Segmentation Fault when starting Muse [sf#1393956]

[Windfisch]

Submitted by nobody on 2005-12-30 23:34:12

haver@bibo:~/Midi$ muse -v No superuser privileges, using system timer fallback muse: Linux Music Editor; Version 0.7.2pre2

haver@bibo:~/Midi$ muse No superuser privileges, using system timer fallback no locale /</usr/share/muse/locale> starting with default template name2route: not found name2route: not found set realtime scheduler: Operation not permitted midi thread 9119 NOT running SCHED_FIFO Segmentation fault haver@bibo:~/Midi$

Unfortunately muse crashes reproducible on my Debian system. This is testing. I have not looked up the Debian bug tracking database yet but I thought since it is so nicely reproducable that I could attach you the strace output for you to improve your program.

Good luck.

[Windfisch]

Submitted by spamatica on 2006-01-01 21:22:14

Hi,

this bugreport is very strange. It's a pity you didn't
specify an e-mail adress.
In the output there's a recurring error,
"No superuser privileges, using system timer fallback". Which looks like it could have something to do with this.
The problem is that there is no such text anywhere in the MusE sources. My guess is that you are using a prebuilt version of MusE that contains some patch, debian?. If the error is related I can not say but it would be good if we could get this

patch.

Before I spotted this I was just going to ask you to try a newer version of MusE (I guess you have to go all the way to CVS because of recent bug...), work has been done in this area.

[Windfisch]

Submitted by fhaver on 2006-01-02 19:07:41

Yes, I am using a Debian version. I do not know if they patched it. It is just convenient to type: apt-get install muse ;-).

If I start muse with -a it does not chrash. Otherwise it crashes. I updated my Debian system from testing to unstable to get a newer version if there would be any.

haver@bibo:~$ muse No superuser privileges, using system timer fallback NO Config File </home/haver/.MusE> found /usr/bin/konqueror no locale /</usr/share/muse/locale> Trying RTC timer... fatal error: open /dev/rtc failed: No such file or directory Trying ALSA timer... got timer = 12 QObject::connect: No such signal PartCanvas::horizontalScroll(int) QObject::connect: (sender name: 'unnamed') QObject::connect: (receiver name: 'unnamed') Arranger::configChanged - no bitmap! starting with default template name2route: not found name2route: not found Arranger::configChanged - no bitmap! AlsaTimer::setTimerTicks(): requested freq 1024 Hz too high for timer (max is 250) freq stays at 250 Hz Segmentation fault haver@bibo:~$

Sorry for the anonymous posting before, I forgot to log in.

haver@bibo:~/Midi$ gdb /usr/bin/muse -c core.5342 ... Reading symbols from /usr/lib/liblcms.so.1... (no debugging symbols found)...done. Loaded symbols for /usr/lib/liblcms.so.1 Reading symbols from /usr/lib/muse/plugins/freeverb.so...(no debugging symbols found)...done. Loaded symbols for /usr/lib/muse/plugins/freeverb.so

0 0x73203a72 in ?? ()

(gdb) bt

0 0x73203a72 in ?? ()

1 0x080a510f in

QValueListPrivate::~QValueListPrivate ()

2 0xb70e4ed0 in __libc_start_main () from /lib/tls/libc.so.6

3 0x080675b1 in ?? ()

(gdb)

[Windfisch]

Submitted by fhaver on 2006-01-02 19:12:00

A complete reinstall helped last time to get it starting again ... strange. I tried to delete the .MusE config file to start with fresh settings, but that did not help.

Is there something strange with my version of QT? I am using a Gnome desktop and not KDE. Let me know if you need more data e.g. core files, configuration I am using etc.

[Windfisch]

Submitted by fhaver on 2006-01-03 22:45:01

Hi Robert,

The hardware I am using is my old T21 Thinkpad. I am very new to the MIDI and music business. My wife and I just bought an e-piano for her and I thought: Let's look what we might be able to do, if we attach a notebook to the device.

I tried playing arround looking for a nice easy to use MIDI Software. To make it some callenge I wanted to try it on my Linux notebook first ;-). The sound card of the T21 seems not really be able to support MIDI nor has it internal sounds. So I installed timitity, got a little frustrated until I figured out that I needed to use some sounds distributed in a different package. I am using now freepats, and was astouned that it was not automatically installed with timitity. But now I can listen to the example midi files, althought the performance is not optimal.

I have not yet started to use jackd but I tried also Rosegarden and it looks as if they have an option to start jackd automatically for beginners, nevertheless I do not use that because of bad performace of my notebook (or wrong settings... e.g. to high sample rate).

Even not using jackd I think Muse should not segfault. By the way I had it segfaulting when moving the progress sliders back and force, but I was not able to reproduce that, so I could not write a bug report and I had core dumping disabled. But the segfault we are talking here about is more serious because it stops me from being able to start Muse.

