Re: [PATCH] Create avr-linux-tdep.c (add reintroduce gdbarch_gdb_signal_{to,from}_target for that target)

[Pedro Alves <palves@redhat.com>]

On 08/12/2013 10:48 PM, Sergio Durigan Junior wrote:
> On Monday, August 12 2013, Joern Rennecke wrote:
> 
>> Quoting Sergio Durigan Junior <sergiodj@redhat.com>:
>>
>>
>>> +avr-*-*linux*)
>>> +	# Target: AVR Linux
>>> +	gdb_target_obs="avr-tdep.o avr-linux-tdep.o linux-tdep.o"
>>> +	gdb_sim=../sim/avr/libsim.a
>>> +	;;
>>> +
>>>  avr-*-*)
>>>  	# Target: AVR
>>>  	gdb_target_obs="avr-tdep.o"
>>>
>>
>> That is creating a new target.
> 
> Yes.
> 
>> So are users using a linux host supposed
>> to configure for this new target?  But why does it look like it's the
>> target that is running linux - there is no such thing for avr.
>> Shouldn't we rather autoconf the host signals?
> 
> These are not host signals, they are the target's.
> 
> Maybe I am confusing things here.  I saw
> <arch/avr32/include/uapi/asm/signal.h> and I assumed one could run Linux
> on AVR, but on a further investigation I saw that it only applies to
> AVR32, right?  I'm definitely not an expert here, sorry about it.
> 
> But well, if that is the case, then I guess we can either (a) treat
> AVR32 separately and make an avr32-linux-tdep.c for it, or (b) drop this
> patch entirely.

Looks like upstream GDB doesn't support AVR32 at all.  At least, I
can't find any hit for avr32 in the sources.  I looked around, and
I can't find GDB in Atmel's current toolchains (on their site).  I found
avr32-gdb-6.7.1.atmel.1.0.4.tar.gz at
http://distribute.atmel.no/tools/avr32/old/avr32gnutoolchain/v2.2.1/source/
though (note "old"), which does contain AVR32 GDB and GDBserver ports.
Guess this was never pushed upstream.  Sounds like we should drop
this patch then.

-- 
Pedro Alves