On Mon, May 06, 2024 at 09:42:18PM +0200, Lukas Wunner wrote: > On Mon, May 06, 2024 at 08:43:43PM +0200, Johannes Stezenbach wrote: > > so, git bisect pointed to: > > > > commit 701d2054fa317188cd4039c84e72c73254013b23 > > Author: Thomas Zimmermann <tzimmermann@xxxxxxx> > > Date: Tue Jun 13 13:07:13 2023 +0200 > > > > fbdev: Make support for userspace interfaces configurable > > > > Add Kconfig option CONFIG_FB_DEVICE and make the virtual fbdev > > ... > > > > I think nothing on my system uses /dev/fb*, thus I had disabled > > CONFIG_FB_DEVICE. Everything worked, only suspend/resume broke. > > I can confirm v6.8.9 with CONFIG_FB_DEVICE=y works fine. > > A user-visible regression like this is always sufficient to ask for a > revert of the offending commit. It's the commit author's duty to find > a fix ASAP. If you could file a bug report that would be helpful but > you've already done more than could be asked from any user. I don't think it is a regression. A new config option was added with default=y and I chose to disable it. The results were unexpected since the Kconfig help text didn't say "can break suspend/resume with i915" ;-/ Johannes
