SYSBIOS aborts

cobsonchael

Other Parts Discussed in Thread: SYSBIOS

I am running SYSBIOS on my custom 6748 board. i keep getting aborts seemingly randomly. i don't know where to look for them exactly.

my question is: How do you debug aborts? i get an abort message from SysMin but i have no idea what any of it means. here is an example from the latest abort:

[C674X_0] {module#39}: line 345: error {id:0xbaa0000, args:[0xc1f23538, 0x2c23768]}
xdc.runtime.Error.raise: terminating execution
{module#44}: line 97: error {id:0x92c0000, args:[0xc301a2f5, 0xc301a2f4]}
xdc.runtime.Error.raise: terminating execution

i looked for a wiki to explain this but i can't seem to find one. how do i debug this? when i looked at the ROV i get 2 tasks that are flagged red saying "invalid state" but none of the stacks, heaps, or system stack is overflowing. this happened several times now and it is not always the same tasks that this is occuring to.

over 13 years ago

0 Steven Connell over 13 years ago

TI__Mastermind 45025 points

Hi Mike,

Do you see any information printed out in the printf logs or raw logs views (e.g. tools -> RTA -> Raw Logs, etc.)?

You can find more information on how to handle these Errors in the XDC tools documentation. For example, on my machine it's here:

"C:\ti\xdctools_3_23_00_32\docs\xdctools.chm"

Once open, then navigate to the Error module:

Lastly, the SYS/BIOS FAQ wiki shows how to backtrace when you get an exception with a register dump in the terminal output:

http://processors.wiki.ti.com/index.php/SYS/BIOS_FAQs#4_Exception_Dump_Decoding_Using_the_CCS_Register_View

Steve

0 cobsonchael over 13 years ago in reply to Steven Connell

Mastermind 7930 points

well that looks like a fine way to catch an error in a block of code that may produce an error. the problem i am getting is that i don't know where the aborts are coming from. as far as i know there is something wrong with the way i have the NDK setup and it is messing things up. also i am not getting exceptions. i am looking for how to debug for the abort messages that dump out from the sysmin.

i also cannot find that document. i have 4 files and 3 folders in my doc folder of my xdc tools doc directory. icons, license, rel_notes_archive folders and tilogo.gif, titagline.gif, README.txt, and docs.zip for files. i looked at the doc they had at the website but all i can find is how to call the API, not how to debug with it.

0 Steven Connell over 13 years ago in reply to cobsonchael

TI__Mastermind 45025 points

Hi Mike,

In ROV, under the BIOS module, there should be a "Scan for errors" tab. Can you check that and see if it brings up any more information about the error messages you're seeing?

Can you try setting a breakpoint at this symbol?

xdc_runtime_Error_raiseX__F

It should hit once your abort happens. Then in CCS you should see the call stack trace to help you see where it came from.

Another option is to define an Error raise hook function (Error.raiseHook in your *.cfg file) and set a break point in there. You can use these functions within your hook to get information about the error:

The .chm file I pointed you to before is the same as the XDC tools API guide that you can find via the CCS help.

Steve

0 Eddie3909 over 13 years ago in reply to Steven Connell

Guru 10385 points

Hi Steve

CCS 4.2.5
CGT 7.2.6.
IPC 1.24.3.32
SYS/BIOS 6.33.5.46
XDCtools 3.23.2.47

I'm trying to do what you suggest and have entered the cfg code below

var Error = xdc.useModule('xdc.runtime.Error');
Error.raiseHook = "&ErrorMgr_tiSysbiosBiosErrorRaiseHook";

But I noticed that the cfg viewer for the "Error" module (selected XDC/Error tab) still shows the "Error Hook" as ti.sysbios.BIOS.errorRaiseHook.

Should it change?

Cheers2u

Processors

Processors forum

SYSBIOS aborts