Rafael Espíndola
2014-Mar-27 18:02 UTC
[LLVMdev] [cfe-dev] Named register variables GNU-style
> Aren't you going to need some kind of "private unnamed_addr" thing > just to make syntactically valid IR?I guess the best would probably be a metadata string, that way we are sure we don't actually output anything. Cheers, Rafael
On 27 March 2014 18:02, Rafael Espíndola <rafael.espindola at gmail.com> wrote:> I guess the best would probably be a metadata string, that way we are > sure we don't actually output anything.Hi Rafael, I don't get the risk here, why would that risk any output? I can see that the @llvm.annotation intrinsics use metadata for text, but they discard them on use. It may be simpler to get an MDNode and convert it into text, but if the intrinsic is *always* converted-or-fail, I don't understand how it could output any unwanted values. cheers, --renato
On 28 March 2014 10:05, Renato Golin <renato.golin at linaro.org> wrote:> On 27 March 2014 18:02, Rafael Espíndola <rafael.espindola at gmail.com> wrote: >> I guess the best would probably be a metadata string, that way we are >> sure we don't actually output anything. > > Hi Rafael, > > I don't get the risk here, why would that risk any output?Ignore me. That was meant for the global variable, not the string on the intrinsic. cheers, --renato