Debian Bug report logs - #835648
wpasupplicant: suspend takes very long due to problematic system-sleep hook

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Tomas Janousek <tomi@nomi.cz>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Sun, 28 Aug 2016 00:39:36 +0200

Package: wpasupplicant
Version: 2.5-2+v2.4-2
Severity: normal

I recently upgraded the wpasupplicant package and my system started taking a
LONG time to suspend. My journal contains:

Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.2710] manager: sleep requested (sleeping: no  enabled: yes)
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.2710] manager: sleeping...
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.2711] device (eth0): state change: unavailable -> unmanaged (reason 'sleeping') [20 10 37]
Aug 27 23:40:53 notes kernel: e1000e: eth0 NIC Link is Down
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.3941] manager: NetworkManager state is now ASLEEP
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.3974] device (wlan1): state change: activated -> deactivating (reason 'sleeping') [100 110 37]
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4048] device (wlan1): state change: deactivating -> disconnected (reason 'sleeping') [110 30 37]
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4232] dhcp4 (wlan1): canceled DHCP transaction, DHCP client pid 18447
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4233] dhcp4 (wlan1): state changed bound -> done
Aug 27 23:40:53 notes kernel: wlan1: deauthenticating from c4:27:95:77:a6:b8 by local choice (Reason: 3=DEAUTH_LEAVING)
Aug 27 23:40:53 notes wpa_supplicant[2812]: wlan1: CTRL-EVENT-DISCONNECTED bssid=c4:27:95:77:a6:b8 reason=3 locally_generated=1
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4548] dns-mgr: Writing DNS information to /sbin/resolvconf
Aug 27 23:40:53 notes wpa_supplicant[2812]: wlan1: CTRL-EVENT-REGDOM-CHANGE init=CORE type=WORLD
Aug 27 23:40:53 notes NetworkManager[1904]: <warn>  [1472334053.5079] sup-iface[0x9947a98,wlan1]: connection disconnected (reason -3)
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.5080] device (wlan1): supplicant interface state: completed -> disconnected
Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.5091] device (wlan1): state change: disconnected -> unmanaged (reason 'sleeping') [30 10 37]
Aug 27 23:40:53 notes systemd[1]: Reached target Sleep.
Aug 27 23:40:53 notes systemd[1]: Starting Suspend...
Aug 27 23:40:54 notes wpa_supplicant[2812]: nl80211: deinit ifname=wlan1 disabled_11b_rates=0
Aug 27 23:41:04 notes systemd-sleep[18924]: Selected interface 'wlan1'
Aug 27 23:41:04 notes systemd-sleep[18924]: 'SUSPEND' command timed out.
Aug 27 23:41:04 notes systemd-sleep[18928]: /lib/systemd/system-sleep/wpasupplicant failed with error code 254.
Aug 27 23:41:04 notes systemd-sleep[18924]: Suspending system...

Disabling the /lib/systemd/system-sleep/wpasupplicant script fixes the problem
for me. From what NetworkManagers prints to the journal it seems to me that
the /lib/systemd/system-sleep/wpasupplicant script is unnecessary, and since
it makes suspend take 10 seconds longer, it is actually not just unnecessary,
but harmful.

I suspend the system by invoking "systemctl suspend" via acpid.

I tried to research this a bit. If I manually invoke

    dbus-send --print-reply --system --dest=org.freedesktop.NetworkManager /org/freedesktop/NetworkManager org.freedesktop.NetworkManager.Sleep boolean:true

and then try to "/sbin/wpa_cli suspend", I always get:

    Failed to connect to non-global ctrl_ifname: (nil)  error: No such file or directory

Seems to me then that the 10s delay I'm getting is a race condition between NM
sleeping and wpa_cli attempting to suspend. It's also interesting that on my
hardware (ThinkPad T420, Intel(R) Centrino(R) Advanced-N 6205 AGN, REV=0xB0)
wpa_cli suspend does nothing at all -- after invoking it and getting OK, wifi
stays connected and pings.

I also looked into dbus-monitor output and in my case NM drops the inhibitor
only after putting both eth and wlan devices offline, unlike in the launchpad
issue that the /lib/systemd/system-sleep/wpasupplicant is trying to fix.

