bug#26158: GNOME starts unreliably

bug#26158: GNOME starts unreliably

[Leo Famulari]

[-- Attachment #1: Type: text/plain, Size: 604 bytes --]

----- Forwarded message from "pelzflorian (Florian Pelz)" <pelzflorian@pelzflorian.de> -----
Date: Sat, 18 Mar 2017 15:50:49 +0100
From: "pelzflorian (Florian Pelz)" <pelzflorian@pelzflorian.de>
To: help-guix@gnu.org
Subject: Re: I can only start an X server in guix as root
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Icedove/45.8.0

[...]

So I’m not going to be of much help, but I want to rule out the obvious:
For me most of the time GNOME starts only on the second log-in attempt
on GuixSD. I assume you have tried logging in multiple times in a row?

[...]

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

bug#26158: Gnome starts unreliably

[Joshua Branson]

Yes sir.  I've tried logging in multiple times.  I did it just now to be
sure.  I tried logging in 3 times for each of the desktops I have
installed: Gnome on X, Gnome (which I assume is on wayland), and on
xfce.    I could not log in.  But when I tried logging in as root, it
worked just fine.  :)   Just to reiterate, I can login to a virtual
console. 


If I could get Emacs' erc to work on a virtual console, I'd complain a
little less perhaps. 


P.S.


Herein follows a slightly related rant, in which I think out loud.  I've
heard about Guix developers logging into a virtual console and working
from that, which sounds totally B.A.  The reason that doesn't work for
me at the moment, is that I've told erc to store my irc passwords in
.authinfo.gpg.  AND emacs from a virtual console does not properly
decrypt the file, so I can't use erc on emacs in a virtual console.  The
error I get says pinentry not installed, but I know I have installed it
before.  I did just reconfigure, so maybe I should have to install it
again.  I suppose that I could write some elisp, to check if the current
system is guix.  If yes, then don't use .authinfo.gpg.   Then I would
have to figure out how to mail via emacs, which I've done before.  That
way I could contribute to guix whilst inside guix.


Joshua

bug#26158: Gnome starts unreliably

[pelzflorian (Florian Pelz)]

Does

guix package -i xf86-video-intel

or

guix package -i xf86-video-nouveau

help (or whatever else your GPU is from `guix package -s xf86-video-`)?

Regards,
Florian

bug#26158: Gnome starts unreliably

[Ludovic Courtès]

"pelzflorian (Florian Pelz)" <pelzflorian@pelzflorian.de> skribis:

> Does
>
> guix package -i xf86-video-intel
>
> or
>
> guix package -i xf86-video-nouveau
>
> help (or whatever else your GPU is from `guix package -s xf86-video-`)?

Xorg wouldn’t pick up video drivers installed in the user’s profile, so
this shouldn’t make any difference AIUI.  (See ‘xorg-configuration-file’
in (gnu services xorg).)

Am I missing something?

Thanks,
Ludo’.

bug#26158: Gnome starts unreliably

[Joshua Branson]

[-- Attachment #1.1: Type: text/plain, Size: 2333 bytes --]


I ran as "joshua"

guix package -I xf86-video-intel

and

guix package -I xf86-video-nouveau


Guix had no output on either of those commands, which means neither is installed for my user joshua right?  BUT I can login to gnome as root.  So obviously one of those xf86-video-* is installed in my root user profile right?


I did go ahead and login to a virtual console and run as "joshua":

guix package -i xf86-video-nouveau;


So now my non-root user "joshua" has xf86-video-nouveau installed.  I tried logging into Gnome, gnome on X, and xfce several times as "joshua".  It did not work.  I rebooted and tried again.  I still cannot login.

Joshua

P.S.  I originally posted this to help-guix, so I'm not sure how it came to be on the guix-bugs mailing list.  But when I look at the bug online ( https://debbugs.gnu.org/cgi/bugreport.cgi?bug=26158 )  my original email is not present.  So I'm attaching the original email as well, which has my config.scm and some relevant errors that I can find.  Thanks for the rapid responses.  Essentially I am dual booting guixSD and Parabola and both share the same /home.  I'm assuming that is part of the problem.


