BlueSCSI Ultra Narrow doesn't work in USB MSC initiator mode

[csdexter]

What works:

• stand-alone initiator mode (imaging): sees the drives, waits for me to load a disc, starts copying, finishes successfully, ejects the disk
• plugging the SD into a PC afterwards and checking the image results in a bit-identical comparison with the original data

What doesn't work:

• USB MSC initiator mode (USB2SCSI bridge): sees the drives, enumerates them on the USB side and dies (watchdog reset) when the PC requests the cache mode page

Configuration:

• Hardware: BlueSCSI Ultra Narrow
• Firmware: 2026.04.27-rel Apr 28 2026 00:36:16
• SD Card: SanDisk Max Endurance 128GB fresh from the blister
• SCSI Targets: Fujitsu MCJ3230SS (MO, 2.3GB) and SONY SMO-F561 (MO, 9.1GB) -- both and the used media known good since they all work fine with an Adaptec AHA-2940UW in a PC, not to mention in stand-alone initiator mode as described above
• Jumpers: USB Initiator, Initiator Mode and Term EN ON; Debug OFF; Back Feed and Power On ON
• Power: via Molex connector, same PC PSU that powers the drive(s)
• bluescsi.ini looks like so:

[SCSI]
InitiatorMSC=true
InitiatorMode=true
EnableUSBMassStorage=false

I'm also attaching log.txt and err.txt from a failed attempt.
The dmesg from the PC looks like this when it fails:

[70928.447393] usb 1-3: new full-speed USB device number 15 using xhci_hcd
[70933.677752] usb 1-3: New USB device found, idVendor=1209, idProduct=2023, bcdDevice= 1.00
[70933.677759] usb 1-3: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[70933.677762] usb 1-3: Product: BlueSCSI
[70933.677765] usb 1-3: Manufacturer: BlueSCSI
[70933.677768] usb 1-3: SerialNumber: 123456
[70933.711834] cdc_acm 1-3:1.0: ttyACM0: USB ACM device
[70933.715795] usb-storage 1-3:1.2: USB Mass Storage device detected
[70933.715974] scsi host11: usb-storage 1-3:1.2
[70934.769864] scsi 11:0:0:0: Direct-Access     FUJITSU  MCR3230SS-S      0020 PQ: 0 ANSI: 2
[70934.770301] sd 11:0:0:0: Attached scsi generic sg7 type 0
[70947.832539] usb 1-3: USB disconnect, device number 15
[70947.842468] sd 11:0:0:0: [sdh] 1063146 2048-byte logical blocks: (2.18 GB/2.03 GiB)
[70947.842482] sd 11:0:0:0: [sdh] Write Protect is off
[70947.842485] sd 11:0:0:0: [sdh] Mode Sense: 00 00 00 00
[70947.842490] sd 11:0:0:0: [sdh] Asking for cache data failed
[70947.842493] sd 11:0:0:0: [sdh] Assuming drive cache: write through
[70947.851634] sd 11:0:0:0: [sdh] Attached SCSI removable disk
[70969.997354] sd 10:0:0:1: [sdd] 249737216 512-byte logical blocks: (128 GB/119 GiB)
[70970.001557] sdd: detected capacity change from 0 to 249737216
[70970.009204]  sdd: sdd1
[71092.104742] sdd: detected capacity change from 249737216 to 0

If there's any more debugging information I can provide, just let me know. Thank you for your time!

log-usbmsc-initiator.txt
errlog-usbmsc-initiator.txt

[erichelgeson]

It should be addressed in #394 please test and let me know.

[erichelgeson]

Found what I think is the issue - when we moved to cmake from PlatformIO/Arudino the buffer for tinyusb was 4096, but the default is 512 - so your 2048 block size cdrom failed here silently - https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/31221228510?pr=394 - I also think I addressed your questions on re-enumeration. Please give that build a try and let me know.

[erichelgeson]

@csdexter friendly ping to see if you can re-test the new build.

[csdexter]

Sorry for the delay, will re-test this evening and update the thread. Thank you!

[erichelgeson]

Latest firmware is now here: https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/31221228510?pr=394 - rebased that branch after merging in a few other items

[csdexter]

Works like a charm:

• removable media devices now enumerate even if empty
• detection of media insertion/requests for ejecting media work as expected from the USB side
• data transfer works just fine with an average of 415 KB/sec
• the debug (CDC-ACM) console stays very much alive during usage, no more tarpitting :-)

Thank you very much for all the help, I now have a working USB2SCSI solution for all my archiving work!

[erichelgeson]

Thank you for taking the time to report and test the issue! I'll merge shortly and then plan to start release regression testing.

[csdexter]

Resolved by firmware version linked from thread.