Bug 869856 - [Arrandale] Text disappearing in Firefox and Terminal
[Arrandale] Text disappearing in Firefox and Terminal
Product: Red Hat Enterprise Linux 6
Classification: Red Hat
Component: kernel (Show other bugs)
Unspecified Unspecified
unspecified Severity unspecified
: rc
: ---
Assigned To: Dave Airlie
Desktop QE
: Regression
Depends On:
  Show dependency treegraph
Reported: 2012-10-24 20:44 EDT by Dave Airlie
Modified: 2013-02-21 01:53 EST (History)
1 user (show)

See Also:
Fixed In Version: kernel-2.6.32-338.el6
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: 852210
Last Closed: 2013-02-21 01:53:16 EST
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

  None (edit)
Description Dave Airlie 2012-10-24 20:44:56 EDT
+++ This bug was initially created as a clone of Bug #852210 +++

Created attachment 607348 [details]

Description of problem:

I don't have a way to reproduce this, and there's nothing significant
in the logs, but this laptop has got itself into a state where text
is disappearing all over the place, seemingly in all applications.
See the attached screenshot.

Version-Release number of selected component (if applicable):

[   180.629] Build Date: 08 August 2012  09:55:23AM
[   180.631] Build ID: xorg-x11-server 

How reproducible:

? Just once.

Steps to Reproduce:
1. Unknown.
Actual results:

Expected results:

Additional info:

--- Additional comment from rjones@redhat.com on 2012-08-27 17:25:48 EDT ---

Created attachment 607349 [details]

--- Additional comment from airlied@redhat.com on 2012-08-27 17:48:13 EDT ---

rpm -q xorg-x11-drv-intel please.

--- Additional comment from rjones@redhat.com on 2012-08-28 03:40:52 EDT ---


Some other points covered on IRC last night:

 - Running XFCE.

 - Font antialising is set to Grayscale.

 - Compositor disabled.

 - Restart of the X server fixed the problem.

 - Same hardware runs F16/F17 just fine for over a year.

--- Additional comment from thingie@bahnhof.cz on 2012-09-14 05:00:12 EDT ---

I have a similar (not sure if same) problem with an Intel Arrandale chipset graphics on both Fedora 17 and 18 on two different computers. Both have the same chipset. It does happen only after few hours of use. There is nothing interesting in the logs either.

Key difference is that it's happening only with 3.5 kernels and newer (3.6-rc on FC 18, even latest 3.6 checkout that I've compiled mysefl). It doesn't happen with the latest F17 3.4 kernel or any older, not even a long period of use without any restart (weeks).

--- Additional comment from thingie@bahnhof.cz on 2012-09-14 05:01:07 EDT ---

I should have also added that I don't use XFCE, but default Gnome3 environment with compositing enabled.

--- Additional comment from rebus@seznam.cz on 2012-09-17 06:41:59 EDT ---

Hello I have got the same issue (using default Gnome3 + gnome-shell). Fonts disappearing especially in firefox, thunderbird, gnome-terminal. It usually disappears or appears back when the part of the screen gets redrawn on some event like mouse-over or scroll with the page, resize a font on the webpage (ctrl+ ctrl-).


I have tried now to downgrade back to xorg-x11-server-Xorg-1.12.0-2.fc17.x86_64 - I believe I was not seeing these issues before upgrade to latest Xorg server.

Michal Ambroz

--- Additional comment from choeger@umpa-net.de on 2012-09-20 09:27:03 EDT ---

Same hier. This is a lenovo X201 with fedora 17 all stable updates.

If there is any particular log that could be helpful, let me know.

--- Additional comment from airlied@redhat.com on 2012-09-20 17:20:24 EDT ---

I'm currently bisecting this but its a long process, since it takes so long to know the kernel is good.

Michael did the old server help?

I'm currently putting it down as a 3.4->3.5 kernel regression.

