libunwind configure script explicitly links libunwind against
libgcc_s. libgcc_s is only guaranteed to be available for toolchains
that supports dynamic linking: pure static linking toolchains only
have libgcc.a, not libgcc_s.so.
Therefore, let's make libunwind unavailable on toolchains that lack
dynamic linking support. We could potentially support linking with
libgcc, but switching to libgcc_s was done upstream because libgcc was
lacking some symbols on ARM
(https://lists.nongnu.org/archive/html/libunwind-devel/2014-06/msg00024.html). Even
though recent gcc versions seem to provide such symbols in libgcc.a,
having libunwind available on static linking configurations is not a
useful enough use-case to do the necessary research to find when this
issue was fixed in gcc.
Since libunwind is not used as a mandatory dependency in any package,
adding this !BR2_STATIC_LIBS dependency is trivial and nicely avoids
the problematic situation.
This fixes two different autobuilder failures:
- Gstreamer 1.x programs failing to link, because libunwind links
against libgcc_s that isn't available (static linking
configuration):
http://autobuild.buildroot.net/results/9d4fbf7167e9afce0eef5c9e0cfd42c966ecba36/
- Gmrender-resurrect, which fails to link, because GStreamer 1.x uses
some libunwind functionality, but does not take into account the
libunwind dependency in its .pc files (static linking
configuration):
http://autobuild.buildroot.net/results/0a3a2485c187a000482c178f1e9c64dd716a858f/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
libunwind use sigreturn() and setcontext() while building for x86 [1]
but this function is not available with musl [2].
This throw a warning during libunwind build:
In file included from x86/Los-linux.c:4:0:
x86/Gos-linux.c: In function ‘_ULx86_local_resume’:
x86/Gos-linux.c:298:7: warning: implicit declaration of function ‘sigreturn’ [-Wimplicit-function-declaration]
sigreturn (sc);
^
But any program trying to link against libunwind-generic.so or
libunwind.so fail to build:
[...]/usr/lib/libunwind-generic.so: undefined reference to `sigreturn'
[...]/usr/lib/libunwind-generic.so: undefined reference to `setcontext'
collect2: error: ld returned 1 exit status
Disable libunwind for x86 target when musl is used.
Reported upstream by Waldemar [3]
[1] http://git.savannah.gnu.org/gitweb/?p=libunwind.git;a=blob;f=src/x86/Gos-linux.c;h=17aebc2974af50eb0bf8292689b2ed22a4c97866;hb=HEAD#l299
[2] http://openwall.com/lists/musl/2016/02/04/3
[3] http://lists.nongnu.org/archive/html/libunwind-devel/2017-04/msg00030.html
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Waldemar Brodkorb <wbx@openadk.org>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
BR2_ARM_CPU_HAS_ARM already embeddeds the fact that this is an ARM chip,
so (BR2_arm || BR2_armeb) is forcibly true when BR2_ARM_CPU_HAS_ARM is.
In other words, BR2_ARM_CPU_HAS_ARM can't be true unless either BR2_arm
or BR2_armeb itslef is true.
Simplify the logic accordingly.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Romain Naour <romain.naour@openwide.fr>
Acked-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
libunwind use sigreturn() while building for x86 [1] but this function
is not available with uClibc-ng.
This throw a warning during libunwind build:
In file included from x86/Los-linux.c:4:0:
x86/Gos-linux.c: In function ‘_ULx86_local_resume’:
x86/Gos-linux.c:298:7: warning: implicit declaration of function ‘sigreturn’ [-Wimplicit-function-declaration]
sigreturn (sc);
^
But any program trying to link against libunwind-generic.so fail to build:
[...]usr/lib/libunwind-generic.so: undefined reference to `sigreturn'
collect2: error: ld returned 1 exit status
Disable libunwind for x86 target when uClibc-ng is used.
