BUG Community

Welcome! Log In

Forums BUGmodules GPS I/O Exception

Subscribe to GPS I/O Exception  26 posts, 8 voices , Tagged: java gps

Log in to reply to this topic
 
Oct 5, 2009 1:32am
Medium limnephilidae 5 posts

Hello,

I am new user and owner of a bugbase module and I’m having a little trouble with the GPS module. I am getting a Java I/O Exception in the concierge.log as show below. I couldn’t find details for this particular error in a search through the community posts. Any suggestions??

 

01/25 11:56:27:277 ERROR   FIXLEDStatusTask: Unable to query gps control on slot 1 java.io.IOException: ioctl BMI_GPS_GETSTAT failed

01/25 11:56:32:277 ERROR   FIXLEDStatusTask: Unable to query gps control on slot 1 java.io.IOException: ioctl BMI_GPS_GETSTAT failed

... continuously.

I am using RS 1.4 as far as I know with a very new BugBase wifi. I was attempting to run the latest version of the WhereAmI sample app.
uname -a:
Linux bug 2.6.27.2 #116 PREEMPT Thu Jul 16 15:38:51 EDT 2009 armv6l unknown

 

I cat’ed the device as described in a different post and I assume I am getting legitimate GPS data:

 

$GPGGA,002023.059,,,,,0,00,,,M,0.0,M,,000059

$GPGLL,,,,,002023.059,V17

$GPGSA,A,1,,,,,,,,,,,,,,,1E

$GPGSV,3,1,12,20,00,000,,10,00,000,,31,00,000,,27,00,000,7C

$GPGSV,3,2,12,19,00,000,,07,00,000,,04,00,000,,24,00,000,76

$GPGSV,3,3,12,16,00,000,,28,00,000,,26,00,000,,29,00,000,78

$GPRMC,002023.059,V,,,,,,,291006,,2C

$GPVTG,,T,,M,,N,,K4E

$GPGGA,002024.061,,,,,0,00,,,M,0.0,M,,000055

$GPGLL,,,,,002024.061,V1B

$GPGSA,A,1,,,,,,,,,,,,,,,1E

$GPGSV,3,1,12,20,00,000,,10,00,000,,31,00,000,,27,00,000,7C

$GPGSV,3,2,12,19,00,000,,07,00,000,,04,00,000,,24,00,000,76

$GPGSV,3,3,12,16,00,000,,28,00,000,,26,00,000,,29,00,000,78

$GPRMC,002024.061,V,,,,,,,291006,,20

$GPVTG,,T,,M,,N,,K4E

$GPGGA,002025.070,,,,,0,00,,,M,0.0,M,,000054

$GPGLL,,,,,002025.070,V1A

$GPGSA,A,1,,,,,,,,,,,,,,,1E

$GPGSV,3,1,12,20,00,000,,10,00,000,,31,00,000,,27,00,000,7C

$GPGSV,3,2,12,19,00,000,,07,00,000,,04,00,000,,24,00,000,76

$GPGSV,3,3,12,16,00,000,,28,00,000,,26,00,000,,29,00,000,78

$GPRMC,002025.070,V,,,,,,,291006,,21

$GPVTG,,T,,M,,N,,K4E

$GPGGA,002026.059,,,,,0,00,,,M,0.0,M,,00005C

$GPGLL,,,,,002026.059,V12

$GPGSA,A,1,,,,,,,,,,,,,,,1E

$GPGSV,3,1,12,20,00,000,,10,00,000,,31,00,000,,27,00,000,7C

$GPGSV,3,2,12,19,00,000,,07,00,000,,04,00,000,,24,00,000,76

$GPGSV,3,3,12,16,00,000,,28,00,000,,26,00,000,,29,00,000,78

$GPRMC,002026.059,V,,,,,,,291006,,29

$GPVTG,,T,,M,,N,,K4E

$GPGGA,002027.059,,,,,0,00,,,M,0.0,M,,00005D

$GPGLL,,,,,002027.059,V13

$GPGSA,A,1,,,,,,,,,,,,,,,1E

$GPGSV,3,1,12,20,00,000,,10,00,000,,31,00,000,,27,00,000,7C

$GPGSV,3,2,12,19,00,000,,07,00,000,,04,00,000,,24,00,000,76

$GPGSV,3,3,12,16,00,000,,28,00,000,,26,00,000,,29,00,000,78

$GPRMC,002027.059,V,,,,,,,291006,,28

$GPVTG,,T,,M,,N,,K4E

$GPGGA,002028.071,,,,,0,00,,,M,0.0,M,,000058

$GPGLL,,,,,002028.071,V16

$GPGSA,A,1,,,,,,,,,,,,,,,1E

