[pending] Desktop freezing occasionally (pentium4)

deep42thought · 2019-09-07 07:21:39

cx wrote:

@deep42thought: pacdiff found pacnew files for /etc/default/grub, /etc/lightdm/lightdm.conf, and /etc/makepkg.conf, which I skipped for now. – Shall I overwrite them?

Merging is the best strategy here: "v" -> edit with vim -> save -> remove/overwrite (depending on which file you edited).
/etc/makepkg.conf is the file which is responsible for building bogus packages.

regards,
deep42thought

cx · 2019-09-07 17:07:15

@deep42thought: Yes, but IMO I don’t need to change/overwrite /etc/makepkg.conf, as the following section is the same compared with /etc/makepkg.conf.pacnew:

CARCH=pentium4
CHOST="i686-pc-linux-gnu"
CPPFLAGS="-D_FORTIFY_SOURCE=2"
CFLAGS="-march=pentium4 -mtune=generic -O2 -pipe -fno-plt"
CXXFLAGS="-march=pentium4 -mtune=generic -O2 -pipe -fno-plt"

levi wrote:

pacman -Qi|less

Well, a package/arch determining routine would be much more comfortable since I already got >1,600 packages installed ...

cx wrote:

...
[  171.767639] perf: interrupt took too long (2510 > 2500), lowering kernel.perf_event_max_sample_rate to 79500
[  248.970680] perf: interrupt took too long (3146 > 3137), lowering kernel.perf_event_max_sample_rate to 63300
[  464.711661] perf: interrupt took too long (3945 > 3932), lowering kernel.perf_event_max_sample_rate to 50400
The "perf: interrupt took too long" lines appear even when there’s no freezing; usually always at the very end, probably because the box simple was idle while waiting for a freeze. I looked for that in the net, and they seem harmless.

Hm, maybe one "perf" line is harmless, but does it mean something when there are three of them?

Meanwhile I believe mousepad is involved. That’s my favorite text editor in X and part of the xfce-goodies group. I use it almost daily, and a desktop freeze happens when it is or was running – just like in the currently frozen session! The funny thing is, yet in mousepad I can still type text and am able to save and quit the program – but that’s it; I can’t switch to another client, even on the same workspace.

The freezing does not occur directly after starting mousepad, but more like after a random period of time afterwards, and not necessarily when I’m working in it. – Couldn’t there be a relation: unloading "libinput" module(s) <--> text editor?

NB:

cx wrote:

I compared xmonad and xmonad-contrib on both machines – and look, they appear to be 'same-same but different':
Arch64 version: 0.15-18
Arch32 version: 0.15-18.0
So, can inconsistant version naming be the problem?

Reminder to myself: RTFM

levi · 2019-09-07 18:02:43

Yes, I also get interrupt taken too long messaging filling my dmesg. I also have a run of three of them in my current dmesg, albeit at times 1582, 2722 and 4525 which just means not much else was happening that got logged over that span, but yours are still over a second apart so I'd still count them as just happening while other logging is quiet, so the fact they're sequential is due to that, not anything intrinsic about them.

Re using less, hit '/' when in less then type 'i686' and enter. After reviewing that dodgy package, hit 'n' to go to the next hit (maybe twice if you went up from the previous hit). You should only have about 60 packages to find, so that should be pretty doable in that interface.

You could probably write a script to spit out just the i686 package names, but it's complicated enough for me that I'd probably end up writing a 15 line python script to do it, because I know how to debug that when it doesn't work first of all, but in shell scripting instead I'd end up staring at it and comparing against other of my script to find I forgot to put a space or something between two things.

cx · 2019-09-07 19:51:24

levi wrote:

Re using less, hit '/' when in less then type 'i686' and enter. After reviewing that dodgy package, hit 'n' to go to the next hit (maybe twice if you went up from the previous hit). You should only have about 60 packages to find, so that should be pretty doable in that interface.

