) && \
     !defined(_WIN32))
@@ -594,7 +667,9 @@ void	tsd_cleanup(void *arg);
 #ifndef JEMALLOC_ENABLE_INLINE
 malloc_tsd_protos(JEMALLOC_ATTR(unused), , tsd_t)
 
+tsd_t	*tsd_fetch_impl(bool init);
 tsd_t	*tsd_fetch(void);
+tsdn_t	*tsd_tsdn(tsd_t *tsd);
 bool	tsd_nominal(tsd_t *tsd);
 #define	O(n, t)								\
 t	*tsd_##n##p_get(tsd_t *tsd);					\
@@ -602,6 +677,9 @@ t	tsd_##n##_get(tsd_t *tsd);					\
 void	tsd_##n##_set(tsd_t *tsd, t n);
 MALLOC_TSD
 #undef O
+tsdn_t	*tsdn_fetch(void);
+bool	tsdn_null(const tsdn_t *tsdn);
+tsd_t	*tsdn_tsd(tsdn_t *tsdn);
 #endif
 
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_TSD_C_))
@@ -609,9 +687,13 @@ malloc_tsd_externs(, tsd_t)
 malloc_tsd_funcs(JEMALLOC_ALWAYS_INLINE, , tsd_t, tsd_initializer, tsd_cleanup)
 
 JEMALLOC_ALWAYS_INLINE tsd_t *
-tsd_fetch(void)
+tsd_fetch_impl(bool init)
 {
-	tsd_t *tsd = tsd_get();
+	tsd_t *tsd = tsd_get(init);
+
+	if (!init && tsd_get_allocates() && tsd == NULL)
+		return (NULL);
+	assert(tsd != NULL);
 
 	if (unlikely(tsd->state != tsd_state_nominal)) {
 		if (tsd->state == tsd_state_uninitialized) {
@@ -628,6 +710,20 @@ tsd_fetch(void)
 	return (tsd);
 }
 
+JEMALLOC_ALWAYS_INLINE tsd_t *
+tsd_fetch(void)
+{
+
+	return (tsd_fetch_impl(true));
+}
+
+JEMALLOC_ALWAYS_INLINE tsdn_t *
+tsd_tsdn(tsd_t *tsd)
+{
+
+	return ((tsdn_t *)tsd);
+}
+
 JEMALLOC_INLINE bool
 tsd_nominal(tsd_t *tsd)
 {
@@ -659,6 +755,32 @@ tsd_##n##_set(tsd_t *tsd, t n)						\
 }
 MALLOC_TSD
 #undef O
+
+JEMALLOC_ALWAYS_INLINE tsdn_t *
+tsdn_fetch(void)
+{
+
+	if (!tsd_booted_get())
+		return (NULL);
+
+	return (tsd_tsdn(tsd_fetch_impl(false)));
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+tsdn_null(const tsdn_t *tsdn)
+{
+
+	return (tsdn == NULL);
+}
+
+JEMALLOC_ALWAYS_INLINE tsd_t *
+tsdn_tsd(tsdn_t *tsdn)
+{
+
+	assert(!tsdn_null(tsdn));
+
+	return (&tsdn->tsd);
+}
 #endif
 
 #endif /* JEMALLOC_H_INLINES */
diff --git a/deps/jemalloc/include/jemalloc/internal/util.h b/deps/jemalloc/include/jemalloc/internal/util.h
index b2ea740fd..4b56d652e 100644
--- a/deps/jemalloc/include/jemalloc/internal/util.h
+++ b/deps/jemalloc/include/jemalloc/internal/util.h
@@ -40,6 +40,14 @@
  */
 #define	MALLOC_PRINTF_BUFSIZE	4096
 
+/* Junk fill patterns. */
+#ifndef JEMALLOC_ALLOC_JUNK
+#  define JEMALLOC_ALLOC_JUNK	((uint8_t)0xa5)
+#endif
+#ifndef JEMALLOC_FREE_JUNK
+#  define JEMALLOC_FREE_JUNK	((uint8_t)0x5a)
+#endif
+
 /*
  * Wrap a cpp argument that contains commas such that it isn't broken up into
  * multiple arguments.
@@ -57,73 +65,21 @@
 #	define JEMALLOC_CC_SILENCE_INIT(v)
 #endif
 