-- System Information:
Debian Release: stretch/sid
  APT prefers testing
  APT policy: (980, 'testing'), (980, 'stable'), (500, 'unstable-debug'), (500, 'testing-debug'), (500, 'unstable'), (500, 'stable'), (200, 'experimental'), (1, 'experimental-debug')
Architecture: i386 (x86_64)
Foreign Architectures: amd64

Kernel: Linux 4.4.19-lis64+ (SMP w/4 CPU cores; PREEMPT)
Locale: LANG=cs_CZ.UTF-8, LC_CTYPE=cs_CZ.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash
Init: systemd (via /run/systemd/system)

Versions of packages wpasupplicant depends on:
ii  adduser           3.115
ii  libc6             2.23-4
ii  libdbus-1-3       1.10.10-1
ii  libnl-3-200       3.2.27-1
ii  libnl-genl-3-200  3.2.27-1
ii  libpcsclite1      1.8.18-1
ii  libreadline6      6.3-8+b4
ii  libssl1.0.2       1.0.2h-1
ii  lsb-base          9.20160629

wpasupplicant recommends no packages.

Versions of packages wpasupplicant suggests:
pn  libengine-pkcs11-openssl  <none>
ii  wpagui                    2.5-2+v2.4-2

-- no debconf information

-- debsums errors found:
debsums: changed file /lib/systemd/system-sleep/wpasupplicant (from wpasupplicant package)

-- 
Tomáš Janoušek, a.k.a. Pivník, a.k.a. Liskni_si, http://work.lisk.in/

Message #10 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Andrew Shadura <andrew@shadura.me>

To: 835648@bugs.debian.org, Martin Pitt <martin.pitt@ubuntu.com>

Cc: Tomas Janousek <tomi@nomi.cz>

Subject: Re: [pkg-wpa-devel] Bug#835648: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Tue, 30 Aug 2016 22:36:48 +0200

On 28/08/16 00:39, Tomas Janousek wrote:
> Package: wpasupplicant
> Version: 2.5-2+v2.4-2
> Severity: normal

Martin, could you please have a look?

