Graphics Resolution- Upgrade /Blank Screen after reboot

[MAFoElffen]

Fresh UBUNTU 11.04 installation on Asus E35M1-M freezes during boot, but only if rebooted after one or more suspend-wakeup cycles. Suspend & wakeup works perfectly, including wakeup from the Chicony wireless USB keyboard. Problems begin if I shutdown or reboot system (no matter which way - GUI Shutdown or Reboot, CLI command 'reboot', CTRL+ALT+DEL or short push of POWER button).

1. connecting computer to power; system goes to standby

2. pressing POWER button; system starts and boots as expected, ending with CLI login. No problems here.

3. clicking System -> Shutdown -> Restart; system reboots as expected.

4. clicking System -> Shutdown -> Suspend; system goes to sleep as expected.

5. clicking any key on keyboard; system wakes up. No problems here, everything restarted as expected.

6. clicking System -> Shutdown -> Restart; Problem here:
after BIOS screen dismiss, no grub menu. Just empty screen with blinking cursor in the top left corner.
After cca 12 seconds cursor jumps one line down and then stays here forever.

7. pressing CTRL+ALT+DEL; system behaves exactly as described in [6].

8. inserting LiveDVD and pressing CTRL+ALT+DEL; system behaves exactly as described in [6].

9. shortly pressing POWER button; system behaves exactly as described in [6].

10. pressing RESET button or switching power completely OFF and ON; system again boots/reboots
as expected, no problems UNTIL NEXT SUSPEND-RESUME.

Eh? Not sure what is this about as I have described just one problem, haven't I ??

VERY IMPORTANT EDIT:
Sorry not being patient enough before. Now I have spent really long time in front of LCD and discovered that after about 4 minutes
of silence, a get the following message:

Reboot and select proper Boot device or Insert Boot Media in selected Boot Device and press a key.

Where this message comes from? grub or BIOS ???

If I enable beep at grub start, I hear this beep just if boot works. In case the system does not boot, I do not hear it, so the message seems to come from BIOS. But how Restart can be affected by the Suspend-Resume ?? Before I do S-R, Restart works as expected. Only after I do S-R, I can not Restart without pressing RESET. Would somebody explain what is going here ??????

I thought last night when I saw this that this problem is off-topic to Graphics Resolution... I earlier opened the door on this on diagnosing non-booting systems. This problem does include grub and video drivers... So I took this on, instead of referring you elsewhere to launchpad.

To make this short and simple- some of this confusion in the posts is "us" trying to translate what "we" describe to each other and get a common understanding of your problem and solution. I apologize for that confusion.

You said you did not have multiple problems, but you asked for help with sometimes "not booting on restart" (albiet in specific circumstances ) and that your Grub menu did not come up... That was the 2 problems that you stressed.

The grub menu- I saw hardcode forcing your Grun menu to display- as a good step to help diagnose your main "chief complaint" as setting a milestone or check-mark- to see "where" your PC is locking up. And if the PC was locking up after grub, giving you options.

So when it locks up now, you just confirmed where it locks up and what is affected.
- Normally restart works fine.
- After a system wakeup, if you choose to restart, it shuts down and upon the restarts power up, it hangs. The BIOS boot messages display, but before any Grub Boot, your BIOS displays an error message, saying:

Reboot and select proper Boot device or Insert Boot Media in selected Boot Device and press a key.

Which means that the BIOS can no longer see your Hard disks///

What is going on? Well, from what you describe, there is memory corruption that affects the BIOS... Have you ever heard someone trying to help (A lot of time the tech support of a hardware vendor) say:

If a laptop--> Shutdown your system. Pull out your battery. Wait 10 minutes, Put Battery back in, Power up.

If a desktop--> Shutdown system. Shut off your PSU power switch. Unplug from wall. Turn off your monitor. Unplug from wall. Unplug your video cable. Wait 10 minutes. Reconnect everything. Power up. Start,

Why and What do these things do... The why is that there may be memory corruption and even though the pc is powered down that problem is resident in a cache somewhere. These "solutions try to remove power to places where that problem might try to stay alive.

Restart can be affected by the Suspend-Resume ??

Yes. See explanation above.

I test dev branches of operation systems. I sometimes have this same problem. There are reasons for this problem. Usually this problem is either an OS problem on wakeup, an APCI setting, a Video driver problem...

