TMS320F280039C: ECSDA verify

I suspect something goes wrong between the call to XCALLOC and the call to calloc.  According to the last screenshot, the arguments to calloc are:

• num: 73730
• size: 542426

Those two numbers multiplied together are a value much larger than the address space of C28x.  How does that happen?

Thanks and regards,

-George

I donot know what happened, the XCALLOC is library function which directly point to calloc().

For the source file that calls XCALLOC, please follow the directions in the article How to Submit a Compiler Test Case.

Thanks and regards,

-George

I have generated  .pp file according to this article, but it seems like I could not send it to you by TI Forum.

Try zipping the file and then attach to this forum:

https://e2e.ti.com/support/site-support-group/site-support/f/site-support-forum/761619/faq-how-do-i-attach-a-file-or-link-to-my-post

ecc_util.zip

I have upload .pp file.

1.  Also, the ECDSA library which I integrated is from C2000Ware_5_00_00_00, is this library tested and verified for use?

Another topic in E2E is related to ECDSA too.

TMS320F28P550SJ: When migrate ECSDA Project from F280049 to F28P55x, sig_r_val and sig_s_val do not have any vaule? - C2000 microcontrollers forum - C2000︎ microcontrollers - TI E2E support forums

Thank you for submitting the preprocessed file.  Based on that, I can see nothing wrong.

Please perform this debug experiment.  Inspect the CPU state at the very beginning of calloc, when it is called from the function init_c.  Confirm you see these values:

• register ACC = 1
• memory location *-SP[4] = 6

If you don't see those values, what do you see instead?

Thanks and regards,

-George

Thank you for answer.

1. At the very beginning of calloc, the value as follow, seems to be different from expectations.

2. When I run example of ecdsa in 280049, the length of sig_r_val  is 64 chars, and the length of sig_s_val is 64 chars too.

I am wonder about the length of signature R and signature S,  I always thought his length was 32，why the length is 64?

I apologize for failing to precisely describe what I mean when I say ...

George Mock said:

at the very beginning of calloc

Here is what I mean.  When calloc is called from the function init_c, and the next instruction to execute is the first one inside of calloc, inspect the CPU state as I describe in my previous post.  Please tell me what you see.  

Thanks and regards,

-George

Alena said:

When I run example of ecdsa in 280049, the length of sig_r_val  is 64 chars, and the length of sig_s_val is 64 chars too.

I am wonder about the length of signature R and signature S,  I always thought his length was 32，why the length is 64?

I presume the ecdsa example comes from C2000Ware.  C2000Ware is maintained by a different team, and I am unable to answer questions about it.  Please start a new thread on E2E. Or, if you prefer, you can wait for the calloc questions in this thread to resolve, then ask me to bring this thread to the attention of the team that supports C2000Ware.

Thanks and regards,

-George

Thanks for your answer.

As you can see, the value of ACC and *-SP[4] seems equal to what you expected.

George Mock said:

Confirm you see these values:

• register ACC = 1
• memory location *-SP[4] = 6

Thank you for your patience.

You have shown that, as of the point where calloc starts, everything is as expected.  I doubt the problem is within calloc.  It has worked correctly, with no changes, for a very long time.  Something else is wrong.  CCS might have a bug in displaying values.  Or some hardware error happens at this point.  Or some other similar problem.

I'm not sure what to suggest.  Try single stepping through calloc.  I am aware you don't know this code.  But if some hardware error happens, maybe you can catch it.  If you still cannot find the problem, I will ask for help from other experts who have experience debugging problems like this one.

Thanks and regards,

-George