Hi Dave,
This one plus the pci_disable_device and related error path cleanups are
part of the series to move the agp setup into drivers.
I've implemented your suggestion to work around the midlayer, so drm/i915
doesn't block for this at all. But I still think this is the right way to
solve this.
There's the neat little problem that some systems die if vga decoding
isn't decoded anywhere, because linux disabled that pci device.
Atm we solve that problem by simple not calling pci_disable_device at
driver unregister time for drm pci devices. Which isn't to great
because it leaks a pci_enable
On Mon, Jun 11, 2012 at 11:42:35AM +0300, Jani Nikula wrote:
> On Fri, 08 Jun 2012, Daniel Vetter wrote:
> > There's the neat little problem that some systems die if vga decoding
> > isn't decoded anywhere, because linux disabled that pci device.
> >
> > Atm we solve that problem by simple not cal
On Fri, 08 Jun 2012, Daniel Vetter wrote:
> There's the neat little problem that some systems die if vga decoding
> isn't decoded anywhere, because linux disabled that pci device.
>
> Atm we solve that problem by simple not calling pci_disable_device at
> driver unregister time for drm pci devices
There's the neat little problem that some systems die if vga decoding
isn't decoded anywhere, because linux disabled that pci device.
Atm we solve that problem by simple not calling pci_disable_device at
driver unregister time for drm pci devices. Which isn't to great
because it leaks a pci_enable
