Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.sys.apple2.programmer > #4287 > unrolled thread

Emulator Detection

Back to article view | Back to comp.sys.apple2.programmer

Contents

  Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-06 14:46 -0800
    Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-06 19:03 -0800
      Re: Emulator Detection gids.rs@sasktel.net - 2018-03-07 01:07 -0800
        Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-07 07:18 -0800
          Re: Emulator Detection TomCh <tomcharlesworth26@gmail.com> - 2018-03-07 12:53 -0800
            Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-07 13:24 -0800
              Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-08 04:42 -0800
                Re: Emulator Detection Tom Porter <laseractiveguy@gmail.com> - 2018-03-08 07:41 -0800
                  Re: Emulator Detection Tom Porter <laseractiveguy@gmail.com> - 2018-03-08 07:46 -0800
                    Re: Emulator Detection Tom Porter <laseractiveguy@gmail.com> - 2018-03-08 08:55 -0800
    Re: Emulator Detection Steve Nickolas <usotsuki@buric.co> - 2018-03-07 05:49 -0500
    Re: Emulator Detection Mark Pilgrim <markpilgrimfastmail@gmail.com> - 2018-03-08 10:21 -0800
      Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-08 22:41 -0800
        Re: Emulator Detection MG <michael@apps.guidero.us> - 2018-03-10 13:24 -0800
          Re: Emulator Detection Zellyn <zellyn@gmail.com> - 2018-03-11 14:21 -0700
      Re: Emulator Detection qkumba <peter.ferrie@gmail.com> - 2018-03-11 14:50 -0700
        Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-11 22:02 -0700
        Re: Emulator Detection TomCh <tomcharlesworth26@gmail.com> - 2018-03-12 13:25 -0700
    Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-11 22:19 -0700
      Re: Emulator Detection gids.rs@sasktel.net - 2018-03-12 10:17 -0700
        Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-12 11:11 -0700
          Re: Emulator Detection mmphosis <mmphosis@macgui.com> - 2018-03-12 22:37 +0000
          Re: Emulator Detection gids.rs@sasktel.net - 2018-03-12 17:03 -0700
            Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-12 17:27 -0700
              Re: Emulator Detection gids.rs@sasktel.net - 2018-03-12 17:39 -0700
                Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-12 20:02 -0700
                  Re: Emulator Detection gids.rs@sasktel.net - 2018-03-12 20:39 -0700
                    Re: Emulator Detection gids.rs@sasktel.net - 2018-03-12 20:40 -0700
                    Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-12 20:58 -0700
                      Re: Emulator Detection qkumba <peter.ferrie@gmail.com> - 2018-03-12 21:51 -0700
                        Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-13 18:45 -0700
                          Re: Emulator Detection gids.rs@sasktel.net - 2018-03-13 22:22 -0700
                            Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-13 22:43 -0700
                              Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-14 00:08 -0700
                                Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-14 00:30 -0700
                                  Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-14 08:38 -0700
                          Re: Emulator Detection Zellyn <zellyn@gmail.com> - 2018-03-14 09:03 -0700
                            Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-14 13:58 -0700
                              Re: Emulator Detection Zellyn <zellyn@gmail.com> - 2018-03-15 06:11 -0700
          Re: Emulator Detection Michael Pohoreski <michael.pohoreski@gmail.com> - 2018-03-12 21:55 -0700
            Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-12 22:50 -0700
        Re: Emulator Detection Zellyn <zellyn@gmail.com> - 2018-03-12 13:40 -0700
    Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-14 08:42 -0700
      Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-14 13:57 -0700
        Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-14 21:07 -0700
          Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-15 14:56 -0700
            Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-16 20:49 -0700
              Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-16 22:20 -0700
                Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-16 22:24 -0700
                  Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-03-17 02:32 -0700
                    Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-17 08:09 -0700
                      Re: Emulator Detection Michael J. Mahon <mjmahon@aol.com> - 2018-03-17 17:01 -0500
                        Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-04-28 02:44 -0700
                          Re: Emulator Detection John Brooks <jbrooks@blueshiftinc.com> - 2018-04-28 22:01 -0700
                            Re: Emulator Detection Antoine Vignau <antoine.vignau@laposte.net> - 2018-04-29 08:16 -0700
    Re: Emulator Detection "Frank M." <frank_o_rama@hotmail.com> - 2018-03-14 21:23 -0700

Page 1 of 3  [1] 2 3  Next page →

#4287 — Emulator Detection

Hi All!

