Replace PAI vtable dispatch with direct calls

The pai_t interface implements C-style polymorphism via function pointers
to abstract over PAC and HPA. This abstraction provides no real benefit:
only two implementations exist, the dispatcher already knows which one to
use, and HPA stubs 2 of 5 operations. Remove the runtime dispatch in
favor of direct calls.

This commit:
- Promotes pac_alloc/expand/shrink/dalloc/time_until_deferred_work to
  external linkage and replaces the pai_t *self parameter with pac_t *pac.
- Promotes hpa_alloc/expand/shrink/dalloc/time_until_deferred_work to
  external linkage and replaces pai_t *self with hpa_shard_t *shard.
- Updates hpa_dalloc_batch's signature to take hpa_shard_t * directly
  and removes the hpa_from_pai container-of helper. Updates internal
  callers in hpa_alloc, hpa_dalloc, and hpa_sec_flush_impl.
- Drops the vtable assignments from pac_init() and hpa_shard_init().
- Replaces pai_alloc/dalloc/etc. dispatch in pa.c with direct calls.
  HPA expand and shrink (which are unconditional failure stubs) are
  skipped entirely for HPA-owned extents.
- Removes the pa_get_pai() helper.
- Updates tests in test/unit/hpa.c and test/unit/hpa_sec_integration.c
  to call hpa_alloc/dalloc/etc. directly.

The pai_t struct field stays as dead weight in pac_t and hpa_shard_t;
it is removed in the next commit along with pai.h itself.

No behavioral changes.

include/jemalloc/internal/hpa.h

@@ -154,6 +154,18 @@ bool hpa_shard_init(tsdn_t *tsdn, hpa_shard_t *shard, hpa_central_t *central,
     emap_t *emap, base_t *base, edata_cache_t *edata_cache, unsigned ind,
     const hpa_shard_opts_t *opts, const sec_opts_t *sec_opts);
 
+edata_t *hpa_alloc(tsdn_t *tsdn, hpa_shard_t *shard, size_t size,
+    size_t alignment, bool zero, bool guarded, bool frequent_reuse,
+    bool *deferred_work_generated);
+bool hpa_expand(tsdn_t *tsdn, hpa_shard_t *shard, edata_t *edata,
+    size_t old_size, size_t new_size, bool zero,
+    bool *deferred_work_generated);
+bool hpa_shrink(tsdn_t *tsdn, hpa_shard_t *shard, edata_t *edata,
+    size_t old_size, size_t new_size, bool *deferred_work_generated);
+void hpa_dalloc(tsdn_t *tsdn, hpa_shard_t *shard, edata_t *edata,
+    bool *deferred_work_generated);
+uint64_t hpa_time_until_deferred_work(tsdn_t *tsdn, hpa_shard_t *shard);
+
 void hpa_shard_stats_accum(hpa_shard_stats_t *dst, hpa_shard_stats_t *src);
 void hpa_shard_stats_merge(
     tsdn_t *tsdn, hpa_shard_t *shard, hpa_shard_stats_t *dst);

include/jemalloc/internal/pa.h

@@ -96,7 +96,7 @@ struct pa_shard_s {
 	/* Allocates from a PAC. */
 	pac_t pac;
 
-	hpa_shard_t hpa_shard;
+	hpa_shard_t hpa;
 
 	/* The source of edata_t objects. */
 	edata_cache_t edata_cache;

include/jemalloc/internal/pac.h

@@ -165,6 +165,17 @@ bool pac_init(tsdn_t *tsdn, pac_t *pac, base_t *base, emap_t *emap,
     ssize_t dirty_decay_ms, ssize_t muzzy_decay_ms, pac_stats_t *pac_stats,
     malloc_mutex_t *stats_mtx);
 
+edata_t *pac_alloc(tsdn_t *tsdn, pac_t *pac, size_t size, size_t alignment,
+    bool zero, bool guarded, bool frequent_reuse,
+    bool *deferred_work_generated);
+bool pac_expand(tsdn_t *tsdn, pac_t *pac, edata_t *edata, size_t old_size,
+    size_t new_size, bool zero, bool *deferred_work_generated);
+bool pac_shrink(tsdn_t *tsdn, pac_t *pac, edata_t *edata, size_t old_size,
+    size_t new_size, bool *deferred_work_generated);
+void pac_dalloc(tsdn_t *tsdn, pac_t *pac, edata_t *edata,
+    bool *deferred_work_generated);
+uint64_t pac_time_until_deferred_work(tsdn_t *tsdn, pac_t *pac);
+
 static inline size_t
 pac_mapped(const pac_t *pac) {
 	return atomic_load_zu(&pac->stats->pac_mapped, ATOMIC_RELAXED);