[turbofan] Don't assume we have receiver maps in preprocessed feedback

This was an oversight in my previous CL.

Bug: chromium:936077, v8:7790
Change-Id: Ic1034c1754d10c72df8f61d1e2c34333e1565e3e
Reviewed-on: https://chromium-review.googlesource.com/c/1491222Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#59923}

src/compiler/access-info.cc

@@ -282,6 +282,8 @@ bool AccessInfoFactory::ComputeElementAccessInfos(
     }
   }
 
+  if (processed.receiver_maps.empty()) return false;
+
   if (access_mode == AccessMode::kLoad) {
     // For polymorphic loads of similar elements kinds (i.e. all tagged or all
     // double), always use the "worst case" code without a transition.  This is

test/mjsunit/regress/regress-936077.js

+// Copyright 2019 the V8 project authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+// Flags: --expose-gc --allow-natives-syntax
+// Flags: --concurrent-inlining --function-context-specialization
+
+function main() {
+  var obj = {};
+  function foo() { return obj[0]; };
+  gc();
+  obj.x = 10;
+  %OptimizeFunctionOnNextCall(foo);
+  foo();
+}
+main();
+main();