I have a use for being able to detect if a program is running under emulation. I found this presentation from Kfest and have a couple of questions.

https://archive.org/details/EmulatorDetectionIn6502AssemblyKansasfest2014Presentation


The code seems to count how many times the emulator "flies over" the data stream as it's presented to the emulator. Each emulator has a different speed that it emulates the disk hardware using, and the detector just calculates what that slowdown speed is, and links that to a specific emulator.

2. The code is written to work on a disk II. Can this method also work using a hard drive? I have only seen documentation for how the disk II rwts calls are made (c08c, etc) and am not sure how to make the same calls to a hard drive.


 ORG $300
 SEC ; Emulator Detection Code
 CLD ; for slot 6-based Disk II
 CLI ; There must be a disk in the drive
 CLV
 LDX #$60 ; Slot 6
 LDY #$00
 LDA $C089,X ; Turn on motor
LOOP1 LDA $C08C,X ; Read data
 BPL LOOP1
 CMP #$DE ; There is 1 "DE" on track 0 sector 0
 BNE LOOP1
LOOP2 LDA $C08C,X
 BPL LOOP2
 CMP #$AA ; There is 1 "AA" on track 0 sector 0
 BNE LOOP1
LOOP3 LDA $C08C,X
 BPL LOOP3
LOOP4 LDA $C08C,X
 BPL LOOP4
 CMP #$FF ; There are 5 "FF"s on track 0 sector 0
 BNE DONE
 INY
 BNE LOOP4
DONE STY $B6FC ; Output in location 46844 / PEEK (46844)
 LDA $C088,X ; turn off motor
 RTS


The basic program shows these timings... (reordered low to high)

I'm assuming that lower numbers = faster virtual drive access. Is this true?
Virtual II, for one, seems to have correct-ish disk ii speeds, while has basically zero waiting for hard drive access. AppleWin seems to be faithful for both. Wondering if this would also be a problem for detecting it via hard drive. 

 80  IF X = 7 THEN  PRINT "WELCOME MESS USER": END 
 30  IF X = 19 THEN  PRINT "WELCOME VIRTUAL ][ USER": END 
 60  IF X = 21 THEN  PRINT "WELCOME KEGS/ACTIVEGS/GSPORT USER": END 
 20  IF X = 27 THEN  PRINT "WELCOME APPLEWIN USER": END 
 70  IF X = 30 THEN  PRINT "WELCOME OPENEMULATOR USER": END 
 40  IF X = 33 THEN  PRINT "WELCOME SWEET16 USER": END 
 50  IF X = 49 THEN  PRINT "WELCOME JACE USER": END 
 90  PRINT "REAL HARDWARE OR UNKNOWN EMULATOR"

Frank

[toc] | [next] | [standalone]

#4289

The above code works with the Kfest disk image in the virtual disk II drive. On the image it's loaded to $0800 with the initial boot sector that's loaded by the disk II controller ROM. I updated the disassembly with my labels and comments.

The BASIC program does X = PEEK(46844)/$B6FC which I didn't include.

f

[toc] | [prev] | [next] | [standalone]

#4293

On Tuesday, March 6, 2018 at 9:03:23 PM UTC-6, Frank M. wrote:
> The above code works with the Kfest disk image in the virtual disk II drive. On the image it's loaded to $0800 with the initial boot sector that's loaded by the disk II controller ROM. I updated the disassembly with my labels and comments.
> 
> The BASIC program does X = PEEK(46844)/$B6FC which I didn't include.
> 
> f


I can confirm it works with Sweet16.
With Mac version of GSPort, it alternates between being recognized and unknown emulator, with alternating values of 21 and 6.

Would have been nice to see what real hardware values are to compare to.  I don't have any real hardware set up right now.

