On 5/26/26 03:35, Peter Maydell wrote:
I think the use of these boolean arguments leaves this looking
less readable than what we had here before, unfortunately
(now you have to go look at the function prototype to figure
out what "false, false, true" means, whereas direct calls
to "set_default_nan_mode()" or "set_flush_to_zero()" are
immediately clear about what they're setting.)
I think it would be clearer to have an arm_init_fp_status()
which doesn't take any arguments and which we use on every
fp_status[] element, before then calling the functions that
override the necessary settings from that default.
Ok, done.
r~
