Add missing EABI_32BIT_DUMMY_ARG arguments for some callOperation(J_JITOperation_EGReoJ, ...) overloads
Created attachment 275542 [details] Patch
Comment on attachment 275542 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=275542&action=review Nice catch! > Source/JavaScriptCore/dfg/DFGSpeculativeJIT.h:1742 > + m_jit.setupArgumentsWithExecState(arg1, arg2, EABI_32BIT_DUMMY_ARG arg3Payload, arg3Tag); Is SH4_32BIT_DUMMY_ARG necessary here too? (BTW, at a glance, there are several missing SH4_32BIT_DUMMY_ARG. I'm not sure there are any active users of SH4 port...) > Source/JavaScriptCore/dfg/DFGSpeculativeJIT.h:1854 > + m_jit.setupArgumentsWithExecState(arg1, arg2, EABI_32BIT_DUMMY_ARG arg3Payload, arg3Tag); Ditto.
Comment on attachment 275542 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=275542&action=review >> Source/JavaScriptCore/dfg/DFGSpeculativeJIT.h:1742 >> + m_jit.setupArgumentsWithExecState(arg1, arg2, EABI_32BIT_DUMMY_ARG arg3Payload, arg3Tag); > > Is SH4_32BIT_DUMMY_ARG necessary here too? > (BTW, at a glance, there are several missing SH4_32BIT_DUMMY_ARG. I'm not sure there are any active users of SH4 port...) I don't know. The comment for SH4_32BIT_DUMMY_ARG suggests yes, but in that case there are a lot of these dummy args missing already, and I'm not aware of any public builder testing for this configuration. Not being able to test on this platform makes me reluctant to try and put args into the proper places.
Comment on attachment 275542 [details] Patch Clearing flags on attachment: 275542 Committed r199052: <http://trac.webkit.org/changeset/199052>
All reviewed patches have been landed. Closing bug.