-#define	JEMALLOC_GNUC_PREREQ(major, minor)				\
-    (!defined(__clang__) &&						\
-    (__GNUC__ > (major) || (__GNUC__ == (major) && __GNUC_MINOR__ >= (minor))))
-#ifndef __has_builtin
-#  define __has_builtin(builtin) (0)
-#endif
-#define	JEMALLOC_CLANG_HAS_BUILTIN(builtin)				\
-    (defined(__clang__) && __has_builtin(builtin))
-
 #ifdef __GNUC__
 #	define likely(x)   __builtin_expect(!!(x), 1)
 #	define unlikely(x) __builtin_expect(!!(x), 0)
-#  if JEMALLOC_GNUC_PREREQ(4, 6) ||					\
-      JEMALLOC_CLANG_HAS_BUILTIN(__builtin_unreachable)
-#	define unreachable() __builtin_unreachable()
-#  else
-#	define unreachable()
-#  endif
 #else
 #	define likely(x)   !!(x)
 #	define unlikely(x) !!(x)
-#	define unreachable()
 #endif
 
-/*
- * Define a custom assert() in order to reduce the chances of deadlock during
- * assertion failure.
- */
-#ifndef assert
-#define	assert(e) do {							\
-	if (unlikely(config_debug && !(e))) {				\
-		malloc_printf(						\
-		    "<jemalloc>: %s:%d: Failed assertion: \"%s\"\n",	\
-		    __FILE__, __LINE__, #e);				\
-		abort();						\
-	}								\
-} while (0)
+#if !defined(JEMALLOC_INTERNAL_UNREACHABLE)
+#  error JEMALLOC_INTERNAL_UNREACHABLE should have been defined by configure
 #endif
 
-#ifndef not_reached
-#define	not_reached() do {						\
-	if (config_debug) {						\
-		malloc_printf(						\
-		    "<jemalloc>: %s:%d: Unreachable code reached\n",	\
-		    __FILE__, __LINE__);				\
-		abort();						\
-	}								\
-	unreachable();							\
-} while (0)
-#endif
+#define unreachable() JEMALLOC_INTERNAL_UNREACHABLE()
 
-#ifndef not_implemented
-#define	not_implemented() do {						\
-	if (config_debug) {						\
-		malloc_printf("<jemalloc>: %s:%d: Not implemented\n",	\
-		    __FILE__, __LINE__);				\
-		abort();						\
-	}								\
-} while (0)
-#endif
-
-#ifndef assert_not_implemented
-#define	assert_not_implemented(e) do {					\
-	if (unlikely(config_debug && !(e)))				\
-		not_implemented();					\
-} while (0)
-#endif
+#include "jemalloc/internal/assert.h"
 
 /* Use to assert a particular configuration, e.g., cassert(config_debug). */
 #define	cassert(c) do {							\
@@ -148,9 +104,9 @@ void	malloc_write(const char *s);
  * malloc_vsnprintf() supports a subset of snprintf(3) that avoids floating
  * point math.
  */
-int	malloc_vsnprintf(char *str, size_t size, const char *format,
+size_t	malloc_vsnprintf(char *str, size_t size, const char *format,
     va_list ap);
-int	malloc_snprintf(char *str, size_t size, const char *format, ...)
+size_t	malloc_snprintf(char *str, size_t size, const char *format, ...)
     JEMALLOC_FORMAT_PRINTF(3, 4);
 void	malloc_vcprintf(void (*write_cb)(void *, const char *), void *cbopaque,
     const char *format, va_list ap);
@@ -163,10 +119,16 @@ void	malloc_printf(const char *format, ...) JEMALLOC_FORMAT_PRINTF(1, 2);
 #ifdef JEMALLOC_H_INLINES
 
 #ifndef JEMALLOC_ENABLE_INLINE
-int	jemalloc_ffsl(long bitmap);
-int	jemalloc_ffs(int bitmap);
-size_t	pow2_ceil(size_t x);
-size_t	lg_floor(size_t x);
+unsigned	ffs_llu(unsigned long long bitmap);
+unsigned	ffs_lu(unsigned long bitmap);
+unsigned	ffs_u(unsigned bitmap);
+unsigned	ffs_zu(size_t bitmap);
+unsigned	ffs_u64(uint64_t bitmap);
+unsigned	ffs_u32(uint32_t bitmap);
+uint64_t	pow2_ceil_u64(uint64_t x);
+uint32_t	pow2_ceil_u32(uint32_t x);
+size_t	pow2_ceil_zu(size_t x);
+unsigned	lg_floor(size_t x);
 void	set_errno(int errnum);
 int	get_errno(void);
 #endif
@@ -174,27 +136,74 @@ int	get_errno(void);
 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_UTIL_C_))
 
 /* Sanity check. */
-#if !defined(JEMALLOC_INTERNAL_FFSL) || !defined(JEMALLOC_INTERNAL_FFS)
-#  error Both JEMALLOC_INTERNAL_FFSL && JEMALLOC_INTERNAL_FFS should have been defined by configure
+#if !defined(JEMALLOC_INTERNAL_FFSLL) || !defined(JEMALLOC_INTERNAL_FFSL) \
+    || !defined(JEMALLOC_INTERNAL_FFS)
+#  error JEMALLOC_INTERNAL_FFS{,L,LL} should have been defined by configure
 #endif
 