[toc] | [prev] | [next] | [standalone]

#4296

Since the program is only counting how many times "FF" shows up in the read stream, on real hardware I would expect that to be 5. Emulators present this value faster than the program can read it, so it takes multiple passes through track 0 sector 0 to sense the FF, hence numbers higher than 5. I haven't tried it on actual hardware though.

Does booting from the archive.org floppy give different results than my included disassembly? On Virtual II for me it always returns the same number (19). I tried to setup the same state, but there are other instructions on track 0, sector 0 that I may be missing.

Also note, if you run CATALOG or other disk command you will change the value, as the head is no longer on track 0 sector 0 as it would be when you boot up.  (the archive.org disk is designed to be booted from--My goal is to repurpose this code and method for virtual hard drive or CFFA/Floppy Emu).

F


On Wednesday, March 7, 2018 at 1:07:08 AM UTC-8, gid...@sasktel.net wrote:
> 
> Would have been nice to see what real hardware values are to compare to.  I don't have any real hardware set up right now.

[toc] | [prev] | [next] | [standalone]

#4298

On Wednesday, March 7, 2018 at 3:18:39 PM UTC, Frank M. wrote:
> Since the program is only counting how many times "FF" shows up in the read stream, on real hardware I would expect that to be 5. Emulators present this value faster than the program can read it, so it takes multiple passes through track 0 sector 0 to sense the FF, hence numbers higher than 5. I haven't tried it on actual hardware though.
> 
> Does booting from the archive.org floppy give different results than my included disassembly? On Virtual II for me it always returns the same number (19). I tried to setup the same state, but there are other instructions on track 0, sector 0 that I may be missing.
> 
> Also note, if you run CATALOG or other disk command you will change the value, as the head is no longer on track 0 sector 0 as it would be when you boot up.  (the archive.org disk is designed to be booted from--My goal is to repurpose this code and method for virtual hard drive or CFFA/Floppy Emu).
> 
> F
> 
> 
> On Wednesday, March 7, 2018 at 1:07:08 AM UTC-8, gid...@sasktel.net wrote:
> > 
> > Would have been nice to see what real hardware values are to compare to.  I don't have any real hardware set up right now.

re. AppleWin

As Steve says, it'll return either 6 (gap2 between addr epilogue & data prologue) or 75 (gap3+gap1 = 27+48) at the end->start of the track.

Of course this is for a .dsk (or .po or .do) and not a .nib image.

AppleWin will consistently give these numbers. It has a buffer of disk nibbles, and just returns the next nibble when you read $C0EC... irrespective of the number of cycles between accesses(*).

(*) Actually this is the default (ie. Disk Access Speed = Enhance Speed). When using Disk Access Speed = Authentic Speed, then there's some crude rotation emulation which will advance the nibble buffer's index by 1 for every 32 cycles between accesses.

NB. The above is correct for AppleWin 1.27.0.0, but shouldn't be relied on for any future versions!

Tom

[toc] | [prev] | [next] | [standalone]

#4299

Gotcha. Interesting. The disk must be relying on the rotation being in a certain position when the test loop is run. It does some other disk commands before calling the subroutine, and the motor has already been turned on.

Any way to sense this using virtual mass storage? I don't know the code to read from slot 7. Do you need the "motor on" command to do a read? Not sure I understand how mass storage devices work in regards to that.

F

[toc] | [prev] | [next] | [standalone]

#4301

You assume a dos 3.3 formatted disk. what if I write more FF/10 than the 5 described above? 5 is the value for the head to be in sync.

[toc] | [prev] | [next] | [standalone]

#4302

I've been thinking about how to detect an emulator for copy protection purposes... didn't think it was possible... wowzers this is great!

[toc] | [prev] | [next] | [standalone]

#4303

I do not know what I am doing most of the time... but here...

 BPL LOOP2 
 CMP #$AA ; There is 1 "AA" on track 0 sector 0 
 BNE LOOP1 
LOOP3 LDA $C08C,X 

shouldn't that be BNE LOOP2?

[toc] | [prev] | [next] | [standalone]

#4304

I got the program working and it tests 6 or 75 with AppleWIN depending on drive speed settings.

