linux-kernel - Re: [PATCH] apple-gmux: Restore switch registers on suspend/resume

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20120731170732.GC19979@thinkpad-t410>
Date:	Tue, 31 Jul 2012 12:07:32 -0500
From:	Seth Forshee <seth.forshee@...onical.com>
To:	Matthew Garrett <mjg59@...f.ucam.org>,
	David Woodhouse <dwmw2@...radead.org>,
	Andreas Heider <andreas@...tr.de>,
	Arun Raghavan <arun.raghavan@...labora.co.uk>,
	linux-kernel@...r.kernel.org, airlied@...hat.com
Subject: Re: [PATCH] apple-gmux: Restore switch registers on suspend/resume

On Tue, Jul 31, 2012 at 10:18:56AM -0500, Seth Forshee wrote:
> On Sun, Jul 29, 2012 at 09:59:00PM +0100, Matthew Garrett wrote:
> > On Sun, Jul 29, 2012 at 09:52:51PM +0100, David Woodhouse wrote:
> > > On Sun, 2012-07-29 at 20:39 +0100, Matthew Garrett wrote:
> > > > And it looks like intel_lvds->edid is only set during intel_lvds_init(). 
> > > > That seems less than ideal. How about something like this entirely 
> > > > untested patch? 
> > > 
> > > Actually, it works if I write 'MIGD' first and then 'IGD'. Looks like we
> > > aren't switching the DDC over to the new gfx card soon enough in the
> > > process?
> > 
> > Yes, the call to the switcheroo code only comes after the card is 
> > powered on. Cc:ing Dave Airlie - what are the expectations here? It 
> > looks like i915 should have a reprobe function.
> 
> I dove into this yesterday and ended up with something similar to what
> you had. I'm rescanning for the EDID from the reprobe callback, and I
> also made it so that intel_lvds_init() will still keep the LVDS
> connector even if the panel isn't attached at boot and
> intel_lvds_detect() returns connected/disconnected based on whether or
> not we found an EDID for the LVDS.
> 
> All of this is working to the extent that I can boot with the Radeon
> card active, switch over to the Intel card, and get the EDID for the
> internal panel and an external monitor (although oddly on an HDMI
> connector, no on the DP like I expected). Both screens are remaining
> blank though. However I'm also getting blank screens if I mux over to
> the Intel GPU from grub before loading the kernel, which used to work
> for the LVDS panel at least.
> 
> I've copied the whole series of patches I'm running on top of 3.5 to
> http://people.canonical.com/~sforshee/apple-gmux-patches if anyone else
> wants to give them a try. The i915 changes are in patches 13-17. I've
> also pushed the i915 patches to my gmux-switcheroo branch.

Aha! I forgot to add i915.lvds_use_ssc=0 to the kernel parameters. So
with that and the above patches I can boot to the DGPU and successfully
switch between it and the integrated GPU. My VT console isn't showing up
with the IGPU, but the desktop does shows up.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/