WBQuad troubleshooting: no video/HDMI output

[Matthew W. Novak]

Hi folks.  I need some help troubleshooting a WBQUAD issue: I'm not getting any video output from the HDMI out, and I'm not sure how to proceed.  Note that I am not a developer, but I am a competent Linux user, so while a lot of things here will be over my head, I should be able to gather troubleshooting and debugging doc without much trouble.

So far, I have . . . 

1. downloaded the WBQuad Android 4.2.2 image
   
   
• I've also walked through the following steps using the wandboard-quad-ubuntu-11.10-20130604.img image with the same results
  
  
2. DD'd that image to my mSD card per the Quick Start Guide: dd if=./android-4.2.2-wand-quad-preview-20130607.img of=/dev/[my_SD_card] bs=1M
   
   
• I've DD'd the image both from Linux and OSX machines with the same results
  
  
3.  inserted that mSD card into the OS mSD card slot on the processor board, plugged in an HDMI-to-DVI cable that's attached to an HDCP-enabled monitor, and plugged in my 5V 2A 10W power supply
   
   
• I've tried this with multiple monitors and multiple HDMI-to-DVI cables with the same results
• I've tried booting with a USB keyboard attached, and/or with an OTG cable attached, with the same results
  
  
4. in every iteration above, a green light on the board comes on and nothing happens at the monitor
   
   
• to reiterate, this is the same result whether booting the Android or Ubuntu images from the Wandboard downloads page

However, at least with Android, the image appears to be booting: with an OTG cable plugged between the board and my laptop, I can do the following:

• "adb devices" and see the board listed
• "adb shell" and get into a functional Android command prompt
• "logcat" from within the adb shell and watch as the boot progresses and generates log entries

If I leave logcat running indefinitely, I eventually get kicked out of the shell and the board disappears from "adb devices".  

I've also noticed that I can only access the device via adb on the first boot after DD'ing an image to the mSD card: the board does not appear in "adb devices" after any subsequent boot of the same mSD card (reset button, reboot from within the adb shell, hard power off/power on, etc.).  

So, I appear to have a functional board with a functional Android image on the mSD card, but I'm not getting any video output via HDMI, and the instability between first and subsequent boots isn't encouraging.  Any advice on where to start?

Note that I don't have a null modem cable available, but will be able to pick one up over the weekend to help with debugging.

--Matt  

[Matthew W. Novak]

Board hooked up to laptop via OTG cable, no SD card inserted, powered on: issuing "system_profiler SPUSBDataType" (akin to lsusb on OSX) shows:

        SE Blank ARIK:

          Product ID: 0x0054

          Vendor ID: 0x15a2  (Freescale Semiconductor, Inc.)

          Version: 0.01

          Speed: Up to 12 Mb/sec

          Manufacturer: Freescale SemiConductor Inc

          Location ID: 0x04100000 / 4

          Current Available (mA): 500

          Current Required (mA): 10

So again: board appears to be alive, just no output to HDMI.

--Matt

[Matthew W. Novak]

Turns out I can successfully SSH into the 20130604 Ubuntu 11.10 image:

[mwnovak@silverBook wandboard-quad-ubuntu-11.10-20130604]$ ssh ro...@192.168.0.11
ro...@192.168.0.11's password:

Welcome to Linaro 11.10 (development branch) (GNU/Linux 3.0.35+ armv7l)

 * Documentation:  https://wiki.linaro.org/

New release '12.04 LTS' available.

Run 'do-release-upgrade' to upgrade to it.

Last login: Thu Jan  1 00:00:55 1970

For anyone else who comes looking, default [login / password] here is [root / linaro].

More evidence of a happy board; just no video.

--Matt

[ArchAngel Azrael]

Hi, any chance to you test using only a HDMI cable and not a HDMI-DVI cable ?

Leonardo

[Matthew W. Novak]

Hi Leonardo.  I can test with straight HDMI, but I'll need to borrow a display.  Bit of a hassle, but I'll see if I can make that test and report back.

Having said that, there really shouldn't be an issue between HDMI and DVI-D, particularly going into a HDCP-compliant monitor.  My understanding is that there's absolutely no manipulation of the digital video signal taking place in this arrangement, though I may be incorrect.

--Matt 

[ArchAngel Azrael]

Another thing to check, are you sure the driver is setting the correct video resolution ?

By the way, when I receive my board ( +/- 60 days, because I live in Brazil) I will use the same cable you are using to connect to my old Samsung 2232BW monitor.

I am very interesting in to know the possible solution for your problem.

Excuse me my bad english.

Leonardo

[Matthew W. Novak]

This is with a Samsung 2433BW, a perfectly decent monitor if we can get it working. :)

Still under the Ubuntu image, xrandr gives the following:

root@wandboard:~# xrandr

Can't open display

I've started Google'ing on that, but most of what I'm getting back either seems unrelated or is starting to go over my head.  For what it's worth, this is from a freshly imaged and booted SD card and the monitor has been connected and powered-on the entire time.  

Any suggestions on how to proceed?

--Matt

 

[Matthew W. Novak]

Just tested using an HDMI-to-HDMI cable between the QUAD and a 32-inch Samsung television: still no video output from either the current Android (4.2.2-wand-quad-preview-20130607) or Ubuntu (wandboard-quad-ubuntu-11.10-20130604) images.  Both images were freshly DD'd to the SD card before testing.  No joy.  --Matt

 

On Friday, June 14, 2013 4:09:42 PM UTC-4, ArchAngel Azrael wrote:

[Matthew W. Novak]

Two questions:

1) is anyone with a WB QUAD getting video output?  If so, could I trouble you for details of your setup?

2) is it possible that the board needs more than a 2-amp power supply to push a video signal?

Somewhat grasping at straws,

--Matt

[Jay Custin]

Same here... WB QUAD using HDMI cable to both my monitor and a TV's HDMI and I get zilch.  Unfortunately I do NOT have the luxury of another cable solution - sigh.

On a side bar, I also never see more than one half of the RJ45 Ethernet LEDs to light.  I tried both the Ubuntu and Android images (via dd).

Jay

[Mark Schiech Jr]

I own the WB QUAD using an HDMI to DVI-D cable for my monitor with video. I used the Yocto image with no changes needed. My monitor model is Dell SP2008WFP using 1680x1050 resolution by default.

[Matthew W. Novak]

Hi Jay.  Sorry to hear that you're in the same boat, but thanks for posting.

Question: what power supply are you using?  What monitor?  Just trying to chase down variables.

Regarding ethernet, I've not paid much attention but it appears to be working on my board: SSH'ing into the Ubuntu image, I'd get green and orange lights while interacting with the CLI, as expected.

--Matt

On Sunday, June 16, 2013 5:48:35 PM UTC-4, Jay Custin wrote:

[Matthew W. Novak]

Hi Mark.  Thanks for posting: I'm glad to hear that you're up and running with your WBQ.  

You're using the yocto-1.4_wandboard_quad_20130606 image currently posted to the downloads page of the wandboard.org site?  I've tried that image as well, but still no video out for me.

What are you using for a power supply?

--Matt

[Mark Schiech Jr]

Yes, I'm using the image from the downloads page. The power adapter I use is a TomTom model A10P1-05MP 5V 2 Amp. The barrel plug is smaller so I used a converter plug.

[Matthew W. Novak]

Alright Mark, thank you for the info.  Not sure how to explain the discrepancy: I've now tried the Yocto image against two different monitors and two different HDMI-to-DVI-D cables, and no video on either.  This with a 5V2A adapter.

Question to the list: how can I verify that my hardware is functioning as intended? specifically the video and HDMI components?  If the hardware I received is functional, I'm content to work toward (or wait for) a software solution . . . but that becomes an exercise in frustration if my hardware is faulty from the start.

Thanks for any  help,

--Matt  

[Mark Schiech Jr]

I would assume it's a faulty board and request an exchange. I don't see how several boards in the same batch could have different components requiring different device drivers.

[ArchAngel Azrael]

Hi, are there any log ? The last lines of dmesg, some xorg (?) output (and log) ?

To enable the xorg log :
http://forum.stmlabs.com/showthread.php?tid=795&pid=7125#pid7125
Try adding
Option "ModeDebug" "true"/var/log/X

To /etc/X11/xorg.conf
Under Section "Device"

Do you tried to plug the hdmi cable * after *  in the the board  after the boot ?

Do you tried some solutions and tips like:

http://askubuntu.com/questions/176300/xrandr-as-root-cannot-find-displays

http://www.solid-run.com/phpbb/viewtopic.php?f=6&t=591

http://forum.xbmc.org/showthread.php?tid=57030

I have read some comment about the error "xrandr Can't open display":  If anyone gets "Can't open display" while running xrandr commands from this (great!) tutorial, you might be SSH'ing in. You need to run those on the actual X environment. Source: http://therd0.blogspot.com.br/2013/01/mk802-lubuntu-change-resolution.html

I really hope you find the solution, because if I get this problem, I will unable to replace the board here in Brazil.

Leonardo

[Matthew W. Novak]

Mark: agreed.  I'll contact Future about an exchange.

Leonardo: hopefully this is an isolated issue and you won't have any problems.  

For what it's worth, I did try plugging in the HDMI connection after powering on the board and waiting for the image to boot: still no video under either the Android or Ubuntu images.  

The xrandr "Can't open display" message was indeed due to being within an SSH session--I should have realized that in the first place.  

However, with Mark's confirmation of HDMI output from the same board, power supply, display (roughly), and software image, I'm going to conclude that the board is faulty and seek a replacement.  I'll update this thread with the eventual outcome.

--Matt  

[www.wandboard.org]

Can someone post the logfile from their serial during bootcycle so we can check the logfile.

Please list at same time exactly which file you downloaded from the wandboard download page.

