On Sat, 25 May 2019, Alex Henrie wrote: > On Sat, May 25, 2019 at 12:34 AM Richard Biener <rguent...@suse.de> wrote: > > > > On May 24, 2019 5:49:38 PM GMT+02:00, Alex Henrie <alexhenri...@gmail.com> > > wrote: > > >As far as I can tell, "fndecl" is a misnomer: these attributes are > > >more accurately called "function definition attributes", i.e. > > >attributes that affect the assembly code of the function but do not > > >affect its calling convention. > > > > Yes. The others are attributes also applicable to function types to > > properly support attributed indirect calls. > > I'm happy to rename ix86_handle_fndecl_attribute to > ix86_handle_fndef_attribute then. Should I do that in this patch or a > separate patch?
I don't see this fixes anything. It's pretty well-know that attributes affecting call semantics have to be accepted both at type and decl position and that they should be added to the type attributes. Attributes that only affect code generation for the function itself should be added to decl attributes and do not necessarily need to be accepted in type position.
