[Buildroot] [PATCH v2 3/6] toolchain: introduce BR2_TOOLCHAIN_HAS_BINUTILS_BUG_27100

[Giulio Benetti]

And since only openal package seemed to be "affected by this bug", 
please drop this patch.

Thank you!
Kind regards
-- 
Giulio Benetti
Benetti Engineering sas

On 3/17/21 5:57 PM, Giulio Benetti wrote:
> Hi Thomas,
> 
> On 3/14/21 11:05 PM, Thomas Petazzoni wrote:
>> On Wed, 10 Mar 2021 10:14:50 +0100
>> Giulio Benetti <giulio.benetti at benettiengineering.com> wrote:
>>
>>> On OpenRisc binutils it still present ld bug 27100 leading to a package
>>> openal to fail building:
>>> http://autobuild.buildroot.net/results/c05/c053b9e191b31f0b1f8f8c902b602e187f714dc0/
>>>
>>> The bug was already reported and it's been updated:
>>> https://sourceware.org/bugzilla/show_bug.cgi?id=27100
>>
>> How do you know it is this bug? The "final link failed: bad value"
>> message is pretty generic.
> 
> You're right, that is not the entire message, it's:
> '''
> /home/buildroot/autobuild/instance-0/output-1/host/opt/ext-toolchain/bin/../lib/gcc/or1k-buildroot-linux-uclibc/9.3.0/../../../../or1k-buildroot-linux-uclibc/bin/ld:
> CMakeFiles/OpenAL.dir/al/source.cpp.o: pc-relative relocation against
> dynamic symbol alSourcePlayv
> /home/buildroot/autobuild/instance-0/output-1/host/opt/ext-toolchain/bin/../lib/gcc/or1k-buildroot-linux-uclibc/9.3.0/../../../../or1k-buildroot-linux-uclibc/bin/ld:
> final link failed: bad value
> '''
> 
> and that a package's issue. I think I've added by mistake openal failure
> as a test-case but it's not.
> 
>> Did you actually check that those issues no
>> longer appear with binutils 2.36 ?
>>> +# https://sourceware.org/bugzilla/show_bug.cgi?id=27100
>>> +# Affect toolchains built with binutils 2.33.1 (fixed in binutils 2.36)
>>> +config BR2_TOOLCHAIN_HAS_BINUTILS_BUG_27100
>>> +	bool
>>> +	default y if BR2_or1k && !BR2_BINUTILS_VERSION_2_36_X
>>
>> Obviously, an annoying problem is that the BR2_BINUTILS_VERSION_2_36_X
>> part of the condition does not take into account external toolchains.
>> So for external toolchains that have binutils >= 2.36, we still won't
>> be able to build those packages.
>>
>> But oh well, for something like OpenRISC, this is perhaps an acceptable
>> trade-off.
>>
>> Thomas
>>
>