Commit Graph

54 Commits

Author SHA1 Message Date
Ed Maste
01b7ad0400 libcompat: disable retpoline when building build tools
These are built with the host toolchain which may not support retpoline.
While here, move the MK_ overrides to a separate line and sort them
alphabetically to support future changes.

MFC with:	r339511
Sponsored by:	The FreeBSD Foundation
2018-11-03 19:31:11 +00:00
Alex Richardson
a00cd9540d Fix non-FreeBSD host lib32 build for TARGET=amd64
When building on non-FreeBSD systems we need to pass an explicit target
triple to clang otherwise it will attempt to build with the host triple.
This also has advantages when building on a FreeBSD host: we now tell
clang that we are targeting at least FreeBSD 12.0 instead of an older
version so it can enable newer features.

Reviewed By:	brooks (mentor)
Approved By:	jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D16842
2018-08-23 18:19:10 +00:00
John Baldwin
52f9bacb0e Install the 32-bit compat sanitizer libraries.
The lib32 build was already building the i386 version of
the clang sanitizers (libclang_rt) but they were not being
installed.  This enables the installation.

MK_TOOLCHAIN=no was originally added to the install make
environment to disable includes so that NO_INCS could be
removed.  The MK_TOOLCHAIN in bsd.incs.mk was subsequently
renamed to MK_INCLUDES, but bsd.lib.mk doesn't even include
bsd.incs.mk when LIBRARIES_ONLY is defined which the install
make environment for compat libs now defines.  However,
setting MK_TOOLCHAIN=no forced MK_CLANG=no which disabled
libclang_rt during the install32 phase.  Remove MK_TOOLCHAIN=no
since LIBRARIES_ONLY is now sufficient.

Since the libcompat environment overrides both LIBDIR and
SHLIBDIR, libclang_rt/Makefile.inc has to set both variables
to force the libraries to be installed to the location
expected by the compiler.

Reviewed by:	bdrewery, dim
MFC after:	2 weeks
Differential Revision:	https://reviews.freebsd.org/D16574
2018-08-03 18:52:51 +00:00
Ed Maste
0d84986542 Makefile.libcompat: correctly append to make variable
Reported by:	hselasky
Pointy hat to:	emaste
2018-07-30 14:42:59 +00:00
Ed Maste
ad91146c66 Makefile.libcompat: use one variable assignment per line
This makes it easier to carry or merge patches that remove gnu/
components.
2018-07-30 14:28:00 +00:00
Alex Richardson
55a5b46aff Don't change directory owner to root when building with -DNO_ROOT
Currently the mtree calls in Makefile.inc1 all change the directory owner
to match the spec file. However, we should not be doing this during
distributeworld if -DNO_ROOT is passed. Additionally, when creating the
WORLDTMP directory hierachy there is no need to change the owner to root so
we now always pass the -W flag when populating WORLDTMP.

This is also required for building FreeBSD on Linux/Mac since the required
groups/users will not exist there which is how I discovered this issue.

Reviewed By:	emaste, bdrewery, imp
Approved By:	brooks (mentor)
Differential Revision: https://reviews.freebsd.org/D14185
2018-06-29 21:15:26 +00:00
John Baldwin
151578dc5f Remove the various build flag hacks for GCC cross-compile.
The xtoolchain GCC packages have not required these flags since ports
commits r465416 and r466701.  The in-tree GCC 4.2.1 has also been patched
in r335716 and r335717 to correctly honor --sysroot when looking for
includes and libraries.

Reviewed by:	bdrewery
Sponsored by:	DARPA / AFRL
Differential Revision:	https://reviews.freebsd.org/D16055
2018-06-28 21:26:14 +00:00
Bryan Drewery
6e6be2bbe2 Reduce exec and fstat overhead for non-build targets.
This is mostly targetting 'make showconfig' and 'make test-system-*' for
the benefit of makeman and universe-one-clang work.

