Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
X logs me out abruptly
View unanswered posts
View posts from last 24 hours
View posts from last 7 days

Goto page Previous  1, 2, 3  Next  
Reply to topic    Gentoo Forums Forum Index Desktop Environments
View previous topic :: View next topic  
Author Message
Zucca
Moderator
Moderator


Joined: 14 Jun 2007
Posts: 3343
Location: Rasi, Finland

PostPosted: Tue Nov 07, 2017 11:38 pm    Post subject: Reply with quote

Hmmm...
Quote:
[ 2568.069883] elogind: 579 output lines suppressed due to ratelimiting
seems odd...
Can you paste the contents of your /etc/elogind/logind.conf?
Also if you can find those ~600 missing lines somewhere... pastenbin them.
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--
Quote:
I am NaN! I am a man!
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Wed Nov 08, 2017 12:07 am    Post subject: Reply with quote

Zucca wrote:
seems odd...
Can you paste the contents of your /etc/elogind/logind.conf?

Thank you for helping.

However, as I said in my post just before, that obviously crossed yours, I can't afford to spend more time pursuing this problem.

But, just as a courtesy to someone so prompt to help his fellow, here is what you asked, it's not that long:
Code:
# cat logind.conf
#  This file is part of elogind.
#
#  elogind is free software; you can redistribute it and/or modify it
#  under the terms of the GNU Lesser General Public License as published by
#  the Free Software Foundation; either version 2.1 of the License, or
#  (at your option) any later version.
#
# Entries in this file show the compile time defaults.
# You can change settings by editing this file.
# Defaults can be restored by simply deleting this file.
#
# See logind.conf(5) for details.

[Login]
#KillUserProcesses=yes
#KillOnlyUsers=
#KillExcludeUsers=root
#InhibitDelayMaxSec=5
#HandlePowerKey=poweroff
HandlePowerKey=ignore
#HandleSuspendKey=suspend
HandleSuspendKey=ignore
#HandleHibernateKey=hibernate
HandleHibernateKey=ignore
#HandleLidSwitch=suspend
HandleLidSwitch=ignore
#HandleLidSwitchDocked=ignore
#PowerKeyIgnoreInhibited=no
#SuspendKeyIgnoreInhibited=no
#HibernateKeyIgnoreInhibited=no
#LidSwitchIgnoreInhibited=yes
#HoldoffTimeoutSec=30s
#IdleAction=ignore
#IdleActionSec=30min
#RuntimeDirectorySize=10%
#RemoveIPC=yes
#InhibitorsMax=8192
#SessionsMax=8192
#UserTasksMax=33%

[Sleep]
#SuspendState=mem standby freeze
#SuspendMode=
#HibernateState=disk
#HibernateMode=platform shutdown
#HybridSleepState=disk
#HybridSleepMode=suspend platform shutdown

The ignore settings exist because Xfce4 and acpi take care of those events.

Quote:
[ 2568.069883] elogind: 579 output lines suppressed due to ratelimiting

This was the last message of the session. After that, I clicked restart on the DM and booted the other partition. When I came back, this is the message that follows the one above:
Code:
Oct 27 19:05:16 mypc kernel: [    0.000000] Linux version 4.13.9-gentoo (root@mypc) (gcc version 6.4.0 (Gentoo 6.4.0 p1.0)) #1 SMP Wed Oct 25 16:39:09 WEST 2017


Thank you again for helping.
Back to top
View user's profile Send private message
LIsLinuxIsSogood
Veteran
Veteran


Joined: 13 Feb 2016
Posts: 1179

PostPosted: Wed Nov 08, 2017 12:50 am    Post subject: Reply with quote

That sounds like a good idea.

Just some final thoughts to consider:
a) with many application or software related crashes...it is helpful to know more about the situation leading up to the incident, yes.
b) from the information you provided the most suspicious messages are not in X logs or the Vbox stuff but the information describing logging in and out...fixing sddm?? (Some other things for you to try would be bypassing sddm, to see if the problems persist).

I think given a bit of time as you said someone with a bit more knowledge of the authentication involved in user/root logins may be able to get you there much faster.

Good luck

EDIT: Another thing I think worth mentioning is that you are 100% in the right to be saying it, about the limited amount of time, but also seeing how you are trying to get the thing working and I may have actually been getting in the way more than anything I'm glad you will continue to hopefully receive help from someone else in this capacity because while I was purposefully misleading you (I assure that is true) that the solution could be a quicker one with someone else's advice than mine (and that I am very sure of)!
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Thu Nov 09, 2017 10:07 am    Post subject: Reply with quote

LIsLinuxIsSogood wrote:
I was purposefully misleading you

You did not mislead me. Although together we couldn't achieve any result, your good-will is enough for my gratitude.

Furthermore, from a distance of a few days, I can conclude that either I'm not ready for Gentoo or Gentoo is not ready for me. I guess most people here will say it's the former, and they are probably right. But anyway, the net result is the same although I tried hard.

Keep well.
Back to top
View user's profile Send private message
Zucca
Moderator
Moderator


Joined: 14 Jun 2007
Posts: 3343
Location: Rasi, Finland

PostPosted: Thu Nov 09, 2017 11:25 am    Post subject: Reply with quote

My guess why this happened is that for some reason elogind thought you were idling and then logging you out.
Some DEs override elogind settings... those that are in logind.conf. logind.conf basically stores the default settings and then your DE can apply its own over it.
To the point: I'd disable idle actions from the logind.conf and set that infamouns KillUserProcs to no.

This is just my guess.

The other guess I have is that the Apple hardware has some quirks that some component of your system reads wrong. Starting from Apple EFI "1.5"... which isn't even a standard.
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--
Quote:
I am NaN! I am a man!
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Thu Nov 09, 2017 11:53 am    Post subject: Reply with quote

Thank you for stepping in again.

Although I gave up trying to fix the issue systematically, and returned to my former distro for daily work, that doesn't mean I gave it up completely. As I said, I'll take a peek once in a while, mainly in the evening and lunch time. I just took a few days off because I was very frustrated and irritated about this issue. I guess I'm used to solve all the problems I lay my hands on, forgive the immodesty.
Zucca wrote:
I'd disable idle actions from the logind.conf and set that infamouns KillUserProcs to no.

I had already heard about the later; isn't this one of the biggest criticisms people make to systemd as well? This I can try in the evening to give it enough time to crash. Then, if I start trusting the system, I put it on daily work for ultimate test.
As to disable idle actions, is this IdleAction=ignore? Isn't this the default?
Quote:
The other guess I have is that the Apple hardware has some quirks that some component of your system reads wrong. Starting from Apple EFI "1.5"... which isn't even a standard.

Again, why do the other distros not have this problem? Well, it actually happened just once since I remember.
Is this dependent on the Apple firmware? I think I have the latest version. I'll investigate this.

I'll post the results maybe tomorrow.
Back to top
View user's profile Send private message
Zucca
Moderator
Moderator


Joined: 14 Jun 2007
Posts: 3343
Location: Rasi, Finland

PostPosted: Thu Nov 09, 2017 3:21 pm    Post subject: Reply with quote

jssilva wrote:
As to disable idle actions, is this IdleAction=ignore? Isn't this the default?
IIRC yes it is.
To be sure I'd put that on the config. Also make sure there isn't anything in your DE/WM settings that would enable it.
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--
Quote:
I am NaN! I am a man!
Back to top
View user's profile Send private message
hceline
n00b
n00b


Joined: 30 Aug 2015
Posts: 36

PostPosted: Thu Nov 09, 2017 6:32 pm    Post subject: Reply with quote

I when you mentioned disabling the radon card, it got me thinking. I experienced the exact same symptoms when I forced the kernel to load i915.ko before radeon.ko while trying to debug an vaapi issue. I've never had X-crashes with no error-message anywhere before, but when I loaded i915.ko before radeon.ko (or had i915 built in and radeon as module) X crashed that way about 4 times a day.
Edit: I'm running X on the i915 and fbcon on radeon.
Back to top
View user's profile Send private message
Zucca
Moderator
Moderator


