Global EEPROM instance always returns 0

[LexNastin]

Board

ESP32 Dev Module

Device Description

N/A

Hardware Configuration

N/A

Version

v2.0.1

IDE Name

Arduino IDE

Operating System

Windows 11

Flash frequency

N/A

PSRAM enabled

no

Upload speed

115200

Description

Expected Behaviour

After running EEPROM.begin(2048) on the global EEPROM instance, EEPROM.length() should return 2048

Current Behaviour

After running EEPROM.begin(2048) on the global EEPROM instance, EEPROM.length() returns 0

Sketch

N/A

Debug Message

N/A

Other Steps to Reproduce

No response

I have checked existing issues, online documentation and the Troubleshooting Guide

• I confirm I have checked existing issues, online documentation and Troubleshooting guide.

[VojtechBartoska]

Hello, can you please retest this on v2.0.3-rc1?

[LexNastin]

I am not currently sure where my second test ESP is, so can't it test at the moment. When I find it I will try it out

[VojtechBartoska]

Thank you!

[VojtechBartoska]

Any news on this?

[LexNastin]

Oh shit sorry, I couldn't find my ESP so I didn't end up testing. Sorry...

[UnicodeMan]

Hi. I tested this today with platformIO, pressed update all (updating both libraries and boards, but the result is still that EEPROM returns 0 when used outside in other classes and such. In other words, behaviour is still not global.

Oh wait! Actually, adding EEPROM.begin(4096) in setup() screws up my serial lcd output. EDIT: its cpu panic
Could this be a clue as to what is happening here?

EndeavourOS (arch-Linux based distribution), Linux 5.17.5.

[lbernstone]

@UnicodeMan Are you testing the result of EEPROM.begin()? If it fails to initialize (perhaps your nvs partition is full), then length will still be 0. Setting the EEPROM length is one of the last things done in the begin function.

[UnicodeMan]

I tested the result. EEPROM.begin(4096) returns 1. EEPROM.length(), however, returns 0.
Im not sure about NVS partition being full or not, I reduced compiled binary by removing BT dependency, so its ~282kBytes.
If there are any specific instructions about checking whether nvs partition is being used by the program (or being used some other way), I'd be happy to cooperate and test it again.

[UnicodeMan]

Seeing as EEPROM is deprecated on ESP32, I decided to use Preferences library. It works great. So anyone having issues with EEPROM should consider migrating to Preferences (though I see why EEPROM is still used).

[LexNastin]

Seeing as EEPROM is deprecated on ESP32, I decided to use Preferences library. It works great. So anyone having issues with EEPROM should consider migrating to Preferences (though I see why EEPROM is still used).

Interesting. The reason I'm using it is because I tried to port over a project for the ESP8266 and got it working, but nothing would persist a restart. I just added lots of Serial.print statements to find what the culprit was in the saving code and it came down to this. Seeing as it wasn't any sort of dynamic number, I hardcoded the expected length, but a fix to the EEPROM library would be useful nonetheless.

[UnicodeMan]

Nevermind, same issue. It doesn't save the values.

Seeing as EEPROM is deprecated on ESP32, I decided to use Preferences library. It works great. So anyone having issues with EEPROM should consider migrating to Preferences (though I see why EEPROM is still used).

[lbernstone]

Checking the nvs status is more difficult than just erasing it. Open a command prompt and find esptool (usually %userdata%/Arduino15/packages/esp32/tools/esptool/) and run esptool erase_flash. That will wipe you whole device, and then you can upload your firmware again. If that fixes Preferences, then please test EEPROM again.