CD-ROM Emulation doesn't work with stock Apple drivers #1

akuker · 2020-07-08T20:16:59Z

The "stock" version of Apple's CD-ROM driver for MacOS 8.x (and probably all versions of "classic" Mac OSs) does not recognize the emulated CD-ROM of RaSCSI. At startup, the CD-ROM driver will send an unsupported command that is not supported by RaSCSI.

I believe this check was done to make sure that only CD-ROMs with Apple firmware were used.

There are a couple work-arounds:

Install ResExcellence's CD-ROM driver hack (https://web.archive.org/web/20040404121044/http://www.resexcellence.com/hack_html_99/12-21-98.shtml)
[Untested] Use a 3rd party CD-ROM driver. The CD driver that was distributed with Toast may be a good one to try.

erichelgeson · 2020-10-12T22:18:39Z

FYI https://macintoshgarden.org/apps/cd-sunrise-v22b did not work with rascsi cd emulation

rdmark · 2021-08-17T23:54:42Z

For the record, the Toast CD Reader driver from https://macintoshgarden.org/apps/astarte-toast-256 does not work either. The extension just fails to load when testing on System 7.1 (Pro and P6) as well as Mac OS 7.6.1, even though a RaSCSI CD drive is attached.

rdmark · 2021-08-17T23:59:22Z

One observation right now, is that since ResExcellence's hack is the only known (?) solution, and the oldest version of Apple CD-ROM that the hack is proven to work with is 5.4.2, effectively the oldest Mac OS version that can be used right now with RaSCSI CD-ROM is 8.1.
Today I spent a few hours trying to hack earlier Apple CD-ROM drivers without success.
Edit: the above statement is from testing on develop. Akuker has informed me that it should work on 7.0.1 or 7.5.3 with master.

rdmark · 2021-08-18T23:13:29Z

I've got some good news: After some research and testing today, I am glad to announce that I've found two 3rd party drivers that work well with RaSCSI: RapidCD and CD/DVD Drive TuneUp 2.0.3
See my detailed test results here: http://shorturl.at/wKNY9

My favorite is RapidCD, since it mounts CDs super fast, also can handle the DOS partitions, and is the only one that works on my stock Mac SE. It even works on System 6.0.7, which blew me away!

uweseimet · 2021-08-21T08:50:45Z

@rdmark It would be helpful to have a log on trace level for those cases where software fails to use the emulated devices.

rdmark · 2021-08-22T06:10:57Z

@uweseimet This is an example when attaching an image and the system not being able to read the disk. Hacked Apple CD ROM 5.3.1, System 7.1.1 on a Mac SE.
The same image can be read by the system when using 3rd party drivers such as CD/DVD TuneUp, or RapidCD.

