i never did try anything with my big ben card, but i can test this. i assume reading works ( this is ogc_readsegment? ) , since it has the same code/bitwise operations there?

@suloku : you seem to have pulled in changes from elsewhere in the branch of your open pull request (in this case main).
you'll need to either remove the commits, or remake this pull request using a different branch that doesn't have all this noise.

Corrected, everything should be fine now.

Actually I was consistently failing to write at block 1024 (or more precisely, content wasn't written to block 1024,). After some tests I noticed the content was being written to block 0 onwards after block 1024, then I noticed card_writepage and card_readsegment had different calculations for the offset card command.
I did some offset calcultations for blocks 0->2048 on a simple C snippet, that's where I confirmed the error and noticed card_readsegment was correctly doing the job. I made the changes and afterwards it worked fine.
I'm pretty sure a simple gci restore would also fail, I'll confirm later by writing a 1025 block file to an empty memory card and then trying to restore another gci file. This will most likely corrupt the card.

confirmed by testing the following :

• make backup with GCMM
• alter backup in hex editor past 0x800000
• restore backup (old libogc gave an error in GCMM, i suspect because it couldn't verify the byte being written correctly)
• backup again
• verify in last backup altered data is there

thanks for the contribution !