Checked with IIE plat with both the A2Heaven SD Floppy and 5.25 Disk Drive returns "0"... is that right?  I transferred the program over 3 times to make sure I did it correctly.  I did have one non "0" reading when I first did it but flicked thru the code/screens so fast I don't remember what that was.  Then I thought I corrupted the code, so kept transferring it over, on the emulator it still shows its numbers. 

[toc] | [prev] | [next] | [standalone]

#4294

I punched it in and half the time AppleWin gives me 6, with an occasional 
75. XD

My results with other emulators:

ApplePC gives me 8 with the occasional 72.
Modapple alternates between 5 and 6 with the occasional 0.
MAME goes 5 31 30 0, with the occasional stray 31.
Agat Emulator alternates (mostly) between 6 and 16 with the occasional 0.

-uso.

[toc] | [prev] | [next] | [standalone]

#4305

On Tuesday, March 6, 2018 at 5:46:40 PM UTC-5, Frank M. wrote:
> Hi All!
> 
> I have a use for being able to detect if a program is running under emulation. I found this presentation from Kfest and have a couple of questions.
> 
> https://archive.org/details/EmulatorDetectionIn6502AssemblyKansasfest2014Presentation

Oh hi, I am the author of this demo.

> The code seems to count how many times the emulator "flies over" the data stream as it's presented to the emulator. Each emulator has a different speed that it emulates the disk hardware using, and the detector just calculates what that slowdown speed is, and links that to a specific emulator.

No, it is counting the number of sync nibbles between the address field and the data field. These nibbles do not exist on a .dsk image (which only stores sector data), so emulators must recreate them. It turns out that each emulator recreates them slightly differently.

> 2. The code is written to work on a disk II. Can this method also work using a hard drive? I have only seen documentation for how the disk II rwts calls are made (c08c, etc) and am not sure how to make the same calls to a hard drive.

I don't know what hard drives look like at the bit or nibble level. There are copy protection routines that act on 3.5-inch disks, including weak/floating bytes (Rastan GS) and misplaced timing bits (Rampage //c), so maybe code could be written to operate on hard drive images. I've never tried it.

> I'm assuming that lower numbers = faster virtual drive access. Is this true?

No, the number it returns is the number of $FF nibbles after the address field epilogue and before the data field prologue.

Also, in response to someone else's question, the original demo disk runs the routine during the boot process of a standard DOS 3.3 disk, after the boot1 code on sector 0 has loaded the 9 sectors of RWTS code into memory. So the head of the drive is always at a specific place on track 0 (having just finished reading sector 0 for the second time), so the results are consistent. Your standalone program might catch the drive head anywhere on the track, so the string of $FF nibbles it finds might be anywhere -- after the address epilogue (like the original demo), after the data epilogue (higher numbers), or at the end of a track (possibly even higher numbers).

To reproduce the results of the original demo disk, you would need to call something like $B944 (to search for and parse the next available address field) in order to position the drive head so the next thing it reads is the nibbles after address epilogue.

Hope this helps.

-Mark

[toc] | [prev] | [next] | [standalone]

#4309

Thanks for chiming in Mark. It's very helpful. I just spent some time reading about the sync bytes and the Disk II. Doesn't look like a hard drive-based method would work, as all of this interleaving/sync bytes/etc is because the Disk II hardware relies on software to work properly.

For what it's worth, the code below seems to provide a working output under either DOS 3.3 or ProDOS, with either a DOS 3.3 or ProDOS disk in drive 1.

Unfortunately, not relevant to my application, but thought the code might be of use to someone else.

