By David Clack on Jan 22, 2010
Just did an image-update on a B128a install to B131.
suspend / resume is working again, compiz is fine also without /etc/X11/xorg.conf.
Looks like engineering has been busy.
So this may be my last blog entry as a Sun employee, looks like Oracle will get the go ahead from the EU next week.
If this blog does go away, I wanted to say what a pleasure it's been working with you all on the Intel Atom Networks with OpenSolaris.
Checking out B130, you can see that work has been done in the Xorg and shutdown menus areas. It looks like it all works apart from the resume failure.
The new login screen is pretty cool.
I've been working on the Samsung NC20 all week, creating a platform for the JavaFX developers in my group.
I've been working on Ubuntu as OpenSolaris does not like the 64 bit VIA Nano processor.
Running Ubuntu through the ugrades from 8.04, 8.10, 9.04 and 9.10, I've gotten everything to work.
My main goal was to allow for 3G network access, I'm using an AT&T 3G Quicksilver USB adapter, you can see from the attached pictures, that 3G networks are totally configurable and controllable via the network manager.
I can turn off the wired ethernet and wifi, and just connect into the AT&T 3G network, I'm on the edge down here at the beach, so the speed is a little slow, but it works.
Happy New Year
I just did a clean install of B130 on the AA1 10" and and image-update to B130 on the MSI U100.
Everything works fine accept for the video on resume and ssh will not let me login as root, I need to go through the regular user first.
I'm working with engineering to see if we can fix the video issue, the systems are still live, the video trashed. I'll keep you all posted.
BTW I'm testing the VIA Nano based Samsung NC20, it's a really cool little netbook, I'm working on Ubuntu 8.04 currently for another project, once that's done, I'll fire up OpenSolaris see how it works with the OpenChrome driver.
I do like the full size keyboard and 12.1" screen (1280 x800), it's really thin and not very heavy at all.
So I've been looking into this some more.
The issue is with the windows folder "My Documents", having a space in the directory name.
The terminal command chmod is not handling it the same as nautilus.
If I create a directory with Windows 7 "My_Documents" then it does come up with read / write permissions correctly at mount time.
So now I have to put the "_" character in all my Windows directories, change all the Windows programs to find "My_Documents" and My Pictures, and My Music and My Videos.
So I did an luugrade to Nevada B112 on the Acer Aspire One 10" last night.
As far as I can tell from a few hours of testing it's a perfect build.
All function key + are working volume control, brightness, external monitor switch.
Compiz is fine, suspend / resume is fine, I've even got virtual terminal working.
I've attached the instructions for turning on virtual terminals
Please follow the procedure to enable it if you want to use this long waited feature, and please read through before you take any action:
As root or privileged user:
# svcadm enable vtdaemon
# svcadm enable console-login:vt2
# svcadm enable console-login:vt3
# svcadm enable console-login:vt4
# svcadm enable console-login:vt5
# svcadm enable console-login:vt6
# svccfg -s vtdaemon setprop options/hotkeys=true
# svcadm refresh vtdaemon
# svcadm restart vtdaemon
I guess most people are unlikely to use the lock screen feature
(which will lock up your screen while you switch to other VTs),
you can disable it by:
# svccfg -s vtdaemon setprop options/secure=false
# svcadm refresh vtdaemon
# svcadm restart vtdaemon
If you are doing the above in X session, please log out and wait for Xorg
to restart. And then you will be able to switch between virtual console
terminals and X sessions via hotkeys (Alt + Ctrl + F#).
By default the first 6 virtual console terminals are reserved for console
login sessions, so Xorg will start on the 7th virtual console terminal.
There is an Xorg keyboard driver bug in build 122 which would occasionally hang Xorg during switch. It has been fixed in build 123. If you are eager to try it from build 122, please replace
with the attached ones.
You may also want to "pkill fbconsole" before you switch back to /dev/console
(Alt + Ctrl + F1) otherwise your console will lose input. It should be done upon
every reboot. It would be fixed when the new GDM comes in.
