On 13.05.2008 15:14, Peter Stuge wrote:
On Tue, May 13, 2008 at 01:56:38PM +0200, Uwe Hermann wrote:
- {"ST", "M25P40@ICH9", ST_ID, ST_M25P40, 512, 64 * 1024, TEST_UNTESTED, probe_ichspi_stm25_sig, erase_ichspi, write_ichspi, read_ichspi},
- {"ST", "M25P80@ICH9", ST_ID, ST_M25P80, 1024, 64 * 1024, TEST_UNTESTED, probe_ichspi_stm25_sig, erase_ichspi, write_ichspi, read_ichspi},
- {"ST", "M25P16@ICH9", ST_ID, ST_M25P16, 2048, 64 * 1024, TEST_UNTESTED, probe_ichspi_stm25, erase_ichspi, write_ichspi, read_ichspi},
- {"ST", "M25P32@ICH9", ST_ID, ST_M25P32, 4096, 64 * 1024, TEST_UNTESTED, probe_ichspi_stm25, erase_ichspi, write_ichspi, read_ichspi},
- {"ST", "M25P64@ICH9", ST_ID, ST_M25P64, 8192, 64 * 1024, TEST_UNTESTED, probe_ichspi_stm25, erase_ichspi, write_ichspi, read_ichspi},
Why "@ICH9", does the chip behave differently depending on where you plug it in?
How to program an SPI flash chip depends on both the chip type, and the SPI master type. (Ie which superio/southbridge it is connected to.)
Currently, flashrom does not have infrastructure that can deal with a non-transparent flash memory bus master, so the support for these flash chips that Claus is proposing will only work on ICH9 systems.
This sucks of course, but it is a flaw of flashrom and I do think it is OK to include Claus' changes because they do work for him.
It is too much to expect a brand new contributor to rewrite the SPI infrastructure in flashrom. That will come, I know Carl-Daniel has ideas.
Once the RES support patch is merged (I'm waiting for Fredrik Tolf to signoff it and Peter Stuge to ack it) and we have a deal about the "flags" field in struct flashchip, it will be easy to integrate ICH9 support into flashrom.
Regards, Carl-Daniel