> -----Original Message----- > From: Richard Henderson <[email protected]> > Sent: Sunday, February 14, 2021 5:15 PM > To: Taylor Simpson <[email protected]>; [email protected] > Cc: [email protected]; [email protected]; [email protected]; > [email protected]; Brian Cain <[email protected]> > Subject: Re: [PATCH v8 17/35] Hexagon (target/hexagon/fma_emu.[ch]) > utility functions > > On 2/7/21 9:46 PM, Taylor Simpson wrote: > > Redundant with softfloat. Is the default nan really -1? I suppose then that > hexagon does not distinguish QNaN from SNaN? > > You'll want to patch fpu/softfloat-specialize.c.inc for both of these choices: > no_signaling_nans and parts_default_nan.
Yes, the default is really -1. We do distinguish SNaN - when a SNaN is fed to an arithmetic operation, it raises the invalid flag and a QNaN is returned. So, it looks like I'll need to patch parts_default_nan and parts_silence_nan as well as set default_nan_mode to 1. > This is duplicating code from include/fpu/softfloat-macros.h, except for the > wrapping to Int128. That said, I don't think you should actually need this, > or, frankly, the vast majority of the rest of your fp code. OK, I will look into this. Thanks, Taylor
