Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally working

Moderators: mdrjr, odroid

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby adbensi » Tue Apr 21, 2015 6:21 pm

Humm
I make a clone of the repository, then You can download it from
https://github.com/Clouded/linux-rt-odr ... 10.72-rt77
Same version ;)
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Wed Apr 22, 2015 8:32 am

Fascinating. I was looking for an answer about the system freezing and there was lots of talk about hdmi to vga converters causing freezes. So I logged it remotely with ssh, disconnect the hdmi convetor and I was able to clone the source without a lock up. Cool. Before doing that I had tried cloning 4 times and system froze up every time. So that should be solvable. Now to get the rt kernel going, and fix the network problem I should be home free.

Thanks for cloning the repository. It should be faster to download than to clone. Takes me about 1/2 hour. (I don't think my network connection to github is very fast.

mark.
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Thu Apr 23, 2015 12:43 am

Ok,
I started from a clean system, and downloaded the zip file of your cloned repository. I extracted the repository and successfully built the modules. But when I try to build the kernel it fails with:
perl: warning: Falling back to the standard locale ("C").
drivers/amlogic/ethernet/am_net8218.c:2781:13: error: ‘ethernet_suspend’ undeclared here (not in a function)
.suspend = ethernet_suspend,
^
drivers/amlogic/ethernet/am_net8218.c:2782:13: error: ‘ethernet_resume’ undeclared here (not in a function)
.resume = ethernet_resume,
^
make[3]: *** [drivers/amlogic/ethernet/am_net8218.o] Error 1
make[2]: *** [drivers/amlogic/ethernet] Error 2
make[1]: *** [drivers/amlogic] Error 2
make: *** [drivers] Error 2

The code it is building is this:
#ifdef CONFIG_SUSPEND
.suspend = ethernet_suspend,
.resume = ethernet_resume,
#endif

ethernet_suspend is a function but comes up undefined. That must mean that CONFIG_PM is true because that code looks like this:
#ifndef CONFIG_PM
static int ethernet_suspend(struct platform_device *dev, pm_message_t event)

The other label ethernet_resume is not defined at all. If CONFIG_SUSPEND was false this would not fail. If CONFIG_PM was false, ethernet_suspend would not faile but ethernet_suspend would still be undefined.
Am I missing a step, or file or is this file bad?
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Thu Apr 23, 2015 1:05 am

I see these values are both defined in .config as yes. These are power management options. Don't I want these disabled anyway?
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Thu Apr 23, 2015 4:37 am

I did a git clone and tried it and it looks like it working. Finished all modules, building kernel now. (takes a while).

Mark.
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby robroy » Thu Apr 23, 2015 5:34 am

markwinger wrote:I did a git clone and tried it and it looks like it working. Finished all modules, building kernel now. (takes a while).

Mark.

I have been waiting for this day for a while! Nice job Mark :)
Rob Roy, Editor-in-Chief of ODROID Magazine (submit an article | make a donation)
User avatar
robroy
 
Posts: 3378
Joined: Mon Feb 25, 2013 2:29 am
Location: San Francisco, California
languages_spoken: English
ODROIDs: Q2, X, X2, U2, U3, XU-E, XU3, XU4, C1, C1+, C2

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Thu Apr 23, 2015 8:39 am

Something went wrong with the setup. After completing the process (and finding a couple of mistakes in the procedure) and completed the install. I then rebooted and:
Linux odroid 3.10.72-78 #1 SMP PREEMPT Mon Mar 23 23:14:36 BRT 2015 armv7l armv7l armv7l GNU/Linux

Shouldn't I see -rt77 in the version?
realtimeconfigquickscan says Kernel with Real-Time Preemption... not found - not good

So now what?
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby adbensi » Thu Apr 23, 2015 9:02 am

Did you copy the rt_config?
cp arch/arm/configs/odroidc_rt_defconfig .config
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Thu Apr 23, 2015 9:11 pm

I think I did the copy. So I did it again making sure I did the copy. Same result. Still get this:

odroid@odroid:~$ uname -a
Linux odroid 3.10.72-78 #1 SMP PREEMPT Mon Mar 23 23:14:36 BRT 2015 armv7l armv7l armv7l GNU/Linux
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Thu Apr 23, 2015 9:25 pm

I figured it out:
Linux odroid 3.10.72-rt77 #3 SMP PREEMPT RT Thu Apr 23 14:35:01 ACST 2015 armv7l armv7l armv7l GNU/Linux

The problem was I follow letter for letter the process in the thread. There was a few mistakes. At the end it said to copy the files to /boot. It should have read /media/boot.

