This will still build the compiler for the target but will not build the
bootstrap cross-compiler in the cross-tools phase. Other toolchain
bootstrapping, such as elftoolchan and binutils, currently still occurs.
This will utilize the default CC (cc, /usr/bin/cc) as an external compiler.
This is planned to be on-by-default eventually.
This will utilize the __FreeBSD_cc_version compiler macro defined in the
source tree and compare it to CC's version. If they match then the
cross-compiler is skipped. If [X]CC is an external compiler (absolute
path) or WITHOUT_CROSS_COMPILER is already set, then this logic is skipped.
If the expected bootstrap compiler type no longer matches the found CC
compiler type (clang vs gcc), then the logic is skipped. As an extra
safety check the version number is also compared from the compiler to
the tree version.
Clang:
The macro FREEBSD_CC_VERSION is defined in:
lib/clang/include/clang/Basic/Version.inc
For clang -target will be used if TARGET_ARCH != MACHINE_ARCH. This
is from the current external toolchain logic. There is currently an
assumption that the host compiler can build the TARGET_ARCH. This
will usually be the case since we don't conditionalize target arch
support in clang, but it will break when introducing new
architectures. This problem is mitigated by incrementing the version
when adding new architectures.
GCC:
The macro FBSD_CC_VER is defined in:
gnu/usr.bin/cc/cc_tools/freebsd-native.h
For GCC there is no simple -target support when TARGET_ARCH !=
MACHINE_ARCH. In this case the opportunistic skip is not done. If we
add proper support for this case in external toolchain logic then it
will be fine to enable.
This relies on the macros being incremented whenever any change occurs
to these compilers that warrant rebuilding files. It also should never
repeat earlier values.
Reviewed by: brooks, bapt, imp
Sponsored by: EMC / Isilon Storage Division
Differential Revision: https://reviews.freebsd.org/D6357
This is mostly a style change so that other code does not duplicate
it. The problem is when META_MODE needs to be disabled but it has
been overridden by .MAKE.MODE.
Sponsored by: EMC / Isilon Storage Division
architectures. There's no definition for it, nobody uses it
and it is unlikely to ever work. We can put it back when someone
demonstrates it running...
The new default for armv6 is hard float, so extend that default
to the external toolchain support.
side effect of replacing /etc/{master.,}passwd and /etc/group.
Submitted by: O. Hartmann
Pointyhat to: gjb (myself)
Sponsored by: The FreeBSD Foundation
The old (^/stable/9) default was yes, not no ("no" was the new default
introduced recently that broke POLA). Restore it to keep POLA like
glebius intended in r299077
MFC after: 3 days
X-MFC with: r299086
Pointyhat to: ngie (research before assuming and committing next time)
Sponsored by: EMC / Isilon Storage Division
Without this the incremental build was broken since .depend.* are not
generated with .MAKE.MODE=meta and .meta files were not created to
track dependencies. Typically meta mode does not create .meta files
when building with curdir==objdir but the kernel build is special.
Reported by: Nikolai Lifanov <lifanov@mail.lifanov.com>
Sponsored by: EMC / Isilon Storage Division
This also protects them from trying to create .meta files
with WITH_META_MODE.
Reported by: Nikolai Lifanov <lifanov@mail.lifanov.com>
Sponsored by: EMC / Isilon Storage Division
created to avoid creating ${.OBJDIR}.
The duplicate REVISION/BRANCH/VERSION evaluation will be addressed
separately.
Sponsored by: The FreeBSD Foundation
When using meta mode with filemon, the build is reliably incremental
safe. Bmake will use the meta files, along with filemon information,
to rebuild targets when their dependencies change, commands change,
or files they generate are missing.
Sponsored by: EMC / Isilon Storage Division
Some of the clang libraries build in this phase and the cross-tools
phase. Later in the cross-tools phase when they build with a default
TOOLS_PREFIX, they see a changed build command in meta mode due to
the changed DEFAULT_SYSROOT. This is avoided by passing along
TOOLS_PREFIX earlier.
Sponsored by: EMC / Isilon Storage Division
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
The change in r284345 moved the creation of openpam_static_modules.o to
lib/libpam/static_modules but never managed to get them into libpam.a.
Move this logic to lib/libpam/static_libpam and have it create a static
library for libpam.a The main lib/libpam/libpam will only create a
shared library. No redundancy in compilation or installation exists
in this solution.
This avoids requiring a pass with -D_NO_LIBPAM_SO_YET.
Sponsored by: EMC / Isilon Storage Division
cause bizarre packaging failures with non-default OBJDIR
and/or MAKEOBJDIRPREFIX and REPODIR set to a location not
within OBJDIR.
Reported by: many
Tested by: sef
Sponsored by: The FreeBSD Foundation
GCC does add <sysroot>/usr/lib to the library search path but it comes after
/usr/local/lib which can find ports libraries such as libedit.so. The
bad path comes in as /usr/local/lib/gcc/x86_64-portbld-freebsd11.0/5.3.0/../../../
which corresponds to <prefix>/lib.
