GPF when interruption occur in ring 3

[aigle]

Hello !

I am doing a little 64 bit kernel (so i am in long mode) and as said in the title I have a problem with interruptions after switching to ring 3. (But all is working before switch)

My gdt descriptors seems ok (null, kernel code, kernel data, user code, user data, tss). My tss is empty.
The switch is ok (with the fake stack and iretq), and the user program is an infinite loop that only bug when an interruption occur (the timer interruption or any other if i call int xx from ring 3 program).

Qemu is raising v=0xd and e=0x0000...so thats a general protection fault but without any supplement informations...

Sorry for no code snippet, i cant post it for the moment.

If you have an idea im taker !

Have a good day.

[Octocontrabass]

My tss is empty.

You need to at least set RSP0 in your TSS.

[MichaelPetch]

v=0xd and e=0x0000

*If* it was a segment related error, E=0000 means that the General Protection Fault occurred accessing descriptor index 0x0000 (a NULL descriptor) in the GDT.

Did you happen to look at what instruction (at EIP) was being executed when you got the fault? There *may* have been other hints as to what might be going wrong if you had given us a full dump of QEMU's processor state at the time the exception occurred.

Another recommendation might be to toss it in BOCHS which at times can be better at identifying low level problems. BOCHS will often throw warning information on the console that can offer hints at where a problem may lie.

Do you have your code available online like Github or other service?

[Gigasoft]

No, it means that the fault happened from some reason other than a segment check or IDT problem.