cc @krytarowski

https://github.com/NetBSD/pkgsrc-wip/blob/master/gdb-netbsd/patches/patch-gdb_arm-nbsd-tdep.c

The patch calls:

+  set_gdbarch_iterate_over_regset_sections
+    (gdbarch, armbsd_iterate_over_regset_sections);

However, that function seems to assume the FPA architecture for floating point registers:
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=blob;f=gdb/arm-bsd-tdep.c;h=8cef77297bba2a28385b9647e44ac7f3768d60ca;hb=69ed9b74b61359d43a6f5f5a295c3cfb7f3ee61c#l49
(it doesn't use ARM_D0_REGNUM, etc.).

Also, I would assume (but don't know) that NetBSD core files store FP registers as a struct fpreg, which does not match the layout assumed by that function.