-JEMALLOC_ALWAYS_INLINE int
-jemalloc_ffsl(long bitmap)
+JEMALLOC_ALWAYS_INLINE unsigned
+ffs_llu(unsigned long long bitmap)
+{
+
+	return (JEMALLOC_INTERNAL_FFSLL(bitmap));
+}
+
+JEMALLOC_ALWAYS_INLINE unsigned
+ffs_lu(unsigned long bitmap)
 {
 
 	return (JEMALLOC_INTERNAL_FFSL(bitmap));
 }
 
-JEMALLOC_ALWAYS_INLINE int
-jemalloc_ffs(int bitmap)
+JEMALLOC_ALWAYS_INLINE unsigned
+ffs_u(unsigned bitmap)
 {
 
 	return (JEMALLOC_INTERNAL_FFS(bitmap));
 }
 
-/* Compute the smallest power of 2 that is >= x. */
-JEMALLOC_INLINE size_t
-pow2_ceil(size_t x)
+JEMALLOC_ALWAYS_INLINE unsigned
+ffs_zu(size_t bitmap)
+{
+
+#if LG_SIZEOF_PTR == LG_SIZEOF_INT
+	return (ffs_u(bitmap));
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG
+	return (ffs_lu(bitmap));
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG_LONG
+	return (ffs_llu(bitmap));
+#else
+#error No implementation for size_t ffs()
+#endif
+}
+
+JEMALLOC_ALWAYS_INLINE unsigned
+ffs_u64(uint64_t bitmap)
+{
+
+#if LG_SIZEOF_LONG == 3
+	return (ffs_lu(bitmap));
+#elif LG_SIZEOF_LONG_LONG == 3
+	return (ffs_llu(bitmap));
+#else
+#error No implementation for 64-bit ffs()
+#endif
+}
+
+JEMALLOC_ALWAYS_INLINE unsigned
+ffs_u32(uint32_t bitmap)
+{
+
+#if LG_SIZEOF_INT == 2
+	return (ffs_u(bitmap));
+#else
+#error No implementation for 32-bit ffs()
+#endif
+	return (ffs_u(bitmap));
+}
+
+JEMALLOC_INLINE uint64_t
+pow2_ceil_u64(uint64_t x)
 {
 
 	x--;
@@ -203,15 +212,39 @@ pow2_ceil(size_t x)
 	x |= x >> 4;
 	x |= x >> 8;
 	x |= x >> 16;
-#if (LG_SIZEOF_PTR == 3)
 	x |= x >> 32;
-#endif
 	x++;
 	return (x);
 }
 
-#if (defined(__i386__) || defined(__amd64__) || defined(__x86_64__))
+JEMALLOC_INLINE uint32_t
+pow2_ceil_u32(uint32_t x)
+{
+
+	x--;
+	x |= x >> 1;
+	x |= x >> 2;
+	x |= x >> 4;
+	x |= x >> 8;
+	x |= x >> 16;
+	x++;
+	return (x);
+}
+
+/* Compute the smallest power of 2 that is >= x. */
 JEMALLOC_INLINE size_t
+pow2_ceil_zu(size_t x)
+{
+
+#if (LG_SIZEOF_PTR == 3)
+	return (pow2_ceil_u64(x));
+#else
+	return (pow2_ceil_u32(x));
+#endif
+}
+
+#if (defined(__i386__) || defined(__amd64__) || defined(__x86_64__))
+JEMALLOC_INLINE unsigned
 lg_floor(size_t x)
 {
 	size_t ret;
@@ -222,10 +255,11 @@ lg_floor(size_t x)
 	    : "=r"(ret) // Outputs.
 	    : "r"(x)    // Inputs.
 	    );
-	return (ret);
+	assert(ret < UINT_MAX);
+	return ((unsigned)ret);
 }
 #elif (defined(_MSC_VER))
-JEMALLOC_INLINE size_t
+JEMALLOC_INLINE unsigned
 lg_floor(size_t x)
 {
 	unsigned long ret;
@@ -237,12 +271,13 @@ lg_floor(size_t x)
 #elif (LG_SIZEOF_PTR == 2)
 	_BitScanReverse(&ret, x);
 #else
-#  error "Unsupported type sizes for lg_floor()"
+#  error "Unsupported type size for lg_floor()"
 #endif
-	return (ret);
+	assert(ret < UINT_MAX);
+	return ((unsigned)ret);
 }
 #elif (defined(JEMALLOC_HAVE_BUILTIN_CLZ))
-JEMALLOC_INLINE size_t
+JEMALLOC_INLINE unsigned
 lg_floor(size_t x)
 {
 
