SlackwareThis Forum is for the discussion of Slackware Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
I'm new to this whole skylake thing (my old motherboard died, so I did a cheap temporary upgrade to an MSI B150M VD-Pro motherboard with an Intel Pentium G4500 that has the Intel HD 530 GPU), but I'm having problems with the latest -current (using Eric's live distro, since my 14.1 install was too old and I wanted to test it).
Everything seemed to boot up ok, except once KMS took over, my right screen was noticeably brighter than the left (right screen is using HDMI via a DVI->HDMI adapter and the left is using a standard VGA plug). Once I got into KDE, again, everything seemed to be ok (except for the brighter right screen). I then went to bed, woke up this morning, and my KDE taskbar is frozen. I can't click on anything to bring programs up (although, I can still alt-tab to programs and use alt+F2 to start programs). It even froze the window preview over Chrome, which won't disappear, however, amusingly enough, the preview continues to update as I access various webpages (and indeed, even as I type this, I can see the faint text showing up in the preview window).
I've been reading this thread as it's progressed, but this seems like unusual behavior that hasn't been said before. I also got several dumps in my dmesg, one occurring on boot:
That second one was occurred at 23:47 and then was repeated again at 10:57. KDE's taskbar clock is frozen at 10:57. My Xorg.0.log doesn't show any dumps or issues with X.
I am currently using no custom xorg conf files. Glancing at the Xorg.0.log, it seems like it is running DRI2 and sna.
Code:
[ 75.105] (II) intel: Driver for Intel(R) Integrated Graphics Chipsets:
[ 75.105] (II) intel: Driver for Intel(R) HD Graphics: 2000-6000
[ 75.105] (II) intel: Driver for Intel(R) Iris(TM) Graphics: 5100, 6100
[ 75.105] (II) intel: Driver for Intel(R) Iris(TM) Pro Graphics: 5200, 6200, P6300
[ 75.110] (II) intel(0): Using Kernel Mode Setting driver: i915, version 1.6.0 20151010
[ 75.110] (--) intel(0): gen9 engineering sample
[ 75.110] (--) intel(0): CPU: x86-64, sse2, sse3, ssse3, sse4.1, sse4.2; using a maximum of 2 threads
[ 75.110] (II) intel(0): Creating default Display subsection in Screen section
[ 75.110] (==) intel(0): Depth 24, (--) framebuffer bpp 32
[ 75.110] (==) intel(0): RGB weight 888
[ 75.110] (==) intel(0): Default visual is TrueColor
[ 75.110] (II) intel(0): Output HDMI1 has no monitor section
[ 75.110] (II) intel(0): Enabled output HDMI1
[ 75.110] (II) intel(0): Output DP1 has no monitor section
[ 75.110] (II) intel(0): Enabled output DP1
[ 75.110] (II) intel(0): Output HDMI2 has no monitor section
[ 75.110] (II) intel(0): Enabled output HDMI2
[ 75.110] (--) intel(0): Using a maximum size of 256x256 for hardware cursors
[ 75.111] (II) intel(0): Output VIRTUAL1 has no monitor section
[ 75.111] (II) intel(0): Enabled output VIRTUAL1
[ 75.111] (--) intel(0): Output HDMI1 using initial mode 1920x1080 on pipe 0
[ 75.111] (--) intel(0): Output DP1 using initial mode 1920x1080 on pipe 1
[ 75.111] (==) intel(0): TearFree disabled
[ 75.111] (==) intel(0): DPI set to (96, 96)
[ 75.111] (II) intel(0): SNA initialized with Skylake (gen9) backend
[ 75.111] (==) intel(0): Backing store enabled
[ 75.111] (==) intel(0): Silken mouse enabled
[ 75.111] (II) intel(0): HW Cursor enabled
[ 75.111] (II) intel(0): RandR 1.2 enabled, ignore the following RandR disabled message.
[ 75.111] (==) intel(0): DPMS enabled
[ 75.111] (==) intel(0): Display hotplug detection enabled
[ 75.111] (II) intel(0): [DRI2] Setup complete
[ 75.111] (II) intel(0): [DRI2] DRI driver: i965
[ 75.111] (II) intel(0): [DRI2] VDPAU driver: va_gl
[ 75.111] (II) intel(0): direct rendering: DRI2 enabled
[ 75.111] (II) intel(0): hardware support for Present enabled
Personally, I'm not terribly concerned with this because I will likely put in a discrete AMD card once I'm done testing, but I wanted to do some additional testing with Skylake, since I now have the hardware and I know it's been a constant struggle with -current. Is my next process to force DRI3 or should I try uxa?
HD530 here. I have some glitches in LibreOffice Calc (the auto completion when typing) but it was the same with the previous 20160418_562ae1f. The glitches appeared after I upgraded AlienBOB's LibreOffice to the latest, so I did not report them here. I do not use resume. Steam games, browsers and everything else work fine.
I was able to borrow a laptop and was able to get the HD5500 working well enough with a USB hard drive with Slackware64-Current.
The Intel DDX driver 20160601_b617f80 works, but under sna acceleration I do get a few artifacts and glitches using LibreOffice from SBo, as well as a few tearing issues using some OpenGL using games and apps. Uxa is a lot slower but less glitches.
