fetch ALL_OPT_URL_FILES as well for GCC 14.

clean ups and gcc 14 support.

- remove some obsolete code
- derive MKNATIVE path from EXTERNAL_GCC_SUBDIR
- add version for netbsd gcc 14

Provide HOST_NM as NM so cross-builds on e.g. macOS and Solaris are ok.

This is an update to PR 58437 which only fixed the original issue on Solaris.

PR toolchain/58437

Ensure that the tools NM is used in MAKE_ARGS.

Building on host systems, e.g. Solaris, where nm is not the GNU nm implementation
is otherwise an issue.

Fixes PR58437.

PR toolchain/58437

introduce a way to build GCC without the isl library

if NOGCCISL is set, don't build or link against isl. should fix PR#58505.

branches: 1.113.2;
tools/gcc/Makefile: force std=c++11 for GCC 12 builds

GCC >= 11 now requires C++11 to build. Impacted individual components
vary whether they test for if std=c++11 is the default or has to be
passed. Regardless, our build infrastructure applies its own overrides.
Force this setting so that we can still build on NetBSD 8.x hosts (and
probably 7.x, too). Build tested on NetBSD 8.3.

tools/gcc/Makefile: fix typo in comment

enable isl support for GCC 12.

Include the local NetBSD GCC version so that bumps to it trigger rebuilds

unfortunately, this will cause one rebuild to get the new name.

tools/gcc: Sprinkle NB_FIX_VAX_BACKEND for vax to enable gcc.old MI fixes

PR port-vax/57646: Import major vax toolchain fix by Kalvis Duckmanton

jump through hoops to make the native build find gmp for gdb.

updates for GCC 12 (relatively tiny).

- gcov-iov.h is now called version.h and we build it directly.
- need to explicitly ask for bversion.h gtyp-input.list to be built.

gcc 10 is now in gcc.old, make mknative-gcc.old gcc-10 happy.

branches: 1.105.2;
Patch the build script and toolchain to allow passing through
HOST_CPPFLAGS, HOST_CXXFLAGS, but not LIBRARY_PATH because it breaks
the build. These variables are necessary to support building NetBSD
from a GNU Guix or NixOS host, where /usr/include, /lib, and all
but /bin/sh do not exist. In many cases, support for HOST_CPPFLAGS
was incomplete. From Ryan Sundberg

Disable again initfini; breaks some archs and not worth dealing with when
we have both gcc's active in the tree.

enable initfini-array since all platforms support it.

add a netbsd version for GCC 10. properly generate omp-device-properties.h.

fetch omp-device-properties.h and don't try to fetch sanitizer
stuff automatically yet.

pass down earm* not earmv5* to mknative-gcc.

disable fetching stuff from libsanitizer for now. it needs libstdc++
to build in mknative using gcc makefiles, and our locale changes break
the "normal" build of this.

make sure MKNATIVE_TARGET is always set. since mknative-gcc doesn't
care about the version any more (we keep two versions for gcc/gcc.old)
simply set it to "gcc".

set MODULE to ${EXTERNAL_GCC_SUBDIR} so this actually works..

add something to auto-detect when GCC has changed from gcc to gcc.old
or back and auto-cleandir tools/gcc.

based upon the code for PREVIOUSTOOLDIR in tools/Makefile.

arm needs more files

remove HAVE_GCC == 6 support, add HAVE_GCC == 8 support

Fix bugurl. PR toolchain/48712

branches: 1.93.2;
enable the workarounds for cortex a53 errata 835769 and 843419 by default.

Remove "--disable-libstdcxx-dual-abi". This option does not do what
it says: Instead of disabling the dual-abi and letting you choose
which one you want, it disables the cxx11 abi and the
"--with-default-libstdcxx-abi=OPTION" becomes a noop. Removing the
old COW compatibility ABI is not easy. For more info see:

https://gcc.gnu.org/onlinedocs/libstdc++/manual/configure.html
https://gcc.gnu.org/onlinedocs/libstdc++/manual/using_dual_abi.html

don't pass --disable-symver to mknative configure. this turns off
the checking for gcc 4.6 and some exception handling routines which
we need on several platforms.

this should fix missing symbol issues seen on current-users.

add || true to the gthr-default.h copy, it can fail first time around

- use --with-default-libstdcxx-abi=new for GCC, otherwise we get the
old ABI with --disable-libstdcxx-dual-abi
- fix libbacktrace mknative
- find gthr-default.h for libstdc++ mknative
- fetch FILESYSTEM_SOURCES

obsolete gcc 5 support.
add gcc 7 support. call it "nb1 20190119"

Pass --without-isl to GCC.

