Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
21 years ago

ISS on Nios II not starting at reset address

Hi there,

I am experiencing a problem with the ISS (Instruction Set Simulator) on a Nios-II design. Since my design would not even reach the first breakpoint in main(), I decided to use the 'trace' option in the ISS to see where the CPU wandered off.

To my surprise, this ISS-generated trace file (trace.out) shows that the Nios-II processor starts executing at address 0x00000008. Strange, since the reset address in my design is 0, and I understand that the reset-address of Nios-II is a non-vectored one.

So my question is: is this normal ISS behavior ?

Details below.

Bye from a puzzled Roland http://forum.niosforum.com/work2/style_emoticons/<#EMO_DIR#>/huh.gif

Info : PTF Setting uart_0/SYSTEM_BUILDER_INFO/Iss_Launch_Telnet="0" detected

Info : &#39;uart_0&#39; character stream will be displayed in this window

Info : The host communication device for stdin is uart_0

Info : The host communication device for stdout is uart_0

Info : The host communication device for stderr is uart_0

Info : Running &#39;rollator&#39; model

[1] 0x000000f8 <_start>: 0x06c00034 movhi sp, 0 [dstData=0x0 dstReg=sp]

Nios2 Register Contents

-----------------------------------

PC = 0x000000f8

r0 = 0x00000000 r1 = 0xdeadbeef r2 = 0xdeadbeef r3 = 0xdeadbeef

r4 = 0xdeadbeef r5 = 0xdeadbeef r6 = 0xdeadbeef r7 = 0xdeadbeef

r8 = 0xdeadbeef r9 = 0xdeadbeef r10 = 0xdeadbeef r11 = 0xdeadbeef

r12 = 0xdeadbeef r13 = 0xdeadbeef r14 = 0xdeadbeef r15 = 0xdeadbeef

r16 = 0xdeadbeef r17 = 0xdeadbeef r18 = 0xdeadbeef r19 = 0xdeadbeef

r20 = 0xdeadbeef r21 = 0xdeadbeef r22 = 0xdeadbeef r23 = 0xdeadbeef

r24 = 0xdeadbeef r25 = 0xdeadbeef gp = 0xdeadbeef sp = 0x00000000

fp = 0xdeadbeef ea = 0xdeadbeef ba = 0xdeadbeef ra = 0xdeadbeef

status = 0x00000000 estatus = 0x00000000 bstatus = 0x00000000 ienable = 0x00000000

ipending= 0x00000000

[2] 0x000000fc <_start+0x4>: 0xded40004 addi sp, sp, 20480 [dstData=0x5000 dstReg=sp]

Nios2 Register Contents

-----------------------------------

PC = 0x000000fc

r0 = 0x00000000 r1 = 0xdeadbeef r2 = 0xdeadbeef r3 = 0xdeadbeef

r4 = 0xdeadbeef r5 = 0xdeadbeef r6 = 0xdeadbeef r7 = 0xdeadbeef

r8 = 0xdeadbeef r9 = 0xdeadbeef r10 = 0xdeadbeef r11 = 0xdeadbeef

r12 = 0xdeadbeef r13 = 0xdeadbeef r14 = 0xdeadbeef r15 = 0xdeadbeef

r16 = 0xdeadbeef r17 = 0xdeadbeef r18 = 0xdeadbeef r19 = 0xdeadbeef

r20 = 0xdeadbeef r21 = 0xdeadbeef r22 = 0xdeadbeef r23 = 0xdeadbeef

r24 = 0xdeadbeef r25 = 0xdeadbeef gp = 0xdeadbeef sp = 0x00005000

fp = 0xdeadbeef ea = 0xdeadbeef ba = 0xdeadbeef ra = 0xdeadbeef

status = 0x00000000 estatus = 0x00000000 bstatus = 0x00000000 ienable = 0x00000000

ipending= 0x00000000

4 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    21 years ago

    I believe it&#39;s just skipping over the bootloader, which *is* at 0x0... Doesn&#39;t make much sense to emulate what the bootloader is doing, does it?

    Create an objdump and see where the _start symbol is located.

    - slackerman
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    21 years ago

    The ISS starts executing at the entry point specified in your .ELF file. This is set by the ENTRY directive in the linker script to the function _start:

    ENTRY( _start )

    The debugger starts at the same place when you download from the IDE.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    21 years ago

    Ok, this makes sense. Indeed the address I referred to (0x000000f8) corresponds to label _start. Of course running a program does not mean that you reset the system. Thank you Slacker and Wombat. I have one more question but will create a new topic on that, in order not to &#39;pollute&#39; this thread with multiple subjects.