RFC: ARM GCC changes

[Sebastian Huber]

On 09/25/2012 01:47 PM, Ralf Corsepius wrote:
> On 09/25/2012 12:36 PM, Chris Johns wrote:
>> Sebastian Huber wrote:
>>> On 09/25/2012 12:04 PM, Ralf Corsepius wrote:
>>>> On 09/25/2012 10:54 AM, Sebastian Huber wrote:
>>>>> This is an updated patch for the GCC 4.8. It renames the target
>>>>> "arm-rtemseabi" to "arm-rtems" to bring the ARM tool chain back to the
>>>>> standard RTEMS target pattern "$ARCH-rtems".
>>>>
>>>> As I tried to tell you several times before, I consider this step
>>>>
>>>> a) to be silly, because it raises abiguities to the "arm-rtems*"
>>>> target tuple
>>>> and breaks the clear distingtion between "arm-rtems" and
>>>> "arm-rtemseabi*" we
>>>> currently have.
>>>
>>> We have only one ARM tool chain configuration in the current RTEMS
>>> development version and this is arm-rtemseabi. The ARM tool chain is the
>>> only exception to the $ARCH-rtems pattern. This is confusing and leads
>>> to special cases in general support scripts.
>
> This doesn't mean much more but your scripts to be lacking generality.

We should keep things simple and don't introduce artificial burdens.

>
>>> There is absolutely no
>>> reason to have two ARM tool chain configuration for new versions of
>>> RTEMS.
>
> I vehemently disagree:
>
> * arm-rtems4.11/gcc-4.8 would be eabi.

Yes, this is the purpose of the patch.

> * arm-rtems/gcc-4.7 would fail to build
> * arm-rtems4.11/gcc-4.7 would expect *-eabi
> * arm-rtems4.11/gcc-4.6 would be abi.

This patch targets GCC 4.8.  The patches for GCC 4.6 and 4.7 are part of a 
previous email.

>
> * arm-rtems* would apply conventions arm*eabi users do not expect.

RTEMS ARM users expect a functional tool chain with a standard name.  If a 
RTEMS ARM user has a problem with this change, he can express his concerns here.

>
>
>> +1
>>
>> Please make the change and remove the ARM ABI mess we have. Sebastian,
>> thanks for the patch.
> There is no mess, IMO, this all is bikesheding for no technical reasons but for
> personal reasons.
>
> All this patch would do is to replace a clear separation with a mess and cause
> further mess. Of course you (Chris) will not experience much of the mess
> Sebastian is causing, because you are not packing the toolchains.

Yes, there is a mess, but only because you ignore the advice from the GCC ARM 
maintainer and me.

There is no need for a separation, because only the EABI variant is useful for 
future RTEMS versions.  Why is this so difficult for you to accept?  The 
transition to the EABI variant was a process of several years and RTEMS was one 
of the last users that stuck with the totally obsolete and unmaintained variant.

-- 
Sebastian Huber, embedded brains GmbH

Address : Obere Lagerstr. 30, D-82178 Puchheim, Germany
Phone   : +49 89 18 90 80 79-6
Fax     : +49 89 18 90 80 79-9
E-Mail  : sebastian.huber at embedded-brains.de
PGP     : Public key available on request.

Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.