Bug#861112: xsane: always crashes on start

[Aaro Koskinen]

Hi,

On Fri, May 19, 2017 at 11:44:56AM +0200, Wolfgang Schweer wrote:
> On Fri, May 19, 2017 at 09:47:51AM +0200, John Paul Adrian Glaubitz wrote:
> > On 05/17/2017 10:57 PM, Andreas Henriksson wrote:
> > >> It's disabling Avahi support (I don't have such daemon)
> 
> IMO a daemon isn't needed, that might be a misleading debug message. On 
> my system (up-to-date stretch, GNOME) the package libavahi-client3 
> (Depends: libavahi-common3) is installed via libreoffice-draw / 
> libgnomevfs2-0 Depends -- and I'm unable to reproduce the bug. 
> 
> Maybe it would be sufficent to add libavahi-client3 (instead of 
> libavahi-common3) as a Depends to sane-utils to solve this problem.

Doesn't help:

$ sudo apt-get install libavahi-client3
Reading package lists... Done
Building dependency tree   
Reading state information... Done
libavahi-client3 is already the newest version (0.6.32-2).
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
$ scanimage -L
scanimage: thread-watch.c:171: avahi_threaded_poll_lock: Assertion `p' failed.
Aborted
$ 

A.

Bug#861112: xsane: always crashes on start

[Andreas Henriksson]

Hello Adrian,

Thanks for looking at this again. More comments below.

On Fri, May 19, 2017 at 09:47:51AM +0200, John Paul Adrian Glaubitz wrote:
> But wouldn't that only address the symptoms instead of the actual cause
> of the problem? If I understood Laurent correctly, the NULL value of
> avahi_thread is a result of a race condition that can be avoided by
> calling net_avahi_init() later inside sane_init().

It might be that Laurents fix should *also* go in. It might make things
work more often, but still...

The net_avahi_init() function handles failures by setting avahi_thread
(etc.) to NULL.

The caller of net_avahi_init() does nothing to catch when net_avahi_init()
returns failure (maybe this should be caught and handled here? But I assumed
it was considered ok for it to fail.)

Thus, If for ANY reason net_avahi_init() failed and we continue running
with avahi_thread==NULL. Sane must make sure to not pass it to something
which does not accept a NULL argument, eg. the avahi lock functions.
Locking avahi when we're not using avahi at all is obviously not needed
as I see it.

Thus the patch I proposed. I'm not familiar with sane code so maybe this
is not the best fix, but either way Laurents change definitely doesn't
cover all theoretical bases (ie. it doesn't handle the failure, just
possibly makes failure happen less often)

Hope this helps make my proposal more clear.

Regards,
Andreas Henriksson

Bug#861112: xsane: always crashes on start

[Wolfgang Schweer]

On Fri, May 19, 2017 at 09:47:51AM +0200, John Paul Adrian Glaubitz wrote:
> On 05/17/2017 10:57 PM, Andreas Henriksson wrote:
> >> It's disabling Avahi support (I don't have such daemon)

IMO a daemon isn't needed, that might be a misleading debug message. On 
my system (up-to-date stretch, GNOME) the package libavahi-client3 
(Depends: libavahi-common3) is installed via libreoffice-draw / 
libgnomevfs2-0 Depends -- and I'm unable to reproduce the bug. 

Maybe it would be sufficent to add libavahi-client3 (instead of 
libavahi-common3) as a Depends to sane-utils to solve this problem.

'apt show libavahi-client3' tells me:
This package contains the library for Avahi's C API which allows you to 
integrate mDNS/DNS-SD functionality into your application. 
 
Wolfgang


signature.asc
Description: PGP signature

Bug#861112: xsane: always crashes on start

[John Paul Adrian Glaubitz]

On 05/17/2017 10:57 PM, Andreas Henriksson wrote:
>> It's disabling Avahi support (I don't have such daemon) but still later
>> calling avahi_threaded_poll_lock() with NULL avahi_thread.
> [...]
> 
> Yes, definitely seems so. Could you please test the attached patch
> which hopefully takes care of your issue?

