Added support for TX lock elision of pthread mutexes on s390 and
s390x. This may improve lock scaling of existing programs on TX
capable systems. The lock elision code is only built with
--enable-lock-elision=yes and then requires a GCC version supporting
the TX builtins. With lock elision default mutexes are elided via
__builtin_tbegin, if the cpu supports transactions. By default lock
elision is not enabled and the elision code is not built.
The SELinux team has indicated to me that glibc's SELinux checks
in nscd are not being carried out as they would expect the API
to be used today. They would like to move away from static header
defines for class and permissions and instead use dynamic checks
at runtime that provide an answer which is dependent on the runtime
status of SELinux i.e. more dynamic.
The following patch is a minimal change that moves us forward in
this direction.
It does the following:
* Stop checking for SELinux headers that define NSCD__SHMEMHOST.
Check only for the presence or absence of the library.
* Don't encode the specific SELinux permission constants into a
table at build time, and instead use the symbolic name for the
permission as expected.
* Lookup the "What do we do if we don't know this permission?"
policy and use that if we find SELinux's policy is older than
the glibc policy e.g. we make a request for a permission that
SELinux doesn't know about.
* Lastly, translate the class and permission and then make
the permission check. This is done every time we lookup
a permission, and this is the expected way to use the API.
SELinux will optimize this for us, and we expect the network
latencies to hide these extra library calls.
Tested on x86, x86-64, and via Fedora Rawhide since November 2013.
See:
https://sourceware.org/ml/libc-alpha/2014-04/msg00179.html
This patch makes the configure adds -D_CALL_ELF=1 when compiler does
not define _CALL_ELF (versions before powerpc64le support). It cleans
up compiler warnings on old compiler where _CALL_ELF is not defined
on powerpc64(be) builds.
It does by add a new config.make variable for configure-deduced
CPPFLAGS and accumulate into that (confix-extra-cppflags). It also
generalizes libc_extra_cflags so it accumulates in sysdeps configure
fragmenets.
Autoconf is tested for and run if needed only when --enable-maintainer-mode
is used on configure. This results in the autom4te.cache directory only
being written in the source directory during configure if automatic
autoconf usage is requested.
Fixes BZ #14120.
Autoconf has been deprecating configure.in for quite a long time.
Rename all our configure.in and preconfigure.in files to .ac.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
http://sourceware.org/ml/libc-alpha/2013-08/msg00096.html
This adds the basic configury bits for powerpc64le and powerpcle.
* configure.in: Map powerpc64le and powerpcle to base_machine/machine.
* configure: Regenerate.
* nptl/shlib-versions: Powerpc*le starts at 2.18.
* shlib-versions: Likewise.
The helper binary pt_chown tricked into granting access to another
user's pseudo-terminal.
Pre-conditions for the attack:
* Attacker with local user account
* Kernel with FUSE support
* "user_allow_other" in /etc/fuse.conf
* Victim with allocated slave in /dev/pts
Using the setuid installed pt_chown and a weak check on whether a file
descriptor is a tty, an attacker could fake a pty check using FUSE and
trick pt_chown to grant ownership of a pty descriptor that the current
user does not own. It cannot access /dev/pts/ptmx however.
In most modern distributions pt_chown is not needed because devpts
is enabled by default. The fix for this CVE is to disable building
and using pt_chown by default. We still provide a configure option
to enable hte use of pt_chown but distributions do so at their own
risk.
Check wheter the compiler has the option -fno-tree-loop-distribute-patterns
to inhibit loop transformation to library calls and uses it on memset
and memmove default implementation to avoid recursive calls.
We no longer support configuring for i386, nor do we
elide such a configuration to i686. Configuring with
i386-* is a failure, and we provide an example of
how to fix that.
---
2013-04-17 Carlos O'Donell <carlos@redhat.com>
* configure.in: Remove i386 configure warning. Remove i386 case.
* configure: Regenerate.