$GPGSV,3,1,12,20,00,000,,10,00,000,,31,00,000,,27,00,000,7C

 

Oct 5, 2009 2:02pm
Medium Bug Labs team jconnolly 285 posts

Hi limnephilidae,

I’ve only seen that happen when either concierge is loaded twice (and two processes are trying to read/write to the gps control device node) or if the driver was loaded incorrectly, possibly from a poor connection resulting from a slow or unfirm insertion of the module.

Some things to look at:

ls -haltr /dev/bmi_gps_control_m* crw-rw---- 1 root root 246, 2 Jan 11 01:22 /dev/bmi_gps_control_m3 


This shows that the device node used for reading/writing control operations from the GPS module exists at the OS level.

 

rmmod bmi_gps modprobe bmi_gps

 

 

This will reload the driver.

cat /var/log/concierge.log 

Paste this into a response and it may help me determine what’s getting messed up.

Oct 5, 2009 7:42pm
User_img_not_found_04_med blprnt 19 posts

As jconnolly knows (thanks for the IRC help!) I’m having the same problem.

My concierge.log results look like this:

Thu Feb 19 08:10:56 GMT 1970 FIXLEDStatusTask: Unable to query gps control on slot 3
Exception: ioctl BMI_GPS_GETSTAT failed
Thu Feb 19 08:11:02 GMT 1970 FIXLEDStatusTask: Unable to query gps control on slot 3
Exception: ioctl BMI_GPS_GETSTAT failed

And the var/log/messages output:

Feb 19 08:11:36 bug user.err kernel: [ 511.590000] ReadByte_ACC() – i2c_transfer() failed.
Feb 19 08:11:36 bug user.debug kernel: [ 511.740000] Bus Arbitration lost
Feb 19 08:11:36 bug user.debug kernel: [ 511.740000] Could not grab Bus ownership
Feb 19 08:11:36 bug user.debug kernel: [ 512.010000] Data not transmitted
Feb 19 08:11:36 bug user.debug kernel: [ 512.010000] I2C Status: Addr: 0xfc08000c Data: 0×81
Feb 19 08:11:36 bug user.err kernel: [ 512.010000] pca9546_port_i2c_xfer() – pca9546_select_channel() failed.

The flow of errors in the concierge.log doesn’t seem to start until a few minutes after the module is plugged in…

 

 

Oct 6, 2009 11:21am
User_img_not_found_06_med williamh 4 posts

I noticed a similar i2c error when I upgraded to the 1.4.1 and 1.4.2 file system. I did not notice it with the 1.4 that came with my bug bundle purchased in June.

[ 3823.850000] Bus Arbitration lost
[ 3824.010000] Data not transmitted
[ 3824.010000] I2C Status: Addr: 0xfc08000c Data: 0xe0
[ 3824.010000] Bus busy
[ 3824.010000] ReadByte_ACC() – i2c_transfer() failed.

No concierge errors/exceptions. Just running the GPSLogger app. The GPS appears to be working based on cat /dev/ttymxc2 and GPSLogger log messages to concierge.log.

The LCD display goes grey.

Oct 6, 2009 3:07pm
Medium Bug Labs team bcruskie 45 posts

We still haven’t been able reproduce this issue, however, we are still trying. I have some questions

  • Other than the GPS/LCD what other modules are attached?
  • Are you guys doing module remove/inserts before seeing this issue?
  • Is it intermittent? does a reboot fix the condition?