Joined: 14 Jun 2007
Posts: 3343
Location: Rasi, Finland

PostPosted: Fri Nov 10, 2017 2:45 pm    Post subject: Reply with quote

I wonder if this need to be configured like Nvidia optimus setups...
On those setups the intel GPU is always the one that's outputting the signal to the screen. The more powerful Nvidia GPU only handles the rendering and sends the image (screen output) to the intel GPU which then does the actual interaction with the screen.

Unfortunately I have no experience on such setups. :(
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--
Quote:
I am NaN! I am a man!
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Fri Nov 10, 2017 3:53 pm    Post subject: Reply with quote

Zucca wrote:
I wonder if this need to be configured like Nvidia optimus setups...

No, that I already know for sure. Optimus is a "false" multiplexer that works as you said, the intel gpu is always working, either as the only gpu in intel mode, or as a vehicule of the discrete gpu (nvidia or radeon), just to render the image to the screen. So, it is in fact a software multiplexer and you must use bumblebee to switch the gpu's.

The macbook pro has a hardware (true) multiplexer and so you can use either gpu independently and switch the other one off (or not), to save battery power. In this case, you cannot use bumblebee which is only for Optimus, you must use vgaswitcheroo which is included in the kernel.

So, the system is working since more than a full day now without crashing! This is the report:

I set KillUserProcesses=no in logind.conf and boot the partition. Sddm wouldn't start, gave me a black screen, X was stopping on:
Code:
[    11.038] (++) using VT number 7
[    11.042] (II) intel(0): Using Kernel Mode Setting driver: i915, version 1.6.0 20170619

I don't understand the connection between my modification and the result, or even if my observation was correct after so much tweaking. The fact is that I tried to boot several times but sddm never started.
This hadn't happened before consistently but it had happened a few times, which were recovered with a reboot.
Then I compared the Xlog of a succeeded boot with a frustrated one and pinpointed a difference: on the good one, the active card (look at the star on it) was the integrated (intel) whereas on the frustrated the active card was the radeon, which was not programmed on xorg.conf:
Code:
[    13.421] (--) PCI: (0:0:2:0) 8086:0116:106b:00dc rev 9, Mem @ 0xb0000000/4194304, 0xa0000000/268435456, I/O @ 0x00003000/64
[    13.421] (--) PCI: *(0:1:0:0) 1002:6741:106b:00e2 rev 0, Mem @ 0x90000000/268435456, 0xb0800000/131072, I/O @ 0x00002000/256, BIOS @ 0x????????/131072

So I concluded that the vgaswitcheroo was not switching the cards soon enough before X started. I had a way to fix this, instead of running the service on default level, I rc-updated to boot level:
Code:
depend()
{
    need sysfs
}

start()
{
    ebegin "Switching to Integrated GPU, shutdown Discrete"
    # Power-on both gpu's
    echo ON > /sys/kernel/debug/vgaswitcheroo/switch
    # Switch to integrated (intel) gpu
    echo IGD > /sys/kernel/debug/vgaswitcheroo/switch
    # Power-off the disabled gpu (radeon in the instance)
    echo OFF > /sys/kernel/debug/vgaswitcheroo/switch
    return 0
}

It worked:
Code:
[    13.421] (--) PCI:*(0:0:2:0) 8086:0116:106b:00dc rev 9, Mem @ 0xb0000000/4194304, 0xa0000000/268435456, I/O @ 0x00003000/64
[    13.421] (--) PCI: (0:1:0:0) 1002:6741:106b:00e2 rev 0, Mem @ 0x90000000/268435456, 0xb0800000/131072, I/O @ 0x00002000/256, BIOS @ 0x????????/131072

So now X is not able to use radeon and must use intel. Both drivers (modules) are loaded though.

A colateral improvement is that lxdm had never worked, and I had tried it several times with several tweakings (I prefer sddm, this was just another experiment), and it started to work, which is a good sign towards system stabilization.

