[PlanetCCRMA] rt-kernel and firewire

Fernando Lopez-Lezcano nando at ccrma.Stanford.EDU
Wed May 2 16:15:05 PDT 2012

On 05/02/2012 01:49 PM, barb wrote:
> hi!
> i am new on this mailinglist and only here to investigate one thing i am
> working on currently. to connect the firewire saffire focusrite to my
> fc16 machine i am using the 3.0.17-1.rt33 kernel.
> the problem is, that i can only start up jack with firewire drivers (i
> use the new stack) when i am root. of course all the programms
> connecting to jack afterwards have to be started as root as well, which
> is not what i am aiming at.

Hmmm, I have not used firewire for a while but you might be missing the 
right permissions on /dev/raw1394 (or an equivalent device - check which 
device files get created in /dev when you plugin the card). Anything you 
can see in /var/log/messages that might be relevant? Or in the output of 
ffado? (ie: permissions are failing for what?)

I don't think you need to rebuild the kernel, this appears to be a 
device permissions problem.

-- Fernando

> if i start them as user, i get a 'permisison denied' on the libffado. i
> checked all the permissions and am finally part of any group that exists
> on the system without success. the /dev/fw0 is set on a group, which i
> am part of (rw permissions for u+g).
> i found a manual proposing to make a custom kernel and to disable
> GROUP_SCHED in it, as otherwise realtime scheduling as normal user would
> be not possible.
> -->  http://en.gentoo-wiki.com/wiki/JACK#Kernel_Configuration
> so i did. as i am not really experienced in compiling my own kernels i
> awfully failed with conflicts with installed kernel-headers on the
> rt-custom-kernel and also saying, that i have a newer version installed
> (but i did change the kernel.spec and the build-id).
> so my questions are if i really need a newly compiled rt kernel with
> GROUP_SCHED disabled to get firewire running under a user or if i am
> doing something else wrong. and in case i need to compile the kernel,
> where can i get help for very probably wrongdoing from my side from
> experienced kernel compilers?
> i'd be happy for advise on this issue
> thank you very much
> b#

More information about the PlanetCCRMA mailing list