VOGONS

Search

Common searches

Search results

Display options

Display options

Re: test386.asm CPU tester

Posted on 2017-11-10, 01:01

Just debugged a single 32-bit shift: it sets the CF, then shifts the original value(=1) and ors in the carry flag shifted left by 31, which results in an incorrect 0xffffffff80000000? Shifting more causes those 1-bits to slide into visible range, causing the invalid result? Can anyone tell me why …

Re: test386.asm CPU tester

Posted on 2017-11-09, 23:33

In the meanwhile fixed some bugs and some instruction debugging(opcodes 05h, 0Dh and C0h still were adding the operand size suffixes(now handled by the modr/m)). That should make it fully common log format compatible again. Btw, the short method of memory accesses is currently used with UniPCemu's …

Re: test386.asm CPU tester

Posted on 2017-11-09, 19:35

One thing that still goes awry is the ROR instruction? Copy code to clipboard 1 RORi D EAX=00000001 EDX=00000000 PS=0044 EAX=FE000000 EDX=00000000 PS=0044 Anyone? EAX is supposed to be 02000000?

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 23:38

The geometry given is generated by the function, which currently follows the appendix C and general rules quoted in my last post. If I were to reduce the cylinders to 1024, it wouldn't match the rules without loss of data, or would it? Do I need to add an additional restraint for 1024 cylinders when …

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 23:05

Just found a bug in the CHS detection algorithm, not counting past 0x3FFF instead of 0xFFFF cylinders. I've fixed this now. 2GB reports as 3968,16,63(CHS) and 100MB reports as 12800,16,1(CHS). That should be correct, according to the rules of ATA-5? I can't rely on disk contents, as disks may start …

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 22:35

The words 54, 55 and 56 are supposed to be words 1,3,6. 54,55,56 are the current setting loaded by INITIALIZE DEVICE PARAMETERS command. Wirds 1,3,6 is the default CHS translation used by the disk, loaded when loading the disk image only. Manual on this(device parameters, see function I've given): …

Re: Finding bugs in 8086-80386 emulation core?

Posted on 2017-11-08, 21:26

That shouldn't be the case? The 20% reported by UniPCemu's CPU speed is a direct one-per-second renderjng of the emulated time(nanoseconds emulated( vs realtime(nanoseconds HniPCemu is running, using gettimeofday). It ticks about once each second, retrieves current Unix-style timestamp from the OS …

Re: test386.asm CPU tester

Posted on 2017-11-08, 21:14

I know that the debugger unit in UniPCemu has such texts(Logging of memory accesses use "%08X=%02X(%c)" with values memaddr, value, value in a sprintf)? Maybe rogue memory(overflow) logs somehow?

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 17:49

Just tried using the new geometry and disk image(which has been reformatted to the new geometry). It still shows up at 4MB only in Dosbox(after booting)? The rules mentioned in the ATA-5 spec are followed, as far as I know, to create the emulated disk image geometry(which is reported at diagnostic …

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 17:29

So I would need to create a new disk and move the files from the old disk to a newly created disk(which has an improved CHS partitioning, which follows the ATA-5 appendix C specification)? The 2GB disk image has no problems, since the geometry is unchanged(although the new geometry might have some …

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 14:45

I simply mounted using: imgmount c .\games_~1.img -t hdd -size 512,63,16,3968 The original geometry I used in UniPCemu (formatting it in that way and using it until recent, MS-DOS 5.0a 2GB sfdimg disk image) was the same(due to forced 16,64 and cylinders being automatically calculated). I'll need to …

Re: test386.asm CPU tester

Posted on 2017-11-08, 13:53

Just tried running MS-DOS 6.22's EMM386.EXE with HIMEM.SYS on MS-DOS 5.0a. I now see it's trying to execute a #PF handler due to some unpaged memory, which eventually causes a #GP handler due to EBX offset calculation overflowing(32-bit offset using 16-bit DS descriptor), which eventually runs into …

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 13:49

Would the fact that the default CHS geometry that's reported in UniPCemu's ATA hard disk emulation at drive diagnostics offsets 54,55,56(Default CHS translation) change and make MS-DOS using of the hard disk image buggy? Or will it continue to work without problems using the newly set geometry( …

Re: MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 12:44

Just modified the geometry formula, then tested my large test games disk image(with a MS-DOS 5.0 OS and various games on it to test functionality of my emulator) by mounting it in dosbox and using it normally(though the dosbox prompt). The new geometry is 3968,16,63(C,H,S format) and inputting that …

MS-DOS disk image CHS compatibility?

Posted on 2017-11-08, 10:55

Are MS-DOS hard disk image storage affected by changing the hard disk physical geometry? E.g. by improving an old algorithm to calculate CHS dimensions of a hard disk to be more efficient, would it break data stored on said disk(which is stored as a flat disk image)? Or does MS-DOS always autodetect …

Re: test386.asm CPU tester

Posted on 2017-11-07, 10:32

Just made a simple diff using diff from the diffutils for Windows: test386-EE-reference.diff Anyone can see what's going wrong? It's using the same algorithm as fake86 and CAPE(at least for the ROL/ROR/RCL/RCR instructions. SHL/SAL/SHR instructions are modfied to give correct results). byte op_grp2 …

Re: Finding bugs in 8086-80386 emulation core?

Posted on 2017-11-06, 17:36

For some odd reason, Windows 95a MS-DOS is insanely slow? The screen starts flashing(updates?) and response time becomes extremely long(typing results after a few seconds instead of instantaneously)? DMA isn't running as far as I can see. Maybe some other cause? Tried checking stuff like interrupts, …

Re: Finding bugs in 8086-80386 emulation core?

Posted on 2017-11-06, 16:15

Yay! Good news: floppy disk booting is working now on the Compaq Deskpro 386! 😁 Finally can boot the Windows 95 floppy disks with required stuff for Windows 95 setup:D (Instead of plain MS-DOS already installed on the hard disk)! Filename 460-Windows 95a boot disk booting on Compaq Deskpro 386.jpg …

Page 177 of 275

Go to top of page Go to top of page
Close window
Close window