So there is where you are at now.

You have an onboard AMD Radeon HD 6310 GPU --> What driver are you using?

Please post your /var/log/Xorg.0.log (Please insert it into code tags during post.)

I don't feel that a system log is going to help here as it is carrying a state from shutdown without causing a detectable error, then locking up in the BIOS POST. (Before Grub and much before Linux.)

Sidenote- A LiveCD does not use Grub to boot. It uses Casper / Isolinux. You will not ever get to a Grub Menu on a standard LiveCD. You can get to the kernel boot parameters of the Advanced Casper Menu through <F6> or through <F1> while in that menu. That is described in the first-half of post 3...

[vaton4]

I thought last night when I saw this that this problem is off-topic to Graphics Resolution... I earlier opened the door on this on diagnosing non-booting systems. This problem does include grub and video drivers... So I took this on, instead of referring you elsewhere to launchpad.

Really gratefull for your support. I am with computers since 1966 so I have experienced a lot of such obscure cases in my life. From general point of view, this is not so serious problem - important is that suspend-resume works well. I am quite used to press RESET button if sometimes something gets wrong. I am just confused what is mechanism of this problem. In those old heroic days we were used to understand things from components to operating system. Today I am just playing with configuration files - without really knowing what I am doing.

OK, back to problem.

The BIOS can no longer see your Hard disks

Ya, its clear, but also most confusing. I do not know boot sequence in details, but expecting the BIOS knows from where to boot every time, not just after powered on or reset by button.

... there is memory corruption that affects the BIOS

Any case no. BIOS checks Flash itself and RAM is thoroughly tested. Absolutely no problems with system for days. And this curious behaviour is 100% reproducible - no exceptions.

You have an onboard AMD Radeon HD 6310 GPU --> What driver are you using?

latest fglrx from Ubuntu 11.04 repo, of course.

Please post your /var/log/Xorg.0.log

sorry but I can not do that as today I reinstalled system and replaced Ubuntu by Fedora 14. This verified the problem is not bound to Ubuntu as the same things happen in Fedora environment.

I don't feel that a system log is going to help here ...

Do agree. BMHO most probably it comes or from BIOS, or from grub, or from both (interaction?)

I have noticed similar problems with Windows (see Google: E350 Reboot and select proper Boot device), but they seem to be bound more likely to HD or CD drive problems. I did not ignore this, of course, and tested system with only HDD connected (Hitachi 2.5" SATA) and toggled some BIOS settings. Nothing changed.

I will try to check suspend-resume-restart using Live Ubuntu CD as you suggest and maybe downgrade BIOS.
And there is also another interesting place to investigate - CMOS RAM, the place where BIOS saves its secrets
Do you (or anybody who reads this) know whether system writes something to CMOS before goes to sleep after SUSPEND??
Or sends something to BIOS any other way?? Remember, things are more complex as we are fighting with EFI BIOS here!

EDIT:
Made giant step forward. "Discovered" pm-utils, so now I can check things in CLI. As expected, pm-suspend command somehow suspends system, but resume fails. Now experimenting with "quirks". Maybe it is time to return back to Ubuntu and read the pm documentation ?

[MAFoElffen]

Really gratefull for your support. I am with computers since 1966 so I have experienced a lot of such obscure cases in my life. From general point of view, this is not so serious problem - important is that suspend-resume works well. I am quite used to press RESET button if sometimes something gets wrong. I am just confused what is mechanism of this problem. In those old heroic days we were used to understand things from components to operating system. Today I am just playing with configuration files - without really knowing what I am doing.

OK, back to problem.

Ya, its clear, but also most confusing. I do not know boot sequence in details, but expecting the BIOS knows from where to boot every time, not just after powered on or reset by button.

Any case no. BIOS checks Flash itself and RAM is thoroughly tested. Absolutely no problems with system for days. And this curious behaviour is 100% reproducible - no exceptions.

latest fglrx from Ubuntu 11.04 repo, of course.

sorry but I can not do that as today I reinstalled system and replaced Ubuntu by Fedora 14. This verified the problem is not bound to Ubuntu as the same things happen in Fedora environment.

Do agree. BMHO most probably it comes or from BIOS, or from grub, or from both (interaction?)

