search

BlueSCSI / BlueSCSI-v2

Open source, open hardware, SCSI emulator using the Pi Pico PR2040
https://bluescsi.com
GNU General Public License v3.0
224 stars 23 forks source link

Random write errors -SDIO SD card error on line 413, error code 2 #34

Open z80micro-mc opened 1 year ago

z80micro-mc commented 1 year ago

I am getting occasional write errors when I copy a number of files.

log.txt

[10ms] Platform: BlueSCSI Pico [10ms] FW Version: 2023.04.10-dev Apr 30 2023 15:09:12 Flash chip size: 2048 kB

=== SD Card Info === SD card detected, FAT32 volume size: 7615 MB SD Name: ASTC, MID: 0xFE, OID: 0x34 0x56

=== Global Config === Reading configuration from bluescsi.ini Active configuration (using system preset "Generic"): -- SelectionDelay: 150 -- EnableSCSI2 is off -- Parity is disabled

=== Finding images in / === == Opening /HD10_512.hda for ID: 1 LUN: 0 ---- Image ready == Opening /HD20_512.hda for ID: 2 LUN: 0 ---- Image ready

=== ROM Drive === Platform supports ROM drive up to 1692 kB ---- ROM drive SCSI id 1 is already in use, not enabling

=== Configured SCSI Devices ===

Initialization complete! SDIO SD card error on line 413, error code 2 SDIO SD card error on line 413, error code 2 SDIO SD card error on line 413, error code 2 SDIO SD card error on line 413, error code 2 SDIO SD card error on line 413, error code 2 SdioCard::stopTransmission() timeout SD card write failed: 0x00 SDIO SD card error on line 413, error code 2 SDIO SD card error on line 413, error code 2 SdioCard::stopTransmission() timeout SD card write failed: 0x00 SDIO SD card error on line 413, error code 2

Thoughts / ideas please

Peter

Konf commented 1 year ago

Hi! I had such errors with old and slow SD card. Replaced it with new a one and no such errors appear anymore.

z80micro-mc commented 1 year ago

Konf,

Thanks for your input.

I am currently using an 8 GB, C10, U1 Kingston micro SD rated at 80 MB/s. I obtain ~ 180 KB/s read or write via DOS. The SD is nearly 200 times faster than the transfer speed. On a large number of writes (copy all the files on a disk to another disk) I intermittently get write errors ~ every once every 5-10 minutes, at the current transfer rates I estimate ~ 1 failure in ~ 50,000 write sector operations.

To my reasoning is that is rather indicative of a timing race condition. A faster SD may reduce the likelyhood of errors but I think that there is a fundamental timing race condition. I would be tempted to slow the SD right down to make the condition worse to investigate further.

To date I have only ever seen, errors on write and repeating the same write subsequently succeeds. I have never, seen a failure on read. The write failure gets mapped to a general failure reported by DOS, if one selects the 'r' repeat operation it proceeds without error until the next random event.

Peter

z80micro-mc commented 1 year ago

Eric,

If it would be helpful, I could put a logic analyser on the SCSI bus. However, as the event is so infrequent, it would be helpful if there is a spare gpio bit which you could toggle on error. I could then trigger the analyser on the bit change.

The analyser is capturing all the time and can be set to stop on trigger event, well 25% buffer before or after the event selectable.

Peter

erichelgeson commented 1 year ago

No free pins but you can re-purpose SDBG, S1, or S2 since they are normally unused. Interested in any results you get.

z80micro-mc commented 1 year ago

Eric,

OK dev version with debug error pulse on S1 / S2 pin please. It needs to have Troys mods in for parity disable.