Thanks in advance!

Brian

 

 

 

Oct 6, 2009 3:15pm
User_img_not_found_04_med blprnt 19 posts

Hi Brian,

Thanks for your attention to this.

I get the error with any combination of modules. Right now it’s coming with just the LCD and GPS modules.

Detaching and reattaching modules doesn’t seem to make a difference. 

Nothing seems to fix the condition – reboot or no reboot, whenever the GPS module is inserted I get this error and obviously no GPS functionality.

-Jer

Oct 6, 2009 3:40pm
User_img_not_found_04_med blprnt 19 posts

Hi Brian,

Not sure if this will help but I followed the instructions from this post:

http://community.buglabs.net/forums/8-BUGmodule…

Here’s a log of the GPS feed:

http://pastie.org/644299

-Jer

 

Oct 6, 2009 3:55pm
Medium limnephilidae 5 posts

We still haven’t been able reproduce this issue, however, we are still trying. I have some questions

  • Other than the GPS/LCD what other modules are attached?
  • Are you guys doing module remove/inserts before seeing this issue?
  • Is it intermittent? does a reboot fix the condition?

Thanks in advance!

Brian

 

I was getting inconsistent error conditions when I booted with the GPS module, when I inserted the module after a clean boot, and when I switched the module to a different slot. Tonight I will recreate the issue in all scenarios. The only thing that is consistent for me is the error I posted in my original post and that the GPS has not worked with concierge. 

- Adam

Oct 6, 2009 4:31pm
User_img_not_found_06_med williamh 4 posts

I am running the standard bug bundle – bugbase, gps, lcd, motion, vonhippel as of June 09. No module manipulation on my part. The problem occurs infrequently. It may occur within 1/2 hour of up time or it may go several hours. But it always occurs. A reboot fixes the issue and appears to be the only way to restore the LCD display. I’ve tried restarting concierge with no effect. I’m not familiar enough with the kernel modules to try and unload/reload specific modules to try and isolate the issue.

If you want to provide a sequence of tasks to help narrow it down, let me know.

I was on the path of unplugging each physical module one at a time to see if/when the problem disappeared, but my real job interferred and I haven’t finished this process. I see blprnt did this. I did re-seat each module. I may go back and run the R1.4 rootfs and verify my claim that this started for me with R1.4.1.

Other info. I have a serial connection so I can watch boot sequences. I have a USB/ethernet attached to the von hippel which shows up as eth1. I’m currently running the R1.4.2 rootfs, plus I did an ipkg upgrade.

Bill

Oct 7, 2009 12:30pm
User_img_not_found_06_med williamh 4 posts

Quick followup on i2c issue. It is independent of rootfs version, as I backed off to R1.4 and had the same message. For the LCD, I can recover by reloading the module:

/etc/init.d/concierge stop

rmmod bmi_lcd_core

modprobe bmi_lcd_core

/etc/init.d/concierge start

Oct 7, 2009 6:13pm
Medium Bug Labs team bcruskie 45 posts

Great info guys, Thanks again!

Just a quick update. I have 4 BUGs set up trying to reproduce this issue and I haven’t had any luck up to this point. I will let you know when I make some progress.

-Brian

Oct 8, 2009 1:53am
Medium limnephilidae 5 posts

In all cases of booting up without the GPS and then adding the module it fails. If I boot with the device, it doesn’t show up in the lcd display and is not acknowledged by Concierge as a valid service. When moving the device or adding the device after a boot, I get the same variation on the error below.

messages:

Jan 28 12:20:51 bug local0.debug udhcpc3882: Sending discover…

Jan 28 12:21:01 bug user.err kernel: [  103.210000] Board Inserted at conn-m3

Jan 28 12:21:01 bug user.err kernel: [  103.760000] bmi_device_work_handler(2): device_attach failed.

Jan 28 12:21:02 bug user.info kernel: [  104.090000] bmi_gps.c: probe slot 2

Jan 28 12:21:02 bug user.debug kernel: [  104.090000] ACK not received

Jan 28 12:21:02 bug user.err kernel: [  104.090000] WriteByte_IOX() – i2c_transfer() failed.