Also to quickly confirm if a board is in working condition. You can read the startup guide. (top post on the forums). There is a way to connect the board over USB OTG and if the board pops up as a device you know that the board is operational and in working order.

[Matthew W. Novak]

Can someone post the logfile from their serial during bootcycle so we can check the logfile.

I do not have a serial cable available at the moment, but I can attempt to post log information later.

 

Please list at same time exactly which file you downloaded from the wandboard download page.

I have used the following images from the wandboard.org downloads page:

• android-4.2.2-wand-quad-preview-20130607.img
• wandboard-quad-ubuntu-11.10-20130604.img
• Yocto-1.4_wandboard_quad_20130606.img

Also to quickly confirm if a board is in working condition. You can read the startup guide. (top post on the forums). There is a way to connect the board over USB OTG and if the board pops up as a device you know that the board is operational and in working order.

On a linux machine (no SD card):

mwnovak@cube$ lsusb | grep Freescale
mwnovak@cube$

On an OSX machine, immediately after powering-on the board (no SD card):

mwnovak@sbook$ system_profiler SPUSBDataType | grep Freescale

Vendor ID: 0x15a2 (Freescale Semiconductor, Inc.)

Manufacturer: Freescale Semiconductor, Inc.
mwnovak@sbook$

On an OSX machine, after leaving the board powered on for several minutes (no SD card):

mwnovak@sbook$ system_profiler SPUSBDataType | grep Freescale
mwnovak@sbook$

Thank you for any help.

--Matt

[Mark Schiech Jr]

Matt or anyone else needing a serial converter, do you have any electronic components that could be used to make your own serial converter?

1) RS232 signal converter IC, several different types of these are available and with different part numbers

2) Female ended jumper wire, for connecting to the serial port on the WB

3) Either a USB to TTL serial adapter, or, a microcontroller with TTL serial and USB, or, a computer with another male ended serial port, if so, just use the female ended wire to connect the two, no circuit needed, google "db9 pinout" you should only need five pins, TX, RX, RTS, CTS, GND. Keep the wires  as short as possible to avoid any interference and loss of data on the line.

I hope this helps in troubleshooting your board as I don't like having to go through all the hassles of exchanging a defective item.

[Matthew W. Novak]

Hi Mark.  Thanks for your time here.  I have a serial/null modem cable; I just didn't have it at hand earlier today.  

Two files attached are the from boot attempts of the current Ubuntu and Android images (clearly labeled); both logs were captured from boots of freshly DD'd images.  Ubuntu appears to fail while attempting to mount the root file system; Android doesn't appear to record any particular failure.  

I don't have another micro-SD card handy, but can pick one up if there's reason to suspect the card.

Also note that I've got the board plugged in via serial with minicom capturing activity from power-on until output stops and remains stopped for a few minutes: please let me know if I should be doing something more/else.

Regarding the hassle of exchanging boards: are you involved with Future Electronics or Wandboard?  Would hate for this to be more of a hassle requesting an RMA and boxing/shipping back the board, if necessary.  

Thanks again,

--Matt

[Mark Schiech Jr]

No, sorry I don't work with Future Elec or WandBoard, I'm just another owner of the board with some free time. I'm glad to hear you were able to get some more information regarding the errors. As for the SD card, I have several. Kingston 8 GB, 16 GB, 32 GB all Class 4. SanDisk 8 GB Class 4. And a few 2 GB no names unknown class. Because I only have slower cards I use the Yocto image. What's the information on the SD card you have? 

[Mark Schiech Jr]

I just noticed in both logs the processor speed is 792 MHz. Wasn't the processor supposed to be 1 GHz? According to the picture on WandBoard's site, this processor is the consumer version, which is 1.2 GHz.

[Ionut Ana]

You most likely have bad board or your monitors do not support the resolution, I've just plugged in my quad wandboard and it has a resolution of 1280 x 720 with no way to change it from the settings. I also believe it has a short time out setting by default to turn of screen, changed that to 30 min.

[Matthew W. Novak]

Mark: I may have misread your note about returns, but I definitely appreciate your help here.  SD card is a new PNY 16gb class 10.  Googling some of the errors from my ubuntu boot log, I came across several references to corrupted filesystems and failing drives, so I'm running badblocks against the card right now and will also try to fsck the partitions after writing an image to it.  May or may not learn anything new.

Not sure what to make of the reported processor speed or whether it's related to my woes.  What does your board report on boot?


Ionut: thanks for chiming in, I appreciate it.  I've tried the board without success against two Samsung monitors (native at 1920x1200 and 1280x1024) and a 32-inch Samsung 720p television (native 1,366x768, with support for 1280x720).  Haven't gotten so much as a flicker. 

If you're running the Ubunutu image, you may be able to get other resolutions by adding modelines to your xorg.conf.  There are probably better/newer/slicker ways to do that, though.

--Matt

[www.wandboard.org]

http://www.wandboard.org/index.php/forums?place=topic%2Fwandboard%2FGfAtmxiBpmY%2Fdiscussion

[Matthew W. Novak]

On Wednesday, June 19, 2013 9:20:57 AM UTC-4, www.wandboard.org wrote:

http://www.wandboard.org/index.php/forums?place=topic%2Fwandboard%2FGfAtmxiBpmY%2Fdiscussion

Hello.  Thanks for your response.

I have followed both the Quick Start and User Guide available from the downloads page, as well as the steps outlined in the "Wandboard startup guide" forum post that you've linked above.  It's possible that I missed something or made a mistake, but to the best of my knowledge I have exhausted the published documentation and support for this board.

Do you have a specific suggestion, question, or request for additional information from me?  Please let me know, and I'll be happy to oblige

Thanks again for any help,

--Matt   

[jth...@gmail.com]

Hi Matt,

I am experiencing the same thing with a brand new board fresk out of the box this morning.  I am using a 16GB SanDisk Ultra Class 10 uSD.

The boot appears successful but no video.  Our boot logs are different though.  Looking through yours I see backtraces and evidence of CPU crashes.  Not sure what could be causing that.  I'm trying to boot the Ubuntu image and what I see is the HDMI is detected but further down at timestamps 3.28/9 are lines "mxc_hdmi: No modes read from edid" and "create default modelist".  The X server is running but nothing on the display.

I have a couple SABRE Lite iMX6Q boards which work just fine with this Dell monitor.  I modified bootargs trying to pick a known working mode (1280x720M@60) but that had no affect.

Evaluating the WBQUAD now as it provides SPI access (which the SABRE Lite lacks).  Definitely need video out for my app though.

My log is attached.

All the best,   Joe

[jth...@gmail.com]

A couple other odd behaviors I should have mentioned -
 - what's with the "error: unexpectedly disconnected from boot status daemon" output?
 - I am using ttermpro as the terminal connected to the serial out.  Immediately after the "* Starting bluetooth" output line, character color changes within tterm to bright yellow on white.  This is consistent and would seem to indicate an odd character sequence being emitted.  A clue perhaps?

[jth...@gmail.com]

Burned Yocto image and tried booting that.  Fast boot.  No video.
Wired ethernet/TCP/IP work fine.  No wireless (wifi, BT) as expected.
PS shows X server and window manager running yet nothing on HDMI monitor.

Wandboard folks - how to troubleshoot HDMI out before exchanging/returning board?

[jth...@gmail.com]

A bit more info.  I am using a Dell 1909W monitor.  Tried two different "high quality" HDMI<->DVI cables without success.  In both cases the "No modes read from edid." message appears.  What's up with that?

Switched to a "banner display" which is native 1280x400 but HDMI input only.  There I do see graphics - though four small desktops rendered simultaneously, each in its own quadrant.  The good news is signal is emitted.  The bad news is this isn't what I expect to see.  The edid read failure messages also do not appear.

These HDMI<->DVI cables work fine with both SABRE Lite and Raspberry Pi SBCs.

[Ionut Ana]

I have not tried HDMI<->DVI, can you try HDMI to HDMI, there are other boards (cubieboard) that had problems using HDMI to DVI.

[www.wandboard.org]

Hi all,

The wandboard is a development platform. The images we post in the download page are provided for convenience. A custom version for selected cases can always be compiled/modified by anyone when downloading the sourcecode.

If a display resolution in a demo image is too high for the monitor native resolution you are using. The screen will remain black. With DVI monitors we cannot detect the resolution over the HDMI interface and unless there is a match you will have to make modifications to the runtime demo image or go the sourcecode route.

The below is an example how to change resolution with the Ubuntu image that is currently on the wandboard homepage.

to change resolution to 480p in Ubuntu on the wandboard quad:

1. Connect serial console
2. When u-boot starts, press any key
3. On the u-boot prompt issue the following command => setenv bootargs 'console=ttymxc0,115200 video=mxcfb0:dev=hdmi,720x480@60,if=RGB32 root=/dev/mmcblk0p1 rootfstype=ext4 rootwait rw'
=> save
=> reset
4. The board reboots and Ubuntu comes up with 480p resolution (which all HDMI compliant devices should support)

[Matthew W. Novak]

Joe: sorry to hear about your troubles, and I wish I could offer some help.  It's encouraging that you got video out over HDMI to that banner display, however.

Ionut: I've tried with both HDMI-DVI and straight HDMI cables with no change.

wandboard.org: I had the same problem over straight HDMI (no DVI) to a 720P television (which natively runs the 1280x720 resolution specified by default in the Ubuntu and Android images).  This was true even when I could still boot to a shell (see below).

Unfortunately, I appear to be chasing multiple issues at this point: when I first posted to this thread, I could boot to a functional shell but lacked video output (like others on the thread); now I can't even boot to a shell (kernel panics; logs attached in previous post).  Will attempt to address the boot issue before resuming with the HDMI output.