@@ -253,11 +288,11 @@ lg_floor(size_t x)
 #elif (LG_SIZEOF_PTR == LG_SIZEOF_LONG)
 	return (((8 << LG_SIZEOF_PTR) - 1) - __builtin_clzl(x));
 #else
-#  error "Unsupported type sizes for lg_floor()"
+#  error "Unsupported type size for lg_floor()"
 #endif
 }
 #else
-JEMALLOC_INLINE size_t
+JEMALLOC_INLINE unsigned
 lg_floor(size_t x)
 {
 
@@ -268,20 +303,13 @@ lg_floor(size_t x)
 	x |= (x >> 4);
 	x |= (x >> 8);
 	x |= (x >> 16);
-#if (LG_SIZEOF_PTR == 3 && LG_SIZEOF_PTR == LG_SIZEOF_LONG)
+#if (LG_SIZEOF_PTR == 3)
 	x |= (x >> 32);
-	if (x == KZU(0xffffffffffffffff))
-		return (63);
-	x++;
-	return (jemalloc_ffsl(x) - 2);
-#elif (LG_SIZEOF_PTR == 2)
-	if (x == KZU(0xffffffff))
-		return (31);
-	x++;
-	return (jemalloc_ffs(x) - 2);
-#else
-#  error "Unsupported type sizes for lg_floor()"
 #endif
+	if (x == SIZE_T_MAX)
+		return ((8 << LG_SIZEOF_PTR) - 1);
+	x++;
+	return (ffs_zu(x) - 2);
 }
 #endif
 
