[ECOS] IS eCos fake and GDB vaporware ?

[ECOS] IS eCos fake and GDB vaporware ?

[Jose Souto]

Hi,
I generated a floppy with RedBoot, booted an old Pentium 100 and RedBoot
showed it's welcome splash :o/

> ...
> RedBoot(tm) bootstrap and debug environment [FLOPPY]
> Non-certified release, version UNKNOWN - built 19:45:23, Apr 16 2002
> ...

After, running the eCos Config Tool and building everything I run the
tests (Ctrl-F5) and the tool cannot "teach gdb old tricks" like the
filename of the test program to debug (!)...

> ...
> //D/ECOS/TRY/TRY_IN~1/TESTS/HAL/COMMON/CURRENT/TESTS/BASIC.EXE: No such host or network path.
> (gdb) set height 0
> ...

My last resource is to run gdb from inside the Cygwin tools (yes, I'm
using a Window box).
Running BASH and starting i386-elf-gdb I get the mythical Insight tool
running. It pops up, connects to the target, download the program, show
memory, registers, everything - like if it were working.

But when I make the target run ... nobody else answers. Only the reset
key may open up their ears.

Is the floppy RedBood a fake stub ? 
Is gdb cheating (cozening) everybody for eons ?

Last question:
- does anybody "really" succeeded doing the above steps in any moment in
history ?
Please be honest !

Jose Luiz

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

[ECOS] I can't see vectors.s in the insight5.0?

[peter_ku]

Hello,


I use insight5.0 + redboot to debug my simple program.  my target is E7T.
My simple file is start.c
and I build the ecos by default. (I didn't change any setting)
gcc -g -IBASE/ecos_install/include/ start.c -LBASE/ecos_install/lib/ -Ttarget.ld -nostdlib

to generate a a.out

arm-elf-gdb a.out     to launch the insight

but when I start to use insight to debug, I found the program stop at main( ) (in main.cxx) , and main( ) call cyg_usr_start( )
then I try to set the breakpoint at reset_vector or start (in vectors.s) , I can't see the complete source code of vectors.s and I can't run vectors.s step by step.

How can I let my code run as follows, and I can see my vectors.s in the insight5.0.

start (or reset_vector) -> cyg_start -> cyg_user_start

other than run through main( )

thanks a lot.

Best Regards,
Peter

RE: [ECOS] IS eCos fake and GDB vaporware ?

[Geoff Patch]

At Last! A couple of questions I can respond to with 100% certainty that I'm
right.

And the answers are:

1. No.
2. No.

:-)

Sheesh...


Geoff


-----------------------------
Geoff Patch
Senior Software Engineer
CEA Technologies
Canberra Australia
02-6213-0141



-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] I can't see vectors.s in the insight5.0?

[Jonathan Larmour]

peter_ku wrote:
> 
> Hello,
> 
> I use insight5.0 + redboot to debug my simple program.  my target is E7T.
> My simple file is start.c
> and I build the ecos by default. (I didn't change any setting)
> gcc -g -IBASE/ecos_install/include/ start.c -LBASE/ecos_install/lib/ -Ttarget.ld -nostdlib
> 
> to generate a a.out
> 
> arm-elf-gdb a.out     to launch the insight
> 
> but when I start to use insight to debug, I found the program stop at main( ) (in main.cxx) , and main( ) call cyg_usr_start( )
> then I try to set the breakpoint at reset_vector or start (in vectors.s) , I can't see the complete source code of vectors.s and I can't run vectors.s step by step.

It's an assembly file. This is difficult. But you can sometimes get it to
work by recompiling vectors.S (see the eCos build output for how) and add
-Wa,--gstabs to the command line.

Jifl
-- 
Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062
--[ "You can complain because roses have thorns, or you ]--
--[  can rejoice because thorns have roses." -Lincoln   ]-- Opinions==mine

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss