Return 'false' on EEPROM connection fail

[trouch]

Requirements

• STM32F1 MCU
• #define SD_DETECT_PIN

Description

With FYSETC Cheetah board, EEPROM file doesn't get loaded when booting.
After debugging with SWD probe, I found the first_load function sets loaded=true no matter of the access return, avoiding to properly load EEPROM file when SD got detected.

Without this fix, here is the console output:

[Connected]
start
echo:PowerUp
Marlin 2.0.0-RC0

echo: Last Updated: 2019-07-19 | Author: (trouch, config)
echo:Compiled: Jul 27 2019
echo: Free Memory: 28631  PlannerBufferBytes: 1728
echo:EEPROM version mismatch (EEPROM= Marlin=V68)
echo:Hardcoded Default Settings Loaded
Testing X connection... OK
Testing Y connection... OK
Testing Z connection... OK
Testing E connection... OK
echo:SD card ok

With that fix:

[Connected]
start
echo:PowerUp
Marlin 2.0.0-RC0

echo: Last Updated: 2019-07-19 | Author: (trouch, config)
echo:Compiled: Jul 27 2019
echo: Free Memory: 28631  PlannerBufferBytes: 1728
echo:Hardcoded Default Settings Loaded
Testing X connection... OK
Testing Y connection... OK
Testing Z connection... OK
Testing E connection... OK
echo:SD card ok
echo:V68 stored settings retrieved (614 bytes; crc 176)

IDK if that's the right approach to fix it, but it's the simpler one I found to avoid extra pre-compiler conditions and make it a general improvement.

[thinkyhead]

This will work. Ultimately it will be best for Marlin to be savvy enough to know when it can talk to EEPROM so it will not even try until that point. Initialization of TMC drivers (and a few other things) is tied up with EEPROM, so the more reliable that can be at startup, the better.