[flashrom] Erase-function fallback (was: Using libflashrom from fwupd)

[Nico Huber]

Hi,

starting a new thread for this because it's somewhat orthogonal.

> [...] The progress would not be
> constant in speed (e.g. because reading is usually much faster than
> writing), and it would sometimes even be non-monotonic (if a write
> fails, we re-read the whole chip to make absolutely sure we get
> everything correct thus increasing the total number of bytes to
> transfer...

That's an internal that I'd really try to avoid to build the progress
interface around. Actually I'd like to get rid of this fallback semantic
at least in the library version of flashrom_flash_write(). Some reasons
why:

* I've seen this fallback (not) working in the following cases:

  1. Unreliable connection – fallback made things even worse, usually
     erasing the whole chip.
  2. Wrong flash chip selected – is that a valid use case?
  3. Locked opcodes in the programmer – we could just ask the programmer
     beforehand if it's supposed to work.

* The complex procedure is very far from what someone would expect when
  he calls a write function.

Instead, the caller could set a hint in the flash context if a specific
erase function should be used (to work around issues that shouldn't
exist, ahem).

Thoughts?

Nico