Fix missing GC_dirty calls for GC-allocated objects used internally

Fix missing GC_dirty calls for GC-allocated objects used internally
(a cherry-pick of commit c664ef4f from 'release-7_6')

This change matters only in case of MANUAL_VDB mode.

Also, GC_reachable_here calls are inserted after GC_dirty.

Also, this commit actually disables multiple objects allocation in
GC_generic_malloc_many if MANUAL_VDB and the incremental mode are on.

* finalize.c (GC_grow_table, GC_register_disappearing_link,
GC_unregister_disappearing_link_inner,
GC_move_disappearing_link_inner,
GC_register_finalizer_inner, ITERATE_DL_HASHTBL_END,
DELETE_DL_HASHTBL_ENTRY, GC_finalize, GC_enqueue_all_finalizers): Call
GC_dirty where needed.
* gcj_mlc.c [GC_GCJ_SUPPORT] (GC_gcj_malloc, GC_debug_gcj_malloc,
GC_gcj_malloc_ignore_off_page): Likewise.
* gcj_mlc.c (GC_gcj_malloc, GC_debug_gcj_malloc,
GC_gcj_malloc_ignore_off_page): Call
REACHABLE_AFTER_DIRTY(ptr_to_struct_containing_descr) after GC_dirty(op).
* mallocx.c [MANUAL_VDB] (GC_generic_malloc_many): Always fall back to
GC_generic_malloc (unless !GC_incremental).
* mallocx.c [MANUAL_VDB] (GC_generic_malloc_many): If
GC_is_heap_ptr(result) then call GC_dirty(result) and
REACHABLE_AFTER_DIRTY(op) after storing op pointer.
* pthread_start.c [GC_PTHREADS && !GC_WIN32_THREADS]
(GC_inner_start_routine): Likewise.
* pthread_support.c [GC_PTHREADS && !GC_WIN32_THREADS] (GC_new_thread,
GC_delete_thread, GC_delete_gc_thread): Likewise.
* specific.c [USE_CUSTOM_SPECIFIC] (GC_setspecific,
GC_remove_specific_after_fork): Likewise.
* typd_mlc.c (GC_make_sequence_descriptor, GC_malloc_explicitly_typed,
GC_malloc_explicitly_typed_ignore_off_page, GC_calloc_explicitly_typed):
Likewise.
* win32_threads.c (GC_new_thread, GC_delete_gc_thread_no_free,
GC_delete_thread, GC_CreateThread): Likewise.
* win32_threads.c [!CYGWIN32 && !MSWINCE && !MSWIN_XBOX1]
(GC_beginthreadex): Likewise.
* win32_threads.c [GC_PTHREADS] (GC_pthread_create,
GC_pthread_start_inner): Likewise.
* typd_mlc.c (GC_make_sequence_descriptor): Call REACHABLE_AFTER_DIRTY
for the stored pointers after GC_dirty(result).
* typd_mlc.c (GC_malloc_explicitly_typed,
GC_malloc_explicitly_typed_ignore_off_page, GC_calloc_explicitly_typed):
Call REACHABLE_AFTER_DIRTY(d) after GC_dirty(op).
* win32_threads.c (GC_CreateThread, GC_beginthreadex,
GC_pthread_create): Call REACHABLE_AFTER_DIRTY for the stored pointer
after GC_dirty.
* include/gc_inline.h (GC_FAST_MALLOC_GRANS): Call
GC_end_stubborn_change(my_fl) after GC_FAST_M_AO_STORE() call unless
kind is GC_I_PTRFREE.
* include/gc_inline.h (GC_FAST_MALLOC_GRANS): Call
GC_reachable_here(next) after GC_end_stubborn_change(my_fl).
* include/gc_inline.h (GC_CONS): Call GC_end_stubborn_change(result).
* include/gc_inline.h (GC_CONS): Call GC_reachable_here for the stored
pointers after GC_end_stubborn_change call.
* include/gc_inline.h (GC_CONS): Declare l and r local variables;
compute first and second expression even in case of
GC_MALLOC_WORDS_KIND failure; pass l and r to GC_reachable_here (instead
of first and second).
* include/private/gc_priv.h (REACHABLE_AFTER_DIRTY): New macro.