Code Composer Studio v4 for MSP430 Beta available

Hi Andy,

Thanks for following up on this. I just now found out that if I click the Step Over button enough times then I will eventually get a PC arrow and all is well from that point. Once I get the PC arrow then the Source Step-over and Step-Into buttons become active after being grayed out and I can single step the source code from that point.

I was not looking at the disassembly window before, but now that I am looking at that it is obvious what is happening. It is halting in the middle of the assembly code between C source lines, and hopefully this will be another issue where I don't have something set right (again). Any clue as to what that might be? I've looked all over, but I may have missed a control button somewhere.

Thanks -

Dave

Dave,

I'm not sure what the issue is yet, but it does seem like a CCS problem.

I have a suspicion that I need to check out here. It might have something to do with how stepping is interacting with ISRs in your application.

Andy

Here are a couple more clues that might help:

1) I'm looping endlessly in a process that executes a fair amount of simple integer 32 and 16-bit math in my application. When it halts in the assembly code it always halts in div32u.asm or mult32.asm (I did not write the asm code).

2) I am using lots of timer and external interrupts, but removing external interrupts and/or clearing GIE has no effect. It still halts in the .asm math processes which are only executed in the main spin-loop.

3) This point is likely moot given that clearing GIE doesn't fix anything, but I will still mention that the ISR processes do not perform any  multiply or divide math functions, and there are no multi-dimensional arrays that might trigger a reentrant multiply operation from within an ISR.

I hope this helps -

Dave

Dave,

I really appreciate your help in tracking this down. 

I don't have the exact device you are using immediately available and I haven't been abel to see this issue on other similar devices. While I get an f2132, would you be able to try the attached nonsensical program to see if it reproduces the problem on your target?

Thanks,

Andy

The project you provided runs, halts, and steps just fine.The interrupt triggers the way I would expect, and I don't see any strange behavior.

Here's one that does not work:

void main(void)
{
  long l = 1;
 
  for(;;)
  {
    l *= 123123;
  }
}

Try running that and you might see what I am seeing. The program halts in the middle of the mult32.asm assembly code every time. If you click on the assembly step-over button enough times you will eventually get to the end of the multiply where the PC arrow shows up and the source step-into and source step-over buttons become active and work just fine. It looks like it happens with any 32-bit multiply or divide. Break points always work properly, it's only the Halt button that does not.

Regards,

Dave

Hello John,

I'm now about testing new CCSv.4, but I'm a little bit woried to connect my FET430UIF with it. As I read there is new firmware used by CCS which is incopatible with old one used by CCE and IAR (and Elprotronics programmer I think).

My apprehension is if eZ430-F2013 and eZ430-RF2500 USB sticks will work. I use them for some occasion. Espacially eZ430-F2013 is wery sensitive to upload new firmware (it leads to destroy it).

Regards

Jarda

Jaromir Subcik said:

I'm now about testing new CCSv.4, but I'm a little bit woried to connect my FET430UIF with it. As I read there is new firmware used by CCS which is incopatible with old one used by CCE and IAR (and Elprotronics programmer I think).

jarda - did you experience a problem with the new beta or are you just worried about the possibility of a problem?

Hello Jaromir

May be a bit of reassurance. During my tests I went back and forth 3 or 4 times between CCE V3 and CCS V4. Each time the tool reflashed the FET430UIF. I found no problem. I know that some people experienced problems. But this has not been my experience so far.

Hope this helps

Bruno

I'm just woried about it with bat experience with my first eZ430-F2013. You can go through yahoo MSP430 conference about one and half year ago - there was a big discussion about this problem and many members described similar problem.

Jarda

Bruno, I'm not affraid not working FET430UIF but eZ430 tools.

Jarda

Hello Jarda

I understand.

Bruno

Hello John

I steadily go through my old projects with CCS V4.0.

I am now porting a project that was developed for the F427 a year ago. On this project I have the following problem:

- When in the C/C++ perspective I can build the project no problem.

