Changeset 237595 in webkit for trunk/Source/JavaScriptCore/assembler/ProbeContext.cpp

Timestamp:

Oct 30, 2018, 11:32:20 AM (7 years ago)

Author:

Ross Kirsling

Message:

useProbeOSRExit causes failures for Win64 DFG JIT
​https://bugs.webkit.org/show_bug.cgi?id=190656

Reviewed by Keith Miller.

• assembler/ProbeContext.cpp:

(JSC::Probe::executeProbe):
If lowWatermark is expected to equal lowWatermarkFromVisitingDirtyPages *regardless* of the input param,
then let's just call lowWatermarkFromVisitingDirtyPages instead.

• dfg/DFGOSRExit.cpp:

(JSC::DFG::OSRExit::executeOSRExit):
The result of VariableEventStream::reconstruct appears to be inappropriate for direct use as a stack pointer offset;
mimic the non-probe case and use requiredRegisterCountForExit from DFGCommonData instead.
(Also, stop redundantly setting the stack pointer twice in a row.)

File:

• trunk/Source/JavaScriptCore/assembler/ProbeContext.cpp (modified) (1 diff)

trunk/Source/JavaScriptCore/assembler/ProbeContext.cpp

@@ -61 +61 @@
     if (context.hasWritesToFlush()) {
         context.stack().setSavedStackPointer(state->cpu.sp());
-        void* lowWatermark = context.stack().lowWatermark(state->cpu.sp());
+        void* lowWatermark = context.stack().lowWatermarkFromVisitingDirtyPages();
         state->cpu.sp() = std::min(lowWatermark, state->cpu.sp());