[Clfs-support] Config question

[Andrew Bradford]

On Fri, Apr 22, 2011 at 3:16 PM, Andrew Bradford <bradfa at gmail.com> wrote:
> On Fri, Apr 22, 2011 at 12:42 PM, boedosoy <boedosoy at gmail.com> wrote:
>> 1) AT91RM9200 (ARM920)
>
> ARM920 is armv4t
>
>> 2) i.MX28 (ARM926)
>
> i.MX28 is a ARM926EJ-S core, which is armv5te

Sorry, upon review, I may not have completely answered your question.

For both processors, I'd recommend little endian, arm mode, and the
EABI (aapcs-linux).

armv4t most likely does not have hard floating point built in.  I'd
recommend using soft floating point.
armv5te may have a hard floating point unit, but I'm not that familiar
with Freescale's parts.  I believe hard floating point was an option
for armv5 cores.  You're safe doing a soft floating point build but if
you can find out from Freescale documentation if they included hard
floating point, use which ever version they have.

Pick your target triplets to be based on your other choices.  Use the
examples in the book as a reference, the examples are definitely not a
complete list of all available options.

-Andrew