> I recently upgraded the wpasupplicant package and my system started taking a
> LONG time to suspend. My journal contains:
> 
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.2710] manager: sleep requested (sleeping: no  enabled: yes)
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.2710] manager: sleeping...
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.2711] device (eth0): state change: unavailable -> unmanaged (reason 'sleeping') [20 10 37]
> Aug 27 23:40:53 notes kernel: e1000e: eth0 NIC Link is Down
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.3941] manager: NetworkManager state is now ASLEEP
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.3974] device (wlan1): state change: activated -> deactivating (reason 'sleeping') [100 110 37]
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4048] device (wlan1): state change: deactivating -> disconnected (reason 'sleeping') [110 30 37]
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4232] dhcp4 (wlan1): canceled DHCP transaction, DHCP client pid 18447
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4233] dhcp4 (wlan1): state changed bound -> done
> Aug 27 23:40:53 notes kernel: wlan1: deauthenticating from c4:27:95:77:a6:b8 by local choice (Reason: 3=DEAUTH_LEAVING)
> Aug 27 23:40:53 notes wpa_supplicant[2812]: wlan1: CTRL-EVENT-DISCONNECTED bssid=c4:27:95:77:a6:b8 reason=3 locally_generated=1
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.4548] dns-mgr: Writing DNS information to /sbin/resolvconf
> Aug 27 23:40:53 notes wpa_supplicant[2812]: wlan1: CTRL-EVENT-REGDOM-CHANGE init=CORE type=WORLD
> Aug 27 23:40:53 notes NetworkManager[1904]: <warn>  [1472334053.5079] sup-iface[0x9947a98,wlan1]: connection disconnected (reason -3)
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.5080] device (wlan1): supplicant interface state: completed -> disconnected
> Aug 27 23:40:53 notes NetworkManager[1904]: <info>  [1472334053.5091] device (wlan1): state change: disconnected -> unmanaged (reason 'sleeping') [30 10 37]
> Aug 27 23:40:53 notes systemd[1]: Reached target Sleep.
> Aug 27 23:40:53 notes systemd[1]: Starting Suspend...
> Aug 27 23:40:54 notes wpa_supplicant[2812]: nl80211: deinit ifname=wlan1 disabled_11b_rates=0
> Aug 27 23:41:04 notes systemd-sleep[18924]: Selected interface 'wlan1'
> Aug 27 23:41:04 notes systemd-sleep[18924]: 'SUSPEND' command timed out.
> Aug 27 23:41:04 notes systemd-sleep[18928]: /lib/systemd/system-sleep/wpasupplicant failed with error code 254.
> Aug 27 23:41:04 notes systemd-sleep[18924]: Suspending system...
> 
> Disabling the /lib/systemd/system-sleep/wpasupplicant script fixes the problem
> for me. From what NetworkManagers prints to the journal it seems to me that
> the /lib/systemd/system-sleep/wpasupplicant script is unnecessary, and since
> it makes suspend take 10 seconds longer, it is actually not just unnecessary,
> but harmful.
> 
> I suspend the system by invoking "systemctl suspend" via acpid.
> 
> I tried to research this a bit. If I manually invoke
> 
>     dbus-send --print-reply --system --dest=org.freedesktop.NetworkManager /org/freedesktop/NetworkManager org.freedesktop.NetworkManager.Sleep boolean:true
> 
> and then try to "/sbin/wpa_cli suspend", I always get:
> 
>     Failed to connect to non-global ctrl_ifname: (nil)  error: No such file or directory
> 
> Seems to me then that the 10s delay I'm getting is a race condition between NM
> sleeping and wpa_cli attempting to suspend. It's also interesting that on my
> hardware (ThinkPad T420, Intel(R) Centrino(R) Advanced-N 6205 AGN, REV=0xB0)
> wpa_cli suspend does nothing at all -- after invoking it and getting OK, wifi
> stays connected and pings.
> 
> I also looked into dbus-monitor output and in my case NM drops the inhibitor
> only after putting both eth and wlan devices offline, unlike in the launchpad
> issue that the /lib/systemd/system-sleep/wpasupplicant is trying to fix.
> 
> -- System Information:
> Debian Release: stretch/sid
>   APT prefers testing
>   APT policy: (980, 'testing'), (980, 'stable'), (500, 'unstable-debug'), (500, 'testing-debug'), (500, 'unstable'), (500, 'stable'), (200, 'experimental'), (1, 'experimental-debug')
> Architecture: i386 (x86_64)
> Foreign Architectures: amd64
> 
> Kernel: Linux 4.4.19-lis64+ (SMP w/4 CPU cores; PREEMPT)
> Locale: LANG=cs_CZ.UTF-8, LC_CTYPE=cs_CZ.UTF-8 (charmap=UTF-8)
> Shell: /bin/sh linked to /bin/bash
> Init: systemd (via /run/systemd/system)
> 
> Versions of packages wpasupplicant depends on:
> ii  adduser           3.115
> ii  libc6             2.23-4
> ii  libdbus-1-3       1.10.10-1
> ii  libnl-3-200       3.2.27-1
> ii  libnl-genl-3-200  3.2.27-1
> ii  libpcsclite1      1.8.18-1
> ii  libreadline6      6.3-8+b4
> ii  libssl1.0.2       1.0.2h-1
> ii  lsb-base          9.20160629
> 
> wpasupplicant recommends no packages.
> 
> Versions of packages wpasupplicant suggests:
> pn  libengine-pkcs11-openssl  <none>
> ii  wpagui                    2.5-2+v2.4-2
> 
> -- no debconf information
> 
> -- debsums errors found:
> debsums: changed file /lib/systemd/system-sleep/wpasupplicant (from wpasupplicant package)
> 


-- 
Cheers,
  Andrew

Message #15 received at 835648-close@bugs.debian.org (full text, mbox, reply):

