Bug 121965

I would at least delay that warning, btw. powersaved is started after network, 
so if your network setup takes a long time (i.e. no network available), you 
will get that warning even though powersaved will start soon after. 
 
So just gray out the kpowersave icon and ignore the rest 

Andi, just disable kpowersave for now (quit => do not start on login).
Coolo's suggestion is good for the next kpowersave versions.

@coolo: I would continue to open a errormessage by default (and additionally 
grey out the icon), because the user need to know if powersave is not running. 

If you have a laptop running on battery and powersave dies IMO you need this 
information to react, befor you machine dies hard because of faster empty 
battery.

I see to ways:

1.) an other errormessage, which could contain a checkbox to block popup again 
the message and a configure option in the kpowersave configure dialog to change 
this behavior

2.) we discussed a additional functionality for kpowersave where you can define 
a action/command for (predefined) events. We could integrate there the behavior 
from point one. 

I would prefer the second solution.

Your opinion?
  

No there are really good reasons to sometimes not run it (e.g. buggy BIOS). 
And in these cases KDE really should not try to second guess the user. 
If it doesn't run just accept it. IMHO it is a case here of KDE messing 
with things it shouldn't mess with. 
 
You also don't check for a zillion other cases. 
 

There are no 'zillion other cases'. KPowersave depends on powersave and if 
powersave daemon is not running we need to inform the user about. 

If you would't display this message, you can disable as descriped above (if 
implemented) or you should stop KPowersave. If your machine has a buggy bios and 
you disable/stop powersaved you make this aware. In this case you also can 
disable the message or need to stop kpowersave. But if there is a other reason 
(e.g. powersave die ...) for missing connection to powersave the user need to be 
inform on default.

Sure, there are lots of other services that can be missing and KDE 
mostly handles not having them silently. 
 
Powersaved normally does not die. When it dies it is very likely 
explicit user action. KDE should not try to second guess this. 
 
 

I can't agree to this. 

We had so many cases in the past where powersave dies and it was well to know 
about to find the bugs. We also popup if dbus dies or stopped (after a timeout). 
IMO it make sence to inform the user if the basic/main funtion of the program is 
lost. 

And at least powersave is a central requirements on a laptop. If this not work 
user must know it. This is the same as if there is e.g. a problem with 
openoffice and you can't open/read and change documents anymore. In this case 
you need to display a message to user.

If you speak about 'explicit user action' what is the problem to disable the 
message or kpowersave in this case explicit. If you disable powersave you must 
live with this message at the first time.

to #1 from coolo: I will add a longer delay and some trials to connect 
powersaved on startup of kpowersave before display the warning

Why do you think every machine SUSE runs is a laptop?  That's clearly 
not the case.  If a customer decides to stop powersaved on their  
server or we recommend it to the customer to work around bugs they definitely 
don't want such a dumb message from KDE. I think you're having a bad case of 
tunnel vision here. 
 
Anyways, I don't care that much about powersaved itself, just the cpufreq 
modules.  In case you're really fixed in your "all world is a laptop"  
view another way might be to split rcpowersaved into two scripts - one 
that loads cpufreq and the other to load the daemon that does the rest of 
the stuff.   
 
Thomas what do you think? That would at least allow the KDE people to stay in 
their laptop centric world without too many bad side effects. 
 
Even with that split I would recommend to get rid of the message though. 
 

Sorry, but come down. ;-)

I know that kpowersave does not run only on laptop. But laptops are the primary 
scope of powersave and this is not "a bad case of tunnel vision". 

Note: KPowersave is not part of kde.org. This is part of the powersave project. 

Back to the facts: If you only would prevent loading cpufreq modules you need 
not to stop or remove powersave from init process. You can change the powersave 
settings in /etc/sysconfig/powersave/cpufreq (to off):

# The powersaved startscript will load this module for CPU frequency
# scaling support. If nothing is specified here, the most common cpufreq
# modules will be tried. Entering the correct module will reduce the
# warnings in syslog during boot and probably speed up booting a bit.
# Entering "off" will skip trying to load cpufreq modules and avoid
# warnings on machines with no cpufreq-capable hardware.
#
POWERSAVE_CPUFREQD_MODULE=""

You see we don't need to splitt of the start script of powersave.

Well my complaint still stands. KDE is messing with things it shouldn't. 
 
That's harder to explain to customers, but ok. 
 
I always disliked all this laptop gunk on my desktops anyways. 
How about you implement a simple check somewhere - let's say  
system has a cardbus/PCMCIA bridge - and if it fails don't run 
laptop specific toys? The KDE power tools don't make much sense 
on desktops and servers. 
 
There is also a field in ACPI FADT that describes laptop/desktop etc. 
but it's not always reliable. 
  
 

As a compromise: 
 
How about you add a button "Never show this message again" to 
that powersaved warning?  That would be still ugly for the customers, 
but at least not really ugly anymore. 
 

The problem to detect laptop is: there is no solid method to detect if the 
machine is a laptop. We discussed the so often in the team, but currently there 
is no 100% secure way to differ between desktop and laptop. (there are also 
desktops with cardbus/PCMCIA, the smbios formfactor could also be wrong ...)

I currently work on a solution for this issue in HAL, but this is also not 100 % 
 perfect.

 

to #12: this is what i proposed in #3

FADT || cardbus/pcmcia bridge should be pretty reliable. 
 
Anyways, you see it's a reason why KDE shouldn't display such messages. 
It cannot be sure so it gives too many false positives. 
 
 

(In reply to comment #15)
> FADT || cardbus/pcmcia bridge should be pretty reliable. 

