[Bug c/77530] optimization prevents excess precision from being removed with x86/amd64 long double and rounding to 53 bits

[joseph at codesourcery dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77530

--- Comment #1 from joseph at codesourcery dot com <joseph at codesourcery dot 
com> ---
The x86 back end uses TARGET_96_ROUND_53_LONG_DOUBLE only for the case of 
32-bit mode on DragonflyBSD and FreeBSD.  From what you're saying here, it 
needs to do so for NetBSD as well (both 32-bit and 64-bit?).