RE: [linux-audio-user] Bristol segfaults when run on real-time kernelwhile works

From: Glen Kirkup <handstandnosemanual@email-addr-hidden>
Date: Mon Dec 04 2006 - 04:08:03 EET

Thank you for your quick reply.

>From: "Nick Copeland" <nickycopeland@email-addr-hidden>
>To: handstandnosemanual@email-addr-hidden
>Subject: RE: [linux-audio-user] Bristol segfaults when run on real-time
>kernelwhile works fine on regular.
>Date: Sun, 03 Dec 2006 18:46:23 +0100
>
>Hi Glen,
>
>I am not sure what the exact cause it, but the audio engine (bristol) could
>not open its TCP socket interface:
>
>>Problem with bind
>>Could not open control listening socket: 0
>>No controlling socket available: anticipating MIDI
>
>This is when the bristol process attempts to open a TCP socket for
>communication to allow the GUI to connect and drive the synth parameters,
>and it fails. The engine stays active and decides to default to a physical
>MIDI interface and expects control signals from there - this also fails for
>other reasons, and the engine exits. The seg fault of the GUI is a side
>effect of not being able to connect to the engine via the TCP socket (it
>cannot connect via MIDI). I can fix the segmentation fault quite easy but I
>will have to look at bit further to find out the reason the engine cannot
>open its control port.
>
>Now the engine appears to want to open sock id '0' which is unusual - the
>default is 5028, and as you can see the GUI does appear to try to connect
>to this port, expecting the engine to be listening. Do you run the same
>software with the two different kernel releases? That should not be a
>problem as bristol is all done in user space, not kernel space.
>
>>./startBristol: line 188: [: too many arguments
>
>I will look into why this is reported, it may be related.
>
>Do you always use the jack audio interface with both kernels?
>
>Regards,
>
>Nick
>

When trying different port options I no longer receive any error messages
before the segfault and everything seems to connect fine. I've found that
using the -jack option with any other option while jack is running is what
causes both './startBristol: line 188: [: too many arguments' and the
program choosing port 0 by default. When running in the non-RT kernel, where
Bristol works as expected, gives './startBristol: line 188: [: too many
arguments' when jack is running and -jack with any other option is used,
though selects the correct default port and works fine. All options I have
tried still result in the segfault on the RT kernel (./startBristol: line
272: 3877 Segmentation fault brighton $*), though, which is the only
real error message left considering that the 'too many arguments' problem
does not effect the working of Bristol on the other kernel.

I use the same software on both kernels and the jack audio interface for low
latency.

p.s. Thank you for Bristol, it's a great joy to play with and it being free
is wonderful.

Thanks,
Glen.

>>From: "Glen Kirkup" <handstandnosemanual@email-addr-hidden>
>>Reply-To: A list for linux audio users
>><linux-audio-user@email-addr-hidden>
>>To: linux-audio-user@email-addr-hidden
>>Subject: [linux-audio-user] Bristol segfaults when run on real-time
>>kernelwhile works fine on regular.
>>Date: Sat, 02 Dec 2006 22:08:27 +0000
>>
>>Unfortunately I've recently come across a problem when running the
>>startBristol script. I switch between two kernels (CK for desktop and
>>RT7-CK for real-time music work). When within the CK kernel the program
>>runs as expected, but when in the RT7-CK kernel, as both user and root,
>>the program crashes and outputs this message with any selection of synth,
>>audio driver and port:
>>
>>[root@email-addr-hidden bin]# ./startBristol -jack
>>./startBristol: line 188: [: too many arguments
>>spawning midi thread
>>parent going into idle loop
>>midi sequencer
>>Problem with bind
>>Could not open control listening socket: 0
>>No controlling socket available: anticipating MIDI
>>connected to :0 (814b2f0)
>>Error opening control device, exiting midi thread
>>display is 1280 by 800 pixels
>>Window is w 1280, h 800, d 24, 0 0 0
>>Using TrueColor display
>>masks are ff0000 ff0000 ff0000
>>Initialise the arp2600 link to bristol: 8153348
>>hostname is localhost, bristol
>>port is 5028
>>./startBristol: line 272: 3877 Segmentation fault brighton $*
>>-engine
>>[root@email-addr-hidden bin]# parent exiting
>>
>>[root@email-addr-hidden bin]#
>>
>>As a Linux novice I do not know what could cause this. I changed the
>>permissions of the entire bristol-0.9.5 directory to '777' after my
>>preliminary tests, but this did not help. I have not experienced problems
>>with other audio software using this kernel. If anyone could give
>>assistance then I would be extremely grateful.
>>
>>Thanks,
>>Glen
>>
>>_________________________________________________________________
>>Windows Live™ Messenger has arrived. Click here to download it for free!
>>http://imagine-msn.com/messenger/launch80/?locale=en-gb
>>
>
>_________________________________________________________________
>Express yourself instantly with MSN Messenger! Download today it's FREE!
>http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/
>

_________________________________________________________________
Find Singles In Your Area Now With Match.com! msnuk.match.com
Received on Mon Dec 4 04:15:03 2006

This archive was generated by hypermail 2.1.8 : Mon Dec 04 2006 - 04:15:03 EET