GNU Arm Embedded Toolchain 4.9-2014-q4-major

Release notes for
*************************************************
GNU Tools for ARM Embedded Processors 4.9
                               - Q4 2014
*************************************************

This release includes the following items:
* Bare metal EABI pre-built binaries for running on a Windows host
* Bare metal EABI pre-built binaries for running on a Linux host
* Bare metal EABI pre-built binaries for running on a Mac OS X host
* Source code package (together with build scripts and instructions to setup
  build environment), composed of:
  * gcc : ARM/embedded-4_9-branch revision 218278
    http://gcc.gnu.org/svn/gcc/branches/ARM/embedded-4_9-branch/

  * binutils : 2.24 with mainline backports
    git://sourceware.org/git/binutils.git commit 136a940ac535e464d2a7a86880ce1f1a5554c484

  * newlib and newlib-nano :
    git://sourceware.org/git/newlib.git commit a7ec2c7cea246a2c40357816a18f70732c7aee23

  * gdb : 7.8 with mainline backport/without target sim support
    git://sourceware.org/git/gdb.git commit bf9c7ab447198986c0dc96afc59dfd61d636ff15

  * cloog 0.18.0 :
    ftp://gcc.gnu.org/pub/gcc/infrastructure/cloog-0.18.0.tar.gz

  * expat 2.0.1 :
    http://jaist.dl.sourceforge.net/project/expat/expat/2.0.1/expat-2.0.1.tar.gz

  * gmp 4.3.2 : ftp://gcc.gnu.org/pub/gcc/infrastructure/gmp-4.3.2.tar.bz2

  * libelf 0.8.13 : http://www.mr511.de/software/libelf-0.8.13.tar.gz

  * libiconv 1.14 :
    http://ftp.gnu.org/gnu/libiconv/libiconv-1.14.tar.gz

  * mpc 0.8.1 : ftp://gcc.gnu.org/pub/gcc/infrastructure/mpc-0.8.1.tar.gz

  * mpfr 2.4.2 : ftp://gcc.gnu.org/pub/gcc/infrastructure/mpfr-2.4.2.tar.bz2

  * isl 0.11.1 : ftp://gcc.gnu.org/pub/gcc/infrastructure/isl-0.11.1.tar.bz2

  * zlib 1.2.8
    http://sourceforge.net/projects/libpng/files/zlib/1.2.8/zlib-1.2.8.tar.gz/download

Supported hosts:
* Windows 7 32/64 bits (with installer and alternative zip package)
* Linux 32/64 bits
  - Ubuntu 8.x or later (tarball)
  - Ubuntu LTS 10.04 and later (PPA)
  - RHEL 4/5/6 (tarball)
* Mac OS X 10.7.3 and newer 64 bits (tarball)

Supported target OS:
* Bare metal EABI only

Features:
* All GCC 4.9 features, plus latest mainline features
  - Cortex-M7 support
  - Small multiply support with options:
    . -mcpu=cortex-m1.small-multiply
    . -mcpu=cortex-m0.small-multiply
    . -mcpu=cortex-m0plus.small-multiply
* Define aeabi_memcpy and aeabi_memcpy4 routines in newlib.
* Thumb-1 library optimizations.
* Additional code size optimizations
* Add gdb python build (arm-none-eabi-gdb-py).
* Object file without attribute section can be linked with any
  others.

Tests:
* Tested on a variety of Cortex-M0/M0+/M3/M4/M7/A9 boards
* Tested on Qemu

Important bugs fixed in 4.9 major release:
* LTO ignores -fno-short-enums
  https://bugs.launchpad.net/gcc-arm-embedded/+bug/1315810
  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61123

* Incorrect folding of bitfield in a union on big endian targets
  https://bugs.launchpad.net/gcc-arm-embedded/+bug/1330388
  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62103

* newlib-nano printf("%.*s",moo) walks off the end of string
  https://bugs.launchpad.net/gcc-arm-embedded/+bug/1385829

Known Changes:
* We now use a gcc pass to implement the function of the plugin mentioned in
  http://gcc.gnu.org/ml/gcc-patches/2013-05/msg00667.html.
  Such pass only exists in embedded-4_9-branch branch and not in upstream
  mainline. A new option -ftree-switch-shortcut can be used to turn on/off
  this pass. By default it is off.

Known Issues:
* Thumb1 code size regression due to new register allocation:
  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61578
  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59535
  A workaround is to disable it by option -mno-lra.