* sysdeps/i386/configure.in: Raise error if config_machine is i386.
Add example to error message.
* sysdeps/i386/configure: Regenerate.
This change does two things:
* Treats a target i386-* as if it were i686.
* Fails configure if the user is generating code
for i386.
We no longer support i386 code-generation because the i386
lacks the atomic operations we need in glibc.
You can still configure for i386-*, but you get i686 code.
You can't build with --march=i386, --mtune=i386 or a compiler
that defaults to i386 code-generation.
I've added two i386 entries in the master todo list to discuss
merging and renaming:
http://sourceware.org/glibc/wiki/Development_Todo/Master#i386
The failure modes are fail-safe here. You compile for i386,
get i686, and try to run on i386 and it fails. The configure
log has a warning saying we elided to i686. There is no situation
that I can see where we run into any serious problems.
The patch makes the current state better in that we get less
confused users and we build successfully in more default
configurations.
The next enhancement would be to add --march=i?86
as suggested in #c20 of BZ#10062 for any i?86-* builds, which
would solve the problem of a 32-bit compiler that defaults to
i386 code-gen and glibc configured for i686-* target. Which
previously failed at build time, and now will fail at configure
time (requires adding --march=i686).
Updated NEWS with BZ #10060 and #10062.
No regressions.
---
2013-04-06 Carlos O'Donell <carlos@redhat.com>
[BZ #10060, #10062]
* aclocal.m4 (LIBC_COMPILER_BUILTIN_INLINED): New macro.
* sysdeps/i386/configure.in: Use LIBC_COMPILER_BUILTIN_INLINED and
fail configure if __sync_val_compare_and_swap is not inlined.
* sysdeps/i386/configure: Regenerate.
* configure.in: Build for i686 when configured for i386.
* configure: Regenerate.
* README: Remove i386 reference.
Stop assuming specific path layouts for C++ headers, and instead
use an autodetection method that looks for paths with '/[cg]++'
in the g++ include list.
The test currently tests the binutils frontend support which passes for
all versions of binutils we currently require (2.20+). It doesn't test
the backend which is required for ifunc to actually work, and which most
targets don't yet support.
Change the assembly code so that when we link it, we get a file that has
ifunc relocations if the backend supports it. That way we can test to
see if binutils supports everything we need.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
New configure option --enable-obsolete-rpc makes the deprecated RPC
headers and functions available at compile time as they were before
version 2.14. This option will be removed at some time in the future
after the TI-RPC library becomes fully sufficient for the needs of
existing applications.
* sysdeps/sparc/configure.in: New file.
* sysdeps/sparc/configure: Generate.
* configure.in (libc_cv_sparc_as_vis3): Substitute.
* configure: Regenerate.
* config.h.in (HAVE_AS_VIS3_SUPPORT): New.
* config.make.in (have-as-vis3): New.
* sysdeps/sparc/sparc32/sparcv9/Makefile (ASFLAGS-*): If VIS3 is
available use -Av9d instead of -Av9a.
* sysdeps/sparc/sparc64/Makefile: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/Makefile: New file.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_copysign-vis3.S: New file.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_copysign.S: New file.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_copysignf-vis3.S: New file.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_copysignf.S: New file.
* sysdeps/sparc/sparc64/fpu/multiarch/Makefile: New file.
* sysdeps/sparc/sparc64/fpu/multiarch/s_signbit-vis3.S: New file.
* sysdeps/sparc/sparc64/fpu/multiarch/s_signbit.S: New file.
* sysdeps/sparc/sparc64/fpu/multiarch/s_signbitf-vis3.S: New file.
* sysdeps/sparc/sparc64/fpu/multiarch/s_signbitf.S: New file.
Some symbols have to be identified process-wide by their name. This is
particularly important for some C++ features (e.g., class local static data
and static variables in inline functions). This cannot completely be
implemented with ELF functionality so far. The STB_GNU_UNIQUE binding
helps by ensuring the dynamic linker will always use the same definition for
all symbols with the same name and this binding.