Jan 28 12:21:02 bug user.warn kernel: [  104.100000] bmi_gps.c: BMI_GPS Driver v1.1

concierge.log:
————————————————————————————
  Framework restarted in 20.298 seconds.
————————————————————————————
Wed Jan 28 12:18:04 GMT 1970 Received new INSERT model change event for module Motion
Wed Jan 28 12:18:04 GMT 1970 Started modlet from factory com.buglabs.bug.module.motion…
Wed Jan 28 12:21:04 GMT 1970 GPSModlet setting active (external) antenna
Wed Jan 28 12:21:04 GMT 1970 Failed to set GPS antenna to active (external) antenna
        Exception: ioctl BMI_GPS_GLEDON failed
Wed Jan 28 12:21:04 GMT 1970 Received new INSERT model change event for module GPS
Wed Jan 28 12:21:04 GMT 1970 Unable to access GPS antenna state.
        Exception: ioctl BMI_GPS_GETSTAT failed
Wed Jan 28 12:21:04 GMT 1970 Unable to access GPS antenna state.
        Exception: ioctl BMI_GPS_GETSTAT failed

 

After booting with the GPS device attached, I did not see the GPS log in the base lcd display. I attempted to remove the bmi_gps kernel module after moving it around but apparently the device was blocked:

root@bug:/var/volatile/log# rmmod bmi_gps

ERROR: Module bmi_gps is in use

A few questions…
1. Any possibility I have a defective GPS unit? (I’m having trouble believing this, it smells like kernel driver woes)
2. How can I tell exactly what RS version I have?
3. Any more guesses on resolution other than reloading the driver? In the cases where I could actually reload the driver it doesn’t help resolve the issue.
4. Does the device attach fail message in the kernel really indicate an error? I get that with the motion though it seems to attach fine (note I haven’t tested this with an app yet)
There are a lot of problems for a device that is supposed to get me prototyping quickly. I’m hoping this has one root cause.
- Adam

Oct 8, 2009 2:22am
Medium limnephilidae 5 posts

One more thing to add:

I tested the motion module and it works perfectly. The LCD works great as well. I think that confirms my only issue is the GPS module. 

Also, even though my motion module functions properly for all apps I tested with, I still get that goofy error from the kernel which (without checking src code), I am assuming is not a major issue or indicative of the root cause of the GPS problem.

Jan 28 12:46:23 bug user.err kernel: [42949559.830000] Board Inserted at conn-m1

Jan 28 12:46:24 bug user.err kernel: [42949560.380000] bmi_device_work_handler(0                                                                             ): device_attach failed.

Jan 28 12:46:24 bug user.warn kernel: [42949560.640000] BMI MDACC Driver v1.0

Motion module still works despite the error above. The error does not recur after moving the module to different slots. 

Oct 12, 2009 10:34am
Medium Bug Labs team bcruskie 45 posts

Hey Adam,

I’m stumped on this one. I’ve been trying to reproduce for the past week with no luck. There could be a possiblitly that we had a bad run of GPS modules but it certainly seems driver related.

Here is the info on your finding your Rootfs version: http://bugcommunity.com/wiki/index.php/Determin…

As far as your other questions I will see if Ken Gilmer, our Dev lead, has any suggestions.

Thanks,

Brian

Oct 12, 2009 11:02am
Medium Bug Labs team bcruskie 45 posts

Hey Guys,

Just discussed this issue with the Dev team and we’d like to get these units and GPS modules back in house to look at the issue.

Please send an email to support@buglabs.net requesting an RMA for the BUGbase and GPS module with the following information:

  • Order number
  • URL to this topic

They will set you up with a Return fedex shipping label and get a new Base and GPS module out to you.

Sorry for the inconvienence and thank you for your patience.

-Brian

 

Oct 13, 2009 3:00pm
User_img_not_found_04_med blprnt 19 posts

Hi Guys,

Sorry for not checking into the thread for a while Long weekend.

