> The call fesetround could be any function in practice, and we never > know if that function might use dynamic rounding mode floating point > operation or not, also we don't know if it will be called fesetround > or not. > > So that's why we want to restore before function call to make sure we > don't break rounding mode for any inner function, also backup the frm > after function call can handle the case if the inner function has > changed global rounding mode.
Ah, that clarifies things a bit. Yes, I figured - the function is opaque anyway. I think we're touching a general question here, I suppose there already have been discussions on the LLVM side about this? What about ABI implications? By saving and restoring before and after a function call we assume the FRM register to be call clobbered instead of letting the function/call save/restore it. Apart from fesetround I can't imagine many cases where we would want to re-use the rounding mode from inside a call so it might be advantageous to leave the burden of saving/restoring to the callee. But that's a similar discussion as with the vector ABI so I don't expect a conclusion ;) Regards Robin