--Matt

[jth...@gmail.com]

Since I had been exercising the WBQUAD here in the office and all of the monitors available are DVI, I took the SBC home last night.  I plugged it into a 22" TV/monitor and, lo and behold, on the second try (initial boot the display remained dark and I hit the reset button) the desktop appears!

So, what is different with the Wandboard HDMI circuitry when compared to the SABRE Lite?  One thing I did notice comparing the schematics is the handling of pin 13, labeled CEC-A, as diagrammed on page 23 of the "Wandboard User Guide".  The SABRE Lite schematic shows this pin as not connected.  Could that be it?

Matt - I'd suggest you reburn your uSD card.  I am using the Ubuntu 11.10 quad-core image at this time.  After writing the image to uSD I noticed it looked, well, odd to me.  There is an 8MB unallocated area at the start followed by a 2.5GB (give or take) primary partition (/dev/sdd1) which is followed by a smaller linux-swap partition (/dev/sdd2).  This last piece my workstation flagged as errored by the way.  Using gparted I deleted /dev/sdd2, grew /dev/sdd1 to 12.5GB for room to work then recreated the /dev/sdd2 linux-swap partition using the remainder of the uSD.  This boots and seems stable.

wandboard.org - the displays I am using easily support 1280x720 and I played with the bootargs values yesterday without seeing any change in behavior.  I'll admit I did not try dropping all the way down to 480p resolutions.  Have you been successful with HDMI<->DVI presentations?  If so could you let us know what has worked for you?

All the best,   Joe

[jth...@gmail.com]

Apologies...  Just noticed I cited the wrong page for the wandboard HDMI connector schematic.  That is on page 26 (last page of the doc I downloaded) NOT 23.

[Matthew W. Novak]

Joe: I'd re-written the SD card several times throughout the troubleshooting proces.  Turns out the SD card itself went south: with a new SD card, I'm once again able to boot images (boot log from the latest Android image attached).

Still no video out to any display, however, even after tweaking the bootargs as suggested above.

Found this and am wondering if it's the culprit: http://boundarydevices.com/hdmi-cea-mode/

Android patch: https://github.com/boundarydevices/linux-imx6/commit/b37d90a14e06f4774782ff8a75189e51c9b96c0d

Linux patch: https://github.com/boundarydevices/linux-imx6/commit/50aa49f52436bde67a81cf137618fa00b367c523

The offending code exists in mxc_hdmi.c for both the current Wandboard Android and Ubuntu sources.  I've applied the Boundary Devices patch, but can't successfully build: anyone able to test this while I get my build environment sorted out?

Pretty far out of my depth here, so give a shout if this is an obvious no-go.

--Matt  

[Matthew W. Novak]

Forgot to attach the boot log from the current Android image: it appears clean, but I would be grateful for a more experienced set of eyes.  --Matt

[www.wandboard.org]

Repost because no response on suggestions.

[Matthew W. Novak]

Repost because no response on suggestions.

We got output at 640x480 using the following video statement in bootargs:

video=mxcfb0:dev=hdmi,1920x1080P@60,if=RGB24

So far no luck getting better than 640x480 via xrandr magic.  

Any thoughts on the following? 

http://boundarydevices.com/hdmi-cea-mode/

--Matt

[jth...@gmail.com]

Hi Matt,

uSD went bad?  I guess "when it rains..."

Looked through your boot log and, yes, looks clean to me too.  Did notice the kernel bootline still specifies 1280x720@60.  To get the 480p output do you interrupt the boot and set bootargs?  I left my board in the office for the weekend.  Nice find on the Boundary Devices link.  I have the source respository in place and synced.  Will get it set up for native compilation when I'm back in the office (with the mxc_hdmi.c CEA mods in place).

Enjoy your weekend!   Joe

[Matthew W. Novak]

Looked through your boot log and, yes, looks clean to me too.  Did notice the kernel bootline still specifies 1280x720@60.  To get the 480p output do you interrupt the boot and set bootargs?  I left my board in the office for the weekend.  Nice find on the Boundary Devices link.  I have the source respository in place and synced.  Will get it set up for native compilation when I'm back in the office (with the mxc_hdmi.c CEA mods in place).

 

Hi Joe. 

I captured that (clean) boot log before I had any video output.  The following video statement in the bootargs is what finally garnered output:

video=mxcfb0:dev=hdmi,1920x1080P@60,if=RGB24

640x480 is a far cry from the specified 1920x1080, but hey . . . at least I've verified that the video out isn't dead. :)

I gave up trying to establish a build environment on my laptop (macports + xcode + OSX update management = rats nest), so I'm setting up a dedicated VM for building Android and will see if that works.  Would be very happy if that mxc_hdmi.c patch gets us going at a decent resolution.

--Matt

[Mark Schiech Jr]

Matt I'm glad to hear you got your board working. I found this tutorial http://eewiki.net/display/linuxonarm/Wandboard on making images for either ubuntu or debian. It builds the bootloader and linux kernel. I've used it with no problems. It seems to have several patches for HDMI and other issues on the i.MX6. Also, check the comments on the page for more information.

[Martin Wild]

Hi, Mark!

Ive als found this tutorial @ http://eewiki.net/display/linuxonarm/Wandboard and gave it a try to build my own up-to-date ubuntu image while waiting for my quad wandboard to arrive.
Have you just used it to build an replacement Kernel/U-boot with it or did you create a whole new running card? I am asking because the layout of the created card is completely different from the one wandboard.org offers. and i wonder if it will boot also.

Regards, Martin

[Mark Schiech Jr]

Check the comments page at http://eewiki.net/display/linuxonarm/Wandboard_Comments, people were saying it was working. I haven't had the time to follow the tutorial to find out if it works.

--

You received this message because you are subscribed to the Google Groups "Wandboard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wandboard+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Robert Nelson]

On Sun, Jun 23, 2013 at 6:20 PM, Mark Schiech Jr
<marksch...@gmail.com> wrote:
> Check the comments page at
> http://eewiki.net/display/linuxonarm/Wandboard_Comments, people were saying
> it was working. I haven't had the time to follow the tutorial to find out if
> it works.

Oh "it works" in so far as one user reported it working after tweaking
the u-boot patch.. I'm just waiting for Fedex to deliver a quad board
week, so that i can personally validate everything on my page. ;)

Regards,

--
Robert Nelson
http://www.rcn-ee.com/

[Mark Schiech Jr]

Are you talking about the problem Joerg Boge was having, they needed a package because the computer used 64 bit hardware. I counted and there are SEVERAL people reporting success. Reread the comments and you'll find out more information, don't just skim through and say you read it. The person that made this tutorial, Robert Nelson, has made several, so I think they know what they are doing.

[Robert Nelson]

On Sun, Jun 23, 2013 at 6:44 PM, Mark Schiech Jr
<marksch...@gmail.com> wrote:
> Are you talking about the problem Joerg Boge was having, they needed a
> package because the computer used 64 bit hardware. I counted and there are
> SEVERAL people reporting success. Reread the comments and you'll find out
> more information, don't just skim through and say you read it. The person
> that made this tutorial, Robert Nelson, has made several, so I think they
> know what they are doing.

That's funny... ;) I'm just saying as "author" of said page, I'm
personally also waiting for hardware. ;) Usually I don't add hardware
to the wiki pages, till I have the actual hardware on my desk.. But a
few users had validated it for me.

[Mark Schiech Jr]

I didn't even notice your name on the post. Great work on the tutorial and all the others you've posted.

[Mark Schiech Jr]

Robert, going by this page http://eewiki.net/display/linuxonarm/BeagleBone+Black  I assume you have this board? If so, can you tell me how true this post is that I made about the board, https://groups.google.com/forum/embed/?place=forum/wandboard&showsearch=true&showpopout=true&showtabs=false&parenturl=http%3A%2F%2Fwww.wandboard.org%2Findex.php%2Fforums#!topic/wandboard/hP2HUErZX3Y.

[Fabio Estevam]

On Sun, Jun 23, 2013 at 9:08 PM, Mark Schiech Jr
<marksch...@gmail.com> wrote:
> Robert, going by this page
> http://eewiki.net/display/linuxonarm/BeagleBone+Black I assume you have
> this board? If so, can you tell me how true this post is that I made about
> the board,
> https://groups.google.com/forum/embed/?place=forum/wandboard&showsearch=true&showpopout=true&showtabs=false&parenturl=http%3A%2F%2Fwww.wandboard.org%2Findex.php%2Fforums#!topic/wandboard/hP2HUErZX3Y.

Sorry, but this is totally off topic in this mailing list.

[Robert Nelson]

On Sun, Jun 23, 2013 at 7:08 PM, Mark Schiech Jr
<marksch...@gmail.com> wrote:
> Robert, going by this page
> http://eewiki.net/display/linuxonarm/BeagleBone+Black I assume you have
> this board? If so, can you tell me how true this post is that I made about
> the board,
> https://groups.google.com/forum/embed/?place=forum/wandboard&showsearch=true&showpopout=true&showtabs=false&parenturl=http%3A%2F%2Fwww.wandboard.org%2Findex.php%2Fforums#!topic/wandboard/hP2HUErZX3Y.

Keeping it short, as this is off topic..

So, in my experience, so far 'every' arm board ever released has some
'issues' when first released.

Support forums are exactly that, a location where users who have
issues with a 'board' can post support questions. No one ever takes
the time to sign up for a forum, then post a thread that everything
worked. And remember, the people who design/support the 'board' are
human too...

[Mark Schiech Jr]

Yes I know all of this. I just can't figure out why CircuitCo can't even make the attempts to correct some of the hardware issues. The BeagleBone Black is the second of the BeagleBone model, why are there so many issues? They should've been fixed in the revisions for the first model. They also had even more experience with the beagleBoard and BeagleBoard-xM. Why such crap quality with the new board, because it's just a competition with the Raspberry-Pi.