We don't provide one, but if we did the build would fail. This can also
happen if there's a system isl that GCC manages to find.

From mgorny.

branches: 1.86.2;
Create a new bsd.hostinit.mk file and put the build definitions for all host
programs there; make all Makefiles that use bsd.hostprog.mk include it.
Namely turn off MKREPRO and don't make lint, man pages, info files etc.
Remove the Makefile.inc files that contained these same settings, and
remove the settings from Makefile.host

PR/53238: Robert Elz: Disable MKREPRO in tools; the host compiler might
not support the necessary options. This is done thusly:

1. Set MKREPRO=no in Makefile.host. This handles all the Makefiles that
use it and don't include bsd.own.mk.
2. Create Makefile.inc and set MKREPRO=no in it. Change the Makefiles that
include bsd.own.mk, to include bsd.init.mk which includes Makefile.inc
first. This will also allow us to control other tools options from a
single location if we need to.

XXX: pullup-8

branches: 1.84.2;
convert HAVE_GCC handling to modern GCC release numbering:
- HAVE_GCC=5 is now the default (vs. HAVE_GCC=53 we've been using for
GCC 5.4 and GCC 5.5.)
- remove some more GCC 4.8 code. we don't support GCC 4 here.
- adjust set lists to gcc=5 from gcc=53.

add some basic HAVE_GCC=6 handling (totally unused so far.)

- gcc.old == gcc-5.3, so copy mknative-gcc{,.old} and adjust
- disable dual libstdc++ abi per joerg
(https://gcc.gnu.org/onlinedocs/libstdc++/manual/using_dual_abi.html)
will require a major libstdc++ major bump.

move the GMP_MACHINE_ARCH definition into external/lgpl3/gmp/Makefile.arch.
soon to be GMP 6 allows us to have only two arm copies.

branches: 1.81.4;
provide a common softfloat fenv implemenation and use it for softfloat builds.
restore ABI compatibility with previous releases for ieeefp.h on sh3.
add namespace.h protection for all the fenv interfaces.
use MKSOFTFLOAT on sh3 instead of assuming softfloat.
standardize on comparing MKSOFTFLOAT with "no".
remove the arm-specific softfloat fenv code (which also had several bugs).
fix logic errors in the arm hardfloat feraiseexcept() and feupdateenv().

branches: 1.80.2; 1.80.4;
use --with-diagnostics-color=auto-if-env as the default for GCC 5.

unwind.h has a new problem, and fix the grabbing of x86 specific files.

- pull i386-builtin-types.inc automatically on x86
- reenable cxx rt support
- don't grab insn-modes.h, we build it
- rename the c98 and c11 subdir codecvt.cc files while extracting
them with mknative
- grab GGC_H as well, and use it in dependency lists

remove some more GCC 4.5 code. make GCC 5.3 be HAVE_GCC=53, so that,
for example, HAVE_GCC >= 48 works properly.

Add missing closing ) and line continuation for mknative

prepare for GCC 5:
- remove GCC 4.5 stuff, add most very basic GCC 5 stuff.
- use mknative-gcc.old for GCC 4.8
- copy mknative-gcc to mknative-gcc.old (should have done this when we
switched to using gcc.old.)

this should not affect current builds.

we have lto...

use --disable-shared on m68000 for mknative.

add --enable-libstdcxx-time=rt for the GCC 4.8 build.

or1k does not support --with-float

branches: 1.70.2;
remove remaining makefile support for GCC < 45 that i found.

commonize the netbsd local version string. (the native one wasn't
updated, and now that won't be a problem again.)

two fixes skrll@ pointed out:
- pass --enable-libstdcxx-threads to enable a bunch of tests and thus
get the right answers
- MKNATIVE_CONFIG_TARGET_LIBS needs to do libgcc before libstdc++-v3,
for the GCC 4.8 case.

with these we now get the right threading options in the headers.

Update with-pkgversion date for gcc 4.8.3 import.

branches: 1.66.2;
define GCCDIST properly for GCC 4.1, and since it is now the same
as GCCSRCDIR remove the latter.

prepare for moving GCC 4.5 into gcc.old:

- convert to using ${EXTERNAL_GCC_SUBDIR}
- define base-external-gpl3-gcc* subdir as GCC_SUBDIR
- use <bsd.init.mk> over <bsd.own.mk> for a bunch of places; mostly
because it arranges for ../Makefile.inc to be included earlier, and
don't bother including the latter if the former is already included.
- move all .PATH: settings after <bsd.{own,lib}.mk> so that all
valid variables are set before it is evaluated
- rename mknative-gcc* to match their subdir name.

