Title: [285314] branches/safari-612-branch/Source/_javascript_Core
- Revision
- 285314
- Author
- [email protected]
- Date
- 2021-11-04 14:42:57 -0700 (Thu, 04 Nov 2021)
Log Message
Revert "Cherry-pick r284036. rdar://problem/84167629"
This reverts commit r285280.
Modified Paths
Diff
Modified: branches/safari-612-branch/Source/_javascript_Core/ChangeLog (285313 => 285314)
--- branches/safari-612-branch/Source/_javascript_Core/ChangeLog 2021-11-04 21:39:52 UTC (rev 285313)
+++ branches/safari-612-branch/Source/_javascript_Core/ChangeLog 2021-11-04 21:42:57 UTC (rev 285314)
@@ -416,51 +416,6 @@
(JSC::Wasm::JSC_DEFINE_JIT_OPERATION):
* wasm/WasmOperations.h:
-2021-11-04 Russell Epstein <[email protected]>
-
- Cherry-pick r284036. rdar://problem/84167629
-
- Fix spec-correctness when inlining __proto__ intrinsic using get_by_id_with_this
- https://bugs.webkit.org/show_bug.cgi?id=231559
-
- Reviewed by Yusuke Suzuki.
-
- My original fix in r283512 has some theoretical spec correctness issues.
- I'm not sure if they can be materialized or not since we only use
- get_by_id_with_this in very limited scenarios. However, this patch just
- makes it so we call the getter instead of attempting to inline it
- when using get_by_id_with_this.
-
- * jit/AssemblyHelpers.cpp:
- (JSC::AssemblyHelpers::emitLoadPrototypeWithoutCheck): Deleted.
- * jit/AssemblyHelpers.h:
- * jit/IntrinsicEmitter.cpp:
- (JSC::IntrinsicGetterAccessCase::canEmitIntrinsicGetter):
- (JSC::IntrinsicGetterAccessCase::emitIntrinsicGetter):
-
-
- git-svn-id: https://svn.webkit.org/repository/webkit/trunk@284036 268f45cc-cd09-0410-ab3c-d52691b4dbfc
-
- 2021-10-12 Saam Barati <[email protected]>
-
- Fix spec-correctness when inlining __proto__ intrinsic using get_by_id_with_this
- https://bugs.webkit.org/show_bug.cgi?id=231559
-
- Reviewed by Yusuke Suzuki.
-
- My original fix in r283512 has some theoretical spec correctness issues.
- I'm not sure if they can be materialized or not since we only use
- get_by_id_with_this in very limited scenarios. However, this patch just
- makes it so we call the getter instead of attempting to inline it
- when using get_by_id_with_this.
-
- * jit/AssemblyHelpers.cpp:
- (JSC::AssemblyHelpers::emitLoadPrototypeWithoutCheck): Deleted.
- * jit/AssemblyHelpers.h:
- * jit/IntrinsicEmitter.cpp:
- (JSC::IntrinsicGetterAccessCase::canEmitIntrinsicGetter):
- (JSC::IntrinsicGetterAccessCase::emitIntrinsicGetter):
-
2021-10-25 Null <[email protected]>
Cherry-pick r283600. rdar://problem/84625586
Modified: branches/safari-612-branch/Source/_javascript_Core/jit/IntrinsicEmitter.cpp (285313 => 285314)
--- branches/safari-612-branch/Source/_javascript_Core/jit/IntrinsicEmitter.cpp 2021-11-04 21:39:52 UTC (rev 285313)
+++ branches/safari-612-branch/Source/_javascript_Core/jit/IntrinsicEmitter.cpp 2021-11-04 21:42:57 UTC (rev 285314)
@@ -48,13 +48,6 @@
bool IntrinsicGetterAccessCase::canEmitIntrinsicGetter(JSFunction* getter, Structure* structure)
{
- // We aren't structure checking the this value, so we don't know:
- // - For type array loads, that it's a typed array.
- // - For __proto__ getter, that the incoming value is an object,
- // and if it overrides getPrototype structure flags.
- // So for these cases, it's simpler to just call the getter directly.
- if (stubInfo.thisValueIsInThisGPR())
- return false;
switch (getter->intrinsic()) {
case TypedArrayByteOffsetIntrinsic:
_______________________________________________
webkit-changes mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-changes
