https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #13 from Tom Lane ---
After further experimentation, it seems to me that:
* There was a behavioral change between gcc 9.3.1 and the later releases I
tested. Specifically, in 9.3.1 a -march switch does not override the
platform-sele
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #12 from Tom Lane ---
This does seem like there is some confusion between arm32 and aarch64.
Specifically, having +nofp and not +fp makes sense on the aarch64 side. It
doesn't make sense for arm32, but nonetheless that's how it's a
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #6 from Tom Lane ---
Further info: I dug up an ancient Fedora 30 ARM32 system, and it does not
exhibit this behavior:
$ touch foo.c
$ gcc -c -march=armv8-a foo.c
[fine]
$ gcc -c -march=armv8-a -mfloat-abi=hard foo.c
[also fine]
Thi
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #5 from Tom Lane ---
As far as that goes: I don't think there are any armv8-a machines without an
FPU, either. The v8a spec technically allows it in 32-bit but not 64-bit mode;
but desultory searching doesn't find any examples.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #10 from Tom Lane ---
Well, I simplified the report a bit compared to the real use-case. What we
actually want to use is "-march=armv8-a+crc" so as to compile hardware CRC
instructions (that will be used only if a runtime check veri
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #15 from Tom Lane ---
(In reply to Richard Earnshaw from comment #14)
> GCC implements the following extensions for the v7 architecture:
>
> +fp - add floating point (as defined by the base architecture)
> +simd - add floating point
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
Bug ID: 117766
Summary: Confusion over whether ARM32 -march=armv8-a implies
floating point support
Product: gcc
Version: 10.5.0
Status: UNCONFIRMED
Severity: n
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #3 from Tom Lane ---
No go on fpu=auto, at least on the NetBSD platform:
$ gcc -c -march=armv8-a -mfpu=auto foo.c
cc1: error: '-mfloat-abi=hard': selected processor lacks an FPU
The simd method does work to stop the error:
$ gcc -
