Hello @Jeremy1 ,
I am afraid you have hit a bug in our probing of the compiler. See [CPP-5313] - Jira
In summary, we forward -mfloat-abi=hard
to GCC, but we do not forward -march=armv8.1-m.main+mve.fp+fp.dp
. The compiler defaults to the cpu arm7tdmi
, which has no FPU support, and errors out.
As a workaround, explicitly adding -mcpu
when compiling may fix the issue, as that one we do forward. For instance, when probing with -mcpu=cortex-m55
we get the macros, and the compiler figures the arch by itself (-march=armv8.1-m.main+fp.dp+mve.fp
), so we will get meaningful predefines.
Please, let me know if this helps.