[yocto] Yocto 1.8.1 - Insert USB VFAT thumb drive results in "Volume was not properly unmounted. Some data may be corrupt. Please run fsck"

[Gary Thomas]

On 2016-08-01 15:06, Brian Hutchinson wrote:
> On Mon, Aug 1, 2016 at 8:56 AM, Brian Hutchinson <b.hutchman at gmail.com> wrote:
>> If the USB drive is cold plugged (inserted before power on) it mounts
>> fine, it is only a hotplug that results in the message "Volume was not
>> properly unmounted. Some data may be corrupt. Please run fsck"
>>
>> The drive is mounted under /run/media/sdb1 but a listing of the
>> directory just shows strange characters for the file names.
>>
>> Installed dosfstools on target and running fsck.vfat doesn't report a
>> problem.  Inserting the exact same USB drive in a Ubuntu 14.04 box
>> works fine so it isn't the USB drive.
>>
>> Found this thread talking about the problem I'm seeing (link below)
>> and a bug report that sounds similar to what I'm seeing but I never
>> really saw what the issue is and it says it's been fixed in Yocto
>> 1.8.3 but I don't see a Yocto 1.8.3 release ... only Yocto 1.8.2.
>>
>> https://lists.yoctoproject.org/pipermail/yocto/2015-September/026563.html
>>
>> https://bugzilla.yoctoproject.org/show_bug.cgi?id=9652/
>>
>> When I run the mount command it says the USB drive is mounted as:
>>
>> type vfat (rw,relatime,gid=6,fmask=0007,dmask=0007,allow_utime=0020,codepage=437,iocharset=iso8859-1,shortname=mixed,errors=remount-ro)
>>
>> I'm running a 4.4.1 kernel built outside of Yocto on a TI 816x SoC.
>>
>> Anyone have any information on what the issue may be?  I've played
>> around with my kernel .config thinking it was something there but that
>> didn't make any difference.  Kinda stumped at the moment.
>>
>> Regards,
>>
>> Brian
>
> This is the output of running fsck.vfat in case anyone cares.  It
> doesn't report anything about a "dirty bit" being set.  Could it be
> this version of dosfsck is too old?
>
> dosfsck 2.11, 12 Mar 2005, FAT32, LFN
> Starting check/repair pass.
> Starting verification pass.
> /dev/sdb1: 284 files, 60465/62894 clusters
>

BTW, this behavior still exists in Poky/Yocto 2.1+  I was thinking
of filing a bug myself (I tried to look into the details, but I
could not figure out why the 'fsck' step didn't make the kernel
happy)

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------