Peter Andersson

Hmm.. If you run a SPIFFS_check, is it recovered? Guess not, as you get ERR_DELETED, but still.. Also, do you have a code example of how to reproduce this? I...

Hi @rojer, looked into it a little. The fact that an untouched file becomes corrupt during an aborted GC is really no mystery, as all files in the block being...

No, that's not very ideal I agree. There is some info on it here: https://github.com/pellepl/spiffs/wiki/FAQ#powerlosses-contd-when-should-i-run-spiffs_check I do not know what system or what application you're running it on, but if...

The GC cannot be modified to not require a check, as it is now. Unfortunately, a check can be very expensive - it depends on how big the fs is...

Sound bad, having a check corrupting things is a tad counter intuitive to say the least. Yes, please file as another issue. All info you can collect is valuable. Cheers...

Good to hear. Did you ind out what the problem was?

Thanks for a detailed report. I'll try to reproduce it here. Cheers / P

Hmm, I cannot reproduce it here. This suggests it is something weird with your spi flash hal functions. Actually, this rings a bell. I remember someone using an Arduino spiflash...

SPIFFS_ALIGNED_OBJECT_INDEX_TABLES should help - but if you're using IAR I guess [this line of code](https://github.com/pellepl/spiffs/blob/master/src/spiffs_nucleus.h#L479) is not interpreted by IAR as alignment. You probably need to change this directive (and...

> If I understand this correctly, __attribute(( aligned(sizeof(spiffs_page_ix)) )) just says that the struct has to start at an aligned address. Actually, this also enforces not only that the struct...