23 Mar
2004
23 Mar
'04
8:10 p.m.
Hi all.
a simple stupid question: in muse the tempo is in percentage?
or is 100 really 100 bpm? what is the rationale here?
Secondly when starting muse as a normal user I get:
cannot open rtc clock /dev/rtc: Permission denied
how can I change the permission permanently?
kernel 2.6 with realtime.0.04
Regards
--
antoine rivoire <antoine.rivoire(a)ntlworld.com>
23 Mar
23 Mar
8:46 p.m.
On Wed, Mar 24, 2004 at 12:04:41AM +0000, antoine rivoire wrote:
Hi all.
a simple stupid question: in muse the tempo is in percentage?
or is 100 really 100 bpm? what is the rationale here?
No.. you can set the tempo in bpm in the transport window (among other
places)... press F11 to see it!
Secondly when starting muse as a normal user I get:
cannot open rtc clock /dev/rtc: Permission denied
how can I change the permission permanently?
kernel 2.6 with realtime.0.04
You need to build muse with the --enable-rtcap option on ./configure
James
9:07 p.m.
On Wednesday 24 March 2004 12:04 am, antoine rivoire wrote:
a simple stupid question: in muse the tempo is in
percentage?
or is 100 really 100 bpm? what is the rationale here?
It's BPM.
Secondly when starting muse as a normal user I get:
cannot open rtc clock /dev/rtc: Permission denied
how can I change the permission permanently?
kernel 2.6 with realtime.0.04
Don't know about 2.6 kernels. Have you tried setting muse suid root? I assume
you are launching as muse -R (?)
tim hall
24 Mar
24 Mar
3:51 a.m.
On Wed, Mar 24, 2004 at 12:49:50AM +0000, tim hall wrote:
What does muse -R do? I just run it as muse, and connect to jack. Not
sure what the realtime option does, just hangs on my computer.. unless I
change muse to suid (which is not necessary if it has been compiled with
rtcap). Don't worry about the kernel version. makes no difference.
James
On Wednesday 24 March 2004 12:04 am, antoine rivoire
wrote:
No its not (AFAIK).
When you start Muse, the default is 100%, 120 bpm.
a simple stupid question: in muse the tempo is in
percentage?
or is 100 really 100 bpm? what is the rationale here?
It's BPM.
Secondly when
starting muse as a normal user I get:
cannot open rtc clock /dev/rtc: Permission denied
how can I change the permission permanently?
kernel 2.6 with realtime.0.04
Don't know about 2.6 kernels. Have you tried setting muse suid root? I assume
you are launching as muse -R (?)
12:45 p.m.
On Wednesday 24 March 2004 7:44 am, James Stone wrote:
On Wed, Mar 24, 2004 at 12:49:50AM +0000, tim hall
wrote:
Right, I was slightly off-beam.
You set the tempo using Edit>Mastertrack>... (in bpm) - ignore this %tempo
widget.
I guess that the rationale for this control is for when you want to record
something at half speed or whatever, without altering the Master track's bpm
settings. The default is 120bpm and this tempo fine-tune is obviously set to
100%. The documentation for MusE is virtually non-existent, what is not
obvious is that this very prominent widget is not designed for setting the
tempo - It's for altering the tempo in % steps. Hope that makes sense.
On Wednesday 24 March 2004 12:04 am, antoine
rivoire wrote:
No its not (AFAIK).
When you start Muse, the default is 100%, 120 bpm. a simple stupid question: in muse the tempo is in
percentage?
or is 100 really 100 bpm? what is the rationale here?
It's BPM. What does muse -R do?
Runs it in realtime, if you're not connected via Jackd.
Maybe I'm kludging a bit, I'm still using 0.6.1, which doesn't work well with
Jackd. (Help not required, I know why). I'm looking forward to upgrading.
I'm a little wary of rolling my own, but MusE will be a likely candidate if I
decide to bite the bullet.
cheers
tim hall
4:05 p.m.
Hi,
onsdagen den 24 mars 2004 17.12 skrev tim hall:
On Wednesday 24 March 2004 7:44 am, James Stone
wrote:
Thanks to Joachim Schiele MusE does now have a Wiki, and we are already
preparing some content for it. Hopefully we will have prepared some more
useful documentation in the near future. Interested people are quite welcome
to add content! :)
On Wed, Mar 24, 2004 at 12:49:50AM +0000, tim
hall wrote:
Right, I was slightly off-beam.
You set the tempo using Edit>Mastertrack>... (in bpm) - ignore this %tempo
widget.
I guess that the rationale for this control is for when you want to record
something at half speed or whatever, without altering the Master track's
bpm settings. The default is 120bpm and this tempo fine-tune is obviously
set to 100%. The documentation for MusE is virtually non-existent, On Wednesday 24 March 2004 12:04 am, antoine
rivoire wrote:
No its not (AFAIK).