I have noticed similar problems with Windows (see Google: E350 Reboot and select proper Boot device), but they seem to be bound more likely to HD or CD drive problems. I did not ignore this, of course, and tested system with only HDD connected (Hitachi 2.5" SATA) and toggled some BIOS settings. Nothing changed.

I will try to check suspend-resume-restart using Live Ubuntu CD as you suggest and maybe downgrade BIOS.
And there is also another interesting place to investigate - CMOS RAM, the place where BIOS saves its secrets
Do you (or anybody who reads this) know whether system writes something to CMOS before goes to sleep after SUSPEND??
Or sends something to BIOS any other way?? Remember, things are more complex as we are fighting with EFI BIOS here!

EDIT:
Made giant step forward. "Discovered" pm-utils, so now I can check things in CLI. As expected, pm-suspend command somehow suspends system, but resume fails. Now experimenting with "quirks". Maybe it is time to return back to Ubuntu and read the pm documentation ?

Try "acpt_osi=linux" as a kernel boot option.

The suspend-resume function in natty is done with HAL. It uses apm and acpi to try to handle the function. I know that HAL started getting a little quirky in natty. I know there is another package that does this via a software solution and they use to use it as an alternative...

The idea is install a tool called "uswsusp""

sudo apt-get install uswsusp

And by typing the below command you check if the suspend function works….

sudo s2ram

Same goes for hibernation

sudo s2disk

Now once this is done and all of the above commands work, they can be replaced with the old non-suspect commands that come with ubuntu.

But attention- before editing the system files, always make sure you back up the files in case something goes completely wrong.

sudo cp /usr/lib/hal/scripts/linux/hal-system-power-suspend-linux /usr/lib/hal/scripts/linux/hal-system-power-suspend-linux.bak
sudo cp /usr/lib/hal/scripts/linux/hal-system-power-hibernate-linux  /usr/lib/hal/scripts/linux/hal-system-power-hibernate-linux.bak

This step is to replace the old commands with the new commands in
hal-system-power-suspend-linux

gksudo gedit /usr/lib/hal/scripts/linux/hal-system-power-suspend-linux

Here is the original file...

#!/bin/sh

. hal-functions

hal_check_priv org.freedesktop.hal.power-management.suspend
hal_exec_backend

Edit it to look like this

#!/bin/sh

/sbin/s2ram –force
hal-system-power-hibernate-linux

Save and exit. Then open hal-system-power-hibernate-linux

gksudo gedit /usr/lib/hal/scripts/linux/hal-system-power-hibernate-linux

Here is the original file

#!/bin/sh

. hal-functions

hal_check_priv org.freedesktop.hal.power-management.hibernate
hal_exec_backend

Edit it to look like this

#!/bin/sh

/sbin/s2disk

See if your hardware handles differently with that... and hopefully in a good way! If not, then you could go back by restoring the backup files

sudo cp /usr/lib/hal/scripts/linux/hal-system-power-suspend-linux.bak /usr/lib/hal/scripts/linux/hal-system-power-suspend-linux
sudo cp /usr/lib/hal/scripts/linux/hal-system-power-hibernate-linux.bak   /usr/lib/hal/scripts/linux/hal-system-power-hibernate-linux

[vaton4]

Try "acpt_osi=linux" as a kernel boot option

No change.

The idea is install a tool called "uswsusp"

Done.
Command s2ram without --force refuses to work because of unknown hardware (I have followed URL from error message and found out they do not support new hardware any more). Command s2ram --force suspends, but if I press key on keyboard, just some hardware wakes up (fans spinning, for example) but no video output - monitor says "No Signal" (HDMI) and nothing happens until I reset or "re-power" system. I wasn't able to find any combination of arguments which will convince the system to resume correctly.

Now once this is done and all of the above commands work ...

seems to be unreachable dream...

Found https://bugs.launchpad.net/archlinux/+bug/767975 where evidently the same problem in ArchLinux. Seems I have baught the E350 mainboard too early.

Many thanks again, MAFoElffen. It really was not useless, as I have learned something new, but I do not feel so strong to solve this for comunity. I will wait instead, until somebody succeeds. Bay.

[MAFoElffen]

No change.

Done. I wasn't able to find any combination of arguments which will convince the system to resume correctly.
seems to be unreachable dream...

