/builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/jit/CCallHelpers.h:171:13: error: no matching function for call to 'JSC::CCallHelpers::swap(JSC::ARMRegisters::FPRegisterID&, JSC::ARMRegisters::FPRegisterID&)' swap(source, dest); ^~~~ In file included from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/assembler/MacroAssembler.h:51, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/runtime/BasicBlockLocation.h:29, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/runtime/ControlFlowProfiler.h:29, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/runtime/VM.h:35, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/bytecode/ExecutableToCodeBlockEdge.h:31, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/runtime/FunctionExecutable.h:28, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/runtime/JSFunctionInlines.h:28, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/bytecode/AccessCase.h:30, from /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/bytecode/AccessCase.cpp:27, from /builddir/build/BUILD/webkitgtk-2.21.1/armv7hl-redhat-linux-gnueabi/DerivedSources/JavaScriptCore/unified-sources/UnifiedSource25.cpp:1: /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/assembler/MacroAssemblerARM.h:646:10: note: candidate: 'void JSC::MacroAssemblerARM::swap(JSC::AbstractMacroAssembler<JSC::ARMAssembler>::RegisterID, JSC::AbstractMacroAssembler<JSC::ARMAssembler>::RegisterID)' void swap(RegisterID reg1, RegisterID reg2) ^~~~ /builddir/build/BUILD/webkitgtk-2.21.1/Source/JavaScriptCore/assembler/MacroAssemblerARM.h:646:10: note: no known conversion for argument 1 from 'JSC::ARMRegisters::FPRegisterID' to 'JSC::AbstractMacroAssembler<JSC::ARMAssembler>::RegisterID' {aka 'JSC::ARMRegisters::RegisterID'}
Interesting. My memory is hazy, but does ARM's EABI pass floating point arguments in FP registers?
I could add that it's ARMv7 hardware without Thumb 2 support.
Created attachment 338756 [details] Patch
(In reply to Mark Lam from comment #1) > Interesting. My memory is hazy, but does ARM's EABI pass floating point > arguments in FP registers? hardfp one uses this.
Created attachment 338760 [details] Patch
Comment on attachment 338760 [details] Patch Nice work. r=me
Thanks for your review!
Comment on attachment 338760 [details] Patch Clearing flags on attachment: 338760 Committed r231011: <https://trac.webkit.org/changeset/231011>
All reviewed patches have been landed. Closing bug.
<rdar://problem/39731210>
Thank you Yusuke for the fix!