• Locked Locked
  • Replies 2 replies
  • Subscribers 119 subscribers
  • Views 3173 views
  • Users 0 members are here
2025 survey

We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.

Options
Related
How was your experience today?
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Corrupted value in register after interrupt on LPC2368

Uros Jerman
Note: This was originally posted on 21st July 2009 at http://forums.arm.com

Hello,

I've hit a wall guys...

My problem is as follows:

1. all the processes run normally
2. then external interrupt occures
3. I handle it with eint3 handler (external interrupt 3)
4. after returning form eint3 handler sometimes - not always -
r4 register containes corrupted value. Insted of containing value
which was in before the eint3 interrupt it contains - as it seems -
the value of stack pointer of user mode

Any ideas?

for clearifing my point
void main(void)
{
   while(1)
   {
     /*some code*/
     results_process(); --> eint3handler()
                                     {
                                       /*some code*/
                                platform_powerdown()-->application_powerdown();
                                       /*end of eint3 handler*/
                                     }
   /*register r4 gets corrupted sometimes*/
   }

}

I use RVCT 2.2
All the code is written in THUMB mode, except for eint3_handler which is in ARM mode.


best regards

uros
  • Note: This was originally posted on 21st July 2009 at http://forums.arm.com

    Hi,

    I've actually solved the problem. Sam, you are right, the problem is in a stack overflow.
    I have missed that there is a call to another function via function pointer in application_powerdown
    and that wrecks a havock to stack :(

    My IRQ stack starts at the location 0x40007fe0 and has been 256 bytes long (descending order)

    User stack started at the location 0x40007e80 and is 4k bytes long.

    The function called via function pointer uses more than 256 bytes. Irq SP stopped at 0x40007e28 :)


    and that was the reason of corrupted r4 register.


    thanks again

    uros
  • Arm Employee Badge
    Note: This was originally posted on 21st July 2009 at http://forums.arm.com

    Uros,

    C code on ARM will naturally preserve r4-r11 (adding __irq to the function definition for armcc will cause it to also preserve r0-r3 etc. also), so either there's an asm veneer at the IRQ handler that's doing something silly with r4, or the r4 corruption is happening indirectly (likely due to stack collision between the r13_irq and r13_usr).

    So, some questions:
    What code is at the IRQ handler location?
    How are the stack pointers for the different modes initialised?

    hth
    s.