Screen blanks and freezes after a few minutes of idling, possibly due to broken DPMS

Bug #136853 reported by Martin-Éric Racine on 2007-09-02
26
This bug affects 1 person
Affects Status Importance Assigned to Milestone
xf86-video-siliconmotion
Confirmed
Medium
xorg (Ubuntu)
Low
Unassigned

Bug Description

Since switching to Gutsy to verify whether the packages I maintain still operate as expected, I have noticed that after a few minutes of idle operation, the screen goes blank and cannot be returned to normal. The only way to get screen content again is to reboot, which is possible either by pressing the ACPI-enabled power button or a remote SSH connection. Broken DPMS is suspected, although I have no way of confirming this.

Purposely downgrading xserver-xorg-core and anything that depend on it, then pining it to that version, while keeping all the rest of the packages to Gutsy versions solves it, which suggests something related to X.org upgrade from 1.2 to 1.3 might be at play.

Note: someone on #ltsp mentions a similar situation using xserver-xorg-video-amd, yet the exact same AMD driver doesn't exhibit this behavior when compiled and used with earlier X releases.

Tormod Volden (tormodvolden) wrote :

Can you please attach xorg.conf and Xorg.0.log ?

Changed in xserver-xorg-video-siliconmotion:
assignee: nobody → tormodvolden
status: New → Incomplete
Martin-Éric Racine (q-funk) wrote :

Tormold Volden pointed me to 12ubuntu4, which indeed appears to solve the "DPMS that never wakes up" issue, but not the vcons switching issue.

Martin-Éric Racine (q-funk) wrote :

Here are the files.

Martin-Éric Racine (q-funk) wrote :
Martin-Éric Racine (q-funk) wrote :

Actually, I take that back: the display cannot be awaken by normal means using 12ubuntu4 either.

Changed in xserver-xorg-video-siliconmotion:
assignee: tormodvolden → nobody
status: Incomplete → Confirmed
Bryce Harrington (bryce) on 2007-10-01
Changed in xserver-xorg-video-siliconmotion:
importance: Undecided → High
Oliver Mihatsch (olli-knaeul) wrote :

Same problem here, after updating to gutsy my Amilo M (Silicon Motion SM720 Lynx3DM) runs awfully slow. The Xorg process took 100% CPU just to redraw a window. Also text console was not working.

I've found two workarounds:
To reenable the text console (from Bug #144178 )
"sudo modprobe vga16fb"
"sudo modprobe fbcon"
and ALT+F1 should work again

To get rid of the corruptions and performance problems change in the /etc/X/xorg.conf from 24bit color depth to 16 bit color depth (Display Section). (from Bug #154514 )

I have exactly the same problems described with Xubuntu Gutsy on a Thinkpad i1300 (1171-350).
1. The screen once idle for a time, goes black and never returns no matter the key that you press.
2. The vga console problem. This was solved adding at the end of /etc/modules:

fbcon
vga16fb

3. One more strange effect: with tuxpaint and some java educational app (jclic) the mouse pointer gets garbbled, if I select the --nofancycursors switch with tuxpaint, then the pointer appears as a normal black arrow, buth with two more "shadow" pointers at the right side.

Attached xorg.conf

And Xorg.0.log

Oliver Mihatsch (olli-knaeul) wrote :

It looks like the new EXA acceleration caused that problem:
https://bugs.freedesktop.org/show_bug.cgi?id=8721

After adding the following entry into the Device section of the xorg.conf, 24 bit depth and xv seems to work:
Option "AccelMethod" "EXA"

The man page says it's not really mature, but instead of the old and 'stable' acceleration this one works. :-(

Bryce Harrington (bryce) wrote :

Is this issue still occurring with hardy?

Martin-Éric Racine (q-funk) wrote :

It still does here, on my old ThinkPad.