OK, these are the current i686 packages (60). As far as I remember, I installed them when my system was quite fresh and before I changed to pentium4. Also seemingly they weren’t updated since then:

acl
audit
db
dhcpcd
diffutils
fakeroot
filesystem
fontconfig
gc
gcc
gcc-libs
gmp
gpm
grep
groff
htop
hwinfo
inetutils
iputils
jfsutils
joe
libaio
libassuan
libedit
libid3tag
libidn2
libjpeg-turbo
libmpc
libnetfilter_conntrack
libssh2
libtiff
libutil-linux
libx11
libxcb
libxext
libxml2
libxpm
logrotate
lynx
lzop
mpfr
mtools
npth
ntp
openssl
pcmciautils
pkgfile
procps-ng
psmisc
sed
tar
texinfo
tzdata
util-linux
vi
w3m
which
xfsprogs
xz
zsh

I could of course reinstall them all in one go. Will this be safe enough to make them pentium4?

Next, I just sent a pkill X over ssh and then started a new X session on the pentium4. Further I’ve launched mousepad from a terminal, which produced following output:

** (mousepad:7691): WARNING **: 20:05:27.008: AT-SPI: Could not obtain desktop path or name
** (mousepad:7691): WARNING **: 20:05:27.053: atk-bridge: GetRegisteredEvents returned message with unknown signature
** (mousepad:7691): WARNING **: 20:05:27.054: atk-bridge: get_device_events_reply: unknown signature
** (mousepad:7691): WARNING **: 20:05:27.056: atk-bridge: get_device_events_reply: unknown signature

Despite they’re just warnings, are they of importance here? So far I haven’t found relevant hits on the net ...

<totally OT>
... but stumbled upon this one! Wow, Arch on a Kindle – too good to miss!

cx · 2019-09-07 20:27:37

It must have something to do with mousepad – because meanwhile the desktop is frozen again!

levi · 2019-09-07 20:29:18

Well, at least dhcpcd is available in pentium4, and plenty of others in your list look like things we've probably got (I've got a number of them installed, and I don't recall seeing them in my list of i686 things). I'd expect reinstalling it make it pentium4, just try it and see what happens!

Regarding Arch on a Kindle, let me know when it's booting into Arch not using proot because presumably with proot it's launching it from the KindleOS environment first. I mean, at least it's not Android, and on x86 systems none of us can really trust their factory installed BIOSes not to be installing any kind of hook into their system calls, but I don't really trust big media companies with my data any more.

Edit: Mousepad? That's a text editor isn't it? Are you usually running an instance of it when it freezes?

Last edited by levi (2019-09-07 20:31:06)

cx · 2019-09-07 20:36:40

levi wrote:

Edit: Mousepad? That's a text editor isn't it? Are you usually running an instance of it when it freezes?

Yes.

((Edit:))
Update: Started another X session, then mousepad again – desktop gets frozen once more about an hour later!
(Reinstalled all old i686 packages before, which have been turned in pentium4 ones.)

((Edit2:))

levi wrote:

Regarding Arch on a Kindle, let me know when it's booting into Arch not using proot because presumably with proot it's launching it from the KindleOS environment first. I mean, at least it's not Android, and on x86 systems none of us can really trust their factory installed BIOSes not to be installing any kind of hook into their system calls, but I don't really trust big media companies with my data any more.

NVM, I don’t even own a kindle (and probably never will), as I prefer real books and also don’t trust big media companies either.
I just got curious and inspired to try something more exotic like Arch for ARM.

Last edited by cx (2019-09-07 22:27:37)

cx · 2019-09-09 19:27:51

(I)

Well, that’s weird! – I thought, an /var/log/Xorg.0.log would be re-created at every X session, like on my Arch64:

ls -all
...
-rw-r--r--  1 root  users             31853  9. Sep 15:52 Xorg.0.log
...

Now compare with my Arch32:

ls -all
...
-rw-r--r--  1 root   root              28996 17. Aug 13:46 Xorg.0.log
...

– No wonder, it’s (a) always the same (b) rather looks like a proper X session shutdown to me, i.e. without any errors but many [II]s!
So (c) where to find my latest Xorg.0.log(s)?! And (d), "17. Aug 13:46" is the date when there was a big update and the desktop freezes have started ...

(II)

The desktop crashes are not limited to mousepad!