This partially reverts r297271.
Pointyhat to: bdrewery
Sponsored by: EMC / Isilon Storage Division
In r296926 the -P <path> option was added to kbdcontrol, which enables
this change for a simplified compile-time default keymap build process.
PR: 193865
Reviewed by: Oliver Pinter
Tested by: Oliver Pinter
MFC After: 1 month
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D5708
The use of XCXXFLAGS is to assign it to CXX in CROSSENV.
XCFLAGS is also assigned here so there is no need to have
--syroot and -B flags again.
Sponsored by: EMC / Isilon Storage Division
Without this the default toolchain in /usr/bin/ would not use
WORLDTMP via --sysroot, and would lack --target if cross-building.
PR: 196193
Related: D3970
Sponsored by: EMC / Isilon Storage Division
- The -L WORLDTMP/usr/lib is not needed as GCC is already adding in
-L =/usr/lib internally with --sysroot. It does not do this for
header include paths though, thus passing -isystem =/usr/include is
still needed.
For the forced libc++ usage:
- Use -isystem rather than -I for libc++ headers.
- Use -std=c++11 rather than gnu++11.
- Use -nostdinc++ to ensure GCC's headers don't leak in.
Sponsored by: EMC / Isilon Storage Division
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
This should fix the 'packages' target when MAKEOBJDIRPREFIX or
DESTDIR is set in the make(1) environment or via command line.
Reported by: kmoore, sef, Marko Turk
Sponsored by: The FreeBSD Foundation
for the armv6 ABI switch. This also make WITH_LIBSOFT functional on
the arm platform. As a transition thing, this seems to work even
without switching the ABI (we basically build the same libraries
twice when MK_LIBSOFT=yes until the ABI cut over next
month). MK_LIBSOFT remains default no.
This is in preparation for LIBSOFT.
This file only supports *1* LIBCOMPAT value currently and must be capitalized.
In Makefile.libcompat given LIBCOMPAT=FOO there can be values set for
LIBFOOCFLAGS, LIBFOOCPUFLAGS, LIBFOOWMAKEENV, LIBFOOWMAKEFLAGS, LIBFOOCPUFLAGS,
and LIBFOODTRACE. These will have the standard cross-build values appended
onto them.
This could be extended to support multiple libcompat libraries in the future
once there is a need.
Reviewed by: imp
Sponsored by: EMC / Isilon Storage Division
Differential Revision: https://reviews.freebsd.org/D5612
similar Makefile.libsoft which will do the same for armv6 soft fp API
libraries in prep for pulling the trigger on moving to armv6 hard
float. Once there's two files, I'll work with bdrewery@ to merge the
two files as they are mostly the same. The high rate of churn for
Makefile* makes it quite difficult to make progress out of tree.
Differential Review: https://reviews.freebsd.org/D5566
Remove building of the legacy makewhatis(1) since it was only needed for
6.0 upgrades.
On my 2.2 GHz system libsqlite3 takes 60-100 seconds to build, which due to
its serialized nature can hold up the build waiting on it to finish in
bootstrap-tools.
makewhatis(1) was only required to be a build tool to support upgrades
from 6.0 which was properly removed in r208324 but later reverted due to
installworld using it. The installworld issue was fixed in r275622
when it was added to ITOOLS. A BOOTSTRAPPING check was missed when
makewhatis(1) was replaced with mandoc in r283777.
Sponsored by: EMC / Isilon Storage Division
When not using NO_ROOT the DESTDIR is implicitly created by the
distrib-dirs call. However with NO_ROOT it is trying to write
to the METALOG right away before running distrib-dirs which
may fail.
Sponsored by: EMC / Isilon Storage Division
shared library tracking.
Note, this requires a patched pkg(8) to do anything, but pkg(8)
does not complain when an option is specified that it does not
recognize.
Sponsored by: The FreeBSD Foundation
After calling the cap_init(3) function Casper will fork from it's original
process, using pdfork(2). Forking from a process has a lot of advantages:
1. We have the same cwd as the original process.
2. The same uid, gid and groups.
3. The same MAC labels.
4. The same descriptor table.
5. The same routing table.
6. The same umask.
7. The same cpuset(1).
From now services are also in form of libraries.
We also removed libcapsicum at all and converts existing program using Casper
to new architecture.
Discussed with: pjd, jonathan, ed, drysdale@google.com, emaste
Partially reviewed by: drysdale@google.com, bdrewery
Approved by: pjd (mentor)
Differential Revision: https://reviews.freebsd.org/D4277
architectures we do not provide upstream pkg(8) packages.
This is not tied to anything as-is, and likely will break
your system if used (based on experience with testing with
powerpc).
There is an overwhelming amount of evil happening here,
so until the issues are fixed, it will not be tied into the
'packages' target.
Sponsored by: The FreeBSD Foundation
While functionally expected to be a no-op on big-iron hardware,
embedded hardware (arm, mips) do not have a GENERIC kernel, so
the KERNCONF value must be included in the package to avoid
conflicting packages for the default kernel (RPI-B versus RPI2,
for example).
