[patch] Add ARMv6 and ARMv7 support

Hello,

here is a patch to add ARMv6 and ARMv7 (Cortex A8/A9) to
RTT. Although this port has only been
tested on a single processor architecture (ARM Cortex-A8, ARMv7 instruction
set), this port should support ARMv6 architecture and also multi-core SMP
architectures (ARM Cortex A9). I will have the opportunity to validate
the operation on a Cortex-A9 MPcore in the coming months.

I would have liked to incorporate support for older ARM architectures (ARMv5
and under), but unfortunately I had a lot of strange problems (assert,
exception, race conditions, etc.) with the arm-port branch
that is available in Peter's github account. I really suspect that there
is a problem with ARM atomic operations that can be foundin this branch.

Philippe

AttachmentSize
0001-Add-ARMv6-and-ARMv7-support.patch10.65 KB

[patch] Add ARMv6 and ARMv7 support

Here a second patch for the arm port that fix a name clashing with boost
thread.

Philippe

2011/4/6 Philippe Hamelin <philippe [dot] hamelin [..] ...>

> Hello,
>
> here is a patch to add ARMv6 and ARMv7 (Cortex A8/A9) to
> RTT. Although this port has only been
> tested on a single processor architecture (ARM Cortex-A8, ARMv7 instruction
> set), this port should support ARMv6 architecture and also multi-core SMP
> architectures (ARM Cortex A9). I will have the opportunity to validate
> the operation on a Cortex-A9 MPcore in the coming months.
>
> I would have liked
> to incorporate support for older ARM architectures (ARMv5 and
> under), but unfortunately I had a lot of strange problems (assert,
> exception, race conditions, etc.) with the arm-port branch
> that is available in Peter's github account. I really suspect that there
> is a problem with ARM atomic operations that can be foundin this branch.
>
> Philippe
>

[patch] Add ARMv6 and ARMv7 support

Hi,

Are those patches integrated somewhere ? I was wondering about arm
situation. The sources from 2.6 doesn't contain anything about arm.

2011/4/7 Philippe Hamelin <philippe [dot] hamelin [..] ...>

> Here a second patch for the arm port that fix a name clashing with boost
> thread.
>
> Philippe
>
>
> 2011/4/6 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>
>> Hello,
>>
>> here is a patch to add ARMv6 and ARMv7 (Cortex A8/A9) to
>> RTT. Although this port has only been
>> tested on a single processor architecture (ARM Cortex-A8, ARMv7 instruction
>> set), this port should support ARMv6 architecture and also multi-core SMP
>> architectures (ARM Cortex A9). I will have the opportunity to validate
>> the operation on a Cortex-A9 MPcore in the coming months.
>>
>> I would have liked
>> to incorporate support for older ARM architectures (ARMv5 and
>> under), but unfortunately I had a lot of strange problems (assert,
>> exception, race conditions, etc.) with the arm-port branch
>> that is available in Peter's github account. I really suspect that there
>> is a problem with ARM atomic operations that can be foundin this branch.
>>
>> Philippe
>>
>
>
> --
> Orocos-Dev mailing list
> Orocos-Dev [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-dev
>
>

[patch] Add ARMv6 and ARMv7 support

Hi Willy,

these patches were not integrated because the support for ARM architecture
is now implemented through the more generic GCC built-in atomic operations
(see RTT "oro_gcc" target).

Philippe

2013/10/14 Willy Lambert <lambert [dot] willy [..] ...>

> Hi,
>
> Are those patches integrated somewhere ? I was wondering about arm
> situation. The sources from 2.6 doesn't contain anything about arm.
>
>
> 2011/4/7 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>
>> Here a second patch for the arm port that fix a name clashing with boost
>> thread.
>>
>> Philippe
>>
>>
>> 2011/4/6 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>>
>>> Hello,
>>>
>>> here is a patch to add ARMv6 and ARMv7 (Cortex A8/A9) to
>>> RTT. Although this port has only been
>>> tested on a single processor architecture (ARM Cortex-A8, ARMv7 instruction
>>> set), this port should support ARMv6 architecture and also multi-core SMP
>>> architectures (ARM Cortex A9). I will have the opportunity to validate
>>> the operation on a Cortex-A9 MPcore in the coming months.
>>>
>>> I would have liked
>>> to incorporate support for older ARM architectures (ARMv5 and
>>> under), but unfortunately I had a lot of strange problems (assert,
>>> exception, race conditions, etc.) with the arm-port branch
>>> that is available in Peter's github account. I really suspect that there
>>> is a problem with ARM atomic operations that can be foundin this branch.
>>>
>>> Philippe
>>>
>>
>>
>> --
>> Orocos-Dev mailing list
>> Orocos-Dev [..] ...
>> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-dev
>>
>>
>

[patch] Add ARMv6 and ARMv7 support

Ok, thanks, so if I want to use an other compilator than gcc I have no way
to do this currently ?

