Title: [294429] trunk/Source/_javascript_Core/assembler/SecureARM64EHashPinsInlines.h
- Revision
- 294429
- Author
- [email protected]
- Date
- 2022-05-18 13:43:25 -0700 (Wed, 18 May 2022)
Log Message
Small clean ups in SecureARM64EHashPins.
https://bugs.webkit.org/show_bug.cgi?id=240587
Reviewed by Yusuke Suzuki.
1. In SecureARM64EHashPins::keyForCurrentThread(), there's no need to add 1 to
the value retrieve from TPIDRRO_EL0. TPIDRRO_EL0 is a pointer to valid memory
and can never be 0. Also, we already code in multiple other places where we
expect to load from it without a null check.
2. SecureARM64EHashPins::forEachPage() should use a do-while loop instead of a
for loop. This is because we are guaranteed to at least have 1 page at
start up if the JIT is enabled. If the JIT is not enabled, this code will
never be called.
* Source/_javascript_Core/assembler/SecureARM64EHashPinsInlines.h:
(JSC::SecureARM64EHashPins::keyForCurrentThread):
(JSC::SecureARM64EHashPins::forEachPage):
Canonical link: https://commits.webkit.org/250711@main
Modified Paths
Diff
Modified: trunk/Source/_javascript_Core/assembler/SecureARM64EHashPinsInlines.h (294428 => 294429)
--- trunk/Source/_javascript_Core/assembler/SecureARM64EHashPinsInlines.h 2022-05-18 20:39:03 UTC (rev 294428)
+++ trunk/Source/_javascript_Core/assembler/SecureARM64EHashPinsInlines.h 2022-05-18 20:43:25 UTC (rev 294429)
@@ -43,17 +43,19 @@
#if !HAVE(SIMPLIFIED_FAST_TLS_BASE)
result = result & ~0x7ull;
#endif
- return result + 1;
+ return result;
}
template <typename Function>
ALWAYS_INLINE void SecureARM64EHashPins::forEachPage(Function function)
{
- for (Page* page = firstPage(); page; page = page->next) {
+ Page* page = firstPage();
+ do {
RELEASE_ASSERT(isJITPC(page));
if (function(*page) == IterationStatus::Done)
return;
- }
+ page = page->next
+ } while (page);
}
template <typename Function>
_______________________________________________
webkit-changes mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-changes