[Buildroot] [PATCH 1/1] package/brotli: fix build on microblaze with -02

Arnout Vandecappelle arnout at mind.be
Wed May 15 11:39:37 UTC 2019



On 15/05/2019 12:42, Giulio Benetti wrote:
> 
> 
> Il 15/05/2019 11:50, Thomas Petazzoni ha scritto:
>> Hello,
>>
>> On Wed, 15 May 2019 11:42:12 +0200
>> Giulio Benetti <giulio.benetti at micronovasrl.com> wrote:
>>
>>>>> +# prevents from triggering GCC ICE
>>>>> +# A bug was reported to the gcc bug tracker:
>>>>> +# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68485
>>>>> +ifeq ($(BR2_microblaze),y)
>>>>> +BROTLI_CONF_OPTS += -DCMAKE_C_FLAGS="$(TARGET_CFLAGS) -O0"
>>>>> +endif
>>>>
>>>> This is not a great fix, but still better than not being able to build the
>>>> package at all, so I think it's okay to merge it. ¯\_(ツ)_/¯
>>>
>>> This is another package added to the list of Microblaze GCC6/7 bug 85180.
>>
>> Be careful: we have two separate issues. Gcc bug 85180 and gcc bug 68485.
>>
>> gcc bug 85180, which you were working on, causes the compiler to got in
>> an infinite loop
>>
>> gcc bug 68485, which this patch is about, causes the compiler to throw
>> an ICE (Internal Compiler Error). I.e the compiler aborts, it does not
>> enter in an infinite loop.
> 
> Right.
> 
>> It would be interesting to see if bug 68485 stills exists with gcc 8.x
>> or 9.x.

 IIUC it was actually introduced in GCC8.x, and we backported the patch that
broke it: 0892-microblaze-Revert.patch

 So, my idea of requiring GCC8 for microblaze may not work out for all failures.

 Regards,
 Arnout

> 
> Then I'm going to test both 8.x and 9.x right now for both bugs.
> 
> It seems hard to find a fix for GCC < 8.x and Microblaze is a niche.
> So IMHO it seems a waste of time now with all these packages that fail.
> 
> Best regards


More information about the buildroot mailing list