Talk:Problem with display remaining black after resume

From ThinkWiki
Revision as of 07:56, 6 May 2006 by Dwilkins (Talk | contribs)
Jump to: navigation, search

T43 with Intel Graphics Media Accelerator

My T43 (1871-FYG) with Intel 915GM is also affected by the problem and the 'ATI' sollution with acpi_sleep=s3_bios works. Please consider updating the page (I don't dare to do it myself).

From my experience, this does not work with TP R50e.

If you want to be able to resume, you need :

  • To start from a VT
  • No option acpi_sleep
  • to save the PCI state like:
cat /proc/bus/pci/00/02.0 > /var/cache/video.config
  • Once you get back, restore it:
cat /var/cache/video.config > /proc/bus/pci/00/02.0

However, the X screen is not clean after the restore, so this is not very useable...

The comment on R50e

There is a comment at the end of the page on R50e, saying that you should switch to console first, and look at a page to learn how to do this.

However. it seems to me that the given script already does it. Should the comment be removed?

The scripts are different. However, if you can approve that the solution provided on this page resolves the problem, please remove the footnote.

Wyrfel 14:46, 17 Oct 2005 (CEST)

On my R50e, the script on this page works fine, while the one linked in the footnote quickly wakes up from sleep.

Thisnukes4u 19:39, 1 Jan 2006 (CET)

Problems with R51 and Intel Graphics

The problems described on the page occur intermittently with an Intel Graphics card on an R51. Here is what happens:

1. One some occassions everything seems to work fine.

2. On other occassions the screen looks OK but the some plane seems to have vanished. New text on the screen overwrites instead of replaces prior text. The "standard" pattern disappears from the X background.

3. On yet other occassions the X server dies and comes out "blank" in the mode (2) above.

4. The above problems appear only with the "i810" driver. The "vesa" driver works fine. Moreover, if the machine is "suspend-to-disk"ed then the problem disappears on resumption once the X server is killed and re-started.

But for 1 and 4 above I would have given up and either avoided suspend to ram altogether or used vesa mode---as it stands the problem is tantalising.

Kapil kapil at imsc dot res dot in

X40 with an Intel 855GM Integrated Graphics Device

I used the acpi_sleep=s3_bios boot parameter and the following ACPI script:

set -e

chvt 12
echo mem >/sys/power/state
vbetool post
chvt $vt

The acpi_sleep=s3_bios boot parameter alone did not fix things, and neither did saving and restoring the information from /proc/bus/pci/00/02.0.

X41 with an Intel 915GM Integrated Graphics Device

It works with the acpi_sleep=s3_bios,s3_mode boot parameter. Resume works great, both on the VESA console and in X.

Interesting--that didn't work for me. I needed to do several things to get it to work. First, I used s3_bios, but not s3_mode. Secondly, I had to edit resume_video() in /etc/pm/functions-intel to comment out the VBE post and restore. Those two changes made it functional, but the improper lid state in the HAL daemon caused it to go back to sleep just as soon as it resumed upon the lid opening, so I added a HAL restart to the front end of /etc/acpi/actions/ This means that I have to restart gnome-power-manager, too. I haven't yet figured out a way to do that in; currently I do it manually. I'll post this to the main page soon, unless I hear comments to the contrary. BrianTung 19:52, 30 April 2006 (CEST) (fixed a little)

Z60t with an Intel 915GM Integrated Graphics Device (PCI Express)

It works with the acpi_sleep=s3_bios boot parameter. Here is the script I use (gentoo). Note: stopping and starting WiFi is not necessary to resume, but it is dead on resume until restarted and the module is reloaded. This is on a kernel with framebuffer console support. The blinking led thing is a nice touch I got from the hibernate-scripts package - it blinks the sleep led (moon icon) until the sleep or resume cycle is finished.

The 'vbetool vbemode get/set' commands are for saving and restoring the console text mode - otherwise the screen becomes garbled.

 # blink sleep led (if ibm_acpi is installed)
 echo 7 blink > /proc/acpi/ibm/led
 # change to console 10 (unused?)
 chvt 10
 # save video state
 VBEMODE=`vbetool vbemode get`
 cat /proc/bus/pci/00/02.0 > /tmp/video_state
 # sync filesystem
 # sync hardware clock with system time
 hwclock --systohc
 # stop networking (atheros chipset)
 /etc/init.d/net.ath0 stop
 rmmod ath_pci
 # go to sleep
 sleep 2 ; echo -n 3 > /proc/acpi/sleep
 # blink sleep led
 echo 7 blink > /proc/acpi/ibm/led
 # restore system clock
 hwclock --hctosys
 # restore video state
 vbetool vbemode set $VBEMODE
 cat /tmp/video_state > /proc/bus/pci/00/02.0
 # change back to X
 # restart networking
 modprobe ath_pci
 /etc/init.d/net.ath0 start
 # clean up behind us
 rm /tmp/video_state
 echo 7 off > /proc/acpi/ibm/led

Fedora Core 5

The vbetool on the Talk page worked great for me, but NetworkManager would not bring the wireless interface back up. I had to add "ath_pci" to the SUSPEND_MODULES variable in /etc/sysconfig/pm.