Problems unmounting/fssyncking extern UFS filesystem

O. Hartmann ohartman at mail.zedat.fu-berlin.de
Sun Nov 26 08:07:43 PST 2006


Roland Smith wrote:
> On Sat, Nov 25, 2006 at 05:08:03PM +0100, O. Hartmann wrote:
>   
>> A while ago since now I receive kernel messages like this in FreeBSD
>> 6.2-PRE/AMD64:
>>
>> fsync: giving up on dirty
>> 0xffffff000362c7c0: tag devfs, type VCHR
>>     usecount 1, writecount 0, refcount 325 mountedhere 0xffffff00504d8400
>>     flags ()
>>     v_object 0xffffff00013c80e0 ref 0 pages 1286
>>      lock type devfs: EXCL (count 1) by thread 0xffffff0050287260 (pid
>> 14109)
>>         dev ufs/BACKUP
>>
>> Filesystem is an external USB attached SATA HD, ohci() driven (due to
>> ehci() is not working stable and properly on FreeBSD 6.2).
>>     
>
> The external USB harddisk I'm using works fine with ehci (VIA VT6202 USB
> 2.0 controller) on 6.2-PRERELEASE amd64: 
>
> Controller /dev/usb4:
> addr 1: high speed, self powered, config 1, EHCI root hub(0x0000), VIA(0x0000), rev 1.00
>  port 1 powered
>  port 2 addr 2: high speed, power 100 mA, config 1, Mass Storage Device(0x3507), Prolific Technology Inc.(0x067b), rev 1.00
>   

My mainboard is a ASUS A8N32-SLI Deluxe and therfore the nForce400/16x
SPP/MCP controller chipset and EHCI isn't working, crahsing the box
sporadically. This happens with the mentioned external drive and with an
USB stcik also.

Crashing the box disappears when using only ohci.

The above shown fsync problems occurs in both case, ohci() and ehci().

>   
>> Filesystem is mounted via amd() and there via the'script' option in
>> amd() due to problems of the amd() mounting process recognizing UFS
>> filesystems. After 30 seconds of inactivity the filesystems gets
>> dismounted. This worked quite well in the past, but now I see this
>> kernel error messages.
>>     
>
> The only problems I ever had wer with the firewire interface, not
> USB. But I don't use amd, and I'm using GEOM_ELI encyption.
>
> If amd doesn't work well with ufs, would using glabel be a workaround?
>
> Roland
>   

Maybe amd() dismounts to early ... Don't know. Maybe the magic
'sync;sync;sync' before dismounting will help, I'll try it. Sorry, I got
this 'idea' right now, I may have bothered you all with possibly crap.

Regards,
Oliver


More information about the freebsd-stable mailing list