From: Andrew Shadura <andrewsh@debian.org>

To: 835648-close@bugs.debian.org

Subject: Bug#835648: fixed in wpa 2.5-2+v2.4-3

Date: Wed, 14 Sep 2016 10:32:55 +0000

Source: wpa
Source-Version: 2.5-2+v2.4-3

We believe that the bug you reported is fixed in the latest version of
wpa, 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 835648@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Andrew Shadura <andrewsh@debian.org> (supplier of updated wpa 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 ftpmaster@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Format: 1.8
Date: Wed, 14 Sep 2016 11:11:01 +0200
Source: wpa
Binary: hostapd wpagui wpasupplicant wpasupplicant-udeb
Architecture: source
Version: 2.5-2+v2.4-3
Distribution: unstable
Urgency: medium
Maintainer: Debian wpasupplicant Maintainers <pkg-wpa-devel@lists.alioth.debian.org>
Changed-By: Andrew Shadura <andrewsh@debian.org>
Description:
 hostapd    - IEEE 802.11 AP and IEEE 802.1X/WPA/WPA2/EAP Authenticator
 wpagui     - graphical user interface for wpa_supplicant
 wpasupplicant - client support for WPA and WPA2 (IEEE 802.11i)
 wpasupplicant-udeb - Client support for WPA and WPA2 (IEEE 802.11i) (udeb)
Closes: 835648 836074
Changes:
 wpa (2.5-2+v2.4-3) unstable; urgency=medium
 .
   [ Helmut Grohne ]
   * Address FTCBFS: Set PKG_CONFIG (Closes: #836074).
 .
   [ Andrew Shadura ]
   * Don't run wpa_cli suspend/resume if /run/wpa_supplicant isn't around
     (Closes: #835648).
Checksums-Sha1:
 1eaf36d77a2cd797dcf6cd1a420ce19495c0ac26 2185 wpa_2.5-2+v2.4-3.dsc
 bf810e2c0c304b55a7cba03ce75c496fd04640c8 86620 wpa_2.5-2+v2.4-3.debian.tar.xz
Checksums-Sha256:
 4dbf17ad34e6296e5df580dcd4fe72f275ee1763af00e5be5be557194a5d9915 2185 wpa_2.5-2+v2.4-3.dsc
 80a51fc9f4ea9ef54dd1dcd4a270f3d2fe1a9efd909199fb8ffab626fded6714 86620 wpa_2.5-2+v2.4-3.debian.tar.xz
Files:
 4c08ab9524538d6655cf21a7b4957718 2185 net optional wpa_2.5-2+v2.4-3.dsc
 e7de48840bdee854a39e8d4197ef27af 86620 net optional wpa_2.5-2+v2.4-3.debian.tar.xz

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJX2RRGAAoJEJ1bI/kYT6UUOF0H/RsoKA5cbl8bMMMSaxfzbjlE
JMWt5wFN5djiKbQ9JiAWB1yw1lyt0ZmVcvgeDTzNFlE6llDWuvUnIoA/yzrSKK8j
4tcEmPxcjY4XbLJgbEJu2ML5+yVe7C98GBZ05ibkiIVhE3/FPp0UPliJpt9pKZnt
BoN1Eva3L6QaSv3GqdyWRtBHZNjp2bDp2J2zcdt8BIIN5SJzg1Du4ryZXD1ViNKJ
souAuVrtDFl5AH0DjCZTR1ahcPG95XhTENxMehU+rMXBYDsGimR+wtsb3FaeEB4T
ZR0MYy/1g5uMjkcHsWJ1ezOqeQVADk1YM4iwekT29F1Qh2gtXYBHjO3M2pTTmxI=
=fGz2
-----END PGP SIGNATURE-----

Message #20 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Tomas Janousek <tomi@nomi.cz>

To: 835648@bugs.debian.org

Cc: control@bugs.debian.org

Subject: Re: Bug#835648 closed by Andrew Shadura <andrewsh@debian.org> (Bug#835648: fixed in wpa 2.5-2+v2.4-3)

Date: Wed, 14 Sep 2016 12:45:08 +0200

reopen 835648
thanks

Hello,

On Wed, Sep 14, 2016 at 10:36:06AM +0000, Debian Bug Tracking System wrote:
>    * Don't run wpa_cli suspend/resume if /run/wpa_supplicant isn't around
>      (Closes: #835648).

