[RFA] Make breakpoint condition detection trace conditional on remote_debug.

[RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Joel Brobecker]

Hello,

I am investigating a problem where a breakpoint with a condition
no longer stops the program when running the program through gdbserver.
One of the things I noticed is this message that keeps popping up in
the gdbserver output:

    % gdbserver :4444 a
    Process a created; pid = 10100
    Listening on port 4444
    Remote debugging from host 127.0.0.1
    Found breakpoint condition.
    Found breakpoint condition.
    Found breakpoint condition.

This looks like a debugging trace that should only be enable if
requested.

What do you guys think of this patch.  I'm a little under pressure,
so I didn't test the patch, but I will if it looks correct. I wasn't
sure whether I should have used debug_threads or remote_debug...

gdb/gdbserver/ChangeLog:

        * server.c (process_point_options): If a conditional expression
        is found, only print a message if remote_debug is nonzero.

Thanks,
-- 
Joel

---
 gdb/gdbserver/server.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/gdb/gdbserver/server.c b/gdb/gdbserver/server.c
index 586581c..3c97dbd 100644
--- a/gdb/gdbserver/server.c
+++ b/gdb/gdbserver/server.c
@@ -2908,7 +2908,8 @@ process_point_options (CORE_ADDR point_addr, char **packet)
 	{
 	  case 'X':
 	    /* Conditional expression.  */
-	    fprintf (stderr, "Found breakpoint condition.\n");
+	    if (remote_debug)
+	      fprintf (stderr, "Found breakpoint condition.\n");
 	    add_breakpoint_condition (point_addr, &dataptr);
 	    break;
 	  default:
-- 
1.7.1

Re: [RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Pedro Alves]

On 03/08/2012 06:42 PM, Joel Brobecker wrote:

> gdb/gdbserver/ChangeLog:
> 
>         * server.c (process_point_options): If a conditional expression
>         is found, only print a message if remote_debug is nonzero.


Yes, please.  Thanks.

-- 
Pedro Alves

Re: [RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Joel Brobecker]

> > gdb/gdbserver/ChangeLog:
> > 
> >         * server.c (process_point_options): If a conditional expression
> >         is found, only print a message if remote_debug is nonzero.
> 
> 
> Yes, please.  Thanks.

Thanks, Pedro. Tested on x86_64-linux, and now checked in.

-- 
Joel

Re: [RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Joel Brobecker]

> > > gdb/gdbserver/ChangeLog:
> > > 
> > >         * server.c (process_point_options): If a conditional expression
> > >         is found, only print a message if remote_debug is nonzero.
> > 
> > 
> > Yes, please.  Thanks.
> 
> Thanks, Pedro. Tested on x86_64-linux, and now checked in.

Actually, I initially pushed the wrong patch by accident:

    [RFA] ax-gdb: Do not treat enums and bools as integers.
    http://www.sourceware.org/ml/gdb-patches/2012-03/msg00287.html

Thanks to Pedro for letting me know.

I pushed the correct patch, now. Yeah. I think...

    http://sourceware.org/ml/gdb-cvs/2012-03/msg00140.html

-- 
Joel

Re: [RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Luis Gustavo]

On 03/08/2012 03:42 PM, Joel Brobecker wrote:
> Hello,
>
> I am investigating a problem where a breakpoint with a condition
> no longer stops the program when running the program through gdbserver.
> One of the things I noticed is this message that keeps popping up in
> the gdbserver output:
>
>      % gdbserver :4444 a
>      Process a created; pid = 10100
>      Listening on port 4444
>      Remote debugging from host 127.0.0.1
>      Found breakpoint condition.
>      Found breakpoint condition.
>      Found breakpoint condition.
>
> This looks like a debugging trace that should only be enable if
> requested.
>
> What do you guys think of this patch.  I'm a little under pressure,
> so I didn't test the patch, but I will if it looks correct. I wasn't
> sure whether I should have used debug_threads or remote_debug...
>
> gdb/gdbserver/ChangeLog:
>
>          * server.c (process_point_options): If a conditional expression
>          is found, only print a message if remote_debug is nonzero.
>
> Thanks,


Thanks for the fix, it is the right thing.

Do you have the testcase handy? I can do some digging...

Did the condition go to the target? How did it evaluate?

Luis

Re: [RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Joel Brobecker]

> Thanks for the fix, it is the right thing.

You are welcome.

> Do you have the testcase handy? I can do some digging...
> Did the condition go to the target? How did it evaluate?

Which testcase are you referring to?

-- 
Joel

Re: [RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Luis Gustavo]

On 03/08/2012 08:32 PM, Joel Brobecker wrote:
>> Thanks for the fix, it is the right thing.
>
> You are welcome.
>
>> Do you have the testcase handy? I can do some digging...
>> Did the condition go to the target? How did it evaluate?
>
> Which testcase are you referring to?
>

You mentioned a problem where a conditional breakpoint no longer stops 
where it should when using gdbserver. I'm assuming it is a regression we 
didn't catch with GDB's testsuite when testing the target-side condition 
patches.

Is this the case or is it something else?

Luis

Re: [RFA] Make breakpoint condition detection trace conditional on remote_debug.

[Joel Brobecker]

> You mentioned a problem where a conditional breakpoint no longer
> stops where it should when using gdbserver. I'm assuming it is a
> regression we didn't catch with GDB's testsuite when testing the
> target-side condition patches.
> 
> Is this the case or is it something else?

Aaaaah :-). That's:

    [RFA] ax-gdb: Do not treat enums and bools as integers.
    http://www.sourceware.org/ml/gdb-patches/2012-03/msg00287.html

The problem is easy to understand. But unfortunately, I am unable
to produce a testcase outside of the proprietary code that shows
the problem (http://www.sourceware.org/ml/gdb-patches/2012-03/msg00302.html).

But thanks for the offer!

-- 
Joel