* config.make.in: Add nss-crypt entry.
* crypt/Makefile: If nss-crypt==yes, don't build md5.c, sha256.c,
sha512.c. Don't run md5test, sha256test, sha512test. Pass -DUSE_NSS
and include path for NSS directory to compiler for md5-crypt,
sha256-crypt, sha512-crypt. Link libcrypt.so with -lfreebl3.
* crypt/md5-crypt.c: If USE_NSS is defined, don't use local hash
function implementation, use NSS. Introduce wrappers around the
hash function calls. Little code size optimization.
* crypt/sha256-crypt.c: Likewise.
* crypt/sha512-crypt.c: Likewise.
* scripts/check-local-headers.sh: Ignore nss3 directory.
* configure.in: Rename pic_default to libc_cv_pic_default.
* config.make.in: Likewise.
smaller scopes.
* elf/dl-dst.h (DL_DST_REQ_STATIC): Add l as macro argument.
(DL_DST_REQUIRED): Adjust user.
* include/dlfcn.h (struct link_map): New forward decl.
* inet/getnameinfo.c: Include stddef.h.
(getnameinfo): Use offsetof.
* time/tst-mktime2.c (do_test): Don't rely on signed wrap.
* stdio-common/vfprintf.c (_itoa): Undef before redefining.
* string/strerror_l.c: Include stdlib.h.
(_dl_lookup_symbol_x): Rename hash to old_hash and don't compute
value here. Compute new-style hash value. Pass new hash value
and reference to variable with the old value to do_lookup_x.
(_dl_setup_hash): If DT_GNU_HASH is defined, use it and not
old-style hash table.
(_dl_debug_bindings): Pass new hash value and reference to variable
with the old value to do_lookup_x.
* elf/do-lookup.h (do_lookup_x): Accept additional parameter with
new-style hash value and change old-style hash value parameter to
be a reference. Reoganize functions to determine whether
new-style hash table is available. Only fall back on old-style
table. If old-style hash value is needed, compute it here.
* elf/dynamic-link.h (elf_get_dynamic_info): Relocate DT_GNU_HASH
entry.
* elf/elf.h: Define SHT_GNU_HASH, DT_GNU_HASH, DT_TLSDEC_PLT,
DT_TLSDEC_GOT. Adjust DT_ADDRNUM.
* include/link.h (struct link_map): Add l_gnu_bitmask_idxbits,
l_gnu_shift, l_gnu_bitmask, l_gnu_buckets and l_gnu_chain_zero.
* Makeconfig: If linker supports --hash-style option add it to all
linker command lines to build DSOs.
* config.make.in: Define have-hash-style.
* configure.in: Test whether linker supports --hash-style option.
* elf/dl-misc.c (_dl_name_match_p): Make MAP parameter const.
* sysdeps/generic/ldsodefs.h: Adjust prototype.
* configure: Regenerated.
* sysdeps/sparc/sparc32/sparcv9v/memcpy.S: New file.
* sysdeps/sparc/sparc32/sparcv9v/memset.S: New file.
* sysdeps/sparc/sparc64/sparcv9v/memcpy.S: New file.
* sysdeps/sparc/sparc64/sparcv9v/memset.S: New file.
($(all-subdirs-targets)): Use it in place of -C option.
* Rules: Use $(..) instead of ../ if it's already defined.
* Makeconfig (subdir-srcdirs): New variable.
* csu/Makefile (all-Banner-files): Use it.
* configure.in (--enable-add-ons): Set to "yes" by default.
Handle absolute add-on directory names when looking for configure
fragments. Also look for sysdeps/*/preconfigure fragments in add-ons.
Require add-on configure to set $libc_add_on_canonical, use
that in $add_ons_sfx. Substitute add_on_subdirs with computed list
of subdir names each add-on configure set in libc_add_on_subdirs.
* configure: Regenerated.