Here is what I did:
sudo ln -s /boot /media/boot2
sudo cp ./arch/arm/boot/uImage /boot
sudo cp ./arch/arm/boot/dts/meson8b_odroidc.dtb /boot
sudo cp ./uInitrd /boot

There was a spelling error too which I figure out earlier:
make firmeware. Should spell firmware. (all references to firmware were spelled wrong.)

It works but as we know the network is almost useless. Tonight I will see if I can get jack running low latency.

Mark
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Fri Apr 24, 2015 12:29 am

I need to get the onboard networking going on this. Could someone who has been working on this tell me where you are on this? I would like to help. I am a c/c++ programmer but never worked on linux device drivers(Or kernels either). So if someone please point me in the right direction and let me know what has been done already so I don't duplicate past efforts on this?

Thanks Mark.
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby patoberli » Fri Apr 24, 2015 6:00 am

Check out this PDF: http://csabakiraly.com/files/preprints/ ... API-cr.pdf
I haven't read it fully, but it shows some variants of RT kernels and it's influence on network performance. There are very big differences (page 5, right graph) and it seems to help if you do some RT55 (what ever that is, didn't read that much).
-
pato
patoberli
 
Posts: 401
Joined: Mon Feb 23, 2015 6:32 pm
Location: Switzerland
languages_spoken: english, german
ODROIDs: C1 with RTC, Bluetooth 4.0, Wireless stick 4.0

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Fri Apr 24, 2015 11:07 am

I'll read the article tonight.

The symptoms are strange. One the RT kernel was installed, it be comes impossible to use remotely. Using remote desktop it takes probably 2 minutes to paint the screen. I can open a terminal session, type ls and it sits until dbp times out. But it work going out. I can browse the net with firefox. Not screaming fast but it does work.

The rt kernel does not help with Jack wither. If I start jack, I cannot stop it. I have to reboot. If I don't the system will freeze up in a minute or to. I cannot confirm if disconnection the hdmi adaptor will help because I cannot login remotely. Right now the system is a about a useful as brick.

Getting really frustrated......
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby patoberli » Fri Apr 24, 2015 3:56 pm

RT is really special about how it handles IRQs and the system resources. The drivers have to be developed for this, or you will notice many strange things. I fear this is happening here at the moment. There are reasons why traditional RT operating systems had a very limited hardware compatibility list (HCL), often with a very rudimentary graphical output (although a very fast if the hardware was on the HCL).
Anyway, that document shows some ways on how to change certain aspects of the RT kernel, to get better speed out of devices that work with buffers (like a network card).
-
pato
patoberli
 
Posts: 401
Joined: Mon Feb 23, 2015 6:32 pm
Location: Switzerland
languages_spoken: english, german
ODROIDs: C1 with RTC, Bluetooth 4.0, Wireless stick 4.0

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Fri Apr 24, 2015 11:00 pm

At this point I will be backing away from the odroid-c1 and working on the bananna pi. Even if the networking was perfect, the system freeze when running jack is an absolute show stopper. Since it works perfectly on the banana pi, I know it's an odroid. And since the odroid-c1 also shows the problem on the standard kernel, I know the rt kernel problems are unrelated. The freeze problem is probably related to usb audio drivers but I have no idea how to debug. If some one has any ideas on how to identify the source of the deadlock, I am all ears.

Since there seems to be little or no interest in pro-audio work in this forum I see no point in waiting for a fix. I sent an email to support and they don't seem to understand my problem. (Probably a language barrier). Since the standard os for the banana pro seems rock solid, I will at least be moving forward for a good starting point instead of fighting pre-existing problems.

So interrupt - to - banana-pro for now.

Mark.
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Fri Apr 24, 2015 11:23 pm

Correction:

Ignore the statement I made in my last post:" I sent an email to support and they don't seem to understand my problem." That was for something else, not the odroid. (I am trying to do too many things going on at once).

Mark
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Mon Apr 27, 2015 10:52 pm

Simple update on my project. To do a proof of concept on what I am doing, I took an old retired windows desktop with a single core amd 1.6g processor and installed ubuntu-studio on. With this I was easily able to get my usb interface to operate at less than 4ms latency. But that system did not have gig ethernet.

Since I will need gig ethernet, I tried using live ubutuntu studio on 2 machines. I configured a master system to use audacious to play media, then routed 18 channels of audio to the slave and then same routed all 18 channels audio back to the master on which I monitored the music with the usb device (Focusrite 18i20). Played music for 6 hours continuously with only 3 xruns. So I was sending 34 total channels across the network simultaneously. These machines are faster than the odroid(Master 3.5g 6 cores, slave 4g 4 cores), but based on other testing I am confident that both the odroid-c1 and banana-pro could handle either end.

The only thing I am missing is a stable low latency kernel. The banana is rock solid stable, but no low latency/rt kernel available. This rt kernel for the odroid-c1 is functional but the system locks up when I try to use low latency audio. I think there is a usb audio driver issue because this problem is evident on the both the rt kernel and the standard kernel. Also, the ethernet is barely functional in the rt kernel.

Trying to decide my next step.

Mark
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby origin » Tue May 12, 2015 11:18 pm

Hi,
I have build a rt kernel with your source code on my C1.
But the problem I am facing is : I cannot apt-get install any packages. It alway stop at conneting to ports.ubuntu.com.
As you mentioned, this rt kernel has low built-in ethernet performance. Is it related to this problem?
Have you ever faced this problem Or do you have any clue about this problem?

Thanks.
origin
 
Posts: 3
Joined: Mon May 11, 2015 11:30 pm
languages_spoken: english
ODROIDs: odroid c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Wed May 13, 2015 2:27 am

I have succeded in proving the concept using a banana pro. I found igorpec on the banana forums who was doing kernell builds. With his scripts I built a low latency kernel. (not Realtime). The mainline kernel contains the parts of the realtime patch to support this. I built 3.19.7 and it works great for me on the banana. I would still like to do this on the odroid though. Since my project requires the network (at better that 100 base_t) the state of this kernel is a problem.

Does anyone know how to get the device specific patch information needed to use a mainline source for the kernel build? Is that currently possible?
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby rooted » Thu May 14, 2015 4:02 am

You should be able to build and use mainline, you just wont have hdmi. I successfully built and booted the 4.0 kernel on the XU3.

Do you need the display or just networking, CPU, and USB?
User avatar
rooted
 
Posts: 5174
Joined: Fri Dec 19, 2014 9:12 am
Location: Gulf of Mexico, US
languages_spoken: english
ODROIDs: C1, C1+, C2
XU3 Lite, XU4
N1
VU7+
HiFi Shield 2
Smart Power (original)

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Thu May 21, 2015 5:43 pm

I will need a physical display eventually. I could use rdp short term, but in actual use the device will require it's own user interface.

I am having more basic problems right now. I have no networking at all on the c1. If I boot up the ubuntu image (ubuntu-14.04.2lts-lubuntu-odroid-c1-20150401.img) no networking works at all. Trying to ping I get an error "netork is unreachable". ls /dev/eth* shows no devices, but this did work originally. It appears that my hardware has failed.
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Wed Jun 10, 2015 5:14 am

This thread has been quiet for a while.

I am still going back and forth between rpi rev2, banana pro, orange pi plus, and the odroid-c1 for my networked multichannel low latency audio application. Status: Banana pro was the first to work but has trouble resampling audio in realtime with just a low latency kernel(probably would work with rt kernel). Orange pi I have had not success at all. Odroid-c1 works sort of but without a good RT kernel it appears I cannot get low latency without serious xrun problems. Raspberry pi I have had the best success with. Using the navio-rt raspbian image it works quit well, but the network shares the usb bus so audio and ethernet are fighting for bandwidth.

The raspberry is the only one I currently have a working RT kernel for. At first it looked like just low latency was enough but apparently not. The odroid-c1 is the best option hardware wise but I need an rt kernel with good network performance. Last time I checked tried this rt-kernel for odroid was network performance was almost non-existent. (The ethernet completely failed on that board so there may have been hardware problems at the time. I now have a new odropid-c1).