But wouldn't that only address the symptoms instead of the actual cause
of the problem? If I understood Laurent correctly, the NULL value of
avahi_thread is a result of a race condition that can be avoided by
calling net_avahi_init() later inside sane_init().

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaub...@debian.org
`. `'   Freie Universitaet Berlin - glaub...@physik.fu-berlin.de
  `-GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Bug#861112: xsane: always crashes on start

[Andreas Henriksson]

On Sun, May 07, 2017 at 11:24:13PM +0300, Aaro Koskinen wrote:
> $ SANE_DEBUG_NET=3 scanimage -L
[...]
> scanimage: thread-watch.c:171: avahi_threaded_poll_lock: Assertion `p' failed.
> Aborted
> 
> It's disabling Avahi support (I don't have such daemon) but still later
> calling avahi_threaded_poll_lock() with NULL avahi_thread.
[...]

Yes, definitely seems so. Could you please test the attached patch
which hopefully takes care of your issue?

Regards,
Andreas Henriksson
From: Andreas Henriksson 
Subject: Avoid assertion failure when net_avahi_init failed

The avahi_threaded_poll_{un,}lock functions does not accept a NULL argument.
The avahi_thread variable might be NULL when the net_avahi_init() function
failed. We thus need to make sure we don't pass a NULL avahi_thread
to any of these functions.

Bug-Debian: https://bugs.debian.org/861112

diff -uriNp sane-backends-1.0.26~git20151121/backend/net.c sane-backends-1.0.26~git20151121.nullcheck/backend/net.c
--- sane-backends-1.0.26~git20151121/backend/net.c	2015-11-13 03:10:49.0 +0100
+++ sane-backends-1.0.26~git20151121.nullcheck/backend/net.c	2017-05-17 22:49:01.922390615 +0200
@@ -1045,12 +1045,12 @@ sane_init (SANE_Int * version_code, SANE
 	  continue;
 	}
 #ifdef WITH_AVAHI
-	  avahi_threaded_poll_lock (avahi_thread);
+	  if (avahi_thread) avahi_threaded_poll_lock (avahi_thread);
 #endif /* WITH_AVAHI */
 	  DBG (2, "sane_init: trying to add %s\n", device_name);
 	  add_device (device_name, 0);
 #ifdef WITH_AVAHI
-	  avahi_threaded_poll_unlock (avahi_thread);
+	  if (avahi_thread) avahi_threaded_poll_unlock (avahi_thread);
 #endif /* WITH_AVAHI */
 	}
 
@@ -1096,12 +1096,12 @@ sane_init (SANE_Int * version_code, SANE
 		  continue;
 #endif /* ENABLE_IPV6 */
 #ifdef WITH_AVAHI
-	  avahi_threaded_poll_lock (avahi_thread);
+	  if (avahi_thread) avahi_threaded_poll_lock (avahi_thread);
 #endif /* WITH_AVAHI */
 	  DBG (2, "sane_init: trying to add %s\n", host);
 	  add_device (host, 0);
 #ifdef WITH_AVAHI
-	  avahi_threaded_poll_unlock (avahi_thread);
+	  if (avahi_thread) avahi_threaded_poll_unlock (avahi_thread);
 #endif /* WITH_AVAHI */
 	}
 	  free (copy);
@@ -1519,11 +1519,11 @@ sane_open (SANE_String_Const full_name,
 	   "sane_open: device %s not found, trying to register it anyway\n",
 	   nd_name);
 #ifdef WITH_AVAHI
-  avahi_threaded_poll_lock (avahi_thread);
+  if (avahi_thread) avahi_threaded_poll_lock (avahi_thread);
 #endif /* WITH_AVAHI */
   status = add_device (nd_name, );
 #ifdef WITH_AVAHI
-  avahi_threaded_poll_unlock (avahi_thread);
+  if (avahi_thread) avahi_threaded_poll_unlock (avahi_thread);
 #endif /* WITH_AVAHI */
   if (status != SANE_STATUS_GOOD)
 	{