I have two bases and two GPS modules – I’m just in the process of testing all 4 combinations to try to figure out what is the limiting factor here. 

I will let you know what happens. Worse comes to worse we can send both bases and both GPS units back, but we are heavy in the middle of dev and would rather not lose these units.

-Jer

Oct 13, 2009 3:32pm
User_img_not_found_04_med blprnt 19 posts

All combinations of base + GPS module resulted in the same error. We have two bases, one of which has had the file system updated and one which hasn’t. 

The error seems to be fairly inconsistent – it always happens but sometimes it takes a long time (10 minutes or so) to start reporting in the concierge log.

I guess we’ll have to send these things back. We’re having an open house tomorrow which we’ll need to have them around for – we’ll try to ship them back on Thursday.

-Jer

Oct 13, 2009 3:38pm
Medium Bug Labs team bcruskie 45 posts

Great, please let us know your results Jer. We can always ship you the replacements and have you send us back the hardware after receive them.

Brian

Oct 17, 2009 1:36am
Medium limnephilidae 5 posts

I will contact support.

As an FYI for anyone else who reads this thread, my bug version is 1.4.2 according to the date:

 

root@bug:~# cat /etc/buildinfo

BUG Linux Build Information

Thu Jul 16 16:27:15 EDT 2009

Build Image: bug-image-production-R1.4

Linux darner 2.6.24-23-server #1 SMP Mon Jan 26 01:36:05 UTC 2009 x86_64 GNU/Linux

Oct 19, 2009 5:09pm
Medium Bug Labs team bcruskie 45 posts

Hey Guys,

We finally reproduced this error in house. I think we can hold off on RMA’ing your units while we continue to look at the issue.

Thanks for everything and sorry for taking your time up on contacting support.

I will keep you updated,

Brian

 

Oct 19, 2009 5:25pm
Medium Bug Labs team agibb 8 posts

Hi,

I reproduced the error, and then I killed java and restarted it for a fix. Not sure if this will fix it for everyone here, after fixing it I couldn’t reproduce the GPS error again. Here are the steps I took:

1. Unplugged GPS

2. in your terminal, as root type the following after the hashtag (root@bug:#): 

root@bug:# killall -9 java-cdc

root@bug:#sync
root@bug:
# rm -fr /usr/share/java/storage
root@bug:# sync
root@bug:
# /etc/init.d/concierge restart && tail -f /var/log/concierge.log

Then I plugged the module back in and it was still complainnig, so I killed it rebooted by typing this in the terminal:


root@bug:# rm -fr /usr/share/java/storage
root@bug:
# sync
root@bug:# killall -9 java-cdc
root@bug:
# reboot

Once it was rebooted, I plugged the gps back in. If any steps are confusing, just email me at alicia@buglabs.net

Let us know if that worked!

Happy Hacking,

Alicia

 

Oct 20, 2009 1:32am
User_img_not_found_04_med blprnt 19 posts

Hi Alicia,

This didn’t work for me… 

I’ve tried lots of combinations of reboot, plug-in module, restart concierge, clear java storage, and nothing seems to work.

Oddly, earlier today my GPS module did work for the first time in about 100 tries, though I hadn’t done anything differently than I had previously (at least I don’t think I had). The little green light started blinking, and I actually got lat/lon data from the location service.

Oct 20, 2009 10:31am
User_img_not_found_06_med williamh 4 posts

Uh oh. I’ve already sent my bugbase back, but I am using the opportunity to upgrade to the WiFi bugbase.

Nov 2, 2009 2:27pm
Medium Bug Labs team bcruskie 45 posts

Adam/limnephilidae,

Did you end up contacting support yet? If not I believe we just need to get you a new GPS module.

Thanks,

Brian

 

Nov 2, 2009 9:48pm
Medium Bug Labs team mcholerton 54 posts

You can email your order number or full name on your invoice to support[at]buglabs.net and I’ll make sure we get you a new module right away.  Sorry for the trouble – I’ll be on the look out! 

Log in to reply to this topic
Forums BUGmodules GPS I/O Exception

Powered by Community Engine

Top