You are not logged in.
llpp is as fast as normal on my xorg 1.16 and latest intel. If you're using llpp 19, check this out: http://repo.or.cz/w/llpp.git/commit/78f … e4bf8683a1. (it is a commit after the release of llpp 19).
So, try to switch to git version and report back.
Offline
I did try both -git and the regular one. I even got some older builds and they're slow too:
$ stat --format "%n %y" *.tar | cut -d' ' -f 1,2
llpp-git-1567.ea98622-1-i686.pkg.tar 2014-05-03
llpp-git-1568.902143d-1-i686.pkg.tar 2014-06-01
llpp-git-1582.09dfafe-1-i686.pkg.tar 2014-07-02
llpp-git-1647.2fee246-1-i686.pkg.tar 2014-07-22
llpp-git-1648.6bf9432-1-i686.pkg.tar 2014-07-23
I'm using an ancient card:
$ lspci | grep -i vga
00:02.0 VGA compatible controller: Intel Corporation 82845G/GL[Brookdale-G]/GE Chipset Integrated Graphics Device (rev 01)
Offline
I can start a second X from another VT (while another X is already running) when I start it with 'sudo startx'.
Why doesn't 'startx' alone work?
Offline
karol, try running startx with `bash -x /usr/bin/startx` and see what the actual command it is trying to run?
This silver ladybug at line 28...
Offline
+ unset DBUS_SESSION_BUS_ADDRESS
+ unset SESSION_MANAGER
+ userclientrc=/home/karol/.xinitrc
+ sysclientrc=/etc/X11/xinit/xinitrc
+ userserverrc=/home/karol/.xserverrc
+ sysserverrc=/etc/X11/xinit/xserverrc
+ defaultclient=xterm
+ defaultserver=/usr/bin/X
+ defaultclientargs=
+ defaultserverargs=
+ defaultdisplay=:0
+ clientargs=
+ serverargs=
+ enable_xauth=1
+ d=0
+ true
+ '[' -e /tmp/.X0-lock ']'
+ d=1
+ true
+ '[' -e /tmp/.X1-lock ']'
+ break
+ defaultdisplay=:1
+ unset d
+ whoseargs=client
+ '[' x '!=' x ']'
+ '[' x = x ']'
+ client=xterm
+ '[' x = x ']'
+ '[' -f /home/karol/.xinitrc ']'
+ client=/home/karol/.xinitrc
+ '[' x = x ']'
+ clientargs=
+ '[' x = x ']'
+ server=/usr/bin/X
+ '[' x = x -a x = x ']'
+ '[' -f /home/karol/.xserverrc ']'
+ '[' -f /etc/X11/xinit/xserverrc ']'
+ server=/etc/X11/xinit/xserverrc
+ '[' x = x ']'
+ serverargs=
+ '[' x = x ']'
+ display=:1
+ '[' x1 = x1 ']'
+ '[' x = x ']'
+ XAUTHORITY=/home/karol/.Xauthority
+ export XAUTHORITY
+ removelist=
+ case `uname` in
++ uname
++ grep GNU
++ hostname --version
+ '[' -z 'hostname (GNU inetutils) 1.9.2
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>.' ']'
++ hostname
+ hostname=black
+ authdisplay=:1
++ /usr/bin/mcookie
+ mcookie=bf4665e6c43fad3ac75788c355e97f02
+ test xbf4665e6c43fad3ac75788c355e97f02 = x
+ dummy=0
++ mktemp -p /tmp serverauth.XXXXXXXXXX
+ xserverauthfile=/tmp/serverauth.7mqslaW9Pr
+ trap 'rm -f '\''/tmp/serverauth.7mqslaW9Pr'\''' HUP INT QUIT ILL TRAP KILL BUS TERM
+ xauth -q -f /tmp/serverauth.7mqslaW9Pr
+ serverargs=' -auth /tmp/serverauth.7mqslaW9Pr'
+ for displayname in '$authdisplay' '$hostname/unix$authdisplay'
++ sed -n 's/.*black\/unix:1[[:space:]*].*[[:space:]*]//p'
++ xauth list :1
+ authcookie=a44a4ae7db8a1ac33c3b5409b87037cf
+ '[' za44a4ae7db8a1ac33c3b5409b87037cf = z ']'
+ dummy=1
+ xauth -q -f /tmp/serverauth.7mqslaW9Pr
+ for displayname in '$authdisplay' '$hostname/unix$authdisplay'
++ sed -n 's/.*black\/unix:1[[:space:]*].*[[:space:]*]//p'
