Prevent infinite loop in lfs_traverse

[TheLoneWolfling]

lfs_traverse has no protection against infinite loops. Consider what happens if a directory entry has dir.d.tail == cwd... you end up hanging forever. Relevant section of code:

int lfs_traverse(lfs_t *lfs, int (*cb)(void*, lfs_block_t), void *data) {
    ...
    lfs_dir_t dir;
    lfs_block_t cwd[2]
    ...
    while (true) {
        ...
        lfs_dir_fetch(lfs, &dir, cwd);
        ...
        cwd[0] = dir.d.tail[0];
        cwd[1] = dir.d.tail[1];
        ...
    }
}

This is mostly a theoretical issue - this can only really happen if the on-flash data is already in an invalid state. However, as littlefs is supposed to be failsafe...

Two fixes come to mind:

1. Keep a counter. If you ever traverse >= the total number of blocks, you know that you've hit a loop. (This may take a while... but this isn't supposed to happen unless something has gone wrong anyway, and it's still better than just hanging forever.)
2. Use something like the tortose-and-hare algorithm. (Advance the hare, do the callback, exit if done, panic if hare == tortose, advance the hare, do the callback, exit if done, panic if hare == tortose, advance the tortose, panic if hare == tortose. Repeat.) Probably overkill.

This change implements #1.

[TheLoneWolfling]

Note: there is an analogous issue in lfs_parent - although it's harder to hit, it can be hit from lfs_deorphan.

Ditto, there's an analogous issue with the open file loop in lfs_traverse.

[geky]

Hi @TheLoneWolfling, thanks for creating a PR. I'm going to wait on this until v2 just because a lot of the code is going to change. But you're right it's probably a good idea to have these checks on every loop.

Related #98

[TheLoneWolfling]

Ok. Is it alright if I start looking at v2? Or should I hold off?

[geky]

Sure thing! It may be rough around the edges, but any feedback is welcome :)

[geky]

Ok! Only a year and one major version late. I've added proper loop detection/recovery to #372.

I'll go ahead and close this once #372 is merged. But let me know if you see anything that needs fixing.

Thanks again for the contribution!