[Martin Wild]

Hi,Again

Thanks for the response! But to quick answer my own question:
I followed the tutorial now ,and build an ubuntu 13.05 Server image for wandboard quad on an X64 ubuntu 13.04 host, with an rootfs from linaro taken from here : http://releases.linaro.org/13.05/ubuntu/raring-images ,and it has been confirmed working! Ive uploaded it here for everyone to test :

http://www12.zippyshare.com/v/19364000/file.html

Regards, Martin

[Arthur Zeidin]

same problem here i just bought Quad and for some reasons my hdmi does not work. and then i tried to plug in my doul it does not work ether but it did work before. i will try to fix it by following instructions on forum if there is no success i will have no choice but return both products.

On Friday, June 14, 2013 12:27:35 PM UTC-4, Matthew W. Novak wrote:

Hi folks.  I need some help troubleshooting a WBQUAD issue: I'm not getting any video output from the HDMI out, and I'm not sure how to proceed.  Note that I am not a developer, but I am a competent Linux user, so while a lot of things here will be over my head, I should be able to gather troubleshooting and debugging doc without much trouble.

So far, I have . . . 

1. downloaded the WBQuad Android 4.2.2 image
   
   
• I've also walked through the following steps using the wandboard-quad-ubuntu-11.10-20130604.img image with the same results
  
  
2. DD'd that image to my mSD card per the Quick Start Guide: dd if=./android-4.2.2-wand-quad-preview-20130607.img of=/dev/[my_SD_card] bs=1M
   
   
• I've DD'd the image both from Linux and OSX machines with the same results
  
  
3.  inserted that mSD card into the OS mSD card slot on the processor board, plugged in an HDMI-to-DVI cable that's attached to an HDCP-enabled monitor, and plugged in my 5V 2A 10W power supply
   
   
• I've tried this with multiple monitors and multiple HDMI-to-DVI cables with the same results
• I've tried booting with a USB keyboard attached, and/or with an OTG cable attached, with the same results
  
  
4. in every iteration above, a green light on the board comes on and nothing happens at the monitor
   
   
• to reiterate, this is the same result whether booting the Android or Ubuntu images from the Wandboard downloads page

However, at least with Android, the image appears to be booting: with an OTG cable plugged between the board and my laptop, I can do the following:

• "adb devices" and see the board listed
• "adb shell" and get into a functional Android command prompt
• "logcat" from within the adb shell and watch as the boot progresses and generates log entries

If I leave logcat running indefinitely, I eventually get kicked out of the shell and the board disappears from "adb devices".  

I've also noticed that I can only access the device via adb on the first boot after DD'ing an image to the mSD card: the board does not appear in "adb devices" after any subsequent boot of the same mSD card (reset button, reboot from within the adb shell, hard power off/power on, etc.).  

So, I appear to have a functional board with a functional Android image on the mSD card, but I'm not getting any video output via HDMI, and the instability between first and subsequent boots isn't encouraging.  Any advice on where to start?

Note that I don't have a null modem cable available, but will be able to pick one up over the weekend to help with debugging.

--Matt  

[Mark Schiech Jr]

I don't think you would've received two different model boards that were both defective. You should read the forums and ask questions, it could just be a software issue. Try different images for the boards and if possible, different monitors and/or TVs.

[Arthur Zeidin]

I have read forum twits and my doul board was working fine until i tried to reconnected  again with updated ubuntu image still there is no hdmi connection.

[Mark Schiech Jr]

Try the Yocto image or Android image. I use the Yocto image. If you have a linux computer you can follow this tutorial to create your own image using Ubuntu/Debian filesystem, http://eewiki.net/display/linuxonarm/Wandboard You will need to create two different images, one for each model board.

[Arthur Zeidin]

thanks i will try

[Martin Wild]

Hi!

Maybe this issues can be fixed by passing new display settings to u-boot. Until now I am using an cuvieboard which also depends on u-boot AND has similar display problems. You can create a file, called uEnv.txt and place it on the first partition of your sdcard to pass various settings to the kernel. Unfortunately the provided wandboard images are using a little bit different layout sui there is only one partition. However if you'd create a card as described inside the eelinux tutorial (search for ubuntu 13.05 inside the forum for correct link) you'll get a working card with the needed layout, and a example uEnv.txt. with this you can search for the needed parameter s to change the default resolution. (Help can be found inside cubieboard.org forums)

I think this is worth a try even if the wandboard uses a different soc for the basic s are the same until now. More information about uEnv.txt and U-boot can be found in the u-boot wiki.

Sorry for the sloppy post, but I am writing from my mobile phone and cannot simply copy& paste... ;)

Regards, Martin

[Martin Wild]

Ok, Here it is a little bit more precise:

The tutorial to create your  custom card with the uEnv.txt support is here: http://eewiki.net/display/linuxonarm/Wandboard

IF EDID is supported at the moment this uEnv.txt should force the default screen mode to 720p: (Thanks to Floris Bos for the help!)

create a uEnv.txt with the following 3 lines:


#mmcboot=fatload mmc 0 0x43000000 script.bin || fatload mmc 0 0x43000000 evb.bin; fatload mmc 0 0x48000000 uImage; if fatload mmc 0 0x43100000 uInitrd; then bootm 0x48000000 0x43100000; else bootm 0x48000000; fi
#uenvcmd=run mmcboot
bootargs=console=ttyS0,115200 console=tty0 disp.screen0_output_mode=EDID:1280x720p60 hdmi.audio=EDID:0


(note that it are 3 long lines starting with "mmcboot=" "uenvcmd=" and
"bootargs=", however the first two are Cubieboard-specific so i would leave them commented them out. )

Maybe this can help.

Regards, Martin

[Matthew W. Novak]

Wanted to follow up on this thread.  Allowing that I'm out of my depth here, I appear to have successfully built Android with the Boundary mxc_hdmi.c CEA patch that I mentioned in an earlier post.  Boot log attached.  Unfortunately, under the Android image (stock or patched), I still can't get video out to my DVI monitor.  I can't even get the 640x480 I'm able to get under the Linux images using the "video=mxcfb0:dev=hdmi,1920x1080P@60,if=RGB24" bootargs.

Any tips or suggestions on how to proceed or at least kick the can down the road a bit?

I may have a whirl with the eewiki tutorial that's been discussed and/or the 13.04 server image that Martin recently posted to the forum, but my real hope is to get Android working at a decent resolution.

Any help is appreciated,

--Matt

[jth...@gmail.com]

Hi Matt,

My video is in the same state as yours as I'm looking at 640x480.  I am now working on setting up the boot as per Robert Benson's instructions.  Yesterday's attempts to rebuild the kernel and update uImage using the base instructions (write uImage into unallocated space) failed repeatedly - checksum errors and incorrect byte count reported at boot.

Good news!  Got things configured and running using Robert Benson's method.  Made the mxc_hdmi.c changes prior to building the kernel.  I see no noticeable difference in behavior.  I'm looking at 640x480 still.  At least it is much easier to update kernels and boot settings!

Hat tip to Mr. Benson.  Thank you very much!

All the best,   Joe

[jth...@gmail.com]

Hey Matt,

I dropped my bootargs video setting all the way down to 720x480P@60 to get the 640x480 output.  The higher one(s) which my monitor supports (1440x900 max) brought no joy.

Even my SABRE Lite gives me an add size - 1152x894 or something like that.  I think we're anxiously awaiting some additional driver work...

Ever onward,   Joe

[Martin Wild]

Hi,

has anyoune here tried to plug his wandboard inside a "real" TV with hdmi-port? many monitors have issues with standard hdmi resolutions like 720p or 1080p ,even if thats their native resolution . also hdmi-audio can be an issue.
if you see something like a pink stripe at one screen edge then this can be badly embedded hdmi-audio which your screen is trying to display. (no joke) .
Oh man, i hope this gets fixed soon i can live with almost every (temporary) bug, but having no usable display output when setting up your system is really annoying...

Regards, Martin

[jth...@gmail.com]

Hi Martin,

From one of my posts of June 20...

Since I had been exercising the WBQUAD here in the office and all of the monitors available are DVI, I took the SBC home last night.  I plugged it into a 22" TV/monitor and, lo and behold, on the second try (initial boot the display remained dark and I hit the reset button) the desktop appears!

That is an inexpensive Sansui TV with a couple HDMI in along with legacy ports.

Best,   Joe

[abr...@gmail.com]

In my own research I came across the linux tools get-edid and parse-edid.  These tools ask connected monitors what resolutions they support and pass the result back for static use in an xorg-conf or dynamically through some mechanism (environmental variable?)   Ubuntu normally does it automagically on the desktop so I don't know.  Anyhow it seems like a lack of EDID inofrmaiton or a lack of using that information is at the core of our display problems with the wandboard.

On a whim I searched the forum here and found a hardware and driver level discussion on the issue in a largely unrelated thread.  Search for "NVCC_MIPI" and scroll towards the bottom.  Sorry I don't know how to link directly.

It seems like EDID information is tricky to collect with the wandboard's baseboard configuration and there may be hardware issues/bugs as well.  Given that it was discussed back in April I'm pretty disappointed that the Quad was released with these issues/problems known.  I was close to buying a sabrelite instead but the cheaper price on the wandboard got me to pull the trigger.  I think it is worth development time to resolve this.  Certainly I have wasted more time trying to get full resolution on my wandboard that the cost differential justifies in retrospect.  At least I am learning something new?

Maybe I just need to use the wandboard headless as a server until this is fixed.

[www.wandboard.org]

Hi,

We created a few test images that can be used to check basic resolutions with a HDMI display. We tested these on a number of LG, Acer, Viewsonic and ChiMei monitors and TV's.

Please check both images below and connect a serial port to post log information.