Now I got another one while composing text in abiword (in the means I wasn’t even able to input more text), and after killing and restarting X (via ssh), one more in a VTE (namely xfce4-terminal; still being able to type text)! At least you can count me lucky as the freezes occured pretty early, say after 5–15 minutes uptime in X. And again, systray and info bars are still displaying, the cursor can be moved, while executing things by mouse clicks or hotkeys doesn’t work.

I then re-discovered journald which repeatedly reveals plenty of repeated dbus-errors:

$ sudo journalctl | grep -i failed
...
Sep 09 15:20:30 arch32box org.gtk.vfs.Daemon[2610]: fusermount: failed to access mountpoint /run/user/1000/gvfs: Permission denied
Sep 09 15:20:30 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:30 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:30 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:30 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:31 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:31 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:31 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:31 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:31 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:32 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:32 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:32 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:32 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:32 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:32 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:32 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:32 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:32 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:32 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:33 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:33 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:33 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:33 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:33 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:33 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:33 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:33 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:33 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:33 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:39 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:39 klappbox dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:39 klappbox dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:39 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:39 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:39 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:40 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:40 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:40 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:40 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:40 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:40 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:41 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:41 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:41 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:41 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:41 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:41 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:41 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:41 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:20:42 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:20:42 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:21:20 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:21:20 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:21:21 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:21:21 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:21:21 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:21:21 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:21:21 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:21:21 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:21:33 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:21:33 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.
Sep 09 15:21:36 arch32box dbus-daemon[2610]: [session uid=1000 pid=2608] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Sep 09 15:21:36 arch32box dbus-broker-launch[2639]: Activation request for 'org.a11y.atspi.Registry' failed.

If I’m not mistaken, the "gvfs-denied permission" can be ignored, right? – But the dbus-stuff looks 'promising' though I don’t understand it in detail.

levi · 2019-09-09 19:51:07

I think it depends how you're starting X. I was surprised when I discovered my Xorg.0.log still has a date of 12 Dec 2016, but then I discovered it was in .local/share/xorg because I start my X as a local user after logging in to the terminal. I expect if I still used a display manager handled by a systemd unit, it would still be writing to /var/log, but that was my assuption I guess.

If I remember rightly, my launch of startx tells me where it's logging to but only seconds before it launches the graphical service over the top. I could configure it to use a different vtty, I guess, but in the mean time I didn't properly notice that log for a few months.

I'd guess the dying of dbus could cause many of the issues you're experiencing. Are there any logs related to that? Try 'journalctl -b -u dbus' for a more concise log. I forget but you're running xfce right? IIRC that depends rather significantly on dbus.

cx · 2019-09-09 20:46:08

levi wrote:

... .local/share/xorg ... I expect if I still used a display manager handled by a systemd unit, it would still be writing to /var/log, but that was my assuption I guess.

Awesome!

While in first time I used lightdm for testing alternative WMs, eventually I’ve started X via tty/startx. – OK then, I’m still connected to the frozen desktop machine. Its output of ~/.local/share/xorg/Xorg.0.log with its final lines including the only error reads like this:

...
[  7557.589] (II) event5  - SIGMACHIP Usb Mouse: Enabling spurious button debouncing, see https://wayland.freedesktop.org/libinput/doc/1.14.1/button-debouncing.html for details
[  8041.449] (EE) client bug: timer event5 debounce short: offset negative (-1ms)

And that’s another thing: I wasn’t aware about wayland on my machine – I don’t use it! I will now uninstall it and see, if that "event5 bug" has gone; the best I what can do for now, I suppose. I mean, I don’t get what they’re talking about in wayland ... button-debouncing exactly, but they mention "libinput" all the time. – Stay tuned!

For completion, according to the journal everything’s fine regarding dbus:

$ journalctl -b -u dbus
-- Logs begin at Thu 2019-07-25 18:07:07 CEST, end at Mon 2019-09-09 21:08:01 CEST. --
-- No entries --

Edit:
There’s another file ~/.xsession-errors, dated "17. Aug 13:20" – which might or might not be of interest here – containing:

xfce4-session: error while loading shared libraries: libxfconf-0.so.2: cannot open shared object file: No such file or directory

xfce4-session has been updated anyway; current version is 4.14.0-2.2, September 3rd. – Just FYI.