Has there been any progress on getting the ethernet problem fixed in the rt kernel?

Mark.
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby adbensi » Wed Jun 10, 2015 7:10 am

I do not mind to try fix the ethernet driver because I am concerned with the application. I think at this time using a USB dongle is an option that doesn't cost horrors and solve the problems. Try this too if You like odroid-c1 with this version of the kernel.
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby markwinger » Wed Jun 10, 2015 10:30 pm

I replaced my odroid-c1 due to hardware. Networking quit working completely on the hardkernel disto. WIth the new odroid I ordered a network dongle so I could try it.

Last night I loaded the RT kernel I had built a few weeks ago(rt77) and plugged in the network dongle. It worked with no problems. I am running networked audio, testing very low latency of 4.5 ms with very few xruns. 18 channels each way. Audio originating from a x86 3.5 gig 6 core ubuntu studio distro. On the network I have the banana pro echoing 18 channels back to the x86 were I can monitor the audio. I then added the ocdroid-c1 to the network. I have it receiving the 18 channels, then resampling 2 channels, output to a usb stereo audio dongle(this resampling it pretty cpu intensive). Audio sounds good. Initial testing is working quite well.

Longer term, I want the odroid connected to the usb audio device (the 18 in 20 out) to route the data. However, the usb network dongle sharing with that many usb audio channels will likely not work well, so I would like to get the onboard networking going.

I am not familiar at all with device drivers on linux, and just recently been trying to build kernels (like this one I built the repository you created). But I do have years of c/c++ experience (application development) so if I can do anything to help solve the network problem let me know.

Mark.
markwinger
 
Posts: 26
Joined: Tue Apr 07, 2015 12:56 am
languages_spoken: english
ODROIDs: c1

XU3 PREEMPT_RT anyone?

Unread postby corvid » Sun Jun 14, 2015 2:59 pm

Has anyone tried building for an XU3 with the preempt_rt patches?
corvid
 
Posts: 1
Joined: Sun Jun 14, 2015 2:56 pm
languages_spoken: english
ODROIDs: xu3

Real time Kernel 3.10.80-rt102 working

Unread postby adbensi » Sat Jan 23, 2016 12:41 am

Hi guys!

Ethernet works on the 10MB HUB; At night I will try on the Gigabit SWITCH.
PING medusa.ufscar.br (200.9.84.70) 56(84) bytes of data.
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=1 ttl=55 time=37.4 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=2 ttl=55 time=54.9 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=3 ttl=55 time=36.0 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=4 ttl=55 time=36.2 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=5 ttl=55 time=36.3 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=6 ttl=55 time=36.2 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=7 ttl=55 time=36.2 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=8 ttl=55 time=36.5 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=9 ttl=55 time=36.2 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=10 ttl=55 time=36.3 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=11 ttl=55 time=37.8 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=12 ttl=55 time=102 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=13 ttl=55 time=86.0 ms
64 bytes from medusa.ufscar.br (200.9.84.70): icmp_seq=14 ttl=55 time=35.9 ms
^C
--- medusa.ufscar.br ping statistics ---
14 packets transmitted, 14 received, 0% packet loss, time 37923ms
rtt min/avg/max/mdev = 35.982/46.100/102.913/20.562 ms

Clx, when I upload the source, please, You can edit the first post to link it?

some tests... :)
-----------------

uname -a
Linux odroid 3.10.80-rt102 #2 SMP PREEMPT RT Wed Jan 20 12:44:26 BRST 2016 armv7l GNU/Linux

Kernel with Thermal and WatchDog Enabled - all drivers and modules
----------------------------
cyclictest --smp -p99 -m -n
# /dev/cpu_dma_latency set to 0us
policy: fifo: loadavg: 0.90 0.82 0.62 1/145 2050
T: 0 ( 2047) P:99 I:1000 C: 226586 Min: 8 Act: 18 Avg: 15 Max: 48
T: 1 ( 2048) P:99 I:1500 C: 151035 Min: 8 Act: 19 Avg: 15 Max: 46
T: 2 ( 2049) P:99 I:2000 C: 113268 Min: 7 Act: 19 Avg: 15 Max: 45
T: 3 ( 2050) P:99 I:2500 C: 90606 Min: 6 Act: 18 Avg: 13 Max: 36

cyclictest -l1000000 -m -n -a0 -t1 -p99 -i400 -q
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2252) P:99 I:400 C:1000000 Min: 7 Act: 7 Avg: 7 Max: 63

stress --cpu 4 --io 4 --vm 2 --hdd 1 --vm-bytes 128M --timeout 2500s
cyclictest -D1d -m -n -a0 -t1 -p99 -i1000
# /dev/cpu_dma_latency set to 0us
policy: fifo: loadavg: 13.98 13.19 8.94 12/204 2531
T: 0 ( 2286) P:99 I:1000 C:1179847 Min: 7 Act: 21 Avg: 18 Max: 74

