Status of GNU tool chain for PowerPC

[sebastianhuber]

Our FAE left your company without out any words of warning, so I post this here.

Here is the list of open PRs for GNU Binutils and GCC as discussed on the embedded world:

https://sourceware.org/bugzilla/show_bug.cgi?id=19554

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69617

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69618

Please note that I put Edmar Wienskoski <edmar@freescale.com> on CC. There is no NXP contact in the Binutils/GCC MAINTAINERS file and the Freescale person gave no response. This basically shows again that the Freescale/NXP PowerPC GNU tool chain is unmaintained.

[nicolaerosia_rd]

Hello,

This question still has no answers after more than a year.

What is the state of the patches found in PowerPC Linux SDK 1703, with regards to gcc 4.9.3 ? Are the patches included upstream?

Is it safe to use upstream gcc ?

In SDK 1703 meta-freescale, I have the following patches:

0001.gcc.e6500-FSF-49x.patch
0002.gcc.builtin_isel-49x.patch
0003.gcc.widen_types-49x.patch
0004.gcc.extelim-v4-49x.patch
0005.gcc.extelim_vrp_kugan-v1-49x.patch
0006.gcc.opt-array-offset-49x.patch
0007.gcc.aeabi-49x.patch
0008.gcc.fix_regalloc_for_482.patch
0009.gcc.rm_slow_tests-47.patch
0010.gcc.fix_mingw32.patch
0011.gcc.no_power_builtins-48.patch
0012.gcc.ld_unaligned-460.patch
0013.gcc.local_unaligned_altivec.patch
0014.gcc.soft_float-470.patch
0015.gcc.case_values-48.patch
0016.gcc.fix_pr63854_pass_manager.patch
0017.gcc.builtin_isel_doc.patch
0018.gcc.experimental_move.patch
0019.gcc.e5500_mfocr.patch
0020.gcc.load_on_store_bypass-48x.patch
0021.gcc.fix_constvector.patch
0022.gcc.fix_pr63908_unwind_info.patch
0023.gcc.fix_pr60158_fixup_table.patch
0024.gcc.have-pre-modify-disp-support-49x.patch
0025.gcc.fix_ENGR00298583_dwarf-vector-reg_49x.patch
0026.gcc.fix_MTWX51605-memset-array-init_48.patch
0027.gcc.fix_altivec_constant_alignment-v2.patch
0028.gcc.fix_altivec_reload_gs8.patch
0029.gcc.fix_postfix_gimplifier.patch
0030.gcc.fix_adjust_address_cost.patch
0031.gcc.fix_adjust_sched_loopinv_cost.patch
0032.gcc.fix_e5500_mulli_pipeline.patch
0033.gcc.fix_e500mc_addi_pipeline.patch
0034.gcc.fix_ENGR00292364_debug_frame.patch
0035.gcc.fix_ENGR00215936_49x.patch
0036.gcc.enable_soft_multilib-49x.patch
0037.gcc.fix_49x-doc.patch
0038.gcc.fix_emulation_spec_48.patch
0039.gcc.create_maeabi.patch
0040.gcc.rm_e500v2_loops_48.patch
0041.gcc.fix_e5500-e6500-aeabi-multi-lib.patch
0042.gcc.fix_ivopts.patch
0043.gcc.sysroot_spec_only_linux.patch

./qoriq-ppc:

0019-64-bit-multilib-hack.patch
0049-Enable-SPE-AltiVec-generation-on-powepc-linux-target.patch
0066-cxxflags-for-build.patch

Best regards,

Nicolae Rosia

[sebastianhuber]

I guess these bugs are also present in your SDK. In case these bugs are not present in your SDK, then why didn't you fix them in the FSF upstream version?

Who maintains the NXP/PowerPC in the FSF upstream (Binutils/GCC) after the NXP take over of Freescale? NXP should select a maintainer and propose them to the FSF steering committee.

[yipingwang]

Hello Sebastian Huber,

So far we are still using binutils 2.25 and GCC 4.9.3, the above PRs are about binutils 2.27 and GCC 6.0. I have already reported these defects to Toolchain and Linux SDK production team, will fix them in the future Linux SDK release.

Have a great day,
Yiping

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------