diff --git a/deps/jemalloc/include/jemalloc/internal/valgrind.h b/deps/jemalloc/include/jemalloc/internal/valgrind.h
index a3380df92..877a142b6 100644
--- a/deps/jemalloc/include/jemalloc/internal/valgrind.h
+++ b/deps/jemalloc/include/jemalloc/internal/valgrind.h
@@ -30,17 +30,31 @@
  * calls must be embedded in macros rather than in functions so that when
  * Valgrind reports errors, there are no extra stack frames in the backtraces.
  */
-#define	JEMALLOC_VALGRIND_MALLOC(cond, ptr, usize, zero) do {		\
-	if (unlikely(in_valgrind && cond))				\
-		VALGRIND_MALLOCLIKE_BLOCK(ptr, usize, p2rz(ptr), zero);	\
+#define	JEMALLOC_VALGRIND_MALLOC(cond, tsdn, ptr, usize, zero) do {	\
+	if (unlikely(in_valgrind && cond)) {				\
+		VALGRIND_MALLOCLIKE_BLOCK(ptr, usize, p2rz(tsdn, ptr),	\
+		    zero);						\
+	}								\
 } while (0)
-#define	JEMALLOC_VALGRIND_REALLOC(maybe_moved, ptr, usize,		\
-    ptr_maybe_null, old_ptr, old_usize, old_rzsize, old_ptr_maybe_null,	\
-    zero) do {								\
+#define	JEMALLOC_VALGRIND_REALLOC_MOVED_no(ptr, old_ptr)		\
+    (false)
+#define	JEMALLOC_VALGRIND_REALLOC_MOVED_maybe(ptr, old_ptr)		\
+    ((ptr) != (old_ptr))
+#define	JEMALLOC_VALGRIND_REALLOC_PTR_NULL_no(ptr)			\
+    (false)
+#define	JEMALLOC_VALGRIND_REALLOC_PTR_NULL_maybe(ptr)			\
+    (ptr == NULL)
+#define	JEMALLOC_VALGRIND_REALLOC_OLD_PTR_NULL_no(old_ptr)		\
+    (false)
+#define	JEMALLOC_VALGRIND_REALLOC_OLD_PTR_NULL_maybe(old_ptr)		\
+    (old_ptr == NULL)
+#define	JEMALLOC_VALGRIND_REALLOC(moved, tsdn, ptr, usize, ptr_null,	\
+    old_ptr, old_usize, old_rzsize, old_ptr_null, zero) do {		\
 	if (unlikely(in_valgrind)) {					\
-		size_t rzsize = p2rz(ptr);				\
+		size_t rzsize = p2rz(tsdn, ptr);			\
 									\
-		if (!maybe_moved || ptr == old_ptr) {			\
+		if (!JEMALLOC_VALGRIND_REALLOC_MOVED_##moved(ptr,	\
+		    old_ptr)) {						\
 			VALGRIND_RESIZEINPLACE_BLOCK(ptr, old_usize,	\
 			    usize, rzsize);				\
 			if (zero && old_usize < usize) {		\
@@ -49,11 +63,13 @@
 				    old_usize), usize - old_usize);	\
 			}						\
 		} else {						\
-			if (!old_ptr_maybe_null || old_ptr != NULL) {	\
+			if (!JEMALLOC_VALGRIND_REALLOC_OLD_PTR_NULL_##	\
+			    old_ptr_null(old_ptr)) {			\
 				valgrind_freelike_block(old_ptr,	\
 				    old_rzsize);			\
 			}						\
-			if (!ptr_maybe_null || ptr != NULL) {		\
+			if (!JEMALLOC_VALGRIND_REALLOC_PTR_NULL_##	\
+			    ptr_null(ptr)) {				\
 				size_t copy_size = (old_usize < usize)	\
 				    ?  old_usize : usize;		\
 				size_t tail_size = usize - copy_size;	\
@@ -81,8 +97,8 @@
 #define	JEMALLOC_VALGRIND_MAKE_MEM_NOACCESS(ptr, usize) do {} while (0)
 #define	JEMALLOC_VALGRIND_MAKE_MEM_UNDEFINED(ptr, usize) do {} while (0)
 #define	JEMALLOC_VALGRIND_MAKE_MEM_DEFINED(ptr, usize) do {} while (0)
-#define	JEMALLOC_VALGRIND_MALLOC(cond, ptr, usize, zero) do {} while (0)
-#define	JEMALLOC_VALGRIND_REALLOC(maybe_moved, ptr, usize,		\
+#define	JEMALLOC_VALGRIND_MALLOC(cond, tsdn, ptr, usize, zero) do {} while (0)
+#define	JEMALLOC_VALGRIND_REALLOC(maybe_moved, tsdn, ptr, usize,	\
     ptr_maybe_null, old_ptr, old_usize, old_rzsize, old_ptr_maybe_null,	\
     zero) do {} while (0)
 #define	JEMALLOC_VALGRIND_FREE(ptr, rzsize) do {} while (0)
diff --git a/deps/jemalloc/include/jemalloc/internal/witness.h b/deps/jemalloc/include/jemalloc/internal/witness.h
new file mode 100644
index 000000000..cdf15d797
--- /dev/null
+++ b/deps/jemalloc/include/jemalloc/internal/witness.h
@@ -0,0 +1,266 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct witness_s witness_t;
+typedef unsigned witness_rank_t;
+typedef ql_head(witness_t) witness_list_t;
+typedef int witness_comp_t (const witness_t *, const witness_t *);
+
+/*
+ * Lock ranks.  Witnesses with rank WITNESS_RANK_OMIT are completely ignored by
+ * the witness machinery.
+ */
+#define	WITNESS_RANK_OMIT		0U
+
+#define	WITNESS_RANK_INIT		1U
+#define	WITNESS_RANK_CTL		1U
+#define	WITNESS_RANK_ARENAS		2U
+
+#define	WITNESS_RANK_PROF_DUMP		3U
+#define	WITNESS_RANK_PROF_BT2GCTX	4U
+#define	WITNESS_RANK_PROF_TDATAS	5U
+#define	WITNESS_RANK_PROF_TDATA		6U
+#define	WITNESS_RANK_PROF_GCTX		7U
+
+#define	WITNESS_RANK_ARENA		8U
+#define	WITNESS_RANK_ARENA_CHUNKS	9U
+#define	WITNESS_RANK_ARENA_NODE_CACHE	10
+
+#define	WITNESS_RANK_BASE		11U
+
+#define	WITNESS_RANK_LEAF		0xffffffffU
+#define	WITNESS_RANK_ARENA_BIN		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_ARENA_HUGE		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_DSS		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_ACTIVE	WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_DUMP_SEQ	WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_GDUMP		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_NEXT_THR_UID	WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_THREAD_ACTIVE_INIT	WITNESS_RANK_LEAF
+
+#define	WITNESS_INITIALIZER(rank) {"initializer", rank, NULL, {NULL, NULL}}
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct witness_s {
+	/* Name, used for printing lock order reversal messages. */
+	const char		*name;
+
+	/*
