Forums

Loading program failure on TMS320LF2407A

Started by Ajay November 16, 2004
I just installed the latest version of Code Composer Studio Free Eval
tools.

I imported my project. It did the conversion from the 4.x version to
the CCS version and compiled perfectly.

However, whenever I try Load Program, I get the error.

----------
Load Failed. A section of your program falls into a memory region that
is not writable. Check your linker configuration and or memory map.
----------
This error didn't occur when I was running Code Composer 4.x.

The memory map is usually defined in the GEL file.  Does the message give 
you an offending address?  If so, check in the data sheet to make sure that 
you are writing to a valid address and check the memory map definition in 
the GEL file to make sure that you are defining that memory.

Maybe the old CC 4.x didn't check to make sure you were in the right range 
and that's why it didn't generate the error.  ?  Either that or else the 
memory map is wrong in CCS as I mentioned above.

Did you see this app note on migrating from CC 4.12 to CCS 2.0?

http://focus.ti.com/docs/apps/catalog/resources/appnoteabstract.jhtml?abstractName=spra830

Have you read this app note as well?  Perhaps you don't have a jumper on the 
eZdsp (if that's what you're using) in the right position or something:

http://focus.ti.com/docs/apps/catalog/resources/appnoteabstract.jhtml?abstractName=spra755a

Brad

"Ajay" <ajaydsouza@gmail.com> wrote in message 
news:1100583082.580128.28540@z14g2000cwz.googlegroups.com...
>I just installed the latest version of Code Composer Studio Free Eval > tools. > > I imported my project. It did the conversion from the 4.x version to > the CCS version and compiled perfectly. > > However, whenever I try Load Program, I get the error. > > ---------- > Load Failed. A section of your program falls into a memory region that > is not writable. Check your linker configuration and or memory map. > ---------- > This error didn't occur when I was running Code Composer 4.x. >
We are using a custom board. I checked the jumper settings and they are
perfect.

Incidently, I uninstalled CCS and switched back to CC4.12.
This time it compiled and loaded and executed perfectly, as before.

hence, this memory problem is only related to CCS.

My project uses files from the second link you posted, ofcourse
modified. It is not very complex. All I do is generate a PWM wave as of
now, so it isn't really large as well.

Is there some custom way of creating a Linker command file (*.cmd) for
CCS, which is different from CC4.12 ?

Ajay escreveu:
> We are using a custom board. I checked the jumper settings and they are > perfect. > > Incidently, I uninstalled CCS and switched back to CC4.12. > This time it compiled and loaded and executed perfectly, as before. > > hence, this memory problem is only related to CCS. > > My project uses files from the second link you posted, ofcourse > modified. It is not very complex. All I do is generate a PWM wave as of > now, so it isn't really large as well. > > Is there some custom way of creating a Linker command file (*.cmd) for > CCS, which is different from CC4.12 ? >
In CCS you have to set in the memory map if your memory is flash or RAM, and it would give such an error in case it's set as flash. Just look if your board is set in the MP mode and verify CCS memory map. You can even disable it for test purposes. Verify the status of the MP/MC bit in the SCSR registers as well.
I worked with this chip a few years ago and had a similar problem
changing over CCS versions. My memory may be wrong but one thing I
would suggest is to check that you are outputing a COFF1 type file. I
think the default was COFF2 in the version 2.x and that did not work
for my project. That setting is in build options, I don't have CCS
intalled at the moment so I cant confirm this.

I beleive it was just the programmer that could not read the new COFF
format, since then I use a different serial programmer, (that works on
XP and is faster). You might want to check it out.
http://tigroups.webcrossing.com/WebX?128@254.9hjEabhrdEn.0@.eedc2e2

I hope this helps, but as I say, my memory is vague since it was a
while ago. If this is not helpful then post a reply and I will check
out my notes on it in my old lab book.

Tony
Tony McKay wrote:
> I worked with this chip a few years ago and had a similar problem > changing over CCS versions. My memory may be wrong but one thing I > would suggest is to check that you are outputing a COFF1 type file. I > think the default was COFF2 in the version 2.x and that did not work > for my project. That setting is in build options, I don't have CCS > intalled at the moment so I cant confirm this. >
I will check on this.
> I beleive it was just the programmer that could not read the new COFF > format, since then I use a different serial programmer, (that works
on
> XP and is faster). You might want to check it out. > http://tigroups.webcrossing.com/WebX?128@254.9hjEabhrdEn.0@.eedc2e2 >
Your link above gives me a Missing Template error.
Where do I check for that I am outputting a COFF1 type file?