RE: [ECOS] building tests

RE: [ECOS] building tests

[Paul Kinzelman]

Thanks very much Gary and Jonathan!
Using 'continue' works with -nw!
However, the View->Console sees only gdb output, not the
program output. How do I view the program output (which
I do get with the -nw version)?

I can run, for instance, 'stress_threads', under -nw, and
I see a bunch of output from the program. But if I run it
under the gui, I see only download messages and breakpoint
messages.

TIA! - PaulK

-----Original Message-----
From: Jonathan Larmour [ mailto:jlarmour@redhat.co.uk ]
Sent: Wednesday, April 12, 2000 4:25 PM
To: Paul Kinzelman
Cc: 'ecos-discuss@sourceware.cygnus.com'; Robert Erskine
Subject: Re: [ECOS] building tests


Paul Kinzelman wrote:
> 
> Second, I can't seem to run the prebuilt 'thread_gdb.exe' under insight:
>      arm-elf-gdb thread_gdb.exe (without the '-nw')
> I select the target, click on 'run' and it downloads properly (as far
> as I can tell), and then just hangs with the little clock and a
> 'stop' sign instead of run forever.
> 
> If I start over, download, then look at registers, I see
reasonable-looking
> numbers in the various registers so insight would appear to be hooking
> up to the gdb stub properly. But it doesn't seem to start/run.

You have to choose View->Console to view the program output.

Robert, we should probably add this to the docs at some point.

Jifl
-- 
Red Hat, 35 Cambridge Place, Cambridge, UK. CB2 1NS  Tel: +44 (1223) 728762
"Plan to be spontaneous tomorrow."  ||  These opinions are all my own fault

Re: [ECOS] building tests

[Jonathan Larmour]

Paul Kinzelman wrote:
> 
> Thanks very much Gary and Jonathan!
> Using 'continue' works with -nw!
> However, the View->Console sees only gdb output, not the
> program output. How do I view the program output (which
> I do get with the -nw version)?
> 
> I can run, for instance, 'stress_threads', under -nw, and
> I see a bunch of output from the program. But if I run it
> under the gui, I see only download messages and breakpoint
> messages.

Are you sure you had the console window up *before* you started running the
program?

Jifl
-- 
Red Hat, 35 Cambridge Place, Cambridge, UK. CB2 1NS  Tel: +44 (1223) 728762
"Plan to be spontaneous tomorrow."  ||  These opinions are all my own fault

Re: [ECOS] Building tests

[sandeep]

> I am trying to build eCso tests for Leon, but (after a good deal of
> compiltation) I get the following error message:
>
> /opt/rtems/sparc-rtems/bin/ld: unrecognized option `--fatal-warnings'
> sparc-rtems-gcc -msoft-float -g -nostdlib -Wl,-static -Wl,--fatal-warnings -
> L/ecos-d/cygnus/leonecos/default_install/lib -Ttarget.ld -o
> /ecos-d/cygnus/leonecos/default_install/tests/infra/v2_0/tests/cxxsupp
> tests/cxxsupp.o
> /opt/rtems/sparc-rtems/bin/ld: use the --help option for usage information
>
> What is it? In configtool, Build->options there is no trece of the
> '--fatal-warnings'. Where does it comes from? Normal compilation works
> correctly and I get the "examples" running well, it is just tests that raise
> that error.

"fatal-warnings" comes from LDFLAGS in generated makefile for infra folder under
your build-tree. as evident from error message, your ld doesn't support this
flag ("ld --help" will confirm it)

you can --
- update LDFLAGS there in that makefile and proceed with current build-tree
- or make updation in your ecc file (used to generate buildtree) as shown below
and regenerate the tree and proceed as you did earlier

cdl_option CYGPKG_INFRA_LDFLAGS_ADD {
    # Flavor: data
    # No user value, uncomment the following line to provide one.
    user_value ""
    # value_source user
    # Default value: -Wl,--fatal-warnings
};