++ xauth list black/unix:1
+ authcookie=a44a4ae7db8a1ac33c3b5409b87037cf
+ '[' za44a4ae7db8a1ac33c3b5409b87037cf = z ']'
+ dummy=2
+ xauth -q -f /tmp/serverauth.7mqslaW9Pr
+ xinit /home/karol/.xinitrc -- /etc/X11/xinit/xserverrc :1 -auth /tmp/serverauth.7mqslaW9Pr
X.Org X Server 1.16.0
Release Date: 2014-07-16
X Protocol Version 11, Revision 0
Build Operating System: Linux 3.15.1-1-ARCH i686
Current Operating System: Linux black 3.15.6-1-ARCH #1 SMP PREEMPT Fri Jul 18 07:07:17 CEST 2014 i686
Kernel command line: BOOT_IMAGE=../vmlinuz-linux root=/dev/sda3 rw noapic fbcon=scrollback:512k zswap.enabled=1 zswap.compressor=lz4 zswap.max_pool_percent=50 i915.semaphores=1 quiet initrd=../initramfs-linux.img
Build Date: 18 July 2014 07:14:09PM
Current version of pixman: 0.32.6
Before reporting problems, check http://wiki.x.org
to make sure that you have the latest version.
Markers: (--) probed, (**) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/home/karol/.local/share/xorg/Xorg.1.log", Time: Wed Jul 23 17:18:00 2014
(==) Using config directory: "/etc/X11/xorg.conf.d"
(==) Using system config directory "/usr/share/X11/xorg.conf.d"
(EE)
Fatal server error:
(EE) xf86OpenConsole: Cannot open virtual console 2 (Permission denied)
(EE)
(EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
(EE) Please also check the log file at "/home/karol/.local/share/xorg/Xorg.1.log" for additional information.
(EE)
(EE) Server terminated with error (1). Closing log file.
xinit: giving up
xinit: unable to connect to X server: Connection refused
xinit: server error
+ retval=1
+ '[' x1 = x1 ']'
+ '[' x '!=' x ']'
+ '[' x/tmp/serverauth.7mqslaW9Pr '!=' x ']'
+ rm -f /tmp/serverauth.7mqslaW9Pr
+ exit 1
Offline
llpp works as expected with uxa acceleration.
Hardware doesn't support glamor and sna started causing problems with xorg 1.16.
Offline
What is the problem with xorg-server 1.16 I can not login with Gnome but if I can do it with lxde and when I downgrade to 1.15 if I loguearme with Gnome.
The truth I'm not sure if I'm the only one with this problem, but I see no solution anywhere.
Offline
https://www.archlinux.org/news/xorg-ser … available/ mentions that you shouldn't redirect to std err, but doesn't say that only one rootless X server can be run at a time, even though I posted this to the ML in response to the draft and it was forwarded / quoted on the arch-dev-public list.
kabbalah, you can't login through a login manager.
Offline
Should this thread be moved out of the testing repo forum now that this has hit the repos?
I start X from a bashrc excerpt:
if [[ -z $DISPLAY && $XDG_VTNR -eq 1 ]]; then
read -s -n 1 -t 0.25 key
[[ $? -eq 142 ]] && exec xinit -- vt1 >~/.xorg.log 2>&1
case $key in ""|;; *) exec xinit -- vt1 >~/.xorg.log 2>&1;; esac
fi
Don't mind the case stuff, that's just so I can hit a key to prevent the autostarting - it's functionally equivalent to the following:
[[ -z $DISPLAY && $XDG_VTNR -eq 1 ]] && exec xinit -- vt1 >~/.xorg.log 2>&1
The only revision needed to get it working here was to remove the redirection - I can keep the 'exec'.
