Created attachment 126024 [details] Xorg.0.log Motherboard is an Asus H110M-K. OS is a Debian stretch amd64, with vanilla kernel 4.7.2. It happened the same with Debian jessie and xserver-xorg-video-intel backported. The machine boots in text mode. This is expected, is how I want it to be. If I want graphical mode, I log on and then "startx". I have several machines connected through a KVM switch (PS2 Mouse + PS2 Keyboard+ VGA). This is one of them. If the focus of KVM switch is on this machine, everything works fine. After the boot sequence, I can move back and forth and I always find this machine working. But if the focus is on another machine *during the boot*, this machine stops sending signal to monitor. If I then come back, I find the monitor is receiving no signal. The machine is not stuck, it's working: if I press CTRL-ALT-CANC it begins a normal shutdown, and if I try to logon by telnet or ssh there is no issue. Only the monitor is off.
Created attachment 126025 [details] dmesg when it works
Created attachment 126026 [details] dmesg when it doesn't work
Created attachment 126027 [details] config kernel 4.7.2
Created attachment 126071 [details] dmesg when it works - drm intel I've tried with drm-intel kernel:
Created attachment 126072 [details] dmesg when it doesn't work - drm-intel And here dmesg when the monitor remains off. One thing I can notice is that this dmesg contains these lines [ 4.166745] [drm:intel_power_well_disable] disabling power well 2 [ 4.166751] [drm:intel_power_well_disable] disabling DC off [ 4.166757] [drm:intel_power_well_disable] disabling always-on and they are not on the other.
Valerio - I'm really sorry about this delay for getting back to you. I'm just wondering here that can kernel do much on a case that one does not get "I'm alive" kind of the signal from KVM switch. Have you tried to "wake up" channel with xrandr? This might cause monitor to wake up (previously showing as disconnected). Assumable you are having debug version on your machine, so you should also be able to produce i915_display_info data (sudo cat /sys/kernel/debug/dri/0/i915_display_info) on both fail and okay cases.
I don't have that PC anymore. If I'll have one with the same chipset / video card, I'll test and report again.
Assuming this to be resolved+fixed. If problem still persist with the latest kernels (preferable drm-tip from git://anongit.freedesktop.org/git/drm-tip) then please change status to REOPENED with the proper data included (see https://01.org/linuxgraphics/documentation/how-report-bugs). Side note: In addition to the diffs mentioned in comment 5, I also noticed that connector DP-1/38 does not get connected on non-working case... ON WORKING USE CASE [drm:drm_setup_crtcs] [drm:drm_enable_connectors] connector 33 enabled? no --> [drm:drm_enable_connectors] connector 38 enabled? yes [drm:intel_fb_initial_config] connector HDMI-A-1 not enabled, skipping [drm:intel_fb_initial_config] looking for cmdline mode on connector DP-1 [drm:intel_fb_initial_config] looking for preferred mode on connector DP-1 0 [drm:intel_fb_initial_config] connector DP-1 on [CRTC:26:pipe A]: 1280x1024 [drm:drm_setup_crtcs] desired mode 1280x1024 set on crtc 26 (0,0) ON NON-WORKING USE CASE [drm:drm_setup_crtcs] [drm:drm_enable_connectors] connector 33 enabled? no --> [drm:drm_enable_connectors] connector 38 enabled? no [drm:intel_fb_initial_config] connector HDMI-A-1 not enabled, skipping [drm:intel_fb_initial_config] connector DP-1 not enabled, skipping [drm:intel_fb_initial_config] Not using firmware configuration [drm:drm_setup_crtcs] picking CRTCs for 8192x8192 config Causing 8K resolution to be used on NON-WORKING CASE (?).
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.