When you start Muse, the default is 100%, 120 bpm. a simple stupid question: in muse the tempo is in
percentage?
or is 100 really 100 bpm? what is the rationale here?
It's BPM. what is
not obvious is that this very prominent widget is not designed for setting
the tempo - It's for altering the tempo in % steps. Hope that makes sense.
Right. MusE does not run /well/ if it does not have access to rtc and realtime
capabilities. Timing will be off.
What does muse -R do?
Runs it in realtime, if you're not connected via Jackd. Maybe I'm kludging a bit, I'm still using
0.6.1, which doesn't work well
with Jackd. (Help not required, I know why). I'm looking forward to
upgrading.
You should definitively upgrade, 0.6.3 is heaps better than 0.6.1 at audio and
jack.
I'm a little wary of rolling my own, but MusE will
be a likely
candidate if I decide to bite the bullet.
You mean that there are none available for your system? Not that it's that
difficult to build MusE. Autoconf is always a problem though... :-/
Regards,
Robert
4:20 p.m.
What does muse -R do?
Runs it in realtime, if you're not connected via Jackd.
4:49 p.m.
Hi,
Yep, it's a definitive bug. '-R' does not handle very well the special case
when realtime capabilities are not allowed.
Right. MusE does not run /well/ if it does not have access to rtc and
realtime capabilities. Timing will be off.
Yes.. I had found this before, but I _think_ when you build muse with
rtcap, it automatically uses realtime. The -R option is a bit unclear.
For example, in my system when running muse (without jack), it does not
start, I get:
no locale <muse_en_GB>/</usr/local/share/muse/locale>
no JACK audio server found
ALSA: samplerate 44100 not available, using 48000
WARNING: Cannot lock memory:: Operation not permitted
WARNING: Cannot lock memory:: Operation not permitted
cannot set tick on /dev/rtc: Permission denied
precise timer not available
(muse was compiled for jack though, and with rtcap support).
when I run it with jack running, it starts fine:
no locale <muse_en_GB>/</usr/local/share/muse/locale>
JACK: sample rate changed: 44100
Also timing seems fine (as long as I have given the kernel option
sysctl -w dev.rtc.max-user-freq=1024)...
When I try to run it connected to jack with the -R startup option, it
does not start and gives me the error:
no locale <muse_en_GB>/</usr/local/share/muse/locale>
JACK: sample rate changed: 44100
creating watchdog thread failed:: Invalid argument
creating thread failed:: Illegal seek
I'm not certain, of the top of my head, what this is, which version of MusE is
it? I run MusE 0.6.3 always with -R. Is jack perhaps not started with -R ?
(it's pretty much a requirement there too)
My question is: do I need to start muse with the -R option, and what
does it do?
-R instructs MusE to set certain threads to realtime-scheduling (SCHED_FIFO)
to improve performance. This applies to the audio-thread and the midi-thread.
When you are running through Jack the audio-thread is implicitly set to
SCHED_FIFO thus removing the need. I'm not sure, however, if the same applies
to the midi-thread. In MusE-cvs '-R' is removed (since MusE now requires
Jack) so, there the question is moot.
In any case, if you are running with jack at a fairly small buffersize (<512)
the jitter will probably not be very noticable even if it's there. To make
sure there are no problems lurking about, run jack with 2048 buffers and play
something monotone, e.g. a hihat line.
/Robert
6:18 p.m.
On Wed, Mar 24, 2004 at 09:44:51PM +0100, Robert Jonsson wrote:
Hi,
Yep, it's a definitive bug. '-R' does not handle very well the special case
when realtime capabilities are not allowed.
Yep.. its 0.6.3 and jack is run through qjackctl with realtime option.
Muse is compiled with rtcap so it is not suid (thought it was safer..)
but if I do change it to suid and then start with -R, it starts okay
with no error messages...?? Maybe rtcap does not enable enough realtime
capabilities!!?
Any hints?
James
Right. MusE does not run /well/ if it does not have access to rtc and
realtime capabilities. Timing will be off.
Yes.. I had found this before, but I _think_ when you build muse with
rtcap, it automatically uses realtime. The -R option is a bit unclear.
For example, in my system when running muse (without jack), it does not
start, I get:
no locale <muse_en_GB>/</usr/local/share/muse/locale>
no JACK audio server found
ALSA: samplerate 44100 not available, using 48000
WARNING: Cannot lock memory:: Operation not permitted
WARNING: Cannot lock memory:: Operation not permitted
cannot set tick on /dev/rtc: Permission denied
precise timer not available
(muse was compiled for jack though, and with rtcap support).
when I run it with jack running, it starts fine:
no locale <muse_en_GB>/</usr/local/share/muse/locale>
JACK: sample rate changed: 44100
Also timing seems fine (as long as I have given the kernel option
sysctl -w dev.rtc.max-user-freq=1024)...
