run. This makes the 'ctfconvert' and 'ctfmerge' programs show up during
a build when compiling a kernel, a library or a program.
Sponsored by: The FreeBSD Foundation
are some problems with static executables), make.conf (would also
affect ports which do not use GNU make and do not override the
compile targets) or in the kernel config (via "makeoptions
WITH_CTF=yes").
Additional (related) changes:
- propagate WITH_CTF to module builds
- do not add -g to the linker flags, it's a noop there anyway
(at least according to the man page of ld)
- do not add -g to CFLAGS unconditionally
we need to have a look if it is really needed (IMO not) or if there
is a way to add it only when WITH_CTF is used
Note: ctfconvert / ctfmerge lines will not appear in the build output,
to protect the innocent (those which do not build with WITH_CTF would
see the shell-test and may think WITH_CTF is used).
Reviewed by: imp, jhb, scottl (earlier version)
Discussed on: arch@
compiled with stack protector.
Use libssp_nonshared library to pull __stack_chk_fail_local symbol into
each library that needs it instead of pulling it from libc. GCC
generates local calls to this function which result in absolute
relocations put into position-independent code segment, making dynamic
loader do extra work every time given shared library is being relocated
and making affected text pages non-shareable.
Reviewed by: kib
Approved by: re (kib)
system makefiles.
Note that the CTF conversion defaults to off. We may choose to change this
default later if DTrace proves popular and people are prepared to wear
the compilation performance impact of compiling with debug symbols all the
time.
Setting NO_CTF in the make args or user environment turns off CTF conversion.
Even if we choose to default CTF generation to on later, we still need
NO_CTF so that the buildworld process can bootstrap the tools without
needlessly generating CTF data for temporary tools.
Setting WITH_CTF in the make args or user environment (and _NOT_ in
/etc/make.conf) is the only way to enable CTF data conversion. Nore that
this can't be implemented the same way that the WITH_ and WITHOUT_ stuff
is implemented throughout the buildworld because the CTF conversion needs
to work when building a simple object without a Makefile, using the
default rules in sys.mk.
Typing 'make test.o' with no makefile and just a source file test.c
should work. Also, typing 'make WITH_CTF=1 test.o without a makefile and
just a source file test.c should work and produce an object with a CTF
elf section. Typing 'make WITH_CTF=1 CFLAGS=-g test.o' without a makefile
and just a source file test.c should produce an object with both a CTF
elf section and the debug elf sections.
In the FreeBSD build where more .mk files are used than just sys.mk
which is included my make by default, the use of DEBUG_FLAGS is the
correct way to enable a debug build. The important thing to note here
is that it is the DEBUG_FLAGS setting that prevents libraries and
programs from being stripped on installation. So, for the addition of
CTF data conversion, setting DEBUG_FLAGS to contain -g, without NO_CTF,
will cause the ctfconvert and ctfmerge build programs to be executed
also with the -g arg so that debug symbols are retained rather than
being removed after the CTF data elf section has been added.
Add DTrace libraries to the list of libnames.
test MK_INSTALLLIB, users can set WITHOUT_INSTALLLIB. The old
NO_INSTALLLIB is still supported as several makefiles set it.
- While here, fix an install when instructed not to install libs
(usr.bin/lex/lib/Makefile).
PR: bin/114200
Submitted by: Henrik Brix Andersen
Warning, after symbol versioning is enabled, going back is not easy
(use WITHOUT_SYMVER at your own risk).
Change the default thread library to libthr.
There most likely still needs to be a version bump for at least the
thread libraries. If necessary, this will happen later.
asserts and makes binaries smaller. The binaries also become
repeatable again. As it was, without this md5's of binaries built
with different paths differed.
# Where do I document this?
override NM in Makefiles when, for example, cross compiling and have
that value be used by lorder. NM normally isn't defined, so we pass a
null value to lorder. lorder says 'NM=${NM-nm}' which causes it to
pickup the default value.
a set of symbol definitions (VERSION_DEF) and symbol mappings
(SYMBOL_MAPS).
Add an awk script to generate the version map.
Suggested by: ru
Helped by: ru
Reviewed by: ru
bsd.incs.mk, and use it when installing 32-bit compat libraries
on amd64. This causes it to *not* overwrite native headers with
i386 versions, which was the case with <fenv.h> and <vgl.h>.
PR: amd64/83806
Prodded by: bde
MFC after: 1 week
Fix 'broken' ifdefs.
icc does not support profiling yet so remove unfinished code which was
supposed to help.
Submitted by: netchild (original version)
Reviewed by: ru
Intel C/C++ compiler (lang/icc) to build the kernel.