* Makefile (%/preconfigure: %/preconfigure.in): New pattern rule.
* config.make.in (add-on-subdirs): New substituted variable.
* Makeconfig (all-subdirs): Include $(add-on-subdirs).
Remove $(add-ons), $(sysdep-subdirs).
Don't filter out $(sysdep-inhibit-subdirs).
($(common-objpfx)sysd-dirs): Target removed. Don't include it.
($(common-objpfx)sysd-sorted): Rewritten to feed Depend and Subdirs
files together to gen-sorted.awk, and $(subdirs) via -v.
(subdirs): Remove magic reordering for mach and hurd.
* scripts/gen-sorted.awk: Use subdirs from command line.
Process Subdirs and Depend files directly.
Let Subdirs files use "first dir" and "inhibit dir".
Always move elf to the end of the list.
* hurd/Depend: New file.
* sysdeps/mach/Subdirs: Use "first mach".
* Makefile (dist-separate): Remove linuxthreads.
(dist-separate-linuxthreads): Variable removed.
(glibc-%.tar rule): Use $(sysdeps-add-ons).
* Makerules ($(common-objpfx)Versions.v.i): Use $(subdirs),
not $(all-subdirs).
(sysdep-makefiles): Use $(sysdirs).
(sysdirs): Remove export.
($(+sysdir_pfx)sysd-rules): Handle absolute directory names in
$(config-sysdirs).
(+sysdir_pfx): Variable removed.
(sysd-rules): Use $(common-objpfx) in place of it.
(sysdirs): Variable moved to ...
* Makeconfig (sysdirs): ... here.
Handle absolute directory names in $(config-sysdirs).
(full_config_sysdirs): Variable removed.
* csu/Makefile: Use $(sysdirs) in vpath directive.
* math/Makefile (ulps-file): Use $(sysdirs).
* sysdeps/gnu/Makefile (errlist-c): Likewise.
($(objpfx)errlist-compat.c): Likewise.
* Makeconfig (all-Subdirs-files): Likewise.
($(common-objpfx)config.status): Likewise.
* configure.in (sysnames): Handle absolute add-on directory names.
(sysdeps_add_ons): New variable, AC_SUBST it.
Compute which add-ons contributed sysdeps directories.
* configure: Regenerated.
* config.make.in (sysdeps-add-ons): New substituted variable.
* Makerules (+sysdep_dirs, +sysdep-includes): Variables moved ...
* Makeconfig: ... to here.
(+sysdep_dirs): Append $(sysdeps-add-ons) here.
(+includes): Remove $(objpfx) include, already in $(+sysdep_dirs).
Remove $(includes).
(sysdep-makeconfigs): Use $(+sysdep_dirs).
($(common-objpfx)shlib-versions.v.i): Likewise.
* Makeconfig: Remove hair to set Makeconfig-add-on.
* sysdeps/unix/Makefile (sysdirs): Remove export.
(asm_CPP): Variable removed.
($(common-objpfx)sysd-syscalls): Pass them directly for the script.
* sysdeps/posix/Makefile: New file.
* Makerules (L_tmpnam, TMP_MAX, L_ctermid, L_cuserid): Set non-posix
values here with ?=.
* stdlib/gen-mpn-copy: File removed.
* stdlib/Makefile (distribute): Remove it.
* configure.in: Don't grok --with-gmp.
* configure: Regenerated.
* configure.in (libc_cv_idn): Don't check it; libidn/configure does it.
* configure: Regenerated.
* bare: Directory removed, saved in ports repository.
Roland McGrath <roland@redhat.com>
* configure.in (fno_unit_at_a_time): Set to
-fno-toplevel-reorder if the compiler supports it.
Otherwise use -fno-unit-at-a-time unconditionally.
* configure: Regenerated.
sysdeps/.../machine/CPU subdirs, and give error if none are found.
* configure: Regenerated.
* config.make.in (with-cpu): New substituted variable.