XXX the relationship between the Makefile.inc/Makefile.gcc_path files
is kind of sketchy, it would be great if this was fixed.

put back some cppflags we lost.

Merge gcc-4.8 mknative bits.

The mpfr sources moved down into a src directory - deal with this in
mknative.

update to MPC 1.0.1.

Move the GNUHOSTDIST assignment back to the top of the Makefile.

Group the mknative stuff together - no functional change

g/c VAX_CONFIGURE_ARGS

m68k doesn't support with-float=soft

branches: 1.56.2;
pass both NEWCONFIGDIR and NETBSDSRCDIR to mknative-gcc.

use NETBSDSRCDIR to find mpc/mpfr/gmp, not NEWCONFIGDIR, the
latter which is for finding a r/w directory to write to if
we're running from a read-only tree.

Avoid populating mknative output with gmp,mpfr, and mpc source
directories

Use --with-{mpc,mpfr,gmp}-{lib,include} to avoid the LIBISPRIVATE=no
dance in mknative and avoid poluting the DESTDIR.

Update the README to reflect this.

Update for gcc 4.5.4

branches: 1.51.2;
Always use system zlib. Avoids the need for 32bit build environment when
bootstrapping on some Linux distros. We require zlib to be present anyway.

branches: 1.50.2;
Fix bootstrap-libgcc for gcc 4.5

Fix (harmless?) HAVE_GCC comparison version.

Explicitly request TLS support for mknative.

Pull across the fix for target/50691

ok'ed by mrg@

branches: 1.46.2;
remove pre-gcc-4.5 mknative remnants that have long ceased to function.

add libgomp support. needed by "cc -fopenmp", that koffice 2.3 wants to use.

force HAVE_AS_SPARC_GOTDATA_OP to false, since those relocs seem
busted with our binutils. now sparc64 seems to work fine.

-I$DESTDIR/usr/include is escaping into the generated files from
the native-gcc run. add some sed commands to remove them.

- call each ${MKNATIVE_CONFIG_TARGET_LIBS} target separately
- comment out maybe-all-target-libgcc
- cope with the special transformation needed for gtyp-input.list

- repurpose MKNATIVE_CONFIG_TARGET_LIBS to list all the targets
- disable building pch for the mknative run

get libdecnumber info out of GCC.

get this some what working with GCC 4.5. i at least have managed to
get crtstuff and libgcc to build for amd64, and have linked and tested
/bin/sh with all these GCC 4.5 compiled part.

use --sysroot with GCC 4.5. (maybe could with 4.1 too? the 4.1 code
here is broken right now, anyway, ever since the sysroot changes were
merged into our tree.)

initial port to HAVE_GCC = 45. mknative-gcc parts are from me
and are largely untested, most of the rest is from chuq.

If building the cross gcc for softfloat, add --with-float=soft so it will
default to it.

branches: 1.35.2;
Enable support for --sysroot in binutils and gcc.

revert the previous; not intended to be part of that commit, or any commit

rework the main SUBDIR list here as suggested by enami. put .WAIT on
the same line as the target that needs to wait, but before the target,
rather than on the previous line, which led to confusion and an error
being introduced in the previous change.

Enable -mtune=nocona by default on x86. I have tested on a range of CPUs
and this seems to give the best results.

-introduce a "COMMON_CONFIGURE_ARGS" variable which is included by
both CONFIGURE_ARGS and NATIVE_CONFIGURE_ARGS to reduce duplication
between tool and native configuration
-allow to pass a "--with-arch" argument to both configurations

branches: 1.30.4; 1.30.12;
Use ${TOOL_SED} instead of just "sed" while building other tools.

Where sed is used in a != assignment, use
.if make(depend) || make(all) || make(dependall) || make(install)
to ensure the command doesn't run at "make obj" time when TOOL_SED will
not have been built.

Use ${TOOL_AWK} instead of ${AWK} or plain "awk" in make commands.
Pass AWK=${TOOL_AWK:Q} to shell scripts that use awk.

- add a i386/t-netbsd64 makefile fragment to build 32/64 bit
libgcc/multilib on amd64
- enable the 32/64 bit libgcc/multilib support on sparc64
- adapt mknative-gcc to grab multilib.h
- use --enable-multilib on amd64 and sparc64

none of this affects the installed tools yet, just the src/tools one.

Force use of __cxa_atexit by a configure command line argument
(for booth the tool build and mknative). gcc's "config.gcc" sets
it if the target is netbsd[2-9], but since the target is w/o version
number in our builds, it has no effect. Found by Nick Hudson,
patch approved by mrg.

branches: 1.26.4;
Instead of inventing a new environment variable to override the
SSP functionality location test, use gcc_cv_libc_provides_ssp instead,
as hinted by freza@ and uwe@.

