Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2019-01-02 Thread Ian Campbell
On Wed, 2019-01-02 at 15:43 +, Ian Campbell wrote:
> On Wed, 2019-01-02 at 10:32 -0500, Ted To wrote:
> > My experience was that a reboot allowed qcontrol to be installed but it 
> > certainly makes sense that reloading devices could fix the install as 
> > well.  I'm willing to help test this -- would I install the stretch 
> > qcontrol (not backported), reboot and then try to install the backported 
> > qcontrol?
> 
> I think so, yes.

Actually, I think I now have enough of a handle on what is going on to
figure this out from here (or at least I don't need more input right
now). Thanks for the offer though!

Ian.



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2019-01-02 Thread Ian Campbell
On Wed, 2019-01-02 at 10:32 -0500, Ted To wrote:
> My experience was that a reboot allowed qcontrol to be installed but it 
> certainly makes sense that reloading devices could fix the install as 
> well.  I'm willing to help test this -- would I install the stretch 
> qcontrol (not backported), reboot and then try to install the backported 
> qcontrol?

I think so, yes. After installing the stretch qcontrol and rebooting
you should run:

udevadm info -q all /dev/input/by-path/platform-gpio_keys-event

And confirm that it does _not_ have a line like:

E: TAGS=:power-switch:systemd:

specifically either there should be no TAGS line or if there is one it
should not have "systemd" in it. I think that correctly replicates the
state before the failing update.

I guess you've have seen it while I was writing this mail, but note my
follow up where I decided that the correct command was:

udevadm trigger -s platform -y gpio_keys && dpkg --configure -a

Thanks!

Ian.



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2019-01-02 Thread Ted To
My experience was that a reboot allowed qcontrol to be installed but it 
certainly makes sense that reloading devices could fix the install as 
well.  I'm willing to help test this -- would I install the stretch 
qcontrol (not backported), reboot and then try to install the backported 
qcontrol?


On 1/2/19 10:21 AM, Ian Campbell wrote:

On Wed, 2019-01-02 at 15:14 +, Ian Campbell wrote:

But after booting the buster kernel qcontrol could setup
successfully:


I wonder if it is the reboot itself, and not necessarily the switch to
the buster kernel which is the bandaid here, i.e. if rebooting back
into the stretch kernel would work too.


In fact I also wonder if:
 udevadm control --reload-rules && dpkg --configure -a

would do the trick, indicating that the reload is what is missing.





Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2019-01-02 Thread Ian Campbell
On Wed, 2019-01-02 at 15:21 +, Ian Campbell wrote:
> On Wed, 2019-01-02 at 15:14 +, Ian Campbell wrote:
> > > But after booting the buster kernel qcontrol could setup
> > > successfully:
> > 
> > I wonder if it is the reboot itself, and not necessarily the switch
> > to
> > the buster kernel which is the bandaid here, i.e. if rebooting back
> > into the stretch kernel would work too.
> 
> In fact I also wonder if:
> udevadm control --reload-rules && dpkg --configure -a
> 
> would do the trick, indicating that the reload is what is missing.

Probably not, according to [0] which mentions that udev monitors the
directory and does so automatically (which is what I expected).

That mail does mention `udevadm trigger` to apply new rules to existing
devices though -- that does sound like it might apply here, so I wonder
if:

udevadm trigger -s platform -y gpio_keys && dpkg --configure -a

Does what is required... I have a strong feeling it will and so it
should be done from the qcontrol postinst.

Ian.

[0] https://lists.debian.org/debian-devel/2018/06/msg00047.html 



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2019-01-02 Thread Ian Campbell
On Wed, 2019-01-02 at 15:14 +, Ian Campbell wrote:
> > But after booting the buster kernel qcontrol could setup
> > successfully:
> 
> I wonder if it is the reboot itself, and not necessarily the switch to
> the buster kernel which is the bandaid here, i.e. if rebooting back
> into the stretch kernel would work too.