-------- Forwarded Message --------
Subject:        I can only start an X server in guix as root
Date:   Fri, 17 Mar 2017 20:50:07 -0400
From:   Joshua Branson <bransoj@hotmail.com><mailto:bransoj@hotmail.com>
To:     help-guix@gnu.org<mailto:help-guix@gnu.org>





    Hello,


I currently dual boot guixSD and Parabola.  I use a macbook 7,1, and I
am unable to start an X server in guix as my normal user.  I can login
to gnome via my root user, but obviously this is not ideal.  I can login
to a virtual console as my normal user and access Emacs (yeah), but it
would be nice to be able to start gnome as my normal user.  Just today I ran

guix pull; sudo guix pull; guix system reconfigure /etc/gnome-xfce.scm


It updated fine.  I rebooted and the problem still persists.  Parabola
and guix have seperate / partitions, but they share the same /home.
I've set up guix to create my normal user (joshua) with a uid of 2000,
because that is the uid that the user joshua has on parabola.

I'm attaching my config.scm file and some errors that I've found that
are probably relevant.  Any help would be nice.






[-- Attachment #1.2: Type: text/html, Size: 3441 bytes --]

[-- Attachment #2: guix-X-errors.txt --]
[-- Type: text/plain, Size: 3880 bytes --]

grep WW Xorg.0.log
[  6357.578] (WW) xf86CloseConsole: VT_ACTIVATE failed: Input/output error
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[    47.212] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/100dpi/".
[    47.212] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/75dpi/".
[    47.237] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[    47.385] (WW) Warning, couldn't open module nv
[    47.393] (WW) Warning, couldn't open module fbdev
[    47.393] (WW) Warning, couldn't open module vesa
[    47.394] (WW) Falling back to old probe method for modesetting
[    47.395] (WW) VGA arbiter: cannot open kernel arbiter, no multi-card support
[  6357.578] (WW) xf86CloseConsole: KDSETMODE failed: Input/output error
[  6357.578] (WW) xf86CloseConsole: VT_GETMODE failed: Input/output error
[  6357.578] (WW) xf86CloseConsole: VT_ACTIVATE failed: Input/output error


grep EE Xorg.1.log
[    47.080] Current Operating System: Linux parabola 4.10.1-gnu-1 #1 SMP PREEMPT Mon Feb 27 17:40:02 UYT 2017 x86_64
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[    47.385] (EE) Failed to load module "nv" (module does not exist, 0)
[    47.393] (EE) Failed to load module "fbdev" (module does not exist, 0)
[    47.393] (EE) Failed to load module "vesa" (module does not exist, 0)
[    47.580] (EE) NOUVEAU(0): [COPY] failed to allocate class.
[  6357.259] (EE) systemd-logind: failed to release device: Connection is closed
[  6357.284] (EE) systemd-logind: failed to release device: Connection is closed
[  6357.309] (EE) systemd-logind: failed to release device: Connection is closed
[  6357.346] (EE) systemd-logind: failed to release device: Connection is closed
[  6357.366] (EE) systemd-logind: failed to release device: Connection is closed
[  6357.379] (EE) systemd-logind: failed to release device: Connection is closed
[  6357.393] (EE) systemd-logind: failed to release device: Connection is closed
[  6357.578] (EE) systemd-logind: ReleaseControl failed: Connection is closed

tail -n 10000 /var/log/messages | grep WARNING -m 5 -n

223:Jan  3 15:14:42 localhost avahi-daemon[326]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns!
691:Feb  4 10:27:09 localhost avahi-daemon[326]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns!
1694:Feb  4 11:07:13 localhost avahi-daemon[324]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns!
2626:Feb  4 11:12:58 localhost avahi-daemon[326]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns!
3364:Feb  4 17:58:05 localhost avahi-daemon[326]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns!

tail -n 10000 /var/log/messages | grep ERROR -m 5 -n

1225:Feb  4 10:33:46 localhost ntpd[321]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
3909:Feb  4 18:07:39 localhost ntpd[322]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
6121:Mar 14 15:51:34 localhost ntpd[320]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
6916:Mar 15 16:17:38 localhost ntpd[318]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
7687:Mar 16 13:29:37 localhost ntpd[319]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized


some other old errors

Dec  5 21:09:54 localhost gnome-session-binary[443]: WARNING: Could not parse desktop file xscreensaver.desktop or it references a not found TryExec binary
Dec  5 21:09:55 localhost gnome-keyring-daemon[475]: couldn't access control socket: /run/user/0/keyring/control: No such file or directory
Dec  5 21:09:55 localhost gnome-keyring-daemon[476]: couldn't access control socket: /run/user/0/keyring/control: No such file or directory
Dec  5 21:09:55 localhost gnome-keyring-daemon[477]: couldn't access control socket: /run/user/0/keyring/control: No such file or directory