I don't know which of these two changes (hopefully) fixed the problem, or if the two are needed. I'll have to let it work for a lot longer before feeling safe.

Thank you for your advices. I'll be posting news in a few days, hopefully good.

hceline wrote:
I'm running X on the i915 and fbcon on radeon

Thank you for your information. In fact it strengthens my above theory.

I don't think you can use both gpu's at the same time, and I'm positively sure I read somewhere (reliable) that you cannot switch when X is loaded. You must issue the commands to vgaswitcheroo (see above) with X unloaded. There's a way though, the delayed command DIGD or DDIS which action remains suspended until X is reloaded. I also read that the active gpu status is stored on nvram and that's why macos reaches it, and that's why you must use refind, or another tool which name I don't recall now, to be able to use the intel gpu when you boot anything else than macos. Refind has a setting "spoof_osx_version" for that.

Look at Xorg log, at the PCI line where the cards are discovered, to check which is ready, or both, then it's a matter of your xorg.conf. In my case, I don't want or need the radeon card, I would take it out if I could; but then, I'm not a gamer.
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Tue Nov 14, 2017 1:02 pm    Post subject: Reply with quote

Report:

The system had been working without crashes for a full day and then, just for test, I tried to suppress the KillUserProcesses=no that I had set. Another crash after a short time.

Set the KillUserProcesses=no as it was, and the sytem worked normally for a few days, making me very happy, I thought I made it. Suddenly, just a while ago, another crash.

So, this is enough for me. End-of-session.
Back to top
View user's profile Send private message
LIsLinuxIsSogood
Veteran
Veteran


Joined: 13 Feb 2016
Posts: 1179

PostPosted: Wed Nov 15, 2017 11:07 am    Post subject: Reply with quote

Hey jssilva, here's some more thoughts on the situation. Although not trying to make another lecture happen :wink:
The thing about other OS's is that yes while they are there to be able to work (for you the user), it isn't without its trade-offs that it works like this out of the box. You would be hard pressed to find a more efficient and customizable use of the linux kernel in a system overall, and for myself, having made the switch about 2 years ago from larger distro's of Red Hat and OpenSuse I first hand can tell you that my current system is so much more stable!!!

But that doesn't mean it is any easier to operate. Quite the opposite, it is takes some serious dedication so I don't want to mislead you about that. If you are willing to put up with the learning curve aspects, these are the suggestions for how you might proceed to diagnose a very big problem with no clear solution (someone else can correct any of this if they think it doesn't make sense to do it or alter it however you see fit):

Suggestion 1: Try at least 1 other login manager,

Suggestion 2: Go without a login manager using startx, xinit or just plain old X (make sure it is launched from an accountin which the user has access to the video group)

Suggestion 3: In general, be patient and keep using it just at night like you said until you find that the system (along with X) is in a more stable place. The thing tht is likely going to get you there the fastest is the help from this forum, but it doesn't happen automatically you will need to locate and provide the proper debugging information, or be able to ask the right questions to elicit the quickest response....

Suggestion 4: It may not be right away, or ever even with the particular Macbook that you're using, but when you've successfully implemented some of the steps to get past whatever basic trial of putting X to the test on your Gentoo linux machine, then it will be natural for that to lead you to take greater advantage of it for your other work too.

That's it from me, I hope you are able to bounce back from after this last blow! Just think of it this way, from a standpoint of essential functionality X is either very critical or not at all, since if you want a GUI/DE to be working it is 100% critical, but if you don't it is 0%. And then if you are looking for some interesting things to do while testing the stability of the graphics drivers with X, and the DM then I suggest finding out a bit more about the former, and leaving the latter out of the equation for now. Therefore, my rant is now changed, I'm no longer just anti polkit and anti consoleki, I'm actually anti Desktop environment. And I think the section of the forum could be removed. Jk, maybe just improved.

