Lenovo X1 Skylake GT2 520 rev 7, i915 driver; openSUSE Tumbleweed 190730, plasma 5.16.3, Qt 5.13.0, kernel-default 5.1.16-1 Booting into sddm fails following update from 5.1.16 to 5.2.1 (distribution provided kernel). I have tried up to 5.2.4 and the problem persists. It is impossible to boot without an attached external screen (via DP). After booting, interacting through the external screen, the system behaves normal (but with an unusable internal screen - it is frozen). Reverting back to 5.1.6 fixed the issue. With kernel 5.2.3 booting reaches all the way to sddm but screen is frozen and unresponsive. best, /mikael
Quick test is to add i915.enable_psr=0 to the kernel commandline (grub)
Indeed, i915.enable_psr=0 cured the problem. Thanks, /mikael
José any suggestion on which patches to test, or is everything queued up for stable by this point?
Hi Mikael If you want you can compile your own kernel with this fix: https://patchwork.freedesktop.org/series/63774/#rev4 Otherwise keep i915.enable_psr=0 as Chris said until that patch gets to kernel stable, the last updated that I got is that it would be included on Linux 5.2.6 to be released in this coming week.
Thanks, I'll wait for 5.2.6. cheers, /mikael
I tried the 5.2.6 kernel from the openSUSE stable repository (kernel-default-5.2.6-1.1.g187170c) and changed to i915.enable_psr=2. Problem still there. Went back to 5.2.3 and changed back to i915.enable_psr=0 for a working system.
Hi Mikael The fix was not included on 5.2.6, hopefully it will on 5.2.7 as Greg asked the backported version.
Apparently 5.2.7 doesn't seem solving the problem, according to openSUSE bugzilla entry: https://bugzilla.opensuse.org/show_bug.cgi?id=1143139#c30 > Martin Wilck 2019-08-07 13:32:47 UTC > > I have the same problem as in bug 1143755. It is *not* solved in 5.2.7-1.ga3e04b1 (the workaround "i915.enable_psr=0" helps), but it is solved in 5.3.rc3-1.1.g571863b. I have Skylake hardware: > > > $ cat /sys/kernel/debug/dri/0/i915_capabilities > > gen: 9 > > platform: SKYLAKE > > pch: 4 > > is_mobile: no > > ... > > has_psr: yes > > I've tried to figure out the relevant commit but I haven't yet. Note that 5.3-rc3 does not include the patch which disables PSR2 ( https://patchwork.freedesktop.org/patch/319173/?series=63774&rev=4). > > > $ git log --oneline --grep PSR -i v5.2..v5.3-rc3 -- drivers/gpu/drm/i915/ > > 6d61f71 drm/i915/vbt: Fix VBT parsing for the PSR section > > 417f254 Merge tag 'drm-intel-next-2019-06-19' of git://anongit.freedesktop.org/drm/drm-intel into drm-next > > 381f8a2 drm/i915/psr: Force manual PSR exit in older gens > > 4d432f95 drm: Rename struct edp_vsc_psr to struct dp_sdp > > 8e9d645 drm/i915/dp: Add a config function for YCBCR420 outputs > > 6f08ebe drm/i915: Remove the unneeded AUX power ref from intel_dp_hpd_pulse() > > I suspected 381f8a2 but that makes little sense as that affects gen < 9 only, and I have gen = 9. So the only one that remains would be 6d61f71, no idea if that makes sense.
We sent the backport of "drm/i915/vbt: Fix VBT parsing for the PSR section" to Greg but it did not made to 5.2.7.
I backported the patch into openSUSE kernel, and it was confirmed to work. Thanks.
Fix merged on Linux 5.2.8
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.