Unfortunately this doesn't fix the issue at all. As I said, if I manually do
the dbus-send, then wpa_cli doesn't wait, and /run/wpa_supplicant indeed isn't
present, but during suspend things happen a bit quicker than when invoked
manually and apparently /run/wpa_supplicant _is_ present when the script
checks for it. :-(

-- 
Tomáš Janoušek, a.k.a. Pivník, a.k.a. Liskni_si, http://work.lisk.in/

Message #29 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Andrew Shadura <andrew@shadura.me>

To: Tomas Janousek <tomi@nomi.cz>, 835648@bugs.debian.org

Subject: Re: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Wed, 14 Sep 2016 13:16:21 +0200

On 14/09/16 12:45, Tomas Janousek wrote:
> On Wed, Sep 14, 2016 at 10:36:06AM +0000, Debian Bug Tracking System wrote:
>> >    * Don't run wpa_cli suspend/resume if /run/wpa_supplicant isn't around
>> >      (Closes: #835648).
> Unfortunately this doesn't fix the issue at all. As I said, if I manually do
> the dbus-send, then wpa_cli doesn't wait, and /run/wpa_supplicant indeed isn't
> present, but during suspend things happen a bit quicker than when invoked
> manually and apparently /run/wpa_supplicant _is_ present when the script
> checks for it. :-(

Hmm, interesting. It doesn't exist here when systemd-sleep is being run.

-- 
Cheers,
  Andrew

Message #34 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Martin Pitt <mpitt@debian.org>

To: Tomas Janousek <tomi@nomi.cz>, 835648@bugs.debian.org

Subject: Re: Bug#835648: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Wed, 21 Sep 2016 09:08:03 +0200

Hello all,

FTR, I only added this /lib/systemd/system-sleep/wpasupplicant quirk
to have a counterpart of /usr/lib/pm-utils/sleep.d/60_wpa_supplicant
under systemd, as several users complain about not having wifi after
resuming, and re-adding this hack clearly improved things; OTOH there
are still cases where it doesn't (https://launchpad.net/bugs/1422143).

However, it is just that -- a half-working hack. On most systems (my
own included) this isn't necessary at all, and it just works around
either a kernel bug or a race condition in wpa somewhere. So if you
want to completely drop it from Debian unstable again this is totally
fine IMHO. This at least puts the pressure back to where the real root
cause is :-)

Martin

-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)

Message #39 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Alban Browaeys <prahal@yahoo.com>

To: Debian Bug Tracking System <835648@bugs.debian.org>

Subject: Re: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Thu, 24 Nov 2016 16:18:51 +0100

Package: wpasupplicant
Version: 2.5-2+v2.4-3+b1
Followup-For: Bug #835648

Adding :
"
RuntimeDirectory=wpa_supplicant
RuntimeDirectoryMode=0750
"
to the [Service] section of wpa_supplicant.service file
fixed it for me.

The rationale behind this change is that I observed on resume
/run/wpa_supplicant is there but deleted soon afterwards.
This by mean of "ls -l /run".
I hastily checked tmpfiles.d for tmpclean but got not further. The
above change was suggested by tmpfiles.d man page.

Alban

-- System Information:
Debian Release: stretch/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'unstable'), (500, 'testing'), (1, 'experimental-debug'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.9.0-rc5prahal+ (SMP w/4 CPU cores)
Locale: LANG=fr_FR.utf8, LC_CTYPE=fr_FR.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages wpasupplicant depends on:
ii  adduser           3.115
ii  libc6             2.24-5
ii  libdbus-1-3       1.11.6-1
ii  libnl-3-200       3.2.27-1
ii  libnl-genl-3-200  3.2.27-1
ii  libpcsclite1      1.8.18-1
ii  libreadline7      7.0-1
ii  libssl1.0.2       1.0.2j-4
ii  lsb-base          9.20161101

wpasupplicant recommends no packages.

