Created attachment 401018 [details] full lspci, the 3d controller is my nvidia card. Hi, here a bug with long time, the problem is when we have a discrete card as 3d controller, and the kernel don't recognize it as vga device, its very common with for example notebooks with 2 video cards for example a intel with nvidia. Here the issue in bumblebee project: https://github.com/Bumblebee-Project/Bumblebee/issues/159 and here i found a patch, i test it and works fine at the moment: http://pastebin.com/wpmFi38k my super log of the error: abr 10 03:59:34 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 03:59:34 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 03:59:50 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 03:59:50 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:00:18 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:00:18 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:00:40 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:00:40 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:38:17 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:38:17 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:45:18 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:45:19 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:49:00 pipe-PC kernel: vgaarb: this pci device is not a vga device abr 10 04:49:00 pipe-PC kernel: vgaarb: this pci device is not a vga device continue with the same to much.... using gentoo-sources-3.19.3 Thx.
Can you attach the patch? I'm currently blocked from pastebin And emerge --info also, please.
Created attachment 401482 [details, diff] Patch
Created attachment 401484 [details] emerge --info
i have applied the patch, and my overlays are masked excepted with the packages i use, i don't use system packages of overlays.
Is this a warning or does it have a negative affect?
Changing vga arbitration on such systems may cause side effects that are not directly visible. Adding the 3D controller to vgaarb certainly is contra-productive. It would be better if the code that calls vgaarb for such a 3D controller did not do so (either nvidia closed driver or some userspace code switching GPUs). See also https://bugzilla.kernel.org/show_bug.cgi?id=86121 for a case where vgaarb caused havoc to backlight controls, triggered by nvidia closed driver.
I'm not comfortable with carrying this in gentoo-sources. Feel free to submit upstream if you like.