Sponsored by:	Dell EMC
2018-06-20 17:20:39 +00:00
Bryan Drewery
c4273800ae lib32: Fix lib/libpmc/pmu-events files ending up in source directory.
This could happen with either WITHOUT_AUTO_OBJ=yes or MAKELEVEL>0 for
the initial 'make buildworld' command.

This now ensures that build-tools targets have 'make obj' ran if needed.

This is especially problematic for pmu-events since it is not directly
connected in the build.  Normally the 'make includes' call right before
this implicitly creates the objdir with a 'make obj' already but
misses pmu-events because it is disconnected from lib/libpmc.  Fixing that
would make this new 'make obj' pointless but it is being added to avoid
this problem in the future should another tool be connected like this.

Reported by:	rgrimes, kib, kevans
Sponsored by:	Dell EMC
2018-06-15 18:50:24 +00:00
Bryan Drewery
127b666cd1 libcompat: Only build libmagic build tool if MK_FILE is yes.
Sponsored by:	Dell EMC
2018-06-15 17:03:57 +00:00
Bryan Drewery
6f1f2b1931 Only build jevents if MK_PMC is yes and only for amd64 in libcompat.
Sponsored by:	Dell EMC
2018-06-15 16:48:09 +00:00
Matt Macy
7d1c2b74a0 libpmc/pmu: enable for i386 as well 2018-05-31 22:26:55 +00:00
Bryan Drewery
1c0cac23cf X_COMPILER_* may not be defined.
Sponsored by:	Dell EMC
2018-01-24 18:08:37 +00:00
John Baldwin
a1554bba05 Check for GCC first rather than clang in the MIPS lib32 rules.
This works around a bug with X_COMPILER_TYPE and permits mips64 to build
again with in-tree gcc as well as clang and external gcc.

Sponsored by:	DARPA / AFRL
2018-01-16 01:05:04 +00:00
John Baldwin
8d286bb992 Allow MIPS lib32 to build with clang.
Don't try to set -march as clang doesn't permit use of the O32 ABI
with newer CPU microarchitectures like 'mips3'.  In addition, clang
doesn't permit the O32 ABI with the default N64 target, so use an
explicit O32 -target for clang.

Sponsored by:	DARPA / AFRL
2018-01-11 19:24:44 +00:00
Bryan Drewery
996e475962 Fix nested MAKEOBJDIRPREFIX breaking various release/buildworld/toolchain targets.
This problem was caused by r325329 and r325350.

For the release(7) targets, some will run mm-mtree.sh which itself runs make
with a MAKEOBJDIRPREFIX.  The execution of that script leaks OBJROOT,
MAKEOBJDIR, and MAKELEVEL=1 in the environment.  This causes the mm-mtree makes
to not do some basic setup of OBJROOT and only use this special
MAKEOBJDIRPREFIX case which fails to empty out MAKEOBJDIRPREFIX for further
nested makes, such as a tree walk.  If that tree walk sets OBJROOT/OBJTOP such
as r325329 is doing, then the wrong OBJDIRs end up being used due to the
unemptied MAKEOBJDIRPREFIX being preferred over the proper MAKEOBJDIR.

Pointyhat to:	bdrewery
Sponsored by:	Dell EMC Isilon
2017-11-05 00:11:51 +00:00
Bryan Drewery
1031b94fc4 Set OBJROOT/OBJTOP as make arguments. Missed in r325329.
Sponsored by:	Dell EMC Isilon
2017-11-05 00:11:47 +00:00
Bryan Drewery
4dc89c4d51 Reduce MAKEOBJDIRPREFIX path spam by specifying a direct objdir to use.
Sponsored by:	Dell EMC Isilon
2017-11-02 18:08:36 +00:00
Bryan Drewery
dfa099890c Add option UNIFIED_OBJDIR, on by default, which moves the default build OBJDIR.
This changes the build OBJDIR from the older style of /usr/obj/<srcdir> for
native builds, and /usr/obj/<target>.<target_arch>/<srcdir> for cross builds to
a new simpler format of /usr/obj/<srcdir>/<target>.<target_arch>.  This
new format is used regardless of cross or native build.  It allows
easier management of multiple source tree object directories.