Found https://bugs.launchpad.net/archlinux/+bug/767975 where evidently the same problem in ArchLinux. Seems I have baught the E350 mainboard too early.

Many thanks again, MAFoElffen. It really was not useless, as I have learned something new, but I do not feel so strong to solve this for comunity. I will wait instead, until somebody succeeds. Bay.

I'm thinking that also. I looked at the bug report you found, but I'm thinking closer to yours was this one:
https://bugs.launchpad.net/ubuntu/+s...ux/+bug/301353

I toyed with apm and apci boot switches, but didn't see any that would help a disk or controller in those circumstances. I think as ID'ed in that Bug report, that it may be an issue with the BIOS. I would encourage starting a bug report... It may or not fix yours, but it will get it logged.

I alos see that Oneiric (11.10) did fix a lot of apm and apci issues... Releases today. You might try dist-upgrade and see if that issue is still there.

[bobwdn]

An unanswered question.

I am a little confused. Jockey has, and still is broken, as I understand.

If I am mis-understanding, I am sorry, please correct me.

What we are doing here is installing Nvidia drivers (in my case) manually. After they (Nvidia) are installed, the Nvidia splash screen displays on restart. That tells me the the driver 'is' installed. However, when accessing Jockey, Jockey still claims the driver is inactive and yet Jockey is not displaying the 'additional drivers are available' warning that typically appears in our upper right desktop.

So, do I ignore the 'activate button' that is within Jockey?

OR

Do I active the (apparently) functional driver via Jockey GUI?

[MAFoElffen]

An unanswered question.

I am a little confused. Jockey has, and still is broken, as I understand.

If I am mis-understanding, I am sorry, please correct me.

What we are doing here is installing Nvidia drivers (in my case) manually. After they (Nvidia) are installed, the Nvidia splash screen displays on restart. That tells me the the driver 'is' installed. However, when accessing Jockey, Jockey still claims the driver is inactive and yet Jockey is not displaying the 'additional drivers are available' warning that typically appears in our upper right desktop.

So, do I ignore the 'activate button' that is within Jockey?

OR

Do I active the (apparently) functional driver via Jockey GUI?

Sort of and sort of not... My apologies if you misunderstood... Let me explain a little differently.

Look at the attached picture. See circled in the lower right. The remove button says that nivia-current, the driver highlighted above, is installed.

Look at circled in the lower left. The green highlighted radio button and the phrase "This driver is activated" is correct/okay and indicates that the driver is installed and active. The rest of the phrase "but not currently in use." is incorrect and very misleading to users... Leading them to mistakenly think that their driver in not working, not selcted and not is use- when it "is."

Now look at the same area in the second picture. I've highlighted a driver that is not installed. Looking in the lower left, you can see that the radio button is not selected and the text says "The driver is not activated." If you look in the lower right, the button says "Activate."

The functionality of Jockey, works. The reporting of the installed driver being active is a little obscure, so users and Ubuntu consider that part of Jockey being broke and the confusion it is causing users a problem.

About installing NVidia binary drivers manually? Look at Post 280 of this thread: HERE

Depending on time since being released, these are basically the same drivers in different installation format. For instance today--> nvidia-current is 280.13 whereas the newest at NVidia is 285.05... nvidia current will be 285.o5 in Ubuntu after it gets tested in edgers and xswat to be considered by them as stable enough for common users. (not to say it isn't / just not proven yet...)

The drivers installed through Additional Drivers (Jockey) are Ubuntu targeted debian packages. built with the vendor binaries, sometimes with a little extra to correct issues that binary might have with Ubuntu. They are meant as an automated install. Upgrades to the driver are also automated.

The Nvidia (and ATI) binaries and a manual install, meant to work on Linux... Versions don't always work without error with all distro's of Linux, so be aware they you have to be aware o that. For instance ATI's driver v11.9 doesn't work with Unity but v11.8 did... Nvidia 180,13 had issues with Unity on older NVidia GPU's w/ older Hardware, but 285.05 has that corrected... Upgrades for vendor binaries are also a manual affair.

What is better? It all boils down to personal preference. What do I recommend to people? That changes every day, based on the individual I'm recommending to, their skill set and experience, the hardware they have, what they are trying to do and if they have issues what they are trying to correct. That is a moving and ever changing target.

The "easiest" for the common user without any uncommon issues is to install the drivers from Jockey.