Kernel without Thermal and WatchDog - all drivers and modules
-----------------------------
cyclictest --smp -p99 -m -n
# /dev/cpu_dma_latency set to 0us
policy: fifo: loadavg: 1.03 0.84 0.59 2/147 2930
T: 0 ( 2916) P:99 I:1000 C: 240008 Min: 7 Act: 13 Avg: 15 Max: 58
T: 1 ( 2917) P:99 I:1500 C: 159995 Min: 7 Act: 16 Avg: 15 Max: 51
T: 2 ( 2918) P:99 I:2000 C: 119984 Min: 6 Act: 19 Avg: 13 Max: 41
T: 3 ( 2919) P:99 I:2500 C: 95979 Min: 6 Act: 15 Avg: 12 Max: 35

cyclictest -l1000000 -m -n -a0 -t1 -p99 -i400 -q
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2945) P:99 I:400 C:1000000 Min: 7 Act: 8 Avg: 7 Max: 33

stress --cpu 4 --io 4 --vm 2 --hdd 1 --vm-bytes 128M --timeout 2500s
cyclictest -D1d -m -n -a0 -t1 -p99 -i1000
# /dev/cpu_dma_latency set to 0us
policy: fifo: loadavg: 13.49 9.08 4.27 9/159 3020
T: 0 ( 2998) P:99 I:1000 C: 329546 Min: 7 Act: 26 Avg: 17 Max: 53

Kernel without Thermal and WatchDog - light kernel /boot/uImage 4.53MB
-------------------------------------------------------------------
cyclictest --smp -p99 -m -n
# /dev/cpu_dma_latency set to 0us
policy: fifo: loadavg: 1.12 0.79 0.37 2/143 1151
T: 0 ( 1139) P:99 I:1000 C: 221280 Min: 6 Act: 14 Avg: 15 Max: 44
T: 1 ( 1140) P:99 I:1500 C: 147505 Min: 7 Act: 20 Avg: 15 Max: 59
T: 2 ( 1141) P:99 I:2000 C: 110617 Min: 6 Act: 13 Avg: 15 Max: 46
T: 3 ( 1142) P:99 I:2500 C: 88484 Min: 6 Act: 12 Avg: 13 Max: 39

cyclictest -l1000000 -m -n -a0 -t1 -p99 -i400 -q
# /dev/cpu_dma_latency set to 0us
T: 0 ( 1153) P:99 I:400 C:1000000 Min: 6 Act: 8 Avg: 7 Max: 32

stress --cpu 4 --io 4 --vm 2 --hdd 1 --vm-bytes 128M --timeout 2500s
root@odroid:~# cyclictest -D1d -m -n -a0 -t1 -p99 -i1000
# /dev/cpu_dma_latency set to 0us
policy: fifo: loadavg: 13.89 10.95 5.59 8/154 1312
T: 0 ( 1280) P:99 I:1000 C: 460545 Min: 7 Act: 11 Avg: 17 Max: 54

-----------------------------------------------

total used free shared buffers cached
Mem: 831M 180M 651M 1.6M 4.8M 28M
-/+ buffers/cache: 147M 684M
Swap: 99M 32M 67M

Filesystem Size Used Avail Use% Mounted on
/dev/mmcblk0p2 7.1G 3.1G 4.0G 44% / (source code 2GB and build packages)
udev 10M 0 10M 0% /dev
tmpfs 167M 5.4M 162M 4% /run
tmpfs 416M 0 416M 0% /dev/shm
tmpfs 5.0M 4.0K 5.0M 1% /run/lock
tmpfs 416M 0 416M 0% /sys/fs/cgroup
tmpfs 10M 1.1M 9.0M 11% /DietPi
tmpfs 20M 36K 20M 1% /var/log
tmpfs 416M 84K 416M 1% /tmp
/dev/mmcblk0p1 129M 39M 90M 31% /boot
tmpfs 84M 0 84M 0% /run/user/0

