[Bug c/84190] [7/8 Regression] double arithmetic on x86 no longer rounds to nearest

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84190

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
For the required part I guess we need Joseph's input, though even if it might
not be technically required, I thought that for volatile our general approach
has been "be conservative when you see volatile".