I use, XFree86 nv
Extras:
* some filehandling (non local root of testing) and
* run with min(SCHED_FIFO)+1 not max
- it is a CPU hog (80%) [the more CPU usage the less prio]
- show that it works to avoid creating a trend...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
diff -Naur latencyte
> where to report it or ask about it?
First - you mention virtual terminals
What video driver are you using? Nvidia?
If that is the case retest with xfree86 (nv)
Second - try to what is happening (Magic SysRq)
Third - 2.4.19 is kind of old now (not that many will be interested...),
upgrade and try to reproduce.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
On Tuesday 02 December 2003 02.50, Jack O'Quin wrote:
> Roger Larsson <[EMAIL PROTECTED]> writes:
> > But it is still a way to see that no client burns cycles where it should
> > not - jackd would not start (or stop). And you _can_ get fewer context
> > switche
obably sleep for
> max (1msec, period_msecs). agree?
Either that or (period_msecs + 1)
You should get a solid lockup if your period_msecs is lower than 1000 us ==
1 ms when running SCHED_FIFO... (unless you run clients at the same
priority and the kernel yields at poll - I don't think it does)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
engine->driver->period_usecs/1000);
- - -
if (poll (pfd, 1, poll_timeout) < 0) {
Isn't this dangerous?
* What happens if period_usecs < 1000?
Well poll_timeout gets 0 since period_usecs is an uint64_t
A poll with time out of zero will return immediately.
OK?
/RogerL
--
Roger Larsson
Skellefteå
Sweden
On Sunday 30 November 2003 05.03, Jack O'Quin wrote:
> Roger Larsson <[EMAIL PROTECTED]> writes:
> > > That's right. But, Paul and I have been working closely with this and
> > > don't have much faith in the correctness of the 2.4 scheduler.
> >
On Sunday 30 November 2003 03.19, Jack O'Quin wrote:
> Roger Larsson <[EMAIL PROTECTED]> writes:
> > If next client is a RT process (SCHED_FIFO/RR) then it WILL be selected
> > unless there is another RT process with higher priority, and lower than
> > jackd otherwi
e jack should handle the priorities for jack clients - priorities is a
relative issue anyway... Then jackd could use this scheme when there
is only one client to run - and the lower priority clients when there are
several.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
>
This was about switching threads within a single process - simulating
threads in a system where they are not handled by the kernel.
Not about switching from one user process to another.
> any thoughts? adding a syscall is a pretty trivial patch to create.
And then the hard part begins - trying to get it accepted by Linus :-)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
hare - but then you need to handle several versions too...
In this situation I do not mind /opt ...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
t; Yes. :)
echo 1 > /proc/sys/kernel/lowlatency ?
If you have enabled "Control low latency with sysctl"
It might be possible that something like lock break is needed...
Lots of good stuff is available at:
http://www.tech9.net/rml/linux/
But most are for 2.4 or early 2.5 kernels
/RogerL
--
Roger Larsson
Skellefteå
Sweden
ts looking at www.djcj.org/LAU/quicktoots
* Ecasound
* AlsaModularSynth
* Legasynth
* SpiralSynth
* Pd
* Soundtracker, cheesetracker
* Alsaplayer
* xmms
/RogerL
--
Roger Larsson
Skellefteå
Sweden
On Wednesday 19 November 2003 15.53, Jack O'Quin wrote:
> Roger Larsson <[EMAIL PROTECTED]> writes:
> > I have an idea!
> >
> > rt_monitor is running mlockall and I avoided
> > * dynamic memory allocation
> > * system calls
> > - If I split the
On Wednesday 19 November 2003 01.40, Jack O'Quin wrote:
> Roger Larsson <[EMAIL PROTECTED]> writes:
> > Problem is - why doesn't most distributions even ship with wrappers suid
> > to be able to start the application with SCHED_FIFO/RT/mlock?
> > - It is du
workarounds.