* Makeconfig (+cflags): Add -mcpu=$(with-cpu) if set.
2005-11-04 Roland McGrath <roland@redhat.com>
[BZ #1089]
* configure.in: Change order of add-on sysdeps directories,
so they are interleaved with main sysdeps directories.
* configure: Regenerated.
* configure.in: Add --enable-stackguard-randomization option.
(ENABLE_STACKGUARD_RANDOMIZE): New define.
* config.h.in (ENABLE_STACKGUARD_RANDOMIZE): Add.
* sysdeps/unix/sysv/linux/dl-osinfo.h: Include stdint.h.
(_dl_setup_stack_chk_guard): New inline function.
* sysdeps/generic/dl-osinfo.h: Include stdint.h.
(_dl_setup_stack_chk_guard): New inline function.
* elf/rtld.c (__stack_chk_guard): New variable.
(dl_main): Remove all traces of TLS_INIT_TP_EXPENSIVE.
Set __stack_chk_guard to _dl_setup_stack_chk_guard (),
use THREAD_SET_STACK_GUARD if defined.
* elf/Versions (ld): Export __stack_chk_guard@@GLIBC_2.4.
* sysdeps/generic/libc-start.c (__stack_chk_guard): New variable.
(__libc_start_main): Set __stack_chk_guard to
_dl_setup_stack_chk_guard (), use THREAD_SET_STACK_GUARD if defined.
* sysdeps/generic/libc-tls.c (__libc_setup_tls): Remove all
traces of TLS_INIT_TP_EXPENSIVE.
* debug/Versions (libc): Export __stack_chk_fail@@GLIBC_2.4.
* debug/Makefile (routines): Add stack_chk_fail.
(static-only-routines): Add stack_chk_fail_local.
* debug/stack_chk_fail_local.c: New file.
* debug/stack_chk_fail.c: New file.
* elf/Makefile: Add rules to build and run tst-stackguard1{,-static}
tests.
* elf/tst-stackguard1.c: New file.
* elf/tst-stackguard1-static.c: New file.
* elf/stackguard-macros.h: New file.
* hesiod/hesiod.c (hesiod_init): Initialize classes. Fail if no
valid classes were given or if both are equal.
(hesiod_resolve): Use ctx->classes instead of hardcoded C_IN
and C_HS order.
(parse_config_file): Handle classes keyword.
* hesiod/README.hesiod: Mention addition of the classes keyword.
* sysdeps/generic/strchrnul.c: Add cast to avoid warning.
* libio/iofwide.c: Add casts to avoid warnings.
* stdio-common/printf-prs.c (parse_printf_format): Introduce new
variable f to avoid warnings.
* sysdeps/unix/sysv/linux/x86_64/makecontext.c (__makecontext):
Fix a few casts to avoid warnings.
* iconv/gconv_simple.c (internal_utf8_loop): Make start unsigned
to avoid warning.
(_dl_out_of_memory): Use rtld_hidden_proto.
* elf/dl-error.c (_dl_out_of_memory): Use rtld_hidden_data_def.
(_dl_signal_error): Don't use INTUSE on _dl_out_of_memory.
* elf/dl-open.c (_dl_open): Likewise.
* elf/dl-deps.c (_dl_map_object_deps): Likewise.
* sunrpc/des_impl.c (des_set_key): Make first argument unsigned char *.
* elf/dl-addr.c (_dl_addr): Add a cast.
* elf/unload3mod4.c: Declare foo.
* elf/testobj2.c: Include <stdio.h>.
* sysdeps/gnu/errlist.awk: Emit NERR macro for _sys_nerr_internal
value constant, and use it in the defn.
* sysdeps/gnu/errlist-compat.awk: Use NERR in array decl.
Use actual size for compat array decls.
* config.make.in (cc-mtune): New substituted variable.
* configure.in (libc_cv_cc_mtune): New test for -mtune= switch,
done only if libc_mtune_example is defined.
* configure: Regenerated.