Remember, don't forget to try using startx to test instead of your login manager, which means removing xdm maybe for the time being from your list of services at boot time, and seeing if that could help.
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Wed Nov 15, 2017 2:23 pm    Post subject: Reply with quote

Thank you for your continued interest and helping will. I'm sincerely touched.

Quote:
Suggestion 1: Try at least 1 other login manager,

Did that already with lightdm and lxdm, see my older posts. Didn't work.
Quote:
Suggestion 2: Go without a login manager using startx

Did that already, see my older posts. Didn't work.
Quote:
Suggestion 3: In general, be patient and keep using it

Did that already, for 19 days to be precise. Look at my op. Didn't work.
Quote:
Suggestion 4: It may not be right away

Yes, I guess not, let's hope for better days.
As I said before, either I'm not ready for Gentoo or Gentoo is not ready for me.
Anyway, I'm resting for now.

Keep well.
Back to top
View user's profile Send private message
LIsLinuxIsSogood
Veteran
Veteran


Joined: 13 Feb 2016
Posts: 1179

PostPosted: Wed Nov 15, 2017 7:22 pm    Post subject: Reply with quote

Quote:
Did that already, see my older posts. Didn't work.


I will have a look at those again, but if you tried with startx, what command sid you issue, what Window manager? Do you know how to configure X in general to get more out of it?

My interpretation of X working out of the box os that either your in luck(clearly not the case) or you are up against a real challenge at times with it.

If you have tried to go GUI without Desk environment and still experienced a crash then the post may better moved to eith hardware and X setting.
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Fri Nov 17, 2017 10:56 am    Post subject: Reply with quote

As I said, I gave up this matter, at least until a solid clue comes up of what's happening. And our conversations, that I praise, might eventually do that.
LIsLinuxIsSogood wrote:
if you tried with startx, what command sid you issue, what Window manager? Do you know how to configure X in general to get more out of it?

As I said before, my DE is Xfce4, my WM is wxfm4, I issued just startx; nevertheless, my ~/.xinitrc, setup by portage, is:
Code:
ck-launch-session dbus-launch --sh-syntax --exit-with-session xfce4-session

Quote:
My interpretation of X working out of the box os that either your in luck(clearly not the case) or you are up against a real challenge at times with it.

On other more standard machines, my later experience backed by many people on the web, is that it works out of the box, without any configuration.

However, my machine is everything but standard. And so, I have configuration, mostly for the touchpad with synaptics driver, not libinput, but also a minimum one for the intel gpu. I always thought that the radeon, being disabled and powered off, wouldn't need one.

But yesterday when I read your post, I thought it could pay-off trying to include a device configuration for radeon with the option "Inactive".

Just a theory: what might be happening is that X has a bug, let's call it disfunctionality, that suddenly makes it call, or switch to, the radeon gpu, that it formally doesn't know is inactive, and not finding a card device, shuts off orderly (that's what is happening) and sends me to the DM which addresses the last card it used, the intel. I'll try this over the weekend.

Quote:
If you have tried to go GUI without Desk environment and still experienced a crash then the post may better moved to eith hardware and X setting.

I'll consider that later, or perhaps the mods will.
Back to top
View user's profile Send private message
Zucca
Moderator
Moderator


Joined: 14 Jun 2007
Posts: 3343
Location: Rasi, Finland

PostPosted: Fri Nov 17, 2017 11:24 am    Post subject: Reply with quote

If you ever find a fix for the problem here, please consider writing about the fix to the wiki article at https://wiki.gentoo.org/wiki/Macbook_Pro_(early_2011) .

(I don't remember the %<hex> -codes for (), so just copy-paste the url.)
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--
Quote:
I am NaN! I am a man!
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Fri Nov 17, 2017 11:30 am    Post subject: Reply with quote

Zucca wrote:
If you ever find a fix for the problem here, please consider writing about the fix to the wiki article at https://wiki.gentoo.org/wiki/Macbook_Pro_(early_2011)

In case I find it, and I sincerely hope so because the philosophy behind gentoo is what best suits my user needs, hopes and concerns in the IT world, I'll do that.

