Commit Graph

623 Commits

Author SHA1 Message Date
des
ff3ab430a7 Fix comment about KERNFAST. 2009-09-10 07:37:36 +00:00
kan
ef443476d9 Second attempt at eliminating .text relocations in shared libraries
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)
2009-07-14 21:19:13 +00:00
kan
a3faeb1b41 Back out previous revision until better tested fix is ready.
Approved by: re (impliciti, by approving previos check-in)
2009-06-29 01:33:59 +00:00
kan
f780ef8f19 Eliminate .text relocations in shared libraries 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 everys
time given shared library is being relocated and making affected text pages
non-shareable.

Reviewed by:        kib
Approved by:        re (kensmith)
2009-06-28 23:51:39 +00:00
markm
90ad5bd1d7 Allow "make update" work with SVN, if you take the time to set up your
/etc/make.conf properly.

Only one SCM?
I do not think this is right.
Now I have fixed it.
2009-06-14 15:16:24 +00:00
imp
2a2dfa7d22 Revert c190007 at the request of Ruslan. It was not correct because
ar had been moved from cross tools to bootstrap tools when bsd ar was
introduced.

Submitted by:	ru@
2009-04-05 00:24:49 +00:00
imp
243a40bf04 Add usr.bin/ar to the list of cross tools. Before, when we had gnu
ar, it was built in the gnu/usr.bin/binutils tree.  Now it isn't.

Submitted by:	John Hein
2009-03-19 00:44:22 +00:00
imp
eb1d81774b Restore missing OSREL definition that accidetnally dropped from an
earlier version of this patch.
2009-03-15 23:52:13 +00:00
gabor
0592710988 - Create the buildworld object directories with mtree instead of various
mkdir calls
- Remove the ugly workaroung from libc NLS, which was to create some of
  these directories
2009-03-15 13:14:06 +00:00
gabor
a9bdc02cf2 - Reenable Native Language Support in libc. This feature was disabled due
to possible breakages in the catalog handling code.  Since then, that
  code has been replaced by the secure code from NetBSD but NLS in libc
  remained turned off.  Tests have shown that the feature is stable and
  working so we can now turn it on again.

- Add several new catalog files:
  - ca_ES.ISO8859-1
  - de_DE.ISO8859-1
  - el_GR.ISO8859-7  (by manolis@ and keramida@)
  - es_ES.ISO8859-1  (kern/123179, by carvay@)
  - fi_FI.ISO8859-1
  - fr_FR.ISO8859-1  (kern/78756, by thierry@)
  - hu_HU.ISO8859-2  (by gabor@)
  - it_IT.ISO8859-15
  - nl_NL.ISO8859-1  (corrections by rene@)
  - no_NO.ISO8859-1
  - mn_MN.UTF-8      (by ganbold@)
  - sk_SK.ISO8859-2
  - sv_SE.ISO8859-1
  (The catalogs without explicit source has been obtained from NetBSD.)

Approved by:	attilio
2009-03-13 10:40:38 +00:00
ru
d09496ec68 Don't put "install-info" to the list of install-tools if we're
installing with -DWITHOUT_INFO, otherwise one can experience a
failure trying to installworld on a system that is built with
-DWITHOUT_INFO (i.e., without /usr/bin/install-info).

Reported by:	bland
MFC after:	3 days
2009-03-13 10:09:08 +00:00
imp
6270d8e9fe Implement the xdev target. When you define XDEV=arch XDEV_ARCH=arch,
you can build the cross development tools and install them as
$XDEV-freebsd-xxx for each tool.  This allows one to use autoconf to
find the tools for cross building scenarios.
2009-03-13 07:23:58 +00:00
ru
21f7074ade Fix build when WITH_SSP is set explicitly.
Submitted by:	Jeremie Le Hen
2009-02-21 15:04:31 +00:00
imp
71d7b5671d Implement an idea from Sam Leffler:
make KERNFAST=blah buildkernel
is now a short cut for
	make KERNCONF=blah -DKERNFAST buildkernel

This change works for all kernel config files not named "1".  I did
that to make sure that
	make -DKERNFAST buildkernel
remains the same as
	make -DKERNFAST KERNCONF=GENERIC buildkernel
rather than trying to build and configure "1".  I've never seen a
kernel config file named "1," so I think this is a good compromise.
2009-01-23 18:23:16 +00:00
imp
bcc2168187 Add KERNFAST define. When defined, it skips all the config, depends
and clean steps.  KERNFAST was selected to complement KERNCONF which
is typically used in these scenarios (especially with cross building).

Reviewed by:	arch@
2009-01-15 04:19:03 +00:00
obrien
f8fc18af0b Add the kerberos5 libs to the install32 target.
(we've been building the all along, but never installing them)
2008-11-12 04:43:55 +00:00
des
56068b0b4c Back out testing aid that was inadvertantly committed as part of r184205.
Noticed by:	rdivacky
2008-10-23 16:06:49 +00:00
des
66f807ed8b Retire the MALLOC and FREE macros. They are an abomination unto style(9).
MFC after:	3 months
2008-10-23 15:53:51 +00:00
obrien
a01a900ec7 Rescue is "not yet" for MIPS. 2008-09-19 16:14:42 +00:00
jhb
76ad1fe909 If a CPUTYPE isn't specified, then don't use -march=k8 when compiling
32-bit compat libs on amd64 since -march=k8 may generate instructions
that are not implemented on Intel EM64T processors.  Instead, use
a simpler set of default flags that should work on all amd64-capable
CPUs.