FILES for the Wandboard QUAD :

www.wandboard.org/images/downloads/wand-quad-hdmi-640x480-test.img.xz

www.wandboard.org/images/downloads/wand-quad-hdmi-1920x1080-test.img.xz

***

You can also take a look at the CEA/EIA-861E standard resolutions, and set the resolution (u-boot parameter) thereafter.
http://en.wikipedia.org/wiki/Extended_display_identification_data#CEA_EDID_Timing_Extension_Version_3_data_format

***

If you have problems. Please post your monitor brand/model, resolution (native), and if this is a DVI or HDMI display.

If you solve your issue. Please also report.

***
A couple things to keep in mind while testing.

It could be that you are using a DVI monitor that lack digital pins. Please check and use a digital interface on your display (ideally a HDMI monitor/TV).

It could be that older DVI monitors require an image format that is not default in our demo software (RGB8888/YCbCr444), and since DVI monitors do not carry the I2C signals we are unable to read EDID and therefore can not change to the right format.

[Matthew W. Novak]

For what it's worth, we're finally up and running with Android over straight HDMI into a 1080p television.  The existing Android image boots to 720p, but we got 1080p by passing the following to u-boot:

setenv bootargs 'console=ttymxc0,115200 init=/init video=mxcfb0:dev=hdmi,1920x1080M@60,if=RGB24,bpp=32 video=mxcfb1:off video=mxcfb2:off fbmem=28M vmalloc=400M androidboot.console=ttymxc0 androidboot.hardware=freescale mem=1792M@0x10000000'

It's not gorgeous on a television, but it works as intended.

Absolutely no joy ever with HDMI-to-DVI into either a Samsung 2433BW (1920x1200) or an older Samsung 193P (1280x1024).  This is with a DVI-D cable, per this illustration (no analog signal present).

  

--Matt 

[jth...@gmail.com]

Hi Matt,

Happy to hear you got it showing at full resolution!

wandboard.org - one thing I notice is the focus on recognized CEA modes.  Is there no support for monitors which are "pure" DMT?  It appears the Dell 1909W monitor I am using does not report any CEA capability.  There are 16 supported DMT modes and zero (0) supported CEA modes.

All the best,   Joe

[Arthur Zeidin]

I have tried everything and it still does not work i guess i will have no choice but to return this board, i just simply dont have time to deal with it and try to fix it. for 135 bucks hdmi should be the least problem to deal with it. By the way, my doul board stop working as well after ubuntu image update and even when it worked i still had problem with bluetooth and hdmi volume. Simply to put this board is not ready yet it still ought to be fixed if they want to compete with Pi. Meanwhile, i will try to get my money back. And wait for my new board from 

 http://www.adapteva.com/ 

[Martin Wild]

Not that it is really my problem, but do you really think the next one gets better? According to my experience with raspberry pi and cubieboard audio and/or video support is always a problem when using a arm soc due to closed source binary drivers. So a little bit patience might be a good choice. The wbquad of is not even a month out...

Regards, Martin

[Liberto Camús]

I've just received the WBQuad and it works, but doesn't detect correctly the resolution of my monitor. It's a ASUS VE228 (1920x1080) but using an HDMI cable I get 1280x720 and can't change it.

I've been investigating but can't figure out how to do it. I don't want to bother nobody with noob questions, but I'm totally lost. How do I change u-boot bootargs? Can it be done from the running ubuntu image in the wbquad?  If you could send me a link to the correct docs I would really appreciate it.

Thanks a lot

Liberto

[jth...@gmail.com]

Hi Liberto,

Read back through this thread and find the discussion of bootargs, specifically setting the "video=" component.  That should put you on the path.  In essence, interrupt the boot process, update the bootargs string and continue.  Once your happy withthe display do a save before completing the boot.

If you're following Robert Nelson's methodology, modify the "video=" line in uEnv.txt and reboot.

Best of luck,   Joe

[Liberto Camús]

Thanks a lot, Joe!

Liberto

[Michele Adduci]

Hi, 

i am just a new user of Wandboard Quad.

I followed this tutorial (http://www.eewiki.net/display/linuxonarm/Wandboard#Wandboard-SetupmicroSDcard) and tuning the Kernel option from the menu that showes up during the building on the kernel, i turned on the graphical option related to HDMI capabilities.
Now i can boot into my Wandboard Quad (a minimal Ubuntu 13.04 setup) and i can display the terminal console via a standard HDMI cable. My monitor is a Samsung TV-LED 1080p.
It just takes a lot of time to boot, maybe because of the microSD card i am using (not so new).

Best

[Arthur Zeidin]

i think we are missing  a point here if i am willing to pay 130 bucks i want it to work without any patience. And not to mention the price of pi is 35 bucks for that price i am willing to put up with  minor issues.  

[Arcanjo Azrael]

Hi,

Do you remind what graphical option you enabled ?

Thanks and excuse me my bad english.

Leonardo

[Jay Custin]

Well hopefully those of you who have experienced issues with HDMI video haven't made the same mistake I did, but the photograph/image in the manual showing the slot for the OS may have mislead you as it did for me.  The Wandboard QUAD really DOES have two microSD slots.  The slot for the OS you will see on p. 5 of the user guide and does NOT match the opening in the expansion box!  Instead you will find it under the large aluminum heat sink on the smaller (you can see my red/gray microSD card in the attached image).

Jay

[abr...@gmail.com]

My wandboard quad was outputting 480p and I was working through xorg.conf to get some higher resolutions working on an HDMI projector and a DVI monitor.  Well HDMI stopped working all together - no output at all now - both displays.  I have reimaged the SD card twice now with the stock image and nothing.  Maybe I will try android as a last change before shipping the board back.  I'm struggling to understand how you can brick a board that boots off SD?  Does the ubuntu image ship with telnet or ssh enabled?  Maybe its just a display issue?

I have to agree with you Arthur - it's just not ready yet.  I have worked with the Dreamplug, RPi, router firmware, media streamer firmware and vanilla x86 linux.  I got everything working very well with a bit of effort.  Nothing has been more full of promise and light on delivery than my wandboard.  The raspberry pi is a slow dog but it works.  My dreamplug and mediastreamers are tanks - always work.  I was going to buy the Sabrelite quad which is also IMX6 and seems more mature, maybe that's the way to go.  I really wanted gigabit ethernet, a sata connector and the ability to run XBMC.  IMX6 fits the bill perfectly.  I could retire my somewhat underpowered dreamplug NAS / server (or at least repurpose it) and stream to my projector off the same box.  Instead I wasted a bunch of time.

Can anyone talk me out of giving up?  Is the GK802 image worth a shot?

[jth...@gmail.com]

Whether you stick it out or not is going to depend on what you really want to do with the board.  It isn't surprising to me that the video output is fragile.  That is going to turn out to be a kernel thing which manifests itself in various ways on different platforms.  I also have a couple SABRE Lite boards.  Their video out is quite limited as well.  So why did I pick up a Wandboard given the SABRE Lites?  This comes back to what you want to do with the eval board.  The wandboard presents usable SPI and multiple I2C interfaces.  Those are valuable to me and are not presented on the SL.  On the other hand the SL provides easy access to PCIe which may prove important later.  But for now it isn't.  Both allow me access to MIPI DSI. 

I've been doing embedded a long, long time and it is usually the case that new SBCs take a while to stabilize and for support of specific, unique hardware to arrive.  Shoot, even the Pi's gpu currently has minimal support though we can expect acceleration via Wayland/Weston in the not too distant future.  Thus, for the time being, the Pi CPU is doing all the graphics heavy lifting, affecting overall performance.

If you really need complete, adaptive graphics/video support it's obvious this isn't there - yet.  Two choices, maybe three, put it on the shelf for a while until the graphics support improves; dig in, figure it out, make it work and release your modifications back to the community - we'd all love it!; send it back and forget about it.

To me, $130 for a WBQUAD is a steal.  2GB RAM, functioning SATA (currently running rootfs from there), working wired and wifi enet, i2c & spi, and much more.  Life is good.

BTW, with the SATA SSD hooked up I have different root file systems in different partitions.  Modifying uEnv.txt allows me to move from one to another.  This morning I loaded the xubuntu_12_04 image mentioned on another thread.  Came right up. 

All the best,   Joe

[Arcanjo Azrael]

Hi,

I got something... when I was trying that Xubuntu image from gk802, I
dont get any image.

So, I try some video args in uEnv.txt and I got video, in a Samsung
2232BW, with HDMI-DVI cable.

video=mxcfb0:dev=hdmi,1440x900M@60,if=RGB32

My monitor reports I am seeing a 1280x1024, 64.2KHz, 60Hz, PP .

However, with the Xubuntu image, I am unable to get to graphical
interface, only text.

Excuse my bad english.

Leonardo

[Arcanjo Azrael]

Hi,

I forgot to add: I get the idea to try the 1440x900 resolution from here:

https://community.freescale.com/thread/302653

[jth...@gmail.com]

Hey Leonardo!

Thank you!  Thank you!  Thank you!  Thank you! 

I'd give you a big hug if I could!  My monitor also reports 1280x1024 but it is much more usable than the 640x480.

Xubuntu came up in full blue color motif for me.  No problem there.

Happy Tuesday!    Joe

[abr...@gmail.com]

Yea - what you said.  I want it to be further along and it isn't...  So few boards have the SATA connector and gigabit.  Any idea why my board would totally conk out on video?  I am really at a loss given that I've reimaged the SD card.  Is there any chance the SD needs to be reformatted / blanked to fully reset.  Maybe just trying a new image will work.  GK802 looks promising.

[jth...@gmail.com]

No, I really don't.  Have had sporadic troubles with bad cables and connectors on occasion. 