+	 * Witness rank, where 0 is lowest and UINT_MAX is highest.  Witnesses
+	 * must be acquired in order of increasing rank.
+	 */
+	witness_rank_t		rank;
+
+	/*
+	 * If two witnesses are of equal rank and they have the samp comp
+	 * function pointer, it is called as a last attempt to differentiate
+	 * between witnesses of equal rank.
+	 */
+	witness_comp_t		*comp;
+
+	/* Linkage for thread's currently owned locks. */
+	ql_elm(witness_t)	link;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+void	witness_init(witness_t *witness, const char *name, witness_rank_t rank,
+    witness_comp_t *comp);
+#ifdef JEMALLOC_JET
+typedef void (witness_lock_error_t)(const witness_list_t *, const witness_t *);
+extern witness_lock_error_t *witness_lock_error;
+#else
+void	witness_lock_error(const witness_list_t *witnesses,
+    const witness_t *witness);
+#endif
+#ifdef JEMALLOC_JET
+typedef void (witness_owner_error_t)(const witness_t *);
+extern witness_owner_error_t *witness_owner_error;
+#else
+void	witness_owner_error(const witness_t *witness);
+#endif
+#ifdef JEMALLOC_JET
+typedef void (witness_not_owner_error_t)(const witness_t *);
+extern witness_not_owner_error_t *witness_not_owner_error;
+#else
+void	witness_not_owner_error(const witness_t *witness);
+#endif
+#ifdef JEMALLOC_JET
+typedef void (witness_lockless_error_t)(const witness_list_t *);
+extern witness_lockless_error_t *witness_lockless_error;
+#else
+void	witness_lockless_error(const witness_list_t *witnesses);
+#endif
+
+void	witnesses_cleanup(tsd_t *tsd);
+void	witness_fork_cleanup(tsd_t *tsd);
+void	witness_prefork(tsd_t *tsd);
+void	witness_postfork_parent(tsd_t *tsd);
+void	witness_postfork_child(tsd_t *tsd);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+bool	witness_owner(tsd_t *tsd, const witness_t *witness);
+void	witness_assert_owner(tsdn_t *tsdn, const witness_t *witness);
+void	witness_assert_not_owner(tsdn_t *tsdn, const witness_t *witness);
+void	witness_assert_lockless(tsdn_t *tsdn);
+void	witness_lock(tsdn_t *tsdn, witness_t *witness);
+void	witness_unlock(tsdn_t *tsdn, witness_t *witness);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
+JEMALLOC_INLINE bool
+witness_owner(tsd_t *tsd, const witness_t *witness)
+{
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	witnesses = tsd_witnessesp_get(tsd);
+	ql_foreach(w, witnesses, link) {
+		if (w == witness)
+			return (true);
+	}
+
+	return (false);
+}
+
+JEMALLOC_INLINE void
+witness_assert_owner(tsdn_t *tsdn, const witness_t *witness)
+{
+	tsd_t *tsd;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	if (witness_owner(tsd, witness))
+		return;
+	witness_owner_error(witness);
+}
+
+JEMALLOC_INLINE void
+witness_assert_not_owner(tsdn_t *tsdn, const witness_t *witness)
+{
+	tsd_t *tsd;
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	witnesses = tsd_witnessesp_get(tsd);
+	ql_foreach(w, witnesses, link) {
+		if (w == witness)
+			witness_not_owner_error(witness);
+	}
+}
+
+JEMALLOC_INLINE void
+witness_assert_lockless(tsdn_t *tsdn)
+{
+	tsd_t *tsd;
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+
+	witnesses = tsd_witnessesp_get(tsd);
+	w = ql_last(witnesses, link);
+	if (w != NULL)
+		witness_lockless_error(witnesses);
+}
+
+JEMALLOC_INLINE void
+witness_lock(tsdn_t *tsdn, witness_t *witness)
+{
+	tsd_t *tsd;
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	witness_assert_not_owner(tsdn, witness);
+
+	witnesses = tsd_witnessesp_get(tsd);
+	w = ql_last(witnesses, link);
+	if (w == NULL) {
+		/* No other locks; do nothing. */
+	} else if (tsd_witness_fork_get(tsd) && w->rank <= witness->rank) {
+		/* Forking, and relaxed ranking satisfied. */
+	} else if (w->rank > witness->rank) {
+		/* Not forking, rank order reversal. */
+		witness_lock_error(witnesses, witness);
+	} else if (w->rank == witness->rank && (w->comp == NULL || w->comp !=
+	    witness->comp || w->comp(w, witness) > 0)) {
+		/*
+		 * Missing/incompatible comparison function, or comparison
+		 * function indicates rank order reversal.
+		 */
+		witness_lock_error(witnesses, witness);
+	}
+
+	ql_elm_new(witness, link);
+	ql_tail_insert(witnesses, witness, link);
+}
+
+JEMALLOC_INLINE void
+witness_unlock(tsdn_t *tsdn, witness_t *witness)
+{
+	tsd_t *tsd;
+	witness_list_t *witnesses;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	/*
+	 * Check whether owner before removal, rather than relying on
+	 * witness_assert_owner() to abort, so that unit tests can test this
+	 * function's failure mode without causing undefined behavior.
+	 */
+	if (witness_owner(tsd, witness)) {
+		witnesses = tsd_witnessesp_get(tsd);
+		ql_remove(witnesses, witness, link);
+	} else
+		witness_assert_owner(tsdn, witness);
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
diff --git a/deps/jemalloc/include/jemalloc/jemalloc_defs.h.in b/deps/jemalloc/include/jemalloc/jemalloc_defs.h.in
index ab13c3758..6d89435c2 100644
--- a/deps/jemalloc/include/jemalloc/jemalloc_defs.h.in
+++ b/deps/jemalloc/include/jemalloc/jemalloc_defs.h.in
@@ -33,5 +33,13 @@
  */
 #undef JEMALLOC_USE_CXX_THROW
 
