On Fri, Jan 29, 2010 at 03:14:20PM -0800, David Hendricks wrote:
The patch to board_enable.c needs to be updated to apply cleanly with the new DMI identification stuff, but otherwise this patch applied cleanly for me and I was able to add a superio entry to do some very basic testing -- My superio isn't fully supported (yet), but I was able to run through the probing steps successfully.
Yeah, but as you can see, it is already older code, as it is from december :) Aaeons in a project like flashrom :)
There are three similar-but-different tables used in this file -- One in superio_chip_name(), one in superio_probe(), and then the larger superios[] table. Maybe we can find a way to use a single table to satisfy the needs of all the helper functions?
I want to get rid of the names table, and incorporate that directly in a wider and unrulier main table. The table for probing needs to stay. It contains the probe routines, and the probe addresses, and the order in which the probes happen.
I do not want the mess where one big table gets all probe routines called (even though 45% + 45% of it are the same two routines), like what you see with the flash chips.
I need a small table here, when an entry returns a result, the result, along with the pointer to the probe routine, will get matched in the main table.
Currently, with the superios in there atm, we run 4 probes:
WB@0x2E, WB@0x4E, ITE@0x2E, ITE@0x4E
Glad that someone is looking into this though :)
Luc Verhaegen.