v8/test/mjsunit/regress/regress-786784.js
Mythri A 97ae101029 Hold on to FeedbackMetadata when allocating feedback vectors
Allocating a new feedback vector happens in two steps: We create an
empty structure and then initialize the array based on the
FeedbackMetadata.When allocating a new feedback array we could trigger
a GC which might flush the bytecode and associated feedback metadata.
This shouldn't happen in normal cases, because we either allocate
feedback vector after compilation or when we reach the expected budget.
In both cases, the age of the feedback vector should be 0 and hence
bytecode shouldn't be flushed. However, with debugger enabled we may
allocate feedback vectors even when the bytecode array is old
for example: when we enable precise invocation counters. This also
causes issues in tests with --stress-flush-bytecode. In the stress mode
we flush bytecode without considering the age. Holding on to the
feedback metadata prevents crashes in such cases.

Bug: v8:10560
Change-Id: Ie806ff4102cb5fcf257c8683d5ca957853e38c05
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2218066
Commit-Queue: Mythri Alle <mythria@chromium.org>
Auto-Submit: Mythri Alle <mythria@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68052}
2020-05-28 16:59:12 +00:00

35 lines
1.1 KiB
JavaScript

// Copyright 2017 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: --allow-natives-syntax
function f() {
function g(arg) { return arg; }
// The closure contains a call IC slot.
return function() { return g(42); };
}
const a = Realm.create();
const b = Realm.create();
// Create two closures in different contexts sharing the same
// SharedFunctionInfo (shared due to code caching).
const x = Realm.eval(a, f.toString() + " f()");
const y = Realm.eval(b, f.toString() + " f()");
// Run the first closure to create SFI::code.
x();
// At this point, SFI::code is set and `x` has a feedback vector (`y` does not).
// Enabling block code coverage deoptimizes all functions and triggers the
// buggy code path in which we'd unconditionally replace JSFunction::code with
// its SFI::code (but skip feedback vector setup).
%DebugToggleBlockCoverage(true);
// Still no feedback vector set on `y` but it now contains code. Run it to
// trigger the crash when attempting to write into the non-existent feedback
// vector.
y();