Forwarding from a Ubuntu bug report by Martin-Éric Racine and confirmed by several others:
https://bugs.launchpad.net/ubuntu/+bug/136853

"After a few minutes of idle operation, the screen goes blank and cannot be returned to normal. The only way to get screen content again is to reboot, which is possible either by pressing the ACPI-enabled power button or a remote SSH connection. Broken DPMS is suspected, although I have no way of confirming this.

Purposely downgrading xserver-xorg-core and anything that depend on it, then pining it to that version, while keeping all the rest of the packages to Gutsy versions solves it, which suggests something related to X.org upgrade from 1.2 to 1.3 might be at play."

http://launchpadlibrarian.net/9095014/Xorg.0.log
http://launchpadlibrarian.net/9095015/xorg.conf

The same issue is seen in Ubuntu Hardy.

The vcons issue is covered by bug 144178, and since it has a more detailed analysis of the issue let's focus this bug report on the broken DPMS issue.

The corruption and performance issues mentioned in comment #6 are covered by bug #154514.
For the tuxpaint issue described in comment #7, bug #144182 seems to cover that.

description: updated
Changed in xf86-video-siliconmotion:
status: Unknown → Confirmed
David Burrows (snadge) wrote :

I can also confirm this issue, unfortunately after moving from Gentoo, to Ubuntu 8.04, I also get the vcons switching issue (which the posted workaround in <a href="https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-siliconmotion/+bug/144178">bug #144178</a> does not seem to work for me. Perhaps this is because I'm using vesafb, and not vga16fb).

I wish I could be of more use. Disabling DPMS is the only workaround I can think of for the moment. Is anyone in a position to actually debug this and investigate the cause?

Provide (sloan2189) wrote :

Confirmed on a Gateway Solo 2550 (8.04.1 Server)

Francisco L. Fernandez: You can get rid of the mouse "artifacts" by adding:

Option "HWCursor" "off"

to the Device section of your xorg.conf.

Adding fbcon and vga16fb to /etc/modules fixed the issue of a blank screen once leaving X.

Tormod Volden (tormodvolden) wrote :

Unfortunately there has been no response to the upstream bug report. If somebody can try latest X packages (Intrepid or even better Jaunty) and confirm the problem is still there, we could update the report.

Is this still a problem with the last driver version?

Daniel Durand (durand-d) wrote :

To try to resolve the problem on my Thinkpad series 1200 (xubuntu 8.10), I installed xubuntu jaunty : same result ....!
Daniel

Tormod Volden (tormodvolden) wrote :

Daniel, can you please attach your Jaunty Xorg.0.log?

Yes, reported to still be an issue with version 1.7.0 on xserver 1.6.

2009/3/18 Tormod Volden <email address hidden>

> Daniel, can you please attach your Jaunty Xorg.0.log?
>
> --
>
M Jaunty log (in late, excuse me...). Thanks for answering
--
Daniel Durand

Francisco Jerez (currojerez) wrote :

Daniel: according to your logs you have a monitor connected to the VGA output, is that right or is monitor detection failing?
If you really have a monitor connected, do you experience the DPMS problem with the external monitor?

Could you try setting:
> Option "UseBIOS" "off"
in your xorg.conf, Section "Device"?

You could also try to build the current development version from sources. To check out a copy from the git repository:
> git clone git://anongit.freedesktop.org/xorg/driver/xf86-video-siliconmotion

Daniel Durand (durand-d) wrote :

2009/3/23 Francisco Jerez <email address hidden>

> Daniel: according to your logs you have a monitor connected to the VGA
> output, is that right or is monitor detection failing?

No, I just have my IBM thinkpad 1200 series and no external monitor
selected....

For other tests, i can't do them before next saturday...

