On Mon, 2010-07-12 at 22:14 +0200, Rudolf Marek wrote:
What about the reverse case? What if you put in the PT chip and tell flashrom that it's a PU? flashrom -VE -c A25L40PU In this case, it should fail when it tries to erase the /upper/ region of the memory, because flashrom is expecting to erase a 64 K chunk, but only getting 4 K since it is really a top boot block chip.
I think i did that too.
Well, okay, it's settled... I'll change my patch back to indicate the same RDID values for these two chips.
If that's really what happens... then I am convinced once and for all that the PT and PU really are distinct chips!
Yes they are ;) the markings tells so.
Haha, indeed :) But what a dumb design... the two chips have the same RDID but different erase layout. And for no good reason, since the company that makes them has an established pattern of distinguishing erase layouts based on a bit in the RDID code.
I don't understand why it hangs... I guess this is an issue with your chipset?
Yes it might be a flaw in VT8237S. Those non-uniform flashes are not used in bioses anyway, so I would not make it a big deal.
Thanks for all the testing.
Yeah, I'm definitely making a mountain of a molehill. I'll bet there is someone somewhere, though, who will mix these chips up in an embedded system, and accidentally trash the boot block... :-(
Dan