On Sun, Mar 25, 2007 at 08:09:22PM +0200, Stefan Reinauer wrote:
maybe chipset_enable?
sounds good.
Done.
The island aruma got renamed to agami aruma at some point, the board specific match has been adjusted to that now. I hope that the pci ids i gathered from its src/mainboards/ directory are ok. I of course have no way of testing this.
The rest is pretty much last patch. Except that board enable functions only get passed the name now. The board enable function needs to claim its own pci device, this is safer as the ordering of both found pci devices (and the need for not having to match this and thus handing the callbacks a NULL) was going to get us into trouble.
The difference for chipset enable and flash_enable is the much cleaner chipset_flash_enable routine. iopl and preparing of the pci_access has been relegated to flash_rom.c. Then there's only some extraneous whitespace removal and replacing // with /* */. I'm not sure how svn handles moving of files, but that's usually a good point to do such function-less changes.
Has been tested on all the same machines again. Needs a quick test on the agami.
Luc Verhaegen.