dhara_map_capacity returns inconsistent values about dhara HOT 4 OPEN

Thanks for replying. Thats exactly what I was thinking of doing! A conservative value that allows for a bunch of bad blocks in the future :-)

from dhara.

Hello there, some updates:

We manage to isolate the issue to dhara_journal_capacity() in particular

const dhara_block_t max_bad = j->bb_last > j->bb_current ?
		j->bb_last : j->bb_current;

For some reasons max_bad is 32 after the first format and then it goes to 0 after we do some writes and we experience the issue afore mentioned.

from dhara.

Just hit a very similar issue to this myself....my disk appeared to get larger after writing some data.

Following your hint above, it seems this behaviour may be normal.

When initialising a map we see a call flow dhara_map_init -> dhara_journal_init -> reset_journal

Inside there we prepare the last and current bad blocks:

/* We don't yet have a bad block estimate, so make a
 * conservative guess.
 */
j->epoch = 0;
j->bb_last = j->nand->num_blocks >> 6;
j->bb_current = 0;

Looks like it may make a guess until it knows. which makes me think as well as going up, it could go back down in the future as blocks fail. Which is slightly more worrying.

Did you find a solution?

Thanks,

Phil

from dhara.

Hey Phil,

Not a solution but as a patch you could just assign a fixed sector count and relay on that. I know it's not the best, but if you know exactly your use case and do sporadic writes it's not a big issue

from dhara.