Are you running  your board with the serial debug output live?  If your SD went bad I'd expect to see evidence of corruption - bad reads and whatnot - like Matt saw last week (or maybe the week before).  I've been using SanDisk Ultra 16GB cards.  They are class 10 and perform well.  I am pretty careful with formal shutdowns and reboots, sync commands when things look a bit wonky or there could be pending writes queued somewhere.  Slower cards (class 4 and below) have been troublesome for me on other projects and platforms.

It wouldn't hurt anything to reformat your uSD and reload it - just time.  If you don't have that debug output live I highly recommend you hook it up.

All the best,   Joe

[abr...@gmail.com]

[Arthur Zeidin]

Hi, 

I just ordered new cubieboard and will test it out, so far what i have read  on web,  it works better and faster then pi and people dont ave problem with hdmi. I will also try to make NAS out of it. Meanwhile, i will read more about mini android stick pc, such as CX-919 RK3188 Quad Core ARM. According to spec.. it equipped with AndroidV4.1.1 on Linux Kernel:3.0.8+. I will try to connect it to kinect and make it work. i will keep you post it.

[Arthur Zeidin]

Here is the link for CX-919 RK3188 http://www.geekbuying.com/item/US-Plug-CX-919-RK3188-Quad-Core-ARM-Cortex-A9-1-6Ghz-Android-4-1-1-Mni-PC-TV-Box-2G-8G-BT-HDMI-Black-314977.html

[Martin Wild]

@Arthur

The cubieboard is a great arm board, and gives a good nas system. I know it because i own one. But don't except too much in your HDMI issues. Until now both systems suffer exactly the same problems. The difference is the cubies ones are older and so there are more fixes ( but i can still use only my TV and one of three monitors). A true HDMI interface is almost a must have. If you want the best results,you should start with flori's berryboot image. This has the best compatibility but also limitations to "real" Linux distros.

Regards, Martin

[Arthur Zeidin]

thanks for an update. meanwhile could you please give me more details about how you were able to set up NAS on cubie... I would help me a lot. Thanks.

[Arthur Zeidin]

@Joe 

I cannot agree more with you it could be a steal if only HDMI worked on the clock. 

On Tuesday, July 2, 2013 12:03:01 PM UTC-4, jth...@gmail.com wrote:

[Martin Wild]

Hi,arthur, To take the thread not too far out of topic...

Building a nas requires always setting up nfs or samba (in my case i use samba) and some shares. Howto do this can be found everywhere in the web. I recommend using gadmin-samba if you planning to have a desktop running. For the quick start get yourself your favourite os here http://linux-sunxi.org/Bootable_OS_images
Also linux-sunxi.org contains info about most cubie related stuff you might need.
Meanwhile my wbquad is running fine maybe ill upload a image with SSH and running vlc from login later.

Regards,Martin

[Arcanjo Azrael]

Hi,

What is your current rootfs ?

Do you have a graphical interface with the correct drivers to video (low
cpu usage) and stable ?

If yes, do you mind to share ? Because I dont have any luck with my
limited knowledge.

Another thing, what speed do you get with samba ? At least for me, samba
gives me only 30MB/s with 100% single core usage. And ftp I got the same
speed.

Leonardo

[Martin Wild]

Hi

i am using currently http://releases.linaro.org/12.11/ubuntu/precise-images/alip/linaro-precise-alip-20121124-519.tar.gz

(see here https://groups.google.com/forum/#!topic/wandboard/00jsrK45TnE )

and i dont have or need any hw accelleration for my wb is on top of my living-room shelf running 24/7 unseen. what doesnt mean i qouldnt appreciate it. Also i cannot say much about my samba speed only that it uses the  full 100mbit my router supports. :)

Regards,Martin

[Zoom_h2]

Hello,

as you ask for this is may boot log for wand-dev-sd-20130204-quad.img.xz  :

With

setenv bootargs 'console=ttymxc0,115200 video=mxcfb0:dev=hdmi,1440x900M@60,if=RGB32 root=/dev/mmcblk0p1 rootfstype=ext4 rootwait rw'  

i get a 1280x1024 screen working only on a hp L2245w with a HDMI to DVI cabel.

If i change the cabel while running i will get no screen on a other monitor.

U-Boot 2013.04-00234-gf6c3b34-dirty (May 28 2013 - 20:37:08)


CPU:   Freescale i.MX6Q rev1.2 at 792 MHz

Reset cause: WDOG

Board: Wandboard

DRAM:  2 GiB

MMC:   FSL_SDHC: 0

In:    serial

Out:   serial

Err:   serial

Net:   FEC

Warning: FEC MAC addresses don't match:

Address in SROM is         00:1f:7b:b4:02:f8

Address in environment is  00:1f:7b:b4:00:04


Hit any key to stop autoboot:  1 0

mmc0 is current device


MMC read: dev # 0, block # 2048, count 4080 ... 4080 blocks read: OK

## Booting kernel from Legacy Image at 10800000 ...

   Image Name:   Linux-3.0.35+

   Image Type:   ARM Linux Kernel Image (uncompressed)

   Data Size:    1960092 Bytes = 1.9 MiB

   Load Address: 10008000

   Entry Point:  10008000

   Verifying Checksum ... OK

   Loading Kernel Image ... OK

OK


Starting kernel ...


