Hi Ville,

thanks for your response!

On 02/05/2023 15:13, Ville Syrjälä wrote:

On Sat, Apr 29, 2023 at 09:55:45PM +0200, Kurt Garloff wrote:

...

Hi Ville,

While 0fc6fea41c71 drm/i915: Disable DC states for all commits (cherry picked from commit 41b4c7fe72b6105a4b49395eea9aa40cef94288d) does look correct to me, it does break modesetting on the GPD Pocket 3, a i7-1195G7 laptop. I run the kernel with fbcon=rotate:1 video=DSI-1:panel_orientation=right_side_up \ mem_sleep_default=s2idle

Note that the mem_sleep_default parm has been dropped meanwhile. It used to be required as the laptop would show a black screen after resuming from S3 until 6.1.5 or so.

...

No special i915 parameters. Hardware is described here: https://wiki.archlinux.org/title/GPD_Pocket_3

I disected this patch which was merged (backported) for 6.1.23. I currently run 6.1.26 with it reverted.

Without reverting it, when fbcon is switched to show the splash screen (GPD logo with Ubuntu added in on working kernels), the screen remains black (backlight on, but nothing displayed) and nothing happens any more.

Please file a bug at https://gitlab.freedesktop.org/drm/intel/issues/new

Find a bug report at

https://gitlab.freedesktop.org/drm/intel/-/issues/8419

Boot both kernels (revert vs. no revert), passing 'drm.debug=0xe log_buf_len=4M' to the kernel cmdline, and attach the resulting dmesg from each to the bug.

Getting a boot log out of the crashing laptop required me to find a cable and get netconsole working. (If you know better ways, I'm open to learn.) Thus the delay ...

Also would be good if you try to reproduce on the latest drm-tip (git://anongit.freedesktop.org/drm-tip drm-tip) as well.

I'll test drm-tip as requested and see what happens. I'll attach the info to the bug report, let nme know if you want eMail in addition.

Best,