FS#34795 - modemmanager 0.7.990-1 networkmanager 0.9.8.0-3 breaks UMTS/LTE Mobile Stick Support.

Attached to Project: Arch Linux
Opened by Marcel Rouvlak (catravic) - Monday, 15 April 2013, 19:44 GMT
Last edited by Jan Alexander Steffens (heftig) - Thursday, 18 April 2013, 14:36 GMT
Task Type Bug Report
Category Packages: Testing
Status Closed
Assigned To Jan de Groot (JGC)
Jan Alexander Steffens (heftig)
Architecture All
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:
Failing at initializing modem.

Additional info:
modemmanager 0.7.990-1
networkmanager 0.9.8.0-3

Downgrade to 0.9.8.0-2; 0.6.0.0-1; works fine.

See https://bbs.archlinux.org/viewtopic.php?id=161441 for logs


Steps to reproduce:
Upgrade to mentoined packages and test with huawei or zte modem.
This task depends upon

Closed by  Jan Alexander Steffens (heftig)
Thursday, 18 April 2013, 14:36 GMT
Reason for closing:  Fixed
Additional comments about closing:  0.7.990-3
Comment by phanisvara das (phani00) - Monday, 15 April 2013, 23:42 GMT
i can confirm this on 3.8.7-1-ARCH x86_64, with a 'ZTE WCDMA Technologies MSM' 3G modem. (same versions of net- and modem-mangers.)

in addition, when i plug in the modem and NetworkManager & ModemManager are running, my sound devices get shuffled around: they appear and disappear in the KDE systemsettings module, confusing pulseaudio. without those ~Managers, that doesn't happen.

also, without Network- and ModemManager, wvdial and the proposed netctl profile mobile_ppp work fine.

for some reason, the 3G modem gets assigned conflicting/wrong resources.
Comment by Jan Alexander Steffens (heftig) - Wednesday, 17 April 2013, 18:38 GMT
Please try 0.7.990-2.
Comment by phanisvara das (phani00) - Wednesday, 17 April 2013, 19:58 GMT
things have changed, the sound devices don't get switched around anymore, but it still doesn't connect.

the firest time i connect the modem, it's recognized as such and appears as option in NM. that it doesn't connect at that time may be due to my lousy provider, who've over-sold their capacity, but i've tried several times and normally this time of night (here) works mostly.

i'm attaching 'everything.log' from the moment of connection until the device disappears again from NM.

on further attempts to re-connect the modem it isn't even recognized anymore, only after restarting ModemManager.service.



Comment by Jan Alexander Steffens (heftig) - Wednesday, 17 April 2013, 20:32 GMT
Could you please get a backtrace from "systemd-coredumpctl gdb 6161"?

Never mind, it's not in the journal. Please stop ModemManager and rerun it under gdb in order to get traces.

Oh, and for the future, note that ModemManager must now be enabled (systemctl enable) in order to get started automatically.
Comment by phanisvara das (phani00) - Wednesday, 17 April 2013, 20:57 GMT
running ModemManager via gdb tells me that many libraries have been loaded, but for each one it complains that no debug symbols are available (surprise). it will probably take a while still until those are available from arch repos?

when i attach gdb later (via PID) the same thing happens, but the 3G modem isn't recognize at all.

attaching gdb after the modem has been recognized but before trying to connect doesn't show any output during the (failed) connect attempts. even though i think it's useless, i'm attaching the output i get.

and after stopping MM & NM, the connection works immediately using netctl profile mobile_ppp.
Comment by phanisvara das (phani00) - Wednesday, 17 April 2013, 21:00 GMT
> Oh, and for the future, note that ModemManager must now be enabled (systemctl enable) in order to get started automatically.

of course; but in general i prefer netcfg, or netctl now.
Comment by Jan Alexander Steffens (heftig) - Wednesday, 17 April 2013, 21:00 GMT
You need to tell gdb to "run" (in case of starting fresh) or "continue" (in case of attaching later) in order for ModemManager to actually run. While the gdb prompt is displayed, the program is paused. Once the prompt returns due to ModemManager crashing, use "bt" to get a backtrace.
Comment by phanisvara das (phani00) - Wednesday, 17 April 2013, 21:24 GMT
(gdb) continue
Continuing.

Program received signal SIGSEGV, Segmentation fault.
0x00007f51346815e6 in __memcpy_ssse3_back () from /usr/lib/libc.so.6
(gdb) bt
#0 0x00007f51346815e6 in __memcpy_ssse3_back () from /usr/lib/libc.so.6
#1 0x00007f5134b3585f in g_array_append_vals () from /usr/lib/libglib-2.0.so.0
#2 0x00007f5134b36869 in g_byte_array_append () from /usr/lib/libglib-2.0.so.0
#3 0x0000000000439458 in mm_sms_part_new_from_binary_pdu ()
#4 0x00000000004399b6 in mm_sms_part_new_from_pdu ()
#5 0x00000000004631f9 in ?? ()
#6 0x00007f51350d0ab7 in g_simple_async_result_complete () from /usr/lib/libgio-2.0.so.0
#7 0x0000000000434d08 in ?? ()
#8 0x000000000048312b in ?? ()
#9 0x0000000000480b13 in ?? ()
#10 0x00000000004826ab in ?? ()
#11 0x00007f5134b5ff06 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#12 0x00007f5134b60258 in ?? () from /usr/lib/libglib-2.0.so.0
#13 0x00007f5134b6065a in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
#14 0x0000000000426da6 in main ()
(gdb)
Comment by Jan Alexander Steffens (heftig) - Wednesday, 17 April 2013, 22:14 GMT Comment by Jan Alexander Steffens (heftig) - Thursday, 18 April 2013, 12:58 GMT
Crash got fixed upstream, patch is in 0.7.990-3.
Comment by phanisvara das (phani00) - Thursday, 18 April 2013, 13:59 GMT
great, now it's working as before!

Loading...