The UNIFIED_OBJDIR option will be removed and its feature made permanent
for the 12.0 release.

Relnotes:	yes (don't note UNIFIED_OBJDIR option since it will be removed)
Prior work:	D3711 D874
Reviewed by:	gjb, sjg
Discussed at:	https://lists.freebsd.org/pipermail/freebsd-arch/2016-May/017805.html
Discussed with:	emaste
Sponsored by:	Dell EMC Isilon
Differential Revision:	https://reviews.freebsd.org/D12840
2017-11-01 21:22:05 +00:00
Bryan Drewery
2aba4f29f9 lib32/libsoft: Store object files in a more consistent structure.
Having objects in world32 and a sysroot in lib32 was confusing and
inconsistent with the normal build.  Now objects are stored in
obj-lib32 (or obj-libsoft) and the sysroot (analagous to WORLDTMP)
is stored in obj-lib32/tmp.

Sponsored by:	Dell EMC Isilon
2017-10-29 01:21:39 +00:00
Bryan Drewery
67d2f51ad6 Move some LIBCOMPAT tmpdir logic setup to its own build target.
Its own build target was already handling mtree extractions
just as _worldtmp did, so the other cleaning of the
tmpdir makes sense here as well.

Sponsored by:	Dell EMC Isilon
2017-10-29 01:21:02 +00:00
Warner Losh
0b972ac92e Support armv7 builds for userland
Make armv7 as a new MACHINE_ARCH.

Copy all the places we do armv6 and add armv7 as basically an
alias. clang appears to generate code for armv7 by default. armv7 hard
float isn't supported by the the in-tree gcc, so it hasn't been
updated to have a new default.

Support armv7 as a new valid MACHINE_ARCH (and by extension
TARGET_ARCH).

Add armv7 to the universe build.

Differential Revision: https://reviews.freebsd.org/D12010
2017-10-05 23:01:33 +00:00
Enji Cooper
90d1f60e27 Adjust r322633 to only apply to libexec/rtld-elf, and not usr.bin/ldd,
when running build32/install32

This unbreaks installing usr.bin/ldd as ldd32 when NO_RTLD is defined.

MFC after:      1 week
MFC with:       r322633
2017-09-30 21:00:08 +00:00
Enji Cooper
4ff30397f5 Honor NO_RTLD for rtld-elf, similar to what's done in libexec/Makefile, with
libexec/rtld-elf/... for MK_{LIB32,LIBSOFT}.

MFC after:	1 week
2017-08-17 20:15:20 +00:00
Dimitry Andric
8422e09e55 Follow-up to r321684 (Don't use libc++ when cross-building for gcc
arches), and handle two more cases where libc++ includes could be
incorrectly enabled, in case the host compiler is clang 5.0.0, and the
target (cross) compiler is gcc 4.2.1.

Noted by:	bdrewery
MFC after:	3 days
X-MFC-With:	321684
2017-08-07 16:23:53 +00:00
Bryan Drewery
35fb812adf buildworld: Define SYSROOT to WORLDTMP.
This is to allow downstream Makefiles to know for sure they are building
against a sysroot rather than only depending on ${DESTDIR} or other
assumptions.

This also exports it into buildenv.

MFC after:	2 weeks
Sponsored by:	Dell EMC Isilon
2017-06-19 18:08:20 +00:00
Bryan Drewery
6d75a7a852 Support skipping 'make obj' tree-walk.
This is part of a larger effort for WITH_AUTO_OBJ and a WORLDFAST
option.

Sponsored by:	Dell EMC Isilon
2017-05-09 20:21:38 +00:00
Bryan Drewery
dbbb03c244 META_MODE: Fix build-tools still sometimes rebuilding during target build.
In a cross-build, the build-tools are native host binaries.  We do not
want to rebuild them when building for the target.  Bmake previously
did not support checking .NOMETA on an existing target, so .NOMETA_CMP
was used here.  However, .NOMETA_CMP still triggers meta mode conditions
if the number of commands or the command changes.  In r312467 the paths
to build ncurses files were modified and thus triggered meta mode to
rebuild the build tools (make_keys, make_hash) in ncurses during the
target build.  Bmake 20160604 committed in r301462 changed .NOMETA to
also skip meta mode logic for an existing .meta file as well, thus it
is now the proper fix here.