Another thing to try is to run memtest86 (SuSE has it in Lilo menu)
let it run for a while (full night)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
does rt_monitor print?
(It could be the memory leak and mlockall(FUTURE) problem,
As I do not check that yet.
Another alternative is a fast forking client - they are hard to stop)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
ke ANY guarantees on "worst case"!
>
> Otherwise, a good solution. Perhaps adequate for some applications.
But at the same time SCHED_FIFO is adequate for most applications.
See my point? :-)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
Like capabilities - do not check uid,
but it can fake uids too!)
+ Can use whatever kernel provides
+ Only requires the rt_monitor to be started by root - can be done by init
+ Protects the system from overuse (and lockups) due to bugs in code.
- Can not provide mlockall since it has no pid parameter - the monitor
can't do it for another process. (I would say that it is unlikely that pages
used in a tight audio loop would be thrown out - big buffers might...
Add additional page reads?)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
s no monitor running.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
rts.zip
Description: Zip archive
n only submit io when it gets
scheduled.
Default IO priority for a new process is 10.
Patch is against bk-current.
--
Roger Larsson
Skellefteå
Sweden
non volatile
pointer dst 5.0 s 4.9 s
array dst 4.5 s 4.1 s
/RogerL
--
Roger Larsson
Skellefteå
Sweden
.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
t; > I think this should work. An example of playing stuff using playobjects
> > and one of how to play a simple sinusoid is attached. You may have to
> > edit the Makefile.
> >
> >
> > *** eSafe scanned this email for malicious content ***
> > *** IMPORTANT: Do not open attachments from unrecognized senders ***
--
Roger Larsson
Skellefteå
Sweden
brary), too,
> as jackstart does.
You can request SCHED_FIFO/RT from my monitor.
Current implementation is working but it can be improved!
It uses static allocations intentionally.
The problem is to get it accepted - should it be added to jack(d)?
/RogerL
--
Roger Larsson
Skellefteå
Sweden
rt_monitor.zip
Description: Zip archive
overused those processes would run
out of their timeslice (SCHED_RR never runs out of their timeslice)
I think this feature would be pretty cool! And adding this for
latency sensitive bandwith limited streaming applications could
simplify lots of stuff for the default scheduler...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
that SCHED_USERFIFO were handled in a strict FIFO manner, no
priorities, then a running process of that class can only be suspended by a
SCHED_FIFO or SCHED_RR
Anyway, I think this patch has its place on linux-kernel. It will most likely
be rejected but it shows that people are interested in these issues...
(Who knows Ingo might get another bright idea... :-)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
> (sorry for the off-topic post, but i want people
> > on these lists to have the first chance at this)
> >
> > Rev 1.5 Hammerfall Digi9652
> > all cables
> > Windows/Mac driver CD
> > original packing materials
> >
> > current street price: about $580
> > asking price: $480 OBO + shipping
>
>
--
Roger Larsson
Skellefteå
Sweden
On Friday 28 February 2003 14:49, Fons Adriaensen wrote:
> Roger Larsson writes:
> > - - -
> > In a pure push model each processing step reads data from
> > one file/pipe/device, processes it, pushes it to a file/pipe/device
> > You get:
> >* lots
e hard to eliminate extra copying.
Note: Arts and the KDE multimedia framework does a lot of things right today.
It even goes one step furter since it moves the pluggins, including input and
output into one process - artsd. But currently it does not work perfectly
together with applications using
fo/kde-multimedia
---
--
Roger Larsson
Skellefteå
Sweden
Intel 386")
And at least use optimization level -O1 (use -O3 to get automatic inlining)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
On Friday 14 February 2003 16:18, Steve Harris wrote:
> On Fri, Feb 14, 2003 at 03:54:35 +0100, Roger Larsson wrote:
> > On Friday 14 February 2003 15:02, Steve Harris wrote:
> > > Hi all,
> > >
> > > I'm trying to hack up a quick app that sends MIDI clo
re out how to send MIDI clock pulses explicitly. The ALSA seq
> interface lets me set it up to go automatically, but I want to send a
> clock pulse every 1/24th of the buffer, so the sync is correct.