While here, correct the kernel name in the metadata.
Sponsored by: The FreeBSD Foundation
Replace `make regress` (legacy test make target) and `make test` (incomplete
test make target added with the FreeBSD test suite) with make check as it's
consistent with other open source projects.
`make check` defaults to running tests from `.OBJDIR`, but can be overridden
with the `CHECKDIR` variable.
Add `make checkworld` target to simplify running the FreeBSD test suite from
`TESTSBASE` (i.e. the top-level tests directory), similar to buildworld.
Document `make check` and `make checkworld` in build(7).
Other minor changes:
- Rename intermediate file (`Kyuafile.auto`) to `Kyuafile` to simplify
`make check`.
- Remove terse warnings attached to `beforetest`/`aftertest`.
- Add kyua binary check to check target in suite.test.mk; error out if it's
not found
The MFC is [partly] contingent on other build related changes being MFCed.
Differential Revision: https://reviews.freebsd.org/D4406
MFC after: 2 months
X-MFC to: stable/10
Relnotes: yes
Reviewed by: bdrewery, Evan Cramer <eccramer@gmail.com>
Sponsored by: EMC / Isilon Storage Division
One of the major pain points with how this was implemented
is the requirement of in-tree, hard-coded <name>.ucl, as
well as <name>-<suffix>.ucl where <suffix> can be lib32,
profile, development, debug, or any combination of the four.
This created significant overhead when adding new packages
and any of the files in any of the combinations were missing.
Instead of test(1)-ing if the <packagename>.ucl file exists,
hand off to a script to figure out what the final ucl file
name should be before invoking pkg(8).
The default behavior is 'template.ucl' is used as a fallback.
This affects only the userland packages, as the kernel code
is already smart enough to handle these variations.
Sponsored by: The FreeBSD Foundation
Only 'installworld' needs to be protected and only when not using
-DNO_ROOT, which implies not installing to / and not needing the
lib dependency protections.
Sponsored by: EMC / Isilon Storage Division
and kernel are staged before invoking the 'create-<foo>-packages'
targets.
Include PKG_VERSION value in the 'create-{world,kernel}-packages'
targets so the value is not redefined when packaging the kernel,
which otherwise results in inconsistent and confusing package
version results.
Sponsored by: The FreeBSD Foundation
with -DNO_ROOT to create the METALOG mtree(8) file.
Separate the default STAGEDIR for world (WSTAGEDIR) and kernel
(KSTAGEDIR).
Fix the 'create-kernel-packages' target to work properly.
Evaluate if 'kernel' is set when invoking mtree-to-plist.awk,
which splits the kernel and kernel.debug into separate plist
files.
Fix METALOG creation when building/packaging multiple kernels.
Sponsored by: The FreeBSD Foundation
to pkg(8) commands.
Move the resulting packages outside STAGEDIR to minimize
pollution.
When invoking 'pkg create', include the ABI in the REPODIR
path so the correct hierarchy is created for cross-builds.
Move the STAGEDIR and REPODIR declarations above the targets
that use them to keep things cleaner, and move the packages
target.
Include '-o ABIFILE=DESTDIR/bin/sh' in pkg(8) invocations in
the create-kernel-packages target.
Sponsored by: The FreeBSD Foundation
r288829 states that lex requires the latest m4, but was not always building it.
Move lex to the same logic as m4 since they are closely tied now.
MFC after: 3 days
Sponsored by: EMC / Isilon Storage Division
Reported by: Slawa Olhovchenkov <slw@zxy.spb.ru>
- Rework MANIFEST generation and parsing via bsdinstall(8).
- Allow selecting debugging distribution sets during install.
- Rework bsdinstall(8) to fetch remote debug distribution sets
when they are not available on the local install medium.
- Allow selecting additional non-GENERIC kernels during install.
At present, GENERIC is still required, and installed by default.
Tested with: head@r293203
Sponsored by: The FreeBSD Foundation
not set. The 'stageworld' target is invoked with -DNO_ROOT, so the
metalog file(s) will be created regardless.
This matches the behavior of 'create-world-packages'.
Sponsored by: The FreeBSD Foundation
This retains the original behavior of release-related targets, which
assume 'buildworld' and 'buildkernel' have already happened.
Sponsored by: The FreeBSD Foundation
Instead of using which(1) to look for doxygen, look for it in <LOCALBASE>/bin .
$PATH gets mangled by make buildenv, etc so it's better to just be explicit
about the path if someone uses that for instance.
Differential Revision: https://reviews.freebsd.org/D4406 (part of a larger diff)
Reviewed by: emaste, Evan Cramer <eccramer@gmail.com>
Sponsored by: EMC / Isilon Storage Division
It does not properly import PATH; the PATH is reset by included profile
files on startup which breaks the biggest feature of buildenv (using
sysrooted cc from WORLDTMP)
Spotted by: smh, kib
Sponsored by: EMC / Isilon Storage Division
RISC-V is a new ISA designed to support computer research and education, and
is now become a standard open architecture for industry implementations.