- After building the project if I try to go to the debugger (by clicking on the little bug in the toolbar, the IDE attempts to rebuild everything (why...?). At this point I receive many error messages "...gmake *** no rule to make target 'Instruments\ccsv4\tools\compiler\msp430\include\.... I don't understand why the tools won't build this time, they just did a minute ago! Is it possible that it does not process paths correctly? - Notice the path starts at "Instruments"... Probably missing the beginning.

- If I simply go to the debugger (by invoking "Launch TI Debugger"), the debugger opens, but won't connect to the target because it is setup for a MSP430F2252, and the real target is an F427. I have not found where I can configure the debugger for a specific target (other than in the project properties where it is correctly set...)

I can provide the project if it helps

Regards

Bruno

Bruno,

It is really odd that clicking the project build button works but clicking the bug button fails.  Clicking the bug button just calls project build on the active project before launching the debugger.  If you could send the project that may help to figure it out, to me it seems like one of the path variables is forgetting quotes and the space in Texas Instruments is messing thing up.

Some other info that may help.  The bug button has a couple of actions associated with it.  The first is debug active project, this is the default action.  Debug active project will look at what your active project is (the one that is noted as active in the C/C++ projects view), it will do an incremental build on it and then it will attempt to launch the debugger.  When launching the debugger it looks for the .ccxml file in your project.  The .ccxml file contains the information about which device you are using and which emulation tool is setup.  For MSP430 we generate this file at the end of the project wizard as we just assume you are using a USB connection, for other device families we can't make that assumption so people have to create the .ccxml file.

When you select the Launch TI debugger option we ignore your projects and try to launch your default configuration.  Basically it looks for a .ccxml file that is marked as default.  Not sure where it is picking up the F2252, do you have another project in your workspace that uses 2252?  For MSP430 we try to hide the fact that these .ccxml files are use because the target configs are simple (i.e. once you have picked your device we have enough info to just generate it).  However if you want to debug without projects then you do need to create one.  If you go to Target -> New Target Configuration, it will open the wizard for creating one.  This will place it in a default location global to your install.  Give it a name, then select the connection type (TI MSP430 USB1) then select your device (MSP430F427), then click save on the main toolbar.  To make sure this is setup as the default config go to View -> Target Configurations, expand the user defined node and you should see the name of the file that you just created, if t does not say [default] beside it then right click on and choose to make it the default config.  While we generally hide this stuff it is very useful if you want to create a config with Multiple FETs connected (i.e. USB1 and USB2), or if you are not working with projects and just want to create these configs so that you can quickly launch different configurations, to do this you just create them and then go to the target configuration view and right click on the config you want and select "Launch Selected Configuration".

John

Hello John

I do not know why the F2252. I have never worked with that target. The last project I tested was using F2272.

I can launch the debugger without specifying the project. I did create a new target configuration specifying the F427. I made it the default as you explained. However when I go to "connect to the target" I still get an error message specifying "The target setup (MSP430F2252) does not match the actual target type (MSP430F427)"

After going back and forth between the C/C++ and debug perspectives it finally recognized the target. I was able to load the program and run. At this point the code runs but does not do what it is supposed to do (probably crashes). But this is another problem altogether.

All zipped the project is a bit over 1MB. Where/how do you want me to send it?

Bruno can you send the zip to here: ccsv4beta@list.ti.com

Thanks,

John

Hello!

Could anybody tell me how to make CCS read a license file?

I have setup the wrong host ID. Then I have asked for a license change, and I got it for the right

host ID this time, but I don't know how to make CCS update its license.

Thanks for any hint!

Pascal

Pascal,

There are a couple ways to do it. 

1. Inside CCS to Help -> Licensing Options.  It lets you specify the path to your license file.
2. Just copy your license file into <installdir>\ccsv4\DebugServer\license.  That is where we read the files from by default.

Regards,

John

Hello John

I looked more closely at the error messages. It looks like the problem comes from spaces in the paths of the make file. Every time there is a space the tools seem to interpret that as the end of the command.