Frank

Here some information about my setup:

haver@bibo:~$ mv dpkg.lst dpkg-l.txt haver@bibo:~$ lspci 0000:00:00.0 Host bridge: Intel Corporation 440BX/ZX/DX - 82443BX/ZX/DX Host bridge (rev 03)0000:00:01.0 PCI bridge: Intel Corporation 440BX/ZX/DX - 82443BX/ZX/DX AGP bridge (rev 03) 0000:00:02.0 CardBus bridge: Texas Instruments PCI1450 (rev 03) 0000:00:02.1 CardBus bridge: Texas Instruments PCI1450 (rev 03) 0000:00:03.0 Ethernet controller: 3Com Corporation 3c556B CardBus [Tornado](rev 20) 0000:00:03.1 Communication controller: 3Com Corporation Mini PCI 56k Winmodem (rev 20) 0000:00:05.0 Multimedia audio controller: Cirrus Logic CS 4614/22/24 [CrystalClear SoundFusion Audio Accelerator](rev 01) 0000:00:07.0 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ISA (rev 02) 0000:00:07.1 IDE interface: Intel Corporation 82371AB/EB/MB PIIX4 IDE (rev 01) 0000:00:07.2 USB Controller: Intel Corporation 82371AB/EB/MB PIIX4 USB (rev 01) 0000:00:07.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 03) 0000:01:00.0 VGA compatible controller: S3 Inc. 86C270-294 Savage/IX-MV (rev 13)

haver@bibo:~$ lsmod Module Size Used by ipv6 249472 8 lp 9540 0 hostap_cs 61080 0 hostap 117636 1 hostap_cs ieee80211_crypt 4484 1 hostap orinoco_cs 11528 1 orinoco 42516 1 orinoco_cs hermes 6528 2 orinoco_cs,orinoco pcmcia 33444 6 hostap_cs,orinoco_cs firmware_class 8192 1 pcmcia snd_seq_dummy 2820 0 snd_seq_oss 33280 0 snd_seq_midi 7200 0 snd_seq_midi_event 6144 2 snd_seq_oss,snd_seq_midi snd_seq 50448 7 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq_midi_event snd_cs46xx 83272 0 snd_rawmidi 21024 2 snd_seq_midi,snd_cs46xx snd_seq_device 7180 5 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq,snd_rawmidi snd_ac97_codec 95484 1 snd_cs46xx snd_ac97_bus 2048 1 snd_ac97_codec snd_pcm_oss 50720 0 snd_mixer_oss 17664 1 snd_pcm_oss snd_pcm 84360 3 snd_cs46xx,snd_ac97_codec,snd_pcm_oss snd_timer 22276 2 snd_seq,snd_pcm 8250_pnp 8448 0 uhci_hcd 31504 0 snd 48612 11 snd_seq_oss,snd_seq,snd_cs46xx,snd_rawmidi,snd_seq_device,snd_ac97_codec,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer snd_page_alloc 8712 2 snd_cs46xx,snd_pcm 8250 21028 1 8250_pnp serial_core 18816 1 8250 3c59x 38440 0 mii 4736 1 3c59x yenta_socket 25100 6 rsrc_nonstatic 12160 1 yenta_socket pcmcia_core 37008 3 pcmcia,yenta_socket,rsrc_nonstatic floppy 57540 0

haver@bibo:~$ cat /dev/sndstat Sound Driver:3.8.1a-980706 (ALSA v1.0.10rc1 emulation code) Kernel: Linux bibo 2.6.14.4 #7 PREEMPT Sun Jan 1 18:34:16 CET 2006 i686 Config options: 0

Installed drivers: Type 10: ALSA emulation

Card config: Sound Fusion CS46xx at 0xe8100000/0xe8000000, irq 11

Audio devices: 0: CS46xx (DUPLEX)

Synth devices: NOT ENABLED IN CONFIG

Midi devices: 0: CS46XX

Timers: 7: system timer

Mixers: 0: Cirrus Logic CS4297A rev 4 haver@bibo:~$

haver@bibo:~$ uname -r 2.6.14.4

On Tue, 2006-01-03 at 19:17 +0100, Robert Jonsson wrote: Hello,

Since I don't know how often you would read the forum I send you a mail instead, hope you don't mind.

It's a strange error you got there, it's apparently not the debian changes that produce this.

I'm pondering the lines: name2route: <alsa...

I have a definite feeling that I have seen this before together with a crash problem.

What kind of soundcard do you have and do you know how it is presented in jack? What output does jack_lsp produce?

Regards, Robert

[Windfisch]

Submitted by nold on 2006-01-04 12:28:40

This report seems to be caused by (the one) Debian-specific patch we apply, and that we probably should just drop. The patch stems from 0.6 days and automatically drops the user into debug mode when muse is started without superuser capabilities. This used to work fine and was done in order to give users a chance to look at muse while defaulting to a secure non-suid install. Apparently, this is no longer a viable option these days. Anyway, in case you still want to debug the segfault, you can remove the suid bit from the muse executable and start with 'muse -d'.

Frank, please run 'dpkg-reconfigure -plow muse', select to install muse with superuser privileges, and see whether it works for you. Unfortunately, pre2 didn't work out for a number of people due to some timer issues. Later pre- releases are meant to cure them, but they're currently stuck in unstable due to a Qt toolchain issue on hppa. Anyway, pre5 will be available from unstable by tonight's mirror push. You can try whether the binary package installs on testing as well, though I think the Xrender dependency cannot be fulfilled at the moment. But in any case you should be able to grab the source package from unstable and rebuild it on your testing system.

