Re: test386.asm CPU tester

Posted on 2017-10-20, 18:28

by

by peterferrie

in PC Emulation

in PC Emulation

Note that ARPL won't write to memory unless it has to alter the value. If the page were write-protected, it would be a test against always-write (cmpxchg does the same). That would be a good test. I broke QEMU many years ago like that. My emulator reached the arithlogic tests on the first attempt. …

---

Re: test386.asm CPU tester

Posted on 2017-10-19, 15:25

by

by peterferrie

in PC Emulation

in PC Emulation

This tester needs some of the corner cases. I shall make a PR for them. 😀

---

Re: Finding bugs in 8086-80386 emulation core?

Posted on 2017-10-17, 18:55

by

by peterferrie

in PC Emulation

in PC Emulation

That's exactly the difficult part: the method of execution is a cycle-based one, whereas most emulators(except mine and vladstamate's afaik) run on an instruction level at most(like Dosbox). So that would mean I need to rebuild some of the instruction-level hardware synchronization instead of cycle …

---

Re: Finding bugs in 8086-80386 emulation core?

Posted on 2017-10-14, 01:59

by

by peterferrie

in PC Emulation

in PC Emulation

Now you're in the realm of requiring an actual debugger or huge trace-logs, and performing side-by-side execution to see where the code paths deviate. Once you see where the paths separate, the problem might be obvious. If not, then we can certainly help to identify it.

---

Re: x86 real mode effects of loading a segment register?

Posted on 2017-10-06, 15:53

by

by peterferrie

in PC Emulation

in PC Emulation

Now that I think about it - HIMEM is using segment value 0 to avoid address shifting while memory in flat mode. So clearly the limits are not reloaded when the segment register is assigned, but I believe that they are if an interrupt occurs.

---

Re: x86 real mode effects of loading a segment register?

Posted on 2017-10-03, 19:15

by

by peterferrie

in PC Emulation

in PC Emulation

Loading 0 into them might be the special case that erases all content. I don't know. You might need to do some tests on hardware to see. For example, if you load a 4Gb segment in DS and then copy a 64kb ES to DS, does DS become limited again?

---

Re: x86 real mode effects of loading a segment register?

Posted on 2017-09-29, 20:21

by

by peterferrie

in PC Emulation

in PC Emulation

It's unclear (some of those bits would be hard to test), but I suppose that it is.

---

Re: x86 real mode effects of loading a segment register?

Posted on 2017-09-27, 23:52

by

by peterferrie

in PC Emulation

in PC Emulation

For real mode, the limit should be loaded, too. That's the basis of unreal mode, after all, and the loadall instruction can set it. Note that Loadall can set bitness and D bits, too, and they are honoured.

---

Re: Can x86 16-bit offsets overflow past 0xFFFF?

Posted on 2017-09-25, 16:43

by

by peterferrie

in PC Emulation

in PC Emulation

So the 80286 won't always perform a word prefetch? Page 3-31, 2nd to last paragraph: A prefetcher will perform only a byte code fetch operation for control transfers to an instruction beginning on a numerically odd physical address. So it prefetches words when at an even address, but starting to …

---

Re: x86 loading GDT entry 0 to which is present into a register?

Posted on 2017-09-05, 19:37

by

by peterferrie

in PC Emulation

in PC Emulation

Since the 0th entry in the GDT is not available for use in protected mode, it's wasted memory unless you put something there, such as the GDTR which happens to fit exactly in that space. That's really the only reason why you might consider doing it. As far as how it's used in different environments, …

---

Re: Finding bugs in 8086-80386 emulation core?

Posted on 2017-09-05, 19:29

by

by peterferrie

in PC Emulation

in PC Emulation

himem uses unreal mode for testing memory. The only time that it ever enters protected mode is to load the selectors before returning to real mode to do everything else, including memory copies.

---

Re: x86 loading GDT entry 0 to which is present into a register?

Posted on 2017-09-05, 03:45

by

by peterferrie

in PC Emulation

in PC Emulation

It doesn't matter if it's present or not. You can't access anything using selector 0 because it's reserved for "this selector is unused but requires an always-available value, (which is zero)". The memory location to which it points can be used to hold the GDTR so that you can LGT from that location …

---

Re: x86 loading GDT entry 0 to which is present into a register?

Posted on 2017-08-31, 18:24

by

by peterferrie

in PC Emulation

in PC Emulation

GDT entry 0 is not usable for any actual memory access. It will fault if you try.

---

Re: x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-18, 19:50

by

by peterferrie

in PC Emulation

in PC Emulation

Yes, the flags are the result of the add, but according to my documentation, it's a 16-bit add, not an 8-bit one. i.e. ax = ((ah * imm8) + al) & 0xff That might explain the difference in flags.

---

Re: x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-16, 20:49

by

by peterferrie

in PC Emulation

in PC Emulation

All flags are affected as expected when CL < 9, when doing RCL or RCR.

---

Re: x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-15, 21:32

by

by peterferrie

in PC Emulation

in PC Emulation

The rotate ROM is acting odd: it expects the RCLW BX,CL instruction to set Overflow flag while CL=8? Since CL<>1, it shouldn't even change the overflow flag? Also, there's still a problem with 80286+ AAA instruction? RCL would require CL=9 to rotate to the original value, otherwise it will shift …

---

Re: x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-15, 21:27

by

by peterferrie

in PC Emulation

in PC Emulation

AAA had for a long time a documentation error: it's single 16-bit add, not two 8-bit adds, just as AAS is a 16-bit subtraction.

---

Re: x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-11, 18:59

by

by peterferrie

in PC Emulation

in PC Emulation

For carry, it's not bit 7 that matters in an 8-bit add, it's bit 8 after the adc (consider 0xff+0xff+carry set, or 0xff+0x00+carry clear, depending on the test case). If the value after the add is less than the value of the left parameter, or if the result equals the left parameter and the right …

---

Re: Fit Up

Posted on 2017-08-11, 18:25

by

by peterferrie

in DOSBox Games/Apps

in DOSBox Games/Apps

Yes, it's been on my "to fix" list since forever. I'll get to it one day.

---

Re: x86 TLB vs protection?

Posted on 2017-07-30, 18:45

by

by peterferrie

in PC Emulation

in PC Emulation

The address has to be matched in the TLB before the protection check can happe nbecause there can be a mixture of supervisor and user addresses in the TLBs. Otherwise you are performing the address lookup that should be avoided by using the TLB in order to find the corresponding protection bits.

---