Switch over m68k soffloat build to libc/softfloat.

m68k would previously use softfloat functions contained in libgcc_s.

This change aligns m68k with other platforms built with softfloat.

It is still possible to build the gcc_s softfloat impementation by setting
MKSOFFTFLOAT=yes MKLIBCSOFTFLOAT=no

Addresses PR/59393.

all: remove 'constant in conditional context' from LINTFLAGS

Lint no longer emits this message, as it produced too many false
positives.

compiler_rt: suppress lint warnings about missing prototypes

more lint for clang.

more clang lint appeasing

branches: 1.50.2;
lib/libc/compiler_rt/Makefile.inc: Sort LINTFLAGS lines.

No functional change intended.

libc: Pacify lint on aarch64.

fix clang lint build.

compiler_rt: allow signed bit shifts for __negv

Standard C defines '1 << 31' as undefined behavior, but
https://gcc.gnu.org/onlinedocs/gcc/Integers-implementation.html allows
it for GCC.

Add vax

more annotation for powerpc

fix more alpha

add lint flags for compiler_rt

libc/compiler_rt: Consistently use LIBC_MACHINE_{ARCH,CPU}

Instead of MACHINE_{ARCH,CPU}, respectively.

With this change, unnecessary comparetf2.c is no longer
compiled in for 32-bit compat libc on aarch64.

libc: switch aarch64 to use softfloat functions from compiler_rt

The old definitions in qp.c being all grouped together in one file
causes problems when static linking with libgcc (i.e. cc --static-libgcc)
due to functions like __trunctfdf2 conflicting with libgcc, as seen in
PR 57021

We can also add some missing functions like __fixdfti for converting
a double to an int128_t, the lack of which is currently preventing webkit
from linking on aarch64, as seen in PR 57022

Unclear to me if libc is the right place for these functions, but
we can avoid breaking compatibility right now and maintain the status
quo while avoiding some obvious immediate problems.
nm output for libc shows no functions being removed by this change.

PR port-arm/55897

Fix ABI mismatch for armhf runtime routines for floating-point arithmetics;
For hard-float arm variants, provide

(1) generic runtime routines with correct calling convention, and
(2) EABI runtime routines at the same time.

I've confirmed that no binary changes for kernels.

LGTM by skrll

Treat aarch64eb the same as aarch64

Restrict usage of m68k assembler versions of {,u}divsi3 and {,u}divsi3 to
kernel and bootloader for 68010.

They requires a special calling convention to udivsi3, and cannot to be
mixed up in normal routines provided by libgcc or compiler_rt. Although,
there's no problem for using them in a controlled situation, i.e., kernel
and standalone programs.

Note that this does not affect at all m68k ports other than sun2, since
codes generated by gcc do not call these routines.

Assembler files are moved from common/lib/libc/arch/m68k/gen to
sys/lib/libkern/arch/m68k in order not to be compiled in libc.

Revert hack introduced to lib/libc/compiler_rt/Makefile.inc rev 1.37:
http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libc/compiler_rt/Makefile.inc#rev1.37

Proposed on port-sun2@ with no response...
(Again, this does not affect m68k ports other than sun2.)
http://mail-index.netbsd.org/port-sun2/2020/03/10/msg000102.html

Fix broken printf(3) %d output for numbers more than two digits, e.g.,

printf("%d\n", 42) ---> "::" instead of "42"

Our __{,u}modsi3 codes assume that __udivsi3 returns remainder to
%d1 (volatile register). __udivsi3 in libgcc does not, and therefore
mixing them up results in mess.

Always include soft-float compare and order logic in libc when using
compiler-rt.

branches: 1.35.4; 1.35.6;
Do not use IEEE 754 specific functions for VAX

Implement __clear_cache using __syncicache on PPC

branches: 1.33.2;
Use MACHINE_ARCH for the sparc check.

Add __clear_cache implementatin for SPARC/SPARC64.

Drop fixxfdi.c for m68k (where double == long double and this isn't needed
but causes duplicate symbols with softfloat, if compiling MKSOFTFLOAT=yes)

Make sure to define __SOFT_FP__ when building a SOFTFLOAT userland

Fix buidling with MKPROFILE=no

Sync with changes in the profiling implementation of compiler-rt.
Switch to the "Linux" runtime, which is kind of misnamed for the mostly
ELF based implementation. Make sure to hook up the runtime source
only for the profile library as it uses constructors for the output
creation.

Changes for OR1K

branches: 1.26.2;
VAX does need __clzsi2

powerpc64 uses the same sources as powerpc and needs the long double as
double-double sources.

branches: 1.24.2;
MKCOMPAT fixes for when compat MACHINE_CPU != normal MACHINE_CPU

Add the not-yet-used IEEE quad support files.

Changes to existing files to enable building AARCH64 userland.
evbarm64-el
This is clang only. While gcc4.8 supports aarch64, no netbsd support has
been written for aarch64 with gcc4.8.

Don't include atomic.c, we have aliases for it in src/common.

branches: 1.20.2;
For the moment, don't complain about const casts.

Prepare in case we want to hook up the profile init code later.

Deal with PGO rename.

branches: 1.17.2;
On sh3 __udivsi3 is gcc "millicode" with special calling convention
(less registers clobbered than usual). Each DSO that needs it gets
its own hidden copy from libgcc.a. Do not compile udivsi3.c for sh3
here since this C version doesn't conform to what gcc expects and
overrides the libgcc.a version.

This makes binaries compiled with gcc4.8 work. The issue was not
exposed with gcc4.5 since it uses different division strategy by
default.

Change remaining "MACHINE_ARCH" to "LIBC_MACHINE_ARCH"

Update compiler-rt glue for the new source layout.

Long double support builds for PPC now.

ARM needs __clear_cache.

Must use LIBC_MACHINE_ARCH for deciding whether 64bit support code is
neded. Otherwise e.g. i386 compat libs are broken.

__fixsfdi is required, __fixdfsi already in softfloat.

Use a separate variable for the use of GCC's unwind code.

Don't use aeabi_dcmp.S and aeabi_fcmp.S on ARM yet, softfloat provides
the same set of symbols.

Comment out softfloat code until it is decided what to use long term.

__floatunsidf and __floatunsisf are provided by softfloat, but
__fixdfdi and __fixdfsi are not. So move them into the corresponding
chunks.

Explicitly enumerate platforms that have unwind support to make
piecewise migration easier.

Don't build common quad routines if the platform is 64bit since they
will not be emitted by the compiler.

Add profiling helper functions to the build.

Don't include bswap for ARM right now, the routines require ARMv6.
Separate the soft-float code and don't include it in the libgcc case.

Add ARM (EABI) specific sources. Split off code that requires C11,
unwind support in libc or overlaps with soft-float in preparation for
using compiler-rt in the non-clang case.

Build glue for including compiler-rt in libc, not hooked up yet.