PR:		amd64/113111
Submitted by:	NIIMI Satoshi  sa2c of sa2c.net
MFC after:	1 week
2008-08-19 14:23:26 +00:00
jhb
78d938b9da Various style fixes in the build32 rules.
Submitted by:	bde (mostly)
2008-08-02 12:14:22 +00:00
jhb
a0f546e88f Tweak the support for using ldd on 32-bit objects a bit further.
Specifically, build a 32-bit /usr/bin/ldd32 on amd64 which handles 32-bit
objects.  Since it is a 32-bit binary, it can fork a child process which
can dlopen() a 32-bit shared library.  The current 32-bit support in ldd
can't do this because it does the dlopen() from a 64-bit process.  In order
to preserve an intuitive interface for users, the ldd binary automatically
execs /usr/bin/ldd32 for 32-bit objects.  The end result is that ldd on
amd64 now transparently handles 32-bit shared libraries in addition to
32-bit binaries.

Submitted by:	ps (indirectly)
2008-08-01 21:52:41 +00:00
ru
8735fdbd4c Enable GCC stack protection (aka Propolice) for userland:
- It is opt-out for now so as to give it maximum testing, but it may be
  turned opt-in for stable branches depending on the consensus.  You
  can turn it off with WITHOUT_SSP.
- WITHOUT_SSP was previously used to disable the build of GNU libssp.
  It is harmless to steal the knob as SSP symbols have been provided
  by libc for a long time, GNU libssp should not have been much used.
- SSP is disabled in a few corners such as system bootstrap programs
  (sys/boot), process bootstrap code (rtld, csu) and SSP symbols themselves.
- It should be safe to use -fstack-protector-all to build world, however
  libc will be automatically downgraded to -fstack-protector because it
  breaks rtld otherwise.
- This option is unavailable on ia64.

Enable GCC stack protection (aka Propolice) for kernel:
- It is opt-out for now so as to give it maximum testing.
- Do not compile your kernel with -fstack-protector-all, it won't work.

Submitted by:	Jeremie Le Hen <jeremie@le-hen.org>
2008-06-25 21:33:28 +00:00
obrien
9b1f265839 Add MIPS to the list of known arches. 2008-06-18 13:52:58 +00:00
jb
aa93e85738 Add the DTrace build tools to the list of tools and pass NO_CTF to all
the stages of the build which don't require CTF conversion.
2008-05-23 04:22:14 +00:00
dfr
be0348cb75 Fix conflicts after heimdal-1.1 import and add build infrastructure. Import
all non-style changes made by heimdal to our own libgssapi.
2008-05-07 13:53:12 +00:00
ru
b388dd7a1d Bootstrap the build of btxld. 2008-04-15 05:14:42 +00:00
ru
0f0375e36a Remove options MK_LIBKSE and DEFAULT_THREAD_LIB now that we no longer
build libkse.  This should fix WITHOUT_LIBTHR builds as a side effect.
2008-03-29 17:44:40 +00:00
ru
d9f8b4ad27 Set AR=gnu-ar and RANLIB=gnu-ranlib on systems where we are forced
to use GNU tools.  Remove the _WITH_GNUAR knob.

Prodded by:	obrien
2008-03-02 11:10:46 +00:00
ru
b9adbf63f9 Sort.
Prodded by:	obrien
2008-02-25 16:29:54 +00:00
ru
26039ac217 Make again BSD ar(1) the default system ar(1), now properly handling
source upgrades by falling back to GNU ar(1) as necessary.  Option
WITH_BSDAR is gone.  Option _WITH_GNUAR to aid in upgrades is *not*
supposed to be set by the user.

Stop bootstrapping BSD ar(1) on the next __FreeBSD_version bump, as
there are no known bugs in it.  Bump __FreeBSD_version to anticipate
this and to flag the switch to BSD ar(1), should it be needed for
something.

Input from:	obrien, des, kaiw
2008-02-25 16:16:17 +00:00
des
f408263bc9 Satisfy a particularly obstinate nit-picker :) 2008-02-05 15:41:58 +00:00
des
a2636de674 Give usr.bin/kdump/kdump_subr.c the same treatment as usr.bin/kdump/ioctl.c
and usr.bin/truss/ioctl.c.  This is the correct way to address the problem
that arises when doing an incremental build after a header used by kdump
has been removed (cf. i4b disconnect a while ago)