Frank


*
* Emulator Detection Code
* for slot 6-based Disk II
* There must be a disk in the drive
* (Can be either DOS 3.3 or ProDOS format)
* Works under DOS 3.3 and ProDOS
* Code to be pasted into Merlin (8) v2.33
*
* Output is at $03CF:
* $07/ 7 = MESS
* $13/19 = VIRTUAL ][
* $15/21 = KEGS/ACTIVEGS/GSPORT
* $1B/27 = APPLEWIN
* $1E/30 = OPENEMULATOR
* $21/33 = SWEET16
* $31/49 = JACE USER
* ANY OTHER VALUE = REAL HARDWARE OR UNKNOWN EMULATOR
*
 ORG $300
 LDX #$60 ; Slot 6
 LDY #$00
 LDA $C089,X ; Turn on motor
 LDA #$FF
 JSR $FCA8 ; DELAY based on accumulator
 JSR RDADR16 ; Wait for sync bytes
LOOP1 LDA $C08C,X ; Read data
 BPL LOOP1
 CMP #$DE ; Find closing byte #1
 BNE LOOP1
LOOP2 LDA $C08C,X
 BPL LOOP2
 CMP #$AA ; Find closing byte #2
 BNE LOOP1
LOOP3 LDA $C08C,X ; Loop to count number of $FF nibbles
 BPL LOOP3
LOOP4 LDA $C08C,X ; after the address field epilogue
 BPL LOOP4
 CMP #$FF ; and before the data field prologue
 BNE DONE
 INY
 BNE LOOP4
DONE STY $03CF ; Output in location 975 / PEEK (975)
 LDA $C088,X ; turn off motor
 RTS
*
* The below section is taken from DOS 3.3 source code
* For the RDADR16 routine located at $B944
* (Needed for ProDOS)
*
COUNT EQU $26 ; 'MUST FIND' COUNT. AKA GBASL
LAST EQU $26 ; 'ODD BIT' NIBLS. AKA HBASL
CSUM EQU $27 ; CHECKSUM BYTE.
CSSTV EQU $2C ; FOUR BYTES,
* CHECKSUM, SECTOR, TRACK, AND VOLUME.
Q6L EQU $C08C
DELAY EQU $FCA8
* 
RDERR SEC ;INDICATE 'ERROR EXIT'.
 RTS ;RETURN FROM READ16 OR RDADR16.
RDADR16 LDY #$FC ; AKA $B944 under DOS 3.3
 STY COUNT 'MUST FIND' COUNT.
RDASYN INY
 BNE RDA1 ;LOW ORDER OF COUNT.
 INC COUNT ;(2K NIBLS TO FIND
 BEQ RDERR ;ADR MARK, ELSE ERR)
RDA1 LDA Q6L,X ;READ NIBL.
 BPL RDA1 ;*** NO PAGE CROSS! ***
RDASN1 CMP #$D5 ;ADR MARK 1?
 BNE RDASYN ;(LOOP IF NOT)
 NOP ;ADDED NIBL DELAY.
RDA2 LDA Q6L,X
 BPL RDA2 ;*** NO PAGE CROSS! ***
 CMP #$AA ;ADR MARK 2?
 BNE RDASN1 ;(IF NOT, IS IT AM1?)
 LDY #$3 ;INDEX FOR 4-BYTE READ.
* (ADDED NIBL DELAY)
RDA3 LDA Q6L,X
 BPL RDA3 ;*** NO PAGE CROSS! ***
 CMP #$96 ;ADR MARK 3?
 BNE RDASN1 ;(IF NOT, IS IT AM1?)
* (LEAVES CARRY SET!)
 LDA #$0 INIT CHECKSUM.
RDAFLD STA CSUM
RDA4 LDA Q6L,X ;READ 'ODD BIT' NIBL.
 BPL RDA4 ;*** NO PAGE CROSS! ***
 ROL ;(Accum) ALIGN ODD BITS, '1' INTO LSB.
 STA LAST ;(SAVE THEM)
RDA5 LDA Q6L,X ;READ 'EVEN BIT' NIBL.
 BPL RDA5 ;*** NO PAGE CROSS! ***
 AND LAST ;MERGE ODD AND EVEN BITS.
 STA CSSTV,Y ;STORE DATA BYTE.
 EOR CSUM ;XOR CHECKSUM.
 DEY
 BPL RDAFLD ;LOOP ON 4 DATA BYTES.
 TAY ;IF FINAL CHECKSUM
 BNE RDERR ;NONZERO, THEN ERROR.
RDA6 LDA Q6L,X ;FIRST BIT-SLIP NIBL.
 BPL RDA6 ;*** NO PAGE CROSS! ***
 CMP #$DE
 BNE RDERR ;ERROR IF NONMATCH.
 NOP ;DELAY BETWEEN NIBLS.
RDA7 LDA Q6L,X ;SECOND BIT-SLIP NIBL.
 BPL RDA7 ;*** NO PAGE CROSS! ***
 CMP #$AA
 BNE RDERR ;ERROR IF NONMATCH.
RDEXIT CLC ;CLEAR CARRY ON
 RTS ;NORMAL READ EXITS.