I explored moving the build-tools output to WORLDTMP/tools with
relatively good success, but have concerns that doing so would be
problematic for downstream vendors who use LOCAL_TOOL_DIRS and
expect the tools to be in current OBJDIR for the target.  It also
adds more complexity into finding the tools during target build
and handling of where they are for rescue/rescue and
mkcsmapper_static/mkesdb_static which should really not be connected in
build-tools anyway.

MFC after:	2 weeks
Reported by:	many
Sponsored by:	Dell EMC Isilon
2017-04-06 18:21:59 +00:00
Ed Maste
bdb691d449 Use cross-NM (XNM) in compat32 build
An attempt to build mips64 using external toolchain failed as it tried
to use the host amd64 nm.

MFC after:	1 month
Sponsored by:	The FreeBSD Foundation
2017-01-27 03:43:18 +00:00
John Baldwin
69b4d461be Enable /usr/lib32 for o32 binaries on mips64.
Build and install an o32 set of libraries on mips64 suitable for
running o32 binaries via COMPAT_FREEBSD32. Enable COMPAT_FREEBSD32 in
MALTA64.

Reviewed by:	jmallett, imp
Sponsored by:	DARPA / AFRL
Differential Revision:	https://reviews.freebsd.org/D9032
2017-01-06 23:30:54 +00:00
John Baldwin
452a310167 Set -m32 in MD LIB32CPUCFLAGS rather than MI LIB32CFLAGS.
Both amd64 and powerpc64 use -m32 to compile 32-bit binaries, but not
all platforms follow this convention.  Move the -m32 compile flag into
the per-architecture flags to accomodate other architectures.

Sponsored by:	DARPA / AFRL
2017-01-06 23:20:59 +00:00
Baptiste Daroussin
646a9763f5 Fix build of lib32 libraries for amd64 and powerpc64 with cross toolchain 2016-12-25 15:39:52 +00:00
Bryan Drewery
4d611b2b2d Fix in-tree GCC builds after r304681.
There were a few issues.
- In-tree GCC won't have X_COMPILER_TYPE defined but will have
  WANT_COMPILER_TYPE==gcc set from the SYSTEM_COMPILER logic that can
  be used.  Make the clang check specific to clang as well to ensure
  -target doesn't leak into a GCC build.
- When using a cross-compiler GCC (with a default sysroot or arch) and also
  passing --sysroot, it basically forgets all internal paths for
  libraries.  We've already worked around this quite a bit for
  the external toolchains.  Now for the in-tree bootstrap cross-compiler
  GCC, also pass in the needed -B${WORLDTMP}/usr/lib to find the crt
  object files, but also -isystem and -L to fix the paths.  This creates
  quite a spammy build log, but it is clear and still achieves the goals
  and stays consistent between internal and external build flags.
  Reducing the spam by using the '=' prefix feature will help and be
  done later.

MFC after:	3 days
X-MFC-With:	r304681
Reported by:	bz
Pointyhat to:	bdrewery
Sponsored by:	EMC / Isilon Storage Division
2016-08-23 19:29:37 +00:00
Bryan Drewery
7337126677 Don't disable binutils/elftoolchain bootstrapping with external compiler.
This was a regression from r300349.

Setting MK_CROSS_COMPILER=no forces the compiler bootstraping *and* the
binutils/elftoolchain bootstrapping to be disabled in share/mk/src.opts.mk.
The only intent with using an external compiler is to disable bootstrapping of
the compiler.  The binutils/elftoolchain bootstrapping must still occur unless
XAS is set.  This did not affect WITH_SYSTEM_COMPILER.