Last edited by cx (2019-09-09 21:24:49)

cx · 2019-09-09 23:00:34

Boy, this sucks~
After getting rid of wayland (also xorg-xserver-xwayland and wayland-protocols), I can’t start X anymore! Or better: X is up for a second, then closing down:

xfce4-session: error while loading shared libraries: libwayland-client.o.0: cannot open shared object file: No such file or directory

What the heck has Xfce to do with wayland?! I removed ~/.cache/sessions/*, reinstalled xfce4-session, logged out, rebooted – to no avail ...

levi · 2019-09-09 23:03:16

cx wrote:

For completion, according to the journal everything’s fine regarding dbus:
$ journalctl -b -u dbus
-- Logs begin at Thu 2019-07-25 18:07:07 CEST, end at Mon 2019-09-09 21:08:01 CEST. --
-- No entries --

I didn't mean for you to not still need to sudo it. Mine for comparison is:

-- Logs begin at Sat 2019-05-18 16:47:02 BST, end at Mon 2019-09-09 19:42:59 BST>
Sep 09 16:39:32 lynx systemd[1]: Started D-Bus System Message Bus.
Sep 09 16:39:57 lynx dbus-daemon[341]: [system] Activating via systemd: service >
Sep 09 16:39:57 lynx dbus-daemon[341]: [system] Successfully activated service '>
Sep 09 16:39:57 lynx dbus-daemon[341]: [system] Activating via systemd: service >
Sep 09 16:39:58 lynx dbus-daemon[341]: [system] Successfully activated service '>

I'd expect yours to at least start with similar, and end with errors.

Edit: I can't help with wayland errors. I have it installed as well, although I do recall trying it out in the past, and may not have removed it afterwards. According to pacman -Qo anyway, that particular lib is owned by the unqualified wayland package.

cx · 2019-09-09 23:12:31

I see. Nonetheless no changes for me:

$ sudo journalctl -b -u dbus
[sudo] Password for root: 
-- Logs begin at Thu 2019-07-25 18:07:07 CEST, end at Tue 2019-09-10 00:08:00 CEST. --
-- No entries --

cx · 2019-09-10 00:56:08

Installed wayland again, and got X back at least.

And meanwhile I’ve gained insight playing with libinput debug-gui as root, when it happened the graphical testing window crashed and the desktop froze once I pressed <ESC> (I think). – I’ll be after it again later as I really need going to sleep now!

levi · 2019-09-10 06:08:59

cx wrote:

I see. Nonetheless no changes for me:

$ sudo journalctl -b -u dbus
[sudo] Password for root: 
-- Logs begin at Thu 2019-07-25 18:07:07 CEST, end at Tue 2019-09-10 00:08:00 CEST. --
-- No entries --

Oh, are you already logged in as root. Usual convention is to present root commands with the root prompt (#) rather than dollar ($) when reporting back.

I dunno why you aren't getting any dbus logs. You do have the dbus package installed I assume? It's a dependency of weston and systemd so I'd assume you have. Is it still running pre-crash? Test with sudo ps -ef|grep dbus

cx · 2019-09-10 13:04:17

levi wrote:

Oh, are you already logged in as root. Usual convention is to present root commands with the root prompt (#) rather than dollar ($) when reporting back.

No, I was not logged in as root when I did the sudo command – hence the "$" prompt.

However, the libinput’s window "crashing" on <ESC> I reported last night was normal behaviour. – Besides, today I caught just some more freezes by pressing <F9>, <ß>, or <F11> within different X sessions, in xfce4-terminal and abiword. In other words, all of those keys (and probably others) work normally until by whatever reason they don’t and make the desktop freeze when pressing.

