GCC version for RTEMS 6?

[Chris Johns]

On 25/5/2022 10:44 pm, Kinsey Moore wrote:
> On 5/24/2022 17:29, Chris Johns wrote:
>> On 24/5/2022 5:01 pm, Sebastian Huber wrote:
>>> On 24/05/2022 08:43, Sebastian Huber wrote:
>>>> On 24/05/2022 02:32, Chris Johns wrote:
>>>>> On 6/5/2022 4:16 pm, Sebastian Huber wrote:
>>>>>>> - aarch64 bsps
>>>>>>>
>>>>>>> These BSPs generated lots of warnings in the testsuite due to this line:
>>>>>>>
>>>>>>> https://git.rtems.org/rtems/tree/bsps/aarch64/include/bsp/start.h#n175
>>>>>>>
>>>>>>> The warning is about comparing arrays. I have not looked deeper.
>>>>>> This can be fixed by using RTEMS_OBFUSCATE_VARIABLE().
>>>>> Sorry, I do not know how to use this macro to remove the warning?
>>>> For this macro you need a stack variable.
>>>>
>>>> I guess the warning could be also removed using this compiler note:
>>>>
>>>> rtems/bsps/aarch64/include/bsp/start.h:175:28: note: use
>>>> '&bsp_vector_table_end[0] != &bsp_vector_table_size[0]' to compare the
>>>> addresses
>>>>
>> I have done this and it is quieter now.
>>
>>>> What is the purpose of AArch64_start_set_vector_base()? Do we need this at all
>>>> on the aarch64 target?
>>> Also why is this function exposed by <bsp.h>?
>> I do not know. I am only the messenger :)
>>
>> Kinsey?
> 
> It's definitely necessary for all AArch64 targets. It sets the active vector
> table so that RTEMS can intercept exceptions of all varieties at EL1. It could
> probably be folded into start.S at the appropriate location if that's what
> you're asking.

Thanks.

> As far as why it's exposed by bsp.h, that's likely incidental since I was
> following the design of the ARM code when I wrote this.

OK. I am preferring less and less inlines these days even in C++.

Chris