Now that setting an external compiler sets both MK_CLANG_BOOTSTRAP and
MK_GCC_BOOTSTRAP to no, and MK_CROSS_COMPILER does the same via
share/mk/src/opts.mk, remove redundant logic that checks for
MK_CROSS_COMPILER.  It will not always be true now that MK_CROSS_COMPILER==no
when an external compiler is used and --sysroot/-target is needed.

Reported by:	sbruno
Pointyhat to:	bdrewery
MFC after:	3 days
Sponsored by:	EMC / Isilon Storage Division
2016-07-24 18:50:11 +00:00
Bryan Drewery
48413367ef WITH_SYSTEM_COMPILER: Avoid using GCC4.8+ logic on native GCC4.2 platforms.
As the XXX notes, these should really be checking MK_GNUCXX since there is
already a version check in share/mk/src.opts.mk to disable it.  Fixing that
here is more complex though.  This could also be using X_COMPILER_FEATURES
but uses X_COMPILER_VERSION to keep in sync with the src.opts.mk logic.

Tested by:	andreast
Sponsored by:	EMC / Isilon Storage Division
Approved by:	re (gjb)
2016-06-29 01:50:30 +00:00
Bryan Drewery
0739409184 Fix native powerpc64 build of lib32 with in-tree GCC.
- This was broken by r300350 and r300885.
- Add some comments around the external GCC logic since it is spread out
  and in need of some cleanup.
- The problem was that X_COMPILER_TYPE is always defined from CC->XCC's
  default, so if /usr/bin/cc is GCC (as it is on native powerpc64) then
  X_COMPILER_TYPE was getting GCC and triggering the external logic in
  Makefile.libcompat.  It was intended to always provide -isystem with
  GCC since --sysroot is used into the lib32 sysroot which won't modify
  the header path without the -isystem.  The use of the libc++/std=c++11
  override was only intended to be used for external compilers though
  (more accurately GCC 4.8+ but that's a separate assumption to
  cleanup).  Apply the same logic from Makefile.inc1 to Makefile.libcompat
  to only add the libc++ override when needed for external compilers.

Pointyhat to:	bdrewery
Tested with:	native ppc64 (swills), universe, ppc64 xtoolchain,
		amd64 xtoolchain, sparc64 cross-build of ppc64 (host GCC 4.2)
Reported by:	andreast, swills
Approved by:	re (gjb)
Sponsored by:	EMC / Isilon Storage Division
2016-06-15 23:58:03 +00:00
Ed Maste
dfaedbb2c9 Use -L to specify compat32 library paths instead of -Y
-Y is an uncommon linker option that is rather similar to -L.  In
discussion with Peter it seems early amd64 development might have
required the -Y-specific behaviour, but it is no longer necessary.

Switch to -L which is more widely supported and much more commonly
used, to make it easier to link the FreeBSD base system with linkers
other than ld.bfd.

Submitted by:	Rafael Ávila de Espíndola
Differential Revision:	https://reviews.freebsd.org/D6681
2016-06-09 17:03:03 +00:00
Bryan Drewery
76c5029cdf Libcompat: Swap CXX/CFLAGS.
This is the same as done for the native build in r300770 to ensure that
the libc++ build reads from SYSROOT/usr/include/c++/v1 before reading
from SYSROOT/usr/include.
2016-05-29 06:20:36 +00:00
Bryan Drewery
f661dbee81 GCC External: Revert r300886, r300904, r300917, r300918
The fix in r300873 is mostly enough.  A fix for lib32 will be
committed.separately.
2016-05-29 06:20:15 +00:00
Bryan Drewery
83e0f7db61 Libcompat: Set build tools in environment rather than make overrides.
This allows the CXX hack in r300917 for external GCC to work for
the lib32 build.  It is also the same pattern as the native
build uses by adding the tools into CROSSENV for external
toolchain, rather than make overrides.

Sponsored by:	EMC / Isilon Storage Division
2016-05-29 01:45:15 +00:00
Bryan Drewery
ce00342bb2 Move external GCC compiler hacks to bsd.sys.mk.
This allows respecting -nostdinc, -nostdinc++ and -nostdlib before
making the decision to add in -isystem, etc.  The -isystem flags
are problematic for building lib/libc++ and lib/libcxxrt which wants
to only use its own headers.

