Re: Newlib bug

[Corinna Vinschen <vinschen at redhat dot com>]

On Apr 22 12:00, Yogesh Wani wrote:
> I am very new to all these things so please bear with me and correct
> me if I am making any mistakes.
> 
> I have already reported an error in the crossgcc@sourceware.org
> mailing list. However i digged up a bit more on the git repository and
> found out the offending changelist which is causing me issues in
> building for gcc with newlib for arm7tdmi cpu is the following.
> 
> author           Corinna Vinschen <corinna@vinschen.de>
>                     Tue, 17 Feb 2015 10:26:53 +0100 (09:26 +0000)

That's an old CVS commit, the original author was Hale Wang.

Hale, do you have any idea?


Thanks,
Corinna


> committer      Corinna Vinschen <corinna@vinschen.de>
>                      Tue, 17 Feb 2015 10:26:53 +0100 (09:26 +0000
> commit          a23da62b7adc4a87378b850f45b41c1795beeb1f
> 
> tree               467780b260ba0a837e827905a298d95dab8f5b67
> parent            9e98e77b6fd19387dd8c3720b01710c0e51c1a1d
> 
> 
> I am trying to upgrade my existing toolchain for arm7tdmi with the
> following packages.
> 
> binutils: binutils-2.25
> newlib:  newlib-2.2.0.20150323
> gcc:      gcc-4.9.2
> gmp:     gmp-6.0.0
> mpc:     mpc-1.0.3
> mpfr:     mpfr-3.1.2
> 
> NOTE: Earlier I had newlib version 1.20.0. and that was fine with a
> local fix in the libgloss/arm/redboot-crt0.S file.This was later fixed
> by the commit # 408b20f1cc9b9030d6a4f0896876824b80b838e7
> 
> and configuring GCC with
> 
> ~/armtoolchain/src/gcc-4.9.2/configure --target=arm-none-eabi
> --prefix=/home/yogesh/armtoolchain/binaries/ --enable-languages=c,c++
> --with-gnu-as --with-gnu-ld --with-newlib
> --with-gxx-include-dir=/home/yogesh/armtoolchain/binaries/arm-none-eabi/include
> --with-cpu=arm7tdmi --with-gmp=/home/yogesh/armtoolchain/binaries
> --with-mpfr=/home/yogesh/armtoolchain/binaries
> --with-mpc=/home/yogesh/armtoolchain/binaries
> 
> 
> The error that I get is as follows:
> /home/yogesh/armtoolchain/src/gcc-4.9.2/newlib/libc/machine/arm/aeabi_memset-arm.S:
> Assembler messages:
> /home/yogesh/armtoolchain/src/gcc-4.9.2/newlib/libc/machine/arm/aeabi_memset-arm.S:43:
> Error: selected processor does not support ARM mode `uxtb lr,r2'
> /home/yogesh/armtoolchain/src/gcc-4.9.2/newlib/libc/machine/arm/aeabi_memset-arm.S:58:
> Error: selected processor does not support ARM mode `uxtb lr,r2'
> /home/yogesh/armtoolchain/src/gcc-4.9.2/newlib/libc/machine/arm/aeabi_memset-arm.S:93:
> Error: selected processor does not support ARM mode `uxtbne r2,r2'
> make[9]: *** [lib_a-aeabi_memset-soft.o] Error 1
> make[9]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/fpu/newlib/libc/machine/arm'
> make[8]: *** [all-recursive] Error 1
> make[8]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/fpu/newlib/libc/machine'
> make[7]: *** [all-recursive] Error 1
> make[7]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/fpu/newlib/libc'
> make[6]: *** [all-recursive] Error 1
> make[6]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/fpu/newlib'
> make[5]: *** [all] Error 2
> make[5]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/fpu/newlib'
> make[4]: *** [multi-do] Error 1
> make[4]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/newlib'
> make[3]: *** [all-multi] Error 2
> make[3]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/newlib'
> make[2]: *** [all] Error 2
> make[2]: Leaving directory
> `/home/yogesh/armtoolchain/builds/gcc-4.9.2/arm-none-eabi/newlib'
> make[1]: *** [all-target-newlib] Error 2
> make[1]: Leaving directory `/home/yogesh/armtoolchain/builds/gcc-4.9.2'
> make: *** [all] Error 2
> 
> The issue is that I am building the toolchain for an older
> architecture and the uxtb/uxtbne instructions in the
> aeabi_memset-arm.S file does not do any architecture checks to see if
> the instruction is valid for architectures below armv6.
> 
> Is it possible that you can do a fix and commit the changes or suggest
> me some alternative so that I can proceed with my toolchain build.
> This is really urgent for me.
> 
> Thanks and Regards,
> Yogesh
> 
> 
> -- 
> Yogesh Wani
> Luck is a looser's excuse for winners position !

-- 
Corinna Vinschen
Cygwin Maintainer
Red Hat

Attachment: pgpYjBZFqDM_L.pgp
Description: PGP signature