Modify the gcc configure script to allow override of the "target libc
has ssp functions built-in" test via the TARGET_LIBC_PROVIDES_SSP environment
variable, to allow us to configure a cross-compiler appropriately without
having to try to find out by looking in the target's source directory.

Tweak our build to tell gcc that the ssp bits are now in libc.

The native compiler appears to already think that the ssp bits
live in libc, so no change appears to be needed there.

The autoconf-generated configure script will be committed separately shortly.

Remove non-shared hooks for VAX.

G/C some HAVE_GCC == 3 stuff

branches: 1.22.2; 1.22.4;
pass RAW_GCC_FOR_TARGET when running configure-target-libstdc++-v3.
now iconv and wchar_t things are properly set. still missing
_GLIBCXX_HAVE_S_IFREG and _GLIBCXX_USE_NLS.

force --disable-shared on vax so that GCC does not think it can use
shared libgcc.

introduce a new CXXADDFLAGS and use it to pass CPPFLAG_ISYSTEMXX to
configure. also, clean up this a little more to make mknative-gcc
work on GCC3 again, while fixing some breakage in the last couple
of revisions.

make this work for GCC3 one again.

configure-libiberty is only good for GCC4.
move a comment next to it's code.

set AR and RANLIB to the host versions in MAKE_ARGS - otherwise the
build of the "build-libiberty" fails due to these being set to "".

support GCC4:
- look in gnu/dist/gcc4
- don't build fortran
- use /usr/bin/env -i so that gmake works (XXX - breaks from MAKEFLAGS
in the build.sh generated makewrapper.)
- use gmake and Makefile.gmakehost
- update .native/.configure_done target

Use the correct LD/NM/OBJDUMP when creating the mknative environment.

Use the correct AS when creating the mknative environment.

Reviewed by mrg@

mknative re-runs will appear soon.

branches: 1.13.6;
Be more gcc like in the way we deal with libgcc so that sh3/hppa millicode
and exception handling have a chance of working properly.

- creates libgcc, libgcc_eh and libgcc_s

- updates LIBGCC_SPEC to use them appropriately.

There's a hack in here at the moment with respect to libgcc_so in that it
is preferable to link against libgcc_so will only when -shared-libgcc is
specified (the c++ frontend does this automatically.) Configurations where
LINK_EH_SPEC is defined already do this. The gcc configuration for
NetBSD/alpha and another NetBSD platform (I forget which) actually define
LINK_EH_SPEC probably by accident rather than design.

- updates share/mk to use the compiler's knowledge of what needs linking into
libraries and executables. This removes an hppa hack.

- updates the sets for the newly created libgcc* files.

- support for linking against the _pg version of libgcc has been removed.

Updates so that mknative gets the right things.

- Disable symbol versioning (for now)
- Make sure that libiberty knows its being configured with a
cross compiler.
- The CXX_* variables are no longer needed/used.
- LIB2FUNCS_EXTRA gets pulled in via LIB2ADD
- Get LIB1ASMFUNCS and LIB2ASMSRC
- MAYBE_USE_COLLECT2 got renamed to USE_COLLECT2 (but might not
be used)
- Get EXTRA_HEADERS so that we get generate the right paths for
CPPFLAGS
- Get some variables related to shared libgcc

fix a comment; normalise.

trim FIND_ARGS. remove dead targets.

be a little more verbose about what we're doing when running mknative.
also grab LIB2FUNCS_EXTRA for libgcc3 (old fix for m68000).

Use ${HOST_SH} instead of `sh'.

If necessary, pull in <bsd.sys.mk> to get the definition of HOST_SH;
Makefiles that pull in one of (most of) <bsd.*.mk> will get this anyway.

Set EXTRA_PARTS to a null value, like we do with LIBGCC. This prevents
the tool build from attempting to build the GCC crtstuff before the
target include files are installed. PR toolchain/23112.

(Note: You would not see the bad behavior described in the PR unless
your .mk files were hacked to configure the toolchain target as
*-*-netbsd*2.0.)

for mknative, make sure we look in the target's /lib directly as the
symlinks in the /usr/lib directory point (absolutely) to "/lib/libfoo.so",
which ends up being quite invalid when target!=host...

force long long support. use -isystem* for CCADDFLAGS

we need config.h an another version of gthr-default.h

pass --enable-threads to ./configure

don't configure with multilibs when doing mknative

- build "binutils" and "gcc" separately if USE_TOOLS_TOOLCHAIN is not set
to "yes".
- adjust dbsym/mdsetimage to look in the correct place for libbfd.
- new mknative-gcc based in toolchain/mknative