or



CALL-151
300:A2 60 A0 00 BD 89 C0 A9
308:FF 20 A8 FC 20 3B 03 BD
310:8C C0 10 FB C9 DE D0 F7
318:BD 8C C0 10 FB C9 AA D0
320:EE BD 8C C0 10 FB BD 8C
328:C0 10 FB C9 FF D0 03 C8
330:D0 F4 8C CF 03 BD 88 C0
338:60 38 60 A0 FC 84 26 C8
340:D0 04 E6 26 F0 F3 BD 8C
348:C0 10 FB C9 D5 D0 F0 EA
350:BD 8C C0 10 FB C9 AA D0
358:F2 A0 03 BD 8C C0 10 FB
360:C9 96 D0 E7 A9 00 85 27
368:BD 8C C0 10 FB 2A 85 26
370:BD 8C C0 10 FB 25 26 99
378:2C 00 45 27 88 10 E7 A8
380:D0 B7 BD 8C C0 10 FB C9
388:DE D0 AE EA BD 8C C0 10
390:FB C9 AA D0 A4 18 60
300G
3CF

[toc] | [prev] | [next] | [standalone]

#4313

On Thursday, March 8, 2018 at 10:41:47 PM UTC-8, Frank M. wrote:
> Doesn't look like a hard drive-based method would work, as all of this interleaving/sync bytes/etc is because the Disk II hardware relies on software to work properly. 

There are other ways to detect various emulators, either because they implement a de-facto standard for identifying them, or because almost none of them get the hardware I/O page exactly right.

https://github.com/mgcaret/davex-mg-utils/blob/master/idemu.s

That only supports a few emulators that I've actually tried.  Others may be added.

MG

[toc] | [prev] | [next] | [standalone]

#4316

On Saturday, March 10, 2018 at 4:24:59 PM UTC-5, MG wrote:
> There are other ways to detect various emulators, either because they implement a de-facto standard for identifying them, or because almost none of them get the hardware I/O page exactly right.
> 
> https://github.com/mgcaret/davex-mg-utils/blob/master/idemu.s

Hmmm. My a2audit project should eliminate that method eventually :-)
https://github.com/zellyn/a2audit

Zellyn

[toc] | [prev] | [next] | [standalone]

#4317

> No, the number it returns is the number of $FF nibbles after the address field epilogue and before the data field prologue.

Actually, it's after the data field epilogue of the next sector and before the address field prologue of the sector after that, because the denibbilisation causes the next sector data to be reached part-way in before the epilogue is polled.

[toc] | [prev] | [next] | [standalone]

#4318

For what it's worth, I removed a couple sections of the DOS 3.3 routine on qkumba's suggestion. It seems to be returning pretty constant values based on what emulator it's in. Again, requires a disk II (real or otherwise) to function.

I also added a 0.6 second wait for actual hardware to spin up once the motor has been turned on. Unfortunately not seeing consistent results on a real A2, but didn't see any of the emulator values I tested in 30 or so iterations.

