Created attachment 53173 [details] dmesg with cc03bbf1 Booting into cc03bbf1 results in a black screen once KMS starts, the system keeps running in the background but the screen remains unusuable.
Created attachment 53174 [details] dmesg with 3.1 mainline
Created attachment 53175 [details] intel_gpu_dump with cc03bbf1
Created attachment 53176 [details] intel_gpu_dump with 3.1 mainline
This regression appears to have made it into 3.2.0-rc1 as well. I'll provide my information as soon as I can gather them.
Created attachment 53321 [details] [michel] dmesg output with working KMS (3.1.0 + unrelated ath9k and btrfs patches)
Created attachment 53322 [details] [michel] dmesg output with black screen after KMS (3.2.0-rc1 + ath9k patch)
Created attachment 53323 [details] [michel] intel_reg_dump output with working KMS (3.1.0+)
Created attachment 53324 [details] [michel] intel_reg_dump output with black screen after KMS (3.2.0-rc1+)
So this is because lvds_use_ssc is somehow re-enabled; booting with i915.lvds_use_ssc=0 works around the issue for me. Keith added a framework for blacklisting models that don't support SSC (435793dfb8) and I added my laptop to this blacklist (070d329ae52e2) so it's weird that the issue is cropping up again. I'll try and see if the relevant code is still being used or not
The problem was introduced in commit 72bbe58cd9568c77 from September 26th: drm/i915: Allow SSC parameter to override VBT value In the function intel_panel_use_ssc, now if i915_panel_use_ssc >= 0 (and it's an unsigned int initialized to -1 so it's always greater than 0), then the check for whether the device is in the quirks table is short-circuited since the function immediately returns true at that point. Shouldn't the quirk table always overrides whatever the default SSC setting is (unless overridden by the user)? In which case perhaps either check for quirks first and short-circuit, or also check that there are no quirks even if i915_panel_use_ssc is greater than 0.
The user overrides the quirk which overrides the probe which overrides the default. commit 4415e63b13c68c2f56d16d400a1ae345f68cf655 Author: Keith Packard <keithp@keithp.com> Date: Wed Nov 9 09:57:50 2011 -0800 drm/i915: Module parameters using '-1' as default must be signed type Testing i915_panel_use_ssc for the default value was broken, so the driver would never autodetect the correct value. Signed-off-by: Keith Packard <keithp@keithp.com> Reviewed-by: Michel Alexandre Salim <salimma@fedoraproject.org> Tested-by: Michel Alexandre Salim <salimma@fedoraproject.org> Cc: stable@kernel.org
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.