[Bug target/70465] [4.9/5/6/7 Regression] Poor code for x87 asm

[law at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70465

--- Comment #10 from Jeffrey A. Law <law at redhat dot com> ---
Coming back to Uros's c#5, the more I think about it, that's the solution. 
reg-stack ought to be able to use dependency information  to determine if it
can swap two insns rather than inserting a fxchg.  

I'm not actually working on this, but wanted to make sure anyone looking at
this in the future is aware that Uros's suggestion is, IMHO, the right way to
go.