--- Additional comment from rebus@seznam.cz on 2012-09-20 18:38:13 EDT ---

Hello Dale,
I have downgraded only xorg-x11-server-Xorg. The issue didn't re-appear since then (4+ days).

I am currently running kernel 3.5.3-1.fc17.x86_64 and I have not changed that part. The issue appeared on the same version of kernel so I would not say it is related.


--- Additional comment from airlied@redhat.com on 2012-09-20 18:47:47 EDT ---

Michael so what server are you running now?

--- Additional comment from thingie@bahnhof.cz on 2012-09-21 03:52:42 EDT ---

I've looked into last and yum history, it seems that it had happened for the first time in the early august ( :-) ) with the first 3.5 kernel f17 update, xorg and intel driver version back then were:

xorg-x11-server-Xorg 1.12.2-4.fc17.x86_64
xorg-x11-drv-intel 2.20.1-1.fc17.x86_64

I haven't tried to downgrade them, but booting 3.4 kernel have always helped (running 3.4.6-2.fc17.x86_64 right now).

It's also happening with my F18 alpha system, but I don't have access to it right now.

--- Additional comment from choeger@umpa-net.de on 2012-09-21 05:53:42 EDT ---

I am pretty confident the issue appeared first after I was back from a conference. That was around the 6th of september. I did not update the x server since then. Relevant yum.log lines:

[choeger@elrond modim]$ sudo egrep "(xorg)|(kernel)" /var/log/yum.log | grep Sep
Sep 06 13:47:44 Updated: xorg-x11-drv-intel-2.20.5-1.fc17.x86_64
Sep 06 13:47:45 Updated: kernel-headers-3.5.3-1.fc17.x86_64
Sep 06 13:48:15 Installed: kernel-devel-3.5.3-1.fc17.x86_64
Sep 06 13:49:00 Installed: kernel-3.5.3-1.fc17.x86_64
Sep 10 11:52:42 Updated: xorg-x11-drv-intel-2.20.6-1.fc17.x86_64

So for me it seems to be either the intel driver (does it still do alot nowadays?) or the kernel itself.

--- Additional comment from thingie@bahnhof.cz on 2012-09-28 08:14:42 EDT ---

Tried with 3.6-rc7 and recent F17 Xorg and drv-intel updates, worked well for two days, but eventually fonts have started to fall apart again.

--- Additional comment from airlied@redhat.com on 2012-10-03 01:17:48 EDT ---


I've built a scratch kernel with a workaround from intel guys, its a hack just to narrow down things, I can't reproduce here with this patch in place, but it might just be luck at this point.

--- Additional comment from airlied@redhat.com on 2012-10-03 04:48:16 EDT ---


what hw are you running on? can you attach an Xorg log please.


--- Additional comment from airlied@redhat.com on 2012-10-03 05:21:01 EDT ---

also guys how much RAM have you got in these machines?

--- Additional comment from rjones@redhat.com on 2012-10-03 05:55:31 EDT ---

This bug has not happened to me since I reported it (I
use the machine 24/7).  However FWIW the machine has 8 GB
of RAM.

$ free -m
             total       used       free     shared    buffers     cached
Mem:          7762       3966       3796          0        647       1298
-/+ buffers/cache:       2021       5741
Swap:         5823         82       5741

--- Additional comment from rebus@seznam.cz on 2012-10-03 13:55:58 EDT ---

Hi Dave,
Sorry - I am not able to reproduce the issue since I downgraded the Xorg to fc17 release version so I might not be able to help to narrow it down.

I have even tried last week to upgrade to latest Xorg and kernel in FC17 - not able to reproduce the error:

(In reply to comment #16)
> also guys how much RAM have you got in these machines?