[2021-08-22 07:07:51.674] [info] Validating: cmd=ATTACH, id=4, unit=0, type=SCCD, filename='/home/pi/images/71.iso', device name='', params=''
[2021-08-22 07:07:51.676] [info] Executing: cmd=ATTACH, id=4, unit=0, type=SCCD, filename='/home/pi/images/71.iso', device name='', params=''
[2021-08-22 07:07:51.679] [info] Media capacity for image file '/home/pi/images/71.iso': 39067 blocks
[2021-08-22 07:07:51.681] [info] Added new SCCD device, ID: 4 unit: 0
[2021-08-22 07:07:51.810] [info] +----+----+------+-------------------------------------
[2021-08-22 07:07:51.813] [info] | ID | UN | TYPE | DEVICE STATUS
[2021-08-22 07:07:51.816] [info] +----+----+------+-------------------------------------
[2021-08-22 07:07:51.816] [info] | 3 | 0 | SCHD | /home/pi/images/600MB.hda
[2021-08-22 07:07:51.824] [info] | 4 | 0 | SCCD | /home/pi/images/71.iso (WRITEPROTECT)
[2021-08-22 07:07:51.825] [info] +----+----+------+-------------------------------------
[2021-08-22 07:07:52.668] [trace] virtual void SCSIDEV::Selection() Selection Phase ID=4 (with device)
[2021-08-22 07:07:52.668] [trace] virtual void SASIDEV::Command() Command Phase
[2021-08-22 07:07:52.668] [trace] virtual void SCSIDEV::Execute() Execution phase command $00
[2021-08-22 07:07:52.669] [debug] ++++ CMD ++++ virtual void SCSIDEV::Execute() ID 4 received CmdTestUnitReady ($00)
[2021-08-22 07:07:52.669] [trace] void SASIDEV::CmdTestUnitReady() TEST UNIT READY Command
[2021-08-22 07:07:52.669] [trace] Setting status: Sense Key: $06, ASC: $28, ASCQ: $00
[2021-08-22 07:07:52.669] [trace] BOOL Disk::CheckReady() Disk in needs attention
[2021-08-22 07:07:52.669] [trace] virtual void SCSIDEV::Error(ERROR_CODES::sense_key, ERROR_CODES::asc) Sense Key and ASC for subsequent REQUEST SENSE: $00, $00
[2021-08-22 07:07:52.669] [trace] virtual void SCSIDEV::Error(ERROR_CODES::sense_key, ERROR_CODES::asc) Error (to status phase)
[2021-08-22 07:07:52.669] [trace] void SASIDEV::Status() Status Phase $02
[2021-08-22 07:07:52.669] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 07:07:52.670] [trace] virtual void SCSIDEV::Send() Move to next phase status (8)
[2021-08-22 07:07:52.670] [trace] void SASIDEV::MsgIn() Starting Message in phase
[2021-08-22 07:07:52.670] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 07:07:52.670] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 07:07:52.670] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 07:07:52.670] [trace] virtual void SCSIDEV::Send() Move to next phase msgin (9)
[2021-08-22 07:07:52.670] [trace] virtual void SCSIDEV::BusFree() Bus free phase
[2021-08-22 07:07:52.672] [trace] virtual void SCSIDEV::Selection() Selection Phase ID=4 (with device)
[2021-08-22 07:07:52.672] [trace] virtual void SASIDEV::Command() Command Phase
[2021-08-22 07:07:52.673] [trace] virtual void SCSIDEV::Execute() Execution phase command $03
[2021-08-22 07:07:52.673] [debug] ++++ CMD ++++ virtual void SCSIDEV::Execute() ID 4 received CmdRequestSense ($03)
[2021-08-22 07:07:52.673] [trace] void SASIDEV::CmdRequestSense() REQUEST SENSE Command
[2021-08-22 07:07:52.673] [trace] virtual int Disk::RequestSense(const DWORD*, BYTE*) size of data = 18
[2021-08-22 07:07:52.673] [trace] Setting status: Sense Key: $00, ASC: $00, ASCQ: $00
[2021-08-22 07:07:52.673] [trace] void SASIDEV::CmdRequestSense() Sense Key $06, ASC $28
[2021-08-22 07:07:52.673] [trace] void SASIDEV::DataIn() Going into Data-in Phase
[2021-08-22 07:07:52.673] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 18
[2021-08-22 07:07:52.674] [trace] virtual void SCSIDEV::Send() Move to next phase datain (6)
[2021-08-22 07:07:52.674] [trace] void SASIDEV::Status() Status Phase $00
[2021-08-22 07:07:52.674] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 07:07:52.674] [trace] virtual void SCSIDEV::Send() Move to next phase status (8)
[2021-08-22 07:07:52.674] [trace] void SASIDEV::MsgIn() Starting Message in phase
[2021-08-22 07:07:52.674] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 07:07:52.674] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 07:07:52.675] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 07:07:52.675] [trace] virtual void SCSIDEV::Send() Move to next phase msgin (9)
[2021-08-22 07:07:52.675] [trace] virtual void SCSIDEV::BusFree() Bus free phase
[2021-08-22 07:07:52.676] [trace] virtual void SCSIDEV::Selection() Selection Phase ID=4 (with device)
[2021-08-22 07:07:52.676] [trace] virtual void SASIDEV::Command() Command Phase
[2021-08-22 07:07:52.677] [trace] virtual void SCSIDEV::Execute() Execution phase command $00
[2021-08-22 07:07:52.677] [debug] ++++ CMD ++++ virtual void SCSIDEV::Execute() ID 4 received CmdTestUnitReady ($00)
[2021-08-22 07:07:52.677] [trace] void SASIDEV::CmdTestUnitReady() TEST UNIT READY Command
[2021-08-22 07:07:52.677] [trace] Setting status: Sense Key: $00, ASC: $00, ASCQ: $00
[2021-08-22 07:07:52.677] [trace] BOOL Disk::CheckReady() Disk is ready!
[2021-08-22 07:07:52.677] [trace] void SASIDEV::Status() Status Phase $00
[2021-08-22 07:07:52.678] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 07:07:52.678] [trace] virtual void SCSIDEV::Send() Move to next phase status (8)
[2021-08-22 07:07:52.678] [trace] void SASIDEV::MsgIn() Starting Message in phase
[2021-08-22 07:07:52.678] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 07:07:52.678] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 07:07:52.678] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 07:07:52.679] [trace] virtual void SCSIDEV::Send() Move to next phase msgin (9)
[2021-08-22 07:07:52.679] [trace] virtual void SCSIDEV::BusFree() Bus free phase
[2021-08-22 07:07:52.680] [trace] virtual void SCSIDEV::Selection() Selection Phase ID=4 (with device)
[2021-08-22 07:07:52.680] [trace] virtual void SASIDEV::Command() Command Phase
[2021-08-22 07:07:52.681] [trace] virtual void SCSIDEV::Execute() Execution phase command $43
[2021-08-22 07:07:52.681] [debug] ++++ CMD ++++ virtual void SCSIDEV::Execute() ID 4 received CmdReadToc ($43)

