Otherwise, stackmaps will think that no registers are in use.
Created attachment 264462 [details] the patch
Comment on attachment 264462 [details] the patch View in context: https://bugs.webkit.org/attachment.cgi?id=264462&action=review LGTM > Source/JavaScriptCore/b3/air/AirGenerate.cpp:79 > + recordUsedRegisters(code); Do we cache the results of the used registers? If not, maybe a better name would be just reportUsedRegisters like the function name? > Source/JavaScriptCore/b3/air/AirRecordUsedRegisters.cpp:67 > + Style: I think we can drop the new lines at the end of the file. > Source/JavaScriptCore/b3/air/AirRecordUsedRegisters.h:35 > +// This allocates StackSlots to places on the stack. It first allocates the pinned ones in index Is this really what this phase does?
(In reply to comment #2) > Comment on attachment 264462 [details] > the patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=264462&action=review > > LGTM > > > Source/JavaScriptCore/b3/air/AirGenerate.cpp:79 > > + recordUsedRegisters(code); > > Do we cache the results of the used registers? > If not, maybe a better name would be just reportUsedRegisters like the > function name? Good point. > > > Source/JavaScriptCore/b3/air/AirRecordUsedRegisters.cpp:67 > > + > > Style: I think we can drop the new lines at the end of the file. OK. > > > Source/JavaScriptCore/b3/air/AirRecordUsedRegisters.h:35 > > +// This allocates StackSlots to places on the stack. It first allocates the pinned ones in index > > Is this really what this phase does? Ooops!
Landed in http://trac.webkit.org/changeset/191845