Refactor atfork handlers

Current implementation (sysdeps/nptl/fork.c) replicates the atfork
handlers list backward to invoke the child handlers after fork/clone
syscall.

The internal atfork handlers is implemented as a single-linked list
so a lock-free algorithm can be used, trading fork mulithread call
performance for some code complexity and dynamic stack allocation
(since the backwards list should not fail).

This patch refactor it to use a dynarary instead of a linked list.
It simplifies the external variables need to be exported and also
the internal atfork handler member definition.

The downside is a serialization of fork call in multithread, since to
operate on the dynarray the internal lock should be used.  However
as noted by Florian, it already acquires external locks for malloc
and libio so it is already hitting some lock contention.  Besides,
posix_spawn should be faster and more scalable to run external programs
in multithread environments.

Checked on x86_64-linux-gnu.

	* nptl/Makefile (routines): Remove unregister-atfork.
	* nptl/register-atfork.c (fork_handler_pool): Remove variable.
	(fork_handler_alloc): Remove function.
	(fork_handlers, fork_handler_init): New variables.
	(__fork_lock): Rename to atfork_lock.
	(__register_atfork, __unregister_atfork, libc_freeres_fn): Rewrite
	to use a dynamic array to add/remove atfork handlers.
	* sysdeps/nptl/fork.c (__libc_fork): Likewise.
	* sysdeps/nptl/fork.h (__fork_lock, __fork_handlers, __linkin_atfork):
	Remove declaration.
	(fork_handler): Remove next, refcntr, and need_signal member.
	(__run_fork_handler_type): New enum.
	(__run_fork_handlers): New prototype.
	* sysdeps/nptl/libc-lockP.h (__libc_atfork): Remove declaration.

sysdeps/nptl/fork.h

@@ -24,29 +24,37 @@ extern unsigned long int __fork_generation attribute_hidden;
 /* Pointer to the fork generation counter in the thread library.  */
 extern unsigned long int *__fork_generation_pointer attribute_hidden;
 
-/* Lock to protect allocation and deallocation of fork handlers.  */
-extern int __fork_lock attribute_hidden;
-
 /* Elements of the fork handler lists.  */
 struct fork_handler
 {
-  struct fork_handler *next;
   void (*prepare_handler) (void);
   void (*parent_handler) (void);
   void (*child_handler) (void);
   void *dso_handle;
-  unsigned int refcntr;
-  int need_signal;
 };
 
-/* The single linked list of all currently registered for handlers.  */
-extern struct fork_handler *__fork_handlers attribute_hidden;
-
-
 /* Function to call to unregister fork handlers.  */
 extern void __unregister_atfork (void *dso_handle) attribute_hidden;
 #define UNREGISTER_ATFORK(dso_handle) __unregister_atfork (dso_handle)
 
+enum __run_fork_handler_type
+{
+  atfork_run_prepare,
+  atfork_run_child,
+  atfork_run_parent
+};
+
+/* Run the atfork handlers and lock/unlock the internal lock depending
+   of the WHO argument:
+
+   - atfork_run_prepare: run all the PREPARE_HANDLER in reverse order of
+			 insertion and locks the internal lock.
+   - atfork_run_child: run all the CHILD_HANDLER and unlocks the internal
+		       lock.
+   - atfork_run_parent: run all the PARENT_HANDLER and unlocks the internal
+			lock.  */
+extern void __run_fork_handlers (enum __run_fork_handler_type who)
+  attribute_hidden;
 
 /* C library side function to register new fork handlers.  */
 extern int __register_atfork (void (*__prepare) (void),
@@ -54,6 +62,3 @@ extern int __register_atfork (void (*__prepare) (void),
 			      void (*__child) (void),
 			      void *dso_handle);
 libc_hidden_proto (__register_atfork)
-
-/* Add a new element to the fork list.  */
-extern void __linkin_atfork (struct fork_handler *newp) attribute_hidden;