- or get the effect of previous step via Configtool

--
sandeep



---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.512 / Virus Database: 309 - Release Date: 8/19/2003


-- 
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] Building tests

[Michele Portolan]

I am trying to build eCso tests for Leon, but (after a good deal of
compiltation) I get the following error message:

/opt/rtems/sparc-rtems/bin/ld: unrecognized option `--fatal-warnings'
sparc-rtems-gcc -msoft-float -g -nostdlib -Wl,-static -Wl,--fatal-warnings -
L/ecos-d/cygnus/leonecos/default_install/lib -Ttarget.ld -o
/ecos-d/cygnus/leonecos/default_install/tests/infra/v2_0/tests/cxxsupp
tests/cxxsupp.o
/opt/rtems/sparc-rtems/bin/ld: use the --help option for usage information

What is it? In configtool, Build->options there is no trece of the
'--fatal-warnings'. Where does it comes from? Normal compilation works
correctly and I get the "examples" running well, it is just tests that raise
that error.

Michele Portolan


-- 
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] building tests

[Jonathan Larmour]

Paul Kinzelman wrote:
> 
> Yes, I did View->Console after starting Insight. I did:
> arm-elf-gdb dhrystone&
> View->Console
> File->target (setup)
> Run->Run (downloads, stops at breakpoint)
> Breakpoint message shows in console window with prompt.
> Then I tried 3 different things after this

Ah, I've tried this myself now and I can confirm this is a GDB/Insight bug,
and that I've found where the problem is. I will get it fixed for the next
Insight snapshot.

Jifl
-- 
Red Hat, 35 Cambridge Place, Cambridge, UK. CB2 1NS  Tel: +44 (1223) 728762
"Plan to be spontaneous tomorrow."  ||  These opinions are all my own fault

RE: [ECOS] building tests

[Paul Kinzelman]

Yes, I did View->Console after starting Insight. I did:
arm-elf-gdb dhrystone&
View->Console
File->target (setup)
Run->Run (downloads, stops at breakpoint)
Breakpoint message shows in console window with prompt.
Then I tried 3 different things after this

1) If I type 'c' to the console, it says
'Continuing' and never comes back. But if I wait
for how long the test takes, and then reset
and wakeup the board, all of a sudden I do get the
printout.

2) If I on the 'continue' button on the gui, the clock
shows up, and hangs forever. When I reset and wake up
the board, I get a box: "Program received signal SIGTRAP"
and no printout.

3) Just for yucks, I tried clicking 'Run->Run' again,
and it downloads again, then hangs forever (no
initial breakpoint). When I try reset and
wakeup, I again get the SIGTRAP.

-----Original Message-----
From: Jonathan Larmour [ mailto:jlarmour@redhat.co.uk ]
Sent: Thursday, April 13, 2000 9:20 AM
To: Paul Kinzelman
Cc: 'ecos-discuss@sourceware.cygnus.com'
Subject: Re: [ECOS] building tests


Paul Kinzelman wrote:
> 
> Thanks very much Gary and Jonathan!
> Using 'continue' works with -nw!
> However, the View->Console sees only gdb output, not the
> program output. How do I view the program output (which
> I do get with the -nw version)?
> 
> I can run, for instance, 'stress_threads', under -nw, and
> I see a bunch of output from the program. But if I run it
> under the gui, I see only download messages and breakpoint
> messages.

Are you sure you had the console window up *before* you started running the
program?

Jifl
-- 
Red Hat, 35 Cambridge Place, Cambridge, UK. CB2 1NS  Tel: +44 (1223) 728762
"Plan to be spontaneous tomorrow."  ||  These opinions are all my own fault

Re: [ECOS] building tests

[Travis Furrer]

Actually, I have always had this problem also (I have been
using the insight snapshots for several months now).  So I
just use the -nw option when I need to see the program's
output.

Even if you pull up the console right away and do exactly
the same thing you would from the -nw version, the program
output does not appear there.  I am using the arm-elf
version, if that makes a difference.