>
How about dividing the buffer in 24 frames. And send one MIDI event each
frame? Doable?
/RogerL
--
Roger Larsson
Skellefteå
Sweden
ne Control and so can be controlled from any MMC
controller, such as the Mackie
Digital 8 Bus mixer and many other modern digital mixers.
Authors:
Paul Davis <[EMAIL PROTECTED]>
Distribution: SuSE Linux 8.1 (i386)
--
Roger Larsson
Skellefteå
Sweden
> P.S. Where can I find the mentioned utility?
>
See my reply to the original message.
I forgot to change Subject... it is late...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
hen the
monitor kicked in.
Maybe it should only reduce the priority of those/that as a first level.
And if that did not work - reduce all...
But I am a little afraid of adding features in this kind of application...
Tomorrow... for now I send the current version... (it is quite small)
Start rt
the window manager keep them with a common width
and split height. Resize and Move all windows together.
- Possible?
/RogerL
--
Roger Larsson
Skellefteå
Sweden
http://www.musicdsp.org/
Needs help with their Linux information...
Any takers?
/RogerL
--
Roger Larsson
Skellefteå
Sweden
t rt prio
<- sleeps for 5 s
<- lockup!!! (for about 4 s)
<- rt_monitor reduces the prio, rt exits when detected
/RogerL
--
Roger Larsson
Skellefteå
Sweden
/* RT user.
Copyright (c) 2002 Roger Larsson <[EMAIL
ng very specific resources to be used or more or less
> exclusively by a single task, even when those resources more or less
> mean "all of the machine". this is not something that POSIX or Linux
> provides at this time.
>
> --p
>
>
I will give this monitor addition a try tonight.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
to intercept external calls - like fopen().
Recent messages show that we even can let the server compile
the plugin - "runtime / compiletime processing proof of concept"
And since it does the compilation it can chose which headers are allowed.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
RT processes.
* it should not reduce the priority of system RT processes since
that might be the reason for the DOS attack. [contradicion!]
* it should have a log for the system administrator to look in.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
not help you.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
/* RT monitor.
Copyright (c) 2002 Roger Larsson <[EMAIL PROTECTED]>
This program is free software; you can redistribute it and/or
modify it under the terms of version 2 of the GNU General Public
License as pub
that I have used is to start a timer when the need_reshed
flag is set, and stop it when schedule is called...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
On Sunday 14 April 2002 22.47, Jörn Nettingsmeier wrote:
> all in all, a nice distro, if a little bloated. if you're a
> pssionate do-it-yourselfer, i'd recommend against it, but if you
> want a quick start and can tolerate excessive eye-candy, suse80
> might be for you.
I am a do it myself - so
/www.cs.utah.edu/~regehr/hourglass
>
> Thanks,
>
> John Regehr
> Postdoc, School of Computing, University of Utah
--
Roger Larsson
Skellefteå
Sweden
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
-------
--
Roger Larsson
Skellefteå
Sweden
ere might be some rough edges though.
Comments, bug reports, suggestions are welcome,
Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
---
--
Roger Larsson
Skellefteå
Sweden
* or a better data structure, since it should be just sorted
+ * by expire time. later :-)
+*/
+ struct list_head fifo_r;
+ struct list_head fifo_w;
+
+ /*
+* available entries
+*/
+ struct list_head free_entry;
+
+ /*
+* I/O scheduler settings
+*/
+ unsigned long expires[2];
+ int batch;
+ int restart : 1;
+};
+
+/*
* use the /proc/iosched interface, all the below is history ->
*/
typedef struct blkelv_ioctl_arg_s {
--
Jens Axboe
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
---
--
Roger Larsson
Skellefteå
Sweden
ean that a 1kB DMA buffer empties
> in less than 1/100th of a second, but at least it should be < 200 irqs/sec
> rather than >400).