This is a minimal set of changes required to run 'make kernel-toolchain'
using external (GNU) toolchain.
The FreeBSD/RISC-V project home: https://wiki.freebsd.org/riscv.
Reviewed by: andrew, bdrewery, emaste, imp
Sponsored by: DARPA, AFRL
Sponsored by: HEIF5
Differential Revision: https://reviews.freebsd.org/D4445
- Don't bother looking up REVISION/BRANCH/etc from release/, or the
CPUTYPE check, as these are not used for makeman and wastes time. The also
invokes auto.obj.mk after I reverted auto.obj.mk ignoring -V in r291312.
- Don't modify CC or PATH when WITH_CCACHE_BUILD or WITH_META_MODE is enabled
as it leads to bsd.compiler.mk errors.
Sponsored by: EMC / Isilon Storage Division
This is to fix 'make all' causing it to recurse on both 'all' and 'buildconfig'
due to 'buildconfig' being in ALL_SUBDIR_TARGETS and being a dependency of
'all'.
This now adds all of the '*includes', '*files' targets as subdir targets,
allowing them to recurse.
This also removes the need for some 'realinstall' hacks in bsd.subdir.mk since
it no longer recurses; only 'install' will recurse and call the proper
'beforeinstall', 'realinstall', and 'afterinstall' in each sub-directory.
This fixes 'make includes' and 'make files' to not be a rerolled ${MAKE}
sub-shell but to rather just recurse on 'inclues' and 'files'. This avoids
various issues such as the one fixed in r289462. As such revert Makefile.inc1
back to using 'includes' which avoids an extra tree walk and parallelizes
the includes phases better.
Makefile.inc1 includes a guard so that 'make all' will not use SUBDIR_PARALLEL,
added in r289438. This is so users do not get a probably broken build if they
run 'make all' from the top-level. Before the change in this commit, the
workaround for 'make everything' was 'par-all' which would depend on 'all' and
cause a proper parallel recursion. Now that will not work so a new
_PARALLEL_SUBUDIR_OK is used to allow it.
This is still part of an effort to combine bsd.(files|incs|confs).mk and move
some of its logic out of bsd.subdir.mk, as attempted in r289282 and reverted in
r289331. This commit fixes the problems found there which was mostly double
recursing during 'includes' which would recurse on itself and 'buildincludes'
and 'installincludes', all in parallel. The logic is still in bsd.subdir.mk
for now.
I've been cautious about this commit but have experienced no breakage on the
tree except for the 'par-all' case which was already a hack. If something foo
is depending on something bar that should recurse, it is very likely that the
foo target is being recursed on already meaning that bar will still effectively
recurse once sub-directories call foo.
Discussed on: arch@
MFC after: never
Sponsored by: EMC / Isilon Storage Division
installed as "kernel". This is relevant for packaging of the kernel when
not wanting a default "kernel.txz".
Submitted by: Russell Cattelan <cattelan@thebarn.com>
MFC after: 2 weeks
Obtained from: OneFS
Sponsored by: EMC / Isilon Storage Division
Say it with me, "I will not chain commands with && in Makefiles"
This was originally fixed and explained quite well by bde@ in r36074. The
initial bmake integration caused 'set -e' to stop being used which lead to
r252419. Later 'set -e' expectations were fixed with bmake in r254980.
Because of the && here, errors would be ignored when building in parallel and
a dependency failed. Such as bootstrap-tools since it builds everything in
parallel. If any tool failed in obj/depend/all, it would just ignore the error
and continue to build. This later would result in cascaded errors that only
confused the real issue. This could also cause commands after the failed
command to still execute, leading to more confusion.
This should be fine if the command is in a sub-shell such as: (cmd1 && cmd2)
This reverts r252419.
MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division
- Add a kvaddr_type to represent kernel virtual addresses instead of
unsigned long.
- Add a struct kvm_nlist which is a stripped down version of struct nlist
that uses kvaddr_t for n_value.
- Add a kvm_native() routine that returns true if an open kvm descriptor
is for a native kernel and memory image.
- Add a kvm_open2() function similar to kvm_openfiles(). It drops the
unused 'swapfile' argument and adds a new function pointer argument for
a symbol resolving function. Native kernels still use _fdnlist() from
libc to resolve symbols if a resolver function is not supplied, but cross
kernels require a resolver.
- Add a kvm_nlist2() function similar to kvm_nlist() except that it uses
struct kvm_nlist instead of struct nlist.
- Add a kvm_read2() function similar to kvm_read() except that it uses
kvaddr_t instead of unsigned long for the kernel virtual address.
- Add a new kvm_arch switch of routines needed by a vmcore backend.
Each backend is responsible for implementing kvm_read2() for a given
vmcore format.
- Use libelf to read headers from ELF kernels and cores (except for
powerpc cores).
- Add internal helper routines for the common page offset hash table used
by the minidump backends.
- Port all of the existing kvm backends to implement a kvm_arch switch and
to be cross-friendly by using private constants instead of ones that
vary by platform (e.g. PAGE_SIZE). Static assertions are present when
a given backend is compiled natively to ensure the private constants
match the real ones.