F


*
* Emulator Detection Code
* for slot 6-based Disk II
* There must be a disk in the drive
* (Can be either DOS 3.3 or ProDOS format)
* Code works under DOS 3.3 or ProDOS
* Code to be pasted into Merlin (8) v2.33
*
* Output is at $03CF:
* $05/ 5 = SWEET16
* $06/ 6 = APPLEWIN
* $07/ 7 = VIRTUAL ][
* $00,$01,$02,$03,$09 ? = REAL HARDWARE

* $XX/ ? = MESS
* $XX/ ? = KEGS/ACTIVEGS/GSPORT
* $XX/ ? = OPENEMULATOR
* $XX/ ? = JACE
* ANY OTHER VALUE = REAL HARDWARE OR UNKNOWN EMULATOR
*
 ORG $300
 LDX #$60 ; Slot 6
 LDY #$00
 LDA $C089,X ; Turn on motor
 LDA #$05 ; do WAIT 6 times
 STA $26
WAIT LDA #$C6
 JSR $FCA8 ; WAIT = .1 sec each time
 DEC $26
 LDA $26
 CMP #$00 ; done waiting 0.6 seconds for drive to spin up
 BNE WAIT
 JSR RDADR16 ; Wait for sync bytes
LOOP1 LDA $C08C,X ; Read data
 BPL LOOP1
 CMP #$DE ; Find closing byte #1
 BNE LOOP1
LOOP2 LDA $C08C,X
 BPL LOOP2
 CMP #$AA ; Find closing byte #2
 BNE LOOP1
LOOP3 LDA $C08C,X ; Loop to count number of $FF nibbles
 BPL LOOP3
LOOP4 LDA $C08C,X ; after the address field epilogue
 BPL LOOP4
 CMP #$FF ; and before the data field prologue
 BNE DONE
 INY
 BNE LOOP4
DONE STY $03CF ; Output in location 975 / PEEK(975)
 LDA $C088,X ; turn off motor
 RTS
*
* The below section is taken from DOS 3.3 source code
* For the RDADR16 routine located at $B944
*
COUNT EQU $26 ; 'MUST FIND' COUNT. AKA GBASL
LAST EQU $26 ; 'ODD BIT' NIBLS. AKA HBASL
CSUM EQU $27 ; CHECKSUM BYTE.
CSSTV EQU $2C ; FOUR BYTES,
* CHECKSUM, SECTOR, TRACK, AND VOLUME.
Q6L EQU $C08C
DELAY EQU $FCA8
* 
RDERR SEC ;INDICATE 'ERROR EXIT'.
 RTS ;RETURN FROM READ16 OR RDADR16.
RDADR16 LDY #$FC ; AKA $B944 under DOS 3.3
 STY COUNT 'MUST FIND' COUNT.
RDASYN INY
 BNE RDA1 ;LOW ORDER OF COUNT.
 INC COUNT ;(2K NIBLS TO FIND
 BEQ RDERR ;ADR MARK, ELSE ERR)
RDA1 LDA Q6L,X ;READ NIBL.
 BPL RDA1 ;*** NO PAGE CROSS! ***
RDASN1 CMP #$D5 ;ADR MARK 1?
 BNE RDASYN ;(LOOP IF NOT)
 NOP ;ADDED NIBL DELAY.
RDA2 LDA Q6L,X
 BPL RDA2 ;*** NO PAGE CROSS! ***
 CMP #$AA ;ADR MARK 2?
 BNE RDASN1 ;(IF NOT, IS IT AM1?)
 LDY #$3 ;INDEX FOR 4-BYTE READ.
* (ADDED NIBL DELAY)
RDA3 LDA Q6L,X
 BPL RDA3 ;*** NO PAGE CROSS! ***
 CMP #$96 ;ADR MARK 3?
 BNE RDASN1 ;(IF NOT, IS IT AM1?)
* (LEAVES CARRY SET!)
 LDA #$0 INIT CHECKSUM.
RDAFLD STA CSUM
RDA4 LDA Q6L,X ;READ 'ODD BIT' NIBL.
 BPL RDA4 ;*** NO PAGE CROSS! ***
 ROL ;(Accum) ALIGN ODD BITS, '1' INTO LSB.
 STA LAST ;(SAVE THEM)
RDA5 LDA Q6L,X ;READ 'EVEN BIT' NIBL.
 BPL RDA5 ;*** NO PAGE CROSS! ***
 AND LAST ;MERGE ODD AND EVEN BITS.
 STA CSSTV,Y ;STORE DATA BYTE.
 EOR CSUM ;XOR CHECKSUM.
 DEY
 BPL RDAFLD ;LOOP ON 4 DATA BYTES.
 TAY ;IF FINAL CHECKSUM
 BNE RDERR ;NONZERO, THEN ERROR.
*
* This section is commented out from original DOS 3.3 source
* Read address epilogue FFs, not data epilogue FFs
*
*RDA6 LDA Q6L,X ;FIRST BIT-SLIP NIBL.
* BPL RDA6 ;*** NO PAGE CROSS! ***
* CMP #$DE
* BNE RDERR ;ERROR IF NONMATCH.
* NOP ;DELAY BETWEEN NIBLS.
*RDA7 LDA Q6L,X ;SECOND BIT-SLIP NIBL.
* BPL RDA7 ;*** NO PAGE CROSS! ***
* CMP #$AA
* BNE RDERR ;ERROR IF NONMATCH.
*
RDEXIT CLC ;CLEAR CARRY ON
 RTS ;NORMAL READ EXITS.