In fact I also wonder if:
udevadm control --reload-rules && dpkg --configure -a

would do the trick, indicating that the reload is what is missing.



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2019-01-02 Thread Ian Campbell
On Mon, 2018-12-31 at 11:51 +0100, Bernhard Übelacker wrote:
> There I can confirm the installation of the qcontrol package
> failed while running the stretch kernel:

Thanks, unfortunately I would need the same set of things I asked Ted
for in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=914160#27 to
really be able to start to determine what is going on here. Are you
able to gather any of that while the system is in this broken halfway
state?

> But after booting the buster kernel qcontrol could setup
> successfully:

I wonder if it is the reboot itself, and not necessarily the switch to
the buster kernel which is the bandaid here, i.e. if rebooting back
into the stretch kernel would work too.


Ian



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2018-12-31 Thread Bernhard Übelacker
Hello Ian, hello Ted,
yesterday I tried to reinstall that device with the latest
buster installer - but unfortunately I was already too late,
the archive moved already too far away:
No kernel modules were found.

Therefore I did the reinstall with the stretch installer
and upgraded to buster, like in my last attempt from november.

There I can confirm the installation of the qcontrol package
failed while running the stretch kernel:
(linux-image-4.9.0-8-marvell:armel 4.9.130-2)

apt dist-upgrade
...
Setting up qcontrol (0.5.6-2) ...
Installing new version of config file /etc/qcontrol/ts209.lua ...
Installing new version of config file /etc/qcontrol/ts219.lua ...
Installing new version of config file /etc/qcontrol/ts41x.lua ...
update-initramfs: deferring update (trigger activated)
Created symlink 
/etc/systemd/system/multi-user.target.wants/qcontrol.service → 
/lib/systemd/system/qcontrol.service.
Created symlink 
/etc/systemd/system/multi-user.target.wants/qcontrold.service → 
/lib/systemd/system/qcontrold.service.
Created symlink /etc/systemd/system/sockets.target.wants/qcontrold.socket → 
/lib/systemd/system/qcontrold.socket.
A dependency job for qcontrold.service failed. See 'journalctl -xe' for 
details.
invoke-rc.d: initscript qcontrold, action "start" failed.
● qcontrold.service - qcontrold
Loaded: loaded (/lib/systemd/system/qcontrold.service; enabled; vendor 
preset: enabled)
Active: inactive (dead)

Dez 31 02:20:17 qnap-119p-ii systemd[1]: Dependency failed for qcontrold.
Dez 31 02:20:17 qnap-119p-ii systemd[1]: qcontrold.service: Job 
qcontrold.service/start failed with result 'dependency'.
dpkg: error processing package qcontrol (--configure):
installed qcontrol package post-installation script subprocess returned 
error exit status 1
Setting up libunbound8:armel (1.8.1-1+b1)
...


journalctl
Dez 31 02:20:17 qnap-119p-ii systemd[1]: 
dev-input-by\x2dpath-platform\x2dgpio_keys\x2devent.device: Job 
dev-input-by\x2dpath-platform\x2dgpio_keys\x2devent.device/start timed out.
Dez 31 02:20:17 qnap-119p-ii systemd[1]: Timed out waiting for device 
dev-input-by\x2dpath-platform\x2dgpio_keys\x2devent.device.
Dez 31 02:20:17 qnap-119p-ii systemd[1]: Dependency failed for qcontrold.
Dez 31 02:20:17 qnap-119p-ii systemd[1]: qcontrold.service: Job 
qcontrold.service/start failed with result 'dependency'.
Dez 31 02:20:17 qnap-119p-ii systemd[1]: 
dev-input-by\x2dpath-platform\x2dgpio_keys\x2devent.device: Job 
dev-input-by\x2dpath-platform\x2dgpio_keys\x2devent.device/start failed with 
result 