Explained by:	ru
MFC after:	2 weeks
2008-02-05 08:56:18 +00:00
ru
77a6831b74 Support source upgrades from at least 6.0-RELEASE.
Reviewed by:	imp, obrien
2008-01-23 22:21:36 +00:00
ache
66dc6c3da8 Conditionally add mklocale to bootstrap-tools 2008-01-23 00:16:37 +00:00
ru
59177070da Cleanup after last commit (remove undefined variables). 2008-01-22 15:57:57 +00:00
obrien
362440a880 Remove 5.x and 6.x cruft - source upgrades to RELENG_8 from versions prior
to RELENG_7 are not supported.
2008-01-21 18:44:55 +00:00
brueffer
3d6e2998f3 Remove WITHOUT_FORTRAN from the XMAKE target, it doesn't exist anymore.
Reviewed by:	imp, kan
Approved by:	rwatson (mentor)
MFC after:	3 days
2008-01-16 15:05:49 +00:00
yar
569ad75e67 - Explicitly verify if all needed libs were found by ldd(1). Do so
through scanning its output as ldd(1) returns a non-zero status only
  for really abnormal conditions such as an improper file format.
  Now cp(1) won't get bogus "not" and "found" arguments if a lib
  is missing. [1]

- Don't guess if an element of a complex pipeline is assigned to the main
  shell or a sub-shell.  Namely use stdio, not vars, to pass lists out from
  loops.  If using vars, there's the risk that a loop will run in a sub-shell
  and the list won't make it to the main shell.  It appears that braces and
  parens give only limited control over the issue while stdio always works
  as intended.  Apply this solution to both $progs and $libs for consistency,
  although I've failed to go without it only in the $libs part.

Requested by:	emaste [1]
2007-12-04 12:55:27 +00:00
yar
c9f53dc499 Decouple the install tools from the main system as much as possible.
I.e., not only copy them to a scratch dir, but also make them use saved
copies of libraries and locale files.  That gives us several benefits:

1) ABI breakages should no longer affect installworld over the live system.

2) It becomes safe to run installworld while still running the old kernel.
   However, it can be reasonable to save the old /rescue before that to be
   able to run the old reboot(8), as the new binaries are rather likely to
   fail with the old kernel.  Anyhow, it's now possible to upgrade a system
   in a single reboot _reliably_.

3) With a bit of hackery around rtld(8), it becomes possible to do destructive
   cross-installs, e.g., i386->amd64 over the live system.

The only shared item left between the old and new systems is rtld(8),
which cannot be run from a saved copy easily because its full
pathname is stored in the respective field of each ELF executable.
(In theory, that field could be overridden, e.g., from the environment,
but this can lead to security issues.)  That's why a destructive
cross-install isn't possible w/o hackery yet.

Fruitful ideas by:	ru
Reviewed by:		ru
Tested with:		audit(4)
2007-10-31 09:26:42 +00:00
emaste
be4cf9548f If a required install tool is missing, provide a more friendly error
message (instead of cp's usage info).
2007-10-15 13:50:46 +00:00
obrien
a1598920aa Tweak the handling of "WITHOUT_LIBPTHREAD". Also remove the accidental
treatment of 'LIBKSE' as an "old style" knob.

Submitted by:	ru
Approved by:	re(kensmith)
2007-10-09 23:31:11 +00:00
obrien
eb3f6a9e74 Repo copy libpthreads to libkse.
This introduces the WITHOUT_LIBKSE nob,
and changes WITHOUT_LIBPTHREADS to mean with neither threading libs.
Approved by:	re(kensmith)
2007-10-09 13:42:34 +00:00
ru
7ef0a53855 Revise the list of directories we manually create under
${WORLDTMP} and ${LIB32TMP}; some of them are no longer
needed, and some were never needed.

Approved by:	re (kensmith)
2007-10-01 18:24:46 +00:00
ru
0ea9576a4d Always install libpthread.* symlinks if at least one of
the threading libraries is built.  This simplifies the
logic in makefiles that need to check if the pthreads
support is present.  It also fixes a bug where we would
build a threading library that we shouldn't have built:
for example, building with WITHOUT_LIBTHR and the default
value of DEFAULT_THREADING_LIB (libthr) would mistakenly
build the libthr library, but not install it.

Approved by:	re (kensmith)
2007-10-01 18:22:32 +00:00
delphij
15ea06f0e4 Add sed(1) to cross tools. We do want newly built version during
buildworld.

Approved by:	re (rwatson)
2007-07-10 10:17:32 +00:00
ru
cbfe2577e0 lex(1) no longer installs headers into GCC-internal dirs
(since src/usr.bin/lex/Makefile,v 1.20).
2007-05-26 20:17:19 +00:00
des
eb4340f8a2 Revision 1.576 removed too much. I didn't notice because my /usr is on
ZFS, which doesn't support flags...
2007-05-19 20:34:29 +00:00
kan
bbc340f2f6 GCC will now properly link libc with all shared objects in order to record
their dependency on libc and its versioned symbols. For that to work,
libc needs to be built before any other shared library that might depend
on it. Add necessary glue to make that happen.
2007-05-19 03:34:59 +00:00
des
9c0abd9ab4 Don't use a subshell where it isn't needed.
Noticed by:	John E Hein <jhein@timing.com>
2007-05-16 20:16:06 +00:00