Bug#861112: xsane: always crashes on start

[Aaro Koskinen]

Hi,

On Fri, May 05, 2017 at 08:04:26PM +0200, John Paul Adrian Glaubitz wrote:
> On 05/05/2017 07:16 PM, Aaro Koskinen wrote:
> > I think Jörg Frings-Fürst has analyzed and fixed this bug already - he
> > provided me test packages (libsane_1.0.25-4~test1_amd64.deb etc.) offline
> > for testing and they seemed to fix the issue.
> 
> No, he didn't "fix" the issue. He completely removed a feature from 
> sane-backends
> and called it a fix, assuming that no one in Debian or Ubuntu or any of the 
> derived
> distributions wants to use Avahi with SANE.

Okay...

> > I'm currently away from my machines, but next week I'll recheck the
> > situation again with up-to-date Debian testing, If the problem is still
> > present I will provide the full backtrace etc.
> 
> Yes, please. I am not going to accept such a rather big cut in functionality
> without proper testing and error analysis first.

I found something when enabling debug outputs:

$ SANE_DEBUG_NET=3 scanimage -L
[sanei_debug] Setting debug level of net to 3.
[net] sane_init: authorize != null, version_code != null
[net] net_avahi_init: could not create Avahi client: Daemon not running
[net] net_avahi_init: Avahi init failed, support disabled
[net] sane_init: SANE net backend version 1.0.14 (AF-indep+IPv6) from 
sane-backends 1.0.25
[net] sane_init: Client has little endian byte order
[net] sane_init: searching for config file
scanimage: thread-watch.c:171: avahi_threaded_poll_lock: Assertion `p' failed.
Aborted

It's disabling Avahi support (I don't have such daemon) but still later
calling avahi_threaded_poll_lock() with NULL avahi_thread.

Here's the backtrace:

$ gdb scanimage
GNU gdb (Debian 7.12-6) 7.12.0.20161007-git
Copyright (C) 2016 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later 
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
.
Find the GDB manual and other documentation resources online at:
.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from scanimage...Reading symbols from 
/usr/lib/debug/.build-id/2a/f2218a71fe370679d900eca9ab1b1e4f575859.debug...done.
done.
(gdb) run -L
Starting program: /usr/bin/scanimage -L
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x76420700 (LWP 6782)]
scanimage: thread-watch.c:171: avahi_threaded_poll_lock: Assertion `p' failed.

Thread 1 "scanimage" received signal SIGABRT, Aborted.
x773c7fcf in raise () from /lib/x86_64-linux-gnu/libc.so.6
(gdb) bt
#0  0x773c7fcf in raise () from /lib/x86_64-linux-gnu/libc.so.6
#1  0x773c93fa in abort () from /lib/x86_64-linux-gnu/libc.so.6
#2  0x773c0e37 in ?? () from /lib/x86_64-linux-gnu/libc.so.6
#3  0x773c0ee2 in __assert_fail () from /lib/x86_64-linux-gnu/libc.so.6
#4  0x70ce9404 in avahi_threaded_poll_lock ()
   from /usr/lib/x86_64-linux-gnu/libavahi-common.so.3
#5  0x7fffeadcac7c in sane_net_init (version_code=, 
authorize=) at net.c:1048
#6  0x77bd4dd0 in init (be=be@entry=0x55762290) at dll.c:614
#7  0x77bd5681 in sane_dll_get_devices (device_list=0x7fffbc18, 
local_only=0) at dll.c:1055
#8  0x62ab in main (argc=2, argv=0x7fffe068)
at scanimage.c:2038

A.

Bug#861112: xsane: always crashes on start

[John Paul Adrian Glaubitz]

On 05/05/2017 07:16 PM, Aaro Koskinen wrote:
> I think Jörg Frings-Fürst has analyzed and fixed this bug already - he
> provided me test packages (libsane_1.0.25-4~test1_amd64.deb etc.) offline
> for testing and they seemed to fix the issue.