But after booting the buster kernel qcontrol could setup successfully:

root@qnap-119p-ii:~# apt install -f
Paketlisten werden gelesen... Fertig
Abhängigkeitsbaum wird aufgebaut.   
Statusinformationen werden eingelesen Fertig
Die folgenden Pakete wurden automatisch installiert und werden nicht mehr 
benötigt:
dh-python libbind9-140 libdns162 libicu57 libisc160 libisccc140 
libisccfg140 liblwres141 libperl5.24 libpython3.5-minimal libpython3.5-stdlib 
python3-distutils python3-lib2to3 python3.5
python3.5-minimal rename sgml-base xml-core
Verwenden Sie »apt autoremove«, um sie zu entfernen.
0 aktualisiert, 0 neu installiert, 0 zu entfernen und 0 nicht aktualisiert.
1 nicht vollständig installiert oder entfernt.
Nach dieser Operation werden 0 B Plattenplatz zusätzlich benutzt.
qcontrol (0.5.6-2) wird eingerichtet ...
update-initramfs: deferring update (trigger activated)
Trigger für initramfs-tools (0.132) werden verarbeitet ...
update-initramfs: Generating /boot/initrd.img-4.19.0-1-marvell
kirkwood-qnap: machine: QNAP TS219 family
Using DTB: kirkwood-ts219-6282.dtb
Installing /usr/lib/linux-image-4.19.0-1-marvell/kirkwood-ts219-6282.dtb 
into /boot/dtbs/4.19.0-1-marvell/./kirkwood-ts219-6282.dtb
Taking backup of kirkwood-ts219-6282.dtb.
Installing new kirkwood-ts219-6282.dtb.
Installing /usr/lib/linux-image-4.19.0-1-marvell/kirkwood-ts219-6282.dtb 
into /boot/dtbs/4.19.0-1-marvell/./kirkwood-ts219-6282.dtb
Taking backup of kirkwood-ts219-6282.dtb.
Installing new kirkwood-ts219-6282.dtb.
flash-kernel: installing version 4.19.0-1-marvell
flash-kernel: appending 
/usr/lib/linux-image-4.19.0-1-marvell/kirkwood-ts219-6282.dtb to kernel
Generating kernel u-boot image... done.
Flashing kernel (using 2047906/2097152 bytes)... done.
Flashing initramfs (using 4477212/9437184 bytes)... done.

Kind regards,
Bernhard



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2018-12-22 Thread Ian Campbell
On Sat, 2018-12-22 at 08:00 -0500, Ted To wrote:
> After reverting it and  rebooting everything is now working fine.

Excellent, thanks for letting me know.

Ian.



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2018-12-22 Thread Ted To

On 12/21/18 7:12 AM, Ian Campbell wrote:

On Fri, 2018-12-21 at 06:25 -0500, Ted To wrote:

Still broken for me.  I used Bernhard's workaround and qcontrol
starts.


Thank you for letting me know. Unfortunately this works for me locally
so I need more information from an affected system with no workarounds
applied to diagnose.

Please can you revert whatever edits you made to workaround the issue
(e.g. put the udev and systemd files back to their default/packaged
state and remove any override files you might have added to /etc/,
likewise revert the qcontrol.conf settings to the default).


Thanks for the response Ian.  It seems that what was missing was a 
reboot although part of the problem was that qcontrol failed to 
completely install because of the error.  What I did to get qcontrol to 
completely install was Bernhard's workaround.  After reverting it and 
rebooting everything is now working fine.


Thank you,
Ted



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2018-12-21 Thread Ian Campbell
On Fri, 2018-12-21 at 06:25 -0500, Ted To wrote:
> Still broken for me.  I used Bernhard's workaround and qcontrol
> starts.

Thank you for letting me know. Unfortunately this works for me locally
so I need more information from an affected system with no workarounds
applied to diagnose.

