SD EEPROM: Create eeprom.dat if not found on SD

[pinchies]

fix for SD as eeprom - create file if no eeprom.dat file found.

[pinchies]

I didn't have any problems with the root directory thing we discussed, but I found there was no eeprom.dat created automatically if it wasnt present

[tpruvot]

seems weird to me, we never had to create it manually...

[pinchies]

I didn't either, until the recent builds. I can't figure out why, but 3am me probably was missing something in one of the revisions. It was repeatable for me however, I do think this PR is required.

[pinchies]

I now guess I wonder if adding O_CREAT permission to the initial open would make it return a success.......

[tpruvot]

I dont see why that could help to store junk and return true

[pinchies]

Because if no file is found, the PersistentStore::access_start() aborts the save eeprom process - the same as if no card was detected, so the PersistentStore::access_finish() is never reached, and so the correct file is never created.

But I wonder if adding O_CREAT permission to the initial file.open would achieve the same result.

Regardless, I'm just sharing my findings after people informed me that they had problems, and I diagnosed the cause. Not sure what the best way to patch this bug is.

[tpruvot]

format the sdcard imo, seems an hardware to me

[pinchies]

Tried multiple sd cards, formatted, with files or completely blank, multiple printers, checked write switch etc etc. Repeatable.

[tpruvot]

so you are saying the eeprom data is reloaded before save ? cant be...

[tpruvot]

hmm maybe it is... need to read the shared ocde

[pinchies]

I'm not claiming to be able to write C++ code well any better than I can claim to speak french, but I can at least read some of it :-P

[tpruvot]

i dont see any abort on access_start failure.. just a serial message warning

[tpruvot]

oh i see it now:
#define EEPROM_START() if (!persistentStore.access_start()) { SERIAL_ECHO_MSG("No EEPROM."); return false; } \

[tpruvot]

commit fdbc733

[tpruvot]

but its not the right solution.. this commit... you dont have to create the file to fix that... just need to return true

[tpruvot]

rmm im currently testing the flash eeprom PR, and maybe my issue is the same one... :p
edit: no

bool PersistentStore::access_start() {
  firstWrite = true;
  return true;
}

[pinchies]

Ahhhh, that was the breaking commit! That makes sense now.

[pinchies]

tagging #14776

[tpruvot]

yep, sorry i made tests recently but was before this change... and i was confused about your issue

[pinchies]

I appreciate that you're looking into flash eeprom, would be even nicer to have that working too.

[tpruvot]

not won for now, the PR didnt work on any of my 2 different boards.. (skrmini/longer3D)
edit: in fact it worked... on the skrmini hmm, maybe some hope

else the sd fix to appy: tpruvot@b97409d

[trouch]

⚠️ test_index isn't initialized to EEPROM_OFFSET
But we won't need it anymore once my PR #14829 get merged.

[pinchies]

@trouch not sure I understand your comment in relation to this PR

I think that simply returning true is the wrong solution as if there is another failure to open the file we do want it to return false.

[tpruvot]

see the other implementations, they return most true by default... doing nothing on start.
if (!card.isDetected()) return false; remains for the common case..

and the write may return false in case of write problem anyway

[pinchies]

Fair enough - I guess there’s no need for it to be in an if statement at all then if we don’t care about the returned value being false?

[yangwenxiong]

The version I downloaded in July. Don't create eeprom. Dat file.

[pinchies]

@yangwenxiong - if you do the top change at this link it should work:
tpruvot@b97409d

[yangwenxiong]

oh i see it now:
#define EEPROM_START() if (!persistentStore.access_start()) { SERIAL_ECHO_MSG("No EEPROM."); return false; } \

For the MCU of STM32F103xx series, it is necessary to build a file EEPROM.DAT in SD card in advance to simulate EEPROM data storage in SD card.Otherwise it's not gonna work and you're going to have a problem with it,

[yangwenxiong]

oh i see it now:
#define EEPROM_START() if (!persistentStore.access_start()) { SERIAL_ECHO_MSG("No EEPROM."); return false; } \

For the MCU of STM32F103xx series, it is necessary to build a file EEPROM.DAT in SD card in advance to simulate EEPROM data storage in SD card.Otherwise it's not gonna work and you're going to have a problem with it,

#define EEPROM_START() int eeprom_index = EEPROM_OFFSET; persistentStore.access_start()

[pinchies]

You don’t need to do that. You do need to change the file containing the access_start()

[tpruvot]

This one can be closed now.. (you can delete the branch pinchies:fix-sd-no-file-eeprom @pinchies to do it)

[pinchies]

I don’t see the issue fixed yet?

[tpruvot]

commit 92c35d1