- Enable all of the existing vmcore backends on all platforms. This means
that libkvm on any platform should be able to perform KVA translation
and read data from a vmcore of any platform.
Tested on: amd64, i386, sparc64 (marius)
Differential Revision: https://reviews.freebsd.org/D3341
The bootstrap-tools are supposed to be host tools, which in most cases, use
host headers and libraries. As such, directly including the src tree's headers
for libmd here causes the need to link libmd in since it will be built with
the new symbols (which /usr/lib/libmd.so) won't have unless it is new enough.
During the target build in buildworld the target headers are staged into
WORLDTMP and used via --sysroot, allowing the target xinstall to be built with
the new/target libmd.
The .PATH here was also not doing anything since xinstall does not use libmd
source files.
Sponsored by: EMC / Isilon Storage Division
MFC after: 2 weeks
This simplifies the logic to always try removing the objdir if it exists
and to fallback on a 'cleandir' if no objdir exists. The reasoning for
this is to avoid rm -rf src/* (r126024)
Sponsored by: EMC / Isilon Storage Division
MFC after: 2 weeks
This leads the way for fixing cross-build cleanup, and eventually replacing
'cleandir' with it during the build.
Sponsored by: EMC / Isilon Storage Division
MFC after: 2 weeks
ccache is mostly beneficial for frequent builds where -DNO_CLEAN is not
used to achieve a safe pseudo-incremental build. This is explained in
more detail upstream [1] [2]. It incurs about a 20%-28% hit to populate the
cache, but with a full cache saves 30-50% in build times. When combined with
the WITH_FAST_DEPEND feature it saves up to 65% since ccache does cache the
resulting dependency file, which it does not do when using mkdep(1)/'CC
-E'. Stats are provided at the end of this message.
This removes the need to modify /etc/make.conf with the CC:= and CXX:=
lines which conflicted with external compiler support [3] (causing the
bootstrap compiler to not be built which lead to obscure failures [4]),
incorrectly invoked ccache in various stages, required CCACHE_CPP2 to avoid
Clang errors with parenthesis, and did not work with META_MODE.
The option name was picked to match the existing option in ports. This
feature is available for both in-src and out-of-src builds that use
/usr/share/mk.
Linking, assembly compiles, and pre-processing avoid using ccache since it is
only overhead. ccache does nothing special in these modes, although there is
no harm in calling it for them.
CCACHE_COMPILERCHECK is set to 'content' when using the in-tree bootstrap
compiler to hash the content of the compiler binary to determine if it
should be a cache miss. For external compilers the 'mtime' option is used
as it is more efficient and likely to be correct. Future work may optimize the
'content' check using the same checks as whether a bootstrap compiler is needed
to be built.
The CCACHE_CPP2 pessimization is currently default in our devel/ccache
port due to Clang requiring it. Clang's -Wparentheses-equality,
-Wtautological-compare, and -Wself-assign warnings do not mix well with
compiling already-pre-processed code that may have expanded macros that
trigger the warnings. GCC has so far not had this issue so it is allowed to
disable the CCACHE_CPP2 default in our port.
Sharing a cache between multiple checkouts, or systems, is explained in
the ccache manual. Sharing a cache over NFS would likely not be worth
it, but syncing cache directories between systems may be useful for an
organization. There is also a memcached backend available [5]. Due to using
an object directory outside of the source directory though you will need to
ensure that both are in the same prefix and all users use the same layout. A
possible working layout is as follows:
Source: /some/prefix/src1
Source: /some/prefix/src2
Source: /some/prefix/src3
Objdir: /some/prefix/obj
Environment: CCACHE_BASEDIR='${SRCTOP:H}' MAKEOBJDIRPREFIX='${SRCTOP:H}/obj'
This will use src*/../obj as the MAKEOBJDIRPREFIX and tells ccache to replace
all absolute paths to be relative. Using something like this is required due
to -I and -o flags containing both SRC and OBJDIR absolute paths that ccache
adds into its hash for the object without CCACHE_BASEDIR.
distcc can be hooked into by setting CCACHE_PREFIX=/usr/local/bin/distcc.
I have not personally tested this and assume it will not mix well with
using the bootstrap compiler.
The cache from buildworld can be reused in a subdir by first running
'make buildenv' (from r290424).
Note that the cache is currently different depending on whether -j is
used or not due to ccache enabling -fdiagnostics-color automatically if
stderr is a TTY, which bmake only does if not using -j.
The system I used for testing was:
WITNESS
Build options: -j20 WITH_LLDB=yes WITH_DEBUG_FILES=yes WITH_CCACHE_BUILD=yes
DISK: ZFS 3-way mirror with very slow disks using SSD l2arc/log.
The arc was fully populated with src tree files and ccache objects.
