Patch coming soon.
Created attachment 220913 [details] the patch.
Comment on attachment 220913 [details] the patch. View in context: https://bugs.webkit.org/attachment.cgi?id=220913&action=review r=me > Source/JavaScriptCore/llint/LLIntOpcode.h:48 > + macro(llint_synthesized_label_1, 1) \ > + macro(llint_synthesized_label_2, 1) \ > + macro(llint_synthesized_label_3, 1) \ > + macro(llint_synthesized_label_4, 1) \ > + macro(llint_synthesized_label_5, 1) \ > + macro(llint_synthesized_label_6, 1) \ > + macro(llint_synthesized_label_7, 1) \ Let's call these "llint_cloop_did_return_from_js_N". > Source/JavaScriptCore/llint/LowLevelInterpreter.cpp:93 > +#if 0 Let's give this a name, like ENABLE_OPCODE_TRACING. > Source/JavaScriptCore/offlineasm/asm.rb:48 > + @numSynthesizedGlobalLabels = 0 Let's call this "didReturnFromJSLabelCounter", and move into cloop.rb's Instruction class, using @@ syntax for a per-class variable. > Source/JavaScriptCore/offlineasm/asm.rb:228 > + def numSynthesizedGlobalLabels() > + @numSynthesizedGlobalLabels > + end > + def incNumSynthesizedGlobalLabels() > + @numSynthesizedGlobalLabels += 1 > + end See above renames.
Thanks for the review. The feedback has been applied. Landed in r161927 on the jsCStack branch: <http://trac.webkit.org/r161927>.
*** Bug 126392 has been marked as a duplicate of this bug. ***
*** Bug 126007 has been marked as a duplicate of this bug. ***
*** Bug 126191 has been marked as a duplicate of this bug. ***
(In reply to comment #3) > Thanks for the review. The feedback has been applied. Landed in r161927 on the jsCStack branch: <http://trac.webkit.org/r161927>. The merge had a conflict that went uncaught by svn. This resulted in a redundant push of lr and a redundant pop of cfr, and this breaks the C loop runs (as well as ARMv7 and other similar architectures). This is now fixed in r161942: <http://trac.webkit.org/r161942>.