Versions of packages wpasupplicant suggests:
pn  libengine-pkcs11-openssl  <none>
pn  wpagui                    <none>

-- Configuration Files:
/etc/wpa_supplicant/ifupdown.sh changed:
set -x
if [ -n "$IF_WPA_MAINT_DEBUG" ]; then
	set -x
fi
if [ "$IFACE" = lo ]; then
	exit 0
fi
if [ -n "$IF_WPA_IFACE" ]; then
	WPA_IFACE="$IF_WPA_IFACE"
else
	WPA_IFACE="$IFACE"
fi
if [ -f /etc/wpa_supplicant/functions.sh ]; then
	. /etc/wpa_supplicant/functions.sh
else
	exit 0
fi
if [ ! -x "$WPA_SUP_BIN" ] || [ ! -x "$WPA_CLI_BIN" ]; then
	exit 0
fi
do_start () {
	if test_wpa_cli; then
		# if wpa_action is active for this IFACE, do nothing
		ifupdown_locked && exit 0
		# if the administrator is calling ifup, say something useful
		if [ "$PHASE" = "pre-up" ]; then
			wpa_msg stderr "wpa_action is managing ifup/ifdown state of $WPA_IFACE"
			wpa_msg stderr "execute \`ifdown --force $WPA_IFACE' to stop wpa_action"
		fi
		exit 1
	elif ! set | grep -q "^IF_WPA"; then
		# no wpa- option defined for IFACE, do nothing
		exit 0
	fi
	# ensure stale ifupdown_lock marker is purged
	ifupdown_unlock
	# preliminary sanity checks for roaming daemon
	if [ -n "$IF_WPA_ROAM" ]; then
		if [ "$METHOD" != "manual" ]; then
			wpa_msg stderr "wpa-roam can only be used with the \"manual\" inet METHOD"
			exit 1
		fi
		if [ -n "$IF_WPA_MAPPING_SCRIPT" ]; then
			if ! type "$IF_WPA_MAPPING_SCRIPT" >/dev/null; then
				wpa_msg stderr "wpa-mapping-script \"$IF_WPA_MAPPING_SCRIPT\" is not valid"
				exit 1
			fi
		fi
		if [ -n "$IF_WPA_MAPPING_SCRIPT_PRIORITY" ] && [ -z "$IF_WPA_MAPPING_SCRIPT" ]; then
			wpa_msg stderr "\"wpa-mapping-script-priority 1\" is invalid without a wpa-mapping-script"
			exit 1
		fi
		IF_WPA_CONF="$IF_WPA_ROAM"
		WPA_ACTION_SCRIPT="/sbin/wpa_action"
	fi
	# master function; determines if ifupdown.sh should do something or not
	if [ -n "$IF_WPA_CONF" ] && [ "$IF_WPA_CONF" != "managed" ]; then
		if [ ! -s "$IF_WPA_CONF" ]; then
			wpa_msg stderr "cannot read contents of $IF_WPA_CONF"
			exit 1
		fi	
		WPA_SUP_CONF_CTRL_DIR=$(sed -n -e 's/[[:space:]]*#.*//g' -e 's/[[:space:]]\+.*$//g' \
			-e 's/^ctrl_interface=\(DIR=\)\?\(.*\)/\2/p' "$IF_WPA_CONF")
		if [ -n "$WPA_SUP_CONF_CTRL_DIR" ]; then
			WPA_CTRL_DIR="$WPA_SUP_CONF_CTRL_DIR"
			WPA_SUP_CONF="-c $IF_WPA_CONF"
		else
			# specify the default ctrl_interface since none was defined in
			# the given IF_WPA_CONF
			WPA_SUP_CONF="-c $IF_WPA_CONF -C $WPA_CTRL_DIR"
		fi
	else
		# specify the default ctrl_interface
		WPA_SUP_CONF="-C $WPA_CTRL_DIR"
	fi
}
do_stop () {
	if test_wpa_cli; then
		# if wpa_action is active for this IFACE and calling ifdown,
		# do nothing
		ifupdown_locked && exit 0
	elif test_wpa_supplicant; then
		# wpa_supplicant process exists for this IFACE, but wpa_cli
		# process does not. Allow stop mode to kill this process.
		:
	else
		exit 0
	fi
}
case "$MODE" in 
	start)
		do_start
		case "$PHASE" in
			pre-up)
				kill_wpa_supplicant
				init_wpa_supplicant	|| exit 1
				conf_wpa_supplicant 	|| { kill_wpa_supplicant; exit 1; }
				;;
			post-up)
				init_wpa_cli 		|| { kill_wpa_supplicant; exit 1; }
				;;
		esac
		;;
	stop)
		do_stop
		case "$PHASE" in
			pre-down)
				kill_wpa_cli
				;;
			post-down)
				kill_wpa_supplicant
				;;
			*)
				wpa_msg stderr "unknown phase: \"$PHASE\""
				exit 1
				;;
		esac
		;;
	
