On Wed, 2014-01-08 at 11:12 +0200, Jani Nikula wrote: > On Tue, 07 Jan 2014, "Lu, Ran" <hephooey@xxxxxxxxx> wrote: > > Hi Jani, > > > > As a matter of fact, I tried kernel from 3.9 to 3.13-rc6, all of them cannot > > read the pci information from 0:2.0, and lspci do not have Iris Pro as a video > > controller. I put in the attachment dmesg output from 3.12.6, I added some > > printk to show the results from pci_bus_read_dev_vendor_id, otherwise it is a > > vanilla kernel. I do not think I get any extra information by adding > > drm.debug=0xe because the pci device is never registered properly, but it was > > there anyway.The weird thing is grub can read pci device fine, I took a picture > > since I do not know how to save outputs in grub console. I did some test with > > another laptop with a working HD 4600. It seems even if I use setpci -s 0:2.0 > > 4.b=0 to disable the device, it is still responsive to further setpci and I > > can bring it back by setting 4.b=7. Now it does not look like the video card > > is disable. I guess maybe something wrong in the ACPI table triggered the > > kernel to read the wrong place. But still strange it only missed that > > particular bus. In case you are interested, I also put the dsdt table in the > > attachment, I can provide other tables if they are important. > > I don't have much clues here. Are there any bios settings you could > tweak? Did you try without the nvidia driver loaded? > > BR, > Jani. > > Might it be unable to allocate enough PCI address or it gets allocated above 4G? There are patches queued up for pci-next (plus a few in my tree for i915; I need to get back to these) to get PCI address space above 4G working. Probably unrelated... _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
