Multiple (EE) in X log

PaulBx1 · 2013-01-19 22:56:59

When I am in X and doing things like scrolling up and down in journalctl, sometimes my console window starts overwriting itself making the characters unreadable. I believe at the same time, my X log fills with many of these messages:

[ 14926.782] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.

There are no other EE's in the log. My system log shows nothing different than was reported before my upgrade. However there are a couple things in there like a problem with drm, but I discount them since they are the same before the upgrade too.

This started after my latest system upgrade yesterday. The previous upgrade had been a couple of months ago.

Once I start getting this, it doesn't stop. It also affects browser windows. It doesn't happen all the time; for example after I rebooted I found no problems in the X log (while I am posting now).

If I go back to single user and use journalctl there, I never see any problem, so I'm assuming it is X.

The machine is an Intel D510MO.

[root@myhost xorg.conf.d]# lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation Atom Processor D4xx/D5xx/N4xx/N5xx Integrated Graphics Controller (rev 02)

So far, googling around, I haven't found much about this error.

PaulBx1 · 2013-01-19 23:08:51

When I grep my pacman log for "xorg" I get this:

[2013-01-18 13:41] upgraded xorg-bdftopcf (1.0.3-2 -> 1.0.4-1)
[2013-01-18 13:41] upgraded xorg-server-common (1.13.0-4 -> 1.13.1-1)
[2013-01-18 13:41] upgraded xorg-server (1.13.0-4 -> 1.13.1-1)

erville · 2013-01-21 14:03:47

I have the exact same problem: http://i.imgur.com/55X33mo.png

% tail /var/log/Xorg.0.log
[ 10766.648] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 10776.961] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 10778.017] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 10778.603] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 10786.625] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 10895.163] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 11033.672] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 11039.513] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 11043.315] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.
[ 11227.474] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Resource deadlock avoided.

This machine's a Lenovo S10-3.

% lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation Atom Processor D4xx/D5xx/N4xx/N5xx Integrated Graphics Controller
% pacman -Q xorg-server xf86-video-intel intel-dri
xorg-server 1.13.1-1
xf86-video-intel 2.20.18-1
intel-dri 9.0.1-1

Anyone have any ideas?

Trilby · 2013-01-21 14:15:26

EDIT: sorry, I *thought* this looked familiar but when I found the other thread I was thinking of it turned out to be quite different.

Last edited by Trilby (2013-01-21 14:18:43)

erville · 2013-01-21 14:56:49

Trilby wrote:

EDIT: sorry, I *thought* this looked familiar but when I found the other thread I was thinking of it turned out to be quite different.

Yeah, looked at it too. Theirs was an NVIDIA/ATI video driver problem. Our case on the other hand involves Intel GMA. The symptoms are similar though. Fonts are rendered wrong (characters are swapped and font colors change) and gtk apps rendering is corrupted (frozen sections of the gui/window are overlayed by new/updated sections).

PaulBx1 · 2013-01-21 18:56:59

Yes, that more or less describes my symptoms. I see windows go black, or the text in them gets frozen in place and overwritten, scrolling stops working, etc. When I reboot it recovers again, for a while.

I have googled for the error message (including the "Resource deadlock avoided") and the only thing found by google in the whole internet space is this thread!

Last edited by PaulBx1 (2013-01-21 18:57:53)

avrelaun · 2013-01-23 14:46:50

Same problem here since a few days ago on a Samsung N148 Plus with Intel Atom graphics.

Apparently it appears immediately after using some fast-refreshing screen apps like mplayer2 or skype video.

EDIT: I can confirm this is a xf86-video-intel bug, downgrading to 2.20.17 fixes it.

Last edited by avrelaun (2013-01-23 15:06:17)

PaulBx1 · 2013-01-23 17:00:27

I filed a bug:
https://bugs.freedesktop.org/show_bug.cgi?id=59769

avrelaun · 2013-01-23 17:08:09

PaulBx1 wrote:

I filed a bug:
https://bugs.freedesktop.org/show_bug.cgi?id=59769

Oops, I did it too: https://bugs.freedesktop.org/show_bug.cgi?id=59771

PaulBx1 · 2013-01-23 18:01:05

Looks like mine is now a dup of yours, so I started posting on yours.

erville · 2013-02-06 18:03:18

Tearing windows seems to be resolved. However, another issue crops up. Video playback in MPlayer becomes completely black (audio plays fine), and playing it from the terminal gives this error message:

X11 error: BadAlloc (insufficient resources for operation)

Xorg log:

% cat /var/log/Xorg.0.log | grep \(EE\)
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[ 10203.594] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption: Resource deadlock avoided.

However, video playback is easily fixed by using x11 video output (as opposed to xv which is what mplayer probably uses by default):

mplayer -vo x11 foo.mkv

I'm guessing we got a temporary fix which reverts graphics back to software emulation whenever the X log (EE) error happens. Is this a correct guess? Does anyone else experience this?

dsoul · 2013-02-09 08:51:07

I've turned on options DebugFlushBatches, DebugFlushCaches, DebugWait in xorg conf for intel driver and problem vanished.

Alaska23 · 2013-03-25 13:32:54

dsoul wrote:

I've turned on options DebugFlushBatches, DebugFlushCaches, DebugWait in xorg conf for intel driver and problem vanished.

Since the update to "xorg-server 1.14" the old and working driver "xf86-video-intel 2.20.17" isn't compatible anymore, so i had to update the driver to "xf86-video-intel 2.21.5", which made the error reappear.
After I enabled these three debug-options, the error didn't occur again - Thanx for the hint.

crequill · 2013-04-26 11:03:59

Since two days, it seems the problem disappear with sna mode:

Section "Device"
    Identifier	"i915"
    Driver	"intel"
    Option	"AccelMethod" "sna"
EndSection

Arch Linux

#1 2013-01-19 22:56:59

Multiple (EE) in X log

#2 2013-01-19 23:08:51

Re: Multiple (EE) in X log

#3 2013-01-21 14:03:47

Re: Multiple (EE) in X log

#4 2013-01-21 14:15:26

Re: Multiple (EE) in X log

#5 2013-01-21 14:56:49

Re: Multiple (EE) in X log

#6 2013-01-21 18:56:59

Re: Multiple (EE) in X log

#7 2013-01-23 14:46:50

Re: Multiple (EE) in X log

#8 2013-01-23 17:00:27

Re: Multiple (EE) in X log

#9 2013-01-23 17:08:09

Re: Multiple (EE) in X log

#10 2013-01-23 18:01:05

Re: Multiple (EE) in X log

#11 2013-02-06 18:03:18

Re: Multiple (EE) in X log

#12 2013-02-09 08:51:07

Re: Multiple (EE) in X log

#13 2013-03-25 13:32:54

Re: Multiple (EE) in X log

#14 2013-04-26 11:03:59

Re: Multiple (EE) in X log

Board footer