	*)
		wpa_msg stderr "unknown mode: \"$MODE\""
		exit 1
		;;
esac
exit 0


-- no debconf information

-- debsums errors found:
debsums: changed file /lib/systemd/system/wpa_supplicant.service (from wpasupplicant package)

Message #44 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Alban Browaeys <prahal@yahoo.com>

To: Debian Bug Tracking System <835648@bugs.debian.org>

Subject: Re: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Thu, 24 Nov 2016 19:24:17 +0100

Package: wpasupplicant
Version: 2.5-2+v2.4-3+b1
Followup-For: Bug #835648


Nevermind the previous "fix" is a red herring.

Otherwise the test on /run_wpasupplicant helps at times (but I confirm
there are instances when the it is still available at system-sleep time).
But it is a bit weird: if we expect the directory never to be there at
system-sleep time then the whole wpasupplicant script has no use.

If we expect it to always exists at this stage then the test is
useless.

All in all system-sleep script are run last, so even when NetworkManager
has not turned wpa_supplicant down completely, wpa_cli cannot do its
job. (ie when /run/wpa_supplicant is not deleted but happens that the sockets
in there are partially deleted). Even if the wlan0 socket exists at test time, 
we are so late that it could have vanished by the time wpa_cli process
the request.

Could we bind those wpa_cli commands to wpa_supplicant events ?
Or an nm-dispatcher script / down script ?


Alban

Message #49 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Ralf Jung <post@ralfj.de>

To: 835648@bugs.debian.org

Subject: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Sun, 8 Jan 2017 13:16:16 +0100

Hi,

I am having the same problem here, with the same error in the journal.
Removing the offending wpasupplicant script reduced the suspend time
from >10sec to <2sec. I was already wondering why suspend too so long on
the new laptop...

Kind regards,
Ralf

Message #54 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Andrew Shadura <andrew@shadura.me>

To: Ralf Jung <post@ralfj.de>, 835648@bugs.debian.org

Subject: Re: [pkg-wpa-devel] Bug#835648: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Sun, 8 Jan 2017 13:36:45 +0100

On 8 January 2017 at 13:16, Ralf Jung <post@ralfj.de> wrote:
> Hi,
>
> I am having the same problem here, with the same error in the journal.
> Removing the offending wpasupplicant script reduced the suspend time
> from >10sec to <2sec. I was already wondering why suspend too so long on
> the new laptop...

I remove the script from the new version in unstable. It would be
great if you tested it.

-- 
Cheers,
  Andrew

Message #59 received at 835648@bugs.debian.org (full text, mbox, reply):

From: Ralf Jung <post@ralfj.de>

To: Andrew Shadura <andrew@shadura.me>, 835648@bugs.debian.org

Subject: Re: Bug#835648: [pkg-wpa-devel] Bug#835648: wpasupplicant: suspend takes very long due to problematic system-sleep hook

Date: Sun, 8 Jan 2017 20:20:08 +0100

Hi,

>> I am having the same problem here, with the same error in the journal.
>> Removing the offending wpasupplicant script reduced the suspend time
>> from >10sec to <2sec. I was already wondering why suspend too so long on
>> the new laptop...
> 
> I remove the script from the new version in unstable. It would be
> great if you tested it.

