VOGONS

Search

Common searches

x86 fault handler within own task stack?

Topic actions

First post, by superfury

Posted on 
User metadata
Rank l33t++
Rank
l33t++

What is the size of an error code pushed on the stack when handling a fault(e.g. Double fault) through a task gate in the IDT?

When does the CPU push 16 bits on the stack(descrementing (E)SP with 2)?

All 80386+ documentation only specifies '32-bit' handling for all protected mode examples, mentioning nothing about 16-bit protected mode (compatibility) and how it's handled on 32-bit processors? Such lack of documentation is extremely confusing.

Edit: Looking at Dosbox-X, it also seems to use the D/B-bit for that, but maybe it's actually bit 3 of the access rights(which also differentiates between a 80286 and 80386 TSS)?

Author of the UniPCemu emulator.
UniPCemu Git repository
UniPCemu for Android, Windows, PSP, Vita and Switch on itch.io

Reply 1 of 1, by superfury

Posted on 
User metadata
Rank l33t++
Rank
l33t++

It must use bit 3 of the access rights byte for that, since the D/B-bit(used by UniPCemu and even Dosbox-X) must be zeroed for TSS descriptors(according to the 80386 programmer's reference manual's description of a 32-bit TSS descriptor). Otherwise, any such TSS-fault handler would push the wrong operand size error code on the stack, even on 32-bit faults(which can't be correct).

Author of the UniPCemu emulator.
UniPCemu Git repository
UniPCemu for Android, Windows, PSP, Vita and Switch on itch.io

Go to top of page Go to top of page
Back to PC Emulation
Close window
Close window