MPC5604B program interrupt being generated for stw

[pradeepmc]

Hi,

I was trying to add some assembly code in my project.
I am using the following Codewarrior 10.7 evaluation edition.
Target MCU is MPC5604B.
The following default settings are ON for the project:
c/c++ build -> settings -> PowerPC CPU -> Processor is set to ZEN (I cant find e200z0h core)
c/c++ build -> settings -> PowerPC CPU -> Compress for PowerPC VLE (zen)
c/c++ build -> settings -> PowerPC Compiler -> Processor -> Translate PPC ASM to VLE ASM (ZEN)

I have created a normal c project. And I have one assembly source file within the project.

When the below code is executed, a Program interrupt is hit (IVOR6).
Snippet of code in the source file.

# Saves context to task stack
.macro portSAVE_CONTEXT
    stw         r1,-0x98 (r1)       # store backchain    ----> code causing the interrupt
    addi      r1,r1, 0x98         # allocate stack
    stw     r0,  0x24 (r1)
    ....
.endm

.text
.align 4
vPortYield:
    portSAVE_CONTEXT

The disassembled code for the above pointed instruction is below. Though I am not quite sure how it is translated by Codewarrior.
267           portSAVE_CONTEXT
40001418:   se_stb r2,0(rsp)
4000141a:   dc.w 0xff68        -----> the actual instruction that is causing the interrupt.
4000141c:   e_lha rsp,152(rsp)
40001420:   se_stb r0,0(rsp)

I couldn't find the what the instruction "dc.w" means.
The following are the register when the interrupt was generated.

SRR0: 0x4000141a
ESR: 0x08000020  --- PIL = 1; VLEMI = 1

 Can somebody help me with the reason for the interrupt being generated?

Regards,

Pradeep M C

[martin_kovar]

Hello,

from your description, it is difficult to say, what should be the source of the issue. Could you please share some projects, which demonstrates the problem?

Regards,

Martin

[pradeepmc · Accepted solution]

Hello Martin,

I was able to resolve the issue.
The issue was related to the below setting and where the code was placed in the elf binary.

1. I set following  to ON for the project:
   c/c++ build -> settings -> PowerPC CPU -> Compress for PowerPC VLE (zen)

     2. Then did a clean of the project. (may be i missed this step earlier, before raising the query in the forum)

     3. Changed the instructions to vle.

macro portSAVE_CONTEXT
      e_stw     r1,-0x98 (r1)        /* store backchain     */
    e_add16i  r1,r1, 0x98       /* allocate stack     */
    e_stw      r0,  0x24 (r1)    /* Store r0 working register  */

Now the disassembler has the same vle instructions written by me and no interrupts were generated.

What exactly happened with this change?
The above code was placed in .text_vle section now in the working case.
In the issue case, it was placed in .text section and probably codewarrior converted my instructions to non-vle.
May be this could have been the reason for the dc.w instructions and correspondingly a program interrupt on it's execution.

Below is the snippet of the linker command file of my project for reference:

GROUP : {
        /* Section used for initialization code: __ppc_eabi_init.c,
        MPC56xx_HWInit.c, MPC56xx_init_*.c and the entry point (__startup).
        */
        .init : {}
        .init_vle (VLECODE) : {
            *(.init)
            *(.init_vle)
        }

        /* ISR handlers code. */
        .__exception_handlers  (VLECODE) : {}

        .text (TEXT) ALIGN(0x10) : {}
        .text_vle (VLECODE) ALIGN(0x10): {
            *(.text)                ------------------> code was placed here in issue case.
            *(.text_vle)            ------------------> code was placed here in working case.
        }

Regards,

Pradeep M C