RAM: 76GiB
CPU: Intel(R) Xeon(R) CPU L5520 @2.27GHz
2 package(s) x 4 core(s) x 2 SMT threads = hw.ncpu=16
The WITH_FAST_DEPEND feature was used for comparison here as well to show
the dramatic time savings with a full cache.
buildworld:
x buildworld-before
+ buildworld-ccache-empty
* buildworld-ccache-full
% buildworld-ccache-full-fastdep
# buildworld-fastdep
+-------------------------------------------------------------------------------+
|% * # +|
|% * # +|
|% * # xxx +|
| |A |
| A|
| A |
|A |
| A |
+-------------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 3 3744.13 3794.31 3752.25 3763.5633 26.935139
+ 3 4519 4525.04 4520.73 4521.59 3.1104823
Difference at 95.0% confidence
758.027 +/- 43.4565
20.1412% +/- 1.15466%
(Student's t, pooled s = 19.1726)
* 3 1823.08 1827.2 1825.62 1825.3 2.0785572
Difference at 95.0% confidence
-1938.26 +/- 43.298
-51.5007% +/- 1.15045%
(Student's t, pooled s = 19.1026)
% 3 1266.96 1279.37 1270.47 1272.2667 6.3971113
Difference at 95.0% confidence
-2491.3 +/- 44.3704
-66.1952% +/- 1.17895%
(Student's t, pooled s = 19.5758)
# 3 3153.34 3155.16 3154.2 3154.2333 0.91045776
Difference at 95.0% confidence
-609.33 +/- 43.1943
-16.1902% +/- 1.1477%
(Student's t, pooled s = 19.0569)
buildkernel:
x buildkernel-before
+ buildkernel-ccache-empty
* buildkernel-ccache-empty-fastdep
% buildkernel-ccache-full
# buildkernel-ccache-full-fastdep
@ buildkernel-fastdep
+-------------------------------------------------------------------------------+
|# @ % * |
|# @ % * x + |
|# @ % * xx ++|
| MA |
| MA|
| A |
| A |
|A |
| A |
+-------------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 3 571.57 573.94 571.79 572.43333 1.3094401
+ 3 727.97 731.91 728.06 729.31333 2.2492295
Difference at 95.0% confidence
156.88 +/- 4.17129
27.4058% +/- 0.728695%
(Student's t, pooled s = 1.84034)
* 3 527.1 528.29 528.08 527.82333 0.63516402
Difference at 95.0% confidence
-44.61 +/- 2.33254
-7.79305% +/- 0.407478%
(Student's t, pooled s = 1.02909)
% 3 400.4 401.05 400.62 400.69 0.3306055
Difference at 95.0% confidence
-171.743 +/- 2.16453
-30.0023% +/- 0.378128%
(Student's t, pooled s = 0.954969)
# 3 201.94 203.34 202.28 202.52 0.73020545
Difference at 95.0% confidence
-369.913 +/- 2.40293
-64.6212% +/- 0.419774%
(Student's t, pooled s = 1.06015)
@ 3 369.12 370.57 369.3 369.66333 0.79033748
Difference at 95.0% confidence
-202.77 +/- 2.45131
-35.4225% +/- 0.428227%
(Student's t, pooled s = 1.0815)
[1] https://ccache.samba.org/performance.html
[2] http://www.mail-archive.com/ccache@lists.samba.org/msg00576.html
[3] https://reviews.freebsd.org/D3484
[5] https://github.com/jrosdahl/ccache/pull/30
PR: 182944 [4]
MFC after: 3 weeks
Sponsored by: EMC / Isilon Storage Division
Relnotes: yes
ZSH considers CPUTYPE a magic variable that will be the output of 'uname -m'
even if already set in environment when starting up.
The CPUTYPE?= check in Makefile.inc1 and supporting overriding CPUTYPE
manually in the buildenv shell make automatic workarounds too tricky
here. ZSH should really respect variables set in the environment before
trashing them.
X-MFC-With: r290423
MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division
Also pass BUILDENV=1 into the sub-shell to allow modifying PS1 in .profile such
as:
if [ -n "${BUILDENV}" ]; then
PS1="(buildenv) ${PS1}"
fi
SHELL defaults to 'sh' in share/mk/sys.mk, but is typically passed down by
the shell invoking make as well. Rather than forcing all 'buildenv' users
to use plain /bin/sh, let them use their favorite shell.
MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division
Discussed with: imp
Move CROSS_TOOLS stuff to top of file (before bsd.compiler.mk) so that
decisions made by bsd.compiler.mk can properly affect the defaults in
src.opts.mk. Move that to after bsd.compiler.mk. Add a comment about
why we include bsd.compiler.mk here despite the fact that src.opts.mk
currently does too. Also remove bsd.arch.inc.mk that's been OBE.
Differential Revision: https://reviews.freebsd.org/D4087
libopenbsd is an internal library which
to bring in compatibility stuff from OpenBSD.
This will allow us to bring in more
OpenBSD utilities into the FreeBSD base system.
We similarly use libnetbsd for bringing in stuff from NetBSD.
Reviewed by: bapt
Differential Revision: https://reviews.freebsd.org/D4078
CXXFLAGS to sub-makes.