However, then the error output is banished never to be seen again.
"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman
Offline
Should this thread be moved out of the testing repo forum now that this has hit the repos?
I've already asked about this in a report.
Now if only a moderator would drop by ... ;-)
Offline
Probably should be moved out as Tribly stated.
However, one other issue is vmware display driver segfaults with latest xorg-server (latest build from today). If others have this issue I'll put in yet another ticket upstream for incompatibility with latest Xorg.
Segmentation fault at address 0x8
--
JSkier
Offline
Moved from [ testing ] subforum.
Had you suggested a target subforum I would have done it right away I was torn between Apps+DEs and Sys Admin. I went with the latter.
"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman
Offline
@jskier
I'm having the VMware issue with the latest Xorg.
"Segmentation fault at address 0x8"
Offline
have you try:
startx -- vt1 -nolisten tcp -quiet
Or, maybe look like this into .bash_profile:
if [ -z "$DISPLAY" ] && [ $(tty) == /dev/tty1 ]; then
startx -- vt1 -nolisten tcp -quiet
fi
Last edited by mastotosugiarto (2014-08-06 03:08:42)
amburegul
emeseyu
bahrelway bahrelway
Offline
Not a bad suggestion, however It will still segfault with the same error. I will get a bug in upstream in a bit and provide the info here in an update.
have you try:
startx -- vt1 -nolisten tcp -quietOr, maybe look like this into .bash_profile:
if [ -z "$DISPLAY" ] && [ $(tty) == /dev/tty1 ]; then startx -- vt1 -nolisten tcp -quiet fi
--
JSkier
Offline
Karol, Are you still fighting this?
I am having the same problem as are you. I have been trying to analyze this and have found some weird stuff. I can open any console as any user and startx an X session. If I switch to a different console, as myself or any other user (except root, root doesn't do GUIs), I cannot get a session to start. I receive a "Permission Denied" opening the vt that I am on. Note the following session and the ownerships of the ttys. Before I started X, I created a file that showed the permissions after login, but prior to starting X. I show the permissions with just the single log in session running X, what happened before starting X, after login in to vt2, and (the interesting one) after I tried to start X on the second vt (tl;dr -- the ownership slips back to root). Finally, I show my groups.
ewaller$@$odin ~ 1002 %ls -l /dev/tty?
crw--w---- 1 root tty 4, 0 Aug 5 21:09 /dev/tty0
crw------- 1 ewaller tty 4, 1 Aug 6 22:02 /dev/tty1
crw--w---- 1 root tty 4, 2 Aug 6 22:00 /dev/tty2
crw--w---- 1 root tty 4, 3 Aug 6 22:00 /dev/tty3
crw--w---- 1 root tty 4, 4 Aug 6 22:00 /dev/tty4
crw--w---- 1 root tty 4, 5 Aug 5 21:09 /dev/tty5
crw--w---- 1 root tty 4, 6 Aug 5 21:09 /dev/tty6
crw--w---- 1 root tty 4, 7 Aug 5 21:09 /dev/tty7
crw--w---- 1 root tty 4, 8 Aug 5 21:09 /dev/tty8
crw--w---- 1 root tty 4, 9 Aug 5 21:09 /dev/tty9
ewaller$@$odin ~ 1003 %cat OwnershipAfterLogin
crw--w---- 1 root tty 4, 0 Aug 5 21:09 /dev/tty0
crw------- 1 ewaller tty 4, 1 Aug 6 22:02 /dev/tty1
crw--w---- 1 root tty 4, 2 Aug 6 22:00 /dev/tty2
crw--w---- 1 root tty 4, 3 Aug 6 22:00 /dev/tty3
crw--w---- 1 root tty 4, 4 Aug 6 22:00 /dev/tty4
crw--w---- 1 root tty 4, 5 Aug 5 21:09 /dev/tty5
crw--w---- 1 root tty 4, 6 Aug 5 21:09 /dev/tty6
crw--w---- 1 root tty 4, 7 Aug 5 21:09 /dev/tty7
crw--w---- 1 root tty 4, 8 Aug 5 21:09 /dev/tty8
crw--w---- 1 root tty 4, 9 Aug 5 21:09 /dev/tty9
ewaller$@$odin ~ 1004 %echo I logged into console 2
I logged into console 2
ewaller$@$odin ~ 1005 %ls -l /dev/tty?