>
Yes, it is probably more reasonable - but if the soundcard can support it?
(I have a vision of lots of linux-audio-dev folks pulling out their new
soundcard and replacing it with their since long forgotten SB16...)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
Increase it
slightly if your execution time is really short. Decrease if long. Why?
Because your process is most likely the only RT process in the system
and will be selected before any non RT process. But if there is another
RT process, you need to cooperate somewhat.
--
Roger Larsson
Skellefteå
Sweden
On Wednesdayen den 14 November 2001 11.36, you wrote:
> Hello,
>
> In reply to D. Stimits and Roger Larsson about compiling
> an LL kernel for AMD Athlon without the 3DNOW optimizations
>
> I still get the undefined references to the
> *_mmx_* functions. I am perfect
much sense to me, and I am recompiling
> > (again) my kernel to make sure. Does anybody have a clue of
> > what might be going on?
> >
> > Maarten
--
Roger Larsson
Skellefteå
Sweden
'
> fs/fs.o: In function `block_symlink':
> fs/fs.o(.text+0x49b4): undefined reference to `_mmx_memcpy'
> fs/fs.o: In function `flush_old_exec':
> fs/fs.o(.text+0x834a): undefined reference to `_mmx_memcpy'
> fs/fs.o: In function `d_alloc':
> fs/fs.o(.text+0x11818): undefined reference to `_mmx_memcpy'
> fs/fs.o(.text+0x11d48): more undefined references to `_mmx_memcpy' follow
> make: *** [vmlinux] Error 1
--
Roger Larsson
Skellefteå
Sweden
On Friday 09 November 2001 09:35, Maarten de Boer wrote:
> Roger Larsson <[EMAIL PROTECTED]> wrote:
> > They are not bad either... (sub 10 ms)
>
> I think it is bad. I want the <3 ms :-)
> I will be added latency in the processing (fft, ifft)
> and it all adds up.
&
nels and patches, as indicated on the main
> page.
>
> fwiw,
> dave
--
Roger Larsson
Skellefteå
Sweden
m generation to actual
sending to soundcard, then add some time for output filters...)
> - - -
> used alsa-cvs from oct 20, oss = alsa's oss-emu
Tests with standard oss would be interesting too...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
am" - God (Exodus 3:14)
[or possibly "I will be", see
http://pub18.ezboard.com/fhavetheologywillarguetheology.showMessage?topicID=329.topic&index=22
I did a seach for Exodus 3:14 and found that page...]
--
Roger Larsson
Skellefteå
Sweden
w (2.4.12) proc parameter in /proc/irc/max_rate that I think is
slightly related - if a driver is in interrupt service to much. The interrupt
will be disabled and it will be polled instead...
(this is really for GigaBit ethernet drivers, they can generate so many
interrupts that no user space work gets done...)
/RogerL
--
Roger Larsson
Skellefteå
Sweden
EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
-------
--
Roger Larsson
Skellefteå
Sweden
returned...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
suitable read ahead for each sequential stream...
More about this later (including a patch) when the jury has returned...
/RogerL
--
Roger Larsson
Skellefteå
Sweden
gt;
> /s jsr
>
> On Thu, 14 Jun 2001, Joakim Verona wrote:
> > ... In my opinion, different applications
> > can never be assumed to be well-behaved and cooperative. Then we will
> > wind up with systems behaving like windows 3.11 and the old Macintoshes.
--
Roger Larsson
Skellefteå
Sweden
Do both csound and timidity run with MAX SCHED_FIFO prio?
Then this will be a big problem...
I guess most people has followed Bennos example code to the
exact word - including running max prio.
* Developers - you should not do this...
(It is fine when testing for the absolute best possible late
64 matches
Mail list logo