[Windfisch]

Submitted by spamatica on 2006-01-04 12:41:57

Thank you Daniel,

Hopefully this issue can be closed then.

As you imply, dropping to debug mode to get muse started should not be needed anymore. MusE should handle more variants and if not possible to start, atleast provide a informative error message.

I will try tonight running with -d as you suggest. It's quite possible this setting is not working as we don't use it.

[Windfisch]

Submitted by fhaver on 2006-01-08 14:20:41

I upgraded the distribution to unstable now and it did not work as before. Yesterday the muse package was updated (maybe result of your work). I tried again and it did not work either. I did the 'dpkg-reconfigure -plow muse' and allowed it to run with superuser privileges as you recommended. Now it starts and complains in a box that jackd has not yet been started. I started jackd and afterwards muse which worked fine, except of the fact that I seem not to have the right settings yet for my 800MHz notebook: The sound is crackling if I am moving any windows arround (do you have a recommendation of jackd parameters for me?).

I tried to compile the CVS version of Muse to be able to give better bugreports, but I failed because it wanted to have qt4 installed but my Debian unstable seems to use qt3.

I am not sure if the bug I reported was really solved by removing the debug stuff, but certainly you found a circumvention for me. When playing with it I killed jackd and found Muse segfaulting because jackd faded away... Mhm, is this something you want to fix too? Or is it just me the as a user not doing things the way they should be done?

Thanks for all the help.

[Windfisch]

Submitted by fhaver on 2006-01-08 14:33:35

I was curious if the superuser stuff only changed the behaviour. I did the dpkg-reconfigure again and picked the option not running as superuser, started Muse again and it failed:

haver@bibo:~/Midi$ apt-get install muse E: Could not open lock file /var/lib/apt/lists/lock - open (13 Permission denied) E: Unable to lock the list directory haver@bibo:~/Midi$ muse No superuser privileges, using system timer fallback no locale /</usr/share/muse/locale> Trying RTC timer... fatal error: open /dev/rtc failed: No such file or directory Trying ALSA timer... got timer = 12 QObject::connect: No such signal PartCanvas::horizontalScroll(int) QObject::connect: (sender name: 'unnamed') QObject::connect: (receiver name: 'unnamed') Arranger::configChanged - no bitmap! starting with default template name2route: not found name2route: not found Arranger::configChanged - no bitmap! AlsaTimer::setTimerTicks(): requested freq 1024 Hz too high for timer (max is 250) freq stays at 250 Hz Segmentation fault (core dumped) haver@bibo:~/Midi$

Switching back to the superuser mode I get the jackd missing message and when starting jackd before it starts ok:

haver@bibo:~/Midi$ muse no locale /</usr/share/muse/locale> Trying RTC timer... fatal error: open /dev/rtc failed: No such file or directory Trying ALSA timer... got timer = 14 QObject::connect: No such signal PartCanvas::horizontalScroll(int) QObject::connect: (sender name: 'unnamed') QObject::connect: (receiver name: 'unnamed') Arranger::configChanged - no bitmap! starting with default template Arranger::configChanged - no bitmap! AlsaTimer::setTimerTicks(): requested freq 1024 Hz too high for timer (max is 250) freq stays at 250 Hz cca_open_socket: could not connect to host 'localhost', service '14541' cca_init: could not connect to server 'localhost' - disabling ladcca

Here the version I am currently using:

bibo:/home/haver# dpkg -l muse Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ Name Version Description +++-=================-=================-================================================== ii muse 0.7.1+0.7.2pre5-1 Qt-based midi/audio sequencer

[Windfisch]

Submitted by spamatica on 2006-01-08 19:13:23

Hi,
Responding to both messages:
for jack to behave well you either need to have big
buffers e.g. -p 1024 (possibly much bigger), or run jack
with realtime priority, the -R switch, possibly both if
you don't have a recent kernel.

That it works when you run as superuser seems consistent.
That it still does not work when running as a user is
puzzeling. Though I note that the text:
"No superuser privileges, using system timer fallback" is
still written so I wonder if the patch really is removed.

About building from CVS, there are several CVS versions,
the HEAD branch will become 1.0 in a while, what you want
is REL07.
Though, if you want to build it yourself I recommend
getting the pre5 source tree instead.

About killing jack and getting muse killed in the process. Was this reproducible? This should normally not happen.

Lastly, I welcome you to join our mailing list, I find it
a bit easier to respond and the information will still end
up in an public archive. (See the [Mail] link at the top)

/Robert

[Windfisch]

Submitted by spamatica on 2006-03-19 21:51:08

Is this still an issue?

[Windfisch]

Submitted by sf-robot on 2006-04-10 04:22:38

This Tracker item was closed automatically by the system. It was previously set to a Pending status, and the original submitter did not respond within 14 days (the time period specified by the administrator of this Tracker).