* sysdeps/i386/configure.in (libc_mtune_example): Set it.
* sysdeps/i386/configure: Regenerated.
* sysdeps/i386/Makefile (CFLAGS-initfini.s): Use $(cc-mtune).
2004-09-08 H.J. Lu <hongjiu.lu@intel.com>
* Makeconfig (libunwind): New.
(libgcc_eh): Add $(libunwind).
(gnulib): Always set to -lgcc $(libgcc_eh).
(static-gnulib): Always set to -lgcc -lgcc_eh $(libunwind).
(libc.so-gnulib): New.
* Makerules (LDLIBS-c.so): Use $(libc.so-gnulib) instead of
$(static-gnulib).
* configure.in (libc_cv_cc_with_libunwind): Set to yes if gcc
uses -lunwind for static binaries.
indicating --enable-add-ons with no explicit list.
(running add-on fragments): Allow a fragment to modify $libc_add_on
and have that affect its place in the list of add-ons to use.
* configure: Regenerated.
Fix typo: $(add_ons) -> $(add-ons).
(Makeconfig-add-on): New variable. When doing $(sysdep-makeconfigs)
include, use black magic to get it set to an add-on's name during the
include of the add-on's Makeconfig.
* configure.in: Use variable name `libc_add_on' when sourcing add-on
configure fragments, so they can refer to this.
* configure: Regenerated.
after AC_CANONICAL_HOST and default setting of $machine et al.
Don't set $subdirs from $add_ons.
Instead, source add-on/configure fragments early on.
(base_machine): If it's already set, don't set it based on $machine.
(libc_config_ok): New variable, set to no. If an add-on fragment sets
it to yes, skip the tuple sanity check as if --enable-hacker-mode.
(sysnames): Try appending add-on names after machine as well.
* configure: Regenerated.
* config.make.in (add-ons): Set from @add_ons@ instead of @subdirs@.
* Makeconfig ($(common-objpfx)config.status): Also depend on configure
files in $(add_ons) dirs.
* sysdeps/unix/sysv/linux/bits/in.h (struct ip_mreq): Remove
definition, now in netinet/in.h proper.
long. Rewrite example code so that it takes into account l64a output
(glibc-%.tar rule): Make separate tar files for add-ons listed there.
Depend on their configure files.
(dist-do-separate-dirs): New canned sequence to do that.
(dist-separate-libidn, dist-separate-linuxthreads): New variables.
(dist): Depend on add-on tar files based on $(dist-separate).
(dist-version): New variable, default to $(version).
(tag-for-dist, dist): Use that in place of $(version) in deps.
* Makefile (dist-separate): New variable.
(glibc-%.tar rule): Make separate tar files for add-ons listed there.
Depend on their configure files.
(dist-do-separate-dirs): New canned sequence to do that.
(dist-separate-libidn, dist-separate-linuxthreads): New variables.
(dist): Depend on add-on tar files based on $(dist-separate).
(dist-version): New variable, default to $(version).
(tag-for-dist, dist): Use that in place of $(version) in deps.
* config.h.in: Add entry for HAVE_LIBIDN.
* configure.in: If libidn add-on present define HAVE_LIBIDN.
* Versions.def: Add entry for libcidn.
2004-03-07 Simon Josefsson <jas@extundo.com>
* resolv/netdb.h [__USE_GNU]: Add new AI_IDN ai_flags for addrinfo.
[__USE_GNU]: Add new error code EAI_IDN_ENCODE for getaddrinfo.
* sysdeps/posix/getaddrinfo.c: Add prototype for __idna_to_ascii_lz
and define IDNA_SUCCESS.
(gaih_inet): If ai_flags have AI_IDN, invoke __idna_to_ascii_lz.
(getaddrinfo): Fix EAI_BADFLAGS test to include AI_IDN.
All changes only applicable when glibc is compiled with the libidn
add-on.
2004-03-07 Ulrich Drepper <drepper@redhat.com>