The icc CPUTYPE CFLAGS use icc v7 syntax, icc v8 moans about them, but
doesn't abort. They also produce CPU specific code (new instructions
of the CPU, not only CPU specific scheduling), so if you get coredumps
with signal 4 (SIGILL, illegal instruction) you've used the wrong
CPUTYPE.
Incarnations of this patch survive gcc compiles and my make universe.
I use it on my desktop.
To use it update share/mk, add
/usr/local/intel/compiler70/ia32/bin (icc v7, works)
or
/usr/local/intel_cc_80/bin (icc v8, doesn't work)
to your PATH, make sure you have a new kernel compile directory
(e.g. MYKERNEL_icc) and run
CFLAGS="-O2 -ip" CC=icc make depend
CFLAGS="-O2 -ip" CC=icc make
in it.
Don't compile with -ipo, the build infrastructure uses ld directly to
link the kernel and the modules, but -ipo needs the link step to be
performed with Intel's linker.
Problems with icc v8:
- panic: npx0 cannot be emulated on an SMP system
- UP: first start of /bin/sh results in a FP exception
Parts of this commit contains suggestions or submissions from
Marius Strobl <marius@alchemy.franken.de>.
Reviewed by: silence on -arch
Submitted by: netchild
missing and there are multiple choices using multiple inference
(suffix transformation) rules.
This is known to fix compilation of s_log1p.o in lib/msun on i386,
as otherwise it attempted to use s_log1p.S as the source (which is
marked broken) instead of legal s_log1p.c which is in CFLAGS. The
normal case where .depend file exists is not affected.
Reviewed by: bde
as it was decided that our toolchain will revert to looking
for libraries in /usr/lib only.
- Make /usr/lib/libfoo.so -> /lib/libfoo.so.X symlinks absolute
so that they still work if /usr is symlinked.
- Remove stale /usr/lib/libfoo.so.X libraries during install.
Discussed with: gordon, obrien, peter
hose your system. You end up with just about everything statically linked
(except for libpam.so), which then causes all the pam users to fail.
eg: login, sshd, su etc all stop working because dlopen no longer works
because there is no libc.so in memory anymore.
gcc passes -L/usr/lib to ld. The /usr/lib/libxxx.so symlink is *not* a
compatability link. It is actually the primary link. There should be no
symlinks in /lib at all. Only /lib/libXX.so.Y.
peter@daintree[9:27pm]/usr/bin-104> file yppasswd
yppasswd: setuid ELF 32-bit LSB executable, Intel 80386, version 1 (FreeBSD), for FreeBSD 5.1.1, dynamically linked (uses shared libs), stripped
peter@daintree[9:27pm]/usr/bin-105> ldd yppasswd
yppasswd:
libpam.so.2 => /usr/lib/libpam.so.2 (0x280d1000)
peter@daintree[9:28pm]/usr/bin-106>
Note no libc.so.5. Hence libpam.so.2 has unresolved dependencies.
I believe this is also the cause of the recent buildworld failures when
pam_krb5.so references -lcrypto stuff etc and when librpcsvc.so references
des_setparity() etc.
This change could not possibly have worked, unless there are other missing
changes to the gcc configuration. It won't work with ports versions of
gcc either.
are created in the correct location. Always make them. For libraries
that live in /lib, this causes a /lib/libfoo.so and a compatibility
/usr/lib/libfoo.so to be created. We may want to drop the
/usr/lib/libfoo.so symlink at some future point.
need relative pathing to work correctly. This is s necessary step
for putting libraries in /lib while the .so symlinks still live
in /usr/lib.
This should be a big NOOP in the case where SHLIBDIR == LIBDIR.
All .s files that need cpp(1) processing (see gcc(1) manpage's
DESCRIPTION section) have been repo-copied to .S files. This
is mostly to bring bsd.lib.mk in agreement with sys.mk.
Desired by: obrien
symbols from object files has bitrotted over the last
thirteen years, and it now does more harm than good.
An attempt to work around the problems caused by using
ld(1) for stripping was to pass LDFLAGS to the ld(1)
command, but this was not right either as ${LDFLAGS}
should, by design, be used with cc(1) and not ld(1).
One of the proposed solutions was to use the objcopy(1)
utility to do the strip work, and the other would be to
use strip(1), but Bruce Evans suggested not stripping
any symbols at all. This works by leaving the grunt
work to the final strip(1) command (when installing the
binary).
Submitted by: bde
components. This is generally considered a non-optimal solution but
it gets the job done for the /rescue case.
Submitted by: Tim Kientzle <kientzle@acm.org>
