Ubuntu 16.04 for UDOO Neo

I have reproduced the issue. I will look at it asap.

 

Wireless modules were missing, sorry!
If you have an USB ethernet adapter / Neo Full owners, please run:

Code:

sudo apt update && sudo apt dist-upgrade

Otherwise copy this file to the sdcard and run:

Code:

sudo dpkg -i /path/to/linux-udooneo_4.1.15-gfcea7a1-dirty_armhf.deb

 

Installing the package didn't change anything for me. Any more steps to do?

 

Comparing lsmod output and /lib/modules of both images, it appears that the kernel modules required to use the WLAN functionality of the TI WL1831 aren't included in the package. (Not sure what's required for the BT part, or if it works already with the new image, didn't get that far yet.)

Using the official 2.1 image, i get:

Code:

udooer@udooneo:/lib/modules$ find | grep wl
./3.14.56-udooneo-02023-g8d371df/kernel/backports/drivers/net/wireless/rndis_wlan.ko
./3.14.56-udooneo-02023-g8d371df/kernel/backports/drivers/net/wireless/ti/wlcore
./3.14.56-udooneo-02023-g8d371df/kernel/backports/drivers/net/wireless/ti/wlcore/wlcore.ko
./3.14.56-udooneo-02023-g8d371df/kernel/backports/drivers/net/wireless/ti/wlcore/wlcore_sdio.ko
./3.14.56-udooneo-02023-g8d371df/kernel/backports/drivers/net/wireless/ti/wl18xx
./3.14.56-udooneo-02023-g8d371df/kernel/backports/drivers/net/wireless/ti/wl18xx/wl18xx.ko

Package contents:

Code:

udooer@udooneo:/otgstorage$ dpkg-deb -c ./linux-udooneo_4.1.15-gfcea7a1-dirty_armhf.deb | grep wire
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rt2x00/
-rw-rw-r-- root/root     18684 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rt2x00/rt2x00usb.ko
-rw-rw-r-- root/root     68024 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rt2x00/rt2800lib.ko
-rw-rw-r-- root/root     56156 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rt2x00/rt2x00lib.ko
-rw-rw-r-- root/root     45824 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rt2x00/rt2800usb.ko
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/brcm80211/
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/brcm80211/brcmutil/
-rw-rw-r-- root/root     12212 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/brcm80211/brcmutil/brcmutil.ko
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/brcm80211/brcmfmac/
-rw-rw-r-- root/root    205996 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/brcm80211/brcmfmac/brcmfmac.ko
-rw-rw-r-- root/root     38144 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rndis_wlan.ko
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rtl818x/
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rtl818x/rtl8187/
-rw-rw-r-- root/root     60544 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/rtl818x/rtl8187/rtl8187.ko
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/ath/
drwxrwxr-x root/root         0 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/ath/ath6kl/
-rw-rw-r-- root/root     22540 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/ath/ath6kl/ath6kl_sdio.ko
-rw-rw-r-- root/root    156300 2016-09-27 09:38 ./lib/modules/4.1.15-gfcea7a1-dirty/kernel/drivers/net/wireless/ath/ath6kl/ath6kl_core.ko

 

@Francesco Browsing Github for a bit, it appears that the package you kindly provided was built from the commit preceding the one enabling the WL18xx stuff.

Could you post a link to the latest one, something with 9f042396? Thanks!

 

Updated to Beta1. Please have a look to the new virtual serial port!

 

@Francesco I tried the ubuntu 16.04 beta 1 image in udoo neo. First of all the red LED is blinking as normal booting procedure.However I cannot find any network connection between the udoo and my laptop( I am using ubuntu xenial). Also, using dmesg shows that the udoo usb (cp210x) is not being connected to any /dev/ttyxxx port. I am sure this is not because of insufficient power.Also the udoo doesn't appear as a mass storage storage device.

Kindly look into this.Thanks

 

OTG USB connection is not tested yet. I use a Serial-USB adapter (this or this) to get a console. Wifi, Ethernet and HDMI are working for me too.

 

@Francesco I tested the ubuntu 16.04 beta on neo version by connecting it through the network uding ssh and ip address.I tried some commands like uname,lsb_release etc. But when i used 'ls' command, the neo just hung up. At first i thought of network errors, but after that the neo is not getting connected via ssh.However, it is showing normal boot procedure ( as indicated by the LEDS). But on connecting it to the ethernet, the udoo neo IP is not showing.( I am using nmap to trace the ip connected in the local network).

Any ideas what happend ?

 

@Francesco Solved the issue.Ubuntu 16.04 working fine in the udoo neo.

Can you mention what changes are required to make the USB OTG connection work ?

 

Don't know I have to look at it.
How did you solve your issue?

 

currently using ifconfig in the ubuntu 16.04 beta version gives "bash: /sbin/ifconfig: Input/output error" .

 

Posted Beta 2. Changed since Beta 1:

• Added OTG USB / g_multi
• Added VNC server
• Rebuild Arduino IDE for Xenial
• Fixed LVDS 7''
• Fixed connection to Wi-Fi networks protected with passwords
• Fixed 9-Axis sysfs permissions

Please let me know if you still get io errors.

 

@Francesco Which linux kernel are you using ? I tried to fix those issues using the freescale linux (https://github.com/Freescale/linux-fslc) but could not build the .dtb files.

 

https://github.com/fmntf/linux_kernel/commits/4.1.15_2.x-udoo

 

Thanks @Francesco for the link. Incidentaly the kernel i was using gives error like :

In function `imx_mcc_tty_remove':
vf610_adc.c text+0x74018): undefined reference to `mcc_destroy_endpoint'
drivers/built-in.o: In function `mcc_read_thread':
vf610_adc.c text+0x740c4): undefined reference to `mcc_recv'

during the make zImage process. Any idead what might be causing this ?

 

MCC has been removed since 3.14.52; there is RPMSG now.

 

Hi everybody,

I'm struggeling with a M4 hard fault on my Udoo Neo boards after upgrading to 16.04. Since the problem is a little more complicated I will give you some background first:

I ported a RTOS which we developed at our department to run on MQX. The MQX Version is 4.1 with the initial patch commit "Updated MQX sources" on https://github.com/UDOOboard/udooneo_mqx41_patch.
If I'm running the RC2 Udoobuntu release on the A9 and the ported RTOS on the M4 everything is working.
My problems started when I tried to get 16.04 running. While the A9 was working fine, the M4 caused a hard fault calling __libc_init_array() in toolchain_startup() in comp.c.

It took me a while to figure out that there is another patch for the Neo MQX that is supposed to take care of the hard fault problem. The patch solution was to move __libc_init_array() from comp.c to init_bsp.c.

If I apply the second patch the M4 does not get stuck in a hard fault. However, my constructors (which are called in __libc_init_array()) are called after the MQX TASK_TEMPLATE_STRUCT is created. In order to get the ported RTOS to work properly, it is necessary to call the constructors before the TASK_TEMPLATE_STRUCT is created.

Despite all that, I'm confused that malloc is called from __libc_init_array(). According to http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.faqs/ka3951.html a way to avoid calling malloc is to retarget __cxa_atexit(), which is already done in MQX comp.c ...

I also do not understand how the A9 update changed the behavior of the M4.

 

I am trying to load the driver for ttyRPMSG but it is not sucesfully loaded.

The ssh terminal does not return any massage and I check /dev/ for ttyRPMSG and it is not there.

FYI:

capture before load drivers with "sudo modprobe imx_rpmsg_tty"


capture loading drivers without any message for feedback


capture after load drivers, there is not ttyRPMSG

 

Someone have the link for beta 1 version?