Fix huge_palloc() regression.

Split arena_choose() into arena_[i]choose() and use arena_ichoose() for
arena lookup during internal allocation.  This fixes huge_palloc() so
that it always succeeds during extent node allocation.

This regression was introduced by
66cd953514 (Do not allocate metadata via
non-auto arenas, nor tcaches.).

include/jemalloc/internal/jemalloc_internal.h.in

@@ -550,7 +550,9 @@ 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(tsd_t *tsd, arena_t *arena, bool internal);
+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_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(tsd_t *tsd, unsigned ind, bool init_if_missing);
@@ -788,7 +790,7 @@ 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, bool internal)
+arena_choose_impl(tsd_t *tsd, arena_t *arena, bool internal)
 {
 	arena_t *ret;
 
@@ -802,6 +804,24 @@ arena_choose(tsd_t *tsd, arena_t *arena, bool internal)
 	return (ret);
 }
 
+JEMALLOC_INLINE arena_t *
+arena_choose(tsd_t *tsd, arena_t *arena)
+{
+
+	return (arena_choose_impl(tsd, arena, false));
+}
+
+JEMALLOC_INLINE arena_t *
+arena_ichoose(tsd_t *tsd, arena_t *arena)
+{
+
+	assert(tsd != NULL || arena != NULL);
+
+	if (tsd != NULL)
+		return (arena_choose_impl(tsd, NULL, true));
+	return (arena);
+}
+
 JEMALLOC_INLINE arena_tdata_t *
 arena_tdata_get(tsd_t *tsd, unsigned ind, bool refresh_if_missing)
 {

include/jemalloc/internal/private_symbols.txt

@@ -10,6 +10,7 @@ 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
@@ -35,6 +36,7 @@ arena_decay_time_set
 arena_dss_prec_get
 arena_dss_prec_set
 arena_get
+arena_ichoose
 arena_init
 arena_lg_dirty_mult_default_get
 arena_lg_dirty_mult_default_set

include/jemalloc/internal/tcache.h

@@ -293,7 +293,7 @@ tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 	assert(tcache_success == (ret != NULL));
 	if (unlikely(!tcache_success)) {
 		bool tcache_hard_success;
-		arena = arena_choose(tsd, arena, false);
+		arena = arena_choose(tsd, arena);
 		if (unlikely(arena == NULL))
 			return (NULL);
 
@@ -354,7 +354,7 @@ tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 		 * Only allocate one large object at a time, because it's quite
 		 * expensive to create one and not use it.
 		 */
-		arena = arena_choose(tsd, arena, false);
+		arena = arena_choose(tsd, arena);
 		if (unlikely(arena == NULL))
 			return (NULL);
 
@@ -460,8 +460,7 @@ 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,
-		    false));
+		elm->tcache = tcache_create(tsd, arena_choose(tsd, NULL));
 	}
 	return (elm->tcache);
 }