12:30:28 up 27 min, 1 user, load average: 0.98, 5.99, 4.70

Tasks: 128 total, 1 running, 127 sleeping, 0 stopped, 0 zombie
%Cpu(s): 1.2 us, 3.1 sy, 0.0 ni, 95.7 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem: 851792 total, 192348 used, 659444 free, 5044 buffers
KiB Swap: 102376 total, 32452 used, 69924 free. 30280 cached Mem

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
89 root -51 0 0 0 0 S 7.9 0.0 2:38.36 irq/63-dwc_otg
90 root -51 0 0 0 0 S 7.3 0.0 2:30.18 irq/63-dwc_otg_
844 root 19 -1 118620 7712 4520 S 5.6 0.9 5:34.87 Xorg
1342 root 20 0 4544 1188 792 R 0.7 0.1 0:00.06 top
20 root -2 0 0 0 0 S 0.3 0.0 0:02.87 ksoftirqd/1
34 root -2 0 0 0 0 S 0.3 0.0 0:02.04 ksoftirqd/3
1018 root 20 0 30568 8420 6884 S 0.3 1.0 0:45.91 lxterminal
1 root 20 0 4932 1704 1524 S 0.0 0.2 0:03.55 systemd
2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd
3 root -2 0 0 0 0 S 0.0 0.0 0:03.03 ksoftirqd/0
5 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0H
7 root rt 0 0 0 0 S 0.0 0.0 0:00.00 posixcputmr/0
8 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kclksetdelayd
9 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kcmosdelayd
10 root rt 0 0 0 0 S 0.0 0.0 0:00.00 migration/0
11 root -2 0 0 0 0 S 0.0 0.0 0:00.00 rcub/0
12 root 20 0 0 0 0 S 0.0 0.0 0:00.44 rcu_preempt
13 root 20 0 0 0 0 S 0.0 0.0 0:00.00 rcu_bh
14 root 20 0 0 0 0 S 0.0 0.0 0:00.00 rcu_sched
15 root -2 0 0 0 0 S 0.0 0.0 0:00.90 rcuc/0
16 root rt 0 0 0 0 S 0.0 0.0 0:00.00 watchdog/0
17 root rt 0 0 0 0 S 0.0 0.0 0:00.14 watchdog/1
18 root -2 0 0 0 0 S 0.0 0.0 0:00.98 rcuc/1
19 root rt 0 0 0 0 S 0.0 0.0 0:00.00 migration/1
21 root rt 0 0 0 0 S 0.0 0.0 0:00.00 posixcputmr/1
23 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 kworker/1:0H
24 root rt 0 0 0 0 S 0.0 0.0 0:00.18 watchdog/2
25 root -2 0 0 0 0 S 0.0 0.0 0:01.00 rcuc/2
26 root rt 0 0 0 0 S 0.0 0.0 0:00.00 migration/2
27 root -2 0 0 0 0 S 0.0 0.0 0:02.92 ksoftirqd/2
28 root rt 0 0 0 0 S 0.0 0.0 0:00.00 posixcputmr/2
30 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 kworker/2:0H
31 root rt 0 0 0 0 S 0.0 0.0 0:00.22 watchdog/3
32 root -2 0 0 0 0 S 0.0 0.0 0:00.71 rcuc/3
33 root rt 0 0 0 0 S 0.0 0.0 0:00.00 migration/3
35 root rt 0 0 0 0 S 0.0 0.0 0:00.00 posixcputmr/3
37 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 kworker/3:0H
38 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 khelper

Module Size Used by
nls_cp437 5125 1
fuse 72291 1
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby adbensi » Sat Jan 23, 2016 3:33 am

Settings for eth0:
Supported ports: [ TP MII ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supported pause frame use: Symmetric Receive-only
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised pause frame use: Symmetric Receive-only
Advertised auto-negotiation: Yes
Speed: 10Mb/s
Duplex: Half
Port: MII
PHYAD: 1
Transceiver: external
Auto-negotiation: on
Supports Wake-on: d
Wake-on: d
Link detected: yes