Why is it not enabled by default?
There are several console bugs like the fbconsole one that prevents the services
from being enabled by default. Our team will continue to fix these bugs until
it is prefect to be enabled by default. Now, if you want to use it, you need
to enable it manually.
So i've been testing all the nightlyies of what will become Neveda B121. On both the AA1 10" and MSI Wind.
desktop-print-management-applet was looping, using 100% of the processor, I renamed to desktop-print-management-applet.orig and kill the running program, cpu went back to normal.
The AA1 10" is working fine on Nevada B121.
The MSI Wind is working fine on Nevada B121, touchpad works after resume.
ACPICA is still locking up cpu 1 to 100% after resume, I'm going to check for different BIOS.
I have the AA1 8.9" on OpenSolaris 1002 B118, I'm going to look at local repositories and development for a customer I have in Turkey, they cannot afford to keep coming to our repository from Istanbul.
Also I need to build a headless distro for the BlueWonder Industrial Automation platform and a headless with Xorg distro for the JavaFX Jukebox / Digital Media projects.
So it looks like luupgrade is not working very well at all, ended up with B112, B115 and B116 on my AA1 10 inch, could not boot anything at all.
Good news is a clean install of B116 fixed the problem :-)
Everything works off the DVD but the Attansic ethernet controller, I downloaded the latest atge driver from http://homepage2.nifty.com/mrym3/taiyodo/eng/, looks like they solve the problem of the netbook hanging up if you did a big file transfer.
Compiz is working fine, the Atheros wifi is fine, sound buttons are working without having to adjust in gnome-keyboard-shortcuts anymore.
Suspend / resume is working fine after setting up the /etc/power.conf, I've posted this is an earlier blog entry.
The suspend works fine from the right click on gnome-power-manager in the gnome panel, closing the lid does get it into suspend mode, but it takes about two minutes before the system actually reacts.
I just downloaded and tested the Intel Moblin v2 and tested on the AA1 8 inch, it looks really good, I'll load up a spare hard drive into the AA11 10 inch this afternoon and do a full install, I also have it installed in VB 2.2.4 and it works fine also.
So I purchased a very small USB BlueTooth dongle of ebay for $2.99.
It came in a brown envelope from China, no instructions :-)
I added it to the Windows XP Home side of the AA1 10 inch, I'm able to activesync my HTC 8900 Tilt.
Then I went back to OpenSolaris (Nevada) B112 and followed the instructions here:
When you wget the installer and try to run it on a B112 system is gives you an error that it cannot find http://www.opensolaris.org/os/project/bluetooth/, what it's really looking for is a file bluetooth_mouse_(uname -v).tar.gz.
So I start playing with the script and get all the way to snv_111, it installs and asks you to reboot the system.
It updates /kernel/sockfs and /kernel/genunix, when the system restarts /kernel/fs/nfs gets unresolved dependencies, so you cannot NFS mount anything.
Anyway I followed the instructions for setting up the bluetooth mouse, and I actually got one to work.
So I went back into /kernel and copied back the backups of sockfs and genunix.
NFS is working again, BT is bust :-)
For $2.99 with free shipping it was an interesting way to add BlueTooth to a non BT netbook.
I'll keep trying to download a B112 version just for fun.
I just noticed the external sound port of the AA1 10" was not working with a clean install of SXCE B112.
I installed the boomer rc3 http://www.opensolaris.org/os/project/opensound/files/.
Everthing is working fine now.
So I was having issues with SDcards mounting on the MSI and AA1-10, so I opened a bug on it.
Guoqing Zhu from the Beijing development group was just logged into the MSI system upstairs, he found the problem.
I had formatted the sdcards on Windows XP Home.
Once I used his suggested mkfs command :
mkfs -F pcfs -o fat=32,spc=64 /dev/rdsk/c4t0d0p0:1
spc=64 being the correct parameter, all my sdcards started mounting and unmounting correctly.
