Cross-debugging interfaces known to still work?

Cross-debugging interfaces known to still work?

[Stan Shebs]

In digging through GDB sources trying to figure out which internal
APIs are OK for Apple's port (yes, yes, need to contribute :-) ),
I see that a number of the oddball-protocol remote interfaces
probably no longer work, and some, such as remote-mips.c, are going
unnoticed because they're not built into any config anymore. At the
same time, I'm not seeing very much mail traffic on anything besides
standard remote protocol, sim, and RDI for Arm.

So my question is, do we have any information about who is actually
using or testing any of these other targets, either with current or
older versions of GDBs?

Stan

Re: Cross-debugging interfaces known to still work?

[Russell Shaw]

Stan Shebs wrote:
> In digging through GDB sources trying to figure out which internal
> APIs are OK for Apple's port (yes, yes, need to contribute :-) ),
> I see that a number of the oddball-protocol remote interfaces
> probably no longer work, and some, such as remote-mips.c, are going
> unnoticed because they're not built into any config anymore. At the
> same time, I'm not seeing very much mail traffic on anything besides
> standard remote protocol, sim, and RDI for Arm.
> 
> So my question is, do we have any information about who is actually
> using or testing any of these other targets, either with current or
> older versions of GDBs?
> 
> Stan

The "odd" protocols are used for adapting gdb to control a hardware
ICE. I made one for an atmel avr ICE, but it's not maintained in the
gdb sources yet.

Re: Cross-debugging interfaces known to still work?

[Daniel Jacobowitz]

On Fri, Jun 24, 2005 at 09:59:14AM +1000, Russell Shaw wrote:
> Stan Shebs wrote:
> >In digging through GDB sources trying to figure out which internal
> >APIs are OK for Apple's port (yes, yes, need to contribute :-) ),
> >I see that a number of the oddball-protocol remote interfaces
> >probably no longer work, and some, such as remote-mips.c, are going
> >unnoticed because they're not built into any config anymore. At the
> >same time, I'm not seeing very much mail traffic on anything besides
> >standard remote protocol, sim, and RDI for Arm.
> >
> >So my question is, do we have any information about who is actually
> >using or testing any of these other targets, either with current or
> >older versions of GDBs?
> >
> >Stan
> 
> The "odd" protocols are used for adapting gdb to control a hardware
> ICE. I made one for an atmel avr ICE, but it's not maintained in the
> gdb sources yet.
> 

But most people who want to control an ICE do it in a sensibly
componentized way nowadays, and just use remote.c.

Stan, no, we don't have much information about this.  The RDI bits
don't really work, but people may be using them anyway.  Beyond that I
strongly suspect the rest of the remote-* interfaces are dead.

-- 
Daniel Jacobowitz
CodeSourcery, LLC

Re: Cross-debugging interfaces known to still work?

[Stan Shebs]

Russell Shaw wrote:

> Stan Shebs wrote:
>
>> In digging through GDB sources trying to figure out which internal
>> APIs are OK for Apple's port (yes, yes, need to contribute :-) ),
>> I see that a number of the oddball-protocol remote interfaces
>> probably no longer work, and some, such as remote-mips.c, are going
>> unnoticed because they're not built into any config anymore. At the
>> same time, I'm not seeing very much mail traffic on anything besides
>> standard remote protocol, sim, and RDI for Arm.
>>
>> So my question is, do we have any information about who is actually
>> using or testing any of these other targets, either with current or
>> older versions of GDBs?
>>
>> Stan
>
>
> The "odd" protocols are used for adapting gdb to control a hardware
> ICE. I made one for an atmel avr ICE, but it's not maintained in the
> gdb sources yet.
>
Several of the ones I wrote were actually to talk to boards with
existing ROM monitors, so there's a variety. Some of the boards
targeted seem to have vanished pretty thoroughly - the only
references Google can find are the mentions of them in the GDB
documentation(!).

Stan

Re: Cross-debugging interfaces known to still work?

[Russell Shaw]

Stan Shebs wrote:
> Russell Shaw wrote:
> 
>> Stan Shebs wrote:
>>
>>> In digging through GDB sources trying to figure out which internal
>>> APIs are OK for Apple's port (yes, yes, need to contribute :-) ),
>>> I see that a number of the oddball-protocol remote interfaces
>>> probably no longer work, and some, such as remote-mips.c, are going
>>> unnoticed because they're not built into any config anymore. At the
>>> same time, I'm not seeing very much mail traffic on anything besides
>>> standard remote protocol, sim, and RDI for Arm.
>>>
>>> So my question is, do we have any information about who is actually
>>> using or testing any of these other targets, either with current or
>>> older versions of GDBs?
>>>
>>> Stan
>>
>> The "odd" protocols are used for adapting gdb to control a hardware
>> ICE. I made one for an atmel avr ICE, but it's not maintained in the
>> gdb sources yet.
>>
> Several of the ones I wrote were actually to talk to boards with
> existing ROM monitors, so there's a variety. Some of the boards
> targeted seem to have vanished pretty thoroughly - the only
> references Google can find are the mentions of them in the GDB
> documentation(!).
> 
> Stan

One thing they're useful for is as a reference to write a new one, because
how to do them is not otherwise documented.