These are the current live logs from

# libinput debug-gui

info: event2 Power Button                   added
info: event3 Video Bus                      added
info: event0 Lid Switch                     added
info: event1 Sleep Button                   added
info: event5 SIGMACHIP Usb Mouse            added
info: event4 AT Translated Set 2 keyboard   added
info: event7 TPPS/2 IBM TrackPoint          added
info: event8 ThinkPad Extra Buttons         added

** (libinput-debug-gui:3825): WARNING **: 13:02:32.129: AT-SPI: Could not obtain desktop path or name


** (libinput-debug-gui:3825): WARNING **: 13:02:32.221: atk-bridge: GetRegisteredEvents returned message with unknown signature

** (libinput-debug-gui:3825): WARNING **: 13:02:32.229: atk-bridge: get_device_events_reply: unknown signature

** (libinput-debug-gui:3825): WARNING **: 13:02:32.241: atk-bridge: get_device_events_reply: unknown signature
client bug: timer event5 debounce: offset negative (-15ms)
client bug: timer event5 debounce short: offset negative (-28ms)
client bug: timer event5 debounce: offset negative (-201ms)
client bug: timer event5 debounce: offset negative (-137ms)
client bug: timer event5 debounce short: offset negative (-150ms)
SIGMACHIP Usb Mouse: speed is 0,10
TPPS/2 IBM TrackPoint: speed is 0,10
client bug: timer event5 debounce: offset negative (-4ms)
client bug: timer event5 debounce short: offset negative (-17ms)
client bug: timer event5 debounce: offset negative (-143ms)
client bug: timer event5 debounce short: offset negative (-156ms)
SIGMACHIP Usb Mouse: speed is 0,20
TPPS/2 IBM TrackPoint: speed is 0,20
client bug: timer event5 debounce short: offset negative (-9ms)
client bug: timer event5 debounce: offset negative (-3ms)
client bug: timer event5 debounce short: offset negative (-16ms)
client bug: timer event5 debounce: offset negative (-54ms)
client bug: timer event5 debounce short: offset negative (-67ms)
client bug: timer event5 debounce short: offset negative (-0ms)
client bug: timer event5 debounce: offset negative (-163ms)
client bug: timer event5 debounce short: offset negative (-176ms)
client bug: timer event5 debounce: offset negative (-136ms)
client bug: timer event5 debounce short: offset negative (-149ms)
SIGMACHIP Usb Mouse: speed is 0,30
TPPS/2 IBM TrackPoint: speed is 0,30
client bug: timer event5 debounce: offset negative (-4ms)
client bug: timer event5 debounce short: offset negative (-17ms)
client bug: timer event5 debounce short: offset negative (-2ms)
client bug: timer event5 debounce: offset negative (-49ms)
client bug: timer event5 debounce short: offset negative (-62ms)

Quitting debugging tool with <ESC>; desktop still working normally.

At first sight of the inconsistencies, the "client bug" lagging is less severe than those warnings failed to determining things though ...

levi wrote:

Test with sudo ps -ef|grep dbus

Here you are:

dbus       384     1  0 12:21 ?        00:00:00 /usr/bin/dbus-broker-launch --scope system --audit
dbus       385   384  0 12:21 ?        00:00:01 dbus-broker --log 4 --controller 9 --machine-id 7f35667b76ac472fa3450e6e13ebc0db --max-bytes 536870912 --max-fds 4096 --max-matches 131072 --audit
cx         647   541  0 12:22 ?        00:00:00 /usr/bin/dbus-daemon --session --address=systemd: --nofork --nopidfile --systemd-activation --syslog-only
cx         788   783  0 12:22 ?        00:00:00 /usr/bin/dbus-broker-launch --config-file=/usr/share/defaults/at-spi2/accessibility.conf --scope user
cx         789   788  0 12:22 ?        00:00:00 dbus-broker --log 4 --controller 9 --machine-id 7f35667b76ac472fa3450e6e13ebc0db --max-bytes 100000000000000 --max-fds 6400000 --max-matches 5000000000
cx        3263     1  0 13:01 tty1     00:00:00 /usr/bin/dbus-launch --sh-syntax --exit-with-session xfce4-session
cx        3264     1  0 13:01 ?        00:00:00 /usr/bin/dbus-daemon --syslog --fork --print-pid 4 --print-address 6 --session
cx        3294  3289  0 13:01 ?        00:00:00 /usr/bin/dbus-broker-launch --config-file=/usr/share/defaults/at-spi2/accessibility.conf --scope user
cx        3295  3294  0 13:01 ?        00:00:00 dbus-broker --log 4 --controller 9 --machine-id 7f35667b76ac472fa3450e6e13ebc0db --max-bytes 100000000000000 --max-fds 6400000 --max-matches 5000000000

cx · 2019-09-10 14:09:06

Maybe the issue is related to this one: [(...) communication error with X].