(In reply to comment #15)
> what hw are you running on? can you attach an Xorg log please.
I have got notebook HP EliteBook 6930p with integrated Intel GPU, using dualhead (notebook LCD + external DVI connected monitor)

$ lspci |grep -i vga
00:02.0 VGA compatible controller: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07)

[    78.682] (--) intel(0): Integrated Graphics Chipset: Intel(R) GM45

Best regards
Michal Ambroz

--- Additional comment from airlied@redhat.com on 2012-10-05 18:49:34 EDT ---

*** Bug 851280 has been marked as a duplicate of this bug. ***

--- Additional comment from jrankin@redhat.com on 2012-10-07 11:47:09 EDT ---

Dave, I can reproduce this within a couple days on F17 whenever I'm on a DVI monitor (via docking station) *and* when I'm on any 3.5 kernel. I've never seen this issue on any 3.4 kernels.

Do you want me to try one of the kernels in comment 14?

--- Additional comment from airlied@redhat.com on 2012-10-07 20:45:13 EDT ---


yes please try that kernel.

--- Additional comment from kparal@redhat.com on 2012-10-08 08:23:38 EDT ---

This might be another duplicate: bug 848099

--- Additional comment from jrankin@redhat.com on 2012-10-12 21:30:24 EDT ---

I'm three days on this new kernel (3.5.5-1.bz852210.fc17.x86_64) now, and the problem has not occurred. In the previous 3.5 kernels, the problem always happened when I have the laptop in a DVI docking station for more than a day. I'll continue using this kernel and will report back if the problem does appear.

--- Additional comment from choeger@umpa-net.de on 2012-10-16 07:46:43 EDT ---

It just happened again on 3.6.1 (attached to a docking station via display link).

I tried to switch to a different VT and got this:

[18668.743604] WARNING: at drivers/gpu/drm/i915/intel_display.c:1394 assert_pch_hdmi_disabled+0xb8/0xf0 [i915]()
[18668.743605] Hardware name: 3323REG
[18668.743606] IBX PCH hdmi port still using transcoder B
[18668.743646] Modules linked in: fuse rfcomm bnep ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack ip6table_filter ip6_tables snd_hda_codec_hdmi snd_hda_codec_conexant snd_hda_intel snd_hda_codec snd_hwdep snd_seq qcserial usb_wwan snd_seq_device snd_pcm ecryptfs arc4 uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_core encrypted_keys videodev iwldvm snd_page_alloc e1000e snd_timer coretemp mei iTCO_wdt media iTCO_vendor_support mac80211 lpc_ich iwlwifi mfd_core btusb bluetooth i2c_i801 kvm trusted intel_ips tpm tpm_bios microcode cfg80211 uinput thinkpad_acpi snd soundcore rfkill crc32c_intel ghash_clmulni_intel wmi i915 video i2c_algo_bit drm_kms_helper drm i2c_core
[18668.743650] Pid: 904, comm: Xorg Not tainted 3.6.1-1.fc17.x86_64 #1
[18668.743651] Call Trace:
[18668.743660]  [<ffffffff8105b84f>] warn_slowpath_common+0x7f/0xc0
[18668.743663]  [<ffffffff8105b946>] warn_slowpath_fmt+0x46/0x50
[18668.743684]  [<ffffffffa0094058>] assert_pch_hdmi_disabled+0xb8/0xf0 [i915]
[18668.743703]  [<ffffffffa009694c>] ironlake_crtc_disable+0x24c/0x7b0 [i915]
[18668.743711]  [<ffffffff812e31f1>] ? vsnprintf+0x461/0x600
[18668.743730]  [<ffffffffa0096f3e>] ironlake_crtc_dpms+0x8e/0xa0 [i915]
[18668.743749]  [<ffffffffa0092118>] intel_crtc_dpms+0x48/0x140 [i915]
[18668.743768]  [<ffffffffa009a735>] intel_crtc_disable+0x35/0xb0 [i915]
[18668.743777]  [<ffffffffa0065825>] drm_helper_disable_unused_functions+0x115/0x170 [drm_kms_helper]
[18668.743785]  [<ffffffffa006711a>] drm_crtc_helper_set_config+0x91a/0xb00 [drm_kms_helper]
[18668.743793]  [<ffffffffa0064f12>] drm_fb_helper_set_par+0x82/0x100 [drm_kms_helper]
[18668.743797]  [<ffffffff8131b605>] fb_set_var+0x1a5/0x470
[18668.743802]  [<ffffffff81091179>] ? try_to_wake_up+0x1d9/0x2c0
[18668.743807]  [<ffffffff81327f91>] fbcon_blank+0x1e1/0x2e0
[18668.743812]  [<ffffffff81396e96>] do_unblank_screen+0xb6/0x1f0
[18668.743817]  [<ffffffff8138c4c5>] complete_change_console+0x65/0xf0
[18668.743820]  [<ffffffff8138d74c>] vt_ioctl+0x11fc/0x1300
[18668.743833]  [<ffffffffa0014fe2>] ? drm_ioctl+0xf2/0x580 [drm]
[18668.743837]  [<ffffffff810ef730>] ? call_rcu_bh+0x20/0x20
[18668.743842]  [<ffffffff81382118>] tty_ioctl+0x258/0xb50
[18668.743847]  [<ffffffff8161e17c>] ? do_page_fault+0x1bc/0x4b0
[18668.743851]  [<ffffffff811a0bc9>] do_vfs_ioctl+0x99/0x580
[18668.743855]  [<ffffffff811a1149>] sys_ioctl+0x99/0xa0
[18668.743858]  [<ffffffff816226e9>] system_call_fastpath+0x16/0x1b
[18668.743860] ---[ end trace fddc7b996825e5f1 ]---