The default modesetting driver with glamoregl enabled has very few glitches at all. It's not as fast as the DDX sna driver, since it's using the egl stack, but it works very well. I did have to configure the ZaphodHeads to match the exact display ports on the laptop. The only glitch I got was a misrendered resized screen using Mupen64Plus.
After some more time with this, it seems I get that dump in dmesg whenever I wake my computer from the monitors going to sleep (I don't actually sleep the machine itself, since it is a server). KDE will pop up a dialog box stating the screen has changed and asking if I want to run a configuration tool. It seems that when the monitors go to sleep (or when they wake up), the second monitor, which is connected via a DVI->HDMI adapter, is lost and then readded (I'm assuming so, because any windows I had on the second screen will be on the first screen when the monitors wake up). I do have a DVI->VGA adapter I can try out if I can find another VGA cable (probably buried in my random computer hardware box out in the garage).
However, looking in some bugreports, it seems like this might've been fixed in drm-intel-next, but hasn't made it to a kernel yet (I guess they're hoping it will make it in the 4.8 because it didn't make the 4.7 RCs). I'm not too sure how to build a kernel with the latest drm-intel-next, but I'll might look into it over the next couple of days. Then I'll need to dig into Eric's liveslak script so I can regenerate an initrd for Slackware Live (since it's different from a normal initrd).
I'm now testing git_20160628_784dc02 version of the driver.
On a side note, previous commit 97be00cb38ac508acc1dbbfda4835d32b0445fa4 has the following description:
Code:
pciids: Add more Kabylake PCI IDs.
The spec has been updated adding new PCI IDs.
In parity with kernel:
commit 33d9391d3020e069dca98fa87a604c037beb2b9e
Author: Rodrigo Vivi <rodrigo.vivi@intel.com>
Date: Thu Jun 23 14:50:35 2016 -0700
drm/i915: Add more Kabylake PCI IDs.
I've been keeping up to date with the commits to driver.
With a recent update to the LibreOffice (5.2.0.4) I noticed strange problem.
When I go to the settings, and I try to check/uncheck checkbox, I can notice that the checkbox's frame disappears for a moment.
This was only happening if the OpenGL in LibreOffice was disabled (which is my case, because if I enable OpenGL, documents are rendered slower).
Later on, I noticed another problem. That is, KSnapshot would capture multiple "selection frames", when trying to select capture area.
I started bisecting the Intel's driver code, and sure enough, I found the commit that was causing those problems.
The winner is:
Code:
commit 1f6dfc9df6780bd3768bf065156b6e8dae05b80c (patch)
tree a8826a476f5a257039059dbf30e9b82f8658a94b
parent 49daf5df124b5ae6c7508e934768c292f4143040 (diff)
sna: Only flush GPU bo for a damage event
Based on xf86-video-ati
commit 9a1afbf61fbb2827c86bd86d295fa0848980d60b
Author: Michel Dänzer <michel.daenzer@amd.com>
Date: Mon Jul 11 12:22:09 2016 +0900
Use EventCallback to avoid flushing every time in the FlushCallback
reports seeing an improvement in reducing flushes at the expense of
checking every event for a DamageNotifyEvent. Since we also mix
rendering with SHM buffers, we have a more diverse set of conditions
under which to flush - but maybe we will see enough of a win for DRI to
merit. So far seeing improvement of ~20% for series of small operations
under the compositor without seeing any regressions, should benefit
composited desktop users. The biggest danger here is missed flushes.
So, the last properly working commit is 49daf5df124b5ae6c7508e934768c292f4143040.
I'm sticking now with 49daf5df124b5ae6c7508e934768c292f4143040 till the problem is resolved.
I tried simply reverting 1f6dfc9df6780bd3768bf065156b6e8dae05b80c, but it generates conflicts, which I couldn't be bothered to resolve :-)
On a side note, I tried xf86-video-modesetting too. It seems to be working most of the time, but as soon as I enable OpenGL in LibreOffice, the UI gets corrupted.
On a side note, I tried xf86-video-modesetting too. It seems to be working most of the time, but as soon as I enable OpenGL in LibreOffice, the UI gets corrupted.
and reboot.
glxgears now shows about 20% more FPS, no issues so far.
Maybe its worth to keep an eye on this alternative also for graphical issues on skylake.
@franzen: Are you still using xf86-video-modesetting? Any side effects or problems?
I removed xf86-video-intel two days ago but it is too short time to say if everything works well. xf86-video-modesetting fixed 2 issues I had with the intel driver: no more glitches in LibreOffice Calc; no more flickering on my left monitor when it displays full screen web content while there is some activity on the right monitor.
@franzen: Are you still using xf86-video-modesetting? Any side effects or problems?
It's my father's Intel, he is using xf86-video-modesetting since then, the pc is up several hours per day, no issues so far. Before, he had freezes in Xscreensaver und glitches in Libreoffice.
I've just swapped to the modesetting driver after getting fed up of the various issues I was having with the xf86-video-intel driver.
Using the compton compositor ("compton -b --vsync drm --backend glx" ) with my dwm window manager seems to be working ok so far and appears to fix the video tearing issues I was having with the intel driver. I'm also not seeing the gui artefacts/corruption I was with the intel driver.
Only downside so far is that xbacklight doesn't seem to work with modesetting so I'm having to use the files in /sys/class/backlight/intel_backlight to change the brightness.
Still early days, but if anything else crops up, I'll come back and report.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.