So I thought I would try something differnet on the AA1 10 inch, Fuse is pretty good for read and writting ntfs but it does seem to kill the netbook sometimes.
With this new netbook I've sliced the disk this way:
40Gb - NTFS - Windows XP Home, still have to run iTunes
60Gb - Fat32 data
40Gb - ZFS - OpenSolaris Nevada (SXCE) B110
I'm trying to setup everything so the data for my user is on the Fat32 partition even if it is Windows or OpenSolaris I'm booted into.
I can mount the Fat32 partition in OpenSolaris and I've link my user home directory to it.
So I'm building compiz 0.8.0, see if I can get the MSI to run it.
Found another bug in gnome-system-monitor, on my big Asus C90S powertop says:
Cn Avg residency P-states (frequencies)
C0 (cpu running) (8.4%) 1596 Mhz 100.0%
C1 3.9ms (91.6%) 2394 Mhz 0.0%
Wakeups-from-idle per second: 237.7 interval: 5.5s
Power usage (ACPI estimate): 0.000W
Top causes for wakeups:
67.6% (160.7) <kernel> : genunix`cv_wakeup
42.1% ( 99.9) <kernel> : genunix`clock
21.4% ( 50.8) sched : <cross calls>
6.5% ( 15.4) <kernel> : uhci`uhci_handle_root_hub_status_change
4.2% ( 9.9) <kernel> : ata`ghd_timeout
2.4% ( 5.7) <interrupt> : ehci#0
2.2% ( 5.1) <interrupt> : ata#0
1.9% ( 4.4) <kernel> : uhci`uhci_cmd_timeout_hdlr
1.7% ( 4.0) <kernel> : genunix`schedpaging
1.6% ( 3.9) <kernel> : ehci`ehci_handle_root_hub_status_change
But gnome-system-monitor only shows one processor ?
So much to test, so little time :-)
So development engineering and I have decided to standardize testing on SXCE, this is to give us a clear view into the OS, without any of the major changes that 0906 makes to the standard SXCE.
I have a new category now the new Acer Aspire One with the 10 inch screen. They made a few changes to the hardware, first it has an Atheros ethernet vs the Realtek ethernet the 8.9" has, second it has a single USB attached SD/MSpro card reader vs the two card slots on the 8.9" on the JMicron controller.
I left my original AA1 down in LA, so we are now focusing on making the AA1 8.9", AA1 10" and MSI 10" work with B110.
Most of yesterday we focused on the new power management built into B110, we needed to change a couple of entries in the /etc/power.conf and add a line into /etc/system.
# Auto-Shutdown Idle(min) Start/Finish(hh:mm) Behavior
autoshutdown 30 9:00 9:00 noshutdown
set cpupm_ti_gov_multiple = 5
I also found out that my old standby command kstat -m cpu_info -s current_clock_Hz only show you a single view of what the cpu is doing, powertop using a 5 second grap of the counters is the best way of finding out what the avarage cpu speed is.
We are now working on the sdcard issues on the AA1, also the USB sdcard issues on the AA1-10 and MSI.
You can see an sdcard or MSPro card inserted the first time, after you unmount, the card reader cannot be used again until the next reboot.
The MSI is now suspending and resuming and the cpu throttling is working after resume, the touchpad is still dead though.
I'm having major compiz problems on all the netbooks.
You can sometimes actually manually startup compiz, but the minute you logoff and try to login into the netbook again you just get a background and cursor, it never gets to the desktop.
I've also got the IGO1000 running 0906-B109 from the 8Gb CF, with 1Gb memory and Intel B/G Mini PCI-E wifi.
I'll replace with SXCE B110 sometime today.
I work for the Java Embedded Global Business Unit of Oracle within Java Industry Solutions group under Alliances and Channels . I'm focused on the Java SE Embedded and BDB for ARM and PPC across multiple verticals