interrupt flagged: intr# 19

Kens Alexandre

Other Parts Discussed in Thread: TMS320F28335

I'm getting the following error after starting my sys/bios application on a TMS320F28335:

gged interrupt flagged: intr# 19
xdc.runtime.Error.raise: terminating execution

Any idea what may be causing it? I saw one other post that talks about this problem, but the suggested solution (moving .stack to a different memory section) did not work for me.

I'm using SYS/BIOS 6_34_02_18, xdctools_3_24_05_48.

Thanks for any help.

Kens

over 12 years ago

0 judahvang over 12 years ago

TI__Mastermind 32475 points

Hi Kens,

I think this is the post you are talking about: http://e2e.ti.com/support/embedded/bios/f/355/p/202567/723067.aspx#723067

Seems like you are running into the same exact problem. Are your stacks in internal RAM?

Judah

0 Kens Alexandre over 12 years ago in reply to judahvang

Intellectual 295 points

Hi Judah,

Thanks for replying to my post. Yes, all my stacks are in internal RAM. My system stack is in MSARAM ( stack base address = 0x60 stack length = 0x790), task stacks (.ebss:taskStackSection) are in L0-L3 SARAM (taskStackSection base address = 0x8000 taskStackSection length = 0x22c0).

Kens

0 judahvang over 12 years ago in reply to Kens Alexandre

TI__Mastermind 32475 points

I really don't have any good suggestions on this. To me it sounds like something isn't initialized properly. Have you checked your stacks to make sure they are getting blown?

Only suggestion I have would be to move things around to see if that makes any difference. Maybe post your .map file and .cmd and I can take a peek at it.

Judah

0 Kens Alexandre over 12 years ago in reply to judahvang

Intellectual 295 points

I hope there's something in here:

0844.CCSV5TMSC320F28335.zip

3817.App_ALIGN_s00000101_test.zip

Kens

0 Mark Lazarewicz over 12 years ago in reply to Kens Alexandre

Expert 1505 points

I has same error same processor I backed out my changes and it went away. been awhile I think I tried added something that used the heap

and had a heap size of zero as our policy is to not use any dynamic memory allocation. try bumping stack sizes and heap

0 Kens Alexandre over 12 years ago in reply to Mark Lazarewicz

Intellectual 295 points

Hi Mark,

I tried your suggestions and had no success. It's that time to start stripping stuff out. Anyway, thanks for the suggestions.

Kens

0 Kens Alexandre over 12 years ago in reply to Kens Alexandre

Intellectual 295 points

Problem solved! The problem was a result of memory corruption.

I had some semaphore code that I leveraged from a DSP/BIOS project that was using SEM_pend(&sem) and SEM_post(&sem), where sem is declared in my DSP/BIOS project <project>cfg.h as SEM_Obj sem. In SYS/BIOS, the variable sem is declared as SEM_Obj *sem in my <project>_p28FP.c file, so referencing the address of sem in SYS/BIOS is "bad". I changed the code to use SEM_pend(sem) and SEM_post(sem) without the ampersand.

Kens

Processors

Processors forum

interrupt flagged: intr# 19