The bad passing also causes bsd.dep.mk's logic to selectively pull only some
flags from C[XX]FLAGS to not apply which can be seen with '-L' being passed to
mkdep when using an external compiler.
Now it can be used to effectively "build in a subdir". It will use the
'cross-tools', 'libraries', and 'includes' phases of 'buildworld' to properly
setup a WORLDTMP to use. Then it will build 'everything' only in the
listed SUBDIR_OVERRIDE directories. It is still required to list custom
library directories in LOCAL_LIB_DIRS if SUBDIR_OVERRIDE is something
that contains libraries outside of the normal area (such as
SUBDIR_OVERRIDE=contrib/ofed needing LOCAL_LIB_DIRS=contrib/ofed/usr.lib)
Without these changes, SUBDIR_OVERRIDE with buildworld was broken or hit
obscure failures due to missing libraries, includes, or cross compiler.
SUBDIR_OVERRIDE with 'make <target that is not buildworld>' will continue to
work as it did before although its usefulness is questionable.
With a fully populated WORLDTMP, building with a SUBDIR_OVERRIDE with
-DNO_CLEAN only takes a few minutes to start building the target
directories. This is still much better than building unneeded things via
'everything' when testing small subset changes. A BUILDFAST or
SKIPWORLDTMP might make sense for this as well.
- Add in '_worldtmp' as we still need to create WORLDTMP as later targets,
such as '_libraries' and '_includes' use it. This probably was avoiding
calling '_worldtmp' to not remove WORLDTMP for debugging purposes, but
-DNO_CLEAN can be used for that.
- '_legacy' must be included since '_build-tools' uses -legacy.
The SUBDIR_OVERRIDE change came in r95509, while -legacy being part
of build-tools came in r113136.
- 'bootstrap-tools' is still skipped as this feature is not for
upgrades.
- Fix buildworld combined with SUBDIR_OVERRIDE not installing all includes.
The original change for SUBDIR_OVERRIDE in r95509 kept '_includes'
and '_libraries' as building everything possible as the SUBDIR_OVERRIDE
could need anything from them. However in r96462 the real 'includes'
target was changed from manual sub-makes to just recursing 'includes'
on SUBDIR, thus not all includes have been installed into WORLDTMP since then
when combined with 'buildworld'.
This is not done unless calling 'make buildworld' as it would be
unexpected to have it go into all directories when doing 'make
SUBDIR_OVERRIDE=mydir includes'.
- Also need to build the cross-compiler so it is used with --sysroot.
If this is burdensome then telling the build to use the local compiler
as an external compiler (thus using a proper --sysroot to WORLDTMP) is
possible by setting CC=/usr/bin/cc, CXX=/usr/bin/c++, etc.
- Don't build the lib32 distribution with SUBDIR_OVERRIDE in buildworld
since it won't contain anything related to SUBDIR_OVERRIDE. Testing
of the lib32 build can be done with 'make build32'.
- Document these changes in build.7
Sponsored by: EMC / Isilon Storage Division
MFC after: 2 weeks
Extend OptionalObsoleteFiles.inc to delete all lib32 files when MK_LIB32 is
set to no on a system that previously had lib32 libraries installed.
Also, to prevent "make delete-old-dirs" from always deleting lib32 directories
after an installworld, move the lib32 subtree to its own mtree file that only
gets applied when MK_LIB32=yes.
Test: Ran "make delete-old" and "make delete-old-libs" on a system that never
had MK_LIB32 enabled, and on a system where MK_LIB32 was enabled and later
disabled. Did this both on amd64 and powerpc64.
Test: Ran "make tinderbox" without errors.
Reviewed by: emaste
Differential Revision: https://reviews.freebsd.org/D3923
This has been handled since r228158 made MK_CTF dependent on MK_CDDL
in share/mk/bsd.opts.mk.
MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division
The 'includes' target is currently a pseudo target in bsd.subdir.mk that
does 'cd ${.CURDIR} && ${MAKE} buildincludes && ${MAKE} installincludes',
versus all over targets that just recurse.
In Makefile.inc1 the older duplicated bsd.subdir.mk logic for calling
'includes' was being executed in each subdir directly, meaning 'cd lib && make
includes' became 'cd lib && make buildincludes && make installincludes'. Now
that the bsd.subdir.mk logic is used it is calling 'make buildincludes && make
installincludes' from the top-level which pulls in the PATH=<default path>
from /Makefile.
The sub-make logic for 'includes' in bsd.subdir.mk was attempted to be removed
in r289282 but turned out to be wrong. I have a working version now but
it is not yet ready for commit. So for now in Makefile.inc1 split out
'includes' to 'buildincludes' and 'installincludes' which will avoid the
problem.
MFC after: 2 weeks
X-MFC-With: r289438
Sponsored by: EMC / Isilon Storage Division
sub-makes.
Some of the world phases that used plain '${MAKE} -f Makefile.inc1' were not
passing this variable along which caused them to look it up again. By
using bmake's .export we can remove it from all of the other environment
lines.