We tested FADT in the past and we had only one laptop where this was correct, 
all other where identified as deskptop (see your post on reseach and the reply 
from seife). --> no solution

cardbus/pcmcia: is also no solid solution because there are laptops without 
cardbus/pcmcia and also desktop machines with cardbus/pcmcia. 

As I sad I work on a detection in HAL and I think we need a combination of more 
then one or two sources. IMO acpi lid-button is a solid indicator for laptops. 

We need two things:

     a) Add a button to disable the "powersave daemon not running" message.

     b) A much longer delay for displaying the message if the powersave daemon 
        is not running. The problem seems to be the autologin that allows 
        KDE/kpowersave to be started before the powersave daemon came up.
        20 seconds should be long enough to start the daemon but still should 
        be enough to ensure that machine does not switch off because battery is 
        empty.

This was never discovered because most people use NIS login here, but is really anoying for a lot of people out there.
We should even think about a YOU update for this -> CC'ing behlert.

We had this problem on exactly one machine but I was not able to verify it on another. I don't think that too much people are affected, because I think it would have been already on one of our public lists (opensuse, suse-laptop, suse-linux). But this is some kind of personal assessment.

Nevertheless, a check box for the popup would be good IMO. But I don't know if it is worth for a you update if there are no other pending bugs in kpowersave. Danny?

I don't see a reason for a YOU update because of only one machine. But as i proposed in #3 I would implement a solution for 10.1 for the message and add a longer delay on startup. 

Btw. Thomas: I only use local user accounts and never NIS and this never happen here at any of the machines I used. Ans note: the machine never shutdown if the daemon is not running. This can't happen if powersave is not running ;)

(In reply to comment #19)
> I don't see a reason for a YOU update because of only one machine. But as i
> proposed in #3 I would implement a solution for 10.1 for the message and add a
> longer delay on startup.

ACK
> 
> Btw. Thomas: I only use local user accounts and never NIS and this never happen
> here at any of the machines I used.

me too.

> Ans note: the machine never shutdown if the
> daemon is not running. This can't happen if powersave is not running ;)

If the battery is empty, the machine will power off. Even without powersaved.

> I only use local user accounts and never NIS and this never happen
> here at any of the machines I used.
How many HT, DC or normal SMP machines were this? SMP machines are commonly used in desktops today, but you won't have much in the mobile team. I agree, that fixing this for head should be enough for the moment.

I only have one bugreport (AMD DC), for that right now and it seems Andi also heard some people complaining, if there are more we should consider an update (adding the "never show this window again" message) as an always popping up window is most annoying...

*** Bug 131847 has been marked as a duplicate of this bug. ***

(In reply to comment #21)
> I only have one bugreport (AMD DC), for that right now and it seems Andi also
> heard some people complaining, if there are more we should consider an update
> (adding the "never show this window again" message) as an always popping up
> window is most annoying...

Add me to the list of complaining people ! :-)

I'm writing you from my brand new OpenSuse 10.0 and each KDE start I have this bloody message popping up :
"The powersave demon isn't running. If you start it you'll get better performances : /usr/sbin/rcpowersavedd start"

Please add a button that allow the userr either to not see this message anymore or start automaticaly the deamon powersaved at KDE boot.

The daemon is already started, but because of (insecure) autologin and races in parallel booting this could happen. 

Btw. we currently don'T plan a you update for this.

If you don't like the type of the message, you can change this to a passive popup without sound: change in /opt/kde3/share/config/kpowersaverc the value of psMsgAsPassivePopup to 'true' (or add psMsgAsPassivePopup=true to ~/.kde3/share/config/kpowersaverc in section [General]).

Really, that´s no solution. Now the window appeares in
the Contol Border on the bottom of my screen.

It´s irritating.
It would be the best to shutting down the whole powersaved
or atleast kpowersave.
A new update is necessary; please do it.

I had this issue on a Athlon 2800+ (on a MSI board) and also on a Athlon 2000+  (on a VIA motherboard). Both were fresh installations of SUSE 10. A friend of mine had no problem on an Intel machine. 

I got this pop up all the time until I updated kde. Now the problem is gone. 

(In reply to comment #27)
> I got this pop up all the time until I updated kde. Now the problem is gone. 
> 

To what version, from what version, Miikka? I saw this very same condition on a desktop system last evening (older ACPI-enabled BIOS, though that shouldn't matter). Once I turned off autologin, the system behaved itself (also note Bug #133411 Comment #4 - did I reference that correctly? - which is also a condition apparently related to autologin).

Lewis

To 3.4.3 from the default version that came with with SUSE 10.0. Was it 3.4.2? 

*** Bug 135603 has been marked as a duplicate of this bug. ***

I have this bug too, or at least I get the message.  I have used 3 or four versions of suse on this machine prior and did not have this problem.  I have upgraded to KDE 3.5 and still have this problem, it makes no sense to me but is very very annoying.  I have an HP NC6000 laptop and I am not running auto-login, it doesn't matter how long I wait before logging in, therefore since powersaved is running before login I do not understand why I am getting the message.  When attempting to do what the message says (rcpowersaved start) I get:
###############################################
# ACPI system but acpid not running.          #
# Start acpid first, then restart powersaved! #
############################################### 
if I then run rcacpid start it says:
acpid: loading ACPI modules ( )                                      done
Starting acpid                                                       failed

This either means my situation is different, or fights for the point that we definately need this notification available on laptops to advise us.  I do get some kind of power management as my fan does slow down at times and I can see the status of various hardware in the /proc/ or wherever it is (can't remember sorry).  I'd love to hear what is being done about this.

there are various reasons why acpid wouldnt start, but these have nothing to do with the kpowersave warning. You are welcome to open a separate bug to fix those acpid problems, assign it to me.

Fixed in SVN! (Added new dialog with checkbox) 

*** Bug 149169 has been marked as a duplicate of this bug. ***