Bug 6455

Summary: [beaglebone] USB devices don't work in second boot
Product: [Build System, Metadata & Runtime] BSPs Reporter: Yi Zhao <yi.zhao>
Component: bsps-meta-yoctoAssignee: Denys Dmytriyenko <denis>
Status: VERIFIED FIXED QA Contact:
Severity: normal    
Priority: Medium CC: alexandru.c.georgescu, bluelightning, bruce.ashfield, denis, ke.zou, Qi.Chen, sgw, yi.zhao
Version: 1.6.1   
Target Milestone: 1.6.2   
Hardware: BeagleBone   
OS: arm   
Whiteboard:
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: No (bug/feature does not impact docs)
Attachments:
Description Flags
dmesg-first-boot
none
dmesg-second-boot
none
/var/log/messages-first-boot
none
/varlog/messages-second-boot
none
lsmod-first-boot
none
lsmod-second-boot none

Description Yi Zhao 2014-06-17 10:29:02 UTC
Image location: http://autobuilder.yoctoproject.org/pub/releases/yocto-1.6.1/machines/beaglebone/

Git rev: daisy/a43dba8c2904f9c1ce0425c53c5a7f4718121e6b

The USB devices (mouse, keyboard, USB stick) are only found during the first boot. After reboot, they don't work. Seems the system doesn't found them. Please see the attached logs.
Comment 1 Yi Zhao 2014-06-17 10:29:53 UTC
Created attachment 2002 [details]
dmesg-first-boot
Comment 2 Yi Zhao 2014-06-17 10:30:25 UTC
Created attachment 2003 [details]
dmesg-second-boot
Comment 3 Yi Zhao 2014-06-17 10:30:54 UTC
Created attachment 2004 [details]
/var/log/messages-first-boot
Comment 4 Yi Zhao 2014-06-17 10:31:22 UTC
Created attachment 2005 [details]
/varlog/messages-second-boot
Comment 5 Yi Zhao 2014-06-17 10:31:49 UTC
Created attachment 2006 [details]
lsmod-first-boot
Comment 6 Yi Zhao 2014-06-17 10:32:18 UTC
Created attachment 2007 [details]
lsmod-second-boot
Comment 7 Yi Zhao 2014-06-17 10:36:57 UTC
When remove the file /etc/udev/cache.data and reboot, the USB devices can be found and work again.
Comment 8 Paul Eggleton 2014-06-19 14:51:36 UTC
I could be wrong but this kind of sounds like a generic udev cache problem rather than being something device-specific. Do we see this with other machines?
Comment 9 Denys Dmytriyenko 2014-06-19 15:35:54 UTC
Yes, it is udev cache related. But I believe it used to handle USB devices properly before... Anyway, I've been looking into it, trying to make it work w/o disabling udev cache completely.
Comment 12 Denys Dmytriyenko 2014-07-03 14:43:30 UTC
The fix is in OE-Core master
Comment 13 Yi Zhao 2014-10-29 06:45:30 UTC
Verified with 1.6.2 rc1
git rev: daisy/313864bf528696a72c3ee6b3dec1a0c7b0b223f6