Bug 514348

Summary: hard disk device node lost for hdd inside eSATA Tevion case.
Product: [openSUSE] openSUSE 11.1 Reporter: Elmar Stellnberger <estellnb>
Component: BasesystemAssignee: E-mail List <bnc-team-screening>
Status: VERIFIED NORESPONSE QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P3 - Medium CC: jeffm
Version: Final   
Target Milestone: ---   
Hardware: x86-64   
OS: openSUSE 11.1   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: barehdd.lshal
preplug.lshal
postplug.lshal
postplug2.lshal
dmesg
lspci -nn

Description Elmar Stellnberger 2009-06-18 12:09:21 UTC
User-Agent:       Mozilla/5.0 (compatible; Konqueror/3.5; Linux) KHTML/3.5.10 (like Gecko) SUSE

  If I dismount the Samsung HDD out of the Tevion case and plug it in with an eSATA to SATA cable everything works fine (barehdd.lshal).
  However as soon as I put it into the Tevion case no /dev/sdc* nodes can be found. Interestingly hal recognizes the hdd well and even thinks that the required device nodes are still in place. It remains obscure what kind of conjuncture makes the /dev- and /sys/../sdc* disappear since they seem to be in place initially (preplug.lshal, postplug*.lshal).

Reproducible: Always
Comment 1 Elmar Stellnberger 2009-06-18 12:33:14 UTC
Created attachment 298978 [details]
barehdd.lshal
Comment 2 Elmar Stellnberger 2009-06-18 12:42:19 UTC
Created attachment 298980 [details]
preplug.lshal
Comment 3 Elmar Stellnberger 2009-06-18 13:07:00 UTC
Created attachment 298988 [details]
postplug.lshal
Comment 4 Elmar Stellnberger 2009-06-18 13:08:04 UTC
Created attachment 298989 [details]
postplug2.lshal
Comment 5 Elmar Stellnberger 2009-06-18 13:11:32 UTC
Unfortunately /var/log/messages does not come up with any output about it.
Comment 6 Elmar Stellnberger 2009-09-04 19:26:42 UTC
Seems having worsened under Opensuse 11.2, Milestone 6. 
lshal -lu /org/freedesktop/Hal/devices/storage_serial_SATA_SAMSUNG_HD753LJS13UJ1KQ301303
does not give any output any more.

The drive works under Windows.
What about Novells` hardware initiative?
Comment 7 Tejun Heo 2009-11-12 16:42:40 UTC
What is a tevion case?  Is it an eSATA cage?
Comment 8 Elmar Stellnberger 2009-11-12 17:54:17 UTC
Yes, Tevion is a synonyme for the manufactorer of the eSATA case.
Comment 9 Elmar Stellnberger 2009-11-12 17:54:50 UTC
.
Comment 10 Tejun Heo 2009-11-12 19:52:38 UTC
Can you please attach the output of "dmesg" after connecting the esata device?  Also, please attach the output of "lspci -nn".

Thanks.
Comment 11 Elmar Stellnberger 2009-11-16 10:15:13 UTC
Created attachment 327660 [details]
dmesg

no messages in dmesg; dmesg is the same before and after plugging.
Comment 12 Elmar Stellnberger 2009-11-16 10:17:28 UTC
Created attachment 327662 [details]
lspci -nn

Sorry for the late reply; I had to do a complete system reinstallation.
Comment 13 Tejun Heo 2009-11-17 05:58:22 UTC
That means that the controller is not seeing the device being attached at all - ie. no communication link is being established between the eSATA port on your computer and the device.  I suggest checking the hardware.

Thanks.
Comment 14 Elmar Stellnberger 2009-11-17 10:15:10 UTC
Strange; why it could create the hal device nodes before. Perhaps I should retest with os11.1.
Comment 15 Elmar Stellnberger 2009-11-17 10:16:31 UTC
.
Comment 16 Tejun Heo 2009-11-17 13:43:04 UTC
Why do you think it created hal nodes?  Which node are you referring to in which file?  That just cannot happen.
Comment 17 Elmar Stellnberger 2009-11-17 16:12:45 UTC
see for the attachements: *.lshal (os11.1)
Comment 18 Tejun Heo 2009-11-17 17:15:15 UTC
Yeah, I looked at them and diff'd them.  I don't know how you did the testing but it looks like the device node for bare hadd was hanging around till you did preplug lshal and later went away before you did post lshals (ATA drive unplugging usually takes from 15 to 20 secs).  If the kernel isn't noticing it, there simply is no way hal would know anything about it.

Does booting with "libata.force=1.5Gbps" make any difference?  Can you please boot with the parameter, hotplug the eSATA device and then attach full output of dmesg?
Comment 19 Elmar Stellnberger 2009-11-24 11:37:59 UTC
  Unfortunately there is still no dmesg-output when I boot with libata.force=1.5Gbps with 2.6.31.6-0.0.0.23.273000e-desktop. I wonder why it had recognized the presence of the hdd with os 11.1.
Comment 20 Tejun Heo 2009-11-24 11:54:28 UTC
I'm getting confused.  Didn't you say the first lshal output was taken by connecting the drive directly to the port without going through the eSATA harness?  Does the drive appear if you connect it directly to the port under 11.2?
Comment 21 Elmar Stellnberger 2009-11-24 12:10:19 UTC
 Yes, by connecting the HDD directly (without the eSATA case) it still works well with 11.2.
Comment 22 Tejun Heo 2009-11-24 12:19:51 UTC
I can't think of anything which can make 11.1 and 11.2 behave differently like that.  At that level, it's basically all hardware and the driver is simply listening for events.  Link powersave mode can interfere with hotplugging but if it were link powersave issue, direct connection should have failed the same way.  Can you please boot 11.1 live media and attach the drive via the eSATA harness and attach the log?

Thanks.
Comment 24 Wei Wang 2011-01-13 09:07:23 UTC
Hi ,all 
I close it because no response ,please reopen it if you can provide more
information .
Thanks ,
wei