Blue SCSI v2 SDCard CRC Errors and Other Issues...

[BFEXTU]

@eric

Hi Eric - I have a new, barely-used Sandisk SDCard and it is generating the CRC errors shown below after formatting with Overwrite, etc. The card is a genuine SanDisk Ultra 16GB Class 10 A1 card that I have confirmed as genuine with SanDisk. It is capable of writing at 20MB/s (on Windows) and I have filled the card on the PC side with no errors of any kind. Also, I didn't see any errors during formatting.

Platform: BlueSCSI
FW Version: 2024.05.21-rel May 21 2024 21:21:44
Flash chip size: 2048 kB

=== SD Card Info ===
SD card detected, exFAT volume size: 15186 MB
SD Name: SC16G, MID: 0x03, OID: 0x53 0x44

=== Global Config ===
Config file bluescsi.ini not found, using defaults
Active configuration:

=== Finding images in / ===
== Opening /HD50_512 Blank4Gb.hda for ID: 5 LUN: 0
---- WARNING: This image does not appear to be a valid Macintosh Device image. See: https://github.com/BlueSCSI/BlueSCSI-v2/wiki/Disk-Images
---- Image ready
== Opening /HD40_512 Compilers2Gb HFS.hda for ID: 4 LUN: 0
---- Image ready
== Opening /HD30_512 Main2Gb HFS.hda for ID: 3 LUN: 0
---- Image ready
== Opening /HD20_512 System2Gb.hda for ID: 2 LUN: 0
---- Image ready
== Opening /HD60_512 Blank4GB.hda for ID: 6 LUN: 0
---- WARNING: This image does not appear to be a valid Macintosh Device image. See: https://github.com/BlueSCSI/BlueSCSI-v2/wiki/Disk-Images
---- Image ready


(n.b. there are 2 unformatted drives above at IDs 5 and 6. Also, these drives are the first and last ones that it tries to open -- see below for misreported capacity bug on ID2 volume)

=== ROM Drive ===
Platform supports ROM drive up to 1692 kB
---- ROM drive image not detected

=== Configured SCSI Devices ===
* ID: 2, BlockSize: 512, Type: Fixed, Quirks: Apple, Size: 2048000kB
* ID: 3, BlockSize: 512, Type: Fixed, Quirks: Apple, Size: 2048000kB
* ID: 4, BlockSize: 512, Type: Fixed, Quirks: Apple, Size: 2048000kB
* ID: 5, BlockSize: 512, Type: Fixed, Quirks: Apple, Size: 4194304kB
* ID: 6, BlockSize: 512, Type: Fixed, Quirks: Apple, Size: 4194304kB

Initialization complete!
INFO: Pico Voltage: 3.273V.
SDIO card reports write CRC error, status 0x501FFFFF
SdioCard::writeSectors(0x0093789D,...,35) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0x501FFFFF
SdioCard::writeSectors(0x00B9C880,...,40) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x007D5BAB,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x007DD402,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x007DDF93,...,45) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x007E7A27,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x007D42BA,...,6) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x007F1721,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x007F8B0A,...,54) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60001FF
SdioCard::writeSectors(0x007F9797,...,41) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x007FA65B,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0080284B,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x00803985,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x00804326,...,26) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x008048C7,...,57) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x00808026,...,26) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0080A414,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0080AC72,...,14) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x0080C784,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0080D2A4,...,28) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x0080E09E,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0080E410,...,48) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0081784A,...,54) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0081BA18,...,40) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0081F199,...,39) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x00824F0D,...,51) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0082B765,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60001FF
SdioCard::writeSectors(0x0082CF44,...,60) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x0083103C,...,4) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x00838C44,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x008431A4,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x00845220,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xD60000FF
SdioCard::writeSectors(0x008518B0,...,16) failed: 7
SD card write failed: 0x00
SDIO card reports write CRC error, status 0xAC0001FF
SdioCard::writeSectors(0x00854AB0,...,16) failed: 7
SD card write failed: 0x00

Other than a (maybe not) bad SDCard, is there anything else that could be causing this issue? Is it possible that BlueSCSI v2 has a write timing or other waveform problem? It looks like a possible BlueSCSI v2 issue to me based on the testing that I have done so far (no problems on Windows at much higher R/W transfer rates). Windo

Also, the problem with slow throughput in my other thread relates to the FWB HDT 3.0.2 driver. Both the older BlueSCSI and the BlueSCSI v2 do NOT work well with FWB 3.0.2. The throughput is about 1/2-1/4 of what it should be. On v2, if I use the Apple 4.3 driver, R/W transfers are 5-6MB/s on my Q950 on the external bus. If I use FWB 3.0.2 from Hard Disk Toolkit as the driver for the test volume, stats drop to 1.5MB/s. However, this same driver works great on a physical Cheetah hard drive and writes at 9MB/s. So, there is some incompatibility between all BlueSCSIs and FWB 3.0.2. You may want to mention this FWB 3.0.2 issue on the wiki.