rdmark · 2021-08-22T20:42:38Z

Thanks to Stinkerton18 on Discord, we now have a patched version of the Apple CD-ROM driver that works at least as far back as on System 6.0.7! I've uploaded both to the Garden: https://macintoshgarden.org/apps/apple-cd-rom-extension-patched

BTW, I've fallen out of love slightly with RapidCD. While highly compatible, it's a major memory hog. On system 7.1.1 one attached CD drive takes up 680kb, while Apple's driver only uses 110kb per drive. On a memory limited system like a 4MB SE, RapidCD is almost unusable in real life.

uweseimet · 2021-08-22T20:48:50Z

@rdmark Regarding the log above, does it stop (no further logging) after the CmdReadToc command? I would have expected more output.

rdmark · 2021-08-22T21:13:30Z

@uweseimet Oh no, a loop of similar looking sequences of log messages keep going at a very rapid pace while using a CD device. I can give you a few thousand lines worth if it helps!

uweseimet · 2021-08-22T21:18:39Z

@rdmark :-) Please just some more lines until it is obvious that there is a loop. If I'm not mistaken CmdReadToc is only called once in the log above, i.e. the loop does not seem to have started.

rdmark · 2021-08-22T21:42:19Z

cd_trace.txt
Here you go!

uweseimet · 2021-08-22T21:56:14Z

Now one can see that a MODE SELECT fails (see logfile excerpt below). Looks as if the CD-ROM software is trying to configure the device but RaSCSI doesn't allow that.
Maybe it is sufficient to just return an OK status in this case, without actually doing something for this specific MODE SELECT. It may be worth trying this at some time in the future.

[2021-08-22 22:40:48.717] [trace] virtual void SCSIDEV::Execute() Execution phase command $15
[2021-08-22 22:40:48.717] [debug] ++++ CMD ++++ virtual void SCSIDEV::Execute() ID 5 received CmdModeSelect ($15)
[2021-08-22 22:40:48.717] [trace] void SCSIDEV::CmdModeSelect() MODE SELECT Command
[2021-08-22 22:40:48.717] [trace] void SASIDEV::DataOut() Data out phase
[2021-08-22 22:40:48.717] [trace] virtual void SCSIDEV::Receive()
[2021-08-22 22:40:48.717] [trace] virtual void SCSIDEV::Receive() length was != 0
[2021-08-22 22:40:48.717] [trace] virtual void SCSIDEV::Receive()
[2021-08-22 22:40:48.717] [trace] virtual void SCSIDEV::Receive() ctrl.phase: 7 (dataout)
[2021-08-22 22:40:48.717] [trace] Setting status: Sense Key: $05, ASC: $26, ASCQ: $00
[2021-08-22 22:40:48.718] [trace] virtual void SCSIDEV::Error(ERROR_CODES::sense_key, ERROR_CODES::asc) Sense Key and ASC for subsequent REQUEST SENSE: $00, $00
[2021-08-22 22:40:48.718] [trace] virtual void SCSIDEV::Error(ERROR_CODES::sense_key, ERROR_CODES::asc) Error (to status phase)
[2021-08-22 22:40:48.718] [trace] void SASIDEV::Status() Status Phase $02
[2021-08-22 22:40:48.718] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 22:40:48.718] [trace] virtual void SCSIDEV::Send() Move to next phase status (8)
[2021-08-22 22:40:48.718] [trace] void SASIDEV::MsgIn() Starting Message in phase
[2021-08-22 22:40:48.718] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 22:40:48.718] [trace] virtual void SCSIDEV::Send() sending handhake with offset 0, length 1
[2021-08-22 22:40:48.718] [trace] void SASIDEV::MsgIn() Transitioning to Send()
[2021-08-22 22:40:48.718] [trace] virtual void SCSIDEV::Send() Move to next phase msgin (9)
[2021-08-22 22:40:48.718] [trace] virtual void SCSIDEV::BusFree() Bus free phase
[2021-08-22 22:40:48.719] [trace] virtual void SCSIDEV::Selection() Selection Phase ID=5 (with device)
[2021-08-22 22:40:48.720] [trace] virtual void SASIDEV::Command() Command Phase
[2021-08-22 22:40:48.720] [trace] virtual void SCSIDEV::Execute() Execution phase command $03
[2021-08-22 22:40:48.720] [debug] ++++ CMD ++++ virtual void SCSIDEV::Execute() ID 5 received CmdRequestSense ($03)