That fixes the suspend time, while wifi continues to work as expected.
Thanks :)

; Ralf

Message #64 received at 835648-close@bugs.debian.org (full text, mbox, reply):

From: Andrew Shadura <andrewsh@debian.org>

To: 835648-close@bugs.debian.org

Subject: Bug#835648: fixed in wpa 2:2.4-1

Date: Fri, 24 Feb 2017 16:03:44 +0000

Source: wpa
Source-Version: 2:2.4-1

We believe that the bug you reported is fixed in the latest version of
wpa, 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 835648@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Andrew Shadura <andrewsh@debian.org> (supplier of updated wpa 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 ftpmaster@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Format: 1.8
Date: Mon, 20 Feb 2017 11:55:11 +0100
Source: wpa
Binary: hostapd wpagui wpasupplicant wpasupplicant-udeb
Architecture: source
Version: 2:2.4-1
Distribution: unstable
Urgency: medium
Maintainer: Debian wpasupplicant Maintainers <pkg-wpa-devel@lists.alioth.debian.org>
Changed-By: Andrew Shadura <andrewsh@debian.org>
Description:
 hostapd    - IEEE 802.11 AP and IEEE 802.1X/WPA/WPA2/EAP Authenticator
 wpagui     - graphical user interface for wpa_supplicant
 wpasupplicant - client support for WPA and WPA2 (IEEE 802.11i)
 wpasupplicant-udeb - Client support for WPA and WPA2 (IEEE 802.11i) (udeb)
Closes: 828601 835648 854719
Launchpad-Bugs-Fixed: 1422143
Changes:
 wpa (2:2.4-1) unstable; urgency=medium
 .
   [ Vincent Danjean ]
   * Build with libssl1.0-dev (Closes: #828601).
   * Add an upstream patch to fix hostapd in SMPS mode (Closes: #854719).
 .
   [ Andrew Shadura ]
   * Don't install debian/system-sleep/wpasupplicant (originally introduced
     to fix LP: #1422143), it doesn't improve the state of the things,
     introduces regressions in some cases, and at all isn't supposed to
     work with how wpa-supplicant is started these days (Closes: #835648).
   * Bump the epoch to 2:, so that we can set the upstream version to
     what we really mean. It also has to be higher than 2.6 in unstable
     and 1:2.6 (what hostapd binary package in unstable has).
   * Drop the binary package epoch override.
Checksums-Sha1:
 179ccc9b04d0a4a56f1fdecf28c7b3d50b905bfc 2233 wpa_2.4-1.dsc
 be9f0c01074cebe981a168eb747eab252eeff5f6 1834600 wpa_2.4.orig.tar.xz
 1d894f8ec73fbe754247228d8af911b7a2fda171 87216 wpa_2.4-1.debian.tar.xz
Checksums-Sha256:
 10226c35e6a43b556f7f20ffbbbe3d84a80069efa22ed4382a5cabd3349e586f 2233 wpa_2.4-1.dsc
 a1e4eda50796b2234a6cd2f00748bbe09f38f3f621919187289162faeb50b6b8 1834600 wpa_2.4.orig.tar.xz
 e76262448f47c205824b967a1a3c862eca91ecaf64486c773438d8144c459fff 87216 wpa_2.4-1.debian.tar.xz
Files:
 6754370c694dd5027bc40fc24f9f97fa 2233 net optional wpa_2.4-1.dsc
 6a77b9fe6838b4fca9b92cb22e14de1d 1834600 net optional wpa_2.4.orig.tar.xz
 af22bdfd566d44914c6a74af6ce00a6a 87216 net optional wpa_2.4-1.debian.tar.xz

-----BEGIN PGP SIGNATURE-----

iQEcBAEBCAAGBQJYsFOkAAoJEJ1bI/kYT6UU4n4H/0CGuWz9WMeA1gk3gaFYXYv+
XXMXbBG5Ok8Mn3o9VVq0cyB3NAGOFX1VzAiTChS9fzcAsYfr1Uhx3woSLFarQqzf
hYpt3OlLJrjzmUjidlYKAS35cNZm3XyOaLxZjupNchuluGEzDZil9P+ois7JvVS7
nrlPn4vpRp8Qqt+g1lk8jZzdNE8ADMbyaXJT+8GNapXpVH5PWSVPqDbCupZ3Mtxm
yjvulKWslk1zBfY/J4I3Va35ACZa3jBFRbL6AjCDlgtaApSefgOv1KNMXNhrh53T
W7MhoCwWra/5a7upQgfYWt0nReuORgQzaer+WzBc3IEbcNqYLAPV6wt2I4OFvAQ=
=s1gJ
-----END PGP SIGNATURE-----

Send a report that this bug log contains spam.