The Microsoft code leak mentioned by one of the comments has been out there for years so might as well paste it here so cut down on some of the speculation? Fai

The Microsoft code leak mentioned by one of the comments has been out there for years so might as well paste it here so cut down on some of the speculation? Fair use - commercial value is zero, historical value for analysis and criticism is high. The relevant code comments seems to be "Fix timing problem??" and "486 bug - must wait till after last "out f0" to clear fp exceptions or IGNNE# will be permanently active." public __fpIRQ13 __fpIRQ13: cli WASTE_TIME 70 push ax xor al, al NULL_JMP out 0f0h, al ; reset busy line. NULL_JMP mov al, 65h NULL_JMP out 0a0h, al ; EOI slave irq 5 NULL_JMP mov al, 62h NULL_JMP out 20h, al ; EOI master irq 2 NULL_JMP pop ax sub sp, 2 push bp mov bp, sp fnstsw [bp+2] WASTE_TIME push ax xor al, al NULL_JMP out 0f0h, al ; reset busy line. NULL_JMP pop ax pop bp ; fnclex ; 486 bug - must wait till after last ; "out f0" to clear fp exceptions ; or IGNNE# will be permanently active. WASTE_TIME push ax xor al, al NULL_JMP out 0f0h, al ; reset busy line. NULL_JMP pop ax ; fnclex ; 486 bug - must wait till after last ; "out f0" to clear fp exceptions ; or IGNNE# will be permanently active. WASTE_TIME push ax xor al, al NULL_JMP out 0f0h, al ; reset busy line. NULL_JMP pop ax fnclex ;Now this is safe. WASTE_TIME 70 ;Fix timing problem?? jmp __FPEXCEPTION87P