> Paul Kinzelman wrote:
> >
> > Thanks very much Gary and Jonathan!
> > Using 'continue' works with -nw!
> > However, the View->Console sees only gdb output, not the
> > program output. How do I view the program output (which
> > I do get with the -nw version)?
> >
> > I can run, for instance, 'stress_threads', under -nw, and
> > I see a bunch of output from the program. But if I run it
> > under the gui, I see only download messages and breakpoint
> > messages.
>
> Are you sure you had the console window up *before* you started
> running the program?
>
> Jifl
> --
> Red Hat, 35 Cambridge Place, Cambridge, UK. CB2 1NS  Tel: +44 (1223)
728762
> "Plan to be spontaneous tomorrow."  ||  These opinions are all my own
fault

______________________________________________
FREE Personalized Email at Mail.com
Sign up at http://www.mail.com/?sr=signup

Re: [ECOS] building tests

[Jonathan Larmour]

Paul Kinzelman wrote:
> 
> Second, I can't seem to run the prebuilt 'thread_gdb.exe' under insight:
>      arm-elf-gdb thread_gdb.exe (without the '-nw')
> I select the target, click on 'run' and it downloads properly (as far
> as I can tell), and then just hangs with the little clock and a
> 'stop' sign instead of run forever.
> 
> If I start over, download, then look at registers, I see reasonable-looking
> numbers in the various registers so insight would appear to be hooking
> up to the gdb stub properly. But it doesn't seem to start/run.

You have to choose View->Console to view the program output.

Robert, we should probably add this to the docs at some point.

Jifl
-- 
Red Hat, 35 Cambridge Place, Cambridge, UK. CB2 1NS  Tel: +44 (1223) 728762
"Plan to be spontaneous tomorrow."  ||  These opinions are all my own fault

RE: [ECOS] building tests

[Gary Thomas]

On 12-Apr-00 Paul Kinzelman wrote:
> I was able to make all the tests (for the Cirrus EDB7211 board) without
> errors, but when I try to download one and run any test:
>     arm-elf-gdb -nw <objname>
>     target...
>     load (shows it downloaded 4 sections OK, with start adr and size)
>     run
> I get:
>     The program being debugged has been started already

This is an anomoly with GDB.  Instead of using 'run' after 'load', you
should use 'continue' (just 'c' works fine).

I'm pretty sure this is in the Getting Started guides.

[ECOS] building tests

[Paul Kinzelman]

I'm building on a Linux RH6.1 PC
arm-elf-gdb --version gives 20000314
gcc is 2.95.2

I downloaded/flashed the gdb stub (not cygmon).

I was able to make all the tests (for the Cirrus EDB7211 board) without
errors, but when I try to download one and run any test:
    arm-elf-gdb -nw <objname>
    target...
    load (shows it downloaded 4 sections OK, with start adr and size)
    run
I get:
    The program being debugged has been started already
    Start it from the beginning? (y or n)
And if I say yes,  it says:
    Starting program: <objname>
    You can't do that without a process to debug

This happens also when I try to run 'thread_gdb' (one of the built tests).

But when I load/run the prebuilt 'thread_gdb.exe', it seems to run fine
(I get the OK printout) under gdb (not insight).

The 2 thread_gdb files are a different size, but a 'files xxxx' says the
same for both (ELF 32-bit LSB, ARM, verision 1, statically linked,
not stripped)

How do I get the built tests to run?

Second, I can't seem to run the prebuilt 'thread_gdb.exe' under insight:
     arm-elf-gdb thread_gdb.exe (without the '-nw')
I select the target, click on 'run' and it downloads properly (as far
as I can tell), and then just hangs with the little clock and a
'stop' sign instead of run forever.

If I start over, download, then look at registers, I see reasonable-looking
numbers in the various registers so insight would appear to be hooking
up to the gdb stub properly. But it doesn't seem to start/run.

TIA!

-PaulK