Hi Rudolf, hi Chunhao, > You see one eeprom dissapeared, your device appeared but 0x55 remained for > some reason. But in fact nothing left there. Without information from ASUS > I cant tell you what happened there. Your assumption that the EEPROM at 0x55 might be wrong. The device at 0x55 might be a different one after the SMBus as been switched. In fact, I think it is. >From the early days, we assumed that Asus was merely trying to hide their hardware monitoring chip. This is probably true, but nevertheless what they are doing is SMBus switching. They might as well be doing that for other reasons, such as the need to have different chips at the same I2C address. Think of the Tyan S4882 example. So it sounds like a more proper approach of the problem would be to hack the SMBus driver in the same way I did for i2c-amd756, and add proper bus multiplexing. (And yes, we would need to integrate the multiplexing deeper in i2c-core in Linux 2.6, but I have not yet taken the time to look at how this can be done cleanly.) > I expected this to show up. Again I cant tell what happened in bus because > I dont know what the multiplexer really did with the bus. Chunhao, could you please ask Asus for a complete diagram or description of the SMBus on this board? We need to know which devices are on each branch of the multiplexer. Looks like memory module EEPROMs are on branch 1, and the W83792D is on branch 2, but there seems to be other EEPROMs on branch 2 as well. If we could have a picture or detailed explanation of the SMBus use on this board, we could modify the kernel code accordingly. Thanks, -- Jean Delvare
