prusa3d / Prusa-Firmware-Buddy

Firmware for the Original Prusa MINI, Original Prusa MK4 and the Original Prusa XL 3D printers by Prusa Research.
Other
1.13k stars 218 forks source link

[BUG] BSOD when trying to print anything #3275

Closed Gyrxiur closed 2 months ago

Gyrxiur commented 11 months ago

Printer type - [MINI]

Printer firmware version - [5.1.0-alpha1]

Original or Custom firmware - [Original]

Optional upgrades - [Filament Runout Sensor]

USB drive Please let us know if you print via USB flash drive or USB/Octoprint

Describe the bug When I try to print any file, with or without input shaper, when I open it in printer menu, I get BSOD, also I get BSOD when trying to print it through PrusaLink.

How to reproduce Update firmware to 5.1.0-alpha1 and try to print anything.

Expected behavior It should open up file preview in printer and allow me to print or be able to print through PrusaLink.

G-code Not needed, it does with every gcode older firmware had no problems, or with new with fully updated Prusaslicer.

Crash dump file https://ulozto.net/file/myRkfsbwy4zV/dump-buddy-bin

Gyrxiur commented 11 months ago

Ok, so I have found out, it has something to do with a lot of files on flash drive, if I have only few, it works.

Gyrxiur commented 11 months ago

And also, it likes to crash when connecting flash drive, better to turn it off, plug it in and turn on.

LukasLendvorsky commented 11 months ago

@Gyrxiur Can you please attach gcode you are trying to print?

Sorry about closing this issue, this it was missclick

murk-sy commented 11 months ago

The BSOD is on task:displayTask? dump_buddy.zip

Removing all other files and generating gcode without previews does not fix the issue.

Reformatting the USB drive and then putting all gcode files back fixes the issue.

It's probably related to long term USB flash drive data deterioration that I've dealt with many many times, see https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/1793

Gyrxiur commented 11 months ago

Yes, task:displayTask. I don't think it is deterioration, on previous firmware files work ok, even oldest ones. It has something to do with many files on drive, I have over 400 files on it, and also with speed, on faster drive, if I only have few files, they work ok, but if I copy all of the files, even the ones, previously working, no longer work. And on that slower drive, even if I keep only few files, some of them load, some don't, randomly. Weird. But as I said, with 4.4.1 same files work.

Gyrxiur commented 11 months ago

@Gyrxiur Can you please attach gcode you are trying to print?

Sorry about closing this issue, this it was missclick

It doesn't matter what file it is. Even that IS benchy doesn't open if I have a lot of files on drive. Also, I have BSOD even when I connect drive with a lot of files, and I have tried reformatting it and copying same files to it again. Drive is ok, works fine with 4.4.1 fw. There is probably some issue when I have a lot of files on flashdrive, which is worse on slower drive.

calrock27 commented 11 months ago

I have the same issue as well. Newly flashed 5.0.1 alpha1 on my Mini+. Every Gcode file generates a crash, even previously known working files and the presliced benchy files in the 5.0.1 alpha1.

BSOD code: task:displayTask?

Dump: dump_buddy.zip

This is utilizing an 11 month old Sandisk 64GB flash drive with approx. 20 sliced files on the drive. https://www.amazon.com/gp/product/B01EZ0X3L8/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1

calrock27 commented 11 months ago

I have the same issue as well. Newly flashed 5.0.1 alpha1 on my Mini+. Every Gcode file generates a crash, even previously known working files and the presliced benchy files in the 5.0.1 alpha1.

BSOD code: task:displayTask?

Dump: dump_buddy.zip

This is utilizing an 11 month old Sandisk 64GB flash drive with approx. 20 sliced files on the drive. https://www.amazon.com/gp/product/B01EZ0X3L8/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1

Interestingly enough - I now have a print that is able to be started. I reformatted the drive using a fat32 specific tool for larger drives and tried again. https://pendriveapps.com/format-fat32-larger-than-32gb/. Does anyone else care to try? Perhaps this is a drive format issue?

jwidess commented 11 months ago

Had the same problem, BSOD with USB drive that had ~100 Gcode files. Tried deleting them all and still had a crash with just the single newly sliced file. I had to reformat my 16GB flash drive for it to work without BSOD.

Dump from my final crash with only the single newly created Gcode file, before formatting. dump_buddy.zip Display Dump

DFEvans commented 11 months ago

I have found that this issue seems to be at least partly caused by the "Format of G-code thumbnails" option for my Minis. Changing it to the new "QOI" option and setting the G-code thumbnails option to 16x16, 220x124, 200x240 (per the new presets for IS printing) seems to be a lot more reliable. PNG thumbnail prints, the old default, are troublesome, even with low file counts - I bumped into this issue with a USB containing only four GCode files.

I'm not sure if this applies to printers other than the Mini, though - did QOI thumbnails get adopted for those, too?

dpenezic commented 11 months ago

I have the same issue, however, I did some more tests and found the following:

Altogether it is impossible to print anything, and I use an original Prusa USB stick.

It is look like some issue with FET stack and manipulating with files ... maybe some buffer overflow

murk-sy commented 11 months ago

@dpenezic Can you try doing what I did? Copy files off the drive, reformat it and then copy all the files back. Afterwards it worked perfectly fine for me.

I've long suspected the filesystem to be the cause of a lot of my woes over the years and I feel this is just another one of them.

DFEvans commented 11 months ago

Further to my above comment, I've now hit this with 4-5 QOI files, so either the thumbnails are barely related, or it was simply chance. I've downgraded my firmware again for the time being, since there's no clear workaround other than cleaning the USB every couple of prints.

dpenezic commented 11 months ago

@murk-sy I did try with a completely new from-box USB stick and with only one test file (benchy 16m), without success. What may be an issue is fact that I do everything on Linux , and don't have any MS Win comp near to try format USB stick.

Additionally, I don't see a direct connection, because I have some experience with micro/midi controllers and most of the time is a work or not-work state of reading ... in this case, because of optimization, it is possible that some stack/buffer overflow happened .

vinscarlata commented 11 months ago

+1 for BSOD with 400+ gcode files. Mine is properly FAT32 formatted (using rufus). I think someone asked that.

In my case, after the BSOD started happening with my main USB stick, I restoring the 4.4.1 firmware and was able to print fine from that USB stick again. So whatever is wrong with the USB stick, it only bothers the alpha firmware, not 4.4.1.

Found this bug report. I reinstalled the alpha firmware, deleted all the extra files from the USB stick and it did not fix it, but I did notice that a bunch of the gcode files that were previously transferred via wifi by Prusa Slicer had jacked up timestamps in 2097.

Copied a failing gcode to another USB stick and it worked fine. Re-formatted the problematic USB stick, and I can print from it fine as well.

sauloquirino commented 11 months ago

Same issue here

michalmiddleton commented 11 months ago

I had this happen as well. BSOD popped up during (or near the end) of upload via PrusaLink. Using the original USB drive with about 417 files on it. I just retried and it worked the next time. Crash dump attached. dump_buddy.bin.zip

Not sure if it matters but I am also affected by https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/3156 , however I am now using the original USB drive and no longer suffer from the data corruption. IMG_3816

github-actions[bot] commented 3 months ago

This issue has been flagged as stale because it has been open for 60 days with no activity. The issue will be closed in 7 days unless someone removes the "stale" label or adds a comment.

github-actions[bot] commented 2 months ago

This issue has been closed due to lack of recent activity. Please consider opening a new one if needed.