[Yxven]

I'm not having any luck with my black screen issue.

11.04 works, but a fresh install of 11.10 or upgrading from 11.04 to 11.10 makes me boot into a black screen before the monitor kicking into power save mode.

Booting with nomodeset allows me to see the kubuntu loading screen, and then it loads some stuff before stopping. It usually stops at pulseaudio, but the amount of other stuff it loads and the order seems to change. The only thing that consistently says fail is "Stopping automatic crash report generation". I can then ctrl-alt-f1 into a shell.

I have a AMD A6-3650 processor with integrated AMD Radeon HD 6530D for graphics and a ASUS F1A75-M motherboard.

I uninstalled flgrlx as per the amd instructions with no luck.
I tried installing the catalyst amd driver with no luck.
I currently have a driver installed this way:
https://wiki.ubuntu.com/X/Troublesho...i_from_scratch


I've tried these options:

 
nomodeset  
xforcevesa  
i915.modeset=0 xforcevesa  
vga=771
vga=791
vga=792
vga=829
acpi_osi="Linux"
acpi=off

sudo hwinfo --framebuffer says:

> hal.1: read hal dataprocess 1680: arguments to dbus_move_error() were incorrect, assertion "(dest) == NULL || !dbus_error_is_set ((dest))" failed in file ../../dbus/dbus-errors.c line 280.
This is normally a bug in some application using the D-Bus library.
libhal.c 3483 : Error unsubscribing to signals, error=The name org.freedesktop.Hal was not provided by any .service files
02: None 00.0: 11001 VESA Framebuffer
  [Created at bios.464]
  Unique ID: rdCR.kBwBDUNBOLF
  Hardware Class: framebuffer
  Model: "(C) 1988-2010, AMD Technologies Inc.  SUMO"
  Vendor: "(C) 1988-2010, AMD Technologies Inc. "
  Device: "SUMO"
  SubVendor: "AMD ATOMBIOS"
  SubDevice:
  Revision: "01.00"
  Memory Size: 16 MB
  Memory Range: 0xc0000000-0xc0ffffff (rw)
  Mode 0x0300: 640x400 (+640), 8 bits
  Mode 0x0301: 640x480 (+640), 8 bits
  Mode 0x0303: 800x600 (+832), 8 bits
  Mode 0x0305: 1024x768 (+1024), 8 bits
  Mode 0x0307: 1280x1024 (+1280), 8 bits
  Mode 0x0310: 640x480 (+1280), 15 bits
  Mode 0x0311: 640x480 (+1280), 16 bits
  Mode 0x0313: 800x600 (+1600), 15 bits
  Mode 0x0314: 800x600 (+1600), 16 bits
  Mode 0x0316: 1024x768 (+2048), 15 bits
  Mode 0x0317: 1024x768 (+2048), 16 bits
  Mode 0x0319: 1280x1024 (+2560), 15 bits
  Mode 0x031a: 1280x1024 (+2560), 16 bits
  Mode 0x030d: 320x200 (+640), 15 bits
  Mode 0x030e: 320x200 (+640), 16 bits
  Mode 0x0320: 320x200 (+1280), 24 bits
  Mode 0x0393: 320x240 (+320), 8 bits
  Mode 0x0395: 320x240 (+640), 16 bits
  Mode 0x0396: 320x240 (+1280), 24 bits
  Mode 0x03b3: 512x384 (+512), 8 bits
  Mode 0x03b5: 512x384 (+1024), 16 bits
  Mode 0x03b6: 512x384 (+2048), 24 bits
  Mode 0x03c3: 640x350 (+640), 8 bits
  Mode 0x03c5: 640x350 (+1280), 16 bits
  Mode 0x03c6: 640x350 (+2560), 24 bits
  Mode 0x0333: 720x400 (+768), 8 bits
  Mode 0x0335: 720x400 (+1472), 16 bits
  Mode 0x0336: 720x400 (+2944), 24 bits
  Mode 0x0353: 1152x864 (+1152), 8 bits
  Mode 0x0355: 1152x864 (+2304), 16 bits
  Mode 0x0356: 1152x864 (+4608), 24 bits
  Mode 0x0363: 1280x960 (+1280), 8 bits
  Mode 0x0365: 1280x960 (+2560), 16 bits
  Mode 0x0366: 1280x960 (+5120), 24 bits
  Mode 0x0321: 640x480 (+2560), 24 bits
  Mode 0x0322: 800x600 (+3200), 24 bits
  Mode 0x0323: 1024x768 (+4096), 24 bits
  Mode 0x0324: 1280x1024 (+5120), 24 bits
  Mode 0x0343: 1400x1050 (+1408), 8 bits
  Mode 0x0345: 1400x1050 (+2816), 16 bits
  Mode 0x0346: 1400x1050 (+5632), 24 bits
  Mode 0x0373: 1600x1200 (+1600), 8 bits
  Mode 0x0375: 1600x1200 (+3200), 16 bits
  Mode 0x0376: 1600x1200 (+6400), 24 bits
  Mode 0x0383: 1792x1344 (+1792), 8 bits
  Mode 0x0385: 1792x1344 (+3584), 16 bits
  Mode 0x0386: 1792x1344 (+7168), 24 bits
  Mode 0x03d3: 1856x1392 (+1856), 8 bits
  Mode 0x03d5: 1856x1392 (+3712), 16 bits
  Mode 0x03d6: 1856x1392 (+7424), 24 bits
  Mode 0x03e3: 1920x1440 (+1920), 8 bits
  Mode 0x03e5: 1920x1440 (+3840), 16 bits
  Mode 0x03e6: 1920x1440 (+7680), 24 bits
  Config Status: cfg=new, avail=yes, need=no, active=unknown