[    0.000000] Linux version 3.0.35+ (tapani@triceratops) (gcc version 4.5.1 (Sourcery G++ Lite 2010.09-50) ) #5 SMP PREEMPT Thu Jun 6 21:41:26 CST 2013
[    0.000000] CPU: ARMv7 Processor [412fc09a] revision 10 (ARMv7), cr=10c53c7d
[    0.000000] CPU: VIPT nonaliasing data cache, VIPT aliasing instruction cache
[    0.000000] Machine: Wandboard
[    0.000000] Memory policy: ECC disabled, Data cache writealloc
[    0.000000] CPU identified as i.MX6Q, silicon rev 1.2
[    0.000000] PERCPU: Embedded 7 pages/cpu @d100e000 s5408 r8192 d15072 u32768
[    0.000000] Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 475136
[    0.000000] Kernel command line: console=ttymxc0,115200 video=mxcfb0:dev=hdmi,1440x900M@60,if=RGB32 root=/dev/mmcblk0p1 rootfstype=ext4 rootwait rw
[    0.000000] PID hash table entries: 2048 (order: 1, 8192 bytes)
[    0.000000] Dentry cache hash table entries: 65536 (order: 6, 262144 bytes)
[    0.000000] Inode-cache hash table entries: 32768 (order: 5, 131072 bytes)
[    0.000000] Memory: 80MB 1792MB = 1872MB total
[    0.000000] Memory: 1894424k/1894424k available, 202728k reserved, 1409024K highmem
[    0.000000] Virtual kernel memory layout:
[    0.000000]     vector  : 0xffff0000 - 0xffff1000   (   4 kB)
[    0.000000]     fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
[    0.000000]     DMA     : 0xf4600000 - 0xffe00000   ( 184 MB)
[    0.000000]     vmalloc : 0xea800000 - 0xf2000000   ( 120 MB)
[    0.000000]     lowmem  : 0xc0000000 - 0xea000000   ( 672 MB)
[    0.000000]     pkmap   : 0xbfe00000 - 0xc0000000   (   2 MB)
[    0.000000]     modules : 0xbf000000 - 0xbfe00000   (  14 MB)
[    0.000000]       .init : 0xc0008000 - 0xc0032000   ( 168 kB)
[    0.000000]       .text : 0xc0032000 - 0xc050f000   (4980 kB)
[    0.000000]       .data : 0xc0510000 - 0xc054bca0   ( 240 kB)
[    0.000000]        .bss : 0xc054bcc4 - 0xc0568784   ( 115 kB)
[    0.000000] Preemptible hierarchical RCU implementation.
[    0.000000]  Verbose stalled-CPUs detection is disabled.
[    0.000000] NR_IRQS:496
[    0.000000] MXC GPIO hardware
[    0.000000] sched_clock: 32 bits at 3000kHz, resolution 333ns, wraps every 1431655ms
[    0.000000] arm_max_freq=1GHz
[    0.000000] MXC_Early serial console at MMIO 0x2020000 (options '115200')
[    0.000000] bootconsole [ttymxc0] enabled
[    0.000000] Console: colour dummy device 80x30
[    0.211789] Calibrating delay loop... 1581.05 BogoMIPS (lpj=7905280)
[    0.299912] pid_max: default: 32768 minimum: 301
[    0.304870] Mount-cache hash table entries: 512
[    0.310113] CPU: Testing write buffer coherency: ok
[    0.315313] hw perfevents: enabled with ARMv7 Cortex-A9 PMU driver, 7 counters available
[    0.415941] CPU1: Booted secondary processor
[    0.545942] CPU2: Booted secondary processor
[    0.675945] CPU3: Booted secondary processor
[    0.765367] Brought up 4 CPUs
[    0.781204] SMP: Total of 4 processors activated (6324.22 BogoMIPS).
[    0.818327] print_constraints: dummy:
[    0.822424] NET: Registered protocol family 16
[    0.832802] print_constraints: vddpu: 725 <--> 1300 mV at 1150 mV fast normal
[    0.840456] print_constraints: vddcore: 725 <--> 1300 mV at 1150 mV fast normal
[    0.848279] print_constraints: vddsoc: 725 <--> 1300 mV at 1200 mV fast normal
[    0.856021] print_constraints: vdd2p5: 2000 <--> 2775 mV at 2400 mV fast normal
[    0.863822] print_constraints: vdd1p1: 800 <--> 1400 mV at 1100 mV fast normal
[    0.871551] print_constraints: vdd3p0: 2625 <--> 3400 mV at 3000 mV fast normal
[    0.961465] hw-breakpoint: found 6 breakpoint and 1 watchpoint registers.
[    0.968309] hw-breakpoint: 1 breakpoint(s) reserved for watchpoint single-step.
[    0.975679] hw-breakpoint: maximum watchpoint size is 4 bytes.
[    0.981578] L310 cache controller enabled
[    0.985626] l2x0: 16 ways, CACHE_ID 0x410000c7, AUX_CTRL 0x02070000, Cache size: 1048576 B
[    1.026623] bio: create slab <bio-0> at 0
[    1.031746] print_constraints: VDDA: 2500 mV
[    1.036420] print_constraints: VDDIO: 3300 mV
[    1.041438] vgaarb: loaded
[    1.045475] SCSI subsystem initialized
[    1.050124] spi_imx imx6q-ecspi.0: probed
[    1.054496] spi_imx imx6q-ecspi.1: probed
[    1.059558] usbcore: registered new interface driver usbfs
[    1.065301] usbcore: registered new interface driver hub
[    1.070859] usbcore: registered new device driver usb
[    1.095377] imx-ipuv3 imx-ipuv3.0: IPU DMFC NORMAL mode: 1(0~1), 5B(4,5), 5F(6,7)
[    1.103815] Advanced Linux Sound Architecture Driver Version 1.0.24.
[    1.111441] Bluetooth: Core ver 2.16
[    1.115191] NET: Registered protocol family 31
[    1.119671] Bluetooth: HCI device and connection manager initialized
[    1.126064] Bluetooth: HCI socket layer initialized
[    1.130965] Bluetooth: L2CAP socket layer initialized
[    1.136053] Bluetooth: SCO socket layer initialized
[    1.141538] cfg80211: Calling CRDA to update world regulatory domain
[    1.148115] Switching to clocksource mxc_timer1
[    1.170667] NET: Registered protocol family 2
[    1.175526] IP route cache hash table entries: 16384 (order: 4, 65536 bytes)
[    1.184047] TCP established hash table entries: 65536 (order: 7, 524288 bytes)
[    1.192659] TCP bind hash table entries: 65536 (order: 7, 786432 bytes)
[    1.200719] TCP: Hash tables configured (established 65536 bind 65536)
[    1.207284] TCP reno registered
[    1.210452] UDP hash table entries: 256 (order: 1, 8192 bytes)
[    1.216337] UDP-Lite hash table entries: 256 (order: 1, 8192 bytes)
[    1.223244] NET: Registered protocol family 1
[    1.227821] _regulator_get: etb supply vcore not found, using dummy regulator
[    1.235610] _regulator_get: etm.0 supply vcore not found, using dummy regulator
[    1.243056] _regulator_get: etm.1 supply vcore not found, using dummy regulator
[    1.250473] _regulator_get: etm.2 supply vcore not found, using dummy regulator
[    1.257915] _regulator_get: etm.3 supply vcore not found, using dummy regulator
[    1.265746] Static Power Management for Freescale i.MX6
[    1.270993] wait mode is disabled for i.MX6
[    1.275458] cpaddr = ea880000 suspend_iram_base=ea874000
[    1.280881] PM driver module loaded
[    2.005709] link up failed, DB_R0:0x00c56600, DB_R1:0x08200000!
[    2.011653] IMX PCIe port: link down with power supply 0!
[    2.017616] IMX usb wakeup probe
[    2.021499] IMX usb wakeup probe
[    2.025338] cpu regulator init ldo=0
[    2.029145] i.MXC CPU frequency driver
[    2.055972] highmem bounce pool size: 64 pages
[    2.060828] ashmem: initialized
[    2.065689] msgmni has been set to 948
[    2.069822] io scheduler noop registered
[    2.073924] io scheduler cfq registered (default)
[    2.080776] mxc_sdc_fb mxc_sdc_fb.0: register mxc display driver hdmi
[    2.087292] mxc_hdmi mxc_hdmi: Detected HDMI controller 0x13:0xa:0xa0:0xc1
[    2.094230] fbcvt: 1440x900@60: CVT Name - 1.296MA
[    2.111861] imx-ipuv3 imx-ipuv3.0: WARNING: adapt panel end blank lines
[    2.122899] imx-ipuv3 imx-ipuv3.0: IPU DMFC DP HIGH RESOLUTION: 1(0,1), 5B(2~5), 5F(6,7)
[    2.156249] imx-ipuv3 imx-ipuv3.0: WARNING: adapt panel end blank lines
[    2.192283] Console: switching to colour frame buffer device 160x64
[    2.240376] mxc_sdc_fb mxc_sdc_fb.1: register mxc display driver ldb
[    2.246787] _regulator_get: get() with no identifier
[    2.271094] imx-sdma imx-sdma: loaded firmware 1.1
[    2.281259] imx-sdma imx-sdma: initialized
[    2.285796] Serial: IMX driver
[    2.288960] imx-uart.0: ttymxc0 at MMIO 0x2020000 (irq = 58) is a IMX
[    2.295580] console [ttymxc0] enabled, bootconsole disabled
[    2.295580] console [ttymxc0] enabled, bootconsole disabled
[    2.307029] imx-uart.2: ttymxc2 at MMIO 0x21ec000 (irq = 60) is a IMX
[    2.314315] [drm] Initialized drm 1.1.0 20060810
[    2.319450] [drm] Initialized vivante 1.0.0 20120216 on minor 0
[    2.368351] No sata disk.
[    2.371255] FEC Ethernet Driver
[    2.375415] fec_enet_mii_bus: probed
[    2.380313] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[    2.386987] add wake up source irq 72
[    2.390703] fsl-ehci fsl-ehci.1: Freescale On-Chip EHCI Host Controller
[    2.397500] fsl-ehci fsl-ehci.1: new USB bus registered, assigned bus number 1
[    2.432727] fsl-ehci fsl-ehci.1: irq 72, io base 0x02184200
[    2.452712] fsl-ehci fsl-ehci.1: USB 2.0 started, EHCI 1.00
[    2.458906] hub 1-0:1.0: USB hub found
[    2.462674] hub 1-0:1.0: 1 port detected
[    2.466873] add wake up source irq 75
[    2.473599] fsl-ehci fsl-ehci.0: Freescale On-Chip EHCI Host Controller
[    2.480249] fsl-ehci fsl-ehci.0: new USB bus registered, assigned bus number 2
[    2.512727] fsl-ehci fsl-ehci.0: irq 75, io base 0x02184000
[    2.532718] fsl-ehci fsl-ehci.0: USB 2.0 started, EHCI 1.00
[    2.538940] hub 2-0:1.0: USB hub found
[    2.542731] hub 2-0:1.0: 1 port detected
[    2.547128] Initializing USB Mass Storage driver...
[    2.552217] usbcore: registered new interface driver usb-storage
[    2.558245] USB Mass Storage support registered.
[    2.563211] ARC USBOTG Device Controller driver (1 August 2005)
[    2.569203] udc: request mem region for fsl-usb2-udc failed
[    2.574909] fsl-usb2-udc: probe of fsl-usb2-udc failed with error -16
[    2.581649] mousedev: PS/2 mouse device common for all mice
[    2.587413] i2c /dev entries driver
[    2.591590] Linux video capture interface: v2.00
[    2.596598] mxc_v4l2_output mxc_v4l2_output.0: V4L2 device registered as video16
[    2.604206] mxc_v4l2_output mxc_v4l2_output.0: V4L2 device registered as video17
[    2.611788] mxc_v4l2_output mxc_v4l2_output.0: V4L2 device registered as video18
[    2.619501] usbcore: registered new interface driver uvcvideo
[    2.625268] USB Video Class driver (v1.1.0)
[    2.629740] Bluetooth: HCI UART driver ver 2.2
[    2.634209] Bluetooth: HCI H4 protocol initialized
[    2.639070] cpuidle: using governor ladder
[    2.643221] cpuidle: using governor menu
[    2.647281] sdhci: Secure Digital Host Controller Interface driver
[    2.653519] sdhci: Copyright(c) Pierre Ossman
[    2.658032] _regulator_get: sdhci-esdhc-imx.2 supply vmmc not found, using dummy regulator
[    2.666588] mmc0: SDHCI controller on platform [sdhci-esdhc-imx.2] using DMA
[    2.674741] _regulator_get: sdhci-esdhc-imx.1 supply vmmc not found, using dummy regulator
[    2.685336] mmc1: SDHCI controller on platform [sdhci-esdhc-imx.1] using DMA
[    2.692452] sdhci sdhci-esdhc-imx.0: no write-protect pin available!
[    2.699864] _regulator_get: sdhci-esdhc-imx.0 supply vmmc not found, using dummy regulator
[    2.710370] mmc2: SDHCI controller on platform [sdhci-esdhc-imx.0] using DMA
[    2.717707] mxc_vdoa mxc_vdoa: i.MX Video Data Order Adapter(VDOA) driver probed
[    2.725920] VPU initialized
[    2.730930] mxc_asrc registered
[    2.754839] Thermal calibration data is 0x57c4fc69
[    2.759639] Thermal sensor with ratio = 181
[    2.763844] mxc_hdmi mxc_hdmi: Read EDID again
[    2.782715] Anatop Thermal registered as thermal_zone0
[    2.782743] usb 1-1: new low speed USB device number 2 using fsl-ehci
[    2.796167] anatop_thermal_probe: default cooling device is cpufreq!
[    2.799178] mmc0: new high speed SDHC card at address 59b4
[    2.799652] mmcblk0: mmc0:59b4 00000 7.34 GiB
[    2.800900]  mmcblk0: p1 p2
[    2.816129] usbcore: registered new interface driver usbhid
[    2.821707] usbhid: USB HID core driver
[    2.825579] zram: num_devices not specified. Using default: 1
[    2.831331] zram: Creating 1 devices ...
[    2.836233] usbcore: registered new interface driver snd-usb-audio
[    2.847076] mxc_hdmi_soc mxc_hdmi_soc.0: MXC HDMI Audio
[    2.852505] mxc_spdif mxc_spdif.0: MXC SPDIF Audio
[    2.860036] _regulator_get: 1-000a supply VDDD not found, using dummy regulator
[    2.867695] sgtl5000 1-000a: sgtl5000 revision 17
[    2.872608] print_constraints: 1-000a: 850 <--> 1600 mV at 1200 mV normal
[    2.905654] asoc: sgtl5000 <-> imx-ssi.1 mapping ok
[    2.912852] asoc: mxc-spdif <-> imx-spdif-dai.0 mapping ok
[    2.920274] asoc: mxc-hdmi-soc <-> imx-hdmi-soc-dai.0 mapping ok
[    2.920978] mmc1: queuing unknown CIS tuple 0x80 (50 bytes)
[    2.929611] mmc1: queuing unknown CIS tuple 0x80 (7 bytes)
[    2.931684] mmc1: queuing unknown CIS tuple 0x80 (4 bytes)
[    2.943460] ALSA device list:
[    2.946433]   #0: sgtl5000-audio
[    2.949664]   #1: imx-spdif
[    2.952461]   #2: imx-hdmi-soc
[    2.956674] TCP cubic registered
[    2.959913] NET: Registered protocol family 17
[    2.961871] mmc1: queuing unknown CIS tuple 0x02 (1 bytes)
[    2.969964] Bluetooth: RFCOMM socket layer initialized
[    2.974861] mmc1: new SDIO card at address 0001
[    2.979701] Bluetooth: RFCOMM ver 1.11
[    2.983472] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[    2.988789] Bluetooth: BNEP filters: multicast
[    2.993334] Bluetooth: HIDP (Human Interface Emulation) ver 1.2
[    2.999440] VFP support v0.3: implementor 41 architecture 3 part 30 variant 9 rev 4
[    3.007143] ThumbEE CPU extension supported.
[    3.011837] Bus freq driver module loaded
[    3.015872] Bus freq driver Enabled
[    3.019424] mxc_dvfs_core_probe
[    3.022782] DVFS driver module loaded
[    3.027121] drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
[    3.098093] mmc2: host does not support reading read-only switch. assuming write-enable.
[    3.118771] mmc2: new high speed SDHC card at address 0007
[    3.124917] mmcblk1: mmc2:0007 SD8GB 7.42 GiB
[    3.139221]  mmcblk1: p1 p2 p3 < p5 p6 p7 p8 > p4
[    3.144240] mmcblk1: p4 size 13352010 extends beyond EOD, truncated
[    3.272731] mxc_hdmi mxc_hdmi: No modes read from edid
[    3.277907] mxc_hdmi mxc_hdmi: create default modelist
[    3.283147] fbcvt: 1440x900@60: CVT Name - 1.296MA
[    3.426083] imx-ipuv3 imx-ipuv3.0: WARNING: adapt panel end blank lines
[   10.297350] EXT4-fs (mmcblk0p1): 14 orphan inodes deleted
[   10.303133] EXT4-fs (mmcblk0p1): recovery complete
[   10.432467] EXT4-fs (mmcblk0p1): mounted filesystem with ordered data mode. Opts: (null)
[   10.440693] VFS: Mounted root (ext4 filesystem) on device 179:1.
[   10.446897] Freeing init memory: 168K
 * Starting configure network device [74G[ OK ]
 * Starting Failsafe Boot Delay [74G[ OK ]
 * Starting Bridge socket events into upstart [74G[ OK ]
 * Starting RPC portmapper replacement [74G[ OK ]
 * Starting Start this job to wait until rpcbind is started or fails to start [74G[ OK ]
 * Stopping Start this job to wait until rpcbind is started or fails to start [74G[ OK ]
 * Starting configure network device security [74G[ OK ]
 * Starting configure network device [74G[ OK ]
 * Stopping Failsafe Boot Delay [74G[ OK ]
 * Starting configure network device [74G[ OK ]
 * Starting System V initialisation compatibility [74G[ OK ]