The other obvious issue is that for the ID2 volume, BlueSCSI v2 appears to be reporting incorrect capacity to the OS that belongs to one of the other volumes. There are 2 unformatted 4Gb drives at ID5 and ID6 and you can see from the configuration table that ID2 is only 2Gb. Anyway -- looks like there might be a firmware bug. Is it possible this issue is also related to the CRC problem?

I'll try another SD Card, go through the Overwrite process and then reformat as exFAT (since the formatter only supports FAT32). I'll let you know what I see. I will also look at the pico and card slot soldering. I will also try this same card in a ZuluSCSI board and see if shows any errors.

And finally, one UX thing to double-check -- I put the latest firmware .bin image on the card to make sure that it was updated before first run. However, my board was already current and the boot process just left the .bin image there. I didn't check the log, but my suggestion would be to delete the unneeded image if it is the same as the installed version and put a note in the log. The firmware left the image there even though it wasn't needed and the user expectation was that it should be used for update and deleted. The other defensive option would be to do a compare if the version is the same and reflash if the installed version doesn't match the .bin source (then delete the image).

 

[BFEXTU]

I used the same SDCard (no changes) in a ZuluSCSI with default settings. It correctly sees the capacity as 2Gb on the partition inb question, and I did equivalent transfers to the same area of the SD Card with no reported CRC errors. I duplicated the files on the same partition, then copied them to a physical drive, then trashed the original files, then recopied them from a physical drive to the same partition to try to replicate the original CRC failures. If the card were bad or marginal, would ZuiluSCSI report errors with default settings?

[12ms] Platform: ZuluSCSI RP2040
[12ms] FW Version: 24.05.17-release May 17 2024 16:24:20
[13ms] DIP switch settings: debug log 0, termination 1
[13ms] SCSI termination is enabled
[14ms] Flash chip size: 2048 kB
[14ms] SCSI target/disk mode selected by DIP switch, acting as a SCSI disk
[143ms] SD card detected, FAT64 volume size: 15186 MB
[143ms] SD MID: 0x03, OID: 0x53 0x44
[144ms] SD Name: SC16G
[144ms] SD Date: 6/2022
[144ms] SD Serial: 0x5CF5DC28
[175ms] Config file zuluscsi.ini not found, using defaults
[204ms] Active configuration:
[204ms] -- SelectionDelay = 255
[204ms] -- EnableUnitAttention = No
[204ms] -- EnableSCSI2 = Yes
[205ms] -- EnableSelLatch = No
[205ms] -- MapLunsToIDs = No
[205ms] -- EnableParity = Yes
[402ms] Finding images in directory /:
[402ms] -- Using custom block size, 512 from filename: HD50_512 Blank4Gb.hda
[420ms] -- Opening /HD50_512 Blank4Gb.hda for id:5 lun:0
[422ms] ---- Configuring as disk drive drive
[422ms] ---- Read prefetch enabled: 8192 bytes
[422ms] -- Using custom block size, 512 from filename: HD40_512 Compilers2Gb HFS.hda
[440ms] -- Opening /HD40_512 Compilers2Gb HFS.hda for id:4 lun:0
[442ms] ---- Configuring as disk drive drive
[442ms] ---- Read prefetch enabled: 8192 bytes
[443ms] -- Using custom block size, 512 from filename: HD30_512 Main2Gb HFS.hda
[461ms] -- Opening /HD30_512 Main2Gb HFS.hda for id:3 lun:0
[463ms] ---- Configuring as disk drive drive
[463ms] ---- Read prefetch enabled: 8192 bytes
[463ms] -- Using custom block size, 512 from filename: HD20_512 System2Gb.hda
[481ms] -- Opening /HD20_512 Blank2GbApple.hda for id:2 lun:0
[483ms] ---- Configuring as disk drive drive
[483ms] ---- Read prefetch enabled: 8192 bytes
[484ms] -- Using custom block size, 512 from filename: HD60_512 Blank4GB.hda
[502ms] -- Opening /HD60_512 Blank4GB.hda for id:6 lun:0
[504ms] ---- Configuring as disk drive drive
[504ms] ---- Read prefetch enabled: 8192 bytes
[513ms] -- Platform supports ROM drive up to 1692 kB
[513ms] ---- ROM drive image not detected
[514ms] SCSI ID: 2, BlockSize: 512, Type: 0, Quirks: 0, Size: 2048000kB
[515ms] SCSI ID: 3, BlockSize: 512, Type: 0, Quirks: 0, Size: 2048000kB
[515ms] SCSI ID: 4, BlockSize: 512, Type: 0, Quirks: 0, Size: 2048000kB
[516ms] SCSI ID: 5, BlockSize: 512, Type: 0, Quirks: 0, Size: 4194304kB
[517ms] SCSI ID: 6, BlockSize: 512, Type: 0, Quirks: 0, Size: 4194304kB
[663ms] Initialization complete!
[69190ms] Toolbox enabled = 0

