[Bug target/104455] Cannot select -march=armv7-a using GCC 11

["rearnsha at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104455

--- Comment #5 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
(In reply to Jeffrey Walton from comment #4)

> 
> So stepping back to 10,000 feet, we now need to specify options and ISA's we
> are not using. That seems like a bug to me. I'm not sure I would consider
> this fixed.
> 
> Where is it going to stop? How many non-used options that I am not aware of
> will I need to specify?
> 
> I think either GCC or Debian needs to fix this. This could be a GCC bug
> because GCC apparently knows there's a fp unit but it chooses to ignore it.
> Instead it wants me to say it again. This could be a Debian bug because they
> need to completely (not partially) configure things.

No, it knows it /needs/ a FP unit, but you haven't told it which one is
available.  The Arm architecture has many variants and not all are necessarily
compatible with a base set, so you have to explicitly state which one is
intended.  Assumptions are a potential source of massive misunderstandings.

Note, in the past the FPU variant was baked into the compiler when it was
configured and this lead to all forms of confusion because if you had a Neon
FPU and then started compiling for a cortex-m device you could still end up
with Neon instructions in your code because there was no explicit option to
turn it off.  We've gone as far as we can by allowing you to simply specify
'+fp' rather than having to know exactly which variant of the VFP architecture
goes with armv7-a.

Also note that clang has the same option logic for CPU selection.