--- Additional comment from jrankin@redhat.com on 2012-10-19 11:54:23 EDT ---

This problem just reappeared on the 3.5.5-1.bz852210.fc17.x86_64 kernel, though it's much less severe so far. I'm having small amounts of text in xchat disappear when I mouse over them.

--- Additional comment from jrankin@redhat.com on 2012-10-23 18:28:43 EDT ---

I tried the new 3.6.2-4.fc17.x86_64 kernel, and the problem definitely exists, and may even be worse than before. I literally can't see this text as I'm typing it out.

--- Additional comment from jwboyer@redhat.com on 2012-10-24 08:28:51 EDT ---

Dave believes this is fixed with commit 504c7267a1e84b157cbd7e9c1b805e upstream.  I've added that to the fedora kernel on F16-F18.
Comment 1 Dave Airlie 2012-10-24 20:47:23 EDT
I think we've fixed this now, it was broken upstream in 3.5 kernel and 3.7-rc2 has a fix, that I've gotten into Fedora and stable upstream.

The problem occurs mainly on the Thinkpad Ironlake systems, at least I can't reproduce on my HP one. t410, t410s, t510 I think.

It can take 2 or 3 days of use to provoke the issue.

The fix is a one line kernel patch that I'll post shortly.
Comment 2 RHEL Product and Program Management 2012-10-24 20:50:45 EDT
This request was evaluated by Red Hat Product Management for
inclusion in a Red Hat Enterprise Linux release.  Product
Management has requested further review of this request by
Red Hat Engineering, for potential inclusion in a Red Hat
Enterprise Linux release for currently deployed products.
This request is not yet committed for inclusion in a release.
Comment 4 cblaauw 2012-10-28 16:07:22 EDT
I have invisible text in firefox, thunderbird and gnome-terminal. Sometimes garbled stuff in window title. When moving mouse over the text is appears in thunderbird. Kernel is 3.6.3-1.fc17.x86_64. 

Graphics are intel G41, plain desktop machine.

Is this there a newer kernel that takes care of the issue?
Comment 5 Jarod Wilson 2012-11-01 16:25:28 EDT
Patch(es) available on kernel-2.6.32-338.el6
Comment 10 errata-xmlrpc 2013-02-21 01:53:16 EST
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.


Note You need to log in before you can comment on or make changes to this bug.