[MAFoElffen]

I'm not having any luck with my black screen issue.

11.04 works, but a fresh install of 11.10 or upgrading from 11.04 to 11.10 makes me boot into a black screen before the monitor kicking into power save mode.

Booting with nomodeset allows me to see the kubuntu loading screen, and then it loads some stuff before stopping. It usually stops at pulseaudio, but the amount of other stuff it loads and the order seems to change. The only thing that consistently says fail is "Stopping automatic crash report generation". I can then ctrl-alt-f1 into a shell.

I have a AMD A6-3650 processor with integrated AMD Radeon HD 6530D for graphics and a ASUS F1A75-M motherboard.

I uninstalled flgrlx as per the amd instructions with no luck.
I tried installing the catalyst amd driver with no luck.
I currently have a driver installed this way:
https://wiki.ubuntu.com/X/Troublesho...i_from_scratch

I've tried these options:

 
nomodeset  
xforcevesa  
i915.modeset=0 xforcevesa  
vga=771
vga=791
vga=792
vga=829
acpi_osi="Linux"
acpi=off

sudo hwinfo --framebuffer says:

Two observations:

1. "nomodeset" is a common kernel option for nvidia chipsets. The kernel boot option equivalent to that for a Radeon chipset is "radeon.modeset=0"

2. The install instructions you used are over a year old... The instructions I have here: Post #334 are updated for 11.04 and 11.10 and have all the latest work-arounds.

The one caution there is use the Catalyst v11.8 driver instead of v11.9. 11.9 has problems with Ubuntu. From the link in my instructions, which gets to the current Catalyst driver > Look below the driver and Select "Previous Drivers and Software" > Select "Catalyst 11.8"

[Lizella]

Good Evening. First of all, thank you for this thread which has shed light on the problem I am having. I can see it is a common problem shared by many. But even after reading everything here, I can't solve this graphics problem. I have been using Ubuntu since 8.04 and have had nary a problem (or one I couldn't easily fix) therefore have never joined this forum. Is the only solution after all to install a new type of graphics card? Or can this be solved with some tweaking? I am not a savvy user at the tech level. But here is what I can tell you. Attempting to tryout 11.04, and now 11.10 on 10.10 system with correctly burned live CD. To see if I like it and to see if it works. CD won't boot. Purple screen, then black. Same as others here. Yes, I can find Grub and the kernel boots. nomodeset doesn't work. I have Nvidia GeForce 7300 SE. Installed June 2008. Driver Nvidia current 260.19.06
Is this card simply incompatible with the new Ubuntus? If I updated from update manager instead of from CD, would I have graphics problems? (many people do, so I intuit that I probably would) I would like to know before I go out and get a new graphics card. And in case I need a new one, could someone kindly tell me one that is sure to work?
Thanks a lot. Your hard work is appreciated and ever so informative.
Liz