> > On Wed, 2013-10-16 at 14:15 +0800, Shuduo Sang wrote: > >> I hope at least 3.8 stable tree would like to backport this patch > >> since the patch be merged into v3.9-rc5 and v3.8 is used widely. Loop > >> kernel-team@xxxxxxxxxxxxxxxx. > On Thu, Oct 17, 2013 at 11:58 PM, Kamal Mostafa <kamal@xxxxxxxxxxxxx> > wrote: > > > > but that commit says: > > > > This fixes problems caused by the following commit: > > commit d6dd9eb1d96d2b7345fe4664066c2b7ed86da898 > > Author: Daniel Vetter <daniel.vetter@xxxxxxxx> > > Date: Tue Jan 29 16:35:20 2013 -0200 > > drm/i915: dynamic Haswell display power well support > > > > But 3.8-stable doesn't include d6dd9eb. Is 2124b72 actually necessary > > in 3.8? (Or at least, would it actually be useful?) On Mon, 2013-10-21 at 20:05 +0800, Shuduo Sang wrote: > I think so. Thanks. After looking at it more closely, I'm pretty sure that this commit (2124b72) just isn't relevant for 3.8-stable ... In 3.8, the i915 power wells simply all get enabled by intel_init_power_wells() and nothing ever tries to disable them. This commit commit tries to modify the routine intel_set_power_well() which takes a boolean 'enable' param to enable or disable a power well -- but that routine doesn't exist in 3.8 (it was added to i915 later). So, if I understand the commit description and the code correctly, 3.8 doesn't doesn't need and couldn't make use of a "disable_power_well" override boot parameter. If you still think it would be useful for 3.8-stable, please explain why and supply a backport of the patch. Thanks, -Kamal
Attachment:
signature.asc
Description: This is a digitally signed message part