uweseimet · 2021-08-23T08:43:52Z

@rdmark The current uweseimet_develop branch contains additional debug output, which should reveal which code page MODE SELECT wants to change. Would be great if you provided updated logs whenever you have the time for it.

rdmark · 2021-08-23T19:23:58Z

@uweseimet Do I need to compile rascsi with debug symbols to get the right level of debug output?

uweseimet · 2021-08-23T20:31:56Z

@rdmark Not anymore. One of my changes was to remove the coupling between compile options and logging. Note that you can also change the log level of an already running rascsi instance, e.g. with "rasctl -g trace".

rdmark · 2021-08-23T20:53:17Z

@uweseimet trace using b40c6b1 on my Mac SE, System 6.0.7, Apple CD-ROM driver. Up through booting the system off a non-RaSCSI hard drive, the system complaining that the CD file system can't be read, and me hitting the Eject button.
cd_trace2.txt

uweseimet · 2021-08-23T21:00:54Z

@rdmark Thank you for providing updated logs. At least from the new logs we know now which mode page the driver tries to use:

[2021-08-23 21:48:02.151] [debug] virtual bool Disk::Dispatch(SCSIDEV*) received CmdModeSelect ($15)
[2021-08-23 21:48:02.151] [trace] MODE SELECT for unsupported page 21

When (and if) somebody starts investigating this issue thsi piece of information might be useful.

…ed from HD to Disk

akuker · 2021-12-30T04:12:12Z

There's also an issue with ModeSense6. When the Mac sends the ModeSense6 command, it specifies a size of 0x14 (20 decimal). This graphic shows that RaSCSI doesn't respond to the ModeSense6 for Page 1 request. It transitions to an error status 2.

The current RaSCSI code calculates a response of size 24. This is too big, and I don't think RaSCSI is calculating the response size appropriately. The response size should be 20 bytes (as requested by the Mac).... 4 bytes (for the mode sense data header) + 8 bytes (for the block descriptor) + 8 bytes (for the error recovery parameters).

@uweseimet - I'm curious about your opinions on the risk associated with this change.

…bug logging verbosity

akuker · 2021-12-30T04:33:08Z

Another interesting observation - With the Apple CD drive I was using, I didn't ever see a ModeSelect6 (15h) command in the SCSI data capture. I'm going to do some more data captures from some other drives to make sure this wasn't a fluke.

My wild theory? If the drivers get the data they expect from the Inquiry and/or ModeSense6 commands, it doesn't need to change anything using the ModeSelect command.

uweseimet · 2021-12-30T09:29:05Z

@akuker Well, there is no risk if your change is compliant with the SCSI specification. If it is not the change will break a lot. In cases like this it is very important to check against the specification.
I may be wrong, but I think at least MODE SENSE(10) was fixed by somebody for the 21.11 release?

In general, which commands and parameters drivers send can definitely depend on results from previous commands. It's not unusual that drivers fall back to different commands if they find (based on previous responses, typically from INQUIRY, MODE SENSE, or an SCSI error code) that a certain feature is not available.
Real-world examples:

Device ID/LUN checks may depend on whether a device supports REPORT LUNS. If REPORT LUNS is available the device ID/LUN check can be optimized.
A driver may use READ (6) for small sector numbers instead of READ (10), because READ (6) has less SCSI protocol overhead. But READ (6) was deprecated long ago, and is not supported by all devices anymore. So a driver may dynamically react on this command's availability. Same with READ (12) which is also deprecated and was replaced by READ (16).
A hard disk tool may offer different formatting options depending on drive capabilities. Some drives support changing the physical sector size when formatting, for instance, others don't. MODE SENSE often provides the required information (changeable fields). A tool may then dynamically decide on how to format or which options to offer based on a device's capabilities.

