STEP and CONTINUE in GDB REMOTE

STEP and CONTINUE in GDB REMOTE

[Jan Van Belle]

        Hello all,

        I'm struggling with the Remote Protocol (over TCP, but I guess
that doesn't matter). For the moment I can download my code
onto the (not yet existing --> virtual) target, read registers, write
registers, read code and write code.

So now I arrived for the big job: STEP (single or not) and CONTINUE. The
manual says you have many possibilities to
indicate the target has stopped. For the moment I use: $S05#xx , but
this is just a stub to satisfy GDB.

Can someone explain me (or point me to a good description) the syntax of
'all' the possible responses??
--> $Txxx:yyyy#zzz
--> $Wxxx#yyy
--> $Sxxx#yyy
--> ...

It would be a great help (for me and many others ;-) )

Kind regards,

Jan

Re: STEP and CONTINUE in GDB REMOTE

[Daniel Jacobowitz]

On Mon, Nov 04, 2002 at 03:24:03PM +0100, Jan Van Belle wrote:
>         Hello all,
> 
>         I'm struggling with the Remote Protocol (over TCP, but I guess
> that doesn't matter). For the moment I can download my code
> onto the (not yet existing --> virtual) target, read registers, write
> registers, read code and write code.
> 
> So now I arrived for the big job: STEP (single or not) and CONTINUE. The
> manual says you have many possibilities to
> indicate the target has stopped. For the moment I use: $S05#xx , but
> this is just a stub to satisfy GDB.
> 
> Can someone explain me (or point me to a good description) the syntax of
> 'all' the possible responses??
> --> $Txxx:yyyy#zzz
> --> $Wxxx#yyy
> --> $Sxxx#yyy
> --> ...
> 
> It would be a great help (for me and many others ;-) )

Have you read the chapter detailing the remote protocol, in the fine
manual?

http://sources.redhat.com/gdb/current/onlinedocs/gdb_32.html#SEC630

It's in D.3.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: STEP and CONTINUE in GDB REMOTE

[Jan Van Belle]

Daniel Jacobowitz wrote:

> On Mon, Nov 04, 2002 at 03:24:03PM +0100, Jan Van Belle wrote:
> >         Hello all,
> >
> >         I'm struggling with the Remote Protocol (over TCP, but I guess
> > that doesn't matter). For the moment I can download my code
> > onto the (not yet existing --> virtual) target, read registers, write
> > registers, read code and write code.
> >
> > So now I arrived for the big job: STEP (single or not) and CONTINUE. The
> > manual says you have many possibilities to
> > indicate the target has stopped. For the moment I use: $S05#xx , but
> > this is just a stub to satisfy GDB.
> >
> > Can someone explain me (or point me to a good description) the syntax of
> > 'all' the possible responses??
> > --> $Txxx:yyyy#zzz
> > --> $Wxxx#yyy
> > --> $Sxxx#yyy
> > --> ...
> >
> > It would be a great help (for me and many others ;-) )
>
> Have you read the chapter detailing the remote protocol, in the fine
> manual?
>
> http://sources.redhat.com/gdb/current/onlinedocs/gdb_32.html#SEC630
>
> It's in D.3.
>
> --
> Daniel Jacobowitz
> MontaVista Software                         Debian GNU/Linux Developer

Yes I have. I have a copy printed out on my desk. This document explains us
which possible responses there are,
but not when to send which one, and certainly not what the contents should be
for

TAAn...:r...;n...:r...;n...:r...;

I already did several Google's for this topic, but I guess I'll have to find
it in someone's private collection

Kind regards,

Jan

Re: STEP and CONTINUE in GDB REMOTE

[Daniel Jacobowitz]

On Mon, Nov 04, 2002 at 03:38:05PM +0100, Jan Van Belle wrote:
> Daniel Jacobowitz wrote:
> 
> > On Mon, Nov 04, 2002 at 03:24:03PM +0100, Jan Van Belle wrote:
> > >         Hello all,
> > >
> > >         I'm struggling with the Remote Protocol (over TCP, but I guess
> > > that doesn't matter). For the moment I can download my code
> > > onto the (not yet existing --> virtual) target, read registers, write
> > > registers, read code and write code.
> > >
> > > So now I arrived for the big job: STEP (single or not) and CONTINUE. The
> > > manual says you have many possibilities to
> > > indicate the target has stopped. For the moment I use: $S05#xx , but
> > > this is just a stub to satisfy GDB.
> > >
> > > Can someone explain me (or point me to a good description) the syntax of
> > > 'all' the possible responses??
> > > --> $Txxx:yyyy#zzz
> > > --> $Wxxx#yyy
> > > --> $Sxxx#yyy
> > > --> ...
> > >
> > > It would be a great help (for me and many others ;-) )
> >
> > Have you read the chapter detailing the remote protocol, in the fine
> > manual?
> >
> > http://sources.redhat.com/gdb/current/onlinedocs/gdb_32.html#SEC630
> >
> > It's in D.3.
> >
> > --
> > Daniel Jacobowitz
> > MontaVista Software                         Debian GNU/Linux Developer
> 
> Yes I have. I have a copy printed out on my desk. This document explains us
> which possible responses there are,
> but not when to send which one, and certainly not what the contents should be
> for
> 
> TAAn...:r...;n...:r...;n...:r...;
> 
> I already did several Google's for this topic, but I guess I'll have to find
> it in someone's private collection

The explanation of T isn't the clearest...

