br_cc definition

[Hal Finkel]

----- Original Message -----
> From: "deadal nix" <deadalnix at gmail.com>
> To: "Hal Finkel" <hfinkel at anl.gov>
> Cc: "llvm-commits" <llvm-commits at cs.uiuc.edu>
> Sent: Wednesday, September 24, 2014 12:29:12 AM
> Subject: Re: br_cc definition
> 
> 
> The argument 0 is the condition code type. 1 and 2 are arguments. 3
> it the target of the branch. So SDTCisVT<0, OtherVT> and should not
> be an int.
> 

Ah, you're right: Those are represented by CondCodeSDNode (not integer constants). LGTM.

 -Hal

> 
> 
> 2014-09-23 5:42 GMT-07:00 Hal Finkel < hfinkel at anl.gov > :
> 
> 
> 
> 
> ----- Original Message -----
> > From: "deadal nix" < deadalnix at gmail.com >
> > To: "llvm-commits" < llvm-commits at cs.uiuc.edu >
> > Sent: Tuesday, September 23, 2014 3:20:59 AM
> > Subject: br_cc definition
> > 
> > 
> > 
> > As per
> > http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-August/075321.html
> > ,
> > it look like I'm not the only one to want to match br_cc directly.
> > So I prepared a patch.
> 
> +def SDTBrCC : SDTypeProfile<0, 4, [ // brcc
> + SDTCisVT<0, OtherVT>, SDTCisSameAs<1, 2>, SDTCisVT<3, OtherVT>
> +]>;
> +
> 
> Shouldn't the type constraint for the condition code be SDTCisInt<0>?
> 
> -Hal
> 
> > 
> > _______________________________________________
> > llvm-commits mailing list
> > llvm-commits at cs.uiuc.edu
> > http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
> > 
> 
> --
> Hal Finkel
> Assistant Computational Scientist
> Leadership Computing Facility
> Argonne National Laboratory
> 
> 

-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory