Re: Xtensa port

[Maxim Grigoriev <maxim@tensilica.com>]

Thanks for your comment, Michael,

I think I need to look closely at the "multi-arch" mechanism and see if 
it's adequate to Xtensa configurability.

-- Maxim


Michael Snyder wrote:
> On Fri, 2006-09-22 at 12:10 -0700, Maxim Grigoriev wrote:
>   
>> Michael,
>>
>> Thanks much for the quick response.
>>     
>>> At a quick first glance, it looks pretty good.
>>> Do you have a copyright assignment on file
>>>   
>>>       
>> Once in a while, Daniel Jacobowitz mentioned that he has all the 
>> necessary paper work for Xtensa port. I hope it still true.
>>     
>>> One question -- is there a reason for splitting xtensa-config.c 
>>> into a separate file?  Not necessarily a problem, but it seems
>>> to be mostly stuff that would traditionally go in the xxx-tdep.c
>>> file (such as the struct gdbarch_tdep object).
>>>   
>>>       
>> It's because Xtensa is all about configurability. Historically, 
>> Tensilica uses a special configuration called FSF for all our FSF 
>> submissions. It's been used for Xtensa GCC and BFD work. Now, I'm trying 
>> to use it for GDB. So when we have to deal with a different 
>> configuration of the Xtensa processor all we need is to update 
>> xtensa-config.c file.
>>     
>
> Ah, but that's what we have a thing called multi-arch for!
>
> Wouldn't you rather have a single gdb that could debug *any* 
> of your xtensa configurations (as opposed to having to build
> a separate gdb for each one)?
>
>
>