Transition complex objects to "too complex" shape

When an object becomes "too complex" (in other words it has too many
variations in the shape tree), we transition it to use a "too complex"
shape and use a hash for storing instance variables.

Without this patch, there were rare cases where shape tree growth could
"explode" and cause performance degradation on what would otherwise have
been cached fast paths.

This patch puts a limit on shape tree growth, and gracefully degrades in
the rare case where there could be a factorial growth in the shape tree.

For example:

```ruby
class NG; end

HUGE_NUMBER.times do
  NG.new.instance_variable_set(:"@unique_ivar_#{_1}", 1)
end
```

We consider objects to be "too complex" when the object's class has more
than SHAPE_MAX_VARIATIONS (currently 8) leaf nodes in the shape tree and
the object introduces a new variation (a new leaf node) associated with
that class.

For example, new variations on instances of the following class would be
considered "too complex" because those instances create more than 8
leaves in the shape tree:

```ruby
class Foo; end
9.times { Foo.new.instance_variable_set(":@uniq_#{_1}", 1) }
```

However, the following class is *not* too complex because it only has
one leaf in the shape tree:

```ruby
class Foo
  def initialize
    @a = @b = @c = @d = @e = @f = @g = @h = @i = nil
  end
end
9.times { Foo.new }
``

This case is rare, so we don't expect this change to impact performance
of most applications, but it needs to be handled.

Co-Authored-By: Aaron Patterson <tenderlove@ruby-lang.org>

1 files changed, 29 insertions, 14 deletions

diff --git a/vm_insnhelper.c b/vm_insnhelper.c
index 84fe32c1b7..68b8d89abb 100644
--- a/vm_insnhelper.c
+++ b/vm_insnhelper.c
@@ -1213,6 +1213,8 @@ vm_getivar(VALUE obj, ID id, const rb_iseq_t *iseq, IVC ic, const struct rb_call
     }
 
     if (LIKELY(cached_id == shape_id)) {
+        RUBY_ASSERT(cached_id != OBJ_TOO_COMPLEX_SHAPE_ID);
+
         if (index == ATTR_INDEX_NOT_SET) {
             return Qnil;
         }
@@ -1242,24 +1244,31 @@ vm_getivar(VALUE obj, ID id, const rb_iseq_t *iseq, IVC ic, const struct rb_call
 
         rb_shape_t *shape = rb_shape_get_shape_by_id(shape_id);
 
-        if (rb_shape_get_iv_index(shape, id, &index)) {
-            // This fills in the cache with the shared cache object.
-            // "ent" is the shared cache object
-            fill_ivar_cache(iseq, ic, cc, is_attr, index, shape_id);
-
-            // We fetched the ivar list above
-            val = ivar_list[index];
-            RUBY_ASSERT(!UNDEF_P(val));
+        if (shape_id == OBJ_TOO_COMPLEX_SHAPE_ID) {
+            if (!rb_id_table_lookup(ROBJECT_IV_HASH(obj), id, &val)) {
+                val = Qnil;
+            }
         }
         else {
-            if (is_attr) {
-                vm_cc_attr_index_initialize(cc, shape_id);
+            if (rb_shape_get_iv_index(shape, id, &index)) {
+                // This fills in the cache with the shared cache object.
+                // "ent" is the shared cache object
+                fill_ivar_cache(iseq, ic, cc, is_attr, index, shape_id);
+
+                // We fetched the ivar list above
+                val = ivar_list[index];
+                RUBY_ASSERT(!UNDEF_P(val));
             }
             else {
-                vm_ic_attr_index_initialize(ic, shape_id);
-            }
+                if (is_attr) {
+                    vm_cc_attr_index_initialize(cc, shape_id);
+                }
+                else {
+                    vm_ic_attr_index_initialize(ic, shape_id);
+                }
 
-            val = Qnil;
+                val = Qnil;
+            }
         }
 
     }
@@ -1283,6 +1292,8 @@ general_path:
 static void
 populate_cache(attr_index_t index, shape_id_t next_shape_id, ID id, const rb_iseq_t *iseq, IVC ic, const struct rb_callcache *cc, bool is_attr)
 {
+    RUBY_ASSERT(next_shape_id != OBJ_TOO_COMPLEX_SHAPE_ID);
+
     // Cache population code
     if (is_attr) {
         vm_cc_attr_index_set(cc, index, next_shape_id);
@@ -1309,7 +1320,9 @@ vm_setivar_slowpath(VALUE obj, ID id, VALUE val, const rb_iseq_t *iseq, IVC ic,
 
             shape_id_t next_shape_id = ROBJECT_SHAPE_ID(obj);
 
-            populate_cache(index, next_shape_id, id, iseq, ic, cc, is_attr);
+            if (next_shape_id != OBJ_TOO_COMPLEX_SHAPE_ID) {
+                populate_cache(index, next_shape_id, id, iseq, ic, cc, is_attr);
+            }
 
             RB_DEBUG_COUNTER_INC(ivar_set_ic_miss_iv_hit);
             return val;
@@ -1413,6 +1426,7 @@ vm_setivar(VALUE obj, ID id, VALUE val, shape_id_t dest_shape_id, attr_index_t i
             VM_ASSERT(!rb_ractor_shareable_p(obj) || rb_obj_frozen_p(obj));
 
             shape_id_t shape_id = ROBJECT_SHAPE_ID(obj);
+            RUBY_ASSERT(dest_shape_id != OBJ_TOO_COMPLEX_SHAPE_ID);
 
             if (LIKELY(shape_id == dest_shape_id)) {
                 RUBY_ASSERT(dest_shape_id != INVALID_SHAPE_ID && shape_id != INVALID_SHAPE_ID);
@@ -1440,6 +1454,7 @@ vm_setivar(VALUE obj, ID id, VALUE val, shape_id_t dest_shape_id, attr_index_t i
 
             VALUE *ptr = ROBJECT_IVPTR(obj);
 
+            RUBY_ASSERT(!rb_shape_obj_too_complex(obj));
             RB_OBJ_WRITE(obj, &ptr[index], val);
 
             RB_DEBUG_COUNTER_INC(ivar_set_ic_hit);