[SOLVED] Interrupts only firing once

[dgriffen]

I've been working on my OS and ran into a problem with the interrupts, interrupts from the pic only seem to fire once. I've made sure to send the reset signal to the PIC in the handler code so I'm not sure what else could be causing it.
The interrupt should be printed as 32 instead of 23, that's a known issue with my screen code and is not related to the problem.

Screenshot of the output

problem.png (5.41 KiB) Viewed 2281 times

code is located herehttps://github.com/dgriffen/sigmos/tree ... bdcd601579

[eryjus]

You are not sending EOI to the PIC. See 8259 PIC.

[dgriffen]

You are not sending EOI to the PIC. See 8259 PIC.

I send EOI in my interrupt handler code here https://github.com/dgriffen/sigmos/blob ... /handler.c

[FallenAvatar]

...

Looks to me like your kernel is doing EXACTLY what you told it to do, including disabling interrupts the second kmain returns. To be honest, I'm surprised you even get one interrupt in that time.

Also, your isr_wrapper stuff is all sorts of weird. You probably want pushad and popad unless you are writing a real-mode kernel...

- Monk

[dgriffen]

...

Looks to me like your kernel is doing EXACTLY what you told it to do, including disabling interrupts the second kmain returns. To be honest, I'm surprised you even get one interrupt in that time.

Also, your isr_wrapper stuff is all sorts of weird. You probably want pushad and popad unless you are writing a real-mode kernel...

- Monk

Oh yeah, I forgot I disabled interrupts when I returned from kmain. I tried using pushad and popad, but they were considered invalid commands by the assembler. perhaps gcc got rid of those commands?

[KemyLand]

...

Looks to me like your kernel is doing EXACTLY what you told it to do, including disabling interrupts the second kmain returns. To be honest, I'm surprised you even get one interrupt in that time.

Also, your isr_wrapper stuff is all sorts of weird. You probably want pushad and popad unless you are writing a real-mode kernel...

- Monk

Oh yeah, I forgot I disabled interrupts when I returned from kmain. I tried using pushad and popad, but they were considered invalid commands by the assembler. perhaps gcc got rid of those commands?

What assembler are you using? GNU AS recognizes this piece of code perfectly:

Code: Select all

myLabel:
  pushad
  popad

[dgriffen]

What assembler are you using? GNU AS recognizes this piece of code perfectly:

Code: Select all

myLabel:
  pushad
  popad

I am using GNU AS 2.24 compiled for i686-elf. After some research it looks like it uses pushal and popal.

[KemyLand]

What assembler are you using? GNU AS recognizes this piece of code perfectly:

Code: Select all

myLabel:
  pushad
  popad

I am using GNU AS 2.24 compiled for i686-elf. After some research it looks like it uses pushal and popal.

A sorry, sorry! I just wrote the code without thinking ! It's true, now I remember. What happens is simple, the instructions' names are "pusha" and "popa" for Intel. But prefixes specific to assemblers happen to exist. Both 'd' in NASM and 'l' in GAS mean 32-bits.