Resolving odroid.in (odroid.in)... 96.47.231.244
Connecting to odroid.in (odroid.in)|96.47.231.244|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 34452736 (33M) [application/x-tar]
Saving to: ‘/dev/null’
/dev/null 100%[=====================>] 32.86M 421KB/s in 82s
2016-01-22 16:27:36 (412 KB/s) - ‘/dev/null’ saved [34452736/34452736]

eth0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:25417 errors:0 dropped:0 overruns:0 frame:0
TX packets:13162 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:36483343 (34.7 MiB) TX bytes:973908 (951.0 KiB)
Interrupt:40
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time Kernel 3.10.80-rt102 working

Unread postby adbensi » Sat Jan 23, 2016 8:50 am

Source code:
https://github.com/adbensi/kernel-odroidc-3.10.80-rt102

Best Regards

adbensi wrote:Hi guys!

[...]
Clx, when I upload the source, please, You can edit the first post to include link?

[...]
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby michaelvv » Thu Jan 28, 2016 4:44 am

Hi adbensi.

I have compiled your Real Time Kernel, everything went fine.

I don't know how the right way to make an uInitrd, so I just copied
these files to my /boot directory.

config-3.10.80-rt-Diet
meson8b_odroidc.dtb
uImage

When I boot the image, I can only ping my Odroid C1+.

What am I missing ???

/Thanks Michael.
michaelvv
 
Posts: 9
Joined: Sun Jan 24, 2016 6:28 am
languages_spoken: english
ODROIDs: odroid c1+

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby adbensi » Thu Jan 28, 2016 5:53 am

Hi!
May be your device changed from eth0 to eth1.
Try sudo ifconfig eth0
and sudo ifconfig eth1
Post here please
Best regards
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby michaelvv » Fri Jan 29, 2016 7:24 am

Hi...

I'll try to put in on my screen to see what is happening under the boot process.

Which Distro have you tried your realtime kernel under ???

Best Michael.
michaelvv
 
Posts: 9
Joined: Sun Jan 24, 2016 6:28 am
languages_spoken: english
ODROIDs: odroid c1+

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby adbensi » Fri Jan 29, 2016 7:46 am

debian jessie.. openbox + LXDE.
You run make and module_install?

If odroid team allow, I would like to offer the image RT on the server com.odroid.com to download.
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby odroid » Sat Jan 30, 2016 11:00 am

adbensi wrote:debian jessie.. openbox + LXDE.
You run make and module_install?

If odroid team allow, I would like to offer the image RT on the server com.odroid.com to download.


Send a PM to mdrjr with a short description of your OS images.
And he will give you a direction how to use the file server.
User avatar
odroid
Site Admin
 
Posts: 28292
Joined: Fri Feb 22, 2013 11:14 pm
languages_spoken: English
ODROIDs: ODROID

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby michaelvv » Wed Feb 17, 2016 11:25 pm

Hi...

Can now boot your realtime kernel on my Arch linux on my odroid C1+.

I'm missing some modules for getting USB_SOUND to work. Had added

CONFIG_SND_USB=y
CONFIG_SND_USB_AUDIO=m

will see if it's enough, as a starting point.

/Thanks Michael
michaelvv
 
Posts: 9
Joined: Sun Jan 24, 2016 6:28 am
languages_spoken: english
ODROIDs: odroid c1+

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby adbensi » Thu Feb 18, 2016 7:26 am

Hi,
I didn't receice feedback from mdrjr, then if anyone like to get the image, send pm to me.
Best Regards
adbensi
 
Posts: 80
Joined: Fri Apr 18, 2014 10:29 am
Location: Brazil
languages_spoken: Português
ODROIDs: U3 and C1

Re: Real time kernel 3.10.70-rt74 (PREEMPT_RT), finally work

Unread postby odroid » Sat Mar 19, 2016 10:09 am

@adbensi,
Please send a PM to mdrjr and me again.
I will check it soon.
User avatar
odroid
Site Admin
 
Posts: 28292
Joined: Fri Feb 22, 2013 11:14 pm
languages_spoken: English
ODROIDs: ODROID

Previous

Return to Ubuntu

Who is online

Users browsing this forum: No registered users and 5 guests