Please can you revert whatever edits you made to workaround the issue
(e.g. put the udev and systemd files back to their default/packaged
state and remove any override files you might have added to /etc/,
likewise revert the qcontrol.conf settings to the default).

With that done please reboot to ensure everything has been correctly
reloaded and then post the outputs of all of the following commands  to
this bug (all of which will need to be run as root, some are quite
verbose and some of which will error out, but please include
everything):

uname -a
ls -lRt /dev/input/
udevadm info -q all /dev/input/by-path/platform-gpio_keys-event
systemctl status /dev/input/by-path/platform-gpio_keys-event
systemctl status dev-input-by\x2dpath-platform\x2dgpio_keys\x2devent.device
systemctl status qcontrol.service
systemctl status qcontrold.socket
systemctl status qcontrold.service
grep . /{lib,etc}/udev/rules.d/*qcontrol*
grep . /{lib,etc}/systemd/system/*qcontrol*

Thanks,
Ian.



Bug#914160: Bug#912376: fixed in qcontrol 0.5.6-2

2018-12-21 Thread Ted To

Still broken for me.  I used Bernhard's workaround and qcontrol starts.

Version information:

$ apt show qcontrol
Package: qcontrol
Version: 0.5.6-2~bpo9+1
Priority: optional
Section: utils
Maintainer: Ian Campbell 
Installed-Size: 103 kB
Depends: libc6 (>= 2.15), liblua5.1-0, init-system-helpers (>= 1.18~), udev
Homepage: https://www.hellion.org.uk/qcontrol/
Download-Size: 25.6 kB
APT-Manual-Installed: yes
APT-Sources: http://ftp.us.debian.org/debian stretch-backports/main 
armel Packages


On Sun, 09 Dec 2018 15:20:30 + Ian Campbell  wrote:

Source: qcontrol
Source-Version: 0.5.6-2

We believe that the bug you reported is fixed in the latest version of
qcontrol, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 912...@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Ian Campbell  (supplier of updated qcontrol package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmas...@ftp-master.debian.org)


-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Format: 1.8
Date: Sun, 09 Dec 2018 14:42:37 +
Source: qcontrol
Binary: qcontrol qcontrol-udeb
Architecture: source
Version: 0.5.6-2
Distribution: unstable
Urgency: medium
Maintainer: Ian Campbell 
Changed-By: Ian Campbell 
Description:
 qcontrol   - hardware control for QNAP Turbo Station devices
 qcontrol-udeb - hardware control for QNAP Turbo Station devices (udeb)
Closes: 908525 912376
Changes:
 qcontrol (0.5.6-2) unstable; urgency=medium
 .
   [ Ondřej Nový ]
   * d/changelog: Remove trailing whitespaces
 .
   [ Ian Campbell ]
   * Switch Homepage to https.
   * Fix "qcontrol FTCBFS: upstream Makefile hard codes the build
 architecture pkg-config" by allowing the override of the pkg-config binary
 Thanks to Helmut Grohne for the initial patch (Closes: #908525).
   * Correctly install 60-qcontrol.rules (Closes: #912376).
   * Refresh all patches to apply without fuzz.
   * Catch more installation errors in debian/rules.
Checksums-Sha1:
 70a3b9b459708b73436443539d8be74db2d5c44b 1983 qcontrol_0.5.6-2.dsc
 4381fac43f1801f57f690126ba945bd0f693d1e2 19236 qcontrol_0.5.6-2.debian.tar.xz
 7026d800bc091a301a8ff16b3d8e5ab4716a8c60 5888 qcontrol_0.5.6-2_source.buildinfo
Checksums-Sha256:
 f9a9501b9cf603c155a4f939d8f6d3147cc07e6dddc0512de8047baa21a14cb2 1983 
qcontrol_0.5.6-2.dsc
 565dc48adc871653790632da7ccd3f8576c6d00a29bfaff631fd662fc4e851cc 19236 
qcontrol_0.5.6-2.debian.tar.xz