Regards

Bruno

Hello John

I am now encountering a problem with the code generation tools, in the same project that I sent earlier today:

The problem originates in the interrupt routine SAMP_ISR() inSD16.c. After compiling look into SD16.asm you will find the Samp_ISR interrupt routine. When the optimizer is set for level 2 (the default), the new code-gen tools create a routime called abproc0() (abstracted procedure 0). This procedure is called from within the interrupt routine, yet it ends with an RTI...! This RTI is executed before the ISR has had a chance to restore the context (R14 and R15). This causes the code to crash. If I only optimize at level 1, abproc0() is not created and the code is functional. Also, if I modify the code to make it less symetrical, but still optimize at level 2, this routine abproc0() is not created and the code remains functional.

I believe abproc0() should end with an RET at the very least.

Hope this helps. Don't hesitate if you have any question.

Regards

Bruno

Hello John!

Sorry for the delay.

Thanks, it works now!

Pascal

Hello John

With all the changes in the CCS IDE is there any chance of reviving the visual linker sometime in the future?

I was very disapointed when this tool got canned. I imagine there were others like me...

Regards

Bruno

Wow you used the Visual Linker, that is going back a few years!

We do not have any plans to bring it back.  It was a DSP tool feature and at the time we discontinued it we could not find many people using it.  Also with the switch to cache based architectures it wasn't as relevant anymore.  And for families like MSP430 we try to abstract you from having to deal with linker command files.

John 

Well, at the risk of sounding like a dinosaur I go way-backer (wayer-back...) than that  ;-)  I have worked on the C25 and C30 in my days!

Even on architectures such as the MSP430 where the system map is already defined there is a need to do fine placement more often that you might think. In these circumstances the linker command syntax is very powerful, but for detailed work very complex and not extremely intuitive...

At the time when it was discontinued all our customers were disapointed also (mainly because we had done all our examples using it, and it was so easy to use).

In any case if TI ever decides to revive it I know of a few takers.

Best regards

Bruno

Bruno Paillard said:

Each time the tool reflashed the FET430UIF. I found no problem.

Did you have to do anything in particular to make this happen? I am having lots of problems with CCS relating to halting and stepping as mentioned before which prevents me from using this tool, and I would like to go back to what (sort of) worked between CCE and the FET430UIF.

The CCE debugger is now acting up as well and reloading it has no effect. Is there any way at all to force the reloading of the FET430UIF firmware? Folks have been asking for this for a long time on the MSP430 user forums and there have been no answers that I know of.

Andy: Did you get any mileage out of the simple code snippet I sent before regarding the halt/single step? Can you duplicate the error?

Thanks -

Dave

Hello Dave

Upon debugging in CCS V4 the FET430UIF always managed to detect the firmware was not up to date and reflash it. Then upon debugging in CCE V3.0 it always detected it was not the correct version and reflashed it back.

I must be one of the lucky ones. I have never seen a problem.

One thing that may be done: There is a free tool made by ELPROTRONIC (www.elprotronic.com) for the FET 430UIF. This tool also reflashes the FET430UIF if it is not the correct version to work with their software. My guess is that it uses the same software modules as the TI tool. If everything else fails I can send you an old version of this tool. It will surely reflah the FET430UIF to an earlier version. After that, CCE should be able to upgrade the firmware.

Tell me if you want me it works. If not I can send you this old version.

Regards

Bruno

Bruno Paillard said:

There is a free tool made by ELPROTRONIC (www.elprotronic.com) for the FET 430UIF

That is a very useful tool, and it worked just fine. CCE was able to recognize and re-flash the FET430UIF. It's nice to have a tool like that when all you want to do is program a bunch of parts.

Thank you for that link!

Dave

You are most welcome!

I purchased CCS v3.3 Platinum with a 1 year support package 5 months ago which was DSP centric and did not support MSP430.  Do I have access to CCS v4 with MSP430 support now?

Yes you would receive a CCSv4 Platinum license as part of your 1 year subscription and it would include MSP430 support in addition to the DSP & ARM families supported by your current tools.

Regards,

John

I just bought CCE430 PRO 3.x. About when will I get new license for CCSv4? Should I sit tight  and watch the mailbox?

Will it ship by post outside US or by e-mail ?

I don't know how and where to register CCE430 PRO. Please give me a hint! Thx!