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #3: gnome-xfce.scm --]
[-- Type: text/x-scheme; name="gnome-xfce.scm", Size: 4612 bytes --]

;; This is an operating system configuration template
;; for a "desktop" setup with GNOME and Xfce and awesome.
;; I am using this file as of march 17th

(use-modules (gnu) (gnu system nss))
(use-service-modules desktop)
(use-package-modules certs gnome)

(operating-system
 (host-name "GuixSD")
 (timezone "America/Indianapolis")
 (locale "en_US.UTF-8")
 ;;blacklist this keyboard module.  It might help my mouse to load properly.
 ;;http://debbugs.gnu.org/cgi/bugreport.cgi?bug=20433
;; (kernel-arguments '("modprobe.blacklist=usbkdb"))
 (swap-devices '("/dev/sda3"))

 ;; Assuming /dev/sdX is the target hard disk, and "my-root"
 ;; is the label of the target root file system.
 (bootloader (grub-configuration (device "/dev/sda")
				 ;; This bug was fixed with a newer version of guix.  Cool!
                                 ;; Here is my hack (grub-configuration appends bzImage to `linux')
                                 ;; to add a custom entry for debian on /dev/sda3 to the Grub menu
                                 ;; On Debian, do:
                                 ;;   sudo mkdir -p /boot/latest
                                 ;;   sudo ln -s $(ls -t /boot/vmlinuz*|head -1) /boot/latest/bzImage
                                 ;;   sudo ln -s $(ls -t /boot/initramfs*|head -1) /boot/latest/initrd
                                 (menu-entries
                                  (list (menu-entry
                                         (label "Parabola")
                                         (linux "(hd0,1)/boot/vmlinuz-linux-libre")
                                         (linux-arguments '("root=/dev/sda1 ro"))
                                         (initrd "(hd0,1)/boot/initramfs-linux-libre.img"))))))

 (file-systems (cons* (file-system
                       (device "my-root")
                       (title 'label)
                       (mount-point "/")
                       (type "ext4"))
                      (file-system
                       (mount-point "/home")
                       (device "/dev/sda4")
                       (type "ext4"))
                      %base-file-systems))

 (users (cons (user-account
               (name "joshua")
               (comment "joshua")
	       ;;I need to specify that joshua has the uid of 2000.
               (uid 2000)
               (group "users")
               (supplementary-groups '("wheel" "netdev"
                                       "audio" "video" "home"))
               (home-directory "/home/joshua"))      
              %base-user-accounts))

 (groups (cons (user-group
		(name "home")
		;;I need to specify that there is a home group.  Most of my files in my home directory are owned by the group home.
		;;Though I could probably just make them owned by joshua.
		(id 2000)) 
                %base-groups))
 ;; This is where we specify system-wide packages.
 (packages (cons*
	    ;;awesome
		  nss-certs         ;for HTTPS access
		  ;;emacs
		  ;;icecat
		  ;;adding in extra window managers, so maybe I'll be able to login.
		  ;;i3-wm
		  ;;xmonad
		  ;;ratpoison
                  %base-packages))

 
;; (packages (append (map specification->package
;;			'(			
			;;"awesome"
			;;"brasero"
			;;"devhelp"			
			;;"diction" 
			;;"emacs"
			;;"emacs-async"
			;;"emacs-ace-window"
			;;"emacs-dash"
			;;"emacs-emms"
			;;"emacs-evil"
			;;"emacs-f"
			;;"emacs-flycheck"
			;;"emacs-helm"
			;;"emacs-js2-mode"
			;;"emacs-magit-popup"
			;;"emacs-mmm-mode"
			;;"emacs-org"
			;;"emacs-paredit"
			;;"emacs-pdf-tools"
			;;"emacs-projectile"
			;;"emacs-rainbow-delimiters"
			;;"emacs-smart-mode-line"
			;;"emacs-web-mode"
			;;"emacs-zenburn-theme"
			;;"geiser"
			;;code is guile like in python
			;;"font-hack"
			;;"font-dejavu"	                
	                ;;"icecat"
			;; for HTTPS access
;;			"nss-certs"))
  ;;                %base-packages))

