Re: Ping: [PATCH]: Support new ColdFire variants

[Peter Barada <peter at the-baradas dot com>]

>>Has anyone had a chance to look at my patch to add the new ColdFire
>>variants to binutils and clean up the current ColdFire instruction sets?
>>  
>>
>Sorry - I have been swamped with other stuff. 
>
>Anyway I have now had a chance to look over this patch and I am going to 
>apply it with a few changes:

Thanks for looking it over.

>  * Fix up the formatting in several places.  (In particular you 
>occasionally forgot to leave a space between a function name and the 
>opening parenthesis of its arguments).

Guilty as charged :)

>  * Fix up an apparent typo in the change to s_restore() in 
>gas/config/tc-m68k.c:
>
>*************** s_save (ignore)
>*** 5709,5714 ****
>--- 5825,5831 ----
>    s->keep_locals = flag_keep_locals;
>    s->short_refs = flag_short_refs;
>    s->architecture = current_architecture;
>+   s->architecture = current_chip;
>    s->quick = m68k_quick;
>    s->rel32 = m68k_rel32;
>    s->listing = listing;
>
>   I assumed that you meant to store the 'chip' global variable in the 
>the 'chip' field of the save_opts structure.

Correct.

>  * (most importantly): I ran the GAS testsuite with the patched 
>toolchain - it generated lots of new failures.  The cause was the 
>function select_control_regs() which was now emitting the message 
>"architecture not yet selected: defaulting to 68020" for lots of the 
>tests.  Since it was not clear to me whether this message was actually 
>wrong (ie an ISA/chip had not been selected) I elected to change the 
>code so that the message is only generated if the --verbose switch has 
>been included on the command line.  You may like to change this and 
>arrange for 'current_chip' to be initialised by default.

I'll look at initializing current_chip and send in another patch.

-- 
Peter Barada
peter@the-baradas.com