On Sunday, March 11, 2018 at 2:50:16 PM UTC-7, qkumba wrote:
> > No, the number it returns is the number of $FF nibbles after the address field epilogue and before the data field prologue.
> 
> Actually, it's after the data field epilogue of the next sector and before the address field prologue of the sector after that, because the denibbilisation causes the next sector data to be reached part-way in before the epilogue is polled.

[toc] | [prev] | [next] | [standalone]

#4322

On Sunday, March 11, 2018 at 9:50:16 PM UTC, qkumba wrote:
> > No, the number it returns is the number of $FF nibbles after the address field epilogue and before the data field prologue.
> 
> Actually, it's after the data field epilogue of the next sector and before the address field prologue of the sector after that, because the denibbilisation causes the next sector data to be reached part-way in before the epilogue is polled.

Unless it's an emulator like AppleWin :)
...where the disk only spins when you access $C0EC.

[toc] | [prev] | [next] | [standalone]

#4319

Thanks for the tip MG. I looked into location $C04F and it seemed to have interesting behavior for testing emulators.


$C04F - I/O [reserved location] ("de-facto standard among IIgs emulators")


LDA $C04F seems to always return:

$00 - Virtual II
$30 - Sweet16

AppleWin + real hardware fluctuates (floating bus). You can see the bus fluctuations by entering the following short program...

CALL-151
300:AD 4F c0 20 ED FD 4C 00 03
300G

Probably returning similar values to the "vapor lock" method of detecting the vertical blank period. Not tested on any other emulators, so feedback would be appreciated.


Both Virtual II & Sweet16 totally mess up mass storage emulation, which was what I wanted to detect in the first place.

F

[toc] | [prev] | [next] | [standalone]

#4320

On Sunday, March 11, 2018 at 11:19:13 PM UTC-6, Frank M. wrote:
> Thanks for the tip MG. I looked into location $C04F and it seemed to have interesting behavior for testing emulators.
> 
> 
> $C04F - I/O [reserved location] ("de-facto standard among IIgs emulators")
> 
> 
> LDA $C04F seems to always return:
> 
> $00 - Virtual II
> $30 - Sweet16
> 
> AppleWin + real hardware fluctuates (floating bus). You can see the bus fluctuations by entering the following short program...
> 
> CALL-151
> 300:AD 4F c0 20 ED FD 4C 00 03
> 300G
> 
> Probably returning similar values to the "vapor lock" method of detecting the vertical blank period. Not tested on any other emulators, so feedback would be appreciated.
> 
> 
> Both Virtual II & Sweet16 totally mess up mass storage emulation, which was what I wanted to detect in the first place.
> 
> F


Detecting emulators would be kind of redundant unless every little idiosyncracy
is logged for every problem that happens within each emulator.

Is there a plan to use this code for work-arounds in certain emulators?

I would like to see a list of issues that break the code under each emulator.
 

[toc] | [prev] | [next] | [standalone]

Page 1 of 3  [1] 2 3  Next page →

Back to top | Article view | comp.sys.apple2.programmer

csiph-web