On Fri, 2009-06-26 at 13:50 -0600, Matthew Wilcox wrote: > On Fri, Jun 26, 2009 at 12:04:12PM -0700, Keith Mannthey wrote: > > Hello All, > > I am working with Intel X5500 series processors. The documentation > > for these processors can be found here: > > http://www.intel.com/p/en_US/products/server/processor/xeon5000/technical-documents > > (for anyone else who's curious about this, you only need the smaller > 'volume 2', not the larger 'volume 1'). > > > Theses devices are "hidden" from Linux at is stands now, the kernel does > > not detect them. All standard methods of using the device fail, any > > pci_XXX calls works as the devices are not in the list of devices in the > > kernel. For example lspci (which uses sysfs) does not show the > > devices. > > > There are 2 real questions I have: > > > > 1. Why are these devices hidden? I understand some of the other icore7 > > cpus devices behave as I would expect. > > You'd have to ask your BIOS people. Linux doesn't blindly scan every > single PCI address; it would take far too long. Instead, we rely on > the BIOS to tell us about PCI busses that exist, and we scan those. Which table do the devices need to be present in for the kernel to pick them up? Why does the alternate configuration type work? > > 2. If the devices can not be made to be used as a normal pci device > > what is the best way to access them? > > I doubt you can access them other than through PCI config space. > > I suspect we could add functionality for Linux to notice that we're > running on a Nehalem machine, and then to scan these PCI devices. > But it'd be better to find out from your BIOS people why they don't > report these devices. I will start asking them. Thanks, Keith Mannthey LTC Real-Time -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
