On Sun, 2007-12-09 at 15:56 -0500, Tom Sylla wrote:
Steve Isaacs wrote:
chip southbridge/broadcom/bcm21000 device pci 0.0 on end device pci 1.0 off end device pci 2.0 on end device pci 3.0 off end device pci 4.0 off end endYour HT2100 part looks ok. Here is what I have used successfully on a 2100 board:
chip southbridge/broadcom/bcm21000 # HT2100 device pci 0.0 on end # PCIe 0 device pci 1.0 on end # PCIe 1 device pci 2.0 on end # PCIe 2 device pci 3.0 on end # PCIe 3 device pci 4.0 on end # PCIe 4 end
My board has all 5 PCIe ports in use, so they are all enabled. From your "offs", it looks like you must be using the x16/x8 mux option, but that seems to conflict with below:
Yes, it's x16/x8.
Where are the HT2100 PCIe controllers? I would expect to see 1166:0140 or 142 or 144.
I would suggest that you turn them all "on" as an experiment. The Config.lb tree description has always been confusing, with documentation that doesn't talk much about new chipsets. The only way I ever got anything to work was by trial and error. (and my LB complains about leftover statics on my board if I turn any of the HT2100 ports "off")
I'm not sure how much of the lspci output can be believed since it was produced using a Phoenix BIOS on the board (since I don't have a LinuxBIOS config that will boot yet). Also, I don't believe I can perform the experiment you request since I don't have control of the BIOS that was used to capture the lspci output.
Steve