Of course, all mandatory commands and parameters from the respective standard have to be supported, and each device signals in its INQUIRY data which level of the standard it supported. There are not that many mandatory commands, though.

einstein95 · 2022-06-16T19:03:26Z

It appears you've got the wrong response for page 0x30 comparing it to the one in MAME (which does work with stock AppleCD drivers). See https://github.com/mamedev/mame/blob/badecba0e3aa372b75c738a4c53b2a7f8332064e/src/devices/bus/nscsi/cd.cpp#L427

rdmark · 2022-07-08T03:03:14Z

@einstein95 Thanks for the tip! In addition to the MODE SENSE piece here, there seems to be an aspect of masquerading with a particular product string in the MAME solution: mamedev/mame@c58e4b2

akuker added enhancement New feature or request bug Something isn't working and removed enhancement New feature or request labels Jul 8, 2020

akuker mentioned this issue Jul 10, 2020

RaSCSI crashes when MacOS reboots #2

Closed

akuker added the Hacktoberfest label Sep 30, 2020

akuker pushed a commit that referenced this issue Dec 28, 2021

#1 Apple CD-ROM driver expects a response to mode page 48 and 63. Mov…

7e6e40d

…ed from HD to Disk

akuker pushed a commit that referenced this issue Dec 30, 2021

#1 Fix size of the ModeSense6 response for Page Code 1. Adjust the de…

5fabdbf

…bug logging verbosity

akuker mentioned this issue Apr 1, 2022

Loading issue on Akai S1100 #741

Open

rdmark added compatibility Compatibility with particular computing platforms and removed bug Something isn't working Hacktoberfest labels Aug 19, 2022

rdmark mentioned this issue Sep 4, 2022

[QUESTION] #831

Closed

akuker mentioned this issue Sep 6, 2022

Support for native Apple CDROM driver #835

Merged

uweseimet linked a pull request Sep 6, 2022 that will close this issue

Support for native Apple CDROM driver #835

Merged

uweseimet closed this as completed Sep 10, 2022

rdmark mentioned this issue Oct 24, 2022

Web UI: More file path sanitation, better network bridge warnings, each endpoint return one message #932

Merged

benjamink mentioned this issue Nov 4, 2023

DaynaPort networking no longer working after latest build of 'develop' #1306

Closed

Jeremiahdmass mentioned this issue Apr 8, 2024

Macintosh Plus not Seeing PiSCSI #1439

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

CD-ROM Emulation doesn't work with stock Apple drivers #1

CD-ROM Emulation doesn't work with stock Apple drivers #1

akuker commented Jul 8, 2020

erichelgeson commented Oct 12, 2020

rdmark commented Aug 17, 2021

rdmark commented Aug 17, 2021 •

edited

Loading

rdmark commented Aug 18, 2021

uweseimet commented Aug 21, 2021

rdmark commented Aug 22, 2021 •

edited

Loading

rdmark commented Aug 22, 2021

uweseimet commented Aug 22, 2021

rdmark commented Aug 22, 2021

uweseimet commented Aug 22, 2021

rdmark commented Aug 22, 2021

uweseimet commented Aug 22, 2021

uweseimet commented Aug 23, 2021

rdmark commented Aug 23, 2021

uweseimet commented Aug 23, 2021

rdmark commented Aug 23, 2021

uweseimet commented Aug 23, 2021

akuker commented Dec 30, 2021

akuker commented Dec 30, 2021

uweseimet commented Dec 30, 2021 •

edited

Loading

einstein95 commented Jun 16, 2022

rdmark commented Jul 8, 2022

CD-ROM Emulation doesn't work with stock Apple drivers #1

CD-ROM Emulation doesn't work with stock Apple drivers #1

Comments

akuker commented Jul 8, 2020

erichelgeson commented Oct 12, 2020

rdmark commented Aug 17, 2021

rdmark commented Aug 17, 2021 • edited Loading

rdmark commented Aug 18, 2021

uweseimet commented Aug 21, 2021

rdmark commented Aug 22, 2021 • edited Loading

rdmark commented Aug 22, 2021

uweseimet commented Aug 22, 2021

rdmark commented Aug 22, 2021

uweseimet commented Aug 22, 2021

rdmark commented Aug 22, 2021

uweseimet commented Aug 22, 2021

uweseimet commented Aug 23, 2021

rdmark commented Aug 23, 2021

uweseimet commented Aug 23, 2021

rdmark commented Aug 23, 2021

uweseimet commented Aug 23, 2021

akuker commented Dec 30, 2021

akuker commented Dec 30, 2021

uweseimet commented Dec 30, 2021 • edited Loading

einstein95 commented Jun 16, 2022

rdmark commented Jul 8, 2022

rdmark commented Aug 17, 2021 •

edited

Loading

rdmark commented Aug 22, 2021 •

edited

Loading

uweseimet commented Dec 30, 2021 •

edited

Loading