More information the need of these flags can be found at
https://gcc.gnu.org/ml/gcc/2016-03/msg00219.html

This also reverts r300873.

Sponsored by:	EMC / Isilon Storage Division
2016-05-27 23:03:44 +00:00
Bryan Drewery
47873c16d3 Libcompat: Only pass -isystem =/usr/include for external GCC.
This is the same as the main build logic.  GCC with a cross-compiler
requires using -isystem to =/usr/include to get the search order
correct.

Reported by:	dim, asomers
Sponsored by:	EMC / Isilon Storage Division
2016-05-27 23:03:28 +00:00
Bryan Drewery
83f2907550 Set CPP from XCPP for the libcompat build.
Submitted by:	Mark Millard <markmi@dsl-only.net>
2016-04-15 18:32:05 +00:00
Bryan Drewery
d1dd034d07 META_MODE: Don't rebuild build-tools targets during normal build.
This avoids 'build command changed' due to CFLAGS/CC changes during the
normal build.  Without this the build-tools targets end up rebuilding
for the *target* rather than keeping the native versions built in
build-tools.

Sponsored by:	EMC / Isilon Storage Division
2016-04-14 21:06:10 +00:00
Bryan Drewery
91da76981b Remove leftover _LDSCRIPTROOT missed in r297270.
Sponsored by:	EMC / Isilon Storage Division
2016-04-05 21:12:03 +00:00
Bryan Drewery
ce6c8db5c6 LIB32+WITHOUT_TOOLCHAIN: Fix build by staging includes.
This is the same fix as r297281 for the normal build.

Sponsored by:	EMC / Isilon Storage Division
2016-03-31 17:27:14 +00:00
Bryan Drewery
497e80911e Remove the old depend (mkdep) code and make FAST_DEPEND the one true way.
Reviewed by:	emaste, hselasky (partial), brooks (brief)
Discussed on:	arch@
Sponsored by:	EMC / Isilon Storage Division
Differential Revision:	https://reviews.freebsd.org/D5742
2016-03-30 23:50:23 +00:00
Bryan Drewery
17e3ebb1ad Fix libcompat not handling some external toolchain flags.
- Use libc++ with GCC.
- Use CROSS_BINUTILS_PREFIX with -B (r280980 addressed this mostly already)

Sponsored by:	EMC / Isilon Storage Division
2016-03-25 19:12:44 +00:00
Bryan Drewery
1b337a349d Build libcompat (lib32) with a --sysroot pointing into its stage directory.
This overrides the cross-compiler's default sysroot to use the WORLD32's
sysroot for building the lib32 libraries.  Previously the cross-compiler
would default the sysroot to the 64bit WORLDTMP and -B/-L/-isystem flags
were used to build using the lib32 files.  This leads to multiple issues
discussed later.  Some extra headers are now needed to be staged since the
64bit WORLDTMP is not referenced at all for headers.  The 64bit WORLDTMP
is still used via PATH for build tools.  Overriding the default
target/arch is retained in the CC/CXX overrides.

This allows reverting the LDSCRIPT rewriting in installworld from r296921 and
r235122, thus allowing read-only objdirs to work for installing again.

This removes the need for _LDSCRIPTROOT.

This allows progressing the change to always use --sysroot for the build
rather than only relying on the cross-compiler's default sysroot.  The
work for that is in D3970 and needed to resolve WITHOUT_CROSS_COMPILER
not using a --sysroot [1].

PR:		196193 [1]
Sponsored by:	EMC / Isilon Storage Division
2016-03-25 19:12:38 +00:00
Bryan Drewery
429882e6aa LIBRARIES_ONLY should only be defined during install32.
r245561 added it to prevent extra files from being installed during
the install32 phase (to prevent duplicates in the meta log with -DNO_ROOT).
The flag should not be passed during build32 though since it may
prevent staging of includes during the 'make includes' phase on
library directories.

Sponsored by:	EMC / Isilon Storage Division
2016-03-25 19:12:13 +00:00