2013/10/15 Philippe Hamelin <philippe [dot] hamelin [..] ...>

> Hi Willy,
>
> these patches were not integrated because the support for ARM architecture
> is now implemented through the more generic GCC built-in atomic operations
> (see RTT "oro_gcc" target).
>
> Philippe
>
>
> 2013/10/14 Willy Lambert <lambert [dot] willy [..] ...>
>
>> Hi,
>>
>> Are those patches integrated somewhere ? I was wondering about arm
>> situation. The sources from 2.6 doesn't contain anything about arm.
>>
>>
>> 2011/4/7 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>>
>>> Here a second patch for the arm port that fix a name clashing with
>>> boost thread.
>>>
>>> Philippe
>>>
>>>
>>> 2011/4/6 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>>>
>>>> Hello,
>>>>
>>>> here is a patch to add ARMv6 and ARMv7 (Cortex A8/A9) to
>>>> RTT. Although this port has only been
>>>> tested on a single processor architecture (ARM Cortex-A8, ARMv7 instruction
>>>> set), this port should support ARMv6 architecture and also multi-core SMP
>>>> architectures (ARM Cortex A9). I will have the opportunity to validate
>>>> the operation on a Cortex-A9 MPcore in the coming months.
>>>>
>>>> I would have liked
>>>> to incorporate support for older ARM architectures (ARMv5 and
>>>> under), but unfortunately I had a lot of strange problems (assert,
>>>> exception, race conditions, etc.) with the arm-port branch
>>>> that is available in Peter's github account. I really suspect that there
>>>> is a problem with ARM atomic operations that can be foundin this branch.
>>>>
>>>> Philippe
>>>>
>>>
>>>
>>> --
>>> Orocos-Dev mailing list
>>> Orocos-Dev [..] ...
>>> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-dev
>>>
>>>
>>
>

[patch] Add ARMv6 and ARMv7 support

Yes, you're correct.

2013/10/15 Willy Lambert <lambert [dot] willy [..] ...>

> Ok, thanks, so if I want to use an other compilator than gcc I have no way
> to do this currently ?
>
>
> 2013/10/15 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>
>> Hi Willy,
>>
>> these patches were not integrated because the support for ARM
>> architecture is now implemented through the more generic GCC built-in
>> atomic operations (see RTT "oro_gcc" target).
>>
>> Philippe
>>
>>
>> 2013/10/14 Willy Lambert <lambert [dot] willy [..] ...>
>>
>>> Hi,
>>>
>>> Are those patches integrated somewhere ? I was wondering about arm
>>> situation. The sources from 2.6 doesn't contain anything about arm.
>>>
>>>
>>> 2011/4/7 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>>>
>>>> Here a second patch for the arm port that fix a name clashing with
>>>> boost thread.
>>>>
>>>> Philippe
>>>>
>>>>
>>>> 2011/4/6 Philippe Hamelin <philippe [dot] hamelin [..] ...>
>>>>
>>>>> Hello,
>>>>>
>>>>> here is a patch to add ARMv6 and ARMv7 (Cortex A8/A9) to
>>>>> RTT. Although this port has only been
>>>>> tested on a single processor architecture (ARM Cortex-A8, ARMv7 instruction
>>>>> set), this port should support ARMv6 architecture and also multi-core SMP
>>>>> architectures (ARM Cortex A9). I will have the opportunity to validate
>>>>> the operation on a Cortex-A9 MPcore in the coming months.
>>>>>
>>>>> I would have liked
>>>>> to incorporate support for older ARM architectures (ARMv5 and
>>>>> under), but unfortunately I had a lot of strange problems (assert,
>>>>> exception, race conditions, etc.) with the arm-port branch
>>>>> that is available in Peter's github account. I really suspect that there
>>>>> is a problem with ARM atomic operations that can be foundin this branch.
>>>>>
>>>>> Philippe
>>>>>
>>>>
>>>>
>>>> --
>>>> Orocos-Dev mailing list
>>>> Orocos-Dev [..] ...
>>>> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-dev
>>>>
>>>>
>>>
>>
>