This commit updates the gcc musl patches for gcc 4.7, 4.8 and 4.9 so
that the path to the dynamic linker encoded as "program interpreter"
in the generated binaries actually matches the symbolic link installed
by musl when building for mips soft-float.
Indeed, musl installs a symlink called ld-musl-mipsel-sf.so.1, but gcc
currently generates binaries that use /lib/ld-musl-mips.so as program
interpreter.
The fix is simply the one from
https://bitbucket.org/GregorR/musl-cross/commits/825219202365, i.e
adjust MUSL_DYNAMIC_LINKER in our musl gcc patches.
Thanks to these patches:
$ ./host/usr/bin/mipsel-linux-readelf -a ./target/bin/busybox
[...]
[Requesting program interpreter: /lib/ld-musl-mipsel-sf.so.1]
[...]
gcc 5.x doesn't need any fix because the musl patches already use the
right value.
Fixes bug #7562.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Building with -mtune=e6500 led to build failures in glibc (probably in
uclibc as well) because gcc was built for a 32-bit target even though
the target tuple is powerpc64-*. This lead to a mix of 32-bit and
64-bit support and build errors like:
fatal error: gnu/lib-names-32.h: No such file or directory
The root cause is that the configure script is not handling e6500
correctly, because of stupid typo in the condition.
Change has been submitted upstream.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Alvaro Gamez <alvaro.gamez@hazent.com>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that we are always explicitly passing gcc_cv_libc_provides_ssp,
there is no longer any reason to modify the gcc configure/configure.ac
to take into account the musl case. When a musl toolchain is being
built, BR2_TOOLCHAIN_HAS_SSP is always 'y', and therefore
gcc_cv_libc_provides_ssp=yes is always passed when building
gcc-initial and gcc-final.
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>
With support from assembler this option allows compiling huge functions,
where single literal pool at the beginning of a function may not be
reachable by L32R instructions at its end.
Currently assembler --auto-litpools option cannot deal with literals
used from multiple locations separated by more than 256 KBytes of code.
Don't turn constants into literals, instead use MOVI instruction to load
them into registers and let the assembler turn them into literals as
necessary.
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Now that the calls to gcc always pass through the toolchain wrapper, it
is no longer necessary to patch gcc to support poisoning.
This does have the disadvantage that there is no unsafe path check for
libc, libgcc and libstdc++ (all of these are built before the wrapper
exists). But we can assume that the toolchain components themselves
should be pretty safe.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>