>
> If you really have a monitor connected, do you experience the DPMS problem
> with the external monitor?
>
> Could you try setting:
> > Option "UseBIOS" "off"
> in your xorg.conf, Section "Device"?
>
> You could also try to build the current development version from sources.
> To check out a copy from the git repository:
> > git clone git://
> anongit.freedesktop.org/xorg/driver/xf86-video-siliconmotion
>
> --
> Screen blanks and freezes after a few minutes of idling, possibly due to
> broken DPMS
> https://bugs.launchpad.net/bugs/136853
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in xf86-video-siliconmotion: Confirmed
> Status in “xserver-xorg-video-siliconmotion” source package in Ubuntu:
> Confirmed
>
> Bug description:
> Since switching to Gutsy to verify whether the packages I maintain still
> operate as expected, I have noticed that after a few minutes of idle
> operation, the screen goes blank and cannot be returned to normal. The only
> way to get screen content again is to reboot, which is possible either by
> pressing the ACPI-enabled power button or a remote SSH connection. Broken
> DPMS is suspected, although I have no way of confirming this.
>
> Purposely downgrading xserver-xorg-core and anything that depend on it,
> then pining it to that version, while keeping all the rest of the packages
> to Gutsy versions solves it, which suggests something related to X.org
> upgrade from 1.2 to 1.3 might be at play.
>
> Note: someone on #ltsp mentions a similar situation using
> xserver-xorg-video-amd, yet the exact same AMD driver doesn't exhibit this
> behavior when compiled and used with earlier X releases.
>

--
Daniel Durand

Tormod Volden (tormodvolden) wrote :

I have made xserver-xorg-video-siliconmotion packages for Jaunty which include the ``Stop clearing of "VESA compliance power down mode" bit'' patch from git, in my PPA https://launchpad.net/~tormodvolden/+archive/ppa

Please test them as soon as possible and we might be able to get this into Jaunty if it helps.

Francisco Jerez (currojerez) wrote :

I think it would still be interesting to test the version from git, as DPMS is handled somewhat differently there.

Tormod Volden (tormodvolden) wrote :

Francisco, first of all, thanks for maintaining this driver! I am looking for a minimal fix for Jaunty, since it is already in Beta it will be difficult to get a new git snapshot through, unless it is really needed. Which other commits since 1.7.0 might be needed? 0b34c4a29fcf45ac24361b4691cca0fd99cc3a3f ? If you release a 1.7.1 "stable bug fix release" maybe it would be easier to get it in ;)

Anyway, if we don't get test results from any of the guys here, it will have to wait for Ubuntu 9.10.

Francisco Jerez (currojerez) wrote :

I'm fine with doing a new release after sorting this out, I'm not even sure if any of those commits will do the trick.

Daniel Durand (durand-d) wrote :

2009/3/24 Francisco Jerez <email address hidden>

> I'm fine with doing a new release after sorting this out, I'm not even
> sure if any of those commits will do the trick.
>

Hi,

*I've got the ability to test my laptop on evening : i Installed the two
releases proposed and I havn't any progress....
I attach my log

--
Daniel Durand

Francisco Jerez (currojerez) wrote :

The patch I'm attaching (diff'ed against git master) should stop the monitor from being incorrectly reported as connected and also fix this annoying warning I saw in your logs:

> [ 3.482336] (WW) SMI(0): invalid frequency 157.000 MHz [freq <= 135.000 MHz]

That was preventing the driver from programming the maximum memory clock.

It's unlikely that it will fix the DPMS problem though. In fact, I doubt it is really about DPMS because 1.7.0 switches to DPMS off temporarily during mode setting and if it were broken you would never see anything at all on the screen.

To confirm this you can try:
$ xset dpms force off; sleep 1; xset dpms force on

That should turn the LCD off and then on again.

To find out what is the server doing when the screen remains black, you can try to build the driver with CFLAGS="-DSMI_DEBUG", and then invoke it with some high verbosity level, like:
# Xorg -logverbose 8

Then try to reproduce it.

BTW, has the VT switching problem mentioned above been solved on 1.7.0? I'm quite sure that the "awful slowness" problem with 32bpp was fixed on commit 1fc340dcf0abd7fa0837fef2c711aaec4fa8bcb8.