+#ifdef _MSC_VER
+#  ifdef _WIN64
+#    define LG_SIZEOF_PTR_WIN 3
+#  else
+#    define LG_SIZEOF_PTR_WIN 2
+#  endif
+#endif
+
 /* sizeof(void *) == 2^LG_SIZEOF_PTR. */
 #undef LG_SIZEOF_PTR
diff --git a/deps/jemalloc/include/jemalloc/jemalloc_macros.h.in b/deps/jemalloc/include/jemalloc/jemalloc_macros.h.in
index a7028db34..2956c7bb4 100644
--- a/deps/jemalloc/include/jemalloc/jemalloc_macros.h.in
+++ b/deps/jemalloc/include/jemalloc/jemalloc_macros.h.in
@@ -11,12 +11,13 @@
 #define	JEMALLOC_VERSION_NREV @jemalloc_version_nrev@
 #define	JEMALLOC_VERSION_GID "@jemalloc_version_gid@"
 
-#  define MALLOCX_LG_ALIGN(la)	(la)
+#  define MALLOCX_LG_ALIGN(la)	((int)(la))
 #  if LG_SIZEOF_PTR == 2
-#    define MALLOCX_ALIGN(a)	(ffs(a)-1)
+#    define MALLOCX_ALIGN(a)	((int)(ffs((int)(a))-1))
 #  else
 #    define MALLOCX_ALIGN(a)						\
-	 ((a < (size_t)INT_MAX) ? ffs(a)-1 : ffs(a>>32)+31)
+       ((int)(((size_t)(a) < (size_t)INT_MAX) ? ffs((int)(a))-1 :	\
+       ffs((int)(((size_t)(a))>>32))+31))
 #  endif
 #  define MALLOCX_ZERO	((int)0x40)
 /*
@@ -28,7 +29,7 @@
 /*
  * Bias arena index bits so that 0 encodes "use an automatically chosen arena".
  */
-#  define MALLOCX_ARENA(a)	((int)(((a)+1) << 20))
+#  define MALLOCX_ARENA(a)	((((int)(a))+1) << 20)
 
 #if defined(__cplusplus) && defined(JEMALLOC_USE_CXX_THROW)
 #  define JEMALLOC_CXX_THROW throw()
@@ -36,32 +37,7 @@
 #  define JEMALLOC_CXX_THROW
 #endif
 