I’m off to work now, see you later.

levi · 2019-09-10 20:22:55

I don't get those client bug issues when I just played about with libinput, but I don't see of the top of my head why these would cause dbus to crash. It's still frustrating that you can't get normal looking dbus logs even, and I don't know what's up with that. Could you perhaps run that sudo ps -ef command after an X crash to confirm that dbus has quit, so we can be sure a dbus exit is part of the problem at least?

cx · 2019-09-10 21:13:38

I wouldn’t call it "X crash" but rather an 'input in X crash'.

The symptoms described in this bug report sound pretty much the same: libinput problem, random freezes, unresponsive keyboard and mouse, client bug timer event debounces. Further they seem independent from the running DE/WM, or GPU. – Just too sad they only guessed the bug has gone without telling how it got fixed.

Then it seems there’s nothing wrong with dbus; I’ll post another "sudo ps -ef" anyway.

And thanks for staying with me!

cx · 2019-09-10 22:28:03

Here you are (as usual via ssh; crashed after testing <Fn> combinations):

$ sudo ps -ef|grep dbus
[sudo] Passwort für root:
dbus       393     1  0 19:16 ?        00:00:00 /usr/bin/dbus-broker-launch --scope system --audit
dbus       396   393  0 19:16 ?        00:00:03 dbus-broker --log 4 --controller 9 --machine-id 7f35667b76ac472fa3450e6e13ebc0db --max-bytes 536870912 --max-fds 4096 --max-matches 131072 --audit
cx        2539   565  0 22:10 ?        00:00:00 /usr/bin/dbus-daemon --session --address=systemd: --nofork --nopidfile --systemd-activation --syslog-only
cx        2600     1  0 22:10 tty1     00:00:00 /usr/bin/dbus-launch --sh-syntax --exit-with-session xfce4-session
cx        2601     1  0 22:10 ?        00:00:00 /usr/bin/dbus-daemon --syslog --fork --print-pid 4 --print-address 6 --session
cx        2611  2606  0 22:10 ?        00:00:00 /usr/bin/dbus-broker-launch --config-file=/usr/share/defaults/at-spi2/accessibility.conf --scope user
cx        2612  2611  0 22:10 ?        00:00:00 dbus-broker --log 4 --controller 9 --machine-id 7f35667b76ac472fa3450e6e13ebc0db --max-bytes 100000000000000 --max-fds 6400000 --max-matches 5000000000

levi · 2019-09-10 23:08:45

Okay, I still don't understand why it's asking you for the root password if you claim to be logging on as a user. That's not how sudo works whenever I've used it, but I only poke at it until it works for me, so maybe it's working in a mode I've never used.

But in any case, it seems those dbus errors aren't fatal. They're still curious to me, and look likely to have some adverse effects. but what those effects are are hard to say. That 15:20 time; is that when you booted the machine or when it stopped responding properly, or some other time?

You may be right its a libinput related problem rather than a dbus error. I have libinput and xf86-input-libinput installed from doing a 'pacman -Qs libinput' (doesn't need sudoing).

cx · 2019-09-10 23:40:24

levi wrote:

Okay, I still don't understand why it's asking you for the root password if you claim to be logging on as a user. That's not how sudo works whenever I've used it, but I only poke at it until it works for me, so maybe it's working in a mode I've never used.

Erm, if I was already root I wouldn’t need to sudo (with or without password), right? With respect, I don’t get your point.

$ pacman -Qs libinput

local/libinput 1.14.1-1.0
    Input device management and event handling library
local/xf86-input-libinput 0.29.0-1.0 (xorg-drivers)
    Generic input driver for the X.Org server based on libinput

And you’re welcome, I know it works without sudo.

What "15:20 time" btw?

cx · 2019-09-11 00:48:25

Looks like I can now reproduce the freeze – at least it happens pressing the <Fn> key (alone)!

Due to the hanging xev is useless here. If I remember correctly, that key once was mapped as <Insert>. – On my Arch64 machine in comparison, <Fn> itself gets ignored by xev, but of course does not cause a freeze; I’d expect this to be normal.

