wm-logout or :Quit end into nirwana

[jsfarinet]

I created a new user (btw, i gave a time limit for its existence, since it's only for testing purposes: Does that mean it will be automatically deleted or will i be prompted for?). I logged in as normally, then i switched (using "Switch-User" (=gdmflexiserver) ) user and logged in as the "test user". From there i tried the several logouts.

Here the results:

1) Exit (=wm-logout) > logout works; brings me back to the first logged in user wherefrom i switched to the new one
2) Ctrl-Alt-Del works exactly the same way.

That seems to me correct. So, i think the first conclusion is: The problems with 'wm-logout' and (Fluxbox internal command) ':Quit' has nothing to do with the fglrx (amd catalyst) video driver (?)

But i notice: when i login first as the "test user" then switch to my normal user, from there, i.e. the normal/standard user, i can use wm-logout > logout and Ctrl-Alt-Del and it brings me to the first (logged-in) user. But if the first logged-in user (in this case the "test-user") does the same commands they do not bring to the console. I see the console text appearing (the logout msgs on the screen) but then, instead of the new (console) login prompt or the gdm login window (i'm not sure which of the both should appear here?) i end on a "dead screen" with the blinking underline cursor on top left.

[gapan]

That's nice, but it's not what I asked for.

[jsfarinet]

Essentially, answering to your 2 questions:

1) Logout hangs with the new user too (if it is started as the first user).
2) I'm afraid about uninstalling fglrx because i remember the hassle it was in debian (got no X anymore and had to re-install X completely new). But, there's no other way to figure out, isn't it?

(If so, sadly, i'll have to try ... ).

[jsfarinet]

Ok, i did it: I uninstalled the amd catalyst driver (aka: fglrx). And yes, now wm-logout > logout as well as CTRL-ALT-DEL work. They bring me back to the gdm-login screen.

Problem remains here:
- No suspend is working (pm-suspend brings the computer to suspend but the screen does not awake anymore)
- Computer is overheating.

In the boot-up i see a msg saying: "No UMS support in radeon kernel module". Does this say something to you?

Essentially, i'd prefer to stick with the fglrx driver - if it is possible to figure out the logout problem.

Thanks a lot for any pointer.

[jsfarinet]

Answering to myself: There was nomodeset in lilo which caused the UMS statement. That's solved.

For the overheating problem i found the advice to set (in grub but as you stated it's the same for lilo) "radeon.dpm=1" (in append). Doing so, i get the msg `1' is an invalid value for dpm. Why that?

For the black screen problem after suspend (it's really only the screen that's dead/black, X is still active: When i do Ctrl-Alt-Del i hear the welcome sound of gdm (which i set) i tried this: https://wiki.archlinux.org/index.php/AT ... spend_in_X. But this made it impossible to start X.

[jsfarinet]

. . .
For the black screen problem after suspend (it's really only the screen that's dead/black, X is still active: When i do Ctrl-Alt-Del i hear the welcome sound of gdm (which i set) i tried this: https://wiki.archlinux.org/index.php/AT ... spend_in_X. But this made it impossible to start X.

Seems i found a solution: Now, after resume from suspend the screen effectively is coming back after i edited the lilo append line with following parameters:

Code: Select all

append="quiet vt.default_utf8=1 rootfstype=ext4 radeon.dpm=1 acpi_osi=linux acpi_backlight=vendor acpi_sleep=s3_mode resume=/dev/sda3"

If now, if there will be a way to control the overheating problem the open source driver would be fine ("radeon.dpm=?")?

Moreover, there's still pending the problem, how and where to configure close_lid so that it does pm-suspend.

[jsfarinet]

I installed the kernel 3.14 and indeed that seems to solve the error statement about "radeon.dpm=1". And, as a welcome side effect, system temperature significantly went down (from constantly >80°to between 50° - 65° (more or less the same as with the fglrx driver).

There is one annoying instance though: If the laptop is going really hot and i try to reboot - as i did after the installation of the newer kernel - then the boot process does not go thru. It does not even come to build the log files (in /var/log); the screen is like exploding into a white "flickering" cloud. When i wait that the laptop cooled down, boot is fine. The earlier kernel (3.10) did not have this problem (i experienced exactly the same with debian with any kernel > 3.10 or 3.11 ,i do not remember exactly). In my simple mind, i think that might be due to the fact the earlier kernel is not temperature sensitive but the newer are (????).

In any case, actually things seem relatively fine but i've a question: Is there needed any further tweaking? I found this thread in askubuntu which pretty much reflects my problems: https://askubuntu.com/questions/403965/ ... eon-driver.

Later on, i'll try suspend and reboot to see if it works from this kernel and the generally lower temperatures. If so, the last step would be to get close_lid working. Hope someone can help me out here.

In the boot process i see the following msg:

Code: Select all

Warning: Configuration /etc/laptop-mode.con.d/board-specific/*conf ist not readable, skipping
Laptop mode
enabled, not active
   OK

But there is no '/etc/laptop-mode.conf.d' but only '/etc/laptop-mode.conf/' ??

[jsfarinet]

. . . Later on, i'll try suspend and reboot to see if it works from this kernel and the generally lower temperatures. If so, the last step would be to get close_lid working. . . .

Negative: Suspend (or better: RESUME from suspend) returned to *NOT* work (with kernel 3.14)
Positive: Staying relatively cool with kernel 3.14 reboot works.
Positive: Logout, Ctrl-Alt-Del and (within fluxbox) ":Quit" work.

Clueless at this point about suspend ...

[jsfarinet]

Alternatively, there is a new amd-catalyst package on their site. I installed that to test. Now, it seems to work all, also logout/Ctrl-Alt-Del.

But, may be that's also, because i tweaked lilo.conf accordingly to what is said on ArchWiki about fglrx (needs nomodeset). I do not know, wether the tweaking of lilo or the new driver or may be both are the reason, but the initial problem seems to be considered solved.

For the free radeon driver still remains the resume from suspend problem. I looked around a lot, there seem to be many similar problems but frankly out of my limited knowledge i did not find any viable solution (for me). Generally, with the newer kernel it seems to work fine, overheating is definitely solved. Safe the problem above, which is a big one for a laptop though.

[gapan]

You seem to have come a long way! These problems are hardware specific and it's difficult for anyone without the exact same hardware to help. It is possible that an even newer kernel could have solved the open source drivers issue. Try compiling one, it really isn't that hard. There is a howto in the wiki.

Do I understand correctly that you have found a way to have everything working?