 * Starting Advanced Power Management daemon...       [80G
[74G[ OK ]
speech-dispatcher disabled; edit /etc/default/speech-dispatcher
Checking for running unattended-upgrades:
Last login: Thu Jan  1 01:00:47 CET 1970 on tty1
Welcome to Linaro 11.10 (development branch) (GNU/Linux 3.0.35+ armv7l)

 * Documentation:  https://wiki.linaro.org/

New release '12.04 LTS' available.
Run 'do-release-upgrade' to upgrade to it.

[   40.686059] brcmfmac: ERROR @brcmf_cfg80211_event_handler : event queue empty...
[   41.189431] imx-ipuv3 imx-ipuv3.0: WARNING: adapt panel end blank lines
 * Starting bluetooth       [80G root@wandboard:~#
[74G[ OK ]
  [33m* [39;49m PulseAudio configured for per-user sessions
saned disabled; edit /etc/default/saned
Setting up swapspace version 1, size = 204796 KiB
no label, UUID=d98a9fa7-4ce7-4dc6-b278-e98a332a04c9
Setting up swapspace version 1, size = 1535996 KiB
no label, UUID=52be9118-e9e1-4551-8363-5ce3a4d93d13
Device setup complete
 5 Jul 20:49:50 ntpdate[4212]: step time server 217.114.59.66 offset 1373050128.082994 sec
[   65.673369] We cannot prepare for the RX slave dma!

Am Freitag, 14. Juni 2013 18:27:35 UTC+2 schrieb Matthew W. Novak:

Hi folks.  I need some help troubleshooting a WBQUAD issue: I'm not getting any video output from the HDMI out, and I'm not sure how to proceed.  Note that I am not a developer, but I am a competent Linux user, so while a lot of things here will be over my head, I should be able to gather troubleshooting and debugging doc without much trouble.

So far, I have . . . 

1. downloaded the WBQuad Android 4.2.2 image
   
   
• I've also walked through the following steps using the wandboard-quad-ubuntu-11.10-20130604.img image with the same results
  
  
2. DD'd that image to my mSD card per the Quick Start Guide: dd if=./android-4.2.2-wand-quad-preview-20130607.img of=/dev/[my_SD_card] bs=1M
   
   
• I've DD'd the image both from Linux and OSX machines with the same results
  
  
3.  inserted that mSD card into the OS mSD card slot on the processor board, plugged in an HDMI-to-DVI cable that's attached to an HDCP-enabled monitor, and plugged in my 5V 2A 10W power supply
   
   
• I've tried this with multiple monitors and multiple HDMI-to-DVI cables with the same results
• I've tried booting with a USB keyboard attached, and/or with an OTG cable attached, with the same results
  
  
4. in every iteration above, a green light on the board comes on and nothing happens at the monitor
   
   
• to reiterate, this is the same result whether booting the Android or Ubuntu images from the Wandboard downloads page

However, at least with Android, the image appears to be booting: with an OTG cable plugged between the board and my laptop, I can do the following:

• "adb devices" and see the board listed
• "adb shell" and get into a functional Android command prompt
• "logcat" from within the adb shell and watch as the boot progresses and generates log entries

If I leave logcat running indefinitely, I eventually get kicked out of the shell and the board disappears from "adb devices".  

I've also noticed that I can only access the device via adb on the first boot after DD'ing an image to the mSD card: the board does not appear in "adb devices" after any subsequent boot of the same mSD card (reset button, reboot from within the adb shell, hard power off/power on, etc.).  

So, I appear to have a functional board with a functional Android image on the mSD card, but I'm not getting any video output via HDMI, and the instability between first and subsequent boots isn't encouraging.  Any advice on where to start?

Note that I don't have a null modem cable available, but will be able to pick one up over the weekend to help with debugging.

--Matt  

[jth...@gmail.com]

Hi All,

Am back chasing this again.  It looks like the basic trouble is the EDID read fails - clear down in the i2c-core.c::i2c_transfer() invocation from mxc_hdmi.c::mxc_edid_readblk.c.  The return code maps to -ETIMEDOUT.  Attached is the dmesg output for my boot with debug enabled within mxc_hdmi.c.  Search for mxc_edid_read to get to the interesting stuff quickly.

The hdmi controller is detected at timestamp 2.104740.
First attempt to read edid occurs at 2.269963 and fails at 2.779919.
Retry occurs at 2.779947 and is failed at 3.289994.
This drops us into default video mode processing.

wandboard.org folks - looking for some guidance as to how to approach the trouble.  How is i2c managed during boot?  Could there be contention issues?

Happy Wednesday!   Joe

[John Weber]

Joe,

Thanks for looking into this. I think HDMI EDID/DDC has been a question for
people for some time.

I don't have many monitors to test myself, but I do have a small monitor I use
to just display HDMI signals. I believe EDID should work on it, but I do need
to test a different monitor or TV with a known-good DDC channel.

What I did do is confirm that the I2C transaction does appear on the DDC SCL and
SDA signal lines. What I see on my end is the start of a single read
transaction from address 0x50. There is no ACK from the I2C slave (monitor).
An ACK should occur on the 9th clock after the start condition. I changed the
clock speed on I2C1 from 400kHz to 100k and then to 50k with no difference in
the result.

The I2C1 signals look good on both sides of the DDC level shifter. They appear
to be good to me. Fairly clean edges. The only strange part of the waveform is
a small glitch that occurs on the SDA line when the ACK should be occuring. I
see the glitch on both sides of the TXS0102 (I2C1 side and the DCC/5V side). It
is a high-going glitch that occurs just after the rising edge of SCL. It could
be that the TXS0102 is getting confused about the signal direction. I don't
think this should cause a problem with the I2C bus, though, as the gitch does
not appear on the SCL signal.

If I find a different monitor to test, I'll report the results.

John