TMS320F28379D: DSLite issues - interrupted programming

Hello,

I assume UniFlash is being used. What exact version of UniFlash? Is the dslite being called directly or is the generated command-line package being used?

Piotr Romaniuk said:

Unfortunatelly, this process sometimes gets interrupted,

How often is "sometimes"?

Thanks

ki

Hi Ki,

DSLite version 12.7.0.3313
Linux command line package was generated from Uniflash.

Statistics: 15%

Regards,
Piotr Romaniuk 

Piotr Romaniuk said:

DSLite version 12.7.0.3313

What is the UniFlash version? I ask because we discovered some issues with the generated command line packages for multi-core environments. It was fixed in UniFlash 8.8.0

https://sir.ext.ti.com/jira/browse/EXT_EP-11930

Piotr Romaniuk said:

DSLite version 12.7.0.3313

I believe this is from UniFlash 8.7, which is affected by the bug mentioned above.

Though if the above bug is indeed the culprit, I would expect the failure rate to be higher than 15%. But it is still worth updating to the latest Uniflash 8.8.1.

Hi Ki,

I will check it tomorow and let you know.
I need to get access to first device where I generated command line package.

Regards,
Piotr Romaniuk

Please note that the issue is with the generated package from 8.7. You will need to regenerate the package using 8.8.

Hi Ki,

I used probably UniFlash 8.6.0.4688.

As you remember on my device there was issues in GUI - it was very slow and not usable.
How can I generate command line package without running UniFlash? For example by running some script is UniFlash filesystem tree.

Regards,

Piotr Romaniuk

Piotr Romaniuk said:

How can I generate command line package without running UniFlash?

You do need the UniFlash GUI to generate the package. There is no CLI to generate this.

Hi Ki,

can you send be the algorithm of creation command line package?

I would like to write a script that will do the same.

Regards,

Piotr Romaniuk

The process involved in generating the package is quite involved. I am not sure I would recommend this path. I will discuss with engineering however to see if there is another option.

Hi Ki,

I don't need a general solution but only for one microcontroller type and one JTAG emulator type (i.e. xds110).
Maybe this will simplify the process.

Regards,
Piotr Romaniuk

There is a command you can run from CLI that can pull all the required debugger files needed for a specified ccxml file. You would still need to add the additional files like out file and create the main dslite bat file for the package. If this works for you I will try to get the details on the command. 

Hi Ki,

Yes, it looks very promising. I am interested in details.

Regards,
Piotr Romaniuk

Hi Ki,

I updated uniflash to 8.8.1 and regenerated CLI package.

Unfortunately problem still exists:

Executing: /home/pi/uniflash_linux-8.8.1-xds110/ccs_base/DebugServer/bin/DSLite flash -c /mnt/programator/settings/dev-01/f28379d.ccxml -l /mnt/programator/settings/dev-01/generated.ufsettings -e -f -v "/mnt/programator/Bootloader_cpu01-part2.hex"
DSLite version 12.8.0.3522
Configuring Debugger (may take a few minutes on first launch)...
        Initializing Register Database...
        Initializing: IcePick_C_0
        Executing Startup Scripts: IcePick_C_0
        Initializing: C28xx_CPU1
        Executing Startup Scripts: C28xx_CPU1
        Initializing: CPU1_CLA1
        Executing Startup Scripts: CPU1_CLA1
        Initializing: C28xx_CPU2
        Executing Startup Scripts: C28xx_CPU2
        Initializing: CPU2_CLA1
        Executing Startup Scripts: CPU2_CLA1
Connecting...
C28xx_CPU1: GEL Output:
Memory Map Initialization Complete
info: C28xx_CPU1: If erase/program (E/P) operation is being done on one core, the other core should not execute from shared-RAM (SR) as they are used for the E/P code. Also, CPU1 will be halted to determine SR ownership for the CPU which will run the Flash Plugin code, after which CPU1 will be set to run its application. User code execution from SR could commence after both flash banks are programmed.
Loading Program: /mnt/programator/Bootloader_cpu01-part2.hex
        Preparing ...
        0 of 29386 at 0x80000
error: C28xx_CPU1: Flash Programmer: Error erasing Sector A. OTP checksum mismatch detected. Operation Cancelled (4).
error: C28xx_CPU1: File Loader: Memory write failed: Unknown error
        Finished

This happened after succesfull programming flash and OTP of cpu02. Chip was empty at the beginning.

Althrough it printed out "finished" it looks like DSLite process is hang [?]:

pi@lopi-config-tool-3160:/var/www/html/tmp-get$ ps -axu |grep xds
...         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
...      161770  199  0.8 493936 69464 ?        Sl   14:17  76:53 /home/pi/uniflash_linux-8.8.1-xds110/ccs_base/DebugServer/bin/DSLite flash -c /mnt/programator/settings/dev-01/f28379d.ccxml -l /mnt/programator/settings/dev-01/generated.ufsettings -e -f -v /mnt/programator/Bootloader_cpu01-part2.hex

Statistics: 1 error for 5 programmings.

Regards,

Piotr Romaniuk

I am not too familiar with that error. When I searched for it, I found a few threads:

https://e2e.ti.com/support/tools/code-composer-studio-group/ccs/f/code-composer-studio-forum/700911/ccs-tms320f28377s-c28xx_cpu1-flash-programmer-warning-the-configured-device-tms320f28377s-does-not-match-the-detected-device

https://e2e.ti.com/support/microcontrollers/c2000-microcontrollers-group/c2000/f/c2000-microcontrollers-forum/1372368/tms320f28377s-flash-err

The error is coming from the Flash Programmer. I will bring this thread to the attention of the device experts for further investigation.

Thank you Ki. This is very urgent for us.

Regards,

Piotr Romaniuk

Hi Piotr,

As mentioned in the above related thread, have you checked the voltage range for operating conditions when programming the device's flash (erase/program operations)?

Thanks and regards,

Charles

Hi Charles,

we will verify power supply.
Does f28379d require more current during OTP programming that for FLASH?

Regards,
Piotr Romaniuk

Hi Ki,

I developed own bootloader.
It only depends on DSS(JS) & DebugServer(JTAG) during 1st stage for downloading iniatial bootloader into RAM.
After that serial interface is used. In the 2nd stage there is only my software and f28~ hardware features.

Next week I will check if problems still exists on this solution.

Regards,
Piotr Romaniuk

PS
I suspect that there sometimes there is a problem with switching current core after OTP programming. This is hipothesis for now.
The tests with my solution allow for further conclussions.

Hi Piotr,

F28379D requires the same current programming for Flash and OTP.

Regards,

Charles