-#ifdef JEMALLOC_HAVE_ATTR
-#  define JEMALLOC_ATTR(s) __attribute__((s))
-#  define JEMALLOC_ALIGNED(s) JEMALLOC_ATTR(aligned(s))
-#  ifdef JEMALLOC_HAVE_ATTR_ALLOC_SIZE
-#    define JEMALLOC_ALLOC_SIZE(s) JEMALLOC_ATTR(alloc_size(s))
-#    define JEMALLOC_ALLOC_SIZE2(s1, s2) JEMALLOC_ATTR(alloc_size(s1, s2))
-#  else
-#    define JEMALLOC_ALLOC_SIZE(s)
-#    define JEMALLOC_ALLOC_SIZE2(s1, s2)
-#  endif
-#  ifndef JEMALLOC_EXPORT
-#    define JEMALLOC_EXPORT JEMALLOC_ATTR(visibility("default"))
-#  endif
-#  ifdef JEMALLOC_HAVE_ATTR_FORMAT_GNU_PRINTF
-#    define JEMALLOC_FORMAT_PRINTF(s, i) JEMALLOC_ATTR(format(gnu_printf, s, i))
-#  elif defined(JEMALLOC_HAVE_ATTR_FORMAT_PRINTF)
-#    define JEMALLOC_FORMAT_PRINTF(s, i) JEMALLOC_ATTR(format(printf, s, i))
-#  else
-#    define JEMALLOC_FORMAT_PRINTF(s, i)
-#  endif
-#  define JEMALLOC_NOINLINE JEMALLOC_ATTR(noinline)
-#  define JEMALLOC_NOTHROW JEMALLOC_ATTR(nothrow)
-#  define JEMALLOC_SECTION(s) JEMALLOC_ATTR(section(s))
-#  define JEMALLOC_RESTRICT_RETURN
-#  define JEMALLOC_ALLOCATOR
-#elif _MSC_VER
+#if _MSC_VER
 #  define JEMALLOC_ATTR(s)
 #  define JEMALLOC_ALIGNED(s) __declspec(align(s))
 #  define JEMALLOC_ALLOC_SIZE(s)
@@ -87,6 +63,31 @@
 #  else
 #    define JEMALLOC_ALLOCATOR
 #  endif
+#elif defined(JEMALLOC_HAVE_ATTR)
+#  define JEMALLOC_ATTR(s) __attribute__((s))
+#  define JEMALLOC_ALIGNED(s) JEMALLOC_ATTR(aligned(s))
+#  ifdef JEMALLOC_HAVE_ATTR_ALLOC_SIZE
+#    define JEMALLOC_ALLOC_SIZE(s) JEMALLOC_ATTR(alloc_size(s))
+#    define JEMALLOC_ALLOC_SIZE2(s1, s2) JEMALLOC_ATTR(alloc_size(s1, s2))
+#  else
+#    define JEMALLOC_ALLOC_SIZE(s)
+#    define JEMALLOC_ALLOC_SIZE2(s1, s2)
+#  endif
+#  ifndef JEMALLOC_EXPORT
+#    define JEMALLOC_EXPORT JEMALLOC_ATTR(visibility("default"))
+#  endif
+#  ifdef JEMALLOC_HAVE_ATTR_FORMAT_GNU_PRINTF
+#    define JEMALLOC_FORMAT_PRINTF(s, i) JEMALLOC_ATTR(format(gnu_printf, s, i))
+#  elif defined(JEMALLOC_HAVE_ATTR_FORMAT_PRINTF)
+#    define JEMALLOC_FORMAT_PRINTF(s, i) JEMALLOC_ATTR(format(printf, s, i))
+#  else
+#    define JEMALLOC_FORMAT_PRINTF(s, i)
+#  endif
+#  define JEMALLOC_NOINLINE JEMALLOC_ATTR(noinline)
+#  define JEMALLOC_NOTHROW JEMALLOC_ATTR(nothrow)
+#  define JEMALLOC_SECTION(s) JEMALLOC_ATTR(section(s))
+#  define JEMALLOC_RESTRICT_RETURN
+#  define JEMALLOC_ALLOCATOR
 #else
 #  define JEMALLOC_ATTR(s)
 #  define JEMALLOC_ALIGNED(s)
diff --git a/deps/jemalloc/include/msvc_compat/strings.h b/deps/jemalloc/include/msvc_compat/strings.h
index f01ffdd18..a3ee25063 100644
--- a/deps/jemalloc/include/msvc_compat/strings.h
+++ b/deps/jemalloc/include/msvc_compat/strings.h
@@ -21,7 +21,37 @@ static __forceinline int ffs(int x)
 	return (ffsl(x));
 }
 
+#  ifdef  _M_X64
+#    pragma intrinsic(_BitScanForward64)
+#  endif
+
+static __forceinline int ffsll(unsigned __int64 x)
+{
+	unsigned long i;
+#ifdef  _M_X64
+	if (_BitScanForward64(&i, x))
+		return (i + 1);
+	return (0);
+#else
+// Fallback for 32-bit build where 64-bit version not available
+// assuming little endian
+	union {
+		unsigned __int64 ll;
+		unsigned   long l[2];
+	} s;
+
+	s.ll = x;
+
+	if (_BitScanForward(&i, s.l[0]))
+		return (i + 1);
+	else if(_BitScanForward(&i, s.l[1]))
+		return (i + 33);
+	return (0);
+#endif
+}
+
 #else
+#  define ffsll(x) __builtin_ffsll(x)
 #  define ffsl(x) __builtin_ffsl(x)
 #  define ffs(x) __builtin_ffs(x)
 #endif
diff --git a/deps/jemalloc/include/msvc_compat/windows_extra.h b/deps/jemalloc/include/msvc_compat/windows_extra.h
index 0c5e323ff..3008faa37 100644
--- a/deps/jemalloc/include/msvc_compat/windows_extra.h
+++ b/deps/jemalloc/include/msvc_compat/windows_extra.h
@@ -1,26 +1,6 @@
 #ifndef MSVC_COMPAT_WINDOWS_EXTRA_H
 #define	MSVC_COMPAT_WINDOWS_EXTRA_H
 
-#ifndef ENOENT
-#  define ENOENT ERROR_PATH_NOT_FOUND
-#endif
-#ifndef EINVAL
-#  define EINVAL ERROR_BAD_ARGUMENTS
-#endif
-#ifndef EAGAIN
-#  define EAGAIN ERROR_OUTOFMEMORY
-#endif
-#ifndef EPERM
-#  define EPERM  ERROR_WRITE_FAULT
-#endif
-#ifndef EFAULT
-#  define EFAULT ERROR_INVALID_ADDRESS
-#endif
-#ifndef ENOMEM
-#  define ENOMEM ERROR_NOT_ENOUGH_MEMORY
-#endif
-#ifndef ERANGE
-#  define ERANGE ERROR_INVALID_DATA
-#endif
+#include <errno.h>
 
 #endif /* MSVC_COMPAT_WINDOWS_EXTRA_H */