GDB macros file

[Joel Sherrill]

leonp at plris.com wrote:
> Hello, all.
>
> In July Joel published the new version of the GDB macros file.
> I tried to use it and found it really very useful and significantly improved. 
> Many thanks, Joel!
>
> Now, some issues I encountered...
>
> 1. First question is - I feel a bit uncomfortable with the command names so 
> long and therefore hard to remember and to type. May be it can be something 
> simpler? For example, instead of rtems_classic_message_queue - rtems_mq, 
> while for rtems_posix_semaphores - rtems_ps, etc...
>
>   
You can always add function definitions at the bottom which provide 
shorthand
names and pass all arguments to the longer name.

> 2. For my gdb 6.1.1 I have several problems with some macros:
> - rtems_classic_semaphores prints out one LIBIO semaphore and the gdb crashes;
>   
Can you post the structure dump for the semaphore causing this?  It 
should be something
like:

p/x *(Semaphore_Control *)(_Semaphore_Information.local_table[2])

> - rtems_classic_task 2 
> =====================================================================
>  # | Name |    ID    | Pri |  Ticks   | State
> ---+------+----------+-----+----------+------------------------------
>  2 | LOGR | 0a010002 | 250 |        1 | Wmsg 
>   

> BACKTRACE
> 13:37, 11 Jul 2006 (CDT)[[User:JoelSherrill|Dr. Joel]] 13:37, 11 Jul 2006 
> (CDT)
> Invalid cast.
> ^^^^^^^^^^_________________________________
> Is it normal?
>
>   

No. :)

Hmmm.. is the backtrace function implemented on the CPU you are using?  Look
at the source and see if it is using the right register for the 
stack/base pointer?

The only way I know to debug this is to add debug messages to the macros and
see which value is bogus. 
>   

> Again, a lot of thanks to Joel!
>