No, he didn't "fix" the issue. He completely removed a feature from 
sane-backends
and called it a fix, assuming that no one in Debian or Ubuntu or any of the 
derived
distributions wants to use Avahi with SANE.

> I'm currently away from my machines, but next week I'll recheck the
> situation again with up-to-date Debian testing, If the problem is still
> present I will provide the full backtrace etc.

Yes, please. I am not going to accept such a rather big cut in functionality
without proper testing and error analysis first.

Thanks,
Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaub...@debian.org
`. `'   Freie Universitaet Berlin - glaub...@physik.fu-berlin.de
  `-GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Bug#861112: xsane: always crashes on start

[Aaro Koskinen]

Hi,

On Fri, May 05, 2017 at 03:33:34PM +0200, John Paul Adrian Glaubitz wrote:
> There previously has been a race condition in the sane-backends net code
> which Laurent Vivier found [1]. He had come up with a suggested patch to
> address the issue, but the patch was never merged to the Debian package
> or upstream, it seems.
> 
> So, before we go ahead and completely remove a feature that some people
> might be using, I would like to ask that we actually test the patch that
> Laurent suggested and see it helps with the problem. Not having verified
> that we have actually properly fixed the related bug #743420 [2], we cannot
> be exclude this particular bug as the problem Aaro is seeing here.
> 
> @Aaro: It would be nice if you could provide us with the backtrace of your
>crash. sane-backends still has a dedicated -dbg package, for the
>Avahi debug packages you will probably have to use Debian's debug
>repository [2].

I think Jörg Frings-Fürst has analyzed and fixed this bug already - he
provided me test packages (libsane_1.0.25-4~test1_amd64.deb etc.) offline
for testing and they seemed to fix the issue.

I'm currently away from my machines, but next week I'll recheck the
situation again with up-to-date Debian testing, If the problem is still
present I will provide the full backtrace etc.

Thanks,

A.

Bug#861112: xsane: always crashes on start

[John Paul Adrian Glaubitz]

Hi!

There previously has been a race condition in the sane-backends net code
which Laurent Vivier found [1]. He had come up with a suggested patch to
address the issue, but the patch was never merged to the Debian package
or upstream, it seems.

So, before we go ahead and completely remove a feature that some people
might be using, I would like to ask that we actually test the patch that
Laurent suggested and see it helps with the problem. Not having verified
that we have actually properly fixed the related bug #743420 [2], we cannot
be exclude this particular bug as the problem Aaro is seeing here.

@Aaro: It would be nice if you could provide us with the backtrace of your
   crash. sane-backends still has a dedicated -dbg package, for the
   Avahi debug packages you will probably have to use Debian's debug
   repository [2].

Thanks,
Adrian

> [1] 
> https://bugs.launchpad.net/ubuntu/+source/sane-backends/+bug/1208091/comments/10
>  [2] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=743420
> [2] https://wiki.debian.org/AutomaticDebugPackages


-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaub...@debian.org
`. `'   Freie Universitaet Berlin - glaub...@physik.fu-berlin.de
  `-GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Bug#861112: xsane: always crashes on start

[Wolfgang Schweer]

On Sun, Apr 30, 2017 at 06:53:32PM +0200, Jörg Frings-Fürst wrote:
> I have a bugfix release for sane-backend ready and I send it to you.
> 
> Please can you test the packages?

I did tests on different DEs (KDE, MATE, GNOME, Xfce) after installing 
the new packages in each case.

Both 'scanimage -L' and actually scanning via frontends 'Xsane' and 
'Skanlite' worked like expected.

To summarize: I can't reproduce the reported crash neither with the 
packages in stretch nor with the new packages.

Wolfgang


signature.asc
Description: PGP signature

Bug#861112: xsane: always crashes on start

[Jörg Frings-Fürst]

Hello Wolfgang,

thank you for spending your time helping to make Debian better with
this bug report.


