Video Problem Since 2015-04-01 Update

Home » CentOS » Video Problem Since 2015-04-01 Update
CentOS 13 Comments

Hello!

No updates for days until 2015-04-01. Then presented with huge steaming pile of updates all at once, apparently related to CentOS 7.1 release.

Not thinking about it being April Fools Day, I immediately did:

yum -v check-update, then

yum -v update

then rebooted.

After reboot, there is a problem with the video. It seems to flicker when the mouse pointer touches the screen edges. Also, sometimes when viewing pages in Firefox, the window becomes partially covered with horizontal black lines.

During reboots, before the login screen appears, there are quickly disappearing messages seeming to include references to “drm:9” and some sort of “underrun”.

In the terminal window, I get a message:

ABRT has detected 3 problem(s). For more info run: abrt-cli list –since
1428078184

When I do that, I get:

id (blah, blah)
reason: WARNING: at drivers/gpu/drm/i915/intel_display.c:869
intel_wait_for_vblank+0x211/0x220 [i915]()
time: Fri 03 Apr 2015 11:41:49 AM CDT
cmdline: BOOT_IMAGE=/vmlinuz-3.10.0-229.1.2.el7.x86_64
root=UUID=(blah, blah) ro vconsole.keymap=us vconsole.font=latarcyrheb-sun16 rhgb quiet LANG=en_US.UTF-8
count: 1
Directory: /var/tmp/abrt/oops-2015-04-03-11:41:49-590-0

id (blah, blah)
reason: WARNING: at drivers/gpu/drm/i915/intel_display.c:869
intel_wait_for_vblank+0x211/0x220 [i915]()
time: Wed 01 Apr 2015 12:02:12 PM CDT
cmdline: BOOT_IMAGE=/vmlinuz-3.10.0-229.1.2.el7.x86_64
root=UUID=(blah, blah) ro vconsole.keymap=us vconsole.font=latarcyrheb-sun16 rhgb quiet LANG=en_US.UTF-8
package: kernel count: 11
Directory: /var/tmp/abrt/oops-2015-04-01-12:02:11-1220-1
Reported: cannot be reported

id (blah, blah)
reason: WARNING: at drivers/gpu/drm/i915/intel_display.c:869
intel_wait_for_vblank+0x211/0x220 [i915]()
time: Fri 03 Apr 2015 11:41:49 AM CDT
cmdline: BOOT_IMAGE=/vmlinuz-3.10.0-229.1.2.el7.x86_64
root=UUID=(blah, blah) ro vconsole.keymap=us vconsole.font=latarcyrheb-sun16 rhgb quiet LANG=en_US.UTF-8
Directory: /var/tmp/abrt/oops-2015-04-03-11:41:49-590-1
Reported: cannot be reported

The Autoreporting feature is disabled. Please consider enabling it by issuing
‘abrt-auto-reporting enabled’ as a user with root privileges.

So is this line where the problem may be?:

drivers/gpu/drm/i915/intel_display.c:869 intel_wait_for_vblank+0x211/0x220
[i915]()

Started to search online for information, but quickly decided:

1) available info is over my head

2) when in a hole, stop digging

So, any obvious answers, other than a complete reinstall?

Note: I really thought CentOS is about reinstalling after years, not months. Grrr . . .

13 thoughts on - Video Problem Since 2015-04-01 Update

  • Well, it could be the same, but the bug report does seem to refer to the bug being in kernel series 3.18. I am just using whatever 3.10 series kernel comes standard with CentOS 7.x.

    I don’t know anything about bug reports, and I an not familiar with swapping out kernels in CentOS.

    BTW, is El Repo another name for the EPEL repository?

  • Well, the CentOS kernel is a 3.10 kernel in name (or number) only. Red Hat backport much stuff from later kernels, and in this case that includes much broken code for the Intel i915 driver. The code is broken from around 3.15 and still not completely fixed in the current 4.0-rc6
    code base. Updating to a newer kernel, and hence a newer i915 driver, may or may not help, but it doesn’t hurt to try.

    No, they are two completely separate repositories:

    http://wiki.CentOS.org/AdditionalResources/Repositories

  • Okay, thanks to all who took the time to reply.

    What a shame; I hadn’t planned on spending Easter weekend learning to do kernel upgrades, but . . . here it goes.

  • Is there a way in CentOS 7 to boot into an alternate video setup?

    Some distributions have an option to boot into something called VESA (or something similar). Something that might not have the fanciest features, but more likely to at least just work.

  • Check the kernel parameters – ISTR long ago using a paramer to the kernel added to the grub boot line that had one of my machines go into VESA mode. ISTR some numbers after “=” that specied which VESA mode to use to get better resolution (more lines/screen).

    Bill

  • I am getting the same problem as the OP but I am on nvidia.
     
    —–Original message—–
    From:Bill Maltby (C4B)
    Sent:Sat 04-04-2015 12:32 pm Subject:Re: [CentOS] video problem since 2015-04-01 update To:CentOS@CentOS.org;
    Check the kernel parameters – ISTR long ago using a paramer to the kernel added to the grub boot line that had one of my machines go into VESA mode. ISTR some numbers after “=” that specied which VESA mode to use to get better resolution (more lines/screen).

    Bill

  • Jay Warren wrote:
    I can confirm the problem on Nvidia GTX 750/760 with kmod-nvidia 346.47
    proprietary driver from elrepo. As a workaround you could try to set PowerMizer mode to maximum performance using nvidia-settings utility.

    I am not sure if that related to OP, or if that relates to elrepo release, but it looks like the bug was introduced in update to 7.1.

  • Yes, that is very much like the problem I encountered, but in my case it did not need multiple monitors to happen, just one. The config file suggested did seem to fix the problem, at least serving as a (hopefully)
    temporary workaround.

    Johnny, thanks for the pointer. And thanks to all who replied.

    P.S. – the workaround seems to work with both the latest “standard” CentOS
    7 kernel and the ML kernel from the El Repo repository. But today’s X11
    updates (done before the config file was done) made no difference.

  • I spoke too soon. Although the screen flicker has been worked around, I
    still intermittently get lots of horizontal black lines when going to gmail. No other distribution shows this problem, only CentOS 7, and only after the 1503
    updates.

  • I am experiencing the same issue. My screen is flickering, in terminal or when I use Mozilla. Can it be solved somehow? Will it be solved in the future?

LEAVE A COMMENT