Bug 1823

Summary: [Bealgeboard C4] keyboard could not work with 20111207 build
Product: [Build System, Metadata & Runtime] BSPs Reporter: Yi Zhao <yi.zhao>
Component: bsps-configurationAssignee: Paul Eggleton <bluelightning>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: Medium CC: bruce.ashfield, dvhart, jessica.zhang, jingdong.lu, liang.li2, liming.wang, ning.zhang, richard.purdie, ross.burton, sgw, wahlm, yi.zhao, yp.bsp.watcher, yp.watcher
Version: 1.3.2   
Target Milestone: 1.4.1   
Hardware: Beagleboard   
OS: arm   
Whiteboard: Needs to be reproduced
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: ---
Attachments:
Description Flags
dmesg
none
Xorg.0.log none

Description Yi Zhao 2011-12-12 03:02:57 UTC 
Created attachment 293 [details]
dmesg

Tree/branch: poky/1.2_M1
Commit: b9d3a5224c7f76f273d5919a88b7460d35ad2574
Image location: http://autobuilder.pokylinux.org/pub/nightly/20111207-2/machines/beagleboard/

Steps:
1. Connect a usb hub to the EHCI port on the board. Then connect a keyboard and a mice to the hub.  

2. Boot up the board with u-boot cmd:
   setenv bootcmd 'mmc rescan 0; mmc init; fatload mmc 0:1 0x80300000 uImage; bootm 0x80300000' 
   setenv bootargs 'console=tty0 console=ttyO2,115200n8 root=/dev/mmcblk0p2 rootwait rootfstype=ext3 ro' 
   boot

3. When the X startup, the mice can work but keyboard does not work. For example, I can click the "terminal" icon on the desktop to launch it. But can not input anything in this terminal by using keyboard.

This issue doesn't happen on 1.1

You can check the dmesg and Xorg log for more details.
Comment 1 Yi Zhao 2011-12-12 03:03:46 UTC 
Created attachment 294 [details]
Xorg.0.log
Comment 2 Yi Zhao 2011-12-12 03:17:26 UTC 
After the X startup, if I re-plug the keybarod, I can get the following message from serial port:
######### 
root@beagleboard:~# /srv/home/pokybuild/yocto-autobuilder/yocto-slave/nightly-arm/build/build/tmp/work/beagleboard-poky-linux-gnueabi/linux-yocto-2.6.37+git1+aeea99683c7283f1f3320bf2ee7085ee252d4e7e_1+3ddb21
/srv/home/pokybuild/yocto-autobuilder/yocto-slave/nightly-arm/build/build/tmp/work/beagleboard-poky-linux-gnueabi/linux-yocto-2.6.37+git1+aeea99683c7283f1f3320bf2ee7085ee252d4e7e_1+3ddb22772862a8223640fa9751
usb 1-2.6: USB disconnect, address 7

root@beagleboard:~# 
root@beagleboard:~# usb 1-2.6: new low speed USB device using ehci-omap and address 8
input: Dell Dell USB Keyboard as /devices/platform/ehci-omap.0/usb1/1-2/1-2.6/1-2.6:1.0/input/input5
generic-usb 0003:413C:2105.0006: input: USB HID v1.10 Keyboard [Dell Dell USB Keyboard] on usb-ehci-omap.0-2.6/input0

root@beagleboard:~#
############# 

But the keyboard still doesn't work.
Comment 3 Liming Wang 2011-12-13 00:00:55 UTC 
If we change "console=tty0 console=ttyO2" to "console=ttyO2 console=tty0", the keyboard will work well. I think it's not a real bug, but we can have a look how it occurs.
Comment 4 Bruce Ashfield 2012-03-15 12:43:30 UTC 
I don't currently have a C4 available for testing, so moving this to 1.2.1
Comment 5 Darren Hart 2012-05-01 15:21:53 UTC 
It's time we decide what to do with beagleboard issues. I suggest we order 4 of each model we intend to support and make them available to QA, and engineering. Otherwise we might as well close these bugs as WILLNOTFIX.
Comment 6 Bruce Ashfield 2012-05-01 16:17:38 UTC 
Agreed. We need to list a finite set of beagleboards, and make sure we
have them available (and remotely available). At the moment, Wind River
does have more on order, so the xM will be available.

That being said, for 1.3, we'll refresh our ARM reference platform, if
we can get a suitable replacement for the beagleboard, that would solve
the issue as well.
Comment 7 Bruce Ashfield 2012-05-03 12:55:04 UTC 
I put two days for this, but I never did see a comment on our workaround
that was posed by Liming:

  If we change "console=tty0 console=ttyO2" to "console=ttyO2 console=tty0",
  the keyboard will work well. I think it's not a real bug, but we can have a 
  look how it occurs.

Can the submitter try this and let us know ?
Comment 8 Liming Wang 2012-05-03 13:21:12 UTC 
I remember the submitter had taken my sugguestion to boot the board.
Comment 9 Yi Zhao 2012-05-07 05:39:26 UTC 
(In reply to comment #8)
> I remember the submitter had taken my sugguestion to boot the board.

Yes, I have tried it and the keyboard works.
Comment 10 Scott Garman 2012-06-14 21:21:13 UTC 
I have given Scott Rifenbark some documentation on this to include the in Known Issues section of our release notes.
Comment 11 Mark Wahl 2013-03-09 14:44:21 UTC 
I built a beagleboard sato image using the poky-danny-8.0 download archive. Using it on my beagleboard c4 the usb keyboard does not work on X, but the mouse did.
I tried the suggested workaround with swapping the console statement, but it still did not work.

I checked the Xorg.0.log and found:
[1267749.877] (II) config/udev: Adding input device   USB Keyboard (/dev/input/e
vent0)                                                                          
[1267749.877] (II) AutoAddDevices is off - not adding device.                   

After commenting out the line

Option         "AutoAddDevices" "False"

in /etc/X11/xorg.conf the keyboard is recognized and works without problems.
Comment 12 Saul Wold 2013-05-01 16:04:11 UTC 
Can this be reproduced?
Comment 13 Saul Wold 2013-05-01 16:06:05 UTC 
Richard thinks this might be boot loader related.
Comment 14 Ning Zhang 2013-05-02 08:14:47 UTC 
I can not reproduce it on poky-danny branch.
The last poky commit is: 2806646a263527ec0487ea160afd4bdc0a3c1703

I bitbake a core-image-sato image, and use it to bootup the Bealgeboard C4 target. 
Without any bootline changes, the target could bootup, and mouse/keyboard work well.
And when use below bootline, the target could bootup, and mouse/keyboard work well too.
    setenv bootcmd 'mmc rescan 0; fatload mmc 0:1 0x80300000 uImage; bootm 0x80300000' 
    setenv bootargs 'console=ttyO2,115200n8 console=tty0 root=/dev/mmcblk0p2 rootwait rootfstype=ext3 ro' 
    boot
Comment 15 Ross Burton 2013-05-29 10:51:02 UTC 
I'm pretty certain we can drop AutoAddDevices=False from all BSPs.  Can anyone remember why it was added in the first place?  With fast booting and USB probing taking time, it's entirely possible for the X server to start before input devices have been enumerated.