Tormod Volden (tormodvolden) wrote :

I uploaded an 1.7.0-1ubuntu1~tormod3 to my PPA which includes Francisco's patch, for those of you who are not able to build drivers from git. The package is built with CFLAGS="-DSMI_DEBUG". Please test it, as well as the xset test suggested by Francisco.

Daniel Durand (durand-d) wrote :

I've just updated jaunty_tormod.
Great, I see the screen !
I see the toolbar on top
the 2nd half of the screen is block with xdhmc marked
but the screen seems freezed and I've no mouse cursor.
But it's a great progress to have a xwindow !
I've not the time today to send a log but I'll send to morrow afternoon...
Tanks
Daniel

2009/3/30 Tormod Volden <email address hidden>

> I uploaded an 1.7.0-1ubuntu1~tormod3 to my PPA which includes
> Francisco's patch, for those of you who are not able to build drivers
> from git. The package is built with CFLAGS="-DSMI_DEBUG". Please test
> it, as well as the xset test suggested by Francisco.
>
> --
> Screen blanks and freezes after a few minutes of idling, possibly due to
> broken DPMS
> https://bugs.launchpad.net/bugs/136853
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in xf86-video-siliconmotion: Confirmed
> Status in “xserver-xorg-video-siliconmotion” source package in Ubuntu:
> Confirmed
>
> Bug description:
> Since switching to Gutsy to verify whether the packages I maintain still
> operate as expected, I have noticed that after a few minutes of idle
> operation, the screen goes blank and cannot be returned to normal. The only
> way to get screen content again is to reboot, which is possible either by
> pressing the ACPI-enabled power button or a remote SSH connection. Broken
> DPMS is suspected, although I have no way of confirming this.
>
> Purposely downgrading xserver-xorg-core and anything that depend on it,
> then pining it to that version, while keeping all the rest of the packages
> to Gutsy versions solves it, which suggests something related to X.org
> upgrade from 1.2 to 1.3 might be at play.
>
> Note: someone on #ltsp mentions a similar situation using
> xserver-xorg-video-amd, yet the exact same AMD driver doesn't exhibit this
> behavior when compiled and used with earlier X releases.
>

--
Daniel Durand

Daniel Durand (durand-d) wrote :

2009/3/30 Daniel Durand <durand.d@9online.fr>

> I've just updated jaunty_tormod.
>

I join my xorg file
--
Daniel Durand

Tormod Volden (tormodvolden) wrote :

Daniel, that log was not from Jaunty!?

Daniel Durand (durand-d) wrote :

2009/3/31 Tormod Volden <email address hidden>

> Daniel, that log was not from Jaunty!?
>
> --

Hum, little error
I hope this is good

--
Daniel Durand

Daniel Durand <durand.d@9online.fr> writes:

> I've just updated jaunty_tormod.
> Great, I see the screen !
> I see the toolbar on top
> the 2nd half of the screen is block with xdhmc marked
> but the screen seems freezed and I've no mouse cursor.
> But it's a great progress to have a xwindow !
> I've not the time today to send a log but I'll send to morrow afternoon...
> Tanks
> Daniel

I don't understand. Do you mean that the X server wasn't coming up at
all before? I thought your problem was about DPMS.

How are your symptoms then? Are they somewhat different after
upgrading to 1.7.0?

Is the screen already "freezed" when you start a plain X server with
no clients?
# Xorg -retro -logverbose 8

Don't you get an X-shaped cursor after running that? If you don't,
then there may be some problem with hardware cursor rendering. Try
adding this to Section "Device":
> Option "HWCursor" "off"

Daniel Durand (durand-d) wrote :

On Wednesday 01 April 2009 10:35:40 Francisco Jerez wrote:
>
> I don't understand. Do you mean that the X server wasn't coming up at
> all before?
yes, I'd a black screen with my IBM series and silicommotion card