 ;; Add GNOME and/or Xfce---we can choose at the log-in
 ;; screen with F1.  Use the "desktop" services, which
 ;; include the X11 log-in service, networking with Wicd,
 ;; and more.
 (services (cons* (gnome-desktop-service)
                  (xfce-desktop-service)
		  ;; apparently syslog is provided by default
		  ;;(syslog-service)
                  ;;make the console use the dvorak keyboard layout
                  (console-keymap-service "dvorak")
                  %desktop-services))

 ;; Allow resolution of '.local' host names with mDNS.
 (name-service-switch %mdns-host-lookup-nss))

bug#26158: Gnome starts unreliably

[Florian Pelz]

I believe my unreliable launching of GNOME is unrelated. For me,
/var/log/messages shows output like

Mar 21 11:53:31 localhost vmunix: [   86.999947] .gnome-session-[442]:
segfault at 0 ip 00007f7b3141fce9 sp 00007ffd69b89560 error 4 in
libgtk-3.so.0.2200.6[7f7b3114a000+6ed000]

after a crash which I will investigate further with debug output for
gnome-session at a later time.

On Sun, 2017-03-19 at 17:02 +0100, Ludovic Courtès wrote:
> "pelzflorian (Florian Pelz)" <pelzflorian@pelzflorian.de> skribis:
> 
> > Does
> > 
> > guix package -i xf86-video-intel
> > 
> > or
> > 
> > guix package -i xf86-video-nouveau
> > 
> > help (or whatever else your GPU is from `guix package -s xf86-video-`)?
> 
> Xorg wouldn’t pick up video drivers installed in the user’s profile, so
> this shouldn’t make any difference AIUI.  (See ‘xorg-configuration-file’
> in (gnu services xorg).)
> 

Ah, I did not know. Good to know. Now my manifest is shorter. ;)

bug#26158: Gnome starts unreliably

[Joshua Branson]

So I’ve gotten responses for this on the help-guix mailing list and in the guix-bugs list.  I’ll try to just respond to the bug mailing list so the conversation is in one place.

I did login as root, run guix system vm /etc/gnome-xfce.scm;   A vm popped up.  I could not login as joshua, but I could also not login as root either.   When I tried to login as joshua, after I typed in my password, it prompted me for a user name.  So I can’t login as joshua in the vm.  Outside the vm, when I try logging in as joshua I see Logging in, I see lots of text, and then I’m back at the login screen.   In the vm, I tried logging in a root, and it tries to log me in, but inevitably I end up at the login screen again.   Outside the vm, when I try logging in as root, it works fine.

I tried looking for the x-session errors file and couldn’t find anything.

cat /home/joshua/.xsession-errors;

No file.


cat /root/.xsession-errors;

No file. 

cat / .xsession-errors;

no file.


Also it looks like Parabola X errors are listed in ~/.local/share/xorg/, because X is rootless in Parabola.   Does X start as root in guixSD?

If so,

pwd;  -> /var/log


grep -e Log -e tty Xorg.0.log ->  March 22  16:22:29    (today about an hour ago)

grep EE Xorg.0.log

43.680  (EE) NOUVEAU (0):  [COPY] failed to allocate class.

grep WW Xorg.0.log

(WW)  Open ACPI failed (/var/run/aspid.socket)  (No such file or directory)
(WW)  Falling back to old probe method for mode setting
(WW)  Falling back to old probe method for fbdev
(WW) Falling back to old probe method vesa


I feel like if someone tries to make a vm of my scheme file, they’re going to be able to log in just fine….Wait I didn’t set my normal user password in the vm….  So I probably could login to the vm, because my normal user password was not set right?  BUT I can’t get a tty in a vm?  I tried, and I was in outside the vm in a normal tty.

IDK.  

Joshua

P.S.  Most of the time when I login to gnome my mouse cannot move left to right.  It’s probably a different issue, but I felt I should probably mention it.