Fixes:
http://autobuild.buildroot.net/results/54a/54afac8148cff5f3c17e83f80917fd9006948fe0//build-end.log
[1] http://git.savannah.gnu.org/gitweb/?p=libunwind.git;a=blob;f=src/x86/Gos-linux.c;h=17aebc2974af50eb0bf8292689b2ed22a4c97866;hb=HEAD#l299
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It's been deprecated for quite some time now.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Due to a misplaced parenthesis, the libunwind dependency were
incorrect. Indeed, they were of the form:
default y if LIBC_TYPE && (arm dependency) || (list of other architectures)
Which meant that that the LIBC_TYPE dependency was not taken into
account for the "other architectures". This commit changes that to:
default y if LIBC_TYPE && ((arm dependency) || list of other architectures)
Without this fix, libunwind can be selected for example on
PowerPC/uClibc configurations, while only PowerPC/glibc should be
allowed. This fixes:
http://autobuild.buildroot.net/results/96c53ff240dcd68fcc9e3d32c4026c9886edcbe7/
Similarly, it also fixes cases where SuperH/uClibc was allowed, while
only SuperH/glibc should be allowed. This fixes:
http://autobuild.buildroot.net/results/95a6cca21aea9914ba8bb3f571760fc054c53ecd/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The libunwind package currently dependency on glibc *or* uclibc
snapshot. However, we are going to remove the support for uclibc
snapshot, and uClibc-ng has equivalent functionality as uclibc
snapshot. Moreover, musl is also capable of building libunwind for
certain architectures.
Therefore, this commit reworks the architecture dependencies of
libunwind, to make it available on all architectures for which it is
supported, depending on the capabilities of the different C libraries,
and the implementation of libunwind on each architecture.
On some architectures, libunwind uses the C library *context()
functions, which are not provided by musl at all, and not provided by
uClibc on all architectures. But on some other architectures,
libunwind does not use the C library *context() functions, which
explains why it can be built with musl on such architectures.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since libatomic_ops does not implement real atomic operations for
ARMv4 and ARMv5, libunwind must define AO_REQUIRE_CAS do indicate it
requires compare-and-swap operations, even if not available as real
atomic operations for the current architecture. In this case,
libatomic_ops will rely on emulated atomic operations, which also
require linking against libatomic_ops, which was until now not done by
libunwind.
This fixes the mysterious ltrace build issue:
http://autobuild.buildroot.org/results/e1b/e1b330abfa2d80f3f30bc3359428ea429c690eb8/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When a package A depends on config option B and toolchain option C, then
the comment that is given when C is not fulfilled should also depend on B.
For example:
config BR2_PACKAGE_A
depends on BR2_B
depends on BR2_LARGEFILE
depends on BR2_WCHAR
comment "A needs a toolchain w/ largefile, wchar"
depends on !BR2_LARGEFILE || !BR2_WCHAR
This comment should actually be:
comment "A needs a toolchain w/ largefile, wchar"
depends on BR2_B
depends on !BR2_LARGEFILE || !BR2_WCHAR
or if possible (typically when B is a package config option declared in that
same Config.in file):
if BR2_B
comment "A needs a toolchain w/ largefile, wchar"
depends on !BR2_LARGEFILE || !BR2_WCHAR
[other config options depending on B]
endif
Otherwise, the comment would be visible even though the other dependencies
are not met.
This patch adds such missing dependencies, and changes existing such
dependencies from
depends on BR2_BASE_DEP && !BR2_TOOLCHAIN_USES_GLIBC
to
depends on BR2_BASE_DEP
depends on !BR2_TOOLCHAIN_USES_GLIBC
so that (positive) base dependencies are separate from the (negative)
toolchain dependencies. This strategy makes it easier to write such comments
(because one can simply copy the base dependency from the actual package
config option), but also avoids complex and long boolean expressions.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
(untested)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This patch lines up the comments in Config.in files that clarify which
toolchain options the package depends on.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[Thomas: added patch to disable build of tests that fail on uClibc,
added dependencies for architectures that are supported.]
Signed-off-by: Andy Kennedy <andy.kennedy@adtran.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>