crw--w---- 1 root tty 4, 0 Aug 5 21:09 /dev/tty0
crw------- 1 ewaller tty 4, 1 Aug 6 22:02 /dev/tty1
crw------- 1 ewaller tty 4, 2 Aug 6 22:03 /dev/tty2
crw--w---- 1 root tty 4, 3 Aug 6 22:00 /dev/tty3
crw--w---- 1 root tty 4, 4 Aug 6 22:00 /dev/tty4
crw--w---- 1 root tty 4, 5 Aug 5 21:09 /dev/tty5
crw--w---- 1 root tty 4, 6 Aug 5 21:09 /dev/tty6
crw--w---- 1 root tty 4, 7 Aug 5 21:09 /dev/tty7
crw--w---- 1 root tty 4, 8 Aug 5 21:09 /dev/tty8
crw--w---- 1 root tty 4, 9 Aug 5 21:09 /dev/tty9
ewaller$@$odin ~ 1006 %echo I tried to start X on vt2
I tried to start X on vt2
ewaller$@$odin ~ 1007 %echo I tried to start X on vt2
I tried to start X on vt2
ewaller$@$odin ~ 1008 %ls -l /dev/tty?
crw--w---- 1 root tty 4, 0 Aug 5 21:09 /dev/tty0
crw------- 1 ewaller tty 4, 1 Aug 6 22:04 /dev/tty1
crw------- 1 root tty 4, 2 Aug 6 22:05 /dev/tty2
crw--w---- 1 root tty 4, 3 Aug 6 22:00 /dev/tty3
crw--w---- 1 root tty 4, 4 Aug 6 22:00 /dev/tty4
crw--w---- 1 root tty 4, 5 Aug 5 21:09 /dev/tty5
crw--w---- 1 root tty 4, 6 Aug 5 21:09 /dev/tty6
crw--w---- 1 root tty 4, 7 Aug 5 21:09 /dev/tty7
crw--w---- 1 root tty 4, 8 Aug 5 21:09 /dev/tty8
crw--w---- 1 root tty 4, 9 Aug 5 21:09 /dev/tty9
ewaller$@$odin ~ 1009 %groups
lp wheel uucp http games lock network users vboxusers adbusers systemd-journal
ewaller$@$odin ~ 1010 %
As it sits right now, I can only have one instance of X. Often, I've my account on one instance, Windows 7 without a WM on a second, and often a different user on a third.
Another thing that broke -- I used to (sometimes), login, start a tmux session, start X, then attach to the tmux session from a a urxvt terminal. This way I could watch stdout on my console while the X session runs. Now, I am politely told I can only startx from a console.
Last edited by ewaller (2014-08-07 15:13:36)
Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way
Offline
Me2, no second X.
startx results in:
[ 310.831] (EE) xf86OpenConsole: Cannot open virtual console 2 (Permission denied)
(here it is VT2 which tries to open a second X while the first one is running on VT1, different users)
any progress so far?
Offline
Is there no bug reported upstream concerning the inability to start a second X server at all? Where is upstream anyway, or is it this?
This silver ladybug at line 28...
Offline
The issue isn't in X but in systemd (logind) actually, which while making X the "controller" of the session (i.e. allowing it to access devices) actually fails in "preparing/muting" the VT, and then therefore resets it back, including ownership of /dev/ttyX back to root, hence the permission denied error.