ZuluSCSI writes outperformed BlueSCSI v2 by 2x ~1000 vs. ~500 and write throughput at > 7MB/s vs. 6ish for BlueSCSI, but reads were slower than those of BlueSCSI...but there must be some setting because the access time was screwy -- 4000-something.

 

[BFEXTU]

I tried to reproduce the capacity bug with BlueSCSI v2, but could not. So -- who knows -- it could have been a system, driver, os or other problem. But, definitely look into it. It was strange that a good volume was reporting the capacity of one of the unformatted volumes.

However, I still see lots of CRC errors that ZuluSCSI did not report. I will check their .ini file to see if there is a setting and/or if they are reported by default.

See attached logfile.

 

[V.Yakob]

I have a similar problem on how it turned out on two BlueSCSIv2 different hardware versions.
At the same time, I contacted the seller and returned to him the device on which my PM8100 was hanging, he checked it on Windows -- and there was not a single problem.
Now I'm following this problem and notice that for the most part these errors are floating. That is, sometimes, even after long-term operation of computers, there may not be these errors in the log. And sometimes there are a lot of them.

I haven't bought other cards to check yet, but I'm going to do it. Now I have two 128 Gb SCXC V10

At the moment, it seems to me that such errors are more common on PM8100 when I use OS 7.1.2, 7.5.3, but I hardly see them when I use OS 8.6 or 9.1. But that's just an assumption.

What OS do you use?

 

[eric]

There's only been a few people who've had this error so it's a bit hard to diagnose (less than 10 total) - even when we've got a unit that was said to experience this issue we cant reproduce it with our SD cards or on our machines - so it's been a bit of an enigma. Luckily it seems pretty small set of users (no relief for you when it's you having the issue, but to me I'm glad it's just a few )

SanDisk has always been a terrible brand - even back in v1 days we'd recommend no one use them - so may issues go away when people switched to a different brand. The author of the SDFat library we use also recommends to avoid SanDisk.

To all the other behaviors you reported - until the first issue is resolved I don't think we need to investigate them. Speed I test on my 9600 each release with FWB and we're always on top by 1-2%.

I'm going to send you both a DM to get some details and start tracking so I can hopefully narrow down what factor is causing the issue.


Also I assume you've already gone through the troubleshooting guide, done a full overwrite format with the SD Card Associations SD Card Formatter tool, etc, etc.

 

[BFEXTU]

@eric Thanks. I replied. And thanks for the update @V.Yakob

It's always possible that I screwed something up or the card has an assembly issue. I just ordered 5 new SanDisk Max Endurance cards from Western Digital for comparison. If they don't work, I will return them. And I will double-check my build today.

All of my testing is on a Quadra 950 and my board is the newer/short DB-25 external.

FWB v3.0.2 works reliably with all the virtual emulators, but shows poor 25-50% performance, including with ZuluSCSI. So, whatever is happening relating to the slow-down inherits from the shared code base.

Why avoid SanDisk? They should (hopefully) be one of the most reliable brands (and also one of the brands with the most counterfeits). So - does "avoid SanDisk" mean to avoid SanDisk or "avoid the many SanDisk counterfeits?" However, it's true that marketing and perception may not always equal reality. Has there been some issue following the WDC acquisition in 2016? I wasn't aware of any problem -- other than the fact that there are many knock-offs -- like wtih the BluePill.

The most interesting datapoint for me is that BlueSCSI v2 reports CRC errors and ZuluSCSI/Windows do not. Do you know if ZuluSCSI reports CRC errors by default in the log when they occur? I will do some card swapping and see if the problems continue.

 

[eric]

FWB v3.0.2 works reliably with all the virtual emulators

What emulators? Most do not emulate SCSI so do not use those drivers. Only MAME, PCE, QEMU actually emulate SCSI.

 

[BFEXTU]

Sorry about that. I was referring to BlueSCSI v1, v2 and ZuluSCSI. I think of them as "emulators" because flash is mimicking legacy hard drives on these old machines and my roots are in the era of purely mechanical SCSI. But...a controller is a controller, even if it is attached to a non-moving object like an SD card.

Has the asteroid hit yet? -