When I try to run it connected to jack with the -R startup option, it
does not start and gives me the error:
no locale <muse_en_GB>/</usr/local/share/muse/locale>
JACK: sample rate changed: 44100
creating watchdog thread failed:: Invalid argument
creating thread failed:: Illegal seek
I'm not certain, of the top of my head, what this is, which version of MusE is
it? I run MusE 0.6.3 always with -R. Is jack perhaps not started with -R ?
(it's pretty much a requirement there too)
25 Mar
25 Mar
3:34 a.m.
Yep.. its 0.6.3 and jack is run through qjackctl with
realtime option.
Muse is compiled with rtcap so it is not suid (thought it was safer..)
but if I do change it to suid and then start with -R, it starts okay
with no error messages...?? Maybe rtcap does not enable enough realtime
capabilities!!?
This is a possibility, it's a bug in that case, I have a vague memory of
something similar being reported some time ago.
As for running MusE suid, this is in itself not a problem unless your machine
is compromised, in which case you someone use it to lock up the machine
(which is equally true for rtcap:ed muse). The prospect that someone has
bothered to find a security hole in MusE to gain control of a machine through
it's suid usage seems a tad unlikely.
/Robert
11:47 a.m.
Robert Jonsson <robert.jonsson(a)dataductus.se> writes:
Have you tried using the realtime LSM instead of rtcap? I think it
ought to work, but haven't tested it with Muse, myself. That is safer
than setuid root, IMHO.
Yep.. its
0.6.3 and jack is run through qjackctl with realtime option.
Muse is compiled with rtcap so it is not suid (thought it was safer..)
but if I do change it to suid and then start with -R, it starts okay
with no error messages...?? Maybe rtcap does not enable enough realtime
capabilities!!?
This is a possibility, it's a bug in that case, I have a vague memory of
something similar being reported some time ago. As for running MusE suid, this is in itself not a
problem unless your machine
is compromised, in which case you someone use it to lock up the machine
(which is equally true for rtcap:ed muse). The prospect that someone has
bothered to find a security hole in MusE to gain control of a machine through
it's suid usage seems a tad unlikely.
There are a few people on the net with way too much time on their
hands. :-(
Regards,
--
joq
3:39 p.m.
On Thu, Mar 25, 2004 at 09:41:55AM -0600, Jack O'Quin wrote:
Robert Jonsson <robert.jonsson(a)dataductus.se>
writes:
Have you tried using the realtime LSM instead of rtcap? I think it
ought to work, but haven't tested it with Muse, myself. That is safer
than setuid root, IMHO.
Does the realtime LSM work with 2.4.x kernels ('cos thats what I'm using)?
James
Yep.. its
0.6.3 and jack is run through qjackctl with realtime option.
Muse is compiled with rtcap so it is not suid (thought it was safer..)
but if I do change it to suid and then start with -R, it starts okay
with no error messages...?? Maybe rtcap does not enable enough realtime
capabilities!!?
This is a possibility, it's a bug in that case, I have a vague memory of
something similar being reported some time ago. 
3:55 p.m.
On Thu, Mar 25, 2004 at 07:32:02PM +0000, James Stone wrote:
On Thu, Mar 25, 2004 at 09:41:55AM -0600, Jack
O'Quin wrote:
You need the security module infrastructure which is included in the
2.6.x series. There is a large patchset (from the NSA's SE-linux project)
that provides this for the 2.4.x kernel. But, I don't know of anyone who
is using the realtime lsm that way. (I haven't been following the lists
closely for the past couple of weeks, so I may have missed something.)
-Eric Rz.
Robert Jonsson
<robert.jonsson(a)dataductus.se> writes:
Have you tried using the realtime
LSM instead of rtcap? I think it
ought to work, but haven't tested it with Muse, myself. That is safer
than setuid root, IMHO.
Does the realtime LSM work with 2.4.x kernels ('cos
thats what I'm using)?
James Yep.. its
0.6.3 and jack is run through qjackctl with realtime option.
Muse is compiled with rtcap so it is not suid (thought it was safer..)
but if I do change it to suid and then start with -R, it starts okay
with no error messages...?? Maybe rtcap does not enable enough realtime
capabilities!!?
This is a possibility, it's a bug in that case, I have a vague
memory of
something similar being reported some time ago.
4:05 p.m.
James Stone <stone1(a)btinternet.com> writes:
Does the realtime LSM work with 2.4.x kernels
('cos thats what I'm using)?
No. Not unless you install that large SELinux backport patch. If you
do that, it should work, but no one has tested it in that environment.
I consider that overkill for running this small module. For 2.4, I
would recommend applying the two-line capabilities patch, instead.
Sorry for the digression. Someone mentioned 2.6 earlier in this
thread, and I was curious if anyone had tried Muse with the LSM yet.
--
joq
7641
days inactive
7643
days old
linux-audio-user@lists.linuxaudio.org
13 comments
6 participants
participants (6)
-
antoine rivoire -
Eric Dantan Rzewnicki -
Jack O'Quin -
James Stone -
Robert Jonsson -
tim hall