I have a bugfix release for sane-backend ready and I send it to you.

Please can you test the packages?


CU
Jörg

-- 
New:
GPG Fingerprint: 63E0 075F C8D4 3ABB 35AB  30EE 09F8 9F3C 8CA1 D25D
GPG key (long) : 09F89F3C8CA1D25D
GPG Key: 8CA1D25D
CAcert Key S/N : 0E:D4:56

Old pgp Key: BE581B6E (revoked since 2014-12-31).

Jörg Frings-Fürst
D-54470 Lieser

Threema: SYR8SJXB
Wire: @joergfringsfuerst

IRC: j_...@freenode.net
 j_...@oftc.net

My wish list: 
 - Please send me a picture from the nature at your home.


signature.asc
Description: This is a digitally signed message part

Bug#861112: xsane: always crashes on start

[Wolfgang Schweer]

On Mon, Apr 24, 2017 at 11:49:46PM +0300, Aaro Koskinen wrote:
> On Mon, Apr 24, 2017 at 10:21:47PM +0200, John Paul Adrian Glaubitz wrote:
> > Control: tags -1 moreinfo
> > > xsane with 1 network scanner defined in /etc/sane.d/net.conf
> > > crashes always on start:

While testing Debian Edu Stretch (virtualbox network 10.0.0.0/8 with 
gateway to a 192.168.178.0/24 LAN)) I tested to access a scanner 
connected to 192.168.178.2 (Jessie workstation, USB scanner). It worked 
like expected (scanimage, scanlite, xsane).

> > Would you mind sharing your net.conf file with this bug report so
> > that we can try to reproduce the problem?

> ## net backend options
> # Timeout for the initial connection to saned. This will prevent the backend
> # from blocking for several minutes trying to connect to an unresponsive
> # saned host (network outage, host down, ...). Value in seconds.
> # connect_timeout = 60
> 
> ## saned hosts
> # Each line names a host to attach to.
> # If you list "localhost" then your backends can be accessed either
> # directly or through the net backend.  Going through the net backend
> # may be necessary to access devices that need special privileges.
> # localhost
> 192.168.1.100

Same here, only the IP differs; see above.

This is the 'scanimage -L' output ('tjener' is the Debian Edu main 
server):

[jdoe@tjener ~]$ scanimage -L
device `net:192.168.178.2:gt68xx:libusb:001:014' is a Mustek ScanExpress 
1248 UB flatbed scanner

Output on the Jessie workstation for comparison:

[jdoe@tjener ~]$ ssh jdoe@192.168.178.2 scanimage -L
device `gt68xx:libusb:001:014' is a Mustek ScanExpress 1248 UB flatbed 
scanner

Same succes on a workstation inside the 10.0.0.0/8 network.

So I suspect there must be something else in addition causing the 
issue. 

Wolfgang


signature.asc
Description: PGP signature

Processed: Re: Bug#861112: xsane: always crashes on start

[Debian Bug Tracking System]

Processing commands for cont...@bugs.debian.org:

> reassign 861112 libsane
Bug #861112 [xsane] xsane: always crashes on start
Bug reassigned from package 'xsane' to 'libsane'.
No longer marked as found in versions xsane/0.999-5.
Ignoring request to alter fixed versions of bug #861112 to the same values 
previously set
> tags 861112 + moreinfo
Bug #861112 [libsane] xsane: always crashes on start
Added tag(s) moreinfo.
> thanks
Stopping processing here.

Please contact me if you need assistance.
-- 
861112: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=861112
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems

Bug#861112: xsane: always crashes on start

[Jörg Frings-Fürst]

reassign 861112 libsane
tags 861112 + moreinfo
thanks


Hello Aaro,

thank you for spending your time helping to make Debian better with
this bug report.

First I think this bug comes from sane-backends. 

I have a bugfix release for sane-backend ready and I send it to you.
Please can you test the packages?


Many thanks and have a nice weekend.