> I thought your problem was about DPMS.
>
> How are your symptoms then? Are they somewhat different after
> upgrading to 1.7.0?
Yes, by now I get a screen : on the top, a toolbar with "restart - shutdown -
language - session - remote login with XDMCP"
and in the secund half-screen below : a blanck rectangle with on teh left
XDMCP
The screen is freezed and i have no cursor

--
Cordialement,

Daniel Durand

Francisco Jerez (currojerez) wrote :

Does your X server remain completely frozen? I mean, is it still responsive to client requests? Does it still respond to the keyboard? (e.g. does it die when you hit Ctrl+Alt+backspace? But note that for this to work you will probably need to set Option "DontZap" "off" in Section "ServerFlags").

In case it is really frozen and you have access to gdb, a backtrace at the place it hangs would be useful.

It might also be interesting to know if some combination of the following Section "Device" options makes any difference:
> Option "NoAccel"
> Option "UseBIOS" "true"
> Option "HWCursor" "false"
> Option "mclk" "0Hz"

2009/4/2 Francisco Jerez <email address hidden>

> Does your X server remain completely frozen?

Yes, completly.

i'll do the tests on next Saturday.
But, I believed that the new version of Xorg.conf was reinitialised at every
start ? Where put the options proposed ?

Please note that I have no screen and xwindow server with the last liveCD of
Knoppix nor with the liveCD of Puppy-linux.
But, I really like xubuntu and want to keep it on this old laptop.
Thanks
--
Daniel Durand

Daniel Durand <durand.d@9online.fr> writes:

> But, I believed that the new version of Xorg.conf was reinitialised at every
> start ? Where put the options proposed ?

Yes, xorg.conf is what I meant.

Daniel Durand (durand-d) wrote :

On Monday 30 March 2009 21:51:03 Daniel Durand wrote:
> I've just updated jaunty_tormod.
> Great, I see the screen !
> I see the toolbar on top
> the 2nd half of the screen is block with xdhmc marked
> but the screen seems freezed and I've no mouse cursor.
> But it's a great progress to have a xwindow !
> I've not the time today to send a log but I'll send to morrow afternoon...
> Tanks
> Daniel
>

Attached a screen picture showing the freeze screen
--
Cordialement,

Daniel Durand

Tormod Volden (tormodvolden) wrote :

Can anyone answer Francisco's questions in comment 34 please?

korgman (korgie) wrote :

@daniel durand, To have X non-frozen, add the option in driver EXA as already stated.
https://bugs.launchpad.net/xf86-video-siliconmotion/+bug/136853/comments/9

@Tormod Volden, yes, the system went completely frozen. No terminal, no alt+sysreq with "stock" xubuntu-9.04, it works with EXA although yesterday I installed the system for a friend don't know if it works flawless. I will report relative problems from the user.

tags: added: patch-forwarded-upstream

(In reply to comment #3)
> A patch has been prepared downstream:
>http://launchpadlibrarian.net/24327386/fix_sm712_clock_limit_and_monitor_detection.patch

I pushed that patch more than a year ago. It didn't address the problem this bug is about though, it just fixes a couple of related issues.

Bryce Harrington (bryce) on 2010-07-10
tags: added: gutsy
Changed in xf86-video-siliconmotion:
importance: Unknown → Medium
Changed in xf86-video-siliconmotion:
importance: Medium → Unknown
Changed in xf86-video-siliconmotion:
importance: Unknown → Medium

Martin-Éric Racine, thank you for reporting this bug to Ubuntu. Gutsy reached EOL on April 18th, 2009.
See this document for currently supported Ubuntu releases: https://wiki.ubuntu.com/Releases

Is this an issue on a supported release? If so, could you please execute the following command, as it will automatically gather debugging information, in a terminal:
apport-collect 136853

affects: xserver-xorg-video-siliconmotion (Ubuntu) → xorg (Ubuntu)
Changed in xorg (Ubuntu):
importance: High → Low
status: Confirmed → Incomplete
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.