http://bugzilla.opensuse.org/show_bug.cgi?id=973129 http://bugzilla.opensuse.org/show_bug.cgi?id=973129#c1 Egbert Eich changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |UPSTREAM --- Comment #1 from Egbert Eich --- Please report your issue upstream to Intel directly: http://bugs.freedesktop.org. If you provide the upstream ticket URL here we can monitor it and help you. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=973129 http://bugzilla.opensuse.org/show_bug.cgi?id=973129#c3 Egbert Eich changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |calinuswork@gmail.com Resolution|FIXED |UPSTREAM Flags| |needinfo?(calinuswork@gmail | |.com) --- Comment #3 from Egbert Eich --- Let's leave this marked 'Upstream' for a bit. I've looked at fdo#94567 you've posted: I don't think this is related. Your system seems to be an Optimus system and it is not unlikely that your external HDMI port is connected to NVIDIA chip. I may not be able to help much with Optimus as I never had one to play with, but I may be able to help you to find the information you'd need for an upstream bug report. Please boot the system in multi-user mode (not graphical mode) log into a console as root: 1. Run $ lspci -nn > /tmp/lspci.log 2. Start the Xserver (with the external display connected): $ /usr/bin/Xorg -retro & sleep 1; xterm -display :0 This will give you an Xserver with just 'xterm' running. The advantage is that this system may not hang so quickly. 3. In the xterm, please run: $xrandr -q --verbose > /tmp/xrandr.log Please provide the files /tmp/lspci.log, /tmp/xrandr.log and /var/log/Xorg.0.log. You can kill the Xserver by pressing Ctrl-Alt-Del twice. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=973129 http://bugzilla.opensuse.org/show_bug.cgi?id=973129#c5 --- Comment #5 from Egbert Eich --- (In reply to Calin Calin from comment #4)

I will try this and get back to you, thank you.

For reference this was working fine before the 4.5 update.

Ok, TW should keep your previous kernel on an update. Can you boot the old kernel and see if it still works? If it does - please provide the version of this kernel. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=973129 http://bugzilla.opensuse.org/show_bug.cgi?id=973129#c7 --- Comment #7 from Egbert Eich --- (In reply to Calin Calin from comment #6)

It works with kernel 4.4.3-1-default (not great, I can't change the resolution but it does net freeze)

It does not work with 4.5.0-2 and 4.5.0-1

Since an older version worked (better) for you, you could bisect the issue using the upstream kernel repo to find the change which broke things for you.

I am not sure how to boot with no X (sorry)

What do you mean? You don't know how to reboot the machine from the console? Try: ctrl-alt-del or enter 'reboot' on the command line after logging in as 'root'. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=973129 http://bugzilla.opensuse.org/show_bug.cgi?id=973129#c9 --- Comment #9 from Egbert Eich --- (In reply to Calin Calin from comment #8)

What I meant is I don't know how to boot the system in non graphical mode.

Ok, understood. There are two ways to do this: 1. If you want it mermanent, ie until you change it again: a. log in as root. b. run: # systemctl set-default multi-user.target or to set it back to graphical: # systemctl set-default graphical.target 1. To do it temporarily: a In grub, press 'e' to get to the editable menu screen b. Add a '3' somewhere to the kernel command line.

I found this tutorial online related to git-bisect https://wiki.gentoo.org/wiki/Kernel_git-bisect do you have a better resource?

Partly. 1. You can do this from any directory on a file system with a lot of space. If there is a lot of space left in the fs with your home directory for example, do something like: $ mkdir $HOME/kernel-source $ cd kernel-source $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git (don't use kernel-stable.git) $ cd linux 2. I would recommend to build a 'vanilla' 4.5 kernel first to see if this also exhibits the problem: in the directory 'linux' do: $ git checkout v4.5 v4.5 is a version tag. To list all available version tags, do 'git tag' 3. If the v4.5 kernel also has the problem, you should do the same for the 4.4 kernel to see if this one works 'better': $ git checkout v4.4 4. After these steps, (ie if you veriried that v4.4 is ok while v4.5 is broken) you can start bisecting. I the tree, do $ git bisect start $ git bisect good v4.4 $ git bisect bad v4.5 Build the kernel. Building: 0. If you don't have a .config from a previous build, create a .config file: $ zcat /proc/config.gz .config 1. To build the kernel & modules, do $ make olddefconfig $ make bzImage modules modules_install 2. Once the build is done, as root do: # make install modules-install # mkinitrd 3. Reboot, select the kernel you've just built in the boot menu and test. 4. After testing, continue bisecting: $ cd $HOME/kernel-source/linux If problem didn't occur: $ git bisect good If you saw the problem: $ git bisect bad Go to step 1. above and redo. 5. You may want to get rid of old test kernels: The 'quick-n-dirty' way I use: # rm /boot/*{VERSION}* # rm -rf /lib/modules/*{VERSION} What I often do is the delete the current test-kernel before I call 'mkinitrd' in step 2 above. ${VERSION} can be obtained by calling 'uname -r': # VERSION=$(uname -r} There are a few things I do recommend to speed up building: 1. Do you have multiple CPU cores and/or hyper-threading? Build with multiple threads: add -j <N> to the make command line: If n is the number of cores: - with hyper-threading: <N> = (n + 1) * 2 - without hyper-threading <N> = (n + 1) 2. Are you building on the machine you are using to run the kernel? Limit the modules you build to the ones you really need: In the kernel sources, do $ zcat /proc/config.gz > .config $ make olddefconfig $ make localmodconfig The .config file generated should be saved for later builds (just in case): $ cp .config /my_kernel_config 3. Make sure /tmp is mounted to tmpfs. do: $ mount | grep tmp If the output doesn't contain a line like: tpmfs on /tmp type tmpfs (rw,relatime) Add the line: tmpfs /tmp tmpfs defaults 0 0 to your /etc/fstab and reboot. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=973129 http://bugzilla.opensuse.org/show_bug.cgi?id=973129#c11 --- Comment #11 from Calin Calin --- I did not had the time to look into this yet, I only have a second monitor in the office to test it and when I get in the office I always have other stuff to do. The latest updates kernel 4.5.0-3 made things worst, now there is no tearing the system just locks when I plug the hdmi. -- You are receiving this mail because: You are on the CC list for the bug.