CU
Jörg 


-- 
New:
GPG Fingerprint: 63E0 075F C8D4 3ABB 35AB  30EE 09F8 9F3C 8CA1 D25D
GPG key (long) : 09F89F3C8CA1D25D
GPG Key: 8CA1D25D
CAcert Key S/N : 0E:D4:56

Old pgp Key: BE581B6E (revoked since 2014-12-31).

Jörg Frings-Fürst
D-54470 Lieser

Threema: SYR8SJXB
Wire: @joergfringsfuerst

IRC: j_...@freenode.net
 j_...@oftc.net

My wish list: 
 - Please send me a picture from the nature at your home.


signature.asc
Description: This is a digitally signed message part

Bug#861112: xsane: always crashes on start

[Aaro Koskinen]

Hi,

Seems to happen also with scanimage:

$ scanimage -L
scanimage: thread-watch.c:171: avahi_threaded_poll_lock: Assertion `p' failed.
Aborted

A.

Bug#861112: xsane: always crashes on start

[Aaro Koskinen]

Hi,

On Mon, Apr 24, 2017 at 10:21:47PM +0200, John Paul Adrian Glaubitz wrote:
> Control: tags -1 moreinfo
> 
> Hi Aaro!
> 
> > xsane with 1 network scanner defined in /etc/sane.d/net.conf
> > crashes always on start:
> 
> Would you mind sharing your net.conf file with this bug report so
> that we can try to reproduce the problem?

---8<---
# This is the net backend config file.

## net backend options
# Timeout for the initial connection to saned. This will prevent the backend
# from blocking for several minutes trying to connect to an unresponsive
# saned host (network outage, host down, ...). Value in seconds.
# connect_timeout = 60

## saned hosts
# Each line names a host to attach to.
# If you list "localhost" then your backends can be accessed either
# directly or through the net backend.  Going through the net backend
# may be necessary to access devices that need special privileges.
# localhost
192.168.1.100
--->8---

A.

Bug#861112: xsane: always crashes on start

[John Paul Adrian Glaubitz]

Control: tags -1 moreinfo

Hi Aaro!

> xsane with 1 network scanner defined in /etc/sane.d/net.conf
> crashes always on start:

Would you mind sharing your net.conf file with this bug report so
that we can try to reproduce the problem?

Thanks,
Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaub...@debian.org
`. `'   Freie Universitaet Berlin - glaub...@physik.fu-berlin.de
  `-GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Bug#861112: xsane: always crashes on start

[Aaro Koskinen]

Package: xsane
Version: 0.999-5
Severity: grave
Justification: renders package unusable

Dear Maintainer,

xsane with 1 network scanner defined in /etc/sane.d/net.conf
crashes always on start:

$ xsane
xsane: thread-watch.c:171: avahi_threaded_poll_lock: Assertion `p' failed.
Aborted

Before the crash a window appears briefly telling "scanning for devices".

-- System Information:
Debian Release: 9.0
  APT prefers testing
  APT policy: (500, 'testing')
Architecture: amd64
 (x86_64)

Kernel: Linux 4.9.0-2-amd64 (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages xsane depends on:
ii  libc62.24-9
ii  libgimp2.0   2.8.18-1
ii  libglib2.0-0 2.50.3-2
ii  libgtk2.0-0  2.24.31-2
ii  libjpeg62-turbo  1:1.5.1-2
ii  liblcms2-2   2.8-4
ii  libpng16-16  1.6.28-1
ii  libsane  1.0.25-3
ii  libtiff5 4.0.7-6
ii  xsane-common 0.999-5
ii  zlib1g   1:1.2.8.dfsg-5

Versions of packages xsane recommends:
pn  cups-client
ii  firefox-esr [www-browser]  45.8.0esr-1
ii  lynx [www-browser] 2.8.9dev11-1

Versions of packages xsane suggests:
pn  gimp  
pn  gocr | cuneiform | tesseract-ocr | ocrad  
pn  gv
pn  hylafax-client | mgetty-fax   

-- no debconf information