SAA: stop with signal AA
TAA: stop with signal AA, and return the values of registers that we
     know will be queried, or other information:

n... = register number (hex), r... = target byte ordered register
contents, size defined by REGISTER_RAW_SIZE;

n... = `thread', r... = thread process ID, this is a hex integer;

n... = (`watch' | `rwatch' | `awatch', r... = data address, this is a hex integer;

n... = other string not starting with valid hex digit. GDB should ignore this n...,
r... pair and go on to the next.

NAA is obsolete and Oxx is... frowned upon.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: STEP and CONTINUE in GDB REMOTE

[Jan Van Belle]

Daniel Jacobowitz wrote:

> >
> > Yes I have. I have a copy printed out on my desk. This document explains us
> > which possible responses there are,
> > but not when to send which one, and certainly not what the contents should be
> > for
> >
> > TAAn...:r...;n...:r...;n...:r...;
> >
> > I already did several Google's for this topic, but I guess I'll have to find
> > it in someone's private collection
>
> The explanation of T isn't the clearest...
>
> SAA: stop with signal AA
> TAA: stop with signal AA, and return the values of registers that we
>      know will be queried, or other information:
>
> n... = register number (hex), r... = target byte ordered register
> contents, size defined by REGISTER_RAW_SIZE;
>
> n... = `thread', r... = thread process ID, this is a hex integer;
>
> n... = (`watch' | `rwatch' | `awatch', r... = data address, this is a hex integer;
>
> n... = other string not starting with valid hex digit. GDB should ignore this n...,
> r... pair and go on to the next.
>
> NAA is obsolete and Oxx is... frowned upon.
>

Aha, now we're talking
Maybe this could be added in the manual (hint hint)

Thanks a lot !!

Regards,

Jan

Re: STEP and CONTINUE in GDB REMOTE

[Daniel Jacobowitz]

On Mon, Nov 04, 2002 at 03:54:02PM +0100, Jan Van Belle wrote:
> Daniel Jacobowitz wrote:
> 
> > >
> > > Yes I have. I have a copy printed out on my desk. This document explains us
> > > which possible responses there are,
> > > but not when to send which one, and certainly not what the contents should be
> > > for
> > >
> > > TAAn...:r...;n...:r...;n...:r...;
> > >
> > > I already did several Google's for this topic, but I guess I'll have to find
> > > it in someone's private collection
> >
> > The explanation of T isn't the clearest...
> >
> > SAA: stop with signal AA
> > TAA: stop with signal AA, and return the values of registers that we
> >      know will be queried, or other information:
> >
> > n... = register number (hex), r... = target byte ordered register
> > contents, size defined by REGISTER_RAW_SIZE;
> >
> > n... = `thread', r... = thread process ID, this is a hex integer;
> >
> > n... = (`watch' | `rwatch' | `awatch', r... = data address, this is a hex integer;
> >
> > n... = other string not starting with valid hex digit. GDB should ignore this n...,
> > r... pair and go on to the next.
> >
> > NAA is obsolete and Oxx is... frowned upon.
> >
> 
> Aha, now we're talking
> Maybe this could be added in the manual (hint hint)
> 
> Thanks a lot !!

Except for the "stop with" after TAA, I was quoting the manual. 
Perhaps you shold look at the URL I gave you?

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: STEP and CONTINUE in GDB REMOTE

[Andrew Cagney]

> TAAn...:r...;n...:r...;n...:r...;
>> 
>> I already did several Google's for this topic, but I guess I'll have to find
>> it in someone's private collection
> 
> 
> The explanation of T isn't the clearest...
> 
> SAA: stop with signal AA
> TAA: stop with signal AA, and return the values of registers that we
>      know will be queried, or other information:
> 
> n... = register number (hex), r... = target byte ordered register
> contents, size defined by REGISTER_RAW_SIZE;
> 
> n... = `thread', r... = thread process ID, this is a hex integer;
> 
> n... = (`watch' | `rwatch' | `awatch', r... = data address, this is a hex integer;
> 
> n... = other string not starting with valid hex digit. GDB should ignore this n...,
> r... pair and go on to the next.
> 
> NAA is obsolete and Oxx is... frowned upon.

Why not do the reply in texinfo?  That way it can be dropped into the 
manual :-)

Andrew