This fellow Archer experienced similar symptoms also on a Thinkpad – but he was lucky as the problem somehow vanished.

I consider my Arch32 to be uptodate though, so maybe it’s an input driver clash, which requires extra configuration.

Let’s see, I will work out these wikis tomorrow: [ 1 ] [ 2 ].

Edit: @levi, as this bothers you so much, I’ll omit the password line from now on. I just copied it from the terminal, but in the end it’s the same.

Last edited by cx (2019-09-11 01:07:31)

levi · 2019-09-11 05:27:49

cx wrote:

levi wrote:
Okay, I still don't understand why it's asking you for the root password if you claim to be logging on as a user. That's not how sudo works whenever I've used it, but I only poke at it until it works for me, so maybe it's working in a mode I've never used.
Erm, if I was already root I wouldn’t need to sudo (with or without password), right? With respect, I don’t get your point.

Just this:

$ sudo journalctl -b -u dbus
[sudo] Password for root:

On my system I get

[sudo] Password for [username]:

Under that I can enter my user password.

What "15:20 time" btw?

The time from this post

cx · 2019-09-11 11:16:46

levi wrote:

...
Just this:
$ sudo journalctl -b -u dbus
[sudo] Password for root:
On my system I get
[sudo] Password for [username]:
Under that I can enter my user password.

You’re welcome, I prefer targetpw though.

levi wrote:

What "15:20 time" btw?
The time from this post

Then that was the output until the desktop froze.

Last edited by cx (2019-09-12 01:08:58)

Arch Linux

#51 2019-09-07 07:21:39

Re: [pending] Desktop freezing occasionally (pentium4)

#52 2019-09-07 17:07:15

Re: [pending] Desktop freezing occasionally (pentium4)

#53 2019-09-07 18:02:43

Re: [pending] Desktop freezing occasionally (pentium4)

#54 2019-09-07 19:51:24

Re: [pending] Desktop freezing occasionally (pentium4)

#55 2019-09-07 20:27:37

Re: [pending] Desktop freezing occasionally (pentium4)

#56 2019-09-07 20:29:18

Re: [pending] Desktop freezing occasionally (pentium4)

#57 2019-09-07 20:36:40

Re: [pending] Desktop freezing occasionally (pentium4)

#58 2019-09-09 19:27:51

Re: [pending] Desktop freezing occasionally (pentium4)

#59 2019-09-09 19:51:07

Re: [pending] Desktop freezing occasionally (pentium4)

#60 2019-09-09 20:46:08

Re: [pending] Desktop freezing occasionally (pentium4)

#61 2019-09-09 23:00:34

Re: [pending] Desktop freezing occasionally (pentium4)

#62 2019-09-09 23:03:16

Re: [pending] Desktop freezing occasionally (pentium4)

#63 2019-09-09 23:12:31

Re: [pending] Desktop freezing occasionally (pentium4)

#64 2019-09-10 00:56:08

Re: [pending] Desktop freezing occasionally (pentium4)

#65 2019-09-10 06:08:59

Re: [pending] Desktop freezing occasionally (pentium4)

#66 2019-09-10 13:04:17

Re: [pending] Desktop freezing occasionally (pentium4)

#67 2019-09-10 14:09:06

Re: [pending] Desktop freezing occasionally (pentium4)

#68 2019-09-10 20:22:55

Re: [pending] Desktop freezing occasionally (pentium4)

#69 2019-09-10 21:13:38

Re: [pending] Desktop freezing occasionally (pentium4)

#70 2019-09-10 22:28:03

Re: [pending] Desktop freezing occasionally (pentium4)

#71 2019-09-10 23:08:45

Re: [pending] Desktop freezing occasionally (pentium4)

#72 2019-09-10 23:40:24

Re: [pending] Desktop freezing occasionally (pentium4)

#73 2019-09-11 00:48:25

Re: [pending] Desktop freezing occasionally (pentium4)

#74 2019-09-11 05:27:49

Re: [pending] Desktop freezing occasionally (pentium4)

#75 2019-09-11 11:16:46

Re: [pending] Desktop freezing occasionally (pentium4)

Board footer