I've had a look at this, and there is actually a bug report upstream about the cause behind it: basically, when logind assign a controller (e.g. rootless X) to a session, it also does a few things in the VT, such as KDSETMODE, VT_SETMODE, etc
And one of the thing needed is set up a handler to respond to SIGUSR1, which will be used on VT switching. But because it can only have one handler per signal, when another rootless X wants to become controller of its session, things fail, ownership is reset and X can't start. (I'm thinking a fix might be to remove the handler when session goes inactive, and restore it back when activated again, but trying to work on that led to other issues... I'll see if I can do better tomorrow...)
(Note that at least the error is obvious, if owner wasn't reset other things would probably break, as the VT would not be in correct/expected state anyways.)
ewaller: At least you know why you see such ownerhsip changes now: upon failure of "preparing" the VT logind does reset things, which happen to include setting ownership back to root. You should in fact see the corresponding error in your journal, something like "cannot mute VT 2 for session c2 (-16/17)"
Offline
I decided to run X the old way, via the wrapper. This way I can start a second or third X without a problem. I've been a bit busy and I had no time to look into this.
I've also noticed I can't seem to set the acceleration method for my intel card, it's set to UXA no matter what.
It's all very interesting - the 'may you live in interesting times' kind of way.
Offline
The issue isn't in X but in systemd (logind) actually
<snip>
ewaller: At least you know why you see such ownerhsip changes now: upon failure of "preparing" the VT logind does reset things, which happen to include setting ownership back to root. You should in fact see the corresponding error in your journal, something like "cannot mute VT 2 for session c2 (-16/17)"
That was a superb explanation/analysis of what is going on. It makes complete sense and is in line with my observations. Thank you
Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way
Offline
I have started to get along with systemd and all the new ideas it spawns,
But i must still say that the tendency to hide/convulute logging (the old std/terminal output way) is iffy to me.
I won't complain overly, what is the new default for linux tends to be the new default for arch - this is known as we choose arch.
Though those of us that either has very personalised ways, or "older" ways of keeping our system under our control,
We have to actually put in extra work to get back to what used to be "our" default, this is the only sad noggin' for me.
But, as long as there is some way to wrestle back "live" X logging to another terminal, then i will try to make do as always.
PS. And i still appreciate all the good work of the Arch devs, no denying that
Last edited by PReP (2014-08-08 18:52:00)
. Main: Intel Core i5 6600k @ 4.4 Ghz, 16 GB DDR4 XMP, Gefore GTX 970 (Gainward Phantom) - Arch Linux 64-Bit
. Server: Intel Core i5 2500k @ 3.9 Ghz, 8 GB DDR2-XMP RAM @ 1600 Mhz, Geforce GTX 570 (Gainward Phantom) - Arch Linux 64-Bit
. Body: Estrogen @ 90%, Testestorone @ 10% (Not scientific just out-of-my-guesstimate-brain)
Offline
So I had another look at this, and I think I came up with a patch fixing the issue. It's not perfect (notably if you have two active sessions (in different seats) then it still doesn't work (for the same reason)), but it seems to work fine otherwise (at least I've been VT switching between a few rootless X without issues for a little bit now:
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
jjacky 898 1.1 0.4 160576 19028 tty3 S<l 17:13 0:04 /usr/bin/Xorg.bin -nolisten tcp vt03 :1 -auth /tmp/serverauth.w1omwguu9A
jjacky 1661 1.0 0.4 161180 18636 tty2 S<l 17:16 0:02 /usr/bin/Xorg.bin -nolisten tcp vt02 :4 -auth /tmp/serverauth.EujmtSLVkn
jjacky 2168 1.3 0.4 165812 18564 tty4 S<l 17:19 0:00 /usr/bin/Xorg.bin -nolisten tcp vt04 :5 -auth /tmp/serverauth.bVXpM9TuR8
Just sent patches upstream, we'll see what they think of it.
Offline
jjacky you're awesome! dumb question, where is the patch?
This silver ladybug at line 28...
Offline
I sent them to the ML: http://lists.freedesktop.org/archives/s … 21809.html
Offline