[Intel-gfx] [PATCH v2] drm/i915: enable memory self refresh on 945

Eric Anholt eric at anholt.net
Fri Jan 15 23:55:54 CET 2010


On Wed, 6 Jan 2010 12:40:00 -0800, Jesse Barnes <jbarnes at virtuousgeek.org> wrote:
> On Tue, 05 Jan 2010 10:10:38 -0800
> Eric Anholt <eric at anholt.net> wrote:
> 
> > On Thu, 10 Dec 2009 11:52:01 -0800, Jesse Barnes
> > <jbarnes at virtuousgeek.org> wrote:
> > > On Wed, 09 Dec 2009 13:25:39 +0800
> > > Li Peng <peng.li at linux.intel.com> wrote:
> > > 
> > > > I did some test and found that memory self refresh on 945
> > > > isn't really enabled in commit 652c39. On 945, we need to use
> > > > bit 31 of FW_BLC_SELF to enable the write to self refresh bit
> > > > and bit 16 to enable the write of self refresh watermark, and
> > > > memory self refresh will take effect when CPU enters C3+ state.
> > > > 
> > > > This patch enables memory self refresh on 945 when graphics
> > > > is idle, and disable self refresh when it is busy. In my test
> > > > on a netbook of 945GSE chipset, it saves about 0.8W idle power.
> > > > 
> > > > Signed-off-by: Li Peng <peng.li at intel.com>
> > > > ---
> > > >  drivers/gpu/drm/i915/i915_reg.h      |    4 +++-
> > > >  drivers/gpu/drm/i915/intel_display.c |   23
> > > > ++++++++++++++++++++++- 2 files changed, 25 insertions(+), 2
> > > > deletions(-)
> > > 
> > > I'd like to figure out why I don't see hangs on my AspireOne without
> > > the idle/busy change you have here, since SR should be totally
> > > automatic afaik. Also, we may as well include the i915 path as well
> > > (after testing some more of course), since it affects this same code
> > > path.
> > 
> > Where are we at on this one?
> 
> Peng, have you done any more testing?  What's your memory configuration?
> 
> I'm ok with Peng's patch (though it probably needs a refresh) since it
> should have effectively the same amount of power savings as leaving SR
> enabled at all times.  And I'm pretty sure there are some hardware bugs
> related to SR; maybe I'm just not hitting them.

So, this conflicts with the render clock removal, and will need to be
fixed up for that.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20100115/9ab52b59/attachment.sig>


More information about the Intel-gfx mailing list