Peter

------ Original Message ------ From: "Eric Helgeson" @.> To: "BlueSCSI/BlueSCSI-v2" @.> Cc: "z80micro-mc" @.>; "Author" @.> Sent: Saturday, 6 May, 23 At 17:32 Subject: Re: [BlueSCSI/BlueSCSI-v2] Random write errors -SDIO SD card error on line 413, error code 2 (Issue #34)

No free pins but you can re-purpose SDBG, S1, or S2 since they are normally unused. Interested in any results you get. — Reply to this email directly, view it on GitHub https://github.com/BlueSCSI/BlueSCSI-v2/issues/34#issuecomment-1537177104 , or unsubscribe https://github.com/notifications/unsubscribe-auth/ATCK64FFPENKTEQZHO4Z7NLXEZ4K5ANCNFSM6AAAAAAXVEKVAI . You are receiving this because you authored the thread.Message ID: @.***>

erichelgeson commented 1 year ago

I had to use the SDBG (pin 16) - it will go high on an SD error. If you'd like a pulse (how long) or delay or something let me know. https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/4902832099 <-- binaries at the bottom (includes troy's patch, and the typo fixed for ParityEnable) https://github.com/BlueSCSI/BlueSCSI-v2/commit/227baab4b654d54f0d333320cdba79beb3886591 <-- change

z80micro-mc commented 1 year ago

Eric,

That should be fine.

When is the bit cleared? Might I presume on the next valid operation

Peter .

------ Original Message ------ From: "Eric Helgeson" @.> To: "BlueSCSI/BlueSCSI-v2" @.> Cc: "z80micro-mc" @.>; "Author" @.> Sent: Saturday, 6 May, 23 At 18:39 Subject: Re: [BlueSCSI/BlueSCSI-v2] Random write errors -SDIO SD card error on line 413, error code 2 (Issue #34)

I had to use the SDBG (pin 16) - it will go high on an SD error. If you'd like a pulse (how long) or delay or something let me know. https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/4902832099 https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/4902832099 <-- binaries at the bottom (includes troy's patch) 227baab https://github.com/BlueSCSI/BlueSCSI-v2/commit/227baab4b654d54f0d333320cdba79beb3886591 <-- change — Reply to this email directly, view it on GitHub https://github.com/BlueSCSI/BlueSCSI-v2/issues/34#issuecomment-1537188867 , or unsubscribe https://github.com/notifications/unsubscribe-auth/ATCK64DT53ROY2V3RY7C7O3XE2EDXANCNFSM6AAAAAAXVEKVAI . You are receiving this because you authored the thread.Message ID: @.***>

erichelgeson commented 1 year ago

This build will clear the assertion on next success read or write - https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/4902950485

z80micro-mc commented 1 year ago

Eric,

Many thanks for the debug version, I'll dig out the analyser tomorrow and see what I can capture.

Are there any particular signals that you would like to see in the capture?

I have 8, 16, 32, or 48 channels depending upon which analyser I use. The 8, 16 & 32 chanels I can get into Sigrok Pulseview so you can browse them your self. The 48 channel analyser is a Tek 1225, which is perhaps easier to use but has limited export capabilities.

Peter

------ Original Message ------ From: "Eric Helgeson" @.> To: "BlueSCSI/BlueSCSI-v2" @.> Cc: "z80micro-mc" @.>; "Author" @.> Sent: Saturday, 6 May, 23 At 19:10 Subject: Re: [BlueSCSI/BlueSCSI-v2] Random write errors -SDIO SD card error on line 413, error code 2 (Issue #34)

This build will clear the assertion on next success read or write - https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/4902950485 https://github.com/BlueSCSI/BlueSCSI-v2/actions/runs/4902950485 — Reply to this email directly, view it on GitHub https://github.com/BlueSCSI/BlueSCSI-v2/issues/34#issuecomment-1537194552 , or unsubscribe https://github.com/notifications/unsubscribe-auth/ATCK64GWTQU6JS3WGQZKL43XE2H3HANCNFSM6AAAAAAXVEKVAI . You are receiving this because you authored the thread.Message ID: @.***>

erichelgeson commented 1 year ago

Please test with the latest release and report back, there were various fixes related to SD cards.

erichelgeson commented 3 months ago

Hi, any update or anything for us to follow up on? We've seen this a few times and usually a full overwrite format with SD formatter as mentioned takes care of it, or cleaning the contacts on the SD card (can get dirty!), or trying another SD card.