TDA2P-ACD: Details on the register dump

Following up in different thread https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1399871/tda2p-acd-need-information-on-the-errorraisedhook-error-id-elaboration-and-reasons-caused-for-it/5364345#5364345

-Josue

Hello Josue,

Thanks for the documents , from the references the IRP we can see that the current execution when the crash happens it was pointing to the BIOS object

at (.text:ti_sysbios_knl_Swi_Instance_finalize__E) please provide what can lead to crash from this bios module.

Hello Satish,

Due to bandwidth, it will take me a day or two to get back to you.

Thank you for your patience.

-Josue

Satish,

Do you have the information from the sysErrHook? like module and file it is pointing to? What assert did it trigger?

Please review the following documentation: https://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/bios/sysbios/6_46_04_53/exports/bios_6_46_04_53/docs/cdoc/index.html

Read the "Details" for the module and constraints in particular for the "Instance Deletion" , it seems most likely related to 

Satish Hiremath said:

ti_sysbios_knl_Swi_Instance_finalize__E

Best,

-Josue

Satish,

C66 Register Dump analysis:

[2024-09-14 02:19:24.321]   efr=0x2 ------------------------------------ Exception Flag Register (EFR) =0010 Internal exception has been detected.
[2024-09-14 02:19:24.321]   nrp=0x90b38b44-------------------------- Nonmaskable Interrupt (NMI) Return Pointer Register (NRP) contains the return pointer that directs the CPU to the proper location to continue program execution after NMI processing
[2024-09-14 02:19:24.337]   ntsr=0x1020f ------------------------------ NMI/Exception Task State Register (NTSR) is used to store the contents of the task state register (TSR) and the conditions under which an exception occurred in the event of a nonmaskable interrupt (NMI) or an exception.
[2024-09-14 02:19:24.337]   ierB##=0x0--------------------------------- Internal Exception Report Register (IERR) contains flags that indicate the cause of the internal exception.  
[2024-09-14 02:19:24.337]   D eSPx1cC_oDntext=0x928abad8 ------- execution context? 
[2024-09-14 02:19:24.337] A0=0xE0LA AY1=0x0 ----------------------- A register file A0-A31 Data Path
[2024-09-14 02:19:24.337] A2=
[2024-09-14 02:19:24.337] 0x0 A3=0xc31eaa94
[2024-09-14 02:19:24.337] A4=0x40 A5=0x0
[2024-09-14 02:19:24.337] A6=0x1 A7=0x1c80
[2024-09-14 02:19:24.337] A8=0x21c A9=0x0
[2024-09-14 02:19:24.337] A10=0x2 A11=0xc31ea9a4
[2024-09-14 02:19:24.353] A12=0x40 A13=0x805400
[2024-09-14 02:19:24.353] A14=0x803208 A15=0x811400
[2024-09-14 02:19:24.353] A16=0x805408 A17=0x10
[2024-09-14 02:19:24.353] A18=0x805400 A19=0x0
[2024-09-14 02:19:24.353] A20=0x3b A21=0x3d
[2024-09-14 02:19:24.353] A22=0x0 A23=0x0
[2024-09-14 02:19:24.353] A24=0x0 A25=0xc32174e8
[2024-09-14 02:19:24.353] A26=0x261 A27=0x3f3
[2024-09-14 02:19:24.368] A28=0x151 A29=0x4c000
[2024-09-14 02:19:24.368] A30=0x1 A31=0xc0
[2024-09-14 02:19:24.368] B0=0x1 B1=0x1 ----------------------------- B register file B0-B31 Data Path
[2024-09-14 02:19:24.368] B2=0x0 B3=0x90b38b2c
[2024-09-14 02:19:24.368] B4=0x40 B5=0xc
[2024-09-14 02:19:24.368] B6=0x803208 B7=0x1
[2024-09-14 02:19:24.368] B8=0x811400 B9=0x202
[2024-09-14 02:19:24.368] B10=0x800000 B11=0x21c
[2024-09-14 02:19:24.368] B12=0x40 B13=0x2
[2024-09-14 02:19:24.384] B14=0x92a3bb00 B15=0xc314e108
[2024-09-14 02:19:24.384] B16=0x102 B17=0x201
[2024-09-14 02:19:24.384] B18=0x101 B19=0x18
[2024-09-14 02:19:24.384] B20=0x100 B21=0x0
[2024-09-14 02:19:24.384] B22=0x4 B23=0xff
[2024-09-14 02:19:24.384] B24=0x80 B25=0x80
[2024-09-14 02:19:24.384] B26=0xfc B27=0xc31ea9bc
[2024-09-14 02:19:24.384] B28=0xff B29=0xff
[2024-09-14 02:19:24.384] B30=0x18 B31=0x1
[2024-09-14 02:19:24.401] NTSR=0x1020f ------------------------------ NMI/Exception Task State Register (NTSR) is used to store the contents of the task state register (TSR) and the conditions under which an exception occurred in the event of a nonmaskable interrupt (NMI) or an exception.
[2024-09-14 02:19:24.401] ITSR=0x20f ---------------------------------- Interrupt Task State Register (ITSR) ) is used to store the contents of the task state register (TSR) in the event of an interrupt
[2024-09-14 02:19:24.401] IRP=0x90d22928 ---------------------------- Interrupt Return Pointer Register (IRP) contains the return pointer that directs the CPU to the proper location to continue program execution after processing a maskable interrupt
[2024-09-14 02:19:24.401] SSR=0x0 ------------------------------------- Saturation Status Register (SSR) provides saturation flags for each functional unit, making it possible for the program to distinguish between saturations caused by different instructions in the same execute packet
[2024-09-14 02:19:24.401] AMR=0x0 ------------------------------------ Addressing Mode Register (AMR) For each of the eight registers (A4-A7, B4-B7) that can perform linear or circular addressing, the addressing mode register (AMR) specifies the addressing mode
[2024-09-14 02:19:24.401] RILC=0x0 ------------------------------------- SPLOOP Reload Inner Loop Count Register (RILC) See link
[2024-09-14 02:19:24.401] ILC=0x0 --------------------------------------- SPLOOP Inner Loop Count Register (ILC) The SPLOOP or SPLOOPD instructions use the SPLOOP inner loop count register

I do not have much experience with C66 exception handling but the analysis above should help you perhaps gain some more insight. I am consulting internally for additional help.
Looks like some of the registers of interest would be NRP and IRP perhaps. IERR I believe would be helpful except it is 0 in this case.

Best

Josue

Hi Josue, 

Thanks for the clarification, but we are looking more for the program counter (PC) when the exception happened, can we extract the PC information from any of the registers provided in the exception data ?

Satish,

Please see the original response: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1401366/tda2p-acd-details-on-the-register-dump/5366786#5366786

Unfortunately PC cannot be deduced from this crash report as far as I know. The way to find this information is in the document I already share using the CCS tool:

Josue Zamitiz-Ayala said:

6.2 C66x DSP PC Trace

The fellow engineer that helped me also recommended watching the following video that uses the newest CCS version:
How to Debug an Application Crash Using Core Trace in CCS Theia (youtube.com)

And there's also  this one as well done by the team that used to support TI RTOS/SYSBIOS and is no longer with TI:

-Josue

Hi Josue , 

On A15 core   we have a exception  where we are noticing the stack pointer and program counter  address to be same , What are reasons for this behavior, 

Does the TI OS  have any protection to handle these cases ? 

Adding to the above question on the A15 Exception ,  DFSR = 0x00000221 ,  DFAR = 0x4a3d70dd

could also provide information of this what exception can lead to these faults