fix-mmc-busy-loop-on-bytes.patch and Fix hangup in do_pio_write() patch

[Werner Almesberger]

Andy Green wrote:
> At a guess, they may be trying to fix the same bug different ways,

Yes, I can confirm this.

> but
> if it's true, then the second patch should also do the same to
> do_pio_read loop?  So maybe just the first patch will solve it.

Yes, I think also do_pio_read should be fixed.

The two sides are slightly different: in the do_pio_write case,
you loop forever, so the problem is instantly visible.

In the do_pio_read case you loop until the word is complete. This
should be okay in most cases, but I could imagine error conditions
where this is no longer true and you'd end up with a system that
dies from time to time.

- Werner