Add a comment about the usage for VERSION for ctfmerge.
Sponsored by: EMC / Isilon Storage Division
We pass BOOTSTRAPPING=${OSRELDATE} to some of the sub-makes. Rather than
chase every ${MAKE} invokation, just export it as bmake lets us.
Sponsored by: EMC / Isilon Storage Division
Back in r30113, the 'par-*' targets were added to parallelize portions of
the build in a very similar fashion as the SUBDIR_PARALLEL feature used in
r263778. Calling a target without 'par-' (for 'parallel') resulted in the
standard bsd.subdir.mk handling without parallelization. Given we have
SUBDIR_PARALLEL now there is no reason to duplicate the handling here.
In build logs this will result in the ${dir}.${target}__D targets now showing
as the normal ${target}_subdir_${dir} targets.
I audited all of the uses of Makefile.inc1 and Makefile's targets that use
bsd.subdir.mk and found that all but 'all' and 'install' were fine to use
as always parallel.
- For 'install' (from installworld -j) the ordering of lib/ and libexec/
before the rest of the system (described in r289433), and etc/ being last
(described in r289435), is all that matters. So now a .WAIT is added in
the proper places when invoking any 'install*' target. A parallel
installworld does work and took 46% of the time a non-parallel
install would take on my system with -j15 to ZFS.
- For 'all' I left the default handling for this to not run in parallel. A
'par-all' target is still used by the 'everything' stage of buildworld
to continue building in parallel as it already has been. This works
because most of the dependencies are handled by the early bootstrap
phases as well as 'libraries' and 'includes' phases. This lets
all of the SUBDIR build in parallel fine, such as bin/ and lib/. This
will not work if the user invokes 'all' though as we have dependencies
spread all over the system with no way to depend between them (except
for the dirdeps feature in the META_MODE build). Calling 'make all'
from the top-level is still useful at least when using SUBDIR_OVERRIDE.
MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division
The ordering of 'etc' in the install has a long history dating back to the
first time it was realized it needed to be "last" in r4486. That commit
still left it before LOCAL_DIRS though. By having it before LOCAL_DIRS
any manpages they install were not being added to the whatis database in the
install image. They would likely show up in the file after a periodic
rebuild of the file though.
Currently the whatis file is built by an 'afterinstall' hook in etc/Makefile
that calls share/man's 'makedb' target.
MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division
It was added in r152006 to handle serializing access of info/dir when
installing INFO files. We no longer support INFO files since r276551
though.
Sponsored by: EMC / Isilon Storage Division
The case of make(1) using a new /bin/sh issue was fixed in r173219 when ITOOLS
was introduced.
There are still issues with mid-install errors leaving a system unusable that
are currently non-trivial to solve. The safest ordering requires installing
rtld, libc and libthr (in that order) before anything else. We don't do that
now though. Much improvement is needed here still.
Discussed with: kip and kan (rtld/library ordering)
MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division
This was handled for libraries in r256842 but for some reason was missed
for files (bsd.prog.mk).
MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division
Relnotes: yes
This was causing files to be removed from the objdir when -n was used.
_worldtmp makes no sub-make calls.
A more comprehensive solution is coming involving fine-grained '+' where
appropriate.
Sponsored by: EMC / Isilon Storage Division
MFC after: 1 week
with the current behaviour of calling "distribution" in the etc target.
This allows mergemaster/etcupdate to still work when some configuration will be
moved to be handled in the same directories their source code lives in.
when running the build-tools stage.
The requirement is due to the -P flag used when running m4 from usr.bin/lex
Makefile to generate skel.c. With the old m4 that fails and the failure is
ignored, resulting in an empty(-ish) skel.c, which leads to later build
failures when the misconfigured new lex tool is run.
This enables building -current (and 10-stable after MFC) on a stable-8
system again.
MFC after: 3 days
Bootstrap tools exist for backwards compatibility support. DTrace tools
tools are also needed for cross builds, so belong in cross-tools.
Reviewed by: imp (earlier), markj
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D2144
Stale CTF tools are a frequent source of DTrace issues, and they compile
quickly enough that the increase in build time is negligible.
Reviewed by: emaste, imp
Differential Revision: https://reviews.freebsd.org/D3670
a one word variable, which is quite unexpected from documentation.
So, to avoid double installation of a single kernel, protect the extra
kernels loop with ${BUILDKERNELS:[#]} > 1 conditional.
Sponsored by: Netflix
Sponsored by: Nginx, Inc.
from BUILDKERNELS list. This is more strict, since INSTALLKERNEL by
definition is the first word of BUILDKERNELS list. The previous
code failed if INSTALLKERNEL is a substring of additional kernel name.
Reviewed by: gjb
Sponsored by: Netflix
Sponsored by: Nginx, Inc.
For most cases they are equivalent, but BINUTILS_BOOTSTRAP is a
BROKEN_OPTION on arm64 as the in-tree GNU binutils do not support it,
so we need a separate internal flag for ELF Tool Chain.
Reviewed by: andrew, brooks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D3381
