Document the results of painful reverse-engineering of the "portable TLS"

code.

PyThread_set_key_value():  It's clear that this code assumes the passed-in
value isn't NULL, so document that it must not be, and assert that it
isn't.  It remains unclear whether existing callers want the odd semantics
actually implemented by this function.

Python/thread.c

@@ -142,26 +142,80 @@ void PyThread_init_thread(void)
    This code stolen from "thread_sgi.h", where it was the only
    implementation of an existing Python TLS API.
 */
-/*
- * Per-thread data ("key") support.
- */
+/* ------------------------------------------------------------------------
+Per-thread data ("key") support.
+
+Use PyThread_create_key() to create a new key.  This is typically shared
+across threads.
+
+Use PyThread_set_key_value(thekey, value) to associate void* value with
+thekey in the current thread.  Each thread has a distinct mapping of thekey
+to a void* value.  Caution:  if the current thread already has a mapping
+for thekey, value is ignored.
+
+Use PyThread_get_key_value(thekey) to retrieve the void* value associated
+with thekey in the current thread.  This returns NULL if no value is
+associated with thekey in the current thread.
+
+Use PyThread_delete_key_value(thekey) to forget the current thread's associated
+value for thekey.  PyThread_delete_key(thekey) forgets the values associated
+with thekey across *all* threads.
 
+While some of these functions have error-return values, none set any
+Python exception.
+
+None of the functions does memory management on behalf of the void* values.
+You need to allocate and deallocate them yourself.  If the void* values
+happen to be PyObject*, these functions don't do refcount operations on
+them either.
+
+The GIL does not need to be held when calling these functions; they supply
+their own locking.  This isn't true of PyThread_create_key(), though (see
+next paragraph).
+
+There's a hidden assumption that PyThread_create_key() will be called before
+any of the other functions are called.  There's also a hidden assumption
+that calls to PyThread_create_key() are serialized externally.
+------------------------------------------------------------------------ */
+
+/* A singly-linked list of struct key objects remembers all the key->value
+ * associations.  File static keyhead heads the list.  keymutex is used
+ * to enforce exclusion internally.
+ */
 struct key {
+	/* Next record in the list, or NULL if this is the last record. */
 	struct key *next;
+
+	/* The thread id, according to PyThread_get_thread_ident(). */
 	long id;
+
+	/* The key and its associated value. */
 	int key;
 	void *value;
 };
 
 static struct key *keyhead = NULL;
-static int nkeys = 0;
 static PyThread_type_lock keymutex = NULL;
-
+static int nkeys = 0;  /* PyThread_create_key() hands out nkeys+1 next */
+
+/* Internal helper.
+ * If the current thread has a mapping for key, the appropriate struct key*
+ * is returned.  NB:  value is ignored in this case!
+ * If there is no mapping for key in the current thread, then:
+ *     If value is NULL, NULL is returned.
+ *     Else a mapping of key to value is created for the current thread,
+ *     and a pointer to a new struct key* is returned; except that if
+ *     malloc() can't find room for a new struct key*, NULL is returned.
+ * So when value==NULL, this acts like a pure lookup routine, and when
+ * value!=NULL, this acts like dict.setdefault(), returning an existing
+ * mapping if one exists, else creating a new mapping.
+ */
 static struct key *
 find_key(int key, void *value)
 {
 	struct key *p;
 	long id = PyThread_get_thread_ident();
+
 	for (p = keyhead; p != NULL; p = p->next) {
 		if (p->id == id && p->key == key)
 			return p;
@@ -181,18 +235,27 @@ find_key(int key, void *value)
 	return p;
 }
 
+/* Return a new key.  This must be called before any other functions in
+ * this family, and callers must arrange to serialize calls to this
+ * function.  No violations are detected.
+ */
 int
 PyThread_create_key(void)
 {
+	/* All parts of this function are wrong if it's called by multiple
+	 * threads simultaneously.
+	 */
 	if (keymutex == NULL)
 		keymutex = PyThread_allocate_lock();
 	return ++nkeys;
 }
 
+/* Forget the associations for key across *all* threads. */
 void
 PyThread_delete_key(int key)
 {
 	struct key *p, **q;
+
 	PyThread_acquire_lock(keymutex, 1);
 	q = &keyhead;
 	while ((p = *q) != NULL) {
@@ -207,31 +270,46 @@ PyThread_delete_key(int key)
 	PyThread_release_lock(keymutex);
 }
 
+/* Confusing:  If the current thread has an association for key,
+ * value is ignored, and 0 is returned.  Else an attempt is made to create
+ * an association of key to value for the current thread.  0 is returned
+ * if that succeeds, but -1 is returned if there's not enough memory
+ * to create the association.  value must not be NULL.
+ */
 int
 PyThread_set_key_value(int key, void *value)
 {
-	struct key *p = find_key(key, value);
+	struct key *p;
+
+	assert(value != NULL);
+	p = find_key(key, value);
 	if (p == NULL)
 		return -1;
 	else
 		return 0;
 }
 
+/* Retrieve the value associated with key in the current thread, or NULL
+ * if the current thread doesn't have an association for key.
+ */
 void *
 PyThread_get_key_value(int key)
 {
 	struct key *p = find_key(key, NULL);
+
 	if (p == NULL)
 		return NULL;
 	else
 		return p->value;
 }
 
+/* Forget the current thread's association for key, if any. */
 void
 PyThread_delete_key_value(int key)
 {
 	long id = PyThread_get_thread_ident();
 	struct key *p, **q;
+
 	PyThread_acquire_lock(keymutex, 1);
 	q = &keyhead;
 	while ((p = *q) != NULL) {