On Mon, Jun 18, 2018 at 03:06:18PM -0400, Michael Meissner wrote: > On Fri, Jun 15, 2018 at 04:00:42PM -0500, Segher Boessenkool wrote: > > But those aren't types, those are modes. Why would the modes not work? > > That makes no sense. Just like the __ieee128 and __ibm128 types work > > always, so should the IF, KF, IC and KC modes. > > > > So what happens if you try to use KC? > > The problem is the infrastructure of the compiler. The type and builtin > system > really, really does not like having the same external name for a library > function. So if you have a type that uses KFmode and the function resolved to > "__addkf3", and you have another type that uses TFmode, and it too resolves to > "__addkf3" you get an internal error. Ditto for the name mangling. > > So we have to make sure that there are really only two modes being used.
Or just #define TFmode to KFmode (if that is the type being used), and something similar for the types. Similar to how Pmode works. This might not be a simple change, but we'll need to get there eventually. The way TFmode works currently is no end of problems. Segher
