armv7-unknown-linux-gnueabihf target spec contains apparently nonsensical feature combination

[RalfJung]

The target spec for this target sets the following features: +v7,+vfp3,-d32,+thumb2,-neon. However, we have also set vfp3 to imply d32, which means -d32 implies -vfp3, so this makes little sense -- either the +vfp3 or the -d32 must be wrong. Or did I misunderstand something?

We don't have target maintainers listed for that target so I am not sure whom to ping... Cc @workingjubilee @beetrees

[RalfJung]

@rustbot ping arm

[rustbot]

Hey ARM Group! This bug has been identified as a good "ARM candidate".
In case it's useful, here are some instructions for tackling these sorts of
bugs. Maybe take a look?
Thanks! <3

cc @adamgemmell @davidtwco @hug-dev @jacobbramley @JamieCunliffe @joaopaulocarreiro @raw-bin @Stammark

[workingjubilee]

Huh.

I believe it's trying to indicate specifically the "vfp3-d16" configuration.

[workingjubilee]

vfp3 implying d32 doesn't actually make sense, I believe.

[RalfJung]

vfp3 implying d32 doesn't actually make sense, I believe.

That was apparently added all the way back in #128221. Cc @calebzulawski @Amanieu

[calebzulawski]

The table was adapted from LLVM, though I could have made a mistake. This is the current relationship between vfp version and d32 (I'm having a little trouble parsing it): https://github.com/llvm/llvm-project/blob/8f36135aea5dfbb5f090ca9ad055094c9913f735/llvm/lib/Target/ARM/ARMFeatures.td#L73

[RalfJung]

I don't see that line being about vfp3? vfp3 is defined a few lines down, here. It seems to imply vfp2 but nothing else.