Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed the week of November 23.

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMS320F28374S: The eabi .bin generating issue

Part Number: TMS320F28374S

Hi BU Team,

 

My customer Haier is using our F28374S for their EI projects.

They met a issue that previous they can use tiobj2bin(code shown as below) in post build up steps to generate .bin but now they cannot.

"${CCE_INSTALL_ROOT}/utils/tiobj2bin/tiobj2bin" "${BuildArtifactFileName}" "${BuildArtifactFileBaseName}.bin" "${CG_TOOL_ROOT}/bin/ofd2000" "${CG_TOOL_ROOT}/bin/hex2000" "${CCE_INSTALL_ROOT}/utils/tiobj2bin/mkhex4bin" & "${CG_TOOL_HEX}" -i "${BuildArtifactFileName}" -o "${BuildArtifactFileBaseName}.hex" -order MS -romwidth 32 & "${CG_TOOL_HEX}" --ti_txt "${BuildArtifactFileName}" -o "${BuildArtifactFileBaseName}.txt" -order MS -romwidth 32

The console output is shown as below, showing that eabi are not supported

_cgi-bin_mmwebwx-bin_webwxgetmsgimg__&MsgID=3543751075574496153&skey=@crypt_ad543140_bd824022498379580e6c995e8dd1ac49&mmweb_appid=wx_webfilehelper.jpg

And there is no .bin generated in the project.

 

Then they tried to generate .bin by enabling C2000 HEX utility and choosing the output format to .bin could generate a wrong .bin(the device function did not same with the code they write)

Now I let them using hex2000 command in post build up shown as below:

"${CG_TOOL_ROOT}/bin/hex2000" -q -b -image -o "${BuildArtifactFileBaseName}.bin" "${BuildArtifactFileBaseName}.txt" "${BuildArtifactFileBaseName}.out"

 

And they need to make a roms.txt by themself, shown as below:

ROMS {

all_mem: o = 0x108000, l = 0x74000,fill=0xffff

} 

 

This way the can generate the .bin successfully but the .bin is bigger than their code really are, their code is 120~130k but the .bin is 474k with lots of FFFFF in the end of .bin.

 

Could you help check with this issue and give a solution to my customer?

Thanks for your help!!

 

B.R.

Jackey

  • Hi Jackey,

    The core issue appears to be related to EABI vs. COFF ABI format compatibility and how the post-build tools handle the output. Based on community discussions 1, there are known differences in how EABI-compiled output behaves compared to legacy COFF format, which can affect both tool compatibility and binary size.

    Regarding the oversized .bin file (474k vs. actual 120–130k code): This is typically caused by the ROMS configuration filling the entire memory range you've specified (0x108000 to 0x108000 + 0x74000) with the fill value 0xFFFF. The hex2000 tool is working as designed—it's padding the gap between your actual code and the end of the defined ROM region.

    Questions for customer

    1. Compiler and CCS version: What version of Code Composer Studio and the C2000 compiler (cgtools) is your customer currently using? The tiobj2bin tool's EABI support varies by toolchain version.

    2. Build configuration: Are they building with --abi=eabi explicitly set, or did this change recently? If they were previously on COFF ABI and recently migrated to EABI, that would explain why tiobj2bin suddenly rejects the output.

    3. Actual memory layout: In their linker command file (.cmd), what is the actual memory section defined for code? The ROMS configuration (0x108000 with length 0x74000) seems quite large—is that the entire flash region, or should it be smaller to match their actual code size?

    Thanks,

    Ira