ld segfaulting for crisv32

ld segfaulting for crisv32

[Waldemar Brodkorb]

Hi,

when building a uClibc toolchain for crisv32 ld is segfaulting.
I git bisected binutils and the faulty commit is:
Binutils 2.29.1 is the last workinng released version.

commit e01c16a83893b03433c8edf4ccae536d9b9f4831
Author: H.J. Lu <hjl.tools@gmail.com>
Date:   Sat Oct 14 11:09:29 2017 -0700

    cris: Check UNDEFWEAK_NO_DYNAMIC_RELOC
    
    Don't generate dynamic relocation against weak undefined symbol if it
    is resolved to zero.  FIXME: UNDEFWEAK_NO_DYNAMIC_RELOC may need to be
    checked in more places.
    
    bfd/
    
            PR ld/22269
            * elf32-cris.c (cris_elf_relocate_section): Don't generate
            dynamic relocation if UNDEFWEAK_NO_DYNAMIC_RELOC is true.
            (cris_elf_check_relocs): Don't allocate dynamic relocation
            if UNDEFWEAK_NO_DYNAMIC_RELOC is true.
    
    ld/
    
            PR ld/22269
            * testsuite/ld-cris/weakhiddso.d: Update and remove R_CRIS_NONE.

Anything can be done to fix that?

best regards
 Waldemar

Re: ld segfaulting for crisv32

[H.J. Lu]

On Mon, Sep 7, 2020 at 1:02 AM Waldemar Brodkorb <wbx@openadk.org> wrote:
>
> Hi,
>
> when building a uClibc toolchain for crisv32 ld is segfaulting.
> I git bisected binutils and the faulty commit is:
> Binutils 2.29.1 is the last workinng released version.
>
> commit e01c16a83893b03433c8edf4ccae536d9b9f4831
> Author: H.J. Lu <hjl.tools@gmail.com>
> Date:   Sat Oct 14 11:09:29 2017 -0700
>
>     cris: Check UNDEFWEAK_NO_DYNAMIC_RELOC
>
>     Don't generate dynamic relocation against weak undefined symbol if it
>     is resolved to zero.  FIXME: UNDEFWEAK_NO_DYNAMIC_RELOC may need to be
>     checked in more places.
>
>     bfd/
>
>             PR ld/22269
>             * elf32-cris.c (cris_elf_relocate_section): Don't generate
>             dynamic relocation if UNDEFWEAK_NO_DYNAMIC_RELOC is true.
>             (cris_elf_check_relocs): Don't allocate dynamic relocation
>             if UNDEFWEAK_NO_DYNAMIC_RELOC is true.
>
>     ld/
>
>             PR ld/22269
>             * testsuite/ld-cris/weakhiddso.d: Update and remove R_CRIS_NONE.
>
> Anything can be done to fix that?
>

Please open a binutils bug with a testcase.


-- 
H.J.

Re: ld segfaulting for crisv32

[Hans-Peter Nilsson]

> From: Waldemar Brodkorb <wbx@openadk.org>
> Date: Mon, 7 Sep 2020 10:01:47 +0200

> when building a uClibc toolchain for crisv32 ld is segfaulting.
> I git bisected binutils and the faulty commit is:
> Binutils 2.29.1 is the last workinng released version.

Thank you for the report!  I assume you refer to ld configured for the
crisv32-linux target (canonically crisv32-axis-linux-gnu).

Though, as H.J. mentioned, the preferred form for a bug-report is a
bugzilla entry.
See https://sourceware.org/bugzilla/show_bug.cgi?id=26578 for an
example of a recent entry.

> commit e01c16a83893b03433c8edf4ccae536d9b9f4831
[...]
>            * testsuite/ld-cris/weakhiddso.d: Update and remove R_CRIS_NONE.
> Anything can be done to fix that?

If you could help by constructing a reproducer, preferably in the form
of a test-case for the binutils test-suite (see e.g. the referenced
test-case) that now SEGV's but was supposed to work, and attach it to
the (new) bugzilla entry, that'd help a lot, likely up to the point of
whether the bug be fixed or not.

Also, please beware that crisv32-* (all subtargets, including
crisv32-linux, and all cris*-linux) are "obsoleted" in gcc-10 and
removed in gcc-11 (leaving just cris-elf).

brgds, H-P