EDIT: That wiki page covers the MacBook Pro 8.1, which is 13" and, on the matter that worries me, is very different from mine because it does not have the radeon gpu, just the intel HD3000. I don't think my problem fits there, unless you think differently.

See here: https://en.wikipedia.org/wiki/MacBook_Pro
Back to top
View user's profile Send private message
Zucca
Moderator
Moderator


Joined: 14 Jun 2007
Posts: 3343
Location: Rasi, Finland

PostPosted: Fri Nov 17, 2017 3:06 pm    Post subject: Reply with quote

jssilva wrote:
That wiki page covers the MacBook Pro 8.1, which is 13" and, on the matter that worries me, is very different from mine because it does not have the radeon gpu, just the intel HD3000. I don't think my problem fits there, unless you think differently.
True. I guess that page needs own sections for the non-radeon model information and radeon model information. The page title suggest that the article contains information for both.

Let's think that when the problem is solved...
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--
Quote:
I am NaN! I am a man!
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Mon Nov 20, 2017 9:29 am    Post subject: Reply with quote

FYI, no problems over the weekend.

Back on-line today for real test, hopefully final.
Back to top
View user's profile Send private message
LIsLinuxIsSogood
Veteran
Veteran


Joined: 13 Feb 2016
Posts: 1179

PostPosted: Tue Nov 21, 2017 5:39 am    Post subject: Reply with quote

What DE is it again, and my experience with sddm (plasma's display/login manager) is just like everything in that DE it is difficult to manage without a keen sense of the ins and outs of the entire thing. Maybe a simpler login manager, would be a good cure for the problem at least for now.

can you show cat output to /etc/conf.d/xdm, if you already shared that earlier sorry about it.
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Tue Nov 21, 2017 6:49 am    Post subject: Reply with quote

LIsLinuxIsSogood wrote:
What DE is it again

The DE is Xfce4. As to the DM, the system already crashed using sddm, lightdm and lxdm. So, I think the DM is ruled out as to the cause of the problem.
Quote:
can you show cat output to /etc/conf.d/xdm, if you already shared that earlier sorry about it.

Code:
CHECKVT=7
#DISPLAYMANAGER="xdm"
DISPLAYMANAGER="sddm"
#DISPLAYMANAGER="lightdm"
#DISPLAYMANAGER="lxdm"

Anyway, the system was online the whole of yesterday without problems and, although I'm not convinced yet, I think there's a solid probability that the problem is solved.
I'll post again at the end of the week if, hopefully, nothing happens before.
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Sat Nov 25, 2017 6:13 pm    Post subject: Reply with quote

During this week the system crashed 3 times with the usual symptoms, although on 2 of them it is justified (but not understood). In fact, in these two occasions, it happened about 5 minutes after booting with a modified xorg.conf where I replaced the intel driver with uxa acceleration, by the modesetting driver with glamor acceleration. And I think this is reproducible, but didn't test.

So, despite what I read advising the modesetting driver for SandyBridge chipsets and earlier, apparently it is incompatible with my setup.

So the single relevant crash of the week was unpredictable and silent as before. I had turned off bluetooth and thus concluded that the two messages about it just before the crash were consequences, not reasons. So, almost no logs at all except:
Code:
Nov 24 09:38:15 mypc polkitd[6535]: Unregistered Authentication Agent for unix-session:3 (system bus name :1.29, object path /org/gnome/PolicyKit1/AuthenticationAgent, locale en_GB.utf8) (disconnected from bus)
Nov 24 09:38:15 mypc sddm-helper: pam_unix(sddm:session): session closed for user myuser

Unless I'm disregarding an obscure way of activating more eloquent debug logs, and I think I tried them all, I think the devs could have done better.

Anyway, If the average score of one crash per week holds, I think I can live with that until I find the reasons, if ever, or change my laptop.

To the mods, I'm not marking this "solved" because it isn't; and, if they want to move this thread to an area more compatible with Xorg problems, I fully agree because I don't think this has anything to do with DE's.
Back to top
View user's profile Send private message
Zucca
Moderator
Moderator


Joined: 14 Jun 2007
Posts: 3343
Location: Rasi, Finland

PostPosted: Sat Nov 25, 2017 9:48 pm    Post subject: Reply with quote

Since the crashing happens at random, I'd set kernel log level to maximum and maybe install all the possible hardware logging tools, like mcelog. ... Just to rule out hardware problems.
(But did you mention earlier that you ran some other distro without problems?)

But first I need to address this:
jssilva wrote:
As I said before, my DE is Xfce4, my WM is wxfm4, I issued just startx; nevertheless, my ~/.xinitrc, setup by portage, is:
Code:
ck-launch-session dbus-launch --sh-syntax --exit-with-session xfce4-session
If you have elogind running, you propably shouldn't use ck-launch-session. Afaik elogind takes care of many things. So just by "exec xfce4-session" you should be ok.
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--
Quote:
I am NaN! I am a man!
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Sun Nov 26, 2017 12:56 am    Post subject: Reply with quote

Zucca wrote:
If you have elogind running, you propably shouldn't use ck-launch-session. Afaik elogind takes care of many things. So just by "exec xfce4-session" you should be ok.

Thank you for your quick reaction and support. The change is already in place. I'll keep you posted.

And yes, I have another distro on a side partition which ran for many months (more than one year) with just one similar crash; I can't remember if this was after I decided to take the advice to install the modesetting driver with glamor accelerator. Before that, I used xubuntu 14.04 for about two years without any problems. Actually, I still have it on yet another partition.

I have one data partition and 4 system partitions: macos, xubuntu 14.04, manjaro openrc (now artix), and gentoo; all are fully updated and ready to run although I would like to stick to gentoo as my operational system for the reasons I've already explained.

EDIT: I forgot to say that the call ck-launch-session wasn't my initiative; it was portage that set it up like that and I didn't touch it, as far as I can remember. Furthermore, the wiki advises it in certain circumstances. One of these is having consolekit, which I would like to get rid of but couldn't up to now:
Code:
$ equery d consolekit
 * These packages depend on consolekit:
lxde-base/lxdm-0.5.3-r1 (consolekit ? sys-auth/consolekit)
net-misc/networkmanager-1.10.0 (consolekit ? >=sys-auth/consolekit-1.0.0)
sys-apps/accountsservice-0.6.45 (!elogind ? sys-auth/consolekit)
sys-auth/pambase-20150213-r1 (consolekit ? sys-auth/consolekit[pam])
sys-auth/polkit-0.113-r4 (!elogind ? sys-auth/consolekit[policykit])
x11-misc/sddm-0.16.0-r3 (consolekit ? >=sys-auth/consolekit-0.9.4)
xfce-extra/xfce4-power-manager-1.6.0-r1 (!systemd ? sys-auth/consolekit)
Back to top
View user's profile Send private message
jssilva
Tux's lil' helper
Tux's lil' helper


Joined: 27 Sep 2017
Posts: 106

PostPosted: Sat Dec 02, 2017 5:01 pm    Post subject: Reply with quote

Just to let you know that after one week, the system had no crashes and even no detected instabilities, hurraaa! and many thanks to all that jumped in to help.

Yesterday I could finally get rid of consolekit by re-emerging the dependent xfce4-power-manager with USE=systemd, imagine that, and the expected systemd truckload was not pulled in. If you feel curious, see https://forums.gentoo.org/viewtopic-t-1072936.html

Today, I also migrated the whole system to profile 17.0 desktop, recompiling everything, and I'm now planning to migrate gcc to v7.

EDIT: After some reading, I think it's better to stick to stable v6.4 until v7.2 is declared stable for amd64.

I guess that all these changes, if the system remains stable, will mask the main reason of the felt X crashes, but I don't mind if I can start trusting it.